Paperid:1
Authors:Akhil Premkumar
Title: Neural Entropy
Abstract:
We explore the connection between deep learning and information theory through the paradigm of diffusion models. A diffusion model converts noise into structured data by reinstating, imperfectly, information that is erased when data was diffused to noise. This information is stored in a neural network during training. We quantify this information by introducing a measure called \textit{neural entropy}, which is related to the total entropy produced by diffusion. Neural entropy is a function of not just the data distribution, but also the diffusive process itself. Measurements of neural entropy on a few simple image diffusion models reveal that they are extremely efficient at compressing large ensembles of structured data.



Paperid:2
Authors:James Cheshire, Yann Issartel
Abstract:
Abstract:We consider the problem of active seriation, in which the learner must order $n$ items from $T$ sequentially observed pairwise similarities. Items are ordered according to their pairwise similarities: two items sharing a high similarity are expected to be close in the ordering, and vise-versa. At each round, the learner selects a pair $i,j$ and receives a noisy observation of the similarity between $i$ and $j$. After $T$ rounds, the learner returns an ordering of the $n$ items, whose accuracy is measured by the probability of error of not recovering the true ordering. We characterize the minimax rates of this problem when a minimal gap $\Delta$ separates pairwise similarities, and we provide a polynomial time algorithm achieving this rate, adaptively to the unknown parameter $\Delta$.



Paperid:3
Authors:Satoshi Noguchi, Yoshinobu Kawahara
Abstract:
Transformers have achieved remarkable success across natural language processing (NLP) and computer vision (CV). However, deep transformer models often suffer from an over-smoothing issue, in which token representations converge to similar values as they pass through successive transformer blocks. In this paper, we establish an equivalence between the hidden-state dynamics induced by stacked attention layers and graph neural diffusion on a complete graph. From this perspective, over-smoothing can be interpreted as a consequence of the dissipative nature of the underlying diffusion dynamics. Motivated by this physical interpretation, we propose Wavy Transformer, which consists of a novel attention layer based on second-order wavy dynamics. We also introduce a feed-forward network and a normalization layer designed to preserve the physical state-velocity relationship under the chain rule, thereby extending the transformer architecture. We further validate our proposed techniques on various transformer models for NLP and CV tasks. The results consistently demonstrate that Wavy Transformer improves performance with minimal additional parameters and no extra hyperparameter tuning.



Paperid:4
Authors:Justin Domke
Abstract:
Many domain experts do not have the time or training to write formal Bayesian models. This paper takes an informal problem description as input, and combines a large language model and a probabilistic programming language to create a joint distribution over formal models, latent variables, and data. A posterior over latent variables follows by conditioning on observed data and integrating over formal models. This presents a challenging inference problem. We suggest an inference recipe that amounts to generating many formal models from the large language model, performing approximate inference on each, and then doing a weighted average. This is justified an analyzed as a combination of self-normalized importance sampling, MCMC, and variational inference. We show that this produces sensible predictions without the need to specify a formal model.



Paperid:5
Authors:Linghao Zhang, Shilin He, Chaoyun Zhang, Yu Kang, Bowen Li, Chengxing Xie, Junhao Wang, Maoquan Wang, Yufan Huang, Shengyu Fu, Elsie Nallipogu, Qingwei Lin, Yingnong Dang, Saravan Rajmohan, Dongmei Zhang
Abstract:
The issue-resolving task, where a model proposes patches to fix real-world bugs, has become a key benchmark for evaluating the capabilities of large language models (LLMs). SWE-bench is the prevailing benchmark in this domain, but it suffers from several limitations: it has not been updated since its release, is limited to just 12 repositories, and relies heavily on manual effort for both constructing test instances and setting up runnable environments, making it difficult to scale. We present SWE-bench-Live, a live-updatable benchmark designed to address these limitations. Our initial release includes 1,319 tasks derived from real GitHub issues created since 2024, spanning 93 repositories. Each task is accompanied by a dedicated Docker image to ensure reproducible execution. We introduce an automated curation pipeline that streamlines the entire process from instance creation to environment setup, removing manual bottlenecks and enabling continuous, scalable updates. We evaluate a range of state-of-the-art models and agent frameworks on SWE-bench-Live, offering detailed empirical insights into their real-world bug-fixing capabilities. By providing a fresh, diverse, and executable benchmark grounded in live repository activity, SWE-bench-Live supports reliable, large-scale assessment of code LLMs and code agents in realistic development settings.



Paperid:6
Authors:Xi Ding, Lei Wang, Piotr Koniusz, Yongsheng Gao
Abstract:
We propose Graph Consistency Regularization (GCR), a novel framework that injects relational graph structures, derived from model predictions, into the learning process to promote class-aware, semantically meaningful feature representations. Functioning as a form ofself-prompting, GCR enables the model to refine its internal structure using its own outputs. While deep networks learn rich representations, these often capture noisy inter-class similarities that contradict the model's predicted semantics. GCR addresses this issue by introducing parameter-freeGraph Consistency Layers(GCLs) at arbitrary depths. Each GCL builds a batch-level feature similarity graph and aligns it with a global, class-aware masked prediction graph, derived by modulating softmax prediction similarities with intra-class indicators. This alignment enforces that feature-level relationships reflect class-consistent prediction behavior, acting as asemantic regularizerthroughout the network. Unlike prior work, GCR introduces a multi-layer, cross-space graph alignment mechanism with adaptive weighting, where layer importance is learned from graph discrepancy magnitudes. This allows the model to prioritize semantically reliable layers and suppress noisy ones, enhancing feature quality without modifying the architecture or training procedure. GCR is model-agnostic, lightweight, and improves semantic structure across various networks and datasets. Experiments show that GCR promotes cleaner feature structure, stronger intra-class cohesion, and improved generalization, offering a new perspective on learning from prediction structure.



Paperid:7
Authors:John Hughes, Sara Price, Aengus Lynch, Rylan Schaeffer, Fazl Barez, Arushi Somani, Sanmi Koyejo, Henry Sleight, Erik Jones, Ethan Perez, Mrinank Sharma
Abstract:
We introduce Best-of-N (BoN) Jailbreaking, a simple black-box algorithm that jailbreaks frontier AI systems across modalities. BoN Jailbreaking works by repeatedly sampling variations of a prompt with a combination of augmentations---such as random shuffling or capitalization for textual prompts---until a harmful response is elicited. We find that BoN Jailbreaking achieves high attack success rates (ASRs) on closed-source language models, such as 89% on GPT-4o and 78% on Claude 3.5 Sonnet when sampling 10,000 augmented prompts. Further, it is similarly effective at circumventing state-of-the-art open-source defenses like circuit breakers and reasoning models like o1. BoN also seamlessly extends to other modalities: it jailbreaks vision language models (VLMs) such as GPT-4o and audio language models (ALMs) like Gemini 1.5 Pro, using modality-specific augmentations. BoN reliably improves when we sample more augmented prompts. Across all modalities, ASR, as a function of the number of samples (N), empirically follows power-law-like behavior for many orders of magnitude. BoN Jailbreaking can also be composed with other black-box algorithms for even more effective attacks---combining BoN with an optimized prefix attack achieves up to a 35% increase in ASR. Overall, our work indicates that, despite their capability, language models are sensitive to seemingly innocuous changes to inputs, which attackers can exploit across modalities.



Paperid:8
Authors:Xuan Son Nguyen, Aymeric Histace, Nistor Grozavu
Abstract:
Riemannian symmetric spaces (RSS) such as hyperbolic spaces and symmetric positive definite (SPD) manifolds have become popular spaces for representation learning. In this paper, we propose a novel approach for building discriminative neural networks on Siegel spaces, a family of RSS that is largely unexplored in machine learning tasks. For classification applications, one focus of recent works is the construction of multiclass logistic regression (MLR) and fully-connected (FC) layers for hyperbolic and SPD neural networks. Here we show how to build such layers for Siegel neural networks. Our approach relies on the quotient structure of those spaces and the notation of vector-valued distance onRSS. We demonstrate the relevance of our approach on two applications, i.e., radar signal classification and node classification. Our results successfully demonstrate state-of-the-art performance across all datasets.



Paperid:9
Authors:Ahmet Hamdi Güzel, Matthew T Jackson, Jarek Liesen, Tim Rocktäschel, Jakob Foerster, Ilija Bogunovic, Jack Parker-Holder
Abstract:
Training agents to act in embodied environments typically requires vast training data or access to accurate simulation, neither of which exists for many cases in the real world. Instead, world models are emerging as an alternative–leveraging offline, passively collected data, they make it possible to generate diverse worlds for training agents in simulation. In this work, we harness world models to generate “imagined” environments to train robust agents capable of generalizing to novel task variations. One of the challenges in doing this is ensuring the agent trains on useful generated data. We thus propose a novel approach IMAC (Imagined Autocurricula) leveraging Unsupervised Environment Design (UED), induces an automatic curriculum over generated worlds. In a series of challenging, procedurally generated environments, we show it is possible to achieve strong transfer performance on held-out environments having trained only inside a world model learned from a narrower dataset. We believe this opens the path to utilizing larger-scale, foundation world models for generally capable agents.



Paperid:10
Authors:Lu Xu, Tsai Hor Chan, Lequan Yu, Kwok Lam, Guosheng Yin
Abstract:
Density estimation is essential for generative modeling, particularly with the rise of modern neural networks. While existing methods capture complex data distributions, they often lack interpretability and uncertainty quantification. Bayesian nonparametric methods, especially the polya tree, offer a robust framework that addresses these issues by accurately capturing function behavior over small intervals. Traditional techniques like Markov chain Monte Carlo (MCMC) face high computational complexity and scalability limitations, hindering the use of Bayesian nonparametric methods in deep learning. To tackle this, we introduce the variational polya tree (VPT) model, which employs stochastic variational inference to compute posterior distributions. This model provides a flexible, nonparametric Bayesian prior that captures latent densities and works well with stochastic gradient optimization. We also leverage the joint distribution likelihood for a more precise variational posterior approximation than traditional mean-field methods. We evaluate the model's performance on both real data and images, and demonstrate its competitiveness with other state-of-art deep density estimation methods. We also explore its ability in enhancing interpretability and uncertainty quantification. Code is anonymously available at \url{https://anonymous.4open.science/r/var-polya-tree-8726/}.



Paperid:11
Authors:Kiarash Banihashem, MohammadTaghi Hajiaghayi, Danny Mittal
Abstract:
Abstract:We investigate the problem of sequentially selecting elements of an unknown matroid in an online manner to form an independent set, with the goal of maximizing the minimum probability of acceptance across all elements, a property we define as $f$-fairness. Under adversarial arrival orders, we design an $\alpha(\ln(k)+1)$-fair algorithm, where $\alpha$ is the arboricity of the matroid and $k$ is the rank, a result that is nearly optimal. For laminar matroids, we develop an $(2\alpha-1)$-fair algorithm, which is optimal up to constant factors, achieved through a novel online coloring scheme. In the random arrival order setting, we achieve a $(4+o(1))\alpha$-fair algorithm for graphic matroids, matching the optimal result up to constant factors, relying on a novel technique for learning a degeneracy ordering using a sampled subset of edges. We further generalize our result to $p$-matchoids, obtaining a $\beta(p\ln k+1)$-fair algorithm for the adversarial arrival model, where $\beta$ is the optimal offline fairness. Notably, all our results can be extended to a setting with no prior knowledge of the matroid with only a logarithmic increase in the fairness factor.



Paperid:12
Authors:Ioannis Mavrothalassitis, Pol Puigdemont, Noam Levi, Volkan Cevher
Abstract:
Contrary to common belief, we show that gradient ascent-based unconstrained optimization methods frequently fail to perform machine unlearning, a phenomenon we attribute to the inherent statistical dependence between the forget and retain data sets. This dependence, which can manifest itself even as simple correlations, undermines the misconception that these sets can be independently manipulated during unlearning. We provide empirical and theoretical evidence showing these methods often fail precisely due to this overlooked relationship. For random forget sets, this dependence means that degrading forget set metrics (which, for a retrained model, should mirror test set metrics) inevitably harms overall test performance. Going beyond random sets, we consider logistic regression as an instructive example where a critical failure mode emerges: inter-set dependence causes gradient descent-ascent iterations to progressively diverge from the ideal retrained model. Strikingly, these methods can converge to solutions that are not only far from the retrained ideal but are potentially even further from it than the original model itself, rendering the unlearning process actively detrimental. A toy example further illustrates how this dependence can trap models in inferior local minima, inescapable via finetuning. Our findings highlight that the presence of such statistical dependencies, even when manifest only as correlations, can be sufficient for ascent-based unlearning to fail. Our theoretical insights are corroborated by experiments on complex neural networks, demonstrating that these methods do not perform as expected in practice due to this unaddressed statistical interplay.



Paperid:13
Authors:Boris Bonev, Max Rietmann, Andrea Paris, Alberto Carpentieri, Thorsten Kurth
Abstract:
We introduce a generalized attention mechanism for spherical domains, enabling Transformer architectures to natively process data defined on the two-dimensional sphere - a critical need in fields such as atmospheric physics, cosmology, and robotics, where preserving spherical symmetries and topology is essential for physical accuracy. By integrating numerical quadrature weights into the attention mechanism, we obtain a geometrically faithful spherical attention that is approximately rotationally equivariant, providing strong inductive biases and leading to better performance than Cartesian approaches.To further enhance both scalability and model performance, we propose neighborhood attention on the sphere, which confines interactions to geodesic neighborhoods. This approach reduces computational complexity and introduces the additional inductive bias for locality, while retaining the symmetry properties of our method. We provide optimized CUDA kernels and memory-efficient implementations to ensure practical applicability.The method is validated on three diverse tasks: simulating shallow water equations on the rotating sphere, spherical image segmentation, and spherical depth estimation. Across all tasks, our spherical Transformers consistently outperform their planar counterparts, highlighting the advantage of geometric priors for learning on spherical domains.



Paperid:14
Authors:Jianyu Zhang, Leon Bottou
Abstract:
Memory Mosaics, networks of associative memories, have demonstrated appealing compositional and in-context learning capabilities on medium-scale networks (GPT-2 scale) and synthetic small datasets. This work shows that these favorable properties remain when we scale memory mosaics to large language model sizes (llama-8B scale) and real-world datasets. To this end, we scale memory mosaics to 10B size, we train them on one trillion tokens, we introduce a couple architectural modifications (memory mosaics v2), we assess their capabilities across three evaluation dimensions: training-knowledge storage, new-knowledge storage, and in-context learning. Throughout the evaluation, memory mosaics v2 match transformers on the learning of training knowledge (first dimension) and significantly outperforms transformers on carrying out new tasks at inference time (second and third dimensions). These improvements cannot be easily replicated by simply increasing the training data for transformers. A memory mosaics v2 trained on one trillion tokens still perform better on these tasks than a transformer trained on eight trillion tokens.



Paperid:15
Authors:Jiaxin Guo, Zewen Chi, Li Dong, Qingxiu Dong, Xun Wu, Shaohan Huang, Furu Wei
Abstract:
Reward models play a critical role in guiding large language models toward outputs that align with human expectations. However, an open challenge remains in effectively utilizing test-time compute to enhance reward model performance. In this work, we introduce Reward Reasoning Models (RRMs), which are specifically designed to execute a deliberate reasoning process before generating final rewards. Through chain-of-thought reasoning, RRMs leverage additional test-time compute for complex queries where appropriate rewards are not immediately apparent. To develop RRMs, we implement a reinforcement learning framework that fosters self-evolved reward reasoning capabilities without requiring explicit reasoning traces as training data. Experimental results demonstrate that RRMs achieve superior performance on reward modeling benchmarks across diverse domains. Notably, we show that RRMs can adaptively exploit test-time compute to further improve reward accuracy. Our reward models are available at https://huggingface.co/xxx.



Paperid:16
Authors:Runling Long, Yunlong Wang, Jia Wan, Xiang Deng, Xinting Zhu, Weili Guan, Antoni Chan, Liqiang Nie
Abstract:
Occlusion is one of the fundamental challenges in crowd counting. In the community, various data-driven approaches have been developed to address this issue, yet their effectiveness is limited. This is mainly because most existing crowd counting datasets on which the methods are trained are based on passive cameras, restricting their ability to fully sense the environment.Recently, embodied navigation methods have shown significant potential in precise object detection in interactive scenes. These methods incorporate active camera settings, holding promise in addressing the fundamental issues in crowd counting. However, most existing methods are designed for indoor navigation, showing unknown performance in analyzing complex object distribution in large-scale scenes, such as crowds. Besides, most existing embodied navigation datasets are indoor scenes with limited scale and object quantity, preventing them from being introduced into dense crowd analysis. Based on this, a novel task, Embodied Crowd Counting (ECC), is proposed to count the number of persons in a large-scale scene actively. We then build up an interactive simulator, the Embodied Crowd Counting Dataset (ECCD), which enables large-scale scenes and large object quantities. A prior probability distribution approximating a realistic crowd distribution is introduced to generate crowds. Then, a zero-shot navigation method (ZECC) is proposed as a baseline. This method contains an MLLM-driven coarse-to-fine navigation mechanism, enabling active Z-axis exploration, and a normal-line-based crowd distribution analysis method for fine counting. Experimental results show that the proposed method achieves the best trade-off between counting accuracy and navigation cost.



Paperid:17
Authors:Difan Deng, Marius Lindauer
Abstract:
We present Neural Attention Search (NAtS), an end-to-end learnable sparse transformer that automatically evaluates the importance of each token within a sequence and determines if the corresponding token can be dropped after several steps. To this end, we design a search space that contains three token types: (i) Global Tokens will be preserved and queried by all the following tokens. (ii) Local Tokens survive until the next global token appears. (iii) Sliding Window Tokens have an impact on the inference of a fixed size of the next following tokens. Similar to the One-Shot Neural Architecture Search approach, this token-type information can be learned jointly with the architecture weights via a learnable attention mask. Experiments on both training a new transformer from scratch and fine-tuning existing large language models show that NAtS can efficiently reduce the KV cache and the inference costs for the transformer-based models while maintaining the models' performance.



Paperid:18
Authors:Patrick Cheridito, Aleksey Minabutdinov
Abstract:
We introduce a novel deep learning algorithm for computing convex conjugates of differentiable convex functions, a fundamental operation in convex analysis with various applications in different fields such as optimization, control theory, physics and economics. While traditional numerical methods suffer from the curse of dimensionality and become computationally intractable in high dimensions, more recent neural network-based approaches scale better, but have mostly been studied with the aim of solving optimal transport problems and require the solution of complicated optimization or max-min problems. Using an implicit Fenchel formulation of convex conjugation, our approach facilitates an efficient gradient-based framework for the minimization of approximation errors and, as a byproduct, also provides a posteriori guarantees for the quality of approximations. Numerical experiments demonstrate our method's ability to deliver accurate results across different high-dimensional examples. Moreover, by employing symbolic regression with Kolmogorov–Arnold networks, we are able to achieve exact computation of convex conjugates for specific convex functions.



Paperid:19
Authors:Runyi Hu, Jie Zhang, Shiqian Zhao, Nils Lukas, Jiwei Li, Qing Guo, Han Qiu, Tianwei Zhang
Abstract:
We present MaskMark, a simple, efficient, and flexible framework for image watermarking. MaskMark has two variants: (1) MaskMark-D, which supports global watermark embedding, watermark localization, and local watermark extraction for applications such as tamper detection; (2) MaskMark-ED, which focuses on local watermark embedding and extraction, offering enhanced robustness in small regions to support fine-grined image protection. MaskMark-D builds on the classical encoder-distortion layer-decoder training paradigm. In MaskMark-D, we introduce a simple masking mechanism during the decoding stage that enables both global and local watermark extraction. During training, the decoder is guided by various types of masks applied to watermarked images before extraction, helping it learn to localize watermarks and extract them from the corresponding local areas. MaskMark-ED extends this design by incorporating the mask into the encoding stage as well, guiding the encoder to embed the watermark in designated local regions, which improves robustness under regional attacks. Extensive experiments show that MaskMark achieves state-of-the-art performance in global and local watermark extraction, watermark localization, and multi-watermark embedding. It outperforms all existing baselines, including the recent leading model WAM for local watermarking, while preserving high visual quality of the watermarked images. In addition, MaskMark is highly efficient and adaptable. It requires only 20 hours of training on a single A6000 GPU, achieving 15× computational efficiency compared to WAM. By simply adjusting the distortion layer, MaskMark can be quickly fine-tuned to meet varying robustness requirements.



Paperid:20
Authors:Jonghae Park, Daesol Cho, Jusuk Lee, Dongseok Shim, Inkyu Jang, H. Jin Kim
Abstract:
Unsupervised skill discovery in reinforcement learning (RL) aims to learn diverse behaviors without relying on external rewards. However, current methods often overlook the periodic nature of learned skills, focusing instead on increasing the mutual dependency between states and skills or maximizing the distance traveled in latent space. Considering that many robotic tasks—particularly those involving locomotion—require periodic behaviors across varying timescales, the ability to discover diverse periodic skills is essential. Motivated by this, we propose Periodic Skill Discovery (PSD), a framework that discovers periodic behaviors in an unsupervised manner. The key idea of PSD is to train an encoder that maps states to a circular latent space, thereby naturally encoding periodicity in the latent representation. By capturing temporal distance, PSD can effectively learn skills with diverse periods in complex robotic tasks, even with pixel-based observations. We further show that these learned skills achieve high performance on downstream tasks such as hurdling. Moreover, integrating PSD with an existing skill discovery method offers more diverse behaviors, thus broadening the agent’s repertoire. Our demos are available at https://sites.google.com/view/neurips2025psd/



Paperid:21
Authors:Shivalika Singh, Yiyang Nan, Alex Wang, Daniel Dsouza, Sayash Kapoor, Ahmet Üstün, Sanmi Koyejo, Yuntian Deng, Shayne Longpre, Noah Smith, Beyza Ermis, Marzieh Fadaee, Sara Hooker
Abstract:
Measuring progress is fundamental to the advancement of any scientific field. As benchmarks play an increasingly central role, they also grow more susceptible to distortion.Chatbot Arena has emerged as the go-to leaderboard for ranking the most capable AI systems. Yet, in this work we identify systematic issues that have resulted in a distorted playing field. We find that undisclosed private testing practices benefit a handful of providers who are able to test multiple variants before public release and retract scores if desired. We establish that the ability of these providers to choose the best score leads to biased Arena scores due to selective disclosure of performance results. At an extreme, we found one provider testing 27 private variants before making one model public at the second position on the leaderboard. We also establish that proprietary closed models are sampled at higher rates (number of battles) and have fewer models removed from the arena than open-weight and open-source alternatives. Both these policies lead to large data access asymmetries over time. The top two providers have individually received an estimated 19.2% and 20.4% of all data on the arena. In contrast, a combined 83 open-weight models have only received an estimated 29.7% of the total data. With conservative estimates, we show that access to Chatbot Arena data yields substantial benefits; even limited additional data can result in relative performance gains of up to 112% on ArenaHard, a test set from the arena distribution.Together, these dynamics result in overfitting to Arena-specific dynamics rather than general model quality. The Arena builds on the substantial efforts of both the organizers and an open community that maintains this valuable evaluation platform. We offer actionable recommendations to reform the Chatbot Arena's evaluation framework and promote fairer, more transparent benchmarking for the field.



Paperid:22
Authors:Yihong Dong, Ge Li, Yongding Tao, Xue Jiang, Kechi Zhang, Jia Li, Jinliang Deng, Jing Su, Jun Zhang, Jingjing Xu
Abstract:
Despite the remarkable successes of general-purpose neural networks, such as MLPs and Transformers, we find that they exhibit notable shortcomings in modeling and reasoning about periodic phenomena, achieving only marginal performance within the training domain and failing to generalize effectively to out-of-domain (OOD) scenarios. Periodicity is ubiquitous throughout nature and science. Therefore, neural networks should be equipped with the essential ability to model and handle periodicity. In this work, we propose FAN, a novel neural network that effectively addresses periodicity modeling challenges while offering broad applicability similar to MLP with fewer parameters and FLOPs. Periodicity is naturally integrated into FAN's structure and computational processes by introducing the Fourier Principle. Unlike existing Fourier-based networks, which possess particular periodicity modeling abilities but face challenges in scaling to deeper networks and are typically designed for specific tasks, our approach overcomes this challenge to enable scaling to large-scale models and maintains general-purpose modeling capability. Through extensive experiments, we demonstrate the superiority of FAN in periodicity modeling tasks and the effectiveness and generalizability of FAN across a range of real-world tasks. Moreover, we reveal that compared to existing Fourier-based networks, FAN accommodates both periodicity modeling and general-purpose modeling well.



Paperid:23
Authors:Julian Büchel, Iason Chalas, Giovanni Acampa, An Chen, Omobayode Fagbohungbe, Hsinyu Tsai, Kaoutar El Maghraoui, Manuel Le Gallo, Abbas Rahimi, Abu Sebastian
Abstract:
Analog in-memory computing (AIMC) is a promising compute paradigm to improve speed and power efficiency of neural network inference beyond the limits of conventional von Neumann-based architectures. However, AIMC introduces fundamental challenges such as noisy computations and strict constraints on input and output quantization. Because of these constraints and imprecisions, off-the-shelf LLMs are not able to achieve 4-bit-level performance when deployed on AIMC-based hardware. While researchers previously investigated recovering this accuracy gap on small, mostly vision-based models, a generic method applicable to LLMs pre-trained on trillions of tokens does not yet exist. In this work, we introduce a general and scalable method to robustly adapt LLMs for execution on noisy, low-precision analog hardware. Our approach enables state-of-the-art models — including Phi-3-mini-4k-instruct and Llama-3.2-1B-Instruct — to retain performance comparable to 4-bit weight, 8-bit activation baselines, despite the presence of analog noise and quantization constraints. Additionally, we show that as a byproduct of our training methodology, analog foundation models can be quantized for inference on low-precision digital hardware. Finally, we show that our models also benefit from test-time compute scaling, showing better scaling behavior than models trained with 4-bit weight and 8-bit static input quantization. Our work bridges the gap between high-capacity LLMs and efficient analog hardware, offering a path toward energy-efficient foundation models. Code is available atanonymous.4open.science/r/analog-foundation-models-BB03.



Paperid:24
Authors:Yukito Tajima, Nakamasa Inoue, Yusuke Sekikawa, Ikuro Sato, Rio Yokota
Abstract:
Abstract:Gated Linear Units (GLUs) have become essential components in the feed-forward networks of state-of-the-art Large Language Models (LLMs).However, they require twice as many memory reads compared to feed-forward layers without gating, due to the use of separate weight matrices for the gate and value streams.To address this bottleneck, we introduce Masked Gated Linear Units (MGLUs), a novel family of GLUs with an efficient kernel implementation.The core contribution of MGLUs include:(1) the Mixture of Element-wise Gating (MoEG) architecture that learns multiple binary masks, each determining gate or value assignments at the element level on a single shared weight matrix resulting in reduced memory transfer, and (2) FlashMGLU, a hardware-friendly kernel that yields up to a 19.7$\times$ inference-time speed-up over a na\"ive PyTorch MGLU and is 47\% more memory-efficient and 34\% faster than standard GLUs despite added architectural complexity on an RTX5090 GPU.In LLM experiments, the Swish-activated variant SwiMGLU preserves its memory advantages while matching—or even surpassing—the downstream accuracy of the SwiGLU baseline.



Paperid:25
Authors:Ayoub El Hanchi, Murat Erdogdu, Chris Maddison
Abstract:
Abstract:What property of the data distribution determines the excess risk of principal component analysis? In this paper, we provide a precise answer to this question. We establish a central limit theorem for the error of the principal subspace estimated by PCA, and derive the asymptotic distribution of its excess risk under the reconstruction loss. We obtain a non-asymptotic upper bound on the excess risk of PCA that recovers, in the large sample limit, our asymptotic characterization. Underlying our contributions is the following result: we prove that the negative block Rayleigh quotient, defined on the Grassmannian, is generalized self-concordant along geodesics emanating from its minimizer of maximum rotation less than $\pi/4$.



Paperid:26
Authors:Ben Anson, Xi Wang, Laurence Aitchison
Abstract:
One persistent challenge in LLM research is the development of attention mechanisms that are able to generalise from training on shorter contexts to inference on longer contexts. We propose two conditions that we expect all effective long-context attention mechanisms to have: scale-invariant total attention, and scale-invariant attention sparsity. Under a Gaussian assumption, we show that a simple position-dependent transformation of the attention logits is sufficient for these conditions to hold. Experimentally we find that the resulting scale-invariant attention scheme gives considerable benefits in terms of validation loss when zero-shot generalising from training on short contexts to validation on longer contexts, and is effective at long-context retrieval.



Paperid:27
Authors:Yash Savani, Asher Trockman, Zhili Feng, Yixuan Xu, Avi Schwarzschild, Alexander Robey, Marc Finzi, J. Zico Kolter
Abstract:
Frontier models that generate extended reasoning traces inadvertently produce token sequences that can facilitate model distillation. Recognizing this vulnerability, model owners may seek sampling strategies that limit the effectiveness of distillation without compromising model performance.Antidistillation samplingprovides exactly this capability. By strategically modifying a model's next-token probability distribution, antidistillation sampling poisons reasoning traces, rendering them significantly less effective for distillation while preserving the model's utility.



Paperid:28
Authors:Chang Nie
Abstract:
Originating in quantum physics, tensor networks (TNs) have been widely adopted as exponential machines and parametric decomposers for recognition tasks. Typical TN models, such as Matrix Product States (MPS), have not yet achieved successful application in natural image recognition. When employed, they primarily serve to compress parameters within pre-existing networks, thereby losing their distinctive capability to enhance exponential-order feature interactions. This paper introduces a novel architecture named Deep Tree Tensor Network (DTTN), which captures -order multiplicative interactions across features through multilinear operations, while essentially unfolding into a tree-like TN topology with the parameter-sharing property. DTTN is stacked with multiple antisymmetric interacting modules (AIMs), and this design facilitates efficient implementation. Furthermore, our theoretical analysis demonstrates the equivalence between quantum-inspired TN models and polynomial/multilinear networks under specific conditions. We posit that the DTTN could catalyze more interpretable research within this field. The proposed model is evaluated across multiple benchmarks and application domains, demonstrating superior performance compared to both peer methods and state-of-the-art architectures. Our code will be made publicly available upon publication.



Paperid:29
Authors:Kiarash Banihashem, Samira Goudarzi, MohammadTaghi Hajiaghayi, Mohammadhossein Bateni, Hossein Esfandiari
Abstract:
Abstract:We explore the concept of replicability, which ensures algorithmic consistency despite input data variations, for online pricing problems, specifically prophet inequalities and delegation. Given the crucial role of replicability in enhancing transparency in economic decision-making, we present a replicable and nearly optimal pricing strategy for prophet inequalities, achieving a sample complexity of $\text{poly}(\log^*|\mathcal{X}|)$, where $\mathcal{X}$ is the ground set of distributions. Furthermore, we extend these findings to the delegation problem and establish a matching lower bound, proving the necessity of the $\log^*|\mathcal{X}|$ dependence. En route to obtaining these results, we develop a number of technical contributions which are of independent interest. Most notably, we propose a new algorithm for a variant of the heavy hitter problem, which has a nearly linear dependence on the inverse of the heavy hitter parameter, significantly improving upon existing results which have a cubic dependence.



Paperid:30
Authors:Boyu Chen, Zhixiang Zhou, Liuhua Peng, Zhonglei Wang
Abstract:
Limited labeling budget severely impedes data-driven research, such as medical analysis, remote sensing and population census, and active inference is a solution to this problem. Prior works utilizing independent sampling have achieved improvements over uniform sampling, but its insufficient usage of available information undermines its statistical efficiency. In this paper, we propose balanced active inference, a novel algorithm that incorporates balanced constraints based on model uncertainty utilizing the cube method for label selection. Under regularity conditions, we establish its asymptotic properties and also prove that the statistical efficiency of the proposed algorithm is higher than its alternatives. Various numerical experiments, including regression and classification in both synthetic setups and real data analysis, demonstrate that the proposed algorithm outperforms its alternatives while guaranteeing nominal coverage.



Paperid:31
Authors:Yidong Zhou, SU I IAO, Hans-Georg Müller
Abstract:
Gradient boosting has become a cornerstone of machine learning, enabling base learners such as decision trees to achieve exceptional predictive performance. While existing algorithms primarily handle scalar or Euclidean outputs, increasingly prevalent complex-structured data, such as distributions, networks, and manifold-valued outputs, present challenges for traditional methods. Such non-Euclidean data lack algebraic structures such as addition, subtraction, or scalar multiplication required by standard gradient boosting frameworks. To address these challenges, we introduce Fréchet geodesic boosting (FGBoost), a novel approach tailored for outputs residing in geodesic metric spaces. FGBoost leverages geodesics as proxies for residuals and constructs ensembles in a way that respects the intrinsic geometry of the output space. Through theoretical analysis, extensive simulations, and real-world applications, we demonstrate the strong performance and adaptability of FGBoost, showcasing its potential for modeling complex data.



Paperid:32
Authors:Anupam Gupta, Guru Guruganesh, Renato Leme, Jon Schneider
Abstract:
Abstract:We provide an algorithm with regret $O(C d \log \log T)$ for contextual pricing with $C$ corrupted rounds, improving over the previous bound of $O(d^3 C \log^2(T))$ of Krishnamurthy et al. The result is based on a reduction that calls the uncorrupted algorithm as a black-box, unlike the previous approach that modifies the inner workings of the uncorrupted algorithm. As a result, it leads to a conceptually simpler algorithm.Finally, we provide a lower bound ruling out a $O(C + d\log \log T)$ algorithm. This shows that robustifying contextual pricing is harder than robustifying contextual search with $\epsilon$-ball losses, for which it is possible to design algorithms where corruptions add only an extra additive term $C$ to the regret.



Paperid:33
Authors:Vaggelis Dorovatas, Georgios Paraskevopoulos, Alexandros Potamianos
Abstract:
Deep neural networks with short residual connections have demonstrated remarkable success across domains, but increasing depth often introduces computational redundancy without corresponding improvements in representation quality. We introduce Auto-Compressing Networks (ACNs), an architectural variant where additive long feedforward connections from each layer to the output replace traditional short residual connections. By analyzing the distinct dynamics induced by this modification, we reveal a unique property we coin asauto-compression—the ability of a network to organically compress information during training with gradient descent, through architectural design alone. Through auto-compression, information is dynamically "pushed" into early layers during training, enhancing their representational quality and revealing potential redundancy in deeper ones. We theoretically show that this property emerges from layer-wise training patterns found only in ACNs, where layers are dynamically utilized during training based on task requirements. We also find that ACNs exhibit enhanced noise robustness compared to residual networks, superior performance in low-data settings, improved transfer learning capabilities, and mitigate catastrophic forgetting suggesting that they learn representations that generalize better despite using fewer parameters. Our results demonstrate up to 18\% reduction in catastrophic forgetting and 30-80\% architectural compression while maintaining accuracy across vision transformers, MLP-mixers, and BERT architectures. These findings establish ACNs as a practical approach to developing efficient neural architectures that automatically adapt their computational footprint to task complexity, while learning robust representations suitable for noisy real-world tasks and continual learning scenarios.



Paperid:34
Authors:Saba Ahmadi, Siddharth Bhandari, Avrim Blum
Abstract:
We investigate the concept of algorithmic replicability introduced by Impagliazzo et al.(2022) in an online setting. In our model, the input sequence received by the online learner is generated from time-varying distributions chosen by an adversary (obliviously). Our objective is to design low-regret online algorithms that, with high probability, produce the \emph{exact same sequence} of actions when run on two independently sampled input sequences generated as described above. We refer to such algorithms as adversarially replicable.Previous works explored replicability in the online setting under inputs generated independently from a fixed distribution; we term this notion as iid-replicability. Our model generalizes to capture both adversarial and iid input sequences, as well as their mixtures, which can be modeled by setting certain distributions as point-masses.We demonstrate adversarially replicable online learning algorithms for online linear optimization and the experts problem that achieve sub-linear regret. Additionally, we propose a general framework for converting an online learner into an adversarially replicable one within our setting, bounding the new regret in terms of the original algorithm’s regret. We also present a nearly optimal (in terms of regret) iid-replicable online algorithm for the experts problem, highlighting the distinction between the iid and adversarial notions of replicability.Finally, we establish lower bounds on the regret (in terms of the replicability parameter and time) that any replicable online algorithm must incur.



Paperid:35
Authors:Rodrigo Maulen, Claire Boyer, Pierre Marion
Abstract:
Transformers have emerged as a powerful neural network architecture capable of tackling a wide range of learning tasks. In this work, we provide a theoretical analysis of their ability to automatically extract structure from data in an unsupervised setting. In particular, we demonstrate their suitability for clustering when the input data is generated from a Gaussian mixture model. To this end, we study a simplified two-head attention layer and define a population risk whose minimization with unlabeled data drives the head parameters to align with the true mixture centroids.



Paperid:36
Authors:Junying Wang, Edith Tretschk, Ziyan Wang, Yuanlu Xu, Anastasia Ianina, Aljaz Bozic, Ulrich Neumann, Tony Tung
Abstract:
The creation of photorealistic dynamic hair remains a major challenge in digital human modeling because of the complex motions, occlusions, and light scattering. Existing methods often resort to static capture and physics-based models that do not scale as they require manual parameter fine-tuning to handle the diversity of hairstyles and motions, and heavy computation to obtain high-quality appearance. In this paper, we present Dynamic Gaussian Hair (DGH), a novel framework that efficiently learns hair dynamics and appearance. We propose: (1) a coarse-to-fine model that learns temporally coherent hair motion dynamics across diverse hairstyles; (2) a strand-guided optimization module that learns a dynamic 3D Gaussian representation for hair appearance with support for differentiable rendering, enabling gradient-based learning of view-consistent appearance under motion. Unlike prior simulation-based pipelines, our approach is fully data-driven, scales with training data, and generalizes across various hairstyles and head motion sequences. Additionally, DGH can be seamlessly integrated into a 3D Gaussian avatar framework, enabling realistic, animatable hair for high-fidelity avatar representation. DGH achieves promising geometry and appearance results, providing a scalable, data-driven alternative to physics-based simulation and rendering.



Paperid:37
Authors:Ronen Gradwohl, Eilam Shapira, Moshe Tennenholtz
Abstract:
Algorithmic fairness has emerged as a central issue in ML, and it has become standard practice to adjust ML algorithms so that they will satisfy fairness requirements such as Equal Opportunity. In this paper we consider the effects of adopting such fair classifiers on the overall level ofecosystem fairness. Specifically, we introduce the study of fairness with competing firms, and demonstrate the failure of fair classifiers in yielding fair ecosystems. Our results quantify the loss of fairness in systems, under a variety of conditions, based on classifiers' correlation and the level of their data overlap. We show that even if competing classifiers are individually fair, the ecosystem's outcome may be unfair; and that adjusting biased algorithms to improve their individual fairness may lead to an overall decline in ecosystem fairness. In addition to these theoretical results, we also provide supporting experimental evidence. Together, our model and results provide a novel and essential call for action.



Paperid:38
Authors:Luke Darlow, Ciaran Regan, Sebastian Risi, Jeffrey Seely, Llion Jones
Abstract:
Biological brains demonstrate complex neural activity, where neural dynamics are critical to how brains process information. Most artificial neural networks simplify neurons by abstracting away dynamics. We challenge that paradigm. By incorporating neuron-level processing and synchronization, we reintroduce neural timing as a foundational element. We present the Continuous Thought Machine (CTM), a model designed to leverage neural dynamics as its core representation. The CTM has two innovations: (1) neuron-level temporal processing, where each neuron uses unique weight parameters to process incoming histories; and (2) neural synchronization as a latent representation. The CTM aims to strike a balance between neuron abstractions and biological realism. It operates at a level of abstraction that effectively captures essential temporal dynamics while remaining computationally tractable. We demonstrate the CTM's performance and versatility across a range of tasks, including solving 2D mazes, ImageNet-1K classification, parity computation, and more. Beyond displaying rich internal representations and offering a natural avenue for interpretation owing to its internal process, the CTM is able to perform tasks that require complex sequential reasoning. The CTM can also leverage adaptive compute, where it can stop earlier for simpler tasks, or keep computing when faced with more challenging instances. The goal of this work is to share the CTM and its associated innovations, rather than pushing for new state-of-the-art results. To that end, we believe the CTM represents a significant step toward developing more biologically plausible and powerful artificial intelligence systems. We provide an interactive online demonstration: https://anon-ctm.github.io/ctm/.



Paperid:39
Authors:Pengfei Zhou, Jie Xia, Xiaopeng Peng, Wangbo Zhao, Zilong Ye, Zekai Li, Suorong Yang, Jiadong Pan, Yuanxiang Chen, Ziqiao Wang, Kai Wang, Qian Zheng, Xiaojun Chang, Gang Pan, Shurong Dong, Kaipeng Zhang, Yang You
Abstract:
Traditional image editing typically relies on manual prompting, making it labor-intensive and inaccessible to individuals with limited motor control or language abilities. Leveraging recent advances in brain-computer interfaces (BCIs) and generative models, we propose LoongX, a hands-free image editing approach driven by multimodal neurophysiological signals. LoongX utilizes state-of-the-art diffusion models trained on a comprehensive dataset of 23,928 image editing pairs, each paired with synchronized electroencephalography (EEG), functional near-infrared spectroscopy (fNIRS), photoplethysmography (PPG), and head motion signals that capture user intent.To effectively address the heterogeneity of these signals, LoongX integrates two key modules. The cross-scale state space (CS3) module encodes informative modality-specific features. The dynamic gated fusion (DGF) module further aggregates these features into a unified latent space, which is then aligned with edit semantics via fine-tuning on a diffusion transformer (DiT).Additionally, we pre-train the encoders using contrastive learning to align cognitive states with semantic intentions from embedded natural language.Extensive experiments demonstrate that LoongX achieves performance comparable to text-driven methods (CLIP-I: 0.6605 vs. 0.6558; DINO: 0.4812 vs. 0.4637) and outperforms them when neural signals are combined with speech (CLIP-T: 0.2588 vs. 0.2549). These results highlight the promise of neural-driven generative models in enabling accessible, intuitive image editing and open new directions for cognitive-driven creative technologies. Datasets and code will be released to support future work and foster progress in this emerging area.



Paperid:40
Authors:Binh Vu, Jan Kapar, Marvin Wright, David Watson
Abstract:
We propose a principled method for autoencoding with random forests. Our strategy builds on foundational results from nonparametric statistics and spectral graph theory to learn a low-dimensional embedding of the model that optimally represents relationships in the data. We provide exact and approximate solutions to the decoding problem via constrained optimization, split relabeling, and nearest neighbors regression. These methods effectively invert the compression pipeline, establishing a map from the embedding space back to the input space using splits learned by the ensemble's constituent trees. The resulting decoders are universally consistent under common regularity assumptions. The procedure works with supervised or unsupervised models, providing a window into conditional or joint distributions. We demonstrate various applications of this autoencoder, including powerful new tools for visualization, compression, clustering, and denoising. Experiments illustrate the ease and utility of our method in a wide range of settings, including tabular, image, and genomic data.



Paperid:41
Authors:Tomás González, Giulia Fanti, Aaditya Ramdas
Abstract:
Abstract:Private Evolution (PE) is a promising training-free method for differentially private (DP) synthetic data generation. While it achieves strong performance in some domains (e.g., images and text), its behavior in others (e.g., tabular data) is less consistent. To date, the only theoretical analysis of the convergence of PE depends on unrealistic assumptions about both the algorithm’s behavior and the structure of the sensitive dataset. In this work, we develop a new theoretical framework to explain PE’s practical behavior and identify sufficient conditions for its convergence. For $d$-dimensional sensitive datasets with $n$ data points from a bounded domain, we prove that PE produces an $(\varepsilon, \delta)$-DP synthetic dataset with expected 1-Wasserstein distance $\tilde{O}(d(n\varepsilon)^{-1/d})$ from the original, establishing worst-case convergence of the algorithm as $n \to \infty$. Our analysis extends to general Banach spaces as well. We also connect PE to the Private Signed Measure Mechanism, a method for DP synthetic data generation that has thus far not seen much practical adoption. We demonstrate the practical relevance of our theoretical findings in simulations.



Paperid:42
Authors:Alfin Wijaya Rahardja, Junwei Liu, Weitong Chen, Zhenpeng Chen, Yiling Lou
Abstract:
LLM-based agent systems are emerging as a new software paradigm and have been widely adopted across diverse domains such as medicine, robotics, and programming. However, maintaining these systems requires substantial effort, as they are inevitably prone to bugs and continually evolve to meet changing external requirements. Therefore, automatically resolving agent issues (i.e.,bug reports or feature requests) is a crucial and challenging task. While recent software engineering (SE) agents (e.g., SWE-agent) have shown promise in addressing issues in traditional software systems, it remains unclear how effectively they can resolve real-world issues in agent systems, which differ significantly from traditional software. To fill this gap, we first manually analyze 201 real-world agent issues and identify common categories of agent issues. We then spend 500 person-hours constructing AgentIssue-bench, a reproducible benchmark comprising 50 agent issue resolution tasks (each with an executable environment and failure-triggering tests). We further evaluate state-of-the-art SE agents on AgentIssue-bench and reveal their limited effectiveness (.e., with only 3.33% - 12.67% resolution rates). These results underscore the unique challenges of maintaining agent systems compared to traditional software, highlighting the need for further research to develop advanced SE agents for resolving agent issues.



Paperid:43
Authors:Sanjoy Dasgupta, Syamantak Kumar, Shourya Pandey, Purnamrita Sarkar
Abstract:
Low-Precision Streaming PCA addresses the challenge of estimating the top principal component in a streaming setting under limited precision. We establish an information‐theoretic lower bound on the quantization resolution required to attain a desired accuracy in the top eigenvector. We study Oja's algorithm for streaming PCA under linear and nonlinear stochastic quantization. The quantized variants incorporate unbiased stochastic quantization of the weight vector and the data-induced updates. Under mild moment and spectral-gap assumptions on the data distribution, we show that a minibatch version achieves the lower bound up to logarithmic factors under both schemes. This leads to a \textit{dimension-free} quantization error in the nonlinear quantization setting. Empirical evaluations on synthetic streams validate our theoretical findings and demonstrate that our low-precision methods closely track the performance of standard Oja’s algorithm.



Paperid:44
Authors:Yatong Chen, Safwan Hossain, Yiling Chen
Abstract:
We study hypothesis testing in a principal-agent setting where the agent with a private belief about a product's effectiveness interacts with a principal who evaluates approval based on clinical trial outcomes. The principal's goal is to set a p-value to control for both false positives (Type I errors) and false negatives (Type II errors) while accounting for the agent's decision.Building upon earlier works, we develop a game-theoretic model that characterizes the agent's participation in the trial and the request sample size as a response to the regulator's p-value threshold with minimum assumptions. Despite intricate dynamics, we show that the principal's errors exhibit clear monotonic behavior when segmented by an efficiently computable critical p-value threshold. We use this to characterize the principal's optimal p-value. We also empirically validate our model and insights using publicly available data on drug approvals. Overall, our work offers a comprehensive perspective on strategic interactions within hypothesis testing, providing technical and regulatory insights.



Paperid:45
Authors:Chaoran Cheng, Yusong Wang, Yuxin Chen, Xiangxin Zhou, Nanning Zheng, Ge Liu
Abstract:
Consistency models are a class of generative models that enable few-step generation for diffusion and flow matching models. While consistency models have achieved promising results on Euclidean domains like images, their applications to Riemannian manifolds remain challenging due to the curved geometry. In this work, we propose the Riemannian Consistency Model (RCM), which, for the first time, enables few-step consistency modeling while respecting the intrinsic manifold constraint imposed by the Riemannian geometry. Leveraging the covariant derivative and exponential-map-based parameterization, we derive the closed-form solutions for both discrete- and continuous-time training objectives for RCM. We then demonstrate theoretical equivalence between the two variants of RCM: Riemannian consistency distillation (RCD) that relies on a teacher model to approximate the marginal vector field, and Riemannian consistency training (RCT) that utilizes the conditional vector field for training. We further propose a simplified training objective that eliminates the need for the complicated differential calculation. Finally, we provide a unique kinematics perspective for interpreting the RCM objective, offering new theoretical angles. Through extensive experiments, we manifest the superior generative quality of RCM in few-step generation on various non-Euclidean manifolds, including flat-tori, spheres, and the 3D rotation group SO(3).



Paperid:46
Authors:Baoquan Gong, Xiyuan Gao, Pengfei Zhu, Qinghua Hu, Bing Cao
Abstract:
Multimodal learning systems often encounter challenges related to modality imbalance, where a dominant modality may overshadow others, thereby hindering the learning of weak modalities. Conventional approaches often force weak modalities to align with dominant ones in “Learning to be (the same)” (Positive Learning), which risks suppressing the unique information inherent in the weak modalities. To address this challenge, we offer a new learning paradigm: "Learning Not to be" (Negative Learning). Instead of enhancing weak modalities’ target-class predictions, the dominant modalities dynamically guides the weak modality to suppress non-target classes. This stabilizes the decision space and preserves modality-specific information, allowing weak modalities preserve unique information without being over-aligned. We proceed to reveal the multimodal learning from a robustness perspective and theoretically derive the Multimodal Negative Learning (MNL) framework, which introduces a dynamic guidance mechanism tailored for negative learning. Our method provably tightens the robustness lower bound of multimodal learning by increasing the Unimodal Confidence Margin (UCoM) and reduces the empirical error of weak modalities, particularly under noisy and imbalanced scenarios. Extensive experiments across multiple benchmarks demonstrate the effectiveness and generalizability of our approach against the competing methods. Code and models will be released.



Paperid:47
Authors:zhiwen ren, Qiyi Yao, Wei Fan, Jing Qiu, Weiming Zhang, Nenghai Yu
Abstract:
Vector databases support machine learning tasks using Approximate Nearest Neighbour (ANN) query functionality, making them highly valuable digital assets. However, they also face security threats like unauthorized replication. By embedding stealth information, watermarking technology can be used for ownership authentication. This paper introduces a watermarking scheme specifically designed for vector databases. The scheme consists of four steps: generating identifiers, grouping, cryptographic mapping, and modification. Since watermark embedding requires modification of certain vectors, it may negatively affect the ANN query results. Further investigation reveals that in the widely used Hierarchical Navigable Small World (HNSW) indexing structure for vector databases, heuristic edge selection and pruning strategies result in some vectors having fewer edges or even none at all. These vectors exhibit significantly lower query frequencies than others, which means that modifying these vectors incurs less impact on query results. Based on this observation, we propose the Transparent Vector Priority (TVP) watermarking scheme, which prioritizes embedding the watermark in these low-query-frequency “transparent” vectors to minimize the impact of watermark embedding on query results. Experimental results show that compared to the current most effective and relevant watermarking schemes, the TVP scheme can significantly reduce the number of missed and false queries by approximately 75\%.



Paperid:48
Authors:Arthur da Cunha, Mikael Møller Høgsgaard, Andrea Paudice, Yuxin Sun
Abstract:
Boosting is a key method in statistical learning, allowing for converting weak learners into strong ones.While well studied in the realizable case, the statistical properties of weak-to-strong learning remain less understood in the agnostic setting, where there are no assumptions on the distribution of the labels.In this work, we propose a new agnostic boosting algorithm with substantially improved sample complexity compared to prior works under very general assumptions.Our approach is based on a reduction to the realizable case, followed by a margin-based filtering step to select high-quality hypotheses.Furthermore, we show a lower bound on the sample complexity of agnostic boosting, which matches our upper bound up to logarithmic factors.



Paperid:49
Authors:Kiljae Lee, Ziqi Liu, Weijing Tang, Yuan Zhang
Abstract:
Data Shapley is an important tool for data valuation, which quantifies the contribution of individual data points to machine learning models. In practice, group-level data valuation is desirable when data providers contribute data in batch. However, we identify that existing group-level extensions of Data Shapley are vulnerable to \emph{shell company attacks}, where strategic group splitting can unfairly inflate valuations. We propose Faithful Group Shapley Value (FGSV) that uniquely defends against such attacks. Building on original mathematical insights, we develop a provably fast and accurate approximation algorithm for computing FGSV. Empirical experiments demonstrate that our algorithm significantly outperforms state-of-the-art methods in computational efficiency and approximation accuracy, while ensuring faithful group-level valuation.



Paperid:50
Authors:Etienne Gauthier, Francis Bach, Michael Jordan
Abstract:
Abstract:We introduce *Backward Conformal Prediction*, a method that guarantees conformal coverage while providing flexible control over the size of prediction sets. Unlike standard conformal prediction, which fixes the coverage level and allows the conformal set size to vary, our approach defines a rule that constrains how prediction set sizes behave based on the observed data, and adapts the coverage level accordingly. Our method builds on two key foundations: (i) recent results by Gauthier et al. [2025] on post-hoc validity using e-values, which ensure marginal coverage of the form $\mathbb{P}(Y_{\rm test} \in \hat C_n^{\tilde{\alpha}}(X_{\rm test})) \ge 1 - \mathbb{E}[\tilde{\alpha}]$ up to a first-order Taylor approximation for any data-dependent miscoverage $\tilde{\alpha}$, and (ii) a novel leave-one-out estimator $\hat{\alpha}^{\rm LOO}$ of the marginal miscoverage $\mathbb{E}[\tilde{\alpha}]$ based on the calibration set, ensuring that the theoretical guarantees remain computable in practice. This approach is particularly useful in applications where large prediction sets are impractical such as medical diagnosis. We provide theoretical results and empirical evidence supporting the validity of our method, demonstrating that it maintains computable coverage guarantees while ensuring interpretable, well-controlled prediction set sizes.



Paperid:51
Authors:Shin Kim, Mingi Kwon, Jaeseok Jeong, Youngjung Uh
Abstract:
Rectified flow is a generative model that learns smooth transport mappings between two distributions through an ordinary differential equation (ODE). The model learns a straight ODE by reflow steps which iteratively update the supervisory flow. It allows for a relatively simple and efficient generation of high-quality images. However, rectified flow still faces several challenges. 1) The reflow process is slow because it requires a large number of generated pairs to model the target distribution. 2) It is well known that the use of suboptimal fake samples in reflow can lead to performance degradation of the learned flow model. This issue is further exacerbated by error accumulation across reflow steps and model collapse in denoising autoencoder models caused by self-consuming training.In this work, we go one step further and empirically demonstrate that the reflow process causes the learned model to drift away from the target distribution, which in turn leads to a growing discrepancy in reconstruction error between fake and real images. We reveal the drift problem and design a new reflow step, namely the conic reflow. It supervises the model by the inversions of real data points through the previously learned model and its interpolation with random initial points. Our conic reflow leads to multiple advantages. 1) It keeps the ODE paths toward real samples, evaluated by reconstruction. 2) We use only a small number of generated samples instead of large generated samples, 600K and 4M, respectively. 3) The learned model generates images with higher quality evaluated by FID, IS, and Recall. 4) The learned flow is more straight than others, evaluated by curvature. We achieve much lower FID in both one-step and full-step generation in CIFAR-10. The conic reflow generalizes to various datasets such as LSUN Bedroom and ImageNet.



Paperid:52
Authors:Wen-Chao Hu, Qi-Jie Li, Lin-Han Jia, Cunjing Ge, Yu-Feng Li, Yuan Jiang, Zhi-Hua Zhou
Abstract:
Abductive Learning (ABL) integrates machine learning with logical reasoning in a loop: a learning model predicts symbolic concept labels from raw inputs, which are revised through abduction using domain knowledge and then fed back for retraining. However, due to the nondeterminism of abduction, the training process often suffers from instability, especially when the knowledge base is large and complex, resulting in a prohibitively large abduction space. While prior works focus on improving candidate selection within this space, they typically treat the knowledge base as a static black box. In this work, we propose Curriculum Abductive Learning (C-ABL), a method that explicitly leverages the internal structure of the knowledge base to address the ABL training challenges. C-ABL partitions the knowledge base into a sequence of sub-bases, progressively introduced during training. This reduces the abduction space throughout training and enables the model to incorporate logic in a stepwise, smooth way. Experiments across multiple tasks show that C-ABL outperforms previous ABL implementations, significantly improves training stability, convergence speed, and final accuracy, especially under complex knowledge setting.



Paperid:53
Authors:Hyeonseong Jeon, Cheolhong Min, Jaesik Park
Abstract:
Planning with pretrained diffusion models has emerged as a promising approach for solving test-time guided control problems. However, standard gradient-based guidance typically performs optimally under convex and differentiable reward landscapes, showing substantially reduced effectiveness in real-world scenarios involving non-convex objectives, non-differentiable constraints, and multi-reward structures. We propose a Tree-guided Diffusion Planner (TDP), a flexible test-time planning framework that balances exploration and exploitation through structured trajectory generation. We build a trajectory tree using a two-phase denoising process: (1) diverse parent trajectories are produced via fixed-potential particle guidance to encourage broad exploration, and (2) sub-trajectories are locally refined through fast conditional denoising guided by task objectives. Our bi-level sampling framework offers a robust, training-free solution to test-time planning, addressing the limitations of gradient-based guidance by exploring diverse trajectory regions and harnessing gradient information across this expanded solution space.We evaluate our approach on three tasks, including Maze2D gold-picking, robot arm block manipulation, and AntMaze multi-goal exploration. Our approach consistently outperforms state-of-the-art planning approaches across all three tasks, which consist of a wide range of guidance functions.



Paperid:54
Authors:Isai Silva, Paola Ruiz Puentes, Jillian Pearse, Pablo Arbelaez
Abstract:
Petrography is a branch of geology that analyzes the mineralogical composition of rocks from microscopical thin section samples. It is essential for understanding rock properties across geology, archaeology, engineering, mineral exploration, and the oil industry. However, petrography is a labor-intensive task requiring experts to conduct detailed visual examinations of thin section samples through optical polarization microscopes, thus hampering scalability and highlighting the need for automated techniques. To address this challenge, we introduce the Large-scale Imaging and Thin section Optical-polarization Set (LITHOS), the largest and most diverse publicly available experimental framework for automated petrography. LITHOS includes 211,604 high-resolution RGB patches of polarized light and 105,802 expert-annotated grains across 25 mineral categories. Each annotation consists of the mineral class, spatial coordinates, and expert-defined major and minor axes represented as intersecting vector paths, capturing grain geometry and orientation. We evaluate multiple deep learning techniques for mineral classification in LITHOS and propose a dual-encoder transformer architecture that integrates both polarization modalities as a strong baseline for future reference. Our method consistently outperforms single-polarization models, demonstrating the value of polarization synergy in mineral classification. We plan to release the LITHOS Benchmark, comprising our dataset, code, and pretrained models, to foster reproducibility and further research in automated petrographic analysis.



Paperid:55
Authors:Lingjie Jiang, Xun Wu, Shaohan Huang, Qingxiu Dong, Zewen Chi, Li Dong, Xingxing Zhang, Tengchao Lv, Lei Cui, Furu Wei
Abstract:
Recent Large Reasoning Models (LRMs) have shown substantially improved reasoning capabilities over traditional Large Language Models (LLMs) by incorporating extended thinking processes prior to producing final responses. However, excessively lengthy thinking introduces substantial overhead in terms of token consumption and latency, particularly unnecessary for simple queries. In this work, we introduce introduce Native Hybrid Thinking Models (NHTMs), the first kind of model capable of adaptively determining whether to perform thinking based on the contextual information of user queries. To achieve this, we propose a two-stage training pipeline comprising Hybrid Fine-Tuning (HFT) as a cold start, followed by online reinforcement learning with the proposed Hybrid Group Policy Optimization (HGPO) to implicitly learn to select the appropriate thinking mode. Furthermore, we introduce a metric called Hybrid Accuracy to quantitatively assess the model’s capability for hybrid thinking. Expensive experimental results show that NHTMs can adaptively performs hybrid thinking on queries of varying difficulty and type. It outperforms existing LRMs and LLMs in reasoning and general capabilities while significantly improving efficiency. Together, our work advocates for a reconsideration of the appropriate use of extended thinking processes, and provides a solid starting point for building hybrid thinking systems.



Paperid:56
Authors:Yu Pan, Zhongze Cai, Huaiyang Zhong, Guanting Chen, Chonghuan Wang
Abstract:
Direct Preference Optimization (DPO) has emerged as a simple and effective approach for aligning large language models (LLMs) with human preferences, bypassing the need for a learned reward model. Despite its growing adoption, a fundamental question remains open: what characteristics of preference data are most critical for DPO performance? In this work, we provide a systematic study of how preference data distribution influences DPO, from both theoretical and empirical perspectives. We show that the quality of chosen responses plays a dominant role in optimizing the DPO objective, while the quality of rejected responses may have relatively limited impact. Our theoretical analysis characterizes the optimal response distribution under DPO and reveals how contrastiveness between responses helps primarily by improving the chosen samples. We further study an online DPO setting and show it effectively reduces to supervised fine-tuning on the chosen responses. Extensive experiments across diverse tasks confirm our findings: improving the quality of chosen responses consistently boosts performance regardless of the quality of the rejected responses. We also investigate the benefit of mixing the on-policy data. Our results interpret the mechanism behind some widely adopted strategies and offer practical insights for constructing high-impact preference datasets for LLM alignment.



Paperid:57
Authors:Siye Wu, Jian Xie, yikai zhang, Chen, Kai Zhang, Yu Su, Yanghua Xiao
Abstract:
Abstract:While large reasoning models demonstrate strong performance on complex tasks, they lack the ability to adjust reasoning token usage based on task difficulty. This often leads to the "overthinking" problem—excessive and unnecessary reasoning—which, although potentially mitigated by human intervention to control the token budget, still fundamentally contradicts the goal of achieving fully autonomous AI. In this work, we propose Adaptive Reasoning Model (ARM), a reasoning model capable of adaptively selecting appropriate reasoning formats based on the task at hand. These formats include three efficient ones—Direct Answer, Short CoT, and Code—as well as a more elaborate format, Long CoT. To train ARM, we introduce Ada-GRPO, an adaptation of Group Relative Policy Optimization (GRPO), which addresses the format collapse issue in traditional GRPO. Ada-GRPO enables ARM to achieve high token efficiency, reducing tokens by an average of $\sim$30%, and up to $\sim$70%, while maintaining performance comparable to the model that relies solely on Long CoT. Furthermore, not only does it improve inference efficiency through reduced token generation, but it also brings a $\sim$2$\times$ speedup in training. In addition to the default Adaptive Mode, ARM supports two additional reasoning modes: 1) Instruction-Guided Mode, which allows users to explicitly specify the reasoning format via special tokens—ideal when the appropriate format is known for a batch of tasks. 2) Consensus-Guided Mode, which aggregates the outputs of the three efficient formats and resorts to Long CoT in case of disagreement, prioritizing performance with higher token usage. All the resources will be released.



Paperid:58
Authors:Hoki Kim, Keonwoo Kim, Sungwon Chae, Sangwon Yoon
Abstract:
Machine unlearning aims to remove the influence of specific training data (i.e., forget data) from a trained model while maintaining performance on the remaining data (i.e., retain data). Existing approximate unlearning methods typically rely on fine-tuning or negative gradient updates. However, these methods often degrade performance on the retain data or fail to eliminate the influence of forget data.In this paper, we propose Unlearning-Aware Minimization (UAM), a novel min-max optimization framework that identifies weights with high loss on the forget data and uses their gradients to minimize the loss on the retain data. We derive an efficient optimization method for this min-max problem, enabling the successful removal of the forget data and uncovering optimal solutions that existing methods struggle to find.Our empirical results demonstrate that the proposed framework consistently outperforms existing methods in machine unlearning across various benchmark datasets, including image classification datasets (CIFAR-10, CIFAR-100, and TinyImageNet) and multiple-choice question-answering datasets using large language models (WMDP-Bio and WMDP-Cyber).



Paperid:59
Authors:Jack Hopkins, Mart Bakler, Akbir Khan
Abstract:
Large Language Models (LLMs) are rapidly saturating existing benchmarks, necessitating new open-ended evaluations. We introduce the Factorio Learning Environment (FLE), based on the game of Factorio, that tests agents in long-term planning, spatial reasoning, program synthesis, and resource optimization. FLE provides exponentially scaling challenges---from basic automation to complex factories processing millions of resource units per second. We provide two settings: (1) open-play with the open-ended task of building the largest factory on an procedurally generated map and (2) lab-play consisting of 24 bounded tasks with fixed resources. We demonstrate across both settings that models still lack strong spatial reasoning. In lab-play, we find that LLMs exhibit promising short-horizon skills, yet are unable to operate effectively in constrained environments, reflecting limitations in error analysis. In open-play, while LLMs discover automation strategies that improve growth (e.g electric-powered drilling), they fail to achieve complex automation (e.g electronic-circuit manufacturing). We have released FLE as an open-source platform.



Paperid:60
Authors:Jaehoon Yoo, Wonjung Kim, Seunghoon Hong
Abstract:
Discrete Flow-based Models (DFMs) are powerful generative models for high-quality discrete data but typically suffer from slow sampling speeds due to their reliance on iterative decoding processes. This reliance on a multi-step process originates from the factorization approximation of DFMs, which is necessary for handling high-dimensional data. In this paper, we rigorously characterize the approximation error from factorization using Conditional Total Correlation (TC), which depends on the coupling. To reduce the Conditional TC and enable efficient few-step generation, we propose Rectified Discrete Flow (ReDi), a novel iterative method that reduces factorization error by rectifying the coupling between source and target distributions. We theoretically prove that each ReDi step guarantees a monotonic decreasing Conditional TC, ensuring its convergence. Empirically, ReDi significantly reduces Conditional TC and enables few-step generation. Moreover, we demonstrate that the rectified couplings are well-suited for training efficient one-step models on image generation. ReDi offers a simple and theoretically grounded approach for tackling the few-step challenge, providing a new perspective on efficient discrete data synthesis.



Paperid:61
Authors:Kaicheng Zhang, Sinian Zhang, Doudou Zhou, Yidong Zhou
Abstract:
Transfer learning is a powerful paradigm for leveraging knowledge from source domains to enhance learning in a target domain. However, traditional transfer learning approaches often focus on scalar or multivariate data within Euclidean spaces, limiting their applicability to complex data structures such as probability distributions. To address this, we introduce a novel framework for transfer learning in regression models, where outputs are probability distributions residing in the Wasserstein space. When the informative subset of transferable source domains is known, we propose an estimator with provable asymptotic convergence rates, quantifying the impact of domain similarity on transfer efficiency. For cases where the informative subset is unknown, we develop a data-driven transfer learning procedure designed to mitigate negative transfer. The proposed methods are supported by rigorous theoretical analysis and are validated through extensive simulations and real-world applications.



Paperid:62
Authors:Adam Zweiger, Jyothish Pari, Han Guo, Yoon Kim, Pulkit Agrawal
Abstract:
Large language models (LLMs) are powerful but static; they lack mechanisms to adapt their weights in response to new tasks, knowledge, or examples. We introduceSelf-AdaptingLLMs (SEAL), a framework that enables LLMs to self-adapt by restructuring, filtering, and selectively generating training data for themselves. Given a new input, the model produces aself-edit, which may rewrite or restructure content directly, or invoke tools for data augmentation and finetuning. These self-edits drive targeted gradient updates, and a reinforcement-learning meta-optimization loop refines the policy that generates them, where the reward comes from the post-update model performance. SEAL leverages the model's autoregressive generation to take information and create weight-update directives that steer its own adaptation. Evaluations on knowledge incorporation and few-shot task learning indicate that SEAL is a promising step toward LLMs capable of autonomous adaptation.



Paperid:63
Authors:Eden Saig, Nir Rosenfeld
Abstract:
When a prediction algorithm serves a collection of users, disparities in prediction quality are likely to emerge. If users respond to accurate predictions by increasing engagement, inviting friends, or adopting trends, repeated learning creates a feedback loop that shapes both the model and the population of its users. In this work, we introduce evolutionary prediction games, a framework grounded in evolutionary game theory which models such feedback loops as natural-selection processes among groups of users. Our theoretical analysis reveals a gap between idealized and real-world learning settings: In idealized settings with unlimited data and computational power, repeated learning creates competition and promotes competitive exclusion across a broad class of behavioral dynamics. However, under realistic constraints such as finite data, limited compute, or risk of overfitting, we show that stable coexistence and mutualistic symbiosis between groups becomes possible. We analyze these possibilities in terms of their stability and feasibility, present mechanisms that can sustain their existence, and empirically demonstrate our findings.



Paperid:64
Authors:Dariusz Kalociński, Tomasz Steifer
Abstract:
Understanding when learning is possible is a fundamental task in the theory of machine learning. However, many characterizations known from the literature deal with abstract learning as a mathematical object and ignore the crucial question: when can learning be implemented as a computer program? We address this question for universal learning, a generalist theoretical model of online binary classification, recently characterized by Bousquet et al.~(STOC´21). In this model, there is no 'true' hypothesis fixed in advance; instead, an Adversary---playing the role of Nature---can change their mind for as long as local consistency with the given class of hypotheses is maintained. We require Learner to achieve finite number of mistakes while using a strategy that can be implemented as a computer program. We show that universal learning does not imply computable universal learning, even if the class of hypotheses is presented in computable way. We then study the agnostic variant of computable universal learning and provide an exact characterization of classes learnable in this sense. We also consider a variant of proper universal learning and show exactly when it is possible. Together, our results give a more realistic perspective on the existing theory of online binary classification and the related problem of inductive inference.



Paperid:65
Authors:Michael Cardei, Jacob K Christopher, Bhavya Kailkhura, Tom Hartvigsen, Nando Fioretto
Abstract:
Abstract:Discrete diffusion models are a class of generative models that construct sequences by progressively denoising samples from a categorical noise distribution. Beyond their rapidly growing ability to generate coherent natural language, these models present a new and important opportunity to enforce sequence-level constraints, a capability that current autoregressive models cannot natively provide.This paper capitalizes on this opportunity by introducing $\textit{Constrained Discrete Diffusion}$ (CDD), a novel integration of differentiable constraint optimization within the diffusion process to ensure adherence to constraints, logic rules, or safety requirements for generated sequences. Unlike conventional text generators that often rely on post-hoc filtering or model retraining for controllable generation, CDD directly imposes constraints into the discrete diffusion sampling process, resulting in a training-free and effective approach.Experiments in toxicity-controlled text generation, property-constrained molecule design, and instruction-constrained text completion demonstrate that CDD achieves $\textit{zero constraint violations}$ in a diverse array of tasks while preserving fluency, novelty, and coherence, and outperforming autoregressive and existing discrete diffusion approaches.



Paperid:66
Authors:Jan Quan, Johan Suykens, Panagiotis Patrinos
Abstract:
Abstract:Motivated by the recently shown close connection between self-attention and (kernel) principal component analysis (PCA), we revisit PCA through the framework of difference-of-convex (DC) duality, presenting several novel formulations and providing new theoretical insights. Moreover, we uncover that simultaneous iteration, which is connected to the classical QR algorithm, is an instance of the difference-of-convex algorithm (DCA), offering an optimization perspective on this longstanding method. Further, we develop new algorithms for PCA and empirically compare them with state-of-the-art approaches. Lastly, we introduce a kernelizable dual formulation for a robust variant of PCA that minimizes the $l_1$ deviation of the reconstruction errors.



Paperid:67
Authors:Chaitanya Amballa, Yu-Lin Wei, Sattwik Basu, Zhijian Yang, Mehmet Ergezer, Romit Roy Choudhury
Abstract:
Neural Radiance Fields (NeRFs) have been remarkably successful at synthesizing novel views of 3D scenes by optimizing a volumetric scene function. This scene function models how optical rays bring color information from a 3D object to the camera pixels. Radio frequency (RF) or audio signals can also be viewed as a vehicle for delivering information about the environment to a sensor. However, unlike camera pixels, an RF/audio sensor receives a mixture of signals that contain many environmental reflections (also called “multipath”). Is it still possible to infer the environment using such multipath signals? We show that with redesign, NeRFs can be taught to learn from multipath signals, and thereby “see” the environment. As a grounding application, we aim to infer the indoor floorplan of a home from sparse WiFi measurements made at multiple locations inside the home. Although a difficult inverse problem, our implicitly learnt floorplans look promising, and enables forward applications, such as indoor signal prediction and basic ray tracing.



Paperid:68
Authors:Kang-il Lee, Jahyun Koo, Seunghyun Yoon, Minbeom Kim, Hyukhun Koh, Dongryeol Lee, Kyomin Jung
Abstract:
We introduce transductive program synthesis, a new formulation of the program synthesis task that explicitly leverages test inputs during synthesis. While prior approaches to program synthesis—whether based on natural language descriptions or input-output examples—typically aim to generalize from training examples, they often struggle with robustness, especially in real-world settings where training examples are limited and test inputs involve various edge cases. To address this, we propose a novel framework that improves robustness by treating synthesis as an active learning over a finite hypothesis class defined by programs' outputs. We use an LLM to predict outputs for selected test inputs and eliminate inconsistent hypotheses, where the inputs are chosen via a greedy maximin algorithm to minimize the number of LLM queries required. We evaluate our approach on two real-world datasets: Playgol, a string transformation benchmark, and MBPP+, a Python code generation benchmark. We demonstrate that our method significantly improves program synthesis in both accuracy and efficiency.



Paperid:69
Authors:Daniel Melcer, Sujan Kumar Gonugondla, Pramuditha Perera, Haifeng Qian, Wen-Hao Chiang, Yanjun Wang, Nihal Jain, Pranav Garg, Xiaofei Ma, Anoop Deoras
Abstract:
It is common to reject undesired outputs of Large Language Models (LLMs); however, current methods to do so require an excessive amount of computation to re-sample after a rejection, or distort the distribution of outputs by constraining the output to highly improbable tokens.We present a method, Approximately Aligned Decoding (AprAD), to balance the distortion of the output distribution with computational efficiency, inspired by algorithms from the speculative decoding literature.AprAD allows for the generation of long sequences of text with difficult-to-satisfy constraints, while amplifying low probability outputs much less compared to existing methods.We show through a series of experiments that the task-specific performance of AprAD is comparable to methods that do not distort the output distribution, while being much more computationally efficient.



Paperid:70
Authors:Hussen Abu Hamad, Dan Rosenbaum
Abstract:
Neural processes (NPs) are a class of models that learn stochastic processes directly from data and can be used for inference, sampling and conditional sampling. We introduce a new NP model based on flow matching, a generative modeling paradigm that has demonstrated strong performance on various data modalities. Following the NP training paradigm, the model provides amortized predictions of conditional distributions over any arbitrary points in the data. Compared to previous NP models, our model is simple to implement and can be used to sample from conditional distributions using an ODE solver, without requiring auxiliary conditioning methods. In addition, the model provides a controllable tradeoff between accuracy and running time via the number of steps in the ODE solver. We show that our model outperforms previous state-of-the-art methods on various benchmarks including synthetic 1D Gaussian processes data, 2D images, and real-world weather data.



Paperid:71
Authors:Yuning Cui, Wenqi Ren, Alois Knoll
Abstract:
Image restoration aims to recover sharp, high-quality images from degraded, low-quality inputs. Existing methods have progressively advanced from task-specific designs to general architectures, all-in-one frameworks, and composite degradation handling. Despite these advances, computational efficiency remains a critical factor for practical deployment. In this work, we present BioIR, an efficient and universal image restoration framework inspired by the human visual system. Specifically, we design two bio-inspired modules, Peripheral-to-Foveal (P2F) and Foveal-to-Peripheral (F2P), to emulate the perceptual processes of human vision, with a particular focus on the functional interplay between foveal and peripheral pathways. P2F delivers large-field contextual signals to foveal regions based on pixel-to-region affinity, while F2P propagates fine-grained spatial details through a static-to-dynamic two-stage integration strategy. Leveraging the biologically motivated design, BioIR achieves state-of-the-art performance across three representative image restoration settings: single-degradation, all-in-one, and composite degradation. Moreover, BioIR maintains high computational efficiency and fast inference speed, making it highly suitable for real-world applications.



Paperid:72
Authors:Jing Ma, Hanlin Li, Xiang Xiang
Abstract:
Entropy Minimization (EM) is beneficial to reducing class overlap, bridging domain gap, and restricting uncertainty for various tasks in machine learning, yet its potential is limited. To study the internal mechanism of EM, we reformulate and decouple the classical EM into two parts with opposite effects: cluster aggregation driving factor (CADF) rewards dominant classes and prompts a peaked output distribution, while gradient mitigation calibrator (GMC) penalizes high-confidence classes based on predicted probabilities. Furthermore, we reveal the limitations of classical EM caused by its coupled formulation: 1) reward collapse impedes the contribution of high-certainty samples in the learning process, and 2) easy-class bias induces misalignment between output distribution and label distribution. To address these issues, we propose Adaptive Decoupled Entropy Minimization (AdaDEM), which normalizes the reward brought from CADF and employs a marginal entropy calibrator (MEC) to replace GMC. AdaDEM outperforms DEM*, an upper-bound variant of classical EM, and achieves superior performance across various imperfectly supervised learning tasks in noisy and dynamic environments.



Paperid:73
Authors:Josiah Hanna, Nicholas Corrado
Abstract:
Recent work on large language models has demonstrated the use of model-free reinforcement learning (RL) to train reasoning-like capabilities. The emergence of "thinking" through model-free RL is interesting as thinking actions neither produce reward or change the external world state to where it is more likely to get reward. This paper seeks to build a domain-independent understanding of when model-free RL will lead to "thinking" as a strategy for reward maximization. To build this understanding, we first introduce a theoretical model which we call a \textit{thought Markov decision process}. Thought MDPs minimally extend the classical MDP model to include an abstract notion of thought state and thought action. Using the thought MDP model, we prove the importance of policy initialization in determining whether thinking emerges or not and show formally that thought actions are equivalent to the agent choosing to perform a step of policy improvement before continuing to act. We then show that open-source LLMs satisfy the conditions that our theory predicts are necessary for model-free RL to produce thinking-like behavior. Finally, we hypothesize about conditions that would enable thinking to be learned outside of language generation and introduce a toy domain where a combination of multi-task pre-training and designated thought actions enable more data-efficient RL compared to non-thinking agents.



Paperid:74
Authors:Emile van Krieken, Pasquale Minervini, Edoardo Maria Ponti, Antonio Vergari
Abstract:
Neurosymbolic (NeSy) predictors combine neural perception with symbolic reasoning to solve tasks like visual reasoning. However, standard NeSy predictors assume conditional independence between the symbols they extract, thus limiting their ability to model interactions and uncertainty --- often leading to overconfident predictions and poor out-of-distribution generalisation. To overcome the limitations of the independence assumption, we introduceneurosymbolic diffusion models(NeSyDMs), a new class of NeSy predictors that use discrete diffusion to model dependencies between symbols. Our approach reuses the independence assumption from NeSy predictors at each step of the diffusion process, enabling scalable learning while capturing symbol dependencies and uncertainty quantification. Across both synthetic and real-world benchmarks — including high-dimensional visual path planning and rule-based autonomous driving — NeSyDMs achieve state-of-the-art accuracy among NeSy predictors and demonstrate strong calibration.



Paperid:75
Authors:Nathan Wycoff
Abstract:
Regression trees have emerged as a preeminent tool for solving real-world regression problems due to their ability to deal with nonlinearities, interaction effects and sharp discontinuities. In this article, we rather study regression trees applied to well-behaved, differentiable functions, and determine the relationship between node parameters and the local gradient of the function being approximated. We find a simple estimate of the gradient which can be efficiently computed using quantities exposed by popular tree learning libraries. This allows the tools developed in the context of differentiable algorithms, like neural nets and Gaussian processes, to be deployed to tree-based models. To demonstrate this, we study measures of model sensitivity defined in terms of integrals of gradients and demonstrate how to compute them for regression trees using the proposed gradient estimates. Quantitative and qualitative numerical experiments reveal the capability of gradients estimated by regression trees to improve predictive analysis, solve tasks in uncertainty quantification, and provide interpretation of model behavior.



Paperid:76
Authors:Adiba Ejaz, Elias Bareinboim
Abstract:
Abstract:Greedy Equivalence Search (GES) is a classic score-based algorithm for causal discovery from observational data.In the sample limit, it recovers the Markov equivalence class of graphs that describe the data.Still, it faces two challenges in practice: computational cost and finite-sample accuracy.In this paper, we develop Less Greedy Equivalence Search (LGES), a variant of GES that retains its theoretical guarantees while partially addressing these limitations.LGES modifies the greedy step: rather than always applying the highest-scoring insertion, it avoids edge insertions between variables for which the score implies some conditional independence.This more targeted search yields up to a $10$-fold speed-up and $3$-fold reduction in structural error relative to GES.Moreover, LGES can guide the search using prior assumptions, while correcting these assumptions when contradicted by the data.Finally, LGES can exploit interventional data to refine the learned observational equivalence class.We prove that LGES recovers the true equivalence class in the sample limit from observational and interventional data, even with misspecified prior assumptions.Experiments demonstrate that LGES outperforms GES and other baselines in speed, accuracy, and robustness to misspecified assumptions.



Paperid:77
Authors:Chuanyang Zheng, Jiankai Sun, Yihang Gao, Yuehao Wang, Peihao Wang, Jing Xiong, Liliang Ren, Hao Cheng, Janardhan Kulkarni, yelong shen, Zhangyang "Atlas" Wang, Mac Schwager, Anderson Schneider, Xiaodong Liu, Jianfeng Gao
Abstract:
The attention mechanism is a core component of the Transformer architecture.Various methods have been developed to compute attention scores, including multi-head attention (MHA), multi-query attention, group-query attention and so on. We further analyze the MHA and observe that its performance improves as the number of attention heads increases, provided the hidden size per head remains sufficiently large. Therefore, increasing both the head count and hidden size per head with minimal parameter overhead can lead to significant performance gains at a low cost.Motivated by this insight, we introduce Simulated Attention Score (SAS), whichmaintains a compact model size while simulating a larger number of attention heads and hidden feature dimension per head.This is achieved by projecting a low-dimensional head representation into a higher-dimensional space, effectively increasing attention capacity without increasing parameter count. Beyond the head representations, we further extend the simulation approach to feature dimension of the key and query embeddings, enhancing expressiveness by mimicking the behavior of a larger model while preserving the original model size.To control the parameter cost, we also propose Parameter-Efficient Attention Aggregation (PEAA).Comprehensive experiments on a variety of datasets and tasks demonstrate the effectiveness of the proposed SAS method, achieving significant improvements over different attention variants.



Paperid:78
Authors:Aaron Sun, Subhransu Maji, Grant Van Horn
Abstract:
In the single-positive multi-label (SPML) setting, each image in a dataset is labeled with the presence of a single class, while the true presence of other classes remains unknown. The challenge is to narrow the performance gap between this partially-labeled setting and fully-supervised learning, which often requires a significant annotation budget. Prior SPML methods were developed and benchmarked on synthetic datasets created by randomly sampling single positive labels from fully-annotated datasets like Pascal VOC, COCO, NUS-WIDE, and CUB200. However, this synthetic approach does not reflect real-world scenarios and fails to capture the fine-grained complexities that can lead to difficult misclassifications. In this work, we introduce the ML48S dataset, a fine-grained, real-world multi-label dataset derived from recordings of bird sounds. ML48S provides a natural SPML setting with single-positive annotations on a challenging, fine-grained domain, as well as two extended settings in which domain priors give access to additional negative labels. We benchmark existing SPML methods on ML48S and observe significant performance differences compared to synthetic datasets and analyze method weaknesses, underscoring the need for more realistic and difficult benchmarks.



Paperid:79
Authors:Rattana Pukdee, Ziqi Ke, Chirag Gupta
Abstract:
We study the problem of regression with interval targets, where only upper and lower bounds on target values are available in the form of intervals. This problem arises when the exact target label is expensive or impossible to obtain, due to inherent uncertainties. In the absence of exact targets, traditional regression loss functions cannot be used. First, we study the methodology of using a loss functions compatible with interval targets, for which we establish non-asymptotic generalization bounds based on smoothness of the hypothesis class that significantly relaxing prior assumptions of realizability and small ambiguity degree. Second, we propose a novel min-max learning formulation: minimize against the worst-case (maximized) target labels within the provided intervals. The maximization problem in the latter is non-convex, but we show that good performance can be achieved with the incorporation of smoothness constraints. Finally, we perform extensive experiments on real-world datasets and show that our methods achieve state-of-the-art performance.



Paperid:80
Authors:Cheng Yu, Benedikt Stroebl, Diyi Yang, Orestis Papakyriakopoulos
Abstract:
Abstract:As retrieval-augmented AI agents become more embedded in society, their safety properties and ethical behavior remain insufficiently understood. In particular, the growing integration of LLMs and AI agents raises critical questions about how they engage with and are influenced by their environments.This study investigates how expanding retrieval access—from no external sources to Wikipedia-based retrieval and open web search—affects model reliability, bias propagation, and harmful content generation. Through extensive benchmarking of censored and uncensored LLMs and AI Agents, our findings reveal a consistent degradation in refusal rates, bias sensitivity, and harmfulness safeguards as models gain broader access to external sources, culminating in a phenomenon we term $\textbf{\textit{safety devolution}}$. Notably, retrieval-augmented agents built on aligned LLMs often behave more unsafely than uncensored models without retrieval. This effect persists even under strong retrieval accuracy and prompt-based mitigation, suggesting that the mere presence of retrieved content reshapes model behavior in structurally unsafe ways.These findings underscore the need for robust mitigation strategies to ensure fairness and reliability in retrieval-augmented and increasingly autonomous AI systems.



Paperid:81
Authors:chao zhou, Advait Gadhikar, Tom Jacobs, Rebekka Burkholz
Abstract:
Fine-tuning large pre-trained neural networks is known to be resource-intensive, both in terms of memory and computational cost. To mitigate this, a common approach is to restrict training to a subset of the model parameters. By analyzing the relationship between gradients and weights during fine-tuning, we observe a notable pattern: large gradients are often associated with small-magnitude weights. This correlation is more pronounced in fine-tuning settings than in training from scratch. Motivated by this observation, we propose \textsc{NanoAdam}, which dynamically updates only the small-magnitude weights during fine-tuning and offers several practical advantages: first, the criterion is \emph{gradient-free}—the parameter subset can be determined without gradient computation; second, it preserves large-magnitude weights, which are likely to encode critical features learned during pre-training, thereby reducing the risk of catastrophic forgetting; thirdly, it permits the use of larger learning rates and consistently leads to better generalization performance in experiments. We demonstrate this for both NLP and vision tasks.



Paperid:82
Authors:Alireza Bakhtiari, Alex Ayoub, Samuel Robertson, David Janz, Csaba Szepesvari
Abstract:
We establish a lower bound on the eluder dimension in generalised linear model classes, showing that standard eluder dimension-based analysis cannot lead to first-order regret bounds. To address this, we introduce a localisation method for the eluder dimension; our analysis immediately recovers and improves on classic results for Bernoulli bandits, and allows for the first genuine first-order bounds for finite-horizon reinforcement learning tasks with bounded cumulative returns.



Paperid:83
Authors:Mattie Ji, Amauri Souza, Vikas Garg
Abstract:
Including intricate topological information (e.g., cycles) provably enhances the expressivity of message-passing graph neural networks (GNNs) beyond the Weisfeiler-Leman (WL) hierarchy. Consequently, Persistent Homology (PH) methods are increasingly employed for graph representation learning. In this context, recent works have proposed decorating classical PH diagrams with vertex and edge features for improved expressivity. However, due to their dependence on features, these methods still fail to capture basic graph structural information. In this paper, we propose SpectRe --- a new topological descriptor for graphs that integrates spectral information into PH diagrams. Notably, SpectRe is strictly more expressive than existing descriptors on graphs. We also introduce notions of global and local stability to analyze existing descriptors and establish that SpectRe is locally stable. Finally, experiments on synthetic and real-world datasets demonstrate the effectiveness of SpectRe and its potential to enhance the capabilities of graph models in relevant learning tasks.



Paperid:84
Authors:Ilias Diakonikolas, Jingyi Gao, Daniel Kane, Sihan Liu, Christopher Ye
Abstract:
We initiate a systematic investigation of distribution testing in the framework of algorithmic replicability. Specifically, given independent samples from a collection of probability distributions, the goal is to characterize the sample complexity of replicably testing natural properties of the underlying distributions. On the algorithmic front, we develop new replicable algorithms for testing closeness and independence of discrete distributions. On the lower bound front, we develop a new methodology for proving sample complexity lower bounds for replicable testing that may be of broader interest. As an application of our technique, we establish near-optimal sample complexity lower bounds for replicable uniformity testing---answering an open question from prior work---and closeness testing.



Paperid:85
Authors:Shuze Chen, Tianyi Peng
Abstract:
Abstract:Value decomposition has long been a fundamental technique in multi-agent reinforcement learning and dynamic programming. Specifically, the value function of a global state $(s_1,s_2,\ldots,s_N)$ is often approximated as the sum of local functions: $V(s_1,s_2,\ldots,s_N)\approx\sum_{i=1}^N V_i(s_i)$. This approach has found various applications in modern RL systems. However, the theoretical justification for why this decomposition works so effectively remains underexplored. In this paper, we uncover the underlying mathematical structure that enables value decomposition. We demonstrate that a Markov decision process (MDP) permits value decomposition *if and only if* its transition matrix is not "entangled"—a concept analogous to quantum entanglement in quantum physics. Drawing inspiration from how physicists measure quantum entanglement, we introduce how to measure the "Markov entanglement" and show that this measure can be used to bound the decomposition error in general multi-agent MDPs. Using the concept of Markov entanglement, we proved that a widely-used class of policies, the index policy, is weakly-entangled and enjoys a sublinear $\mathcal O(\sqrt{N})$ scale of decomposition error for $N$-agent systems. Finally, we show Markov entanglement can be efficiently estimated, guiding practitioners on the feasibility of value decomposition.



Paperid:86
Authors:Matthew Macfarlane, Clem Bonnet
Abstract:
General intelligence requires systems that acquire new skills efficiently and generalize beyond their training distributions.Although program synthesis approaches have strong generalization power, they face scaling issues due to large combinatorial spaces that quickly make them impractical and require human-generated DSLs or pre-trained priors to narrow this search space.On the other hand, deep learning methods have had high successes, but they lack structured test-time adaptation and rely on heavy stochastic sampling or expensive gradient updates for fine-tuning.In this work, we propose the Latent Program Network (LPN), a new architecture that builds in test-time search directly into neural models.LPN learns a latent space of implicit programs---neurally mapping inputs to outputs---through which it can search using gradients at test time.LPN combines the adaptability of symbolic approaches and the scalability of neural methods.It searches through a compact latent space at test time and bypasses the need for pre-defined domain-specific languages.On a range of programming-by-examples tasks, LPN either outperforms or matches performance compared to in-context learning and test-time training methods.Tested on the ARC-AGI benchmark, we demonstrate that LPN can both learn a compact program space and search through it at test time to adapt to novel tasks.LPN doubles its performance on out-of-distribution tasks when test-time search is switched on.



Paperid:87
Authors:Charig Yang, Samiul Alam, Shakhrul Iman Siam, Michael Proulx, Lambert Mathias, Kiran Somasundaram, Luis Pesqueira, James Fort, Sheroze Sheriffdeen, Omkar Parkhi, Yuheng Ren, Mi Zhang, Yuning Chai, Richard Newcombe, Hyo Jin Kim
Abstract:
To enable egocentric contextual AI in always-on smart glasses, it is crucial to be able to keep a record of the user's interactions with the world, including during reading. In this paper, we introduce a new task of reading recognition to determine when the user is reading. We first introduce the first-of-its-kind large-scale multimodal Reading in the Wild dataset, containing 100 hours of reading and non-reading videos in diverse and realistic scenarios. We then identify three modalities (egocentric RGB, eye gaze, head pose) that can be used to solve the task, and present a flexible transformer model that performs the task using these modalities, either individually or combined. We show that these modalities are relevant and complementary to the task, and investigate how to efficiently and effectively encode each modality. Additionally, we show the usefulness of this dataset towards classifying types of reading, extending current reading understanding studies conducted in constrained settings to larger scale, diversity and realism. Code, model, and data will be public.



Paperid:88
Authors:Andrew Liu, Axel Elaldi, Nicholas Franklin, Nathan Russell, Gurinder Atwal, Yih-En Ban, Olivia Viessmann
Abstract:
Invariant Point Attention (IPA) is a key algorithm for geometry-aware modeling in structural biology, central to many protein and RNA models. However, its quadratic complexity limits the input sequence length. We introduce FlashIPA, a factorized reformulation of IPA that leverages hardware-efficient FlashAttention to achieve linear scaling in GPU memory and wall-clock time with sequence length. FlashIPA matches or exceeds standard IPA performance while substantially reducing computational costs. FlashIPA extends training to previously unattainable lengths, and we demonstrate this by re-training generative models without length restrictions and generating structures of thousands of residues. FlashIPA is available at https://anonymous.4open.science/r/flash_ipa-07CE.



Paperid:89
Authors:Sourav Pal, Kamyar Azizzadenesheli, Vikas Singh
Abstract:
The growing body of work on Physics-Informed Neural Networks (PINNs) seeks to use machine learning strategies to improve methods for solution discovery of Partial Differential Equations (PDEs). While classical solvers may remain the preferred tool of choice in the short-term, PINNs can be viewed as complementary. The expectation is that in some specific use cases, they can even be effective, standalone. A key step in training PINNs is selecting domain points for loss evaluation, where Monte Carlo sampling remains the dominant but often suboptimal in low dimension settings, common in physics. We leverage recent advances in asymptotic expansions of quadrature nodes and weights (for weight functions belonging to the modified Gauss-Jacobi family) together with suitable adjustments for parameterization towards a data-driven framework for learnable quadrature rules. A direct benefit is a performance improvement of PINNs, relative to existing alternatives, on a wide range of problems studied in the literature. Beyond finding a standard solution for an instance of a single PDE, our construction enables learning rules to predict solutions for a given family of PDEs via hyper-networks, a useful capability for PINNs.



Paperid:90
Authors:Wenyuan Li, Guang Li, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama
Abstract:
To address the computational and storage challenges posed by large-scale datasets in deep learning, dataset distillation has been proposed to synthesize a compact dataset that replaces the original while maintaining comparable model performance. Unlike optimization-based approaches that require costly bi-level optimization, distribution matching (DM) methods improve efficiency by aligning the distributions of synthetic and original data, thereby eliminating nested optimization. DM achieves high computational efficiency and has emerged as a promising solution. However, existing DM methods, constrained to Euclidean space, treat data as independent and identically distributed points, overlooking complex geometric and hierarchical relationships. To overcome this limitation, we propose a novel hyperbolic dataset distillation method, termed HDD. Hyperbolic space, characterized by negative curvature and exponential volume growth with distance, naturally models hierarchical and tree-like structures. HDD embeds features extracted by a shallow network into the Lorentz hyperbolic space, where the discrepancy between synthetic and original data is measured by the hyperbolic (geodesic) distance between their centroids. By optimizing this distance, the hierarchical structure is explicitly integrated into the distillation process, guiding synthetic samples to gravitate towards the root-centric regions of the original data distribution while preserving their underlying geometric characteristics. Furthermore, we find that pruning in hyperbolic space requires only 20% of the distilled core set to retain model performance, while significantly improving training stability. Notably, HDD is seamlessly compatible with most existing DM methods, and extensive experiments on different datasets validate its effectiveness. To the best of our knowledge, this is the first work to incorporate the hyperbolic space into the dataset distillation process. The code will be released upon acceptance.



Paperid:91
Authors:Shen Nie, Fengqi Zhu, Zebin You, Xiaolu Zhang, Jingyang Ou, Jun Hu, Jun Zhou, Yankai Lin, Ji-Rong Wen, Chongxuan LI
Abstract:
The capabilities of large language models (LLMs) are widely regarded as relying on autoregressive models (ARMs). We challenge this notion by introducingLLaDA, a diffusion model trained from scratch under the pre-training and supervised fine-tuning (SFT) paradigm. LLaDA employs a forward data masking process and a reverse generation process, parameterized by a Transformer to predict masked tokens. It provides a principled generative approach for probabilistic inference by optimizing a likelihood lower bound. Across extensive benchmarks on general tasks, math, code, and so on, LLaDA demonstrates strongscalabilityand performs comparably to our self-constructed ARM baselines. Remarkably, LLaDA 8B is competitive with strong LLMs like LLaMA3 8B inin-context learningand, after SFT, exhibits impressiveinstruction-followingabilities in case studies such as multi-turn dialogue. Moreover, LLaDA addresses the reversal curse, surpassing GPT-4o in a reversal poem completion task. Our findings show the promise of diffusion models for language modeling at scale and challenge the common assumption that core LLM capabilities discussed above inherently depend on ARMs.



Paperid:92
Authors:Antonio Orvieto, Robert Gower
Abstract:
Abstract:Understanding the remarkable efficacy of Adam when training transformer-based language models has become a central research topic within the optimization community. To gain deeper insights, several simplifications of Adam have been proposed, such as the signed gradient and signed momentum methods. In this work, we conduct an extensive empirical study — training over 1,300 language models across different data configurations and scales — comparing Adam to several known simplified variants. We find that signed momentum methods consistently underperform relative to Adam, even after careful tuning of momentum parameters, clipping setting and learning rates. However, our analysis reveals a compelling option that preserves optimal performance while allowing for new insightful reformulations: constraining the Adam momentum parameters to be equal, $\beta_1=\beta_2$. Beyond robust performance, this choice affords new theoretical insights, directly highlights the ``secret sauce'' on top of signed momentum, and grants a precise statistical interpretation: we show that Adam in this setting implements a natural online algorithm for estimating the mean and variance of gradients—one that arises from a mean-field Gaussian variational inference perspective.



Paperid:93
Authors:Anastasios Manganaris, Vittorio Giammarino, Ahmed Qureshi
Abstract:
Real-world robotic tasks often require agents to achieve sequences of goals while respecting time-varying safety constraints. However, standard Reinforcement Learning (RL) paradigms are fundamentally limited in these settings. A natural approach to these problems is to combine RL with Linear-time Temporal Logic (LTL), a formal language for specifying complex, temporally extended tasks and safety constraints. Yet, existing RL methods for LTL objectives exhibit poor empirical performance in complex and continuous environments. As a result, no scalable methods support both temporally ordered goals and safety simultaneously, making them ill-suited for realistic robotics scenarios. We propose Automaton Constrained Q-Learning (ACQL), an algorithm that addresses this gap by combining goal-conditioned value learning with automaton-guided reinforcement. ACQL supports most LTL task specifications and leverages their automaton representation to explicitly encode stage-wise goal progression and both stationary and non-stationary safety constraints. We show that ACQL outperforms existing methods across a range of continuous control tasks, including cases where prior methods fail to satisfy either goal-reaching or safety constraints. We further validate its real-world applicability by deploying ACQL on a 6-DOF robotic arm performing a goal-reaching task in a cluttered, cabinet-like space with safety constraints. Our results demonstrate that ACQL is a robust and scalable solution for learning robotic behaviors according to rich temporal specifications.



Paperid:94
Authors:Jiacheng Fu, Yue Li, Xin Dong, Wenming Weng, Yueyi Zhang, Zhiwei Xiong
Abstract:
Event-based structured light (SL) systems have attracted increasing attention for their potential in high-performance 3D measurement. Despite the inherent HDR capability of event cameras, reflective and absorptive surfaces still cause event cluttering and absence, which produce overexposed and underexposed regions that degrade the reconstruction quality. In this work, we present the first HDR 3D measurement framework specifically designed for event-based SL systems. First, we introduce a multi-contrast HDR coding strategy that facilitates imaging of areas with different reflectance. Second, to alleviate inter-frame interference caused by overexposed and underexposed areas, we propose a universal confidence-driven stereo matching strategy. Specifically, we estimate a confidence map as the fusion weight for features via an energy-guided confidence estimation. Further, we propose the confidence propagation volume, an innovative cost volume that offers both effective suppression of inter-frame interference and strong representation capability. Third, we contribute an event-based SL simulator and propose the first event-based HDR SL dataset. We also collect a real-world benchmarking dataset with ground truth. We validate the effectiveness of our method with the proposed confidence-driven strategy on both synthetic and real-world datasets. Experimental results demonstrate that our proposed HDR framework enables accurate 3D measurement even under extreme conditions.



Paperid:95
Authors:Hadi Vafaii, Dekel Galor, Jacob Yates
Abstract:
Abstract:Inference in both brains and machines can be formalized by optimizing a shared objective: maximizing the evidence lower bound (ELBO) in machine learning, or minimizing variational free energy ($\mathcal{F}$) in neuroscience (ELBO = $-\mathcal{F}$). While this equivalence suggests a unifying framework, it leaves open how inference is implemented in neural systems. Here, we show that online natural gradient descent on $\mathcal{F}$, under Poisson assumptions, leads to a recurrent spiking neural network that performs variational inference via membrane potential dynamics. The resulting model—the iterative Poisson variational autoencoder (iP-VAE)—replaces the encoder network with local updates derived from natural gradient descent on $\mathcal{F}$. Theoretically, iP-VAE yields a number of desirable features such as emergent normalization via lateral competition, and hardware-efficient integer spike count representations. Empirically, iP-VAE outperforms both standard VAEs and Gaussian-based predictive coding models in sparsity, reconstruction, and biological plausibility. iP-VAE also exhibits strong generalization to out-of-distribution inputs, exceeding hybrid iterative-amortized VAEs. These results demonstrate how deriving inference algorithms from first principles can yield concrete architectures that are simultaneously biologically plausible and empirically effective.



Paperid:96
Authors:Zekun CAI, Yiheng YAO, Guangji Bai, Renhe Jiang, Xuan Song, Ryosuke Shibasaki, Liang Zhao
Abstract:
Real-world data distributions often shift continuously across multiple latent factors such as time, geography, and socioeconomic context. However, existing domain generalization approaches typically treat domains as discrete or evolving along a single axis (e.g., time), which fails to capture the complex, multi-dimensional nature of real-world variation. This paper introduces the task of Continuous Domain Generalization (CDG), which aims to generalize predictive models to unseen domains defined by arbitrary combinations of continuous variation descriptors. We present a principled framework grounded in geometric and algebraic theory, showing that optimal model parameters across domains lie on a low-dimensional manifold. To model this structure, we propose a Neural Lie Transport Operator (NeuralLTO), which enables structured parameter transitions by enforcing geometric continuity and algebraic consistency. To handle noisy or incomplete domain descriptors, we introduce a gating mechanism to suppress irrelevant dimensions and a local chart-based strategy for robust generalization. Extensive experiments on synthetic and real-world datasets—including remote sensing, scientific documents, and traffic forecasting—demonstrate that our method significantly outperforms existing baselines in generalization accuracy and robustness under descriptor imperfections. Code is available at: \url{https://github.com/dfwsds/CDG}.



Paperid:97
Authors:Claire Vernade, Solenne Gaucher, Nicolas Nguyen
Abstract:
Abstract:We study the problem of non-stationary Lipschitz bandits, where the number of actions is infinite and the reward function, satisfying a Lipschitz assumption, can change arbitrarily over time. We design an algorithm that adaptively tracks the recently introduced notion of significant shifts, defined by large deviations of the cumulative reward function. To detect such reward changes, our algorithm leverages a hierarchical discretization of the action space. Without requiring any prior knowledge of the non-stationarity, our algorithmachieves a minimax-optimal dynamic regret bound of $\mathcal{\widetilde{O}}(\tilde{L}^{1/3}T^{2/3})$, where $\tilde{L}$ is the number of significant shifts and $T$ the horizon. This result provides the first optimal guarantee in this setting.



Paperid:98
Authors:Minghao Yang, Pengteng Li, Yan Liang, Qianyi Cai, Zhihang Zheng, Shichen Zhang, Pengfei ZHANG, Zhi-An Huang, Hui Xiong
Abstract:
Deep learning techniques have driven significant progress in various analytical tasks within 3D genomics in computational biology. However, a holistic understanding of 3D genomics knowledge remains underexplored. Here, we proposeMIX-HIC, the first multimodal foundation model of 3D genome that integrates both 3D genome structure and epigenomic tracks, which obtains unified and comprehensive semantics. For accurate heterogeneous semantic fusion, we design the cross-modal interaction and mapping blocks for robust unified representation, yielding the accurate aggregation of 3D genome knowledge. Besides, we introduce the first large-scale dataset comprising over1 millionpairwise samples of Hi-C contact maps and epigenomic tracks for high-quality pre-training, enabling the exploration of functional implications in 3D genomics. Extensive experiments show that MIX-HIC significantly surpasses existing state-of-the-art methods in diverse downstream tasks. This work provides a valuable resource for advancing 3D genomics research.



Paperid:99
Authors:Xidong Feng, Vivek Veeriah, Marcus Chiam, Michael Dennis, Federico Barbero, Johan Obando Ceron, Jiaxin Shi, Satinder Singh, Shaobo Hou, Nenad Tomasev, Tom Zahavy
Abstract:
While Generative AI rapidly advances in various domains, generating truly creative, aesthetic, and counter-intuitive outputs remains a challenge. This paper presents an approach to tackle these difficulties in the domain of chess puzzles. We start by benchmarking Generative AI architectures, and then introduce an RL framework with novel rewards based on chess engine search statistics to overcome some of those shortcomings. The rewards are designed to enhance a puzzle's uniqueness, counter-intuitiveness, diversity, and realism. Our RL approach dramatically increases counter-intuitive puzzle generation by 10x, from 0.22\% (supervised) to 2.5\%, surpassing existing dataset rates (2.1\%) and the best Lichess-trained model (0.4\%). Our puzzles meet novelty and diversity benchmarks, retain aesthetic themes, and are rated by human experts as more creative, enjoyable, and counter-intuitive than composed book puzzles, even approaching classic compositions. Our final outcome is a curated \href{https://drive.google.com/file/d/1THSJ0QSAY9bHTDuR3zKs3Y8ticBYTDn1/view?usp=sharing}{booklet} of these novel AI-generated puzzles.



Paperid:100
Authors:ZiDong Wang, LEI BAI, Xiangyu Yue, Wanli Ouyang, Yiyuan Zhang
Abstract:
We introduce native-resolution image synthesis, a novel paradigm in generative modeling capable of synthesizing images at arbitrary resolutions and aspect ratios. This approach overcomes the limitations of standard fixed-resolution, square-image methods by inherently handling variable-length visual tokens—a core challenge for conventional techniques. To this end, we propose the Native-resolution diffusion Transformer (NiT), an architecture that explicitly models varying resolutions and aspect ratios within its denoising process. Unconstrained by fixed formats, NiT learns intrinsic visual distributions from images encompassing a wide range of resolutions and aspect ratios. Notably, a single NiT model simultaneously achieves the state-of-the-art performance on both ImageNet-256x256 and 512x512 benchmarks. Surprisingly, akin to the robust zero-shot capabilities seen in advanced Large Language Models, NiT, pretrained solely on ImageNet, demonstrates excellent zero-shot generalization performance. It successfully generates high-fidelity images at previously unseen high resolutions (e.g., 1024x1024, 1536x1536) and diverse aspect ratios (e.g., 16:9,3:1, 4:3), as shown in Figure 1. These findings indicate the significant potential of native-resolution modeling as a bridge between visual generative modeling and advanced LLM methodologies. Code and pretrained models will be made publicly available.



Paperid:101
Authors:Litu Rout, Constantine Caramanis, Sanjay Shakkottai
Abstract:
Diffusion Language Models (DLMs) promise parallel generation and bidirectional context, yet they underperform autoregressive (AR) models in bothlikelihood modelingandgenerated text quality. We identify that this performance gap arises when important tokens (e.g., key words or low-frequency words that anchor a sentence) are masked early in the forward process, limiting contextual information for accurate reconstruction. To address this, we introduce theAnchored Diffusion Language Model (ADLM), a novel two-stage framework that first predicts distributions over important tokens via an anchor network, and then predicts the likelihoods of missing tokens conditioned on the anchored predictions. ADLM significantly improves test perplexity on LM1B and OpenWebText, achieving up to 25.4\% gains over prior DLMs, and narrows the gap with strong AR baselines. It also achieves state-of-the-art zero-shot generalization across seven benchmarks and surpasses AR models in MAUVE score, which marks the first time a DLM generates better human-like text than an AR model. Theoretically, we derive an Anchored Negative Evidence Lower Bound (ANELBO) objective and show that anchoring improves sample complexity and likelihood modeling. Beyond diffusion, anchoring boosts performance in AR models and enhances reasoning in math and logic tasks, outperforming existing chain-of-thought approaches.



Paperid:102
Authors:Qihang Zhou, ShenHao Fang, Shibo He, Wenchao Meng, Jiming Chen
Abstract:
Condensing large datasets into smaller synthetic counterparts has demonstrated its promise for image classification. However, previous research has overlooked a crucial concern in image recognition: ensuring that models trained on condensed datasets are unbiased towards protected attributes (PA), such as gender and race. Our investigation reveals that dataset distillation (DD) fails to alleviate the unfairness towards minority groups within original datasets. Moreover, this bias typically worsens in the condensed datasets due to their smaller size. To bridge the research gap, we propose a novel fair dataset distillation (FDD) framework, namely FairDD, which can be seamlessly applied to diverse matching-based DD approaches, requiring no modifications to their original architectures. The key innovation of FairDD lies in synchronously matching synthetic datasets to PA-wise groups of original datasets, rather than indiscriminate alignment to the whole distributions in vanilla DDs, dominated by majority groups. This synchronized matching allows synthetic datasets to avoid collapsing into majority groups and bootstrap their balanced generation to all PA groups. Consequently, FairDD could effectively regularize vanilla DDs to favor biased generation toward minority groups while maintaining the accuracy of target attributes. Theoretical analyses and extensive experimental evaluations demonstrate that FairDD significantly improves fairness compared to vanilla DD methods, with a promising trade-off between fairness and accuracy. Its consistent superiority across diverse DDs, spanning Distribution and Gradient Matching, establishes it as a versatile FDD approach.



Paperid:103
Authors:Jialong Chen, Tong Wang, Bowen Deng, Luonan Chen, Zibin Zheng, Chuan Chen
Abstract:
Inspired by the workings of biological brains, humans have designed artificial neural networks (ANNs), sparking profound advancements across various fields. However, the biological brain possesses high plasticity, enabling it to develop simple, efficient, and powerful structures to cope with complex external environments. In contrast, the superior performance of ANNs often relies on meticulously crafted architectures, which can make them vulnerable when handling complex inputs. Moreover, overparameterization often characterizes the most advanced ANNs. This paper explores the path toward building streamlined and plastic ANNs. Firstly, we introduce the Graph Perceptron (GP), which extends the most fundamental ANN, the Multi-Layer Perceptron (MLP). Subsequently, we incorporate a self-assembly mechanism on top of GP called Self-Assembling Graph Perceptron (SAGP). During training, SAGP can autonomously adjust the network's number of neurons and synapses and their connectivity. SAGP achieves comparable or even superior performance with only about 5% of the size of an MLP. We also demonstrate the SAGP's advantages in enhancing model interpretability, feature selection, and parameter pruning.



Paperid:104
Authors:Jueyu Chen, Wantao Wen, Yeqiang Wang, Erliang Lin, Yemin Wang, Yuheng Jia
Abstract:
Learning from Label Proportions (LLP) is a weakly-supervised paradigm that uses bag-level label proportions to train instance-level classifiers, offering a practical alternative to costly instance-level annotation. However, the weak supervision makes effective training challenging, and existing methods often rely on pseudo-labeling, which introduces noise. To address this, we propose RLPL, a two-stage framework. In the first stage, we use unsupervised contrastive learning to pretrain the encoder and train an auxiliary classifier with bag-level supervision. In the second stage, we introduce an LLP-OTD mechanism to refine pseudo labels and split them into high- and low-confidence sets. These sets are then used in LLPMix to train the final classifier. Extensive experiments and ablation studies on multiple benchmarks demonstrate that RLPL achieves comparable state-of-the-art performance and effectively mitigates pseudo-label noise.



Paperid:105
Authors:Shashi Kumar, Yacouba Kaloga, John Mitros, Petr Motlicek, Ina Kodrasi
Abstract:
Low-rank adaptation (LoRA) is a widely used method for parameter-efficient finetuning.However, existing LoRA variants lack mechanisms to explicitly disambiguate task-relevant information within the learned low-rank subspace, potentially limiting downstream performance. We propose Factorized Variational Autoencoder LoRA (FVAE-LoRA), which leverages a VAE to learn two distinct latent spaces.Our novel Evidence Lower Bound formulation explicitly promotes factorization between the latent spaces, dedicating one latent space to task-salient features and the other to residual information.Extensive experiments on text, audio, and image tasks demonstrate that FVAE-LoRA consistently outperforms standard LoRA. Moreover, spurious correlation evaluations confirm that FVAE-LoRA better isolates task-relevant signals, leading to improved robustness under distribution shifts.



Paperid:106
Authors:Nic Fishman, Gokul Gowri, Peng Yin, Jonathan Gootenberg, Omar Abudayyeh
Abstract:
Abstract:Many real-world problems require reasoning across multiple scales, demanding models which operate not on single data points, but on entire distributions. We introduce generative distribution embeddings (GDE), a framework that lifts autoencoders to the space of distributions. In GDEs, an encoder acts on sets of samples, and the decoder is replaced by a generator which aims to match the input distribution. This framework enables learning representations of distributions by coupling conditional generative models with encoder networks which satisfy a criterion we call distributional invariance. We show that GDEs learn predictive sufficient statistics embedded in the Wasserstein space, such that latent GDE distances approximately recover the $W_2$ distance, and latent interpolation approximately recovers optimal transport trajectories for Gaussian and Gaussian mixture distributions. We systematically benchmark GDEs against existing approaches on synthetic datasets, demonstrating consistently stronger performance. We then apply GDEs to six key problems in computational biology: learning representations of cell populations from lineage-tracing data (150K cells), predicting perturbation effects on single-cell transcriptomes (1M cells), predicting perturbation effects on cellular phenotypes (20M single-cell images), modeling tissue-specific DNA methylation patterns (253M sequences), designing synthetic yeast promoters (34M sequences), and spatiotemporal modeling of viral protein sequences (1M sequences).



Paperid:107
Authors:Haoqian Liang, Xiaohui Wang, Zhichao Li, Ya Yang, Naiyan Wang
Abstract:
Object concepts play a foundational role in human visual cognition, enabling perception, memory, and interaction in the physical world. Inspired by findings in developmental neuroscience—where infants are shown to acquire object understanding through observation of motion—we propose a biologically inspired framework for learning object-centric visual representations in an unsupervised manner.Our key insight is that motion boundary serves as a strong signal for object-level grouping, which can be used to derive pseudo instance supervision from raw videos. Concretely, we generate motion-based instance masks using off-the-shelf optical flow and clustering algorithms, and use them to train visual encoders via contrastive learning. Our framework is fully label-free and does not rely on camera calibration, making it scalable to large-scale unstructured video data.We evaluate our approach on three downstream tasks spanning both low-level (monocular depth estimation) and high-level (3D object detection and occupancy prediction) vision. Our models outperform previous supervised and self-supervised baselines and demonstrate strong generalization to unseen scenes. These results suggest that motion-induced object representations offer a compelling alternative to existing vision foundation models, capturing a crucial but overlooked level of abstraction: the visual instance.The corresponding code will be released upon paper acceptance.



Paperid:108
Authors:Berker Demirel, Marco Fumero, Francesco Locatello
Abstract:
Deep learning systems deployed in real-world applications often encounter data that is different from their in-distribution (ID). A reliable model should ideally abstain from making decisions in this out-of-distribution (OOD) setting. Existing state-of-the-art methods primarily focus on feature distances, such as k-th nearest neighbors and distances to decision boundaries, either overlooking or ineffectively using in-distribution statistics. In this work, we propose a novel angle-based metric for OOD detection that is computed relative to the in-distribution structure. We demonstrate that the angles between feature representations and decision boundaries, viewed from the mean of in-distribution features, serve as an effective discriminative factor between ID and OOD data. We evaluate our method on nine ImageNet-pretrained models. Our approach achieves the lowest FPR in 5 out of 9 ImageNet models, obtains the best average FPR overall, and consistently ranking among the top 3 across all evaluated models. Furthermore, we highlight the benefits of contrastive representations by showing strong performance with ResNet SCL and CLIP architectures. Finally, we demonstrate that the scale-invariant nature of our score enables an ensemble strategy via simple score summation.



Paperid:109
Authors:Wenxuan Wang, Fan Zhang, Yufeng Cui, Haiwen Diao, Zhuoyan Luo, Huchuan Lu, Jing Liu, Xinlong Wang
Abstract:
Existing vision tokenization isolates the optimization of vision tokenizers from downstream training, implicitly assuming the visual tokens can generalize well across various tasks, e.g., image generation and visual question answering. The vision tokenizer optimized for low-level reconstruction is agnostic to downstream tasks requiring varied representations and semantics. This decoupled paradigm introduces a critical misalignment: The loss of the vision tokenization can be the representation bottleneck for target tasks. For example, errors in tokenizing text in a given image lead to poor results when recognizing or generating them. To address this, we propose ETT, an end-to-end vision tokenizer tuning approach that enables joint optimization between vision tokenization and target autoregressive tasks. Unlike prior autoregressive models that use only discrete indices from a frozen vision tokenizer, ETT leverages the visual embeddings of the tokenizer codebook, and optimizes the vision tokenizers end-to-end with both reconstruction and caption objectives. ETT can be seamlessly integrated into existing training pipelines with minimal architecture modifications. Our ETT is simple to implement and integrate, without the need to adjust the original codebooks or architectures of the employed large language models. Extensive experiments demonstrate that our proposed end-to-end vision tokenizer tuning unlocks significant performance gains, i.e., 2-6% for multimodal understanding and visual generation tasks compared to frozen tokenizer baselines, while preserving the original reconstruction capability. We hope this very simple and strong method can empower multimodal foundation models besides image generation and understanding.



Paperid:110
Authors:Lorenz Vaitl, Leon Klein
Abstract:
Boltzmann Generators have emerged as a promising machine learning tool for generating samples from equilibrium distributions of molecular systems using Normalizing Flows and importance weighting.Recently, Flow Matching has helped speed up Continuous Normalizing Flows (CNFs), scale them to more complex molecular systems, and minimize the length of the flow integration trajectories. We investigate the benefits of using path gradients to fine-tune CNFs initially trained by Flow Matching, in the setting where a target energy is known. Our experiments show that this hybrid approach yields up to a threefold increase in sampling efficiency for molecular systems, all while using the same model, a similar computational budget and without the need for additional sampling. Furthermore, by measuring the length of the flow trajectories during fine-tuning, we show that path gradients largely preserve the learned structure of the flow.



Paperid:111
Authors:Aneesh Barthakur, Luiz Chamon
Abstract:
As machine learning applications grow increasingly ubiquitous and complex, they face an increasing set of requirements beyond accuracy. The prevalent approach to handling this challenge is to aggregate a weighted combination of requirement violation penalties into the training objective. To be effective, this approach requires careful tuning of these hyperparameters (weights), typically involving trial-and-error and cross-validation, which becomes ineffective even for a moderate number of requirements. These issues are exacerbated when the requirements involve parities or equalities, as is the case in fairness. An alternative technique uses constrained optimization to formulate these learning problems. Yet, existing approximation and generalization guarantees do not apply to problems involving equality constraints. In this work, we derive a generalization theory for equality-constrained statistical learning problems, showing that their solutions can be approximated using samples and rich parametrizations. Using these results, we propose a practical algorithm based on solving a sequence ofunconstrained,empiricallearning problems. We showcase its effectiveness and the new formulations enabled by equality constraints in fair learning applications.



Paperid:112
Authors:David Steinmann, Wolfgang Stammer, Antonia Wüst, Kristian Kersting
Abstract:
Developing high-performing, yet interpretable models remains a critical challenge in modern AI. Concept-based models (CBMs) attempt to address this by extracting human-understandable concepts from a global encoding (e.g., image encoding) and then applying a linear classifier on the resulting concept activations, enabling transparent decision-making. However, their reliance on holistic image encodings limits their expressiveness in object-centric real-world settings and thus hinders their ability to solve complex vision tasks beyond single-label classification. To tackle these challenges, we introduce Object-Centric Concept Bottlenecks (OCB), a framework that combines the strengths of CBMs and pre-trained object-centric foundation models, boosting performance and interpretability. We evaluate OCB on complex image datasets and conduct a comprehensive ablation study to analyze key components of the framework, such as strategies for aggregating object-concept encodings. The results show that OCB outperforms traditional CBMs and allows one to make interpretable decisions for complex visual tasks.



Paperid:113
Authors:Gabriel Nobis, Maximilian Springenberg, Arina Belova, Rembert Daems, Christoph Knochenhauer, Manfred Opper, Tolga Birdal, Wojciech Samek
Abstract:
Abstract:We present *Fractional Diffusion Bridge Models* (FDBM), a novel generative diffusion bridge framework driven by the rich and non-Markovian fractional Brownian motion (fBM).Real stochastic processes exhibit a degree of memory effects (correlations in time), long-range dependencies, roughness and anomalous diffusion phenomena that are not captured in standard diffusion or bridge modeling due to the use of Brownian motion. As a remedy, leveraging a recent Markovian approximation (MA-fBM), we construct FDBMs that enable tractable inference while preserving the non-Markovian nature of fBM. We prove the resulting bridge is a coupling-preserving process and leverage it for future state prediction from paired training data. We then extend our formulation to the Schr\"{o}dinger bridge problem and derive a principled loss function to learn the unpaired data translation. We evaluate FDBM on both tasks: predicting future protein conformations from aligned data, and unpaired image translation. In both settings, FDBM achieves superior performance compared to the Brownian baselines, yielding lower root mean squared deviation (RMSD) of C$_\alpha$ atomic positions in protein structure prediction and lower Fréchet Inception Distance (FID) in unpaired image translation.



Paperid:114
Authors:Amit Daniely, Idan Mehalel, Elchanan Mossel
Abstract:
Abstract:We study online learning of feedforward neural networks with the sign activation function that implement functions from the unit ball in $\mathbb{R}^d$ to a finite label set $\mathcal{Y} = \{1, \ldots, Y \}$.First, we characterize a margin condition that is sufficient and in some cases necessary for online learnability of a neural network: Every neuron in the first hidden layer classifies all instances with some margin $\gamma$ bounded away from zero. Quantitatively, we prove that for any net, the optimal mistake bound is at most approximately $\mathtt{TS}(d,\gamma)$, which is the $(d,\gamma)$-totally-separable-packing number, a more restricted variation of the standard $(d,\gamma)$-packing number. We complement this result by constructing a net on which any learner makes $\mathtt{TS}(d,\gamma)$ many mistakes. We also give a quantitative lower bound of approximately $\mathtt{TS}(d,\gamma) \geq \max\{1/(\gamma \sqrt{d})^d, d\}$ when $\gamma \geq 1/2$, implying that for some nets and input sequences every learner will err for $\exp(d)$ many times, and that a dimension-free mistake bound is almost always impossible.To remedy this inevitable dependence on $d$, it is natural to seek additional natural restrictions to be placed on the network, so that the dependence on $d$ is removed. We study two such restrictions. The first is the multi-index model, in which the function computed by the net depends only on $s \ll d$ orthonormal directions. We prove a mistake bound of approximately $(1.5/\gamma)^{s + 2}$ in this model.The second is the extended margin assumption. In this setting, we assume that all neurons (in all layers) in the network classify every ingoing input from previous layer with margin $\gamma$ bounded away from zero. In this model, we prove a mistake bound of approximately $(\log Y)/ \gamma^{O(L)}$, where L is the depth of the network.



Paperid:115
Authors:Panjian Huang, Saihui Hou, Chunshui Cao, Xu Liu, Yongzhen Huang
Abstract:
Abstract:What is a gait? Appearance-based gait networks consider a gait as the human shape and motion information from images. Model-based gait networks treat a gait as the human inherent structure from points. However, the considerations remain vague for humans to comprehend truly. In this work, we introduce a novel paradigm Vocabulary-Guided Gait Recognition, dubbed Gait-World, which attempts to explore gait concepts through human vocabularies with Vision-Language Models (VLMs). Despite VLMs have achieved the remarkable progress in various vision tasks, the cognitive capability regarding gait modalities remains limited. The success element in Gait-World is the proper vocabulary prompt where this paradigm carefully selects gait cycle actions as Vocabulary Base, bridging the gait and vocabulary feature spaces and further promoting human understanding for the gait. How to extract gait features? Although previous gait networks have made significant progress, learning solely from gait modalities on limited gait databases makes it difficult to learn robust gait features for practicality. Therefore, we propose the first Gait-World model, dubbed $\alpha$-Gait, which guides the gait network learning with universal vocabulary knowledge from VLMs. However, due to the heterogeneity of the modalities, directly integrating vocabulary and gait features is highly challenging as they reside in different embedding spaces. To address the issues, $\alpha$-Gait designs Vocabulary Relation Mapper and Gait Fine-grained Detector to map and establish vocabulary relations in the gait space for detecting corresponding gait features. Extensive experiments on CASIA-B, CCPG, SUSTech1K, Gait3D and GREW reveal the potential value and research directions of vocabulary information from VLMs in the gait field.



Paperid:116
Authors:Mohamed Younes, Adnane Boukhayma
Abstract:
2D Gaussian Splatting (2DGS) has recently emerged as a promising method for novel view synthesis and surface reconstruction, offering better view-consistency and geometric accuracy than volumetric 3DGS. However, 2DGS suffers from severe aliasing artifacts when rendering at different sampling rates than those used during training, limiting its practical applications in scenarios requiring camera zoom or varying fields of view. We identify that these artifacts stem from two key limitations: the lack of frequency constraints in the representation and an ineffective screen-space clamping approach. To address these issues, we present AA-2DGS, an antialiased formulation of 2D Gaussian Splatting that maintains its geometric benefits while significantly enhancing rendering quality across different scales. Our method introduces a world space flat smoothing kernel that constrains the frequency content of 2D Gaussian primitives based on the maximal sampling frequency from training views, effectively eliminating high-frequency artifacts when zooming in. Additionally, we derive a novel object space Mip filter by leveraging an affine approximation of the ray-splat intersection mapping, which allows us to efficiently apply proper anti-aliasing directly in the local space of each splat.



Paperid:117
Authors:Annie Marsden, Elad Hazan
Abstract:
We study the problem of preconditioning in the setting of sequential prediction. From the theoretical lens of linear dynamical systems, we show that applying a convolution to the input sequence translates to applying a polynomial to the unknown transition matrix in the hidden space. With this insight, we develop a novel preconditioning method that convolves the input sequence with the coefficients of the Chebyshev or Legendre polynomials. We formally prove that this improves the regret of a wide family of prediction methods. We proceed to apply this preconditioning technique to the method of spectral filtering. This gives the first sublinear regret bound that is also hidden-dimension free (up to logarithmic factors) even when the hidden transition matrix is asymmetric.From rigorous experiments on synthetic data we show that our simple preconditioning method generalizes to both 1) settings where the data is \emph{not} from a linear dynamical system and 2) a broad range of learning algorithms, including recurrent neural networks.



Paperid:118
Authors:Aishwarya Venkataramanan, Joachim Denzler
Abstract:
We propose the Distance-informed Neural Process (DNP), a novel variant of Neural Processes that improves uncertainty estimation by combining global and distance-aware local latent structures. Standard Neural Processes (NPs) often rely on a global latent variable and struggle with uncertainty calibration and capturing local data dependencies. DNP addresses these limitations by introducing a global latent variable to model task-level variations and a local latent variable to capture input similarity within a distance-preserving latent space. This is achieved through bi-Lipschitz regularization, which bounds distortions in input relationships and encourages the preservation of relative distances in the latent space. This modeling approach allows DNP to produce better-calibrated uncertainty estimates and more effectively distinguish in- from out-of-distribution data. Empirical results demonstrate that DNP achieves strong predictive performance and improved uncertainty calibration across regression and classification tasks.



Paperid:119
Authors:Guan-Horng Liu, Jaemoo Choi, Yongxin Chen, Benjamin Miller, Ricky T. Q. Chen
Abstract:
Computational methods for learning to sample from the Boltzmann distribution—where the target distribution is known only up to an unnormalized energy function—have advanced significantly recently. Due to the lack of explicit target samples, however, prior diffusion-based methods, known asdiffusion samplers, often require importance-weighted estimation or complicated learning processes. Both trade off scalability with extensive evaluations of the energy and model, thereby limiting their practical usage. In this work, we proposeAdjoint Schrödinger Bridge Sampler (ASBS), a new diffusion sampler that employs simple and scalable matching-based objectives yet without the need to estimate target samples during training. ASBS is grounded on a mathematical model—the Schrödinger Bridge—which enhances sampling efficiency via kinetic-optimal transportation. Through a new lens of stochastic optimal control theory, we demonstrate how SB-based diffusion samplers can be learned at scale via Adjoint Matching and prove convergence to the global solution. Notably, ASBS generalizes the recent Adjoint Sampling (Havens et al., 2025) to arbitrary source distributions by relaxing the so-called memoryless condition that largely restricts the design space. Through extensive experiments, we demonstrate the effectiveness of ASBS on sampling from classical energy functions, amortized conformer generation, and molecular Boltzmann distributions.



Paperid:120
Authors:Alexey Kurennoy, Majed Dodin, Tural Gurbanov, Ana Peleteiro Ramallo
Abstract:
Online evaluation of machine learning models is typically conducted through randomised experiments. Sequential statistical tests are valuable tools for analysing these experiments, as they enable researchers to stop data collection early without increasing the risk of false discoveries. However, existing sequential tests either limit the number of interim analyses or suffer from low statistical power. In this paper, we introduce a novel sequential test designed for continuous monitoring of online experiments. We validate our method using semi-synthetic simulations and demonstrate that it outperforms current state-of-the-art sequential testing approaches. Our method is derived using a new technique that ``inverts'' a bound on the probability of threshold crossing, based on a classical maximal inequality.



Paperid:121
Authors:Yingqi Yu, Honglin Chen, Jun Wu, Wei Xie, Xiangyang Li
Abstract:
Abstract:Quantum noise constitutes a fundamental obstacle to realizing practical quantum technologies. To address the pivotal challenge of identifying quantum systems least affected by noise, we introduce the purest quantum state identification, which can be used to improve the accuracy of quantum computation and communication. We formulate a rigorous paradigm for identifying the purest quantum state among $K$ unknown $n$-qubit quantum states using total $N$ quantum state copies. For incoherent strategies, we derive the first adaptive algorithm achieving error probability $\exp\left(- \Omega\left(\frac{N H_1}{\log(K) 2^n }\right) \right)$, fundamentally improving quantum property learning through measurement optimization. By developing a coherent measurement protocol with error bound $\exp\left(- \Omega\left(\frac{N H_2}{\log(K) }\right) \right)$, we demonstrate a significant separation from incoherent strategies, formally quantifying the power of quantum memory and coherent measurement. Furthermore, we establish a lower bound by demonstrating that all strategies with fixed two-outcome incoherent POVM must suffer error probability exceeding $ \exp\left( - O\left(\frac{NH_1}{2^n}\right)\right)$. This research advances the characterization of quantum noise through efficient learning frameworks. Our results establish theoretical foundations for noise-adaptive quantum property learning while delivering practical protocols for enhancing the reliability of quantum hardware.



Paperid:122
Authors:Boyuan Zhang, Yingjun Du, Xiantong Zhen, Ling Shao
Abstract:
Task vectors capture how a model changes during fine-tuning by recording the difference between pre-trained and task-specific weights. The composition of task vectors, a key operator in task arithmetic, enables models to integrate knowledge from multiple tasks without incurring additional inference costs. In this paper, we propose variational task vector composition, where composition coefficients are taken as latent variables and estimated in a Bayesian inference framework. Unlike previous methods that operate at the task level, our framework focuses on sample-specific composition. Motivated by the observation of structural redundancy in task vectors, we introduce a Spike-and-Slab prior that promotes sparsity and preserves only the most informative components. To further address the high variance and sampling inefficiency in sparse, high-dimensional spaces, we develop a gated sampling mechanism that constructs a controllable posterior by filtering the composition coefficients based on both uncertainty and importance. This yields a more stable and interpretable variational framework by deterministically selecting reliable task components, reducing sampling variance while improving transparency and generalization. Experimental results demonstrate that our method consistently outperforms existing approaches across all datasets by selectively leveraging the most reliable and informative components in task vectors. These findings highlight the practical value of our approach, establishing a new standard for efficient and effective task vector composition.



Paperid:123
Authors:Tyron Lardy, Christina Katsimerou, Wouter Koolen
Abstract:
In real-world decision making problems one needs to pick among multiple policies the one that performs best while respecting economic constraints. This motivates the problem of constrained best-arm identification for bandit problems where every arm is a joint distribution of reward and cost. We investigate the general case where reward and cost are dependent. The goal is to accurately identify the arm with the highest mean reward among all arms whose mean cost is below a given threshold. We prove information-theoretic lower bounds on the sample complexity for three models: Gaussian with fixed covariance, Gaussian with unknown covariance, and non-parametric distributions of rectangular support. We propose a combination of a sampling and a stopping rule that correctly identifies the constrained best arm and matches the optimal sample complexities for each of the three models. Our simulations demonstrate the empirical performance of our algorithms.



Paperid:124
Authors:Pol Caselles RIco, Francesc Moreno-Noguer
Abstract:
Existing 3D face modeling methods usually depend on 3D Morphable Models, which inherently constrain the representation capacity to fixed shape priors. Optimization-based approaches offer high-quality reconstructions but tend to be computationally expensive. In this work, we introduce GLVD, a hybrid method for 3D face reconstruction from few-shot images that extends Learned Vertex Descent (LVD) by integrating per-vertex neural field optimization with global structural guidance from dynamically predicted 3D keypoints. By incorporating relative spatial encoding, GLVD iteratively refines mesh vertices without requiring dense 3D supervision. This enables expressive and adaptable geometry reconstruction while maintaining computational efficiency. GLVD achieves state-of-the-art performance in single-view settings and remains highly competitive in multi-view scenarios, all while substantially reducing inference time.



Paperid:125
Authors:Jianglin Lu, Hailing Wang, Kuo Yang, Yitian Zhang, Simon Jenni, Yun Fu
Abstract:
Abstract:Recent studies have uncovered an interesting phenomenon: unimodal foundation models tend to learn convergent representations, regardless of differences in architecture, training objectives, or data modalities. However, these representations are essentially internal abstractions of samples that characterize samples independently, leading to structural myopia and limited expressiveness.In this paper, we propose *The Indra Representation Hypothesis*, inspired by the philosophical metaphor of Indra’s Net. We argue that representations from unimodal foundation models are converging to implicitly reflect a shared relational structure underlying reality, akin to the relational ontology of Indra’s Net. We formalize this hypothesis using the $\mathcal{V}$-enriched Yoneda embedding from category theory, defining the Indra representation as a relational profile of each sample with respect to all others. This formulation is shown to be unique, complete, and structure-preserving under a given cost function. We instantiate the Indra representation using angular distance and evaluate it in cross-model and cross-modal scenarios involving vision, language, and audio. Extensive experiments demonstrate that Indra representations consistently enhance matching performance across modalities and architectures, providing a theoretically grounded and practical framework for post-training alignment in unimodal models.



Paperid:126
Authors:Alex Infanger, Bruce W. Lee, Addie Foote, Leni Shor, Harish Kamath, Bryce Woodworth, Alex Cloud, Alexander Turner, Jacob Goldman-Wetzler
Abstract:
Large language models acquire undesired capabilities during pretraining, such as knowledge that could support the development of weapons of mass destruction. Machine unlearning methods aim to remove these capabilities but typically achieve only behavioral suppression: the model appears safe but retains latent capabilities that can be easily recovered through adversarial finetuning. To address this limitation, we show that distilling an unlearned model into a randomly initialized student yields a model that is significantly more robust to such relearning attacks. Next, we introduce Unlearn-Noise-Distill-on-Outputs (UNDO), which approximates network reinitialization by copying and noising the weights of an unlearned teacher model. UNDO attains most of the robustness of gold-standard data filtering at a fraction of the compute cost. In extensive comparisons to unlearning baselines on language and arithmetic tasks, UNDO achieves Pareto optimal robustness and performance. On the Weapons of Mass Destruction Proxy (WMDP) benchmark, UNDO robustifies unlearning to achieve competitive performance. Since LLM developers already apply distillation to reduce inference costs, our results show they can obtain robustness essentially “for free” by incorporating suppression before distillation. This insight offers a practical path to safer model deployment and open-weight release.



Paperid:127
Authors:Shuhan Zhang, Zhi Wang, Rui Gao, Shuang Li
Abstract:
Context-dependent choice behavior poses a fundamental challenge to traditional choice models, which often assume stable, independent utilities across alternatives. While recent context-dependent choice models attempt to capture such effects, they are typically limited by restrictive interaction structures or lack interpretability when scaled with flexible function approximators like neural networks. In this paper, we propose a novel modeling framework that bridges the gap between expressiveness and interpretability in context-dependent choice. We develop a neural architecture that decomposes utility into base, pairwise, and higher-order interaction components, with an inductive bias grounded in permutation-equivariant structure. Our framework supports feature-rich alternatives and enables systematic identification of interaction effects across different orders. We establish the universal approximation property of the model and demonstrate its empirical effectiveness on hypothetical, synthetic, and real-world datasets. The proposed approach achieves strong predictive performance while offering transparent insights into the contextual drivers of human choice.



Paperid:128
Authors:Yvann Le Fay, Nicolas Chopin, Simon Barthelmé
Abstract:
Abstract:Variational inference seeks the best approximation of a target distribution within a chosen family, where "best" means minimizing Kullback-Leibler divergence. When the approximation family is exponential, the optimal approximation satisfies a fixed-point equation.We introduce LSVI (Least Squares Variational Inference), a gradient-free, Monte Carlo-based scheme for the fixed-point recursion, where each iteration boils down to performing ordinary least squares regression on tempered log-target evaluations under the variational approximation.We show that LSVI is equivalent to biased stochastic natural gradient descent and use this to derive convergence rates with respect to the numbers of samples and iterations.When the approximation family is Gaussian, LSVI involves inverting the Fisher information matrix, whose size grows quadratically with dimension $d$.We exploit the regression formulation to eliminate the need for this inversion, yielding $O(d^3)$ complexity in the full-covariance case and $O(d)$ in the mean-field case.Finally, we numerically demonstrate LSVI’s performance on various tasks, including logistic regression, discrete variable selection, and Bayesian synthetic likelihood, showing competitive results with state-of-the-art methods, even when gradients are unavailable.



Paperid:129
Authors:Pierre Clavier, Nathan Grinsztajn, Raphaël Avalos, Yannis Flet-Berliac, Irem Ergun, Omar Darwiche Domingues, Olivier Pietquin, Pierre Richemond, Florian Strub, Matthieu Geist
Abstract:
The fine-tuning of pre-trained large language models (LLMs) using reinforcement learning (RL) is generally formulated as direct policy optimization. This approach was naturally favored as it efficiently improves a pretrained LLM with simple gradient updates. Another RL paradigm, Q-learning methods, has received far less attention in the LLM community while demonstrating major success in various non-LLM RL tasks. In particular, Q-learning effectiveness stems from its sample efficiency and ability to learn offline, which is particularly valuable given the high computational cost of sampling with LLM. However, naively applying a Q-learning–style update to the model’s logits is ineffective due to the specificity of LLMs. Our contribution is to derive theoretically grounded loss functions from Bellman equations to adapt Q-learning methods to LLMs. To do so, we interpret LLM logits as Q-values and carefully adapt insights from the RL literature to account for LLM-specific characteristics. It thereby ensures that the logits become reliable Q-value estimates. We then use this loss to build a practical algorithm, ShiQ for Shifted-Q, that supports off-policy, token-wise learning while remaining simple to implement. Finally, ShiQ is evaluated on both synthetic data and real-world benchmarks, e.g., UltraFeedback, BFCL-V3, demonstrating its effectiveness in both single-turn and multi-turn LLM settings.



Paperid:130
Authors:Julius Erbach, Dominik Narnhofer, Andreas Dombos, Bernt Schiele, Jan Eric Lenssen, Konrad Schindler
Abstract:
Flow-based latent generative models such as Stable Diffusion 3 are able to generate images with remarkable quality, even enabling photorealistic text-to-image generation. Their impressive performance suggests that these models should also constitute powerful priors for inverse imaging problems, but that approach has not yet led to comparable fidelity.There are several key obstacles: (i) the encoding into a lower-dimensional latent space makes the underlying (forward) mapping non-linear; (ii) the data likelihood term is usually intractable; and (iii) learned generative models struggle to recover rare, atypical data modes during inference.We present FLAIR, a novel training free variational framework that leverages flow-based generative models as a prior for inverse problems. To that end, we introduce a variational objective for flow matching that is agnostic to the type of degradation, and combine it with deterministic trajectory adjustments to recover atypical modes. To enforce exact consistency with the observed data, we decouple the optimization of the data fidelity and regularization terms. Moreover, we introduce a time-dependent calibration scheme in which the strength of the regularization is modulated according to off-line accuracy estimates. Results on standard imaging benchmarks demonstrate that FLAIR consistently outperforms existing diffusion- and flow-based methods in terms of reconstruction quality and sample diversity.



Paperid:131
Authors:Nikita Karagodin, Shu Ge, Yury Polyanskiy, Philippe Rigollet
Abstract:
We study the effect of normalization schemes on token representations in deep transformers. Modeling their evolution as interacting particles on the sphere, we show that normalization acts as a form of speed regulation. This perspective enables a unified analysis of several schemes---includingPost-LN,Pre-LN,Mix-LN,Peri-LN,nGPT, andLN-scaling---revealing how they influence clustering dynamics and representation collapse. Our framework clarifies how different schemes shape token representations across layers and provides a principled basis for comparing them, identifyingPeri-LNas a particularly effective choice.



Paperid:132
Authors:Enrique Nueve, Bo Waggoner
Abstract:
When agents trade in a Duality-based Cost Function prediction market, they collectively implement the learning algorithm Follow-The-Regularized-Leader [Abernethy et al., 2013]. We ask whether other learning algorithms could be used to inspire the design of prediction markets. By decomposing and modifying the Duality-based Cost Function Market Maker's (DCFMM) pricing mechanism, we propose a new prediction market, called the Smooth Quadratic Prediction Market, the incentivizes agents to collectively implement general steepest gradient descent. Relative to the DCFMM, the Smooth Quadratic Prediction Market has a better worst-case monetary loss for AD securities while preserving axiom guarantees such as the existence of instantaneous price, information incorporation, expressiveness, no arbitrage, and a form of incentive compatibility. To motivate the application of the Smooth Quadratic Prediction Market, we independently examine agents' trading behavior under two realistic constraints: bounded budgets and buy-only securities. Finally, we provide an introductory analysis of an approach to facilitate adaptive liquidity using the Smooth Quadratic Prediction Market. Our results suggest future designs where the price update rule is separate from the fee structure, yet guarantees are preserved.



Paperid:133
Authors:Yingru Li, Jiawei Xu, Baoxiang Wang, Zhiquan Luo
Abstract:
Abstract:Thompson Sampling is a principled method for balancing exploration and exploitation, but its real-world adoption faces computational challenges in large-scale or non-conjugate settings. While ensemble-based approaches offer partial remedies, they typically require prohibitively large ensemble sizes. We propose Ensemble++, a scalable framework using a novel shared-factor ensemble architecture with random linear combinations. For linear bandits, we provide theoretical guarantees showing that Ensemble++ achieves regret comparable to exact Thompson Sampling with only $\Theta(d \log T)$ ensemble sizes--significantly outperforming prior methods. Crucially, this efficiency holds across both compact and finite action sets with either time-invariant or time-varying contexts without configuration changes. We extend this theoretical foundation to nonlinear rewards by replacing fixed features with learnable neural representations while preserving the same incremental update principle, effectively bridging theory and practice for complex domains. Comprehensive experiments across linear, quadratic, neural, and GPT-based contextual bandits validate our theoretical findings and demonstrate Ensemble++'s superior regret-computation tradeoff versus state-of-the-art methods.



Paperid:134
Authors:Zhiyuan Gao, Jiageng Mao, Hong-Xing Yu, HAOZHE LOU, Emily Jia, Jernej Barbic, Jiajun Wu, Yue Wang
Abstract:
A longstanding goal in computer vision is to model motions from videos, while the representations behind motions, i.e. the invisible physical interactions that cause objects to deform and move, remain largely unexplored. In this paper, we study how to recover the invisible forces from visual observations, e.g., estimating the wind field by observing a leaf falling to the ground. Our key innovation is an end-to-end differentiable inverse graphics framework, which jointly models object geometry, physical properties, and interactions directly from videos. Through backpropagation, our approach enables the recovery of force representations from object motions. We validate our method on both synthetic and real-world scenarios, and the results demonstrate its ability to infer plausible force fields from videos. Furthermore, we show the potential applications of our approach, including physics-based video generation and editing.We hope our approach sheds light on understanding and modeling the physical process behind pixels, bridging the gap between vision and physics.



Paperid:135
Authors:Le Quan, Viet Ta
Abstract:
Many real-world graphs exhibit diverse and complex topological structures that are not well captured by geometric manifolds with uniform global curvature, such as hyperbolic or spherical spaces. Recently, there has been growing interest in embedding graphs into pseudo-Riemannian manifolds, which generalize both hyperbolic and spherical geometries. However, existing approaches face three significant limitations, including the ineffective pseudo-Riemannain framework, the shallow architectures, and the absence of clear guideline for selecting suitable pseudo-Riemannian manifolds. To address these issues, we introduce a novel diffeomorphic framework for graph embedding that aligns with the nature of pseudo-Riemannian manifolds. Subsequently, we propose the pseudo-Riemannian Graph Transformer for learning representations of complex graph structures. Our diffeomorphic framework in pseudo-Riemannian geometry enables the principled definitions of core transformer components, including linear attention, residual connection, and layer normalization. Finally, we develop a lightweight space searching algorithm to automatically identify the most suitable pseudo-Riemannian manifold for an input graph. Extensive experiments on diverse real-world graphs demonstrate that our model consistently outperforms other baselines in both node classification and link prediction tasks.



Paperid:136
Authors:Gongfan Fang, Xinyin Ma, Xinchao Wang
Abstract:
Reasoning Language Models, capable of extended chain-of-thought reasoning, have demonstrated remarkable performance on tasks requiring complex logical inference. However, applying elaborate reasoning for all queries often results in substantial computational inefficiencies, particularly when many problems admit straightforward solutions. This motivates an open question: Can LLMs learn when to think? To answer this, we propose Thinkless, a learnable framework that empowers an LLM to adaptively select between short-form and long-form reasoning based on both task complexity and the model's ability. Thinkless is trained under a reinforcement learning paradigm and employs two control tokens, \emph{} for concise responses and \emph{} for detailed reasoning. At the core of our method is a Decoupled Group Relative Policy Optimization (DeGRPO) algorithm, which decomposes the learning objective of hybrid reasoning into two components: (1) a control token loss that governs the selection of the reasoning mode, and (2) a response loss that improves the accuracy of the generated answers. This decoupled formulation enables fine-grained control over the contributions of each objective, stabilizing training and effectively preventing collapse observed in vanilla GRPO. Empirically, on several benchmarks such as Minerva Algebra, MATH-500, and GSM8K, Thinkless is able to reduce the usage of long-chain thinking by 50\% - 90\%, significantly reducing the computational cost of Reasoning Language Models.



Paperid:137
Authors:Joshua Engels, David Baek, Subhash Kantamneni, Max Tegmark
Abstract:
Scalable oversight, the process by which weaker AI systems supervise stronger ones, has been proposed as a key strategy to control future superintelligent systems. However, it is still unclear how scalable oversight itself scales.To address this gap, we propose a framework that quantifies the probability of successful oversight as a function of the capabilities of the overseer and the system being overseen.Specifically, our framework models oversight as a game between capability-mismatched players; the players have oversight-specific Elo scores that are a piecewise-linear function of their general intelligence, with two plateaus corresponding to task incompetence and task saturation. We validate our framework with a modified version of the game Nim and then apply it to four oversight games: Mafia, Debate, Backdoor Code and Wargames. For each game, we find scaling laws that approximate how domain performance depends on general AI system capability. We then build on our findings in a theoretical study of Nested Scalable Oversight (NSO), a process in which trusted models oversee untrusted stronger models, which then become the trusted models in the next step. We identify conditions under which NSO succeeds and derive numerically (and in some cases analytically) the optimal number of oversight levels to maximize the probability of oversight success. We also apply our theory to our four oversight games, where we find that NSO success rates at a general Elo gap of 400 are 13.5\% for Mafia, 51.7\% for Debate, 10.0\% for Backdoor Code, and 9.4\% for Wargames; these rates decline further when overseeing stronger systems.



Paperid:138
Authors:Chengtang Yao, Zhidan Liu, Jiaxi Zeng, Lidong Yu, Yuwei Wu, Yunde Jia
Abstract:
3D visual illusion is a perceptual phenomenon where a two-dimensional plane is manipulated to simulate three-dimensional spatial relationships, making a flat artwork or object look three-dimensional in the human visual system. In this paper, we reveal that the machine visual system is also seriously fooled by 3D visual illusions, including monocular and binocular depth estimation. In order to explore and analyze the impact of 3D visual illusion on depth estimation, we collect a large dataset containing almost 3k scenes and 200k images to train and evaluate SOTA monocular and binocular depth estimation methods. We also propose a 3D visual illusion depth estimation framework that uses common sense from the vision language model to adaptively fuse depth from binocular disparity and monocular depth. Experiments show that SOTA monocular, binocular, and multi-view depth estimation approaches are all fooled by various 3D visual illusions, while our method achieves SOTA performance.



Paperid:139
Authors:Mykola Lukashchuk, Raphaël Trésor, Wouter Nuijten, Ismail Senoz, Bert Vries
Abstract:
Abstract:This paper introduces the Quotient Bayesian Learning Rule, an extension of natural-gradient Bayesian updates to probability models that fall outside the exponential family. Building on the observation that many heavy-tailed and otherwise non-exponential distributions arise as marginals of minimal exponential families, we prove that such marginals inherit a unique Fisher–Rao information geometry via the quotient-manifold construction. Exploiting this geometry, we derive the Quotient Natural Gradient algorithm, which takes steepest-descent steps in the well-structured covering space and projects them horizontally onto the quotient, thereby guaranteeing parametrization-invariant optimization in the target space. Empirical results on the Student-$t$ distribution confirm that our method converges more rapidly and attains higher-quality solutions than previous variants of the Bayesian Learning Rule. These findings position quotient geometry as a unifying tool for efficient and principled inference across a broad class of latent-variable models.



Paperid:140
Authors:Andreas Theophilou, Özgür Şimşek
Abstract:
Discovering novel states or transitions efficiently requires effective policies or optionsthat direct the agent away from well-visited regions of the state space. This paper introducesa method, Novel Exploration via Orthogonality (NEO), that automatically uncoversnot only which state regions are novel but also how to reach them by leveraging Laplacian-basedrepresentations. Specifically, we use the eigenvectors of a modified graph Laplacian toinduce gradient flows from frequently visited (less novel) states to seldom visited (more novel)states. We show that modifying the Laplacian with weighted sinks yieldseigenvectors whose extreme values align with the most novel regions of the state space.We evaluate reinforcement learning agents with our option discovery method against related methods such as eigen-options and cover options in online, incremental settings. In summary we provide three key contributions, we provide bounds for our Laplacian matrices, we show that the smoothest eigenvectors with real eigenvalue below certain bounds provide guaranteed gradients to novel nodes for undirected and directed graphs, and we show that our option method NEO outperforms related state-of-the-art approaches in undirected and directed domains with varying structures.



Paperid:141
Authors:Ezgi Korkmaz
Abstract:
Following the pivotal success of learning strategies to win at tasks, solely by interacting with an environment without any supervision, agents have gained the ability to make sequential decisions in complex MDPs. Yet, reinforcement learning policies face exponentially growing state spaces in experience collection in high dimensional MDPs resulting in a dichotomy between computational complexity and policy success. In our paper we focus on the agent’s interaction with the environment in a high-dimensional MDP during the learning phase and we introduce a theoretically-founded novel method based on experiences obtained through counteractive actions. Our analysis and method provides a theoretical basis for effective and accelerated experience collection, and further comes with zero additional computational cost while leading to significant acceleration in learning. We conduct extensive experiments in the Arcade Learning Environment with high-dimensional state representation MDPs. The experimental results further verify our theoretical analysis, and our method achieves significant performance increase in high-dimensional environments.



Paperid:142
Authors:Yanbing Mao, Yihao Cai, Lui Sha
Abstract:
This paper introduces the Real-DRL framework for safety-critical autonomous systems, enabling runtime learning of a deep reinforcement learning (DRL) agent to develop safe and high-performance action policies in real plants while prioritizing safety. The Real-DRL consists of three interactive components: a DRL-Student, a PHY-Teacher, and a Trigger. The DRL-Student is a DRL agent that innovates in the dual self-learning and teaching-to-learn paradigm and the safety-status-dependent batch sampling. On the other hand, PHY-Teacher is a physics-model-based design of action policies that focuses solely on safety-critical functions. PHY-Teacher is novel in its real-time patch for two key missions: i) fostering the teaching-to-learn paradigm for DRL-Student and ii) backing up the safety of real plants. The Trigger manages the interaction between the DRL-Student and the PHY-Teacher. Powered by the three interactive components, the Real-DRL can effectively address safety challenges that arise from the unknown unknowns and the Sim2Real gap. Additionally, Real-DRL notably features i) assured safety, ii) automatic hierarchy learning (i.e., safety-first learning and then high-performance learning), and iii) safety-status-dependent batch sampling to address the experience imbalance caused by corner cases. Experiments with a real quadruped robot, a quadruped robot in Nvidia Isaac Gym, and a cart-pole system, along with comparisons and ablation studies, demonstrate the Real-DRL's effectiveness and unique features.



Paperid:143
Authors:Keqi Deng, Phil Woodland
Abstract:
Abstract:While Transformer self-attention offers strong parallelism, the Key-Value (KV) cache grows linearly with sequence length and becomes a bottleneck for inference efficiency. Multi-head latent attention was recently developed to compress the KV cache into a low-rank latent space. This paper proposes Multi-head Temporal Latent Attention (MTLA), which further reduces the KV cache size along the temporal dimension, greatly lowering the memory footprint of self-attention inference. MTLA employs a hyper-network to dynamically merge temporally adjacent KV cache vectors. To address the mismatch between the compressed KV cache and processed sequence lengths, a stride-aware causal mask is proposed to ensure efficient parallel training and consistency with inference behaviour. Experiments across tasks, including speech translation, speech recognition, speech understanding and text summarisation, demonstrate that MTLA achieves competitive performance compared to standard Multi-Head Attention (MHA), while greatly improving inference speed and GPU memory usage. For example, on a English-German speech translation task, MTLA achieves a 5.3$\times$ speedup and a reduction in GPU memory usage by a factor of 8.3 compared to MHA, while maintaining translation quality.



Paperid:144
Authors:Mahmoud Hegazy, Liviu Aolaritei, Michael Jordan, Aymeric Dieuleveut
Abstract:
Conformal prediction offers a distribution-free framework for constructing prediction sets with coverage guarantees. In practice, multiple valid conformal prediction sets may be available, arising from different models or methodologies. However, selecting the most desirable set, such as the smallest, can invalidate the coverage guarantees. To address this challenge, we propose a stability-based approach that ensures coverage for the selected prediction set. We extend our results to the online conformal setting, propose several refinements in settings where additional structure is available, and demonstrate its effectiveness through experiments.



Paperid:145
Authors:Beni Egressy, Jan Stühmer
Abstract:
While large language models (LLMs) demonstrate impressive capabilities across numerous applications, their robustness remains a critical concern. This paper is motivated by a specific vulnerability: the order sensitivity of LLMs. This vulnerability manifests itself as the order bias observed when LLMs decide between possible options (for example, a preference for the first option) and the tendency of LLMs to provide different answers when options are reordered. The use cases for this scenario extend beyond the classical case of multiple-choice question answering to the use of LLMs as automated evaluators in AI pipelines, comparing output generated by different models. We introduce Set-LLM, a novel architectural adaptation for pretrained LLMs that enables the processing of mixed set-text inputs with permutation invariance guarantees. The adaptations involve a new attention mask and new positional encodings specifically designed for sets. We provide a theoretical proof of invariance and demonstrate through experiments that Set-LLM can be trained effectively, achieving comparable or improved performance and maintaining the runtime of the original model, while eliminating order sensitivity.



Paperid:146
Authors:Javier Cembrano, Felix Fischer, Max Klimm
Abstract:
Abstract:We study the selection of agents based on mutual nominations, a theoretical problem with many applications from committee selection to AI alignment. As agents both select and are selected, they may be incentivized to misrepresent their true opinion about the eligibility of others to influence their own chances of selection. Impartial mechanisms circumvent this issue by guaranteeing that the selection of an agent is independent of the nominations cast by that agent. Previous research has established strong bounds on the performance of impartial mechanisms, measured by their ability to approximate the number of nominations for the most highly nominated agents. We study to what extent the performance of impartial mechanisms can be improved if they are given a prediction of a set of agents receiving a maximum number of nominations. Specifically, we provide bounds on the consistency and robustness of such mechanisms, where consistency measures the performance of the mechanisms when the prediction is correct and robustness its performance when the prediction is incorrect. For the general setting where up to $k$ agents are to be selected and agents nominate any number of other agents, we give a mechanism with consistency $1-O\big(\frac{1}{k}\big)$ and robustness $1-\frac{1}{e}-O\big(\frac{1}{k}\big)$. For the special case of selecting a single agent based on a single nomination per agent, we prove that $1$-consistency can be achieved while guaranteeing $\frac{1}{2}$-robustness. A close comparison with previous results shows that (asymptotically) optimal consistency can be achieved with little to no sacrifice in terms of robustness.



Paperid:147
Authors:Huajun Yin, Liyuan Wang, Yingqiu Zhu, Liping Zhu, Danyang Huang
Abstract:
With the rapid development of the big data era, Vertical Federated Learning (VFL) has been widely applied to enable data collaboration while ensuring privacy protection. However, the ultrahigh dimensionality of features and the sparse data structures inherent in large-scale datasets introduce significant computational complexity. In this paper, we propose the Vertical Federated Feature Screening (VFS) algorithm, which effectively reduces computational, communication, and encryption costs. VFS is a two-stage feature screening procedure that proceeds from coarse to fine: the first stage quickly filters out irrelevant feature groups, followed by a more refined screening of individual features. It significantly reduces the resource demands of downstream tasks such as secure joint modeling or federated feature selection. This efficiency is particularly beneficial in scenarios with ultrahigh feature dimensionality or severe class imbalance in the response variable. The statistical and computational properties of VFS are rigorously established. Numerical simulations and real-world applications demonstrate its superior performance.



Paperid:148
Authors:Mert Cemri, Melissa Z Pan, Shuyi Yang, Lakshya A Agrawal, Bhavya Chopra, Rishabh Tiwari, Kurt Keutzer, Aditya Parameswaran, Dan Klein, Kannan Ramchandran, Matei A Zaharia, Joseph Gonzalez, Ion Stoica
Abstract:
Despite enthusiasm for Multi-Agent LLM Systems (MAS), their performance gains on popular benchmarks are often minimal. This gap highlights a critical need for a principled understanding of why MAS fail. Addressing this question requires systematic identification and analysis of failure patterns. We introduce MAD, a comprehensive dataset of 1000+ annotated traces collected across 7 popular MAS frameworks. We create MAD, the first Multi-Agent System Failure Dataset, to outline the failure dynamics in MAS for guiding the development of better future systems. To enable systematic classification of failures for MAD, we build the first Multi-Agent System Failure Taxonomy (MAST). We develop MAST through rigorous analysis of 150 traces, guided closely by expert human annotators andvalidated by high inter-annotator agreement (κ = 0.88). This process identifies 14 unique modes, clustered into 3 categories: (i) specification issues, (ii) inter-agent misalignment, and (iii) task verification. To enable scalable annotation, we develop an LLM-as-a-Judge pipeline with high agreement with human annotations. We leverage MAST and MAD to analyze failure patterns across models (GPT4, Claude 3) and tasks (coding, math, general agent), demonstrating improvement headrooms from better MAS design. Our analysis provides insights revealing that identified failures require more sophisticated solutions, highlighting a clear roadmap for future research. We publicly release our comprehensive dataset (MAD), the MAST, and our LLM annotator to facilitate widespread research and development in MAS.



Paperid:149
Authors:Facheng Yu, Ronak Mehta, Alex Luedtke, Zaid Harchaoui
Abstract:
Stochastic gradient optimization is the dominant learning paradigm for a variety of scenarios, from classical supervised learning to modern self-supervised learning. We consider stochastic gradient algorithms for learning problems whose objectives rely on unknown nuisance parameters, and establish non-asymptotic convergence guarantees. Our results show that, while the presence of a nuisance can alter the optimum and upset the optimization trajectory, the classical stochastic gradient algorithm may still converge under appropriate conditions, such as Neyman orthogonality. Moreover, even when Neyman orthogonality is not satisfied, we also show that an algorithm variant with approximately orthogonalized updates (with an approximately orthogonalized gradient oracle) may achieve similar convergence rates. Examples from orthogonal statistical learning/double machine learning and causal inference are discussed.



Paperid:150
Authors:Junyan Cheng, Peter Clark, Kyle Richardson
Abstract:
Abstract:*Can we leverage LLMs to model the process of discovering novel language model (LM) architectures?* Inspired by real research, we propose a multi-agent LLM approach that simulates the conventional stages of research, from ideation and literature search (proposal stage) to design implementation (code generation), generative pre-training, and downstream evaluation (verification). Using ideas from scaling laws, our system *Genesys* employs a *Ladder of Scales* approach; new designs are proposed, adversarially reviewed, implemented, and selectively verified at increasingly larger model scales (14M$\sim$350M parameters) with a narrowing budget (the number of models we can train at each scale). To help make discovery efficient and factorizable, Genesys uses a novel genetic programming backbone, which we show has empirical advantages over commonly used direct prompt generation workflows (e.g., $\sim$86\% percentage point improvement in successful design generation, a key bottleneck). We report experiments involving 1,162 newly discovered designs (1,062 fully verified) and find the best designs to be competitive with known architectures (e.g., outperform GPT2, Mamba2, etc., on 6/9 common benchmarks). We couple these results with comprehensive system-level ablations and formal results, which give broader insights into the design of effective autonomous discovery systems.



Paperid:151
Authors:Pierre-François Massiani, Christian Fiedler, Lukas Haverbeck, Friedrich Solowjow, Sebastian Trimpe
Abstract:
We propose a framework for hypothesis testing on conditional probability distributions, which we then use to construct conditional two-sample statistical tests. These tests identify the inputs --- called covariates in this context --- where two conditional expectations differ with high probability. Our key idea is to transform confidence bounds of a learning method into a conditional two-sample test, and we instantiate this principle for kernel ridge regression (KRR) and conditional kernel mean embeddings. We generalize existing pointwise-in-time or time-uniform confidence bounds for KRR to previously-inaccessible yet essential cases such as infinite-dimensional outputs with non-trace-class kernels. These bounds enable circumventing the need for independent data in our statistical tests, since they allow online sampling. We also introduce bootstrapping schemes leveraging the parametric form of testing thresholds identified in theory to avoid tuning inaccessible parameters, making our method readily applicable in practice. Such conditional two-sample tests are especially relevant in applications where data arrive sequentially or non-independently, or when output distributions vary with operational parameters. We demonstrate their utility through examples in process monitoring and comparison of dynamical systems. Overall, our results establish a comprehensive foundation for conditional two-sample testing, from theoretical guarantees to practical implementation, and advance the state-of-the-art on the concentration of vector-valued least squares estimation.



Paperid:152
Authors:Changdae Oh, Jiatong Li, Shawn Im, Sharon Li
Abstract:
Despite widespread adoption, multimodal large language models (MLLMs) suffer performance degradation when encountering unfamiliar queries under distribution shifts. Existing methods to improve MLLM generalization typically require either more instruction data or larger advanced model architectures, both of which incur non-trivial human labor or computational costs. In this work, we take an alternative approach to enhance the robustness of MLLMs under distribution shifts, from a representation learning perspective. Inspired by the information bottleneck (IB) principle, we derive a variational lower bound of the IB for MLLMs and devise a practical implementation, Visual Instruction Bottleneck Tuning (Vittle). We then provide a theoretical justification of Vittle by revealing its connection to an information-theoretic robustness metric of MLLM. Empirical validation of three MLLMs on open-ended and closed-form question answering and object hallucination detection tasks over 45 datasets, including 30 shift scenarios, demonstrates that Vittle consistently improves the MLLM's robustness under shifts by pursuing the learning of a minimal sufficient representation.



Paperid:153
Authors:Steve Hanneke, Shay Moran, Maximilian Thiessen
Abstract:
Abstract:We revisit the classical model of nonuniform PAC learning, introduced by Benedek and Itai [1994], where generalization guarantees may depend on the target concept (but not on the marginal distribution). In this work, we propose and study a complementary variant, which we call *marginal-nonuniform learning*. In this setting, guarantees may depend on the marginal distribution over the domain, but must hold uniformly over all concepts. This captures the intuition that some data distributions are inherently easier to learn from than others, allowing for a flexible, distribution-sensitive view of learnability. Our main result is a complete characterization of the achievable learning rates in this model, revealing a trichotomy: exponential rates of the form $e^{-n}$ arise precisely when the hypothesis class is finite; linear rates of the form $d/n$ are achievable when a recently introduced combinatorial parameter, the VC-eluder dimension $d$, is finite; and arbitrarily slow rates may occur when $d = \infty$. We conclude by situating marginal-nonuniform learning within the broader landscape of universal learning, and by discussing its relationship to other distribution-dependent learning paradigms.



Paperid:154
Authors:Hoang Chuong Nguyen, Wei Mao, Jose M. Alvarez, miaomiao Liu
Abstract:
3D Gaussian Splatting (3DGS) has recently emerged as a fast, high-quality method for novel view synthesis (NVS). However, its use of low-degree spherical harmonics limits its ability to capture spatially varying color and view-dependent effects like specular highlights. Existing work augments Gaussians with either a global texture map, which struggles with complex scenes, or per-Gaussian textures, which introduce high memory overhead and limit resolution. We propose Image-Based Gaussian Splatting, an efficient alternative that leverages high-resolution source images for fine details and view-specific color modeling. Specifically, we model each pixel color as a combination of a base color from standard 3DGS rendering and a learned residual, inferred from neighboring training images. This promotes accurate surface alignment and enables rendering images of high-frequency details and accurate view-dependent effects. Experiments on standard NVS benchmarks show that our method significantly outperforms prior Gaussian Splatting approaches in rendering quality, without increasing the storage footprint.



Paperid:155
Authors:Aditya Gangrade, Venkatesh Saligrama
Abstract:
Abstract:We study safe linear bandits (SLBs), where an agent selects actions from a convex set to maximize an unknown linear objective subject to unknown linear constraints in each round. Existing methods for SLBs provide strong regret guarantees, but require solving expensive optimization problems (e.g., second-order cones, NP hard programs). To address this, we propose Constrained Linear Thompson Sampling (COLTS), a sampling-based framework that selects actions by solving perturbed linear programs, which significantly reduces computational costs while matching the regret and risk of prior methods. We develop two main variants: S-COLTS, which ensures zero risk and ${\tilde{O}(\sqrt{d^3 T})}$ regret given a safe action, and R-COLTS, which achieves ${\tilde{O}(\sqrt{d^3 T})}$ regret and risk with no instance information. In simulations, these methods match or outperform state of the art SLB approaches while substantially improving scalability. On the technical front, we introduce a novel coupled noise design that ensures frequent 'local optimism' about the true optimum, and a scaling-based analysis to handle the per-round variability of constraints.



Paperid:156
Authors:Swadesh Sistla, Max Kleiman-Weiner
Abstract:
The emergence of programming and strategic reasoning abilities in large language models (LLMs) enable their participation in \textit{program meta-games}: a game-theoretic setting in which players submit computer programs in lieu of moves. These programs offer various possible advantages, including human interpretability, local transparency, and formal verifiability; additionally, they enable \textit{program equilibria}, solution concepts reliant upon program submission and inaccessible within normal form settings. We evaluate the capabilities of leading open-source and closed-weight LLMs to behaviorally predict and classify program meta-game strategies, implement program equilibrium solution concepts, and play program meta-games against other LLM agents across dyadic and evolutionary settings. We identify the emergence of payoff-maximizing, cooperative, and deceptive strategies, as well as LLM-based implementations of existing program equilibrium solution concepts.



Paperid:157
Authors:Wujiang Xu, Kai Mei, Hang Gao, Juntao Tan, Zujie Liang, Yongfeng Zhang
Abstract:
While large language model (LLM) agents can effectively use external tools for complex real-world tasks, they require memory systems to leverage historical experiences. Current memory systems enable basic storage and retrieval but lack sophisticated memory organization, despite recent attempts to incorporate graph databases. Moreover, these systems' fixed operations and structures limit their adaptability across diverse tasks. To address this limitation, this paper proposes a novel agentic memory system for LLM agents that can dynamically organize memories in an agentic way. Following the basic principles of the Zettelkasten method, we designed our memory system to create interconnected knowledge networks through dynamic indexing and linking. When a new memory is added, we generate a comprehensive note containing multiple structured attributes, including contextual descriptions, keywords, and tags. The system then analyzes historical memories to identify relevant connections, establishing links where meaningful similarities exist. Additionally, this process enables memory evolution -- as new memories are integrated, they can trigger updates to the contextual representations and attributes of existing historical memories, allowing the memory network to continuously refine its understanding. Our approach combines the structured organization principles of Zettelkasten with the flexibility of agent-driven decision making, allowing for more adaptive and context-aware memory management.Empirical experiments on six foundation models show superior improvement against existing SOTA baselines. The code is available at \url{https://anonymous.4open.science/r/AgenticMemory-76B4}.



Paperid:158
Authors:Kaiyang Li, Shaobo Han, Qing Su, Wei Li, Zhipeng Cai, Jonathan Shihao Ji
Abstract:
Low-Rank Adaptation (LoRA) has become the de facto parameter-efficient fine-tuning (PEFT) method for large language models (LLMs) by constraining weight updates to low-rank matrices. Recent works such as Tied-LoRA, VeRA, and VB-LoRA push efficiency further by introducing additional constraints to reduce the trainable parameter space. In this paper, we show that the parameter space reduction strategies employed by these LoRA variants can be formulated within a unified framework, Uni-LoRA, where the LoRA parameter space, flattened as a high-dimensional vector space R^D, can be reconstructed through a projection from a subspace R^d, with d << D. We demonstrate that the fundamental difference among various LoRA methods lies in the choice of the projection matrix, P ∈ R^{D×d}. Most existing LoRA variants rely on layer-wise or structure-specific projections that limit cross-layer parameter sharing, thereby compromising parameter efficiency. In light of this, we introduce an efficient and theoretically grounded projection matrix that is isometric, enabling global parameter sharing and reducing computation overhead. Furthermore, under the unified view of Uni-LoRA, this design requires only a single trainable vector to reconstruct LoRA parameters for the entire LLM -- making Uni-LoRA both a unified framework and a “one-vector-only” solution. Extensive experiments on GLUE, mathematical reasoning, and instruction tuning benchmarks demonstrate that Uni-LoRA achieves state-of-the-art parameter efficiency while outperforming or matching prior approaches in predictive performance.



Paperid:159
Authors:Hyeonseong Jeon, Ainaz Eftekhar, Aaron Walsman, Kuo-Hao Zeng, Ali Farhadi, Ranjay Krishna
Abstract:
Abstract:We introduce LOKI, a compute-efficient framework for co-designing morphologies and control policies that generalize across unseen tasks. Inspired by biological adaptation—where animals quickly adjust to morphological changes—our method overcomes the inefficiencies of traditional evolutionary and quality-diversity algorithms. We propose learning convergent functions: shared control policies trained across clusters of morphologically similar designs in a learned latent space, drastically reducing the training cost per design. Simultaneously, we promote divergent forms by replacing mutation with dynamic local search, enabling broader exploration and preventing premature convergence. The policy reuse allows us to explore $\sim780\times$ more designs using 78\% fewer simulation steps and 40\% less compute per design. Local competition paired with a broader search results in a diverse set of high-performing final morphologies. Using the UNIMAL design space and a flat-terrain locomotion task, LOKI discovers a rich variety of designs—ranging from quadrupeds to crabs, bipedals, and spinners—far more diverse than those produced by prior work. These morphologies also transfer better to unseen downstream tasks in agility, stability, and manipulation domains (e.g. $2 \times$ higher reward on bump and push box incline tasks). Overall, our approach produces designs that are both diverse and adaptable, with substantially greater sample efficiency than existing co-design methods.



Paperid:160
Authors:Guojian Zhan, Likun Wang, Xiangteng Zhang, Jiaxin Gao, Masayoshi TOMIZUKA, Shengbo Li
Abstract:
Online planning has proven effective in reinforcement learning (RL) for improving sample efficiency and final performance. However, using planning for environment interaction inevitably introduces a divergence between the collected data and the policy's actual behaviors, degrading both model learning and policy improvement. To address this, we propose BOOM (Bootstrap Off-policy with WOrld Model), a framework that tightly integrates planning and off-policy learning through a bootstrap loop: the policy initializes the planner, and the planner refines actions to bootstrap the policy through behavior alignment. This loop is supported by a jointly learned world model, which enables the planner to simulate future trajectories and provides value targets to facilitate policy improvement. The core of BOOM is a likelihood-free alignment loss that bootstraps the policy using the planner’s non-parametric action distribution, combined with a soft value-weighted mechanism that prioritizes high-return behaviors and mitigates variability in the planner’s action quality within the replay buffer. Experiments on the high-dimensional DeepMind Control Suite and Humanoid-Bench show that BOOM achieves state-of-the-art results in both training stability and final performance.



Paperid:161
Authors:Ruben Härle, Felix Friedrich, Manuel Brack, Björn Deiseroth, Stephan Waeldchen, Patrick Schramowski, Kristian Kersting
Abstract:
There is growing interest in leveraging mechanistic interpretability and controllability to better understand and influence the internal dynamics of large language models (LLMs). However, current methods face fundamental challenges in reliably localizing and manipulating feature representations. Sparse Autoencoders (SAEs) have recently emerged as a promising direction for feature extraction at scale,yet they, too, are limited by incomplete feature isolation and unreliable monosemanticity. To systematically quantify these limitations, we introduce Feature Monosemanticity Score (FMS), a novel metric to quantify feature monosemanticity in latent representation. Building on these insights, we propose Guided Sparse Autoencoders (G-SAE), a method that conditions latent representations on labeled concepts during training. We demonstrate that reliable localization and disentanglement of target concepts within the latent space improve interpretability, detection of behavior, and control. Specifically, our evaluations on toxicity detection, writing style identification, and privacy attribute recognition show that G-SAE not only enhances monosemanticity but also enables more effective and fine-grained steering with less quality degradation. Our findings provide actionable guidelines for measuring and advancing mechanistic interpretability and control of LLMs.



Paperid:162
Authors:Nina Nellen, Polina Turishcheva, Michaela Vystrčilová, Shashwat Sridhar, Tim Gollisch, Andreas Tolias, Alexander Ecker
Abstract:
Abstract:Deep neural networks trained to predict neural activity from visual input and behaviour have shown great potential to serve as digital twins of the visual cortex.Per-neuron embeddings derived from these models could potentially be used to map the functional landscape or identify cell types. However, state-of-the-art predictive models of mouse V1 do not generate functional embeddings that exhibit clear clustering patterns which would correspond to cell types.This raises the question whether the lack of clustered structure is due to limitations of current models or a true feature of the functional organization of mouse V1. In this work, we introduce DECEMber -- Deep Embedding Clustering via Expectation Maximization-based refinement -- an explicit inductive bias into predictive models that enhances clustering by adding an auxiliary $t$-distribution-inspired loss function that enforces structured organization among per-neuron embeddings.We jointly optimize both neuronal feature embeddings and clustering parameters, updating cluster centers and scale matrices using the EM-algorithm.We demonstrate that these modifications improve cluster consistency while preserving high predictive performance and surpassing standard clustering methods in terms of stability.Moreover, DECEMber generalizes well across species (mice, primates) and visual areas (retina, V1, V4). The code is available at https://anonymous.4open.science/r/DECEMber-clustering/.



Paperid:163
Authors:Pedro O. Pinheiro, Pan Kessel, Aya Ismail, Sai Pooja Mahajan, Kyunghyun Cho, Saeed Saremi, Nataša Tagasovska
Abstract:
Generative modeling is increasingly important for data-driven computational design problems. Conventional approaches typically pair a generative model with a discriminative model to select or guide samples towards optimized designs. However, these discriminative models often struggle when data is scarce, a common scenario in scientific applications, and are unreliable at the tail-ends of distributions where optimal designs usually lie.We introduce generative property enhancer (GPE), an approach that implicitly guides generation by matching samples with lower property values to high-value ones. Our framework, formulated as a conditional density estimation problem, inherently defines a target distribution with improved properties, compelling the generative model to produce enhanced and diverse designs without relying on auxiliary predictors. GPE offers a simple, scalable, and end-to-end approach for design optimization. It is modality-agnostic and can be seamless integrated across diverse generative architectures and losses.We show our model achieves competitive empirical results in standard protein fitness optimizationin silicobenchmarks. Finally, we propose iteratively training on a combination of limited real data and self-generated synthetic data, which we show enables extrapolation beyond the original property ranges.



Paperid:164
Authors:Aref Einizade, Dorina Thanou, Fragkiskos Malliaros, Jhony H. Giraldo
Abstract:
Simplicial complexes provide a powerful framework for modeling high-order interactions in structured data, making them particularly suitable for applications such as trajectory prediction and mesh processing. However, existing simplicial neural networks (SNNs), whether convolutional or attention-based, rely primarily on discrete filtering techniques, which can be restrictive. In contrast, partial differential equations (PDEs) on simplicial complexes offer a principled approach to capture continuous dynamics in such structures. In this work, we introduce continuous sim}plicial neural networks (COSIMO), a novel SNN architecture derived from PDEs on simplicial complexes. We provide theoretical and experimental justifications of COSIMO's stability under simplicial perturbations. Furthermore, we investigate the over-smoothing phenomenon—a common issue in geometric deep learning—demonstrating that COSIMO offers better control over this effect than discrete SNNs. Our experiments on real-world datasets demonstrate that COSIMO achieves competitive performance compared to state-of-the-art SNNs in complex and noisy environments.



Paperid:165
Authors:Felix Koulischer, Florian Handke, Johannes Deleu, Thomas Demeester, Luca Ambrogioni
Abstract:
While Classifier-Free Guidance (CFG) has become standard for improving sample fidelity in conditional diffusion models, it can harm diversity and induce memorization by applying constant guidance regardless of whether a particular sample needs correction. We proposeFeedBackGuidance (FBG), which uses a state-dependent coefficient to self-regulate guidance amounts based on need. Our approach is derived from first principles by assuming the learned conditional distribution is linearly corrupted by the unconditional distribution, contrasting with CFG's implicit multiplicative assumption. Our scheme relies on feedback of its own predictions about the conditional signal informativeness to adapt guidance dynamically during inference, challenging the view of guidance as a fixed hyperparameter. The approach is benchmarked on ImageNet512x512, where it significantly outperforms Classifier-Free Guidance and is competitive to Limited Interval Guidance (LIG) while benefitting from a strong mathematical framework. On Text-To-Image generation, we demonstrate that, as anticipated, our approach automatically applies higher guidance scales for complex prompts than for simpler ones and that it can be easily combined with existing guidance schemes such as CFG or LIG.



Paperid:166
Authors:Lele Fu, Sheng Huang, Tianchi Liao, Bowen Deng, zhangchuanfu, Shirui Pan, Chuan Chen
Abstract:
Federated graph learning (FGL) is a privacy-preserving paradigm for modeling distributed graph data, designed to train a powerful global graph neural network. Existing FGL methods predominantly rely on label information during training, effective FGL in an unsupervised setting remains largely unexplored territory. In this paper, we address two key challenges in unsupervised FGL: 1) Local models tend to converge in divergent directions due to the lack of shared semantic information across clients. Then, how to align representation spaces among multiple clients is first challenge. 2) Conventional federated weighted aggregation easily results in degrading the performance of the global model, then which raises another challenge, namely how to adaptively learn the global model parameters. In response to the two questions, we propose a tailored framework named FedPAM, which is composed of two modules: Representation Space Alignment (RSA) and Adaptive Global Parameter Learning (AGPL). RSA leverages a set of learnable anchors to define the global representation space, then local subgraphs are aligned with them through the fused Gromov-Wasserstein optimal transport, achieving the representation space alignment across clients. AGPL stacks local model parameters into third-order tensors, and adaptively integrates the global model parameters in a low-rank tensor space, which facilitates to fuse the high-order knowledge among clients. Extensive experiments on eight graph datasets are conducted, the results demonstrate that the proposed FedPAM is superior over classical and SOTA compared methods.



Paperid:167
Authors:Christoph Staudt, Mark Blacher, Tim Hoffmann, Kaspar Kasche, Olaf Beyersdorff, Joachim Giesen
Abstract:
Einsum expressions specify an output tensor in terms of several input tensors. They offer a simple yet expressive abstraction for many computational tasks in artificial intelligence and beyond. However, evaluating einsum expressions poses hard algorithmic problems that depend on the representation of the tensors. Two popular representations are multidimensional arrays and coordinate lists. The latter is a more compact representation for sparse tensors, that is, tensors where a significant proportion of the entries are zero. So far, however, most of the popular einsum implementations use the multidimensional array representation for tensors. Here, we show on a non-trivial example that, when evaluating einsum expressions, coordinate lists can be exponentially more efficient than multidimensional arrays. In practice, however, coordinate lists can also be significantly less efficient than multidimensional arrays, but it is hard to decide from the input tensors whether this will be the case. Sparsity evolves dynamically in intermediate tensors during the evaluation of an einsum expression. Therefore, we introduce a hybrid solution where the representation is switched on the fly from multidimensional arrays to coordinate lists depending on the sparsity of the remaining tensors. In our experiments on established benchmark einsum expressions, the hybrid solution is consistently competitive with or outperforms the better of the two static representations.



Paperid:168
Authors:Haoran Zhang, Olawale Salaudeen, Marzyeh Ghassemi
Abstract:
Real-world classification datasets often contain label bias, where observed labels differ systematically from the true labels at different rates for different demographic groups. Machine learning models trained on such datasets may then exhibit disparities in predictive performance across these groups. In this work, we characterize the problem of learning fair classification models with respect to the underlying ground truth labels when given only label biased data. We focus on the particular fairness definition of group sufficiency, i.e. equal calibration of risk scores across protected groups. We theoretically show that enforcing fairness with respect to label biased data necessarily results in group miscalibration with respect to the true labels. We then propose a regularizer which minimizes an upper bound on the sufficiency gap by penalizing a conditional mutual information term. Across experiments on eight tabular, image, and text datasets with both synthetic and real label noise, we find that our method reduces the sufficiency gap by up to 9.1% with no significant decrease in overall accuracy.



Paperid:169
Authors:Shreshth Saini, Shashank Gupta, Alan Bovik
Abstract:
Classifier‑free guidance (CFG) is the workhorse for steering large diffusion models toward text‑conditioned targets, yet its naïve application to rectified flow (RF) based models provokes severe off–manifold drift, yielding visual artifacts, text misalignment, and brittle behaviour. We present Rectified-CFG++, an adaptive predictor–corrector guidance that couples the deterministic efficiency of rectified flows with a geometry‑aware conditioning rule. Each inference step first executes a conditional RF update that anchors the sample near the learned transport path, then applies a weighted conditional correction that interpolates between conditional and unconditional velocity fields. We prove that the resulting velocity field is marginally consistent and that its trajectories remain within a bounded tubular neighbourhood of the data manifold, ensuring stability across a wide range of guidance strengths. Extensive experiments on large‑scale text‑to‑image models (Flux, Stable Diffusion 3/3.5, Lumina) show that Rectified-CFG++ consistently outperforms standard CFG on benchmark datasets such as MS‑COCO, LAION‑Aesthetic, and T2I‑CompBench. The code will be released upon publication.



Paperid:170
Authors:Yuanpeng Tu, Hao Luo, Xi Chen, Xiang Bai, Fan Wang, Hengshuang Zhao
Abstract:
We introduce PlayerOne, the first egocentric realistic world simulator, facilitating immersive and unrestricted exploration within vividly dynamic environments. Given an egocentric scene image from the user, PlayerOne can accurately construct the corresponding world and generate egocentric videos that are strictly aligned with the real-scene human motion of the user captured by an exocentric camera. PlayerOne is trained in a coarse-to-fine pipeline that first performs pretraining on large-scale egocentric text-video pairs for coarse-level egocentric understanding, followed by finetuning on synchronous motion-video data extracted from egocentric-exocentric video datasets with our automatic construction pipeline. Besides, considering the varying importance of different components, we design a part-disentangled motion injection scheme, enabling precise control of part-level movements. In addition, we devise a joint reconstruction framework that progressively models both the 4D scene and video frames, ensuring scene consistency in the long-form video generation. Experimental results demonstrate its great generalization ability in precise control of varying human movements and world-consistent modeling of diverse scenarios. It marks the first endeavor into egocentric real-world simulation and can pave the way for the community to delve into fresh frontiers of world modeling and its diverse applications.



Paperid:171
Authors:Kevin Zhu, AliAsghar MohammadiNasrabadi, Alexander Wong, John McPhee
Abstract:
We introduce MusclePose as an end-to-end learnable physics-infused 3D human pose estimator that incorporates muscle-dynamics modeling to infer human dynamics from monocular video. Current physics pose estimators aim to predict physically plausible poses by enforcing the underlying dynamics equations that govern motion. Since this is an underconstrained problem without force-annotated data, methods often estimate kinetics with external physics optimizers that may not be compatible with existing learning frameworks, or are too slow for real-time inference. While more recent methods use a regression-based approach to overcome these issues, the estimated kinetics can be seen as auxiliary predictions, and may not be physically plausible. To this end, we build on existing regression-based approaches, and aim to improve the biofidelity of kinetic inference with a multihypothesis approach --- by inferring joint torques via Lagrange’s equations and via muscle dynamics modeling with muscle torque generators. Furthermore, MusclePose predicts detailed human anthropometrics based on values from biomechanics studies, in contrast to existing physics pose estimators that construct their human models with shape primitives. We show that MusclePose is competitive with existing 3D pose estimators in positional accuracy, while also able to infer plausible human kinetics and muscle signals consistent with values from biomechanics studies, without requiring an external physics engine.



Paperid:172
Authors:Sridhar Mahadevan
Abstract:
In this paper, we aim to shed light on the universal properties underlying causal inference by formulating it in terms of a topos, a type of category that has set-like properties. We argue that a topos provides three universal properties that make it attractive in formalizing causal inference: a general theory for how to combine local functions that define independent causal mechanisms into a consistent global function building on the theory of sheaves in a topos; a generic way to define causal interventions using a subobject classifier in a topos category; and finally, an internal logical language for causal and counterfactual reasoning that emerges from the topos itself. A striking characteristic of subobject classifiers is that they induce an intuitionistic logic, whose semantics is based on the partially ordered lattice of subobjects. We show that the underlying subobject classifier for causal inference is not Boolean in general, but forms a Heyting algebra. We study the internal language of the topos induced by causal models by defining its corresponding Mitchell-Benabou language, a typed local set theory where causal models define types. We describe the Kripke-Joyal intuitionistic semantics of a topos that can support causal inference. We illustrate the application of our framework to a range of representations, from structural causal models to causal models over Markov categories and simplicial sets, as well as Lewis' possible-worlds logic of counterfactuals.



Paperid:173
Authors:Itai Gat, Neta Shaul, Uriel Singer, Yaron Lipman
Abstract:
Abstract:Autoregressive language models accumulate errors due to their fixed, irrevocable left-to-right token generation. To address this, we propose a new sampling method called Resample-Previous-Tokens (RPT). RPT mitigates error accumulation by iteratively revisiting and potentially replacing tokens in a window of previously generated text. This method can be integrated into existing autoregressive models, preserving their next-token-prediction quality and speed. Fine-tuning a pretrained 8B parameter model with RPT for only 100B resulted in $\sim$10\% relative improvements on reasoning and coding benchmarks compared to the standard sampling.



Paperid:174
Authors:Hoang Tran, Thanh Tran, Thanh Chu, Tam Le, Tan Nguyen
Abstract:
Optimal Transport (OT) has emerged as a fundamental tool in machine learning for comparing probability distributions in a geometrically meaningful manner. However, a key limitation of classical OT is its requirement that the source and target distributions have equal total mass, limiting its use in real-world settings involving imbalanced data, noise, outliers, or structural inconsistencies. Partial Transport (PT) addresses this limitation by allowing only a fraction of the mass to be transported, offering greater flexibility and robustness. Nonetheless, similar to OT, PT remains computationally expensive, as it typically involves solving large-scale linear programs—especially in high-dimensional spaces. To alleviate this computational burden, several emerging works have introduced the Tree-Sliced Wasserstein (TSW) distance, which projects distributions onto tree-metric spaces where OT problems admit closed-form solutions. Building on this line of research, we propose a novel framework that extends the tree-sliced approach to the PT setting, introducing the Partial Tree-Sliced Wasserstein (PartialTSW) distance. Our method is based on the key observation that, within tree-metric spaces, the PT problem can be equivalently reformulated as a standard balanced OT problem between suitably modified measures. This reformulation enables efficient computation while preserving the adaptability and robustness of partial transport. Our method proves effective across challenging tasks such as outlier removal and addressing class imbalance in image-to-image translation.



Paperid:175
Authors:Zeyu Xiao, Xinchao Wang
Abstract:
Modern smartphones often feature asymmetric dual-lens systems, capturing wide-angle and ultra-wide views with complementary perspectives and details. Motion and shake can blur the wide lens, while the ultra-wide lens, despite lower resolution, retains sharper details. This natural complementarity offers valuable cues for video deblurring. However, existing methods focus mainly on single-camera inputs or symmetric stereo pairs, neglecting the cross-lens redundancy in mobile dual-camera systems. In this paper, we propose the first video deblurring method, AsLeD-Net, which recurrently aligns and propagates temporal reference features from ultra-wide views fused with features extracted from wide-angle blurry frames. AsLeD-Net consists of two key modules: the adaptive local matching (ALM) module, which refines blurry features using K-nearest neighbor reference features, and the difference compensation (DC) module, which ensures spatial consistency and reduces misalignment. Additionally, AsLeD-Net uses the reference-guided motion compensation (RMC) module for temporal alignment, further improving frame-to-frame consistency in the deblurring process. We validate the effectiveness of AsLeD-Net through extensive experiments, benchmarking it against potential solutions for asymmetric lens deblurring.



Paperid:176
Authors:Yue Song, Andy Keller, Sevan Brodjian, Takeru Miyato, Yisong Yue, Pietro Perona, Max Welling
Abstract:
Orientation-rich images, such as fingerprints and textures, often exhibit coherent angular directional patterns that are challenging to model using standard generative approaches based on isotropic Euclidean diffusion. Motivated by the role of phase synchronization in biological systems, we propose a score-based generative model built on periodic domains by leveraging stochastic Kuramoto dynamics in the diffusion process. In neural and physical systems, Kuramoto models capture synchronization phenomena across coupled oscillators -- a behavior that we re-purpose here as an inductive bias for structured image generation. In our framework, the forward process performs \textit{synchronization} among phase variables through globally or locally coupled oscillator interactions and attraction to a global reference phase, gradually collapsing the data into a low-entropy von Mises distribution. The reverse process then performs \textit{desynchronization}, generating diverse patterns by reversing the dynamics with a learned score function. This approach enables structured destruction during forward diffusion and a hierarchical generation process that progressively refines global coherence into fine-scale details. We implement wrapped Gaussian transition kernels and periodicity-aware networks to account for the circular geometry. Our method achieves competitive results on general image benchmarks and significantly improves generation quality on orientation-dense datasets like fingerprints and textures. Ultimately, this work demonstrates the promise of biologically inspired synchronization dynamics as structured priors in generative modeling.



Paperid:177
Authors:Xinyi Huang, Kangfei Zhao, Long-Kai Huang
Abstract:
Model editing offers a promising paradigm for efficiently and precisely updating knowledge in pre-trained transformers without costly retraining. While extensively studied in language models (LMs), model editing for vision transformers (ViTs) remains underexplored. Existing methods typically adapt LM-based techniques by modifying the multi-layer perceptron (MLP) modules, overlooking the unique characteristics of ViTs. In this work, we show that ViT predictions are more strongly influenced by the multi-head self-attention (MSA) modules than by the MLPs. Building on this observation, we propose a two-stage framework for editing ViTs. First, we identify which attention heads are most responsible for incorrect predictions. Next, we selectively remove the corresponding features to correct the model’s prediction. To further balance error correction with predictive stability on unrelated data, we learn a projection matrix that refines the image representations. Extensive experiments across multiple real-world datasets and model editing benchmarks demonstrate that our method consistently outperforms existing model editing methods for ViTs, achieving superior generalization and locality.



Paperid:178
Authors:Jundong Xu, Hao Fei, Yuhui Zhang, Liangming Pan, Qijun Huang, Qian Liu, Preslav Nakov, Min-Yen Kan, William Yang Wang, Mong-Li Lee, Wynne Hsu
Abstract:
Multimodal symbolic logical reasoning, which aims to deduce new facts from multimodal input via formal logic, is critical in high-stakes applications such as autonomous driving and medical diagnosis, as its rigorous, deterministic reasoning helps prevent serious consequences. To evaluate such capabilities of current state-of-the-art vision language models (VLMs), we introduce the first benchmark MuSLR for multimodal symbolic logical reasoning grounded in formal logical rules. MuSLR comprises 1,093 instances across 7 domains, including 35 atomic symbolic logic and 976 logical combinations, with reasoning depths ranging from 2 to 9. We evaluate 7 state-of-the-art VLMs on MuSLR and find that they all struggle with multimodal symbolic reasoning, with the best model, GPT-4.1, achieving only 46.8%.Thus, we propose LogiCAM, a modular framework that applies formal logical rules to multimodal inputs, boosting GPT-4.1’s Chain-of-Thought performance by 14.13%, and delivering even larger gains on complex logics such as first-order logic. We also conduct a comprehensive error analysis, showing that around 70% of failures stem from logical misalignment between modalities, offering key insights to guide future improvements.



Paperid:179
Authors:Andrew F. Ilersich, Prasanth Nair
Abstract:
The physical sciences are replete with dynamical systems that require the resolution of a wide range of length and time scales. This presents significant computational challenges since direct numerical simulation requires discretization at the finest relevant scales, leading to a high-dimensional state space. In this work, we propose an approach to learn stochastic multiscale models in the form of stochastic differential equations directly from observational data. Our method resolves the state on a coarse mesh while introducing an auxiliary state to capture the effects of unresolved scales. We learn the parameters of the multiscale model using a modern forward-solver-free amortized variational inference method. Our approach draws inspiration from physics-based multiscale modeling approaches, such as large-eddy simulation in fluid dynamics, while learning directly from data. We present numerical studies to demonstrate that our learned multiscale models achieve superior predictive accuracy compared to direct numerical simulation and closure-type models at equivalent resolution.



Paperid:180
Authors:Liad Erez, Tomer Koren
Abstract:
Abstract:We study the problem of multiclass list classification with (semi-)bandit feedback, where input examples are mapped into subsets of size $m$ of a collection of $K$ possible labels. In each round of the interaction, the learner observes feedback consisting of the predicted labels which lie in some underlying set of ground truth labels associated with the given example. Our main result is for the $(\varepsilon,\delta)$-PAC variant of the problem for which we design an algorithm that returns an $\varepsilon$-optimal hypothesis with high probability using a sample complexity of $\smash{\widetilde{O}} \big( (\mathrm{poly}(K/m) + sm / \varepsilon^2) \log (|\mathcal H|/\delta) \big)$ where $\mathcal H$ is the underlying (finite) hypothesis class and $s$ is an upper bound on the number of true labels for a given example. This bound improves upon known bounds for combinatorial semi-bandits whenever $s \ll K$. Moreover, in the regime where $s = O(1)$ the leading terms in our bound match the corresponding full-information rates, implying that bandit feedback essentially comes at no cost. Our PAC learning algorithm is also computationally efficient given access to an ERM oracle for $\mathcal H$. In the special case of single-label classification corresponding to $s=m=1$, we prove a sample complexity bound of $O \big((K^7 + 1/\varepsilon^2)\log (|\mathcal H|/\delta)\big)$ which improves upon recent results in this scenario (Erez et al. '24). Additionally, we consider the regret minimization setting where data can be generated adversarially, and establish a regret bound of $\smash{\widetilde O(|\mathcal H| + \sqrt{smT \log |\mathcal H|})}$. Our results generalize and extend prior work in the simpler single-label setting (Erez et al. '24), and apply more generally to contextual combinatorial semi-bandit problems with $s$-sparse rewards.



Paperid:181
Authors:Jannis Born, Filip Skogh, Nico Wagner, Kahn Rhrissorrakrai, Filippo Utro, Aleksandros Sobczyk
Abstract:
At the core of the Transformer, the softmax normalizes the attention matrix to be right stochastic.Previous research has shown that this often de-stabilizes training and that enforcing the attention matrix to be doubly stochastic (through Sinkhorn's algorithm) consistently improves performance across different tasks, domains and Transformer flavors.However, Sinkhorn's algorithm is iterative, approximative, non-parametric and thus inflexible w.r.t. the obtained doubly stochastic matrix (DSM). Recently, it has been proven that DSMs can be obtained with a parametric quantum circuit, yielding a novel quantum inductive bias for DSMs with no known classical analogue. Motivated by this, we demonstrate the feasibility of a hybrid classical-quantum doubly stochastic Transformer (QDSFormer) that replaces the softmax in the self-attention layer with a variational quantum circuit.We study the expressive power of the circuit and find that it yields more diverse DSMs that better preserve information than classical operators.Across multiple small-scale object recognition tasks, we find that our QDSFormer consistently surpasses both a standard ViT and other doubly stochastic Transformers. Beyond the Sinkformer, this comparison includes a novel quantum-inspired doubly stochastic Transformer (based on QR decomposition) that can be of independent interest. Our QDSFormer also shows improved training stability and lower performance variation suggesting that it may mitigate the notoriously unstable training of ViTs on small-scale data.



Paperid:182
Authors:Anders Aamand, Justin Chen, Mina Dalirrooyfard, Slobodan Mitrovic, Yuriy Nevmyvaka, Sandeep Silwal, Yinzhan Xu
Abstract:
Abstract:Given an undirected, weighted $n$-vertex graph $G = (V, E, w)$, a Gomory-Hu tree $T$ is a weighted tree on $V$ that preserves the Min-$s$-$t$-Cut between any pair of vertices $s, t \in V$. Finding cuts in graphs is a key primitive in problems such as bipartite matching, spectral and correlation clustering, and community detection. We design a differentially private (DP) algorithm that computes an approximate Gomory-Hu tree. Our algorithm is $\varepsilon$-DP, runs in polynomial time, and can be used to compute $s$-$t$ cuts that are $\tilde{O}(n/\varepsilon)$-additive approximations of the Min-$s$-$t$-Cuts in $G$ for all distinct $s, t \in V$ with high probability. Our error bound is essentially optimal, since [Dalirrooyfard, Mitrovic and Nevmyvaka, Neurips 2023] showed that privately outputting a single Min-$s$-$t$-Cut requires $\Omega(n)$ additive error even with $(\varepsilon, \delta)$-DP and allowing for multiplicative error. Prior to our work, the best additive error bounds for approximate all-pairs Min-$s$-$t$-Cuts were $O(n^{3/2}/\varepsilon)$ for $\varepsilon$-DP [Gupta, Roth, Ullman, TCC 2009] and $\tilde{O}(\sqrt{mn}/ \varepsilon)$ for $(\varepsilon, \delta)$-DP [Liu, Upadhyay and Zou, SODA 2024], both achieved by DP algorithms that preserve all cuts in the graph. To achieve our result, we develop an $\varepsilon$-DP algorithm for the Minimum Isolating Cuts problem with near-linear error, and introduce a novel privacy composition technique combining elements of both parallel and basic composition to handle `bounded overlap' computational branches in recursive algorithms, which maybe of independent interest.



Paperid:183
Authors:Ruoxi Jiang, Xiao Zhang, Karan Jakhar, Peter Y. Lu, Pedram Hassanzadeh, Michael Maire, Rebecca Willett
Abstract:
Neural PDE solvers offer a powerful tool for modeling complex dynamical systems, but often struggle with error accumulation over long time horizons and maintaining stability and physical consistency. We introduce a multiscale implicit neural emulator that enhances long-term prediction accuracy by conditioning on a hierarchy of lower-dimensional future state representations. Drawing inspiration from the stability properties of numerical implicit time-stepping methods, our approach leverages predictions several steps ahead in time at increasing compression rates for next-timestep refinements. By actively adjusting the temporal downsampling ratios, our design enables the model to capture dynamics across multiple granularities and enforce long-range temporal coherence. Experiments on turbulent fluid dynamics show that our method achieves high short-term accuracy and produces long-term stable forecasts, significantly outperforming autoregressive baselines while adding minimal computational overhead.



Paperid:184
Authors:Harry Cheng, Ming-Hui Liu, Yangyang Guo, Tianyi Wang, Liqiang Nie, Mohan Kankanhalli
Abstract:
Deepfake detection models face two critical challenges: generalization to unseen manipulations and demographic fairness among population groups. However, existing approaches often demonstrate that these two objectives are inherently conflicting, revealing a trade-off between them.In this paper, we, for the first time, uncover and formally define a causal relationship between fairness and generalization.Building on the back-door adjustment, we show that controlling for confounders (data distribution and model capacity) enables improved generalization via fairness interventions.Motivated by this insight, we propose Demographic Attribute-insensitive Intervention Detection (DAID), a plug-and-play framework composed of: i) Demographic-aware data rebalancing, which employs inverse-propensity weighting and subgroup-wise feature normalization to neutralize distributional biases; and ii) Demographic-agnostic feature aggregation, which uses a novel alignment loss to suppress sensitive-attribute signals.Across three cross-domain benchmarks, DAID consistently achieves superior performance in both fairness and generalization compared to several state-of-the-art detectors, validating both its theoretical foundation and practical effectiveness.



Paperid:185
Authors:Yuxin Zuo, Kaiyan Zhang, Li Sheng, Shang Qu, Ganqu Cui, Xuekai Zhu, Haozhan Li, yuchen zhang, Xinwei Long, Ermo Hua, Biqing Qi, Youbang Sun, Zhiyuan Ma, Lifan Yuan, Ning Ding, Bowen Zhou
Abstract:
This paper investigates Reinforcement Learning (RL) on data without explicit labels for reasoning tasks in Large Language Models (LLMs). The core challenge of the problem is reward estimation during inference while not having access to ground-truth information. While this setting appears elusive, we find that common practices in Test-Time Scaling (TTS), such as majority voting, yield surprisingly effective rewards suitable for driving RL training. In this work, we introduce Test-Time Reinforcement Learning (TTRL), a novel method for training LLMs using RL on unlabeled data. TTRL enables self-evolution of LLMs by utilizing the priors in the pre-trained models. Our experiments demonstrate that TTRL consistently improves performance across a variety of tasks and models. Notably, TTRL boosts the pass@1 performance of Qwen-2.5-Math-7B by approximately 211% on the AIME 2024 with only unlabeled test data. Furthermore, although TTRL is only supervised by the Maj@N metric, TTRL has demonstrated performance to consistently surpass the upper limit of the initial model, and approach the performance of models trained directly on test data with ground-truth labels. Our experimental findings validate the general effectiveness of TTRL across various tasks and highlight TTRL's potential for broader tasks and domains.



Paperid:186
Authors:Anastasios Gerontopoulos, Spyridon Gidaris, Nikos Komodakis
Abstract:
Multi-token prediction has emerged as a promising objective for improving language model pretraining, but its benefits have not consistently generalized to other settings such as fine-tuning. In this paper, we propose MuToR, a simple and effective approach to multi-token prediction that interleaves learnable register tokens into the input sequence, each tasked with predicting future targets. Compared to existing methods, MuToR offers several key advantages: it introduces only a negligible number of additional parameters, requires no architectural changes—ensuring compatibility with off-the-shelf pretrained language models—and remains aligned with the next-token pretraining objective, making it especially well-suited for supervised fine-tuning. Moreover, it naturally supports scalable prediction horizons. We demonstrate the effectiveness and versatility of MuToR across a range of use cases, including supervised fine-tuning, parameter-efficient fine-tuning (PEFT), and pretraining, on challenging generative tasks in both language and vision domains.



Paperid:187
Authors:Hyungi Lee, Moonseok Choi, Hyunsu Kim, Kyunghyun Cho, Rajesh Ranganath, Juho Lee
Abstract:
Uncertainty-aware meta-learning aims not only for rapid adaptation to new tasks but also for reliable uncertainty estimation under limited supervision. Neural Processes (NPs) offer a flexible solution by learning implicit stochastic processes directly from data, often using a global latent variable to capture functional uncertainty. However, we empirically find that variational posteriors for this global latent variable are frequently miscalibrated, limiting both predictive accuracy and the reliability of uncertainty estimates. To address this issue, we propose Test Time Scaling for Neural Processes (TTSNPs), a sequential inference framework based on Sequential Monte Carlo Sampler (SMCS) that refines latent samples at test time without modifying the pre-trained NP model. TTSNPs iteratively transform variational samples into better approximations of the true posterior using neural transition kernels, significantly improving both prediction quality and uncertainty calibration. This makes NPs more robust and trustworthy, extending applicability to various scenarios requiring well-calibrated uncertainty estimates.



Paperid:188
Authors:Jakob de Raaij, Ariel Procaccia, Alexandros Psomas
Abstract:
A panel satisfiesdescriptive representationwhen its composition reflects the population. We examine the role of descriptive representation in collective decision making through an optimization lens, asking whether representative panels make decisions that maximize social welfare for the underlying population. Our main results suggest that, in general, representation with respect tointersectionsof two or more features guarantees higher social welfare than that achieved by the status quo of proportionally representing individual features. Moreover, an analysis of real data suggests that representation with respect to pairs of features is feasible in practice. These results have significant implications for the design ofcitizens' assemblies, which are gaining prominence in AI governance.



Paperid:189
Authors:Shikai Qiu, Charlie Chen, Hoang Phan, Qi Lei, Andrew Wilson
Abstract:
Abstract:In deep learning, the heavy computational overhead of second-order optimizers has traditionally limited their practical adoption in favor of simpler first-order methods like SGD and Adam. Recent empirical successes, however, suggest that some second-order methods may unlock significant computational efficiency in training large-scale machine learning models. In this work, we systematically investigate several key open questions. Do second-order optimizers outperform first-order methods under a fixed compute budget? How can we optimally scale the hyperparameters as we scale up the model? Do efficiency gains in small-scale experiments to transfer to larger-scale systems, and if so, how does second-order optimization affect nerual scaling laws? Through theoretical and empirical analysis, we derive scaling rules for hyperparameters such as learning rate as we jointly scale up width, depth, and batch size for a wide range of optimizers including Shampoo, SOAP, and Muon. We find that, under the correct scaling, second-order optimizers consistently outperform Adam across scales in the compute-optimal setting for training transformers. Furthermore, we show second-order optimizers alter the Chinchilla scaling law, requiring up to $2\times$ less data per parameter for compute-optimality.



Paperid:190
Authors:Zichen Wang, Anjie Dong, Zaiwen Wen
Abstract:
Abstract:Premise selection is a critical bottleneck in interactive theorem proving, particularly with large libraries. Existing methods, primarily relying on semantic embeddings, often fail to effectively leverage the rich structural information inherent in mathematical expressions. This paper proposes a novel framework for premise selection based on the structure of expression trees. The framework enhances premise selection ability by explicitly utilizing the structural information of Lean expressions and by means of the simplified tree representation obtained via common subexpression elimination. Our method employs a multi-stage filtering pipeline, incorporating structure-aware similarity measures including the Weisfeiler-Lehman kernel, tree edit distance, $\texttt{Const}$ node Jaccard similarity, and collapse-match similarity. An adaptive fusion strategy combines these metrics for refined ranking. To handle large-scale data efficiently, we incorporate cluster-based search space optimization and structural compatibility constraints. Comprehensive evaluation on a large theorem library extracted from Mathlib4 demonstrates that our method significantly outperforms existing premise retrieval tools across various metrics. Experimental analysis, including ablation studies and parameter sensitivity analysis, validates the contribution of individual components and highlights the efficacy of our structure-aware approach and multi-metric fusion.



Paperid:191
Authors:Xinyuan Fan, Feiyan Ma, Chenlei Leng, Weichi Wu
Abstract:
Graphons offer a powerful framework for modeling large-scale networks, yet estimation remains challenging. We propose a novel approach that leverages a low-rank additive representation, yielding both a low-rank connection probability matrix and a low-rank graphon--two goals rarely achieved jointly. Our method resolves identification issues and enables an efficient sequential algorithm based on subgraph counts and interpolation. We establish consistency and demonstrate strong empirical performance in terms of computational efficiency and estimation accuracy through simulations and data analysis.



Paperid:192
Authors:Roman Beltiukov, Satyandra Guthula, Wenbo Guo, Walter Willinger, Arpit Gupta
Abstract:
Abstract:This work presents a systematic investigation into the latent knowledge encoded within Network Foundation Models (NFMs) that focuses on hidden representations analysis rather than pure downstream task performance. Different from existing efforts, we analyze the models through a three-part evaluation: Embedding Geometry Analysis to assess representation space utilization, Metric Alignment Assessment to measure correspondence with domain-expert features, and Causal Sensitivity Testing to evaluate robustness to protocol perturbations. Using five diverse network datasets spanning controlled and real-world environments, we evaluate four state-of-the-art NFMs, revealing that they all exhibit significant anisotropy, inconsistent feature sensitivity patterns, an inability to separate the high-level context, payload dependency, and other properties. Our work identifies numerous limitations across all models and demonstrates that addressing them can significantly improve model performance (by up to +0.35 $F_1$ score without architectural changes).



Paperid:193
Authors:Meshi Bashari, Roy Maor Lotan, Yonghoon Lee, Edgar Dobriban, Yaniv Romano
Abstract:
Conformal prediction is a framework for predictive inference with a distribution-free, finite-sample guarantee. However, it tends to provide uninformative prediction sets when calibration data are scarce. This paper introduces Synthetic-powered predictive inference (SPPI), a novel framework that incorporates synthetic data---e.g., from a generative model---to improve sample efficiency. At the core of our method is a score transporter: an empirical quantile mapping that aligns nonconformity scores from trusted, real data with those from synthetic data. By carefully integrating the score transporter into the calibration process, SPPI provably achieves finite-sample coverage guarantees without making any assumptions about the real and synthetic data distributions. When the score distributions are well aligned, SPPI yields substantially tighter and more informative prediction sets than standard conformal prediction. Experiments on image classification and tabular regression demonstrate notable improvements in predictive efficiency in data-scarce settings.



Paperid:194
Authors:Yifei Zhou, Sergey Levine, Jason Weston, Xian Li, Sainbayar Sukhbaatar
Abstract:
Abstract:Large language models are quickly becoming the foundation for intelligent agents that are capable of using tools. However, training such agents is challenging because it requires human creation and annotation of a diverse set of tasks, tools, and evaluation criteria. In this paper, we propose the Self-Challenging Agent framework for training an agent on high-quality tasks that are generated by itself. The agent first plays the role of challenger and generates a task after interacting with the given tools. The tasks take the form of a novel general class of problems termed Code-as-Task, which are defined by an instruction, a verification function and solution and failure cases which serve as tests, allowing to filter only for high-quality tasks. The agent then takes an executor role and trains on those tasks with reinforcement learning using the evaluation feedback as a reward. We show our method improves the performance of Llama-3.1-8B-Instruct on two existing multi-turn tool-use agent benchmarks, M$^3$ToolEval and TauBench, with a two-fold average success rate increase, despite using only self-generated training data.



Paperid:195
Authors:Mirco Giacobbe, Daniel Kroening, Abhinandan Pal, Michael Tautschnig
Abstract:
Safety verification ensures that a system avoids undesired behaviour, acrucial requirement for the design of reactive systems, especially hardwaresystems. Liveness complements safety, ensuring that the system alsoachieves its desired objectives. A complete specification of functionalcorrectness must combine both safety and liveness aspects. Proving withmathematical certainty that a system satisfies a safety property demandspresenting an appropriate inductive invariant of the system, whereasliveness requires showing a measure of progress also known as rankingfunctions. Neural model checking has recently introduced a data-drivenapproach to the formal verification of reactive systems, albeit focusing onranking functions and thus addressing only liveness properties. In thispaper, we extend and generalise neural model checking to additionallyencompass inductive invariants and thus safety properties as well. Given asystem and a linear temporal logic specification of safety and liveness, ourapproach alternates a learning and a checking component towards theconstruction of a provably sound neural certificate. Our new methodintroduces a neural certificate architecture that jointly representsinductive invariants as proofs of safety, and ranking functions as proofs ofliveness. Moreover, our new architecture is able to leverage constraintsolvers for training, accelerating cognate neural model checking work otherwise basedon gradient descent. We experimentally demonstrate that our method is several orders of magnitude faster than the state-of-the-art model checkers on pure liveness and combined safety and liveness verification tasks written in SystemVerilog, while enabling the verification of richer properties than was previously possible forneural model checking.



Paperid:196
Authors:Yaniv Oren, Moritz Zanger, Pascal van der Vaart, Mustafa Mert Çelikok, Matthijs Spaan, Wendelin Boehmer
Abstract:
To learn approximately optimal acting policies for decision problems, modern Actor Critic algorithms rely on deep Neural Networks (DNNs) to parameterize the acting policy and greedification operators to iteratively improve it. The reliance on DNNs suggests an improvement that is gradient based, which is per step much less greedy than the improvement possible by greedier operators such as the greedy update used by Q-learning algorithms.On the other hand, slow and steady changes to the policy can also be beneficial for the stability of the learning process, resulting in a tradeoff between greedification and stability.To address this tradeoff, we propose to extend the standard framework of actor critic algorithms with value-improvement: a second greedification operator applied only when updating the policy's value estimate.In this framework the agent can evaluate non-parameterized policies and perform much greedier updates while maintaining the steady gradient-based improvement to the parameterized acting policy.We prove that this approach converges in the popular analysis scheme of generalized Policy Iteration in the finite-horizon domain.Empirically, incorporating value-improvement into the popular off-policy actor-critic algorithms TD3 and SAC significantly improves or matches performance over their respective baselines, across different environments from the DeepMind continuous control domain, with negligible compute and implementation cost.



Paperid:197
Authors:Ayush Sawarni, Sahasrajit Sarmasarkar, Vasilis Syrgkanis
Abstract:
This paper investigates the integration of response time data into human preference learning frameworks for more effective reward model elicitation. While binary preference data has become fundamental in fine-tuning foundation models, generative AI systems, and other large-scale models, the valuable temporal information inherent in user decision-making remains largely unexploited. We propose novel methodologies to incorporate response time information alongside binary choice data, leveraging the Evidence Accumulation Drift Diffusion (EZ) model, under which response time is informative of the preference strength. We develop Neyman-orthogonal loss functions that achieve oracle convergence rates for reward model learning, matching the theoretical optimal rates that would be attained if the expected response times for each query were known a priori. Our theoretical analysis demonstrates that for linear reward functions, conventional preference learning suffers from error rates that scale exponentially with reward magnitude. In contrast, our response time-augmented approach reduces this to polynomial scaling, representing a significant improvement in sample efficiency. We extend these guarantees to non-parametric reward function spaces, establishing convergence properties for more complex, realistic reward models. Our extensive experiments validate our theoretical findings in the context of preference learning over images.



Paperid:198
Authors:Siu Lun (Alan) Chau, Michele Caprio, Krikamol Muandet
Abstract:
Quantifying differences between probability measures is central to statistics and machine learning. As interest in Imprecise Probabilistic Machine Learning (IPML) grows---where inference and decision-making are grounded in richer representations of uncertainty with imprecise probabilities (IP)---there is an increasing need for metrics beyond classical probability. This work introduces the Integral Imprecise Probability Metric (IIPM) framework, a Choquet integral-based generalisation of classical Integral Probability Metric (IPM) to the setting of capacities---a broad class of IP models encompassing many existing ones, including lower probabilities, probability intervals, belief functions, and more. Theoretically, we establish conditions under which IIPM serves as a valid metric and metrises a form of weak convergence of capacities. Practically, IIPM not only enables comparison across different IP models but also supports the quantification of epistemic uncertainty (EU) within a single IP model. In particular, by comparing an IP model with its conjugate, IIPM gives rise to a new class of EU measures---Maximum Mean Imprecisions (MMIs)---which satisfy key axiomatic properties proposed in the uncertainty quantification (UQ) literature. We validate MMI through selective classification experiments, demonstrating strong empirical performance against established EU measures, and outperforming them when classical methods struggle to scale to a large number of classes. Our work advances both theory and practice in IPML, offering a principled framework for comparing and quantifying epistemic uncertainty under imprecision.



Paperid:199
Authors:Yung-Sung Chuang, Yang Li, Dong Wang, Ching-Feng Yeh, Kehan Lyu, Ramya Raghavendra, Jim Glass, LIFEI HUANG, Jason Weston, Luke Zettlemoyer, Xinlei Chen, Zhuang Liu, Saining Xie, Scott Yih, Shang-Wen Li, Hu Xu
Abstract:
Contrastive Language–Image Pretraining (CLIP) is a popular foundation model, supporting from zero-shot classification, retrieval to encoders for multimodal large language models (MLLMs). Although CLIP is successfully trained on billion-scale image-text pairs from the English world, scaling CLIP's training further to learning from the worldwide web data is still challenging: (1) no curation method is available to handle data points from non-English world; (2) the English performance from existing multilingual CLIP is worse than its English-only counterpart, i.e., ``curse of multilinguality'' that is common in LLMs. Here, we present MetaCLIP 2, the first recipe training CLIP from scratch on worldwide web-scale image-text pairs. To generalize our findings, we conduct rigorous ablations with minimal changes that are necessary to address the above challenges and present a recipe enabling mutual benefits from English and non-English world data.In zero-shot ImageNet classification, MetaCLIP 2 ViT-H/14 surpasses its English-only counterpart by 0.8% and mSigLIP by 0.7%, and surprisingly sets new state-of-the-art without system-level confounding factors (e.g., translation, bespoke architecture changes) on multilingual benchmarks, such as CVQA with 57.4%, Babel-ImageNet with 50.2% and XM3600 with 64.3% on image-to-text retrieval.Code is in supplementary material and models will be made publicly available.



Paperid:200
Authors:Xiaoliang Hu, Fuyun Wang, Tong Zhang, Zhen Cui
Abstract:
Model-free reinforcement learning (RL) combined with diffusion models has achieved significant progress in addressing complex continuous control tasks. However, a persistent challenge in RL remains the accurate estimation of Q-values, which critically governs policy optimization efficacy. While recent advancements employ parametric distributions to model value distributions for enhanced estimation accuracy, current methodologies predominantly rely on unimodal Gaussian assumptions or quantile representations. These constraints introduce distributional bias between the learned and true value distributions, particularly in some task with non-stationary policy, ultimately degrading performance. To address these limitations, we propose value diffusion reinforcement learning (VDRL), a novel model-free online RL method that utilizes the generative capacity of diffusion models to represent expressive and multimodal value distributions. The core innovation of VDRL lies in the use of the variational loss of diffusion-based value distribution, which is theoretically proven to be a tight lower bound for the optimization objective under the KL-divergence measurement. Furthermore, we introduce double value diffusion learning with sample selection to enhance training stability and further improve value estimation accuracy. Extensive experiments conducted on the MuJoCo benchmark demonstrate that VDRL significantly outperforms some SOTA model-free online RL baselines, showcasing its effectiveness and robustness.



Paperid:201
Authors:Mustafa Shukor, Louis Béthune, Dan Busbridge, David Grangier, Enrico Fini, Alaaeldin El-Nouby, Pierre Ablin
Abstract:
Abstract:Large foundation models are typically trained on data from multiple domains, with the data mixture—the proportion of each domain used—playing a critical role in model performance. The standard approach to selecting this mixture relies on trial and error, which becomes impractical for large-scale pretraining. We propose a systematic method to determine the optimal data mixture for any target domain using scaling laws. Our approach accurately predicts the loss of a model of size $N$ trained with $D$ tokens and a specific domain weight vector $h$. We validate the universality of these scaling laws by demonstrating their predictive power in two distinct and large-scale settings: large language model (LLM) and native multimodal model (NMM) pretraining. We further show that these scaling laws can extrapolate to new data mixtures and across scales: their parameters can be accurately estimated using a few small-scale training runs, and used to estimate the performance at larger scales and unseen domain weights. The scaling laws allow to derive the optimal domain weights for any target domain under a given training budget ($N$,$D$), providing a principled alternative to costly trial-and-error methods.



Paperid:202
Authors:Osayamen Aimuyo, Byungsoo Oh, Rachee Singh
Abstract:
The computational sparsity of Mixture-of-Experts (MoE) models allows them to scale efficiently, achieving sub-linear growth in computational costs as model sizes increase. However, distributing the experts across multiple GPUs introduces significant communication overheads, undermining some of their computational benefits. These overheads arise from frequent GPU kernel launches and synchronous communication across distributed experts. To overcome these limitations, we develop FlashDMoE, a novel MoE operator that fuses all computation and communication tasks into a single persistent GPU kernel. FlashDMoE employs device-initiated asynchronous communication and payload-efficient data transfers to dynamically route computation only to GPUs hosting actively selected experts. By leveraging fine-grained, tile-level parallelism within the kernel, FlashDMoE mitigates performance degradation caused by communication stragglers. Our evaluation on multiple ML servers, each with 8 NVIDIA H100 GPUs, demonstrates that FlashDMoE achieves 1.25-15X speedup over state-of-the-art MoE implementations.



Paperid:203
Authors:Sushil Varma, Irène Waldspurger, Laurent Massoulié
Abstract:
Abstract:We consider the graph alignment problem, wherein the objective is to find a vertex correspondence between two graphs that maximizes the edge overlap. The graph alignment problem is an instance of the quadratic assignment problem (QAP), known to be NP-hard in the worst case even to approximately solve. In this paper, we analyze Birkhoff relaxation, a tight convex relaxation of QAP, and present theoretical guarantees on its performance when the inputs follow the Gaussian Wigner Model. More specifically, the weighted adjacency matrices are correlated Gaussian Orthogonal Ensemble with correlation $1/\sqrt{1+\sigma^2}$. Denote the optimal solutions of the QAP and Birkhoff relaxation by $\Pi^\star$ and $X^\star$ respectively. We show that $\|X^\star-\Pi^\star\|_F^2 = o(n)$ when $\sigma = o(n^{-1})$ and $\|X^\star-\Pi^\star\|_F^2 = \Omega(n)$ when $\sigma = \Omega(n^{-0.5})$. Thus, the optimal solution $X^\star$ transitions from a small perturbation of $\Pi^\star$ for small $\sigma$ to being well separated from $\Pi^\star$ as $\sigma$ becomes larger than $n^{-0.5}$. This result allows us to guarantee that simple rounding procedures on $X^\star$ align $1-o(1)$ fraction of vertices correctly whenever $\sigma = o(n^{-1})$. This condition on $\sigma$ to ensure the success of the Birkhoff relaxation is state-of-the-art.



Paperid:204
Authors:Jongwoo Ko, Sungnyun Kim, Sungwoo Cho, Se-Young Yun
Abstract:
Human-generated reward signals are critical for aligning generative models with human preferences, guiding both training and inference-time evaluations. While large language models (LLMs) employed as proxy evaluators, i.e., LLM-as-a-Judge, significantly reduce the costs associated with manual annotations, they typically require extensive modality-specific training data and fail to generalize well across diverse multimodal tasks. In this paper, we proposeKirby-Judge, a reasoning-guided multimodal judge model that leverages minimal textual reasoning data to robustly generalize across multiple modalities and evaluation formats. Our core intuition is that structured textual reasoning explanations inherently encode generalizable decision-making patterns, enabling an effective transfer to multimodal judgments, e.g., with images or videos. Empirical results demonstrate that Kirby-Judge, despite being trained on significantly fewer text data, achieves competitive or superior performance compared to state-of-the-art commercial APIs and extensively trained multimodal evaluators. Notably, Kirby-Judge presents broad impact in modalities like molecule, where comprehensive evaluation benchmarks are scarce, underscoring its practical value in resource-constrained domains. Our framework highlights reasoning-based text supervision as a powerful, cost-effective alternative to traditional annotation-intensive approaches, substantially advancing scalable multimodal model-as-a-judge.



Paperid:205
Authors:Yingnan Liu, Rui Qiao, Mong-Li Lee, Wynne Hsu
Abstract:
Test-time adaptation aims to improve model robustness under distribution shifts by adapting models with access to unlabeled target samples. A primary cause of performance degradation under such shifts is the model’s reliance on features that lack a direct causal relationship with the prediction target. We introduce Test-time Adaptation by Causal Trimming (TACT), a method that identifies and removes non-causal components from representations for test distributions. TACT applies data augmentations that preserve causal features while varying non-causal ones. By analyzing the changes in the representations using Principal Component Analysis, TACT identifies the highest variance directions associated with non-causal features. It trims the representations by removing their projections on the identified directions, and uses the trimmed representations for the predictions. During adaptation, TACT continuously tracks and refines these directions to get a better estimate of non-causal features. We theoretically analyze the effectiveness of this approach and empirically validate TACT on real-world out-of-distribution benchmarks. TACT consistently outperforms state-of-the-art methods by a significant margin.



Paperid:206
Authors:Liang Wang, Haonan Chen, Nan Yang, Xiaolong Huang, Zhicheng Dou, Furu Wei
Abstract:
Abstract:This paper introduces an approach for training o1-like RAG models that retrieve and reason over relevant information step by step before generating the final answer. Conventional RAG methods usually perform a single retrieval step before the generation process, which limits their effectiveness in addressing complex queries due to imperfect retrieval results. In contrast, our proposed method, CoRAG (Chain-of-Retrieval Augmented Generation), allows the model to dynamically reformulate the query based on the evolving state. To train CoRAG effectively, we utilize rejection sampling to automatically generate intermediate retrieval chains, thereby augmenting existing RAG datasets that only provide the correct final answer. At test time, we propose various decoding strategies to scale the model's test-time compute by controlling the length and number of sampled retrieval chains. Experimental results across multiple benchmarks validate the efficacy of CoRAG, particularly in multi-hop question answering tasks, where we observe more than $10$ points improvement in EM score compared to strong baselines. On the KILT benchmark, CoRAG establishes a new state-of-the-art performance across a diverse range of knowledge-intensive tasks. Furthermore, we offer comprehensive analyses to understand the scaling behavior of CoRAG, laying the groundwork for future research aimed at developing factual and grounded foundation models.



Paperid:207
Authors:Fabian J. Roth, Dominik K. Klein, Maximilian Kannapinn, Jan Peters, Oliver Weeger
Abstract:
In recent years, nonlinear dynamic system identification using artificial neural networks has garnered attention due to its broad potential applications across science and engineering. However, purely data-driven approaches often struggle with extrapolation and may yield physically implausible forecasts. Furthermore, the learned dynamics can exhibit instabilities, making it difficult to apply such models safely and robustly.This article introduces stable port-Hamiltonian neural networks, a machine learning architecture that incorporates physical biases of energy conservation and dissipation while ensuring global Lyapunov stability of the learned dynamics. Through illustrative and real-world examples, we demonstrate that these strong inductive biases facilitate robust learning of stable dynamics from sparse data, while avoiding instability and surpassing purely data-driven approaches in accuracy and physically meaningful generalization.Furthermore, the model's applicability and potential for data-driven surrogate modeling are showcased on multi-physics simulation data.



Paperid:208
Authors:Yichi Zhang, Shengwei Xu, Grant Schoenebeck, David Pennock
Abstract:
Eliciting reliable human feedback is essential for many machine learning tasks, such as learning from noisy labels and aligning AI systems with human preferences. Peer prediction mechanisms incentivize truthful reporting without ground truth verification by scoring agents based on correlations with peers. Traditional mechanisms, which ensure that truth-telling maximizes the \textbf{expected scores} in equilibrium, can elicit honest information while assuming agents' utilities are \textbf{linear functions} of their scores. However, in practice, non-linear payment rules are usually preferred, or agents' utilities are inherently non-linear.We propose \emph{stochastically dominant truthfulness (SD-truthfulness)} as a stronger guarantee: the score distribution of truth-telling stochastically dominates all other strategies, incentivizing truthful reporting for a wide range of monotone utility functions. Our first observation is that no existing peer prediction mechanism naturally satisfies this criterion without strong assumptions. A simple solution - rounding scores into binary lotteries — can enforce SD-truthfulness, but often degrades \emph{sensitivity}, a key property related to fairness and statistical efficiency. We demonstrate how a more careful application of rounding can better preserve sensitivity. Furthermore, we introduce a new enforced agreement (EA) mechanism that is theoretically guaranteed to be SD-truthful in binary-signal settings and, under mild assumptions, empirically achieves the highest sensitivity among all known SD-truthful mechanisms.



Paperid:209
Authors:Ren Yi, Octavian Suciu, Adrian Gascon, Sarah Meiklejohn, Eugene Bagdasarian, Marco Gruteser
Abstract:
We study the ability of language models to reason about appropriate information disclosure--a central aspect of the evolving field of agent privacy. While previous works have focused on evaluating the models' ability to align with human decisions, we examine the role of ambiguity and missing context on model accuracy, e.g., precision and recall, when rendering information sharing decisions. We identify context ambiguity as a crucial barrier for high performance in privacy assessments. By designing a framework for context disambiguation, we show that model-generated decision rationales can reveal ambiguities and that systematically disambiguating context based on these rationales leads to significant accuracy improvements (up to 13.3\% in precision and up to 22.3\% in recall) as well as reductions in prompt sensitivity. Overall, our results indicate that approaches for context disambiguation are a promising way forward to enhance agent privacy reasoning.



Paperid:210
Authors:Sabine Rieder, Stefan Pranger, Debraj Chakraborty, Jan Kretinsky, Bettina Könighofer
Abstract:
Trust in a decision-making system requires both safety guarantees and the ability to interpret and understand its behavior. This is particularly important for learned systems, whose decision-making processes are often highly opaque. Shielding is a prominent model-based technique for enforcing safety in reinforcement learning. However, because shields are automatically synthesized using rigorous formal methods, their decisions are often similarly difficult for humans to interpret. Recently, decision trees became customary to represent controllers and policies. However, since shields are inherently non-deterministic, their decision tree representations become too large to be explainable in practice. To address this challenge, we propose a novel approach for explainable safe RL that enhances trust by providing human-interpretable explanations of the shield's decisions. Our method represents the shielding policy as a hierarchy of decision trees, offering top-down, case-based explanations. At design time, we use a world model to analyze the safety risks of executing actions in given states. Based on this risk analysis, we construct both the shield and a high-level decision tree that classifies states into risk categories (safe, critical, dangerous, unsafe), providing an initial explanation of why a given situation may be safety-critical. At runtime, we generate localized decision trees that explain which actions are allowed and why others are deemed unsafe. Altogether, our method facilitates the explainability of the safety aspect in the safe-by-shielding reinforcement learning. Our framework requires no additional information beyond what is already used for shielding, incurs minimal overhead, and can be readily integrated into existing shielded RL pipelines. In our experiments, we compute explanations using decision trees that are several orders of magnitude smaller than the original shield.



Paperid:211
Authors:Neil Marchant, Benjamin Rubinstein
Abstract:
Reuse of data in adaptive workflows poses challenges regarding overfitting and the statistical validity of results. Previous work has demonstrated that interacting with data via differentially private algorithms can mitigate overfitting, achieving worst-case generalization guarantees with asymptotically optimal data requirements. However, such past work assumes data isstaticand cannot accommodate situations where datagrowsover time. In this paper we address this gap, presenting the first generalization bounds for adaptive analysis on dynamic data. We allow the analyst to adaptivelyscheduletheir queries conditioned on the current size of the data, in addition to previous queries and responses. We also incorporate time-varying empirical accuracy bounds and mechanisms, allowing for tighter guarantees as data accumulates. In a batched query setting, the asymptotic data requirements of our bound grows with the square-root of the number of adaptive queries, matching prior works' improvement over data splitting for the static setting. We instantiate our bound for statistical queries with the clipped Gaussian mechanism, where it empirically outperforms baselines composed from static bounds.



Paperid:212
Authors:Bill Psomas, George Retsinas, Nikos Efthymiadis, Panagiotis Filntisis, Yannis Avrithis, Petros Maragos, Ondrej Chum, Giorgos Tolias
Abstract:
The development of composed image retrieval (CIR), a popular research direction in image retrieval, where a combined visual and textual query is used, is held back by the absence of high-quality training and evaluation data.We introduce a new evaluation dataset, i-CIR, which, unlike existing datasets, focuses on an instance-level class definition. The goal is to retrieve images that contain the same particular object as the visual query, presented under a variety of modifications defined by textual queries. Its design and curation process keep the dataset compact to facilitate future research, while maintaining its challenge through a semi-automated selection of hard negative examples. We show that i-CIR is as challenging as using a database with 20 million distractor images, yet contains on average only 3,000 in query-specific databases.To overcome the challenge of obtaining clean, diverse, and suitable training data, we leverage pre-trained vision-and-language models (VLMs) in a training-free approach called BASIC. The method separately estimates query-image-to-image and query-text-to-image similarities, performing late fusion to upweight images that satisfy both queries, while downweighting those that exhibit high similarity with only one of the two. Each individual similarity is further improved by a set of components that are simple, efficient and intuitive.



Paperid:213
Authors:Guillem Brasó, Aljosa Osep, Laura Leal-Taixé
Abstract:
Uniform downsampling remains the de facto standard for reducing spatial resolution in vision backbones. In this work, we propose an alternative design built around a content-aware spatial grouping layer that dynamically assigns tokens to a reduced set based on image boundaries and their semantic content. Stacking our grouping layer across consecutive backbone stages results in hierarchical segmentation that arisesnativelyin the feature extraction process, resulting in our coined Native Segmentation Vision Transformer.We show that a careful design of our architecture enables the emergence of strong segmentation masks solely from grouping layers, that is, without additional segmentation-specific heads. This sets the foundation for a new paradigm ofnative, backbone-level segmentation, which enables strong zero-shot results without mask supervision, as well as a minimal and efficient standalone model design for downstream segmentation tasks.



Paperid:214
Authors:Cheng Zheng, William Koch, Baiang Li, Felix Heide
Abstract:
Hierarchical structures of motion exist across research fields, including computer vision, graphics, and robotics, where complex dynamics typically arise from coordinated interactions among simpler motion components. Existing methods to model such dynamics typically rely on manually-defined or heuristic hierarchies with fixed motion primitives, limiting their generalizability across different tasks. In this work, we propose a general hierarchical motion modeling method that learns structured, interpretable motion relationships directly from data. Our method represents observed motions using graph-based hierarchies, explicitly decomposing global absolute motions into parent-inherited patterns and local motion residuals. We formulate hierarchy inference as a differentiable graph learning problem, where vertices represent elemental motions and directed edges capture learned parent-child dependencies through graph neural networks. We evaluate our approach on two settings: 1D hierarchical motion reconstruction and dynamic 3D scene deformation via Gaussian splatting. Experimental results show that our method produces more realistic and interpretable deformations compared to the baseline on dynamic 3D Gaussian splatting scenes. By providing an adaptable, data-driven hierarchical modeling paradigm, our method offers a formulation applicable to a broad range of motion-centric tasks.



Paperid:215
Authors:Anita Keshmirian, Razan Baltaji, Babak Hemmatian, Hadi Asghari, Lav Varshney
Abstract:
Moral judgment is integral to large language model (LLM) alignment and social reasoning. As multi-agent systems gain prominence, it becomes crucial to understand how LLMs function collectively during collaboration, compared to individual agents. In human moral judgment, group deliberation leads to a utilitarian boost: a tendency to endorse norm violations that maximize benefits for the greatest number of people despite harms. We study whether a similar dynamic emerges in multi-agent LLM systems. We tested six models on well-established sets of moral dilemmas across two conditions: (1) Solo, where models reasoned independently, and (2) Group, where they engaged in multi-turn discussions in pairs or triads. In personal moral dilemmas, where agents must decide to directly harm one individual to maximize the utility for others, all models found moral violations to be more acceptable when part of a group than individually, similar to human experiments. Some models endorsed actions that maximized overall well-being, even if they benefited strangers over familiar individuals. Others became more willing to violate moral norms in groups. However, while human groups show a similar action bias, the mechanism for their utilitarian boost differs from LLMs. Whereas the human shift comes from heightened sensitivity to decision outcomes, LLM groups show either reduced norm sensitivity or enhanced impartiality. This suggests that while the surface behavior of LLM collectives mimics human group reasoning, the underlying drivers differ. We discuss the implications for AI alignment, multi-agent design, and artificial moral reasoning.



Paperid:216
Authors:Bartlomiej Sobieski, Matthew Tivnan, Yuang Wang, Siyeop yoon, Pengfei Jin, Dufan Wu, Quanzheng Li, Przemyslaw Biecek
Abstract:
Solving inverse problems—recovering signals from incomplete or noisy measurements—is fundamental in science and engineering. Score-based generative models (SGMs) have recently emerged as a powerful framework for this task. Two main paradigms have formed: unsupervised approaches that adapt pretrained generative models to inverse problems, and supervised bridge methods that train stochastic processes conditioned on paired clean and corrupted data. While the former typically assume knowledge of the measurement model, the latter have largely overlooked this structural information. We introduce System-embedded Diffusion Bridge Models (SDBs), a new class of supervised bridge methods that explicitly embed the known linear measurement system into the coefficients of a matrix-valued SDE. This principled integration yields consistent improvements across diverse linear inverse problems and demonstrates robust generalization under system misspecification between training and deployment, offering a promising solution to real-world applications.



Paperid:217
Authors:Reza Shirkavand, Shangqian Gao, Peiran Yu, Heng Huang
Abstract:
Abstract:We study cost-aware routing for large language models across diverse and dynamic pools of models. Existing approaches often overlook prompt-specific context, rely on expensive model profiling, assume a fixed set of experts, or use inefficient trial-and-error strategies. We introduce Cost-Spectrum Contrastive Routing (CSCR), a lightweight framework that maps both prompts and models into a shared embedding space to enable fast, cost-sensitive selection. CSCR uses compact, fast-to-compute logit footprints for open-source models and perplexity fingerprints for black-box APIs. A contrastive encoder is trained to favor the cheapest accurate expert within adaptive cost bands. At inference time, routing reduces to a single $k$‑NN lookup via a FAISS index, requiring no retraining when the expert pool changes and enabling microsecond latency. Across multiple benchmarks, CSCR consistently outperforms baselines, improving the accuracy–cost tradeoff by up to 25\%, while generalizing robustly to unseen LLMs and out-of-distribution prompts.



Paperid:218
Authors:Le Yu, Jun Wu, Bo Gou, Xiangde Min, Lei Zhang, Zhang Yi, Tao He
Abstract:
Abstract:Depthwise-separable convolution has emerged as a significant milestone in the lightweight development of Convolutional Neural Networks (CNNs) over the past decade. This technique consists of two key components: depthwise convolution, which captures spatial information, and pointwise convolution, which enhances channel interactions. In this paper, we propose a novel method to lightweight CNNs through the discretization of Ordinary Differential Equations (ODEs). Specifically, we optimize depthwise-separable convolution by replacing the pointwise convolution with a discrete ODE module, termed the \emph{\textbf{C}hannelwise \textbf{O}DE \textbf{S}olver (COS)}. The COS module is constructed by a simple yet efficient direct differentiation Euler algorithm, using learnable increment parameters. This replacement reduces parameters by over $98.36$\% compared to conventional pointwise convolution. By integrating COS into MobileNet, we develop a new extra lightweight network called MobileODE. With carefully designed basic and inverse residual blocks, the resulting MobileODEV1 and MobileODEV2 reduce channel interaction parameters by $71.0$\% and $69.2$\%, respectively, compared to MobileNetV1, while achieving higher accuracy across various tasks, including image classification, object detection, and semantic segmentation. The code is anonymous at {\url{https://anonymous.4open.science/r/MobileODE-5FB2}}.



Paperid:219
Authors:XiaoPeng Yu, Kefan Su, Zongqing Lu
Abstract:
Planning under opponent uncertainty is a fundamental challenge in multi-agent environments, where an agent must act while inferring the hidden policies of its opponents. Existing type-based methods rely on manually defined behavior classes and struggle to scale, while model-free approaches are sample-inefficient and lack a principled way to incorporate uncertainty into planning. We propose Quantized Opponent Models (QOM), which learn a compact catalog of opponent types via a quantized autoencoder and maintain a Bayesian belief over these types online. This posterior supports both a belief-weighted meta-policy and a Monte-Carlo planning algorithm that directly integrates uncertainty, enabling real-time belief updates and focused exploration. Experiments show that QOM achieves superior performance with significantly lower search cost, offering a tractable and effective solution for belief-aware planning.



Paperid:220
Authors:Wenhong Zhu, Ruobing Xie, Weinan Zhang, Rui Wang
Abstract:
Realignment becomes necessary when a language model (LM) fails to meet expected performance. We propose a flexible realignment framework that supports quantitative control of alignment degree during training and inference. This framework incorporatesTraining-time Realignment (TrRa), which efficiently realigns the reference model by leveraging the controllable fusion of logits from both the reference and already aligned models. For example, TrRa reduces token usage by54.63%on DeepSeek-R1-Distill-Qwen-1.5B without any performance degradation, outperforming DeepScaleR-1.5B’s33.86%. To complement TrRa during inference, we introduce a layer adapter that enablessmooth Inference-time Realignment (InRa). This adapter is initialized to perform an identity transformation at the bottom layer and is inserted preceding the original layers. During inference, input embeddings are simultaneously processed by the adapter and the original layer, followed by the remaining layers, and then controllably interpolated at the logit level. We upgraded DeepSeek-R1-Distill-Qwen-7B from a slow-thinking model to one that supports both fast and slow thinking, allowing flexible alignment control evenduring inference. By encouraging deeper reasoning, it even surpassed its original performance.



Paperid:221
Authors:Francesco Emanuele Stradi, Anna Lunghi, Matteo Castiglioni, Alberto Marchesi, Nicola Gatti
Abstract:
Abstract:In constrained MDPs (CMDPs) with adversarial rewards and constraints, a known impossibility result prevents any algorithm from attaining sublinear regret and constraint violation, when competing against a best-in-hindsight policy that satisfies the constraints on average. In this paper, we show how to ease such a negative result, by considering settings that generalize both stochastic CMDPs and adversarial ones. We provide algorithms whose performances smoothly degrade as the level of environment adverseness increases. In this paper, we show that this negative result can be eased in CMDPs with non-stationary rewards and constraints, by providing algorithms whose performances smoothly degrade as non-stationarity increases. Specifically, they attain $\widetilde{\mathcal{O}} (\sqrt{T} + C)$ regret and positive constraint violation under bandit feedback, where $C$ measures the adverseness of rewards and constraints. This is $C = \Theta(T)$ in the worst case, coherently with the impossibility result for adversarial CMDPs. First, we design an algorithm with the desired guarantees when $C$ is known. Then, in the case $C$ is unknown, we obtain the same results by embedding multiple instances of such an algorithm in a general meta-procedure, which suitably selects them so as to balance the trade-off between regret and constraint violation.



Paperid:222
Authors:Ankit Sonthalia, Arnas Uselis, Seong Joon Oh
Abstract:
We study whether visual embedding models capture continuous, ordinal attributes along linear directions, which we termrank axes. We define a model asrankablefor an attribute if projecting embeddings onto such an axis preserves the attribute's order. Across 7 popular encoders and 9 datasets with attributes like age, crowd count, head pose, aesthetics, and recency, we find that many embeddings are inherently rankable. Surprisingly, a small number of samples, or even just two extreme examples, often suffice to recover meaningful rank axes, without full-scale supervision. These findings open up new use cases for image ranking in vector databases and motivate further study into the structure and learning of rankable embeddings.



Paperid:223
Authors:David Abel, Michael Bowling, Andre Barreto, Will Dabney, Shi Dong, Steven Hansen, Anna Harutyunyan, Khimya Khetarpal, Clare Lyle, Razvan Pascanu, Georgios Piliouras, Doina Precup, Jonathan Richens, Mark Rowland, Tom Schaul, Satinder Singh
Abstract:
Agents are minimally entities that are influenced by their past observations and act to influence future observations. This latter capacity is captured by empowerment, which has served as a vital framing concept across artificial intelligence and cognitive science. This former capacity, however, is equally foundational: In what ways, and to what extent, can an agent be influenced by what it observes? In this paper, we ground this concept in a universal agent-centric measure that we refer to as plasticity, and reveal a fundamental connection to empowerment. Following a set of desiderata on a suitable definition, we define plasticity using a new information theoretic quantity we call the generalized directed information. We show that this new quantity strictly generalizes the directed information introduced by Massey (1990) while preserving all of its desirable properties. Our first finding is that plasticity is the mirror of empowerment: The agent’s plasticity is identical to the empowerment of the environment, and vice versa. Our second finding establishes a tension between the plasticity and empowerment of an agent, suggesting that agent design needs to be mindful of both characteristics. We explore the implications of these findings, and suggest that plasticity, empowerment, and their relationship are essential to understanding agency.



Paperid:224
Authors:Joo Chan Lee, Jong Hwan Ko, Eunbyung Park
Abstract:
3D Gaussian Splatting (3DGS) has emerged as a powerful representation for real-time, high-performance rendering, enabling a wide range of applications. However, representing 3D scenes with numerous explicit Gaussian primitives imposes significant storage and memory overhead. Recent studies have shown that high-quality rendering can be achieved with a substantially reduced number of Gaussians when represented with high-precision attributes. Nevertheless, existing 3DGS compression methods still rely on a relatively large number of Gaussians, focusing primarily on attribute compression. This is because a smaller set of Gaussians becomes increasingly sensitive to lossy attribute compression, leading to severe quality degradation. Since the number of Gaussians is directly tied to computational costs, it is essential to reduce the number of Gaussians effectively rather than only optimizing storage. In this paper, we propose Optimized Minimal Gaussians representation (OMG), which significantly reduces storage while using a minimal number of primitives. First, we determine the distinct Gaussian from the near ones, minimizing redundancy without sacrificing quality. Second, we propose a compact and precise attribute representation that efficiently captures both continuity and irregularity among primitives. Additionally, we propose a sub-vector quantization technique for improved irregularity representation, maintaining fast training with a negligible codebook size. Extensive experiments demonstrate that OMG reduces storage requirements by nearly 50% compared to the previous state-of-the-art and enables 600+ FPS rendering while maintaining high rendering quality.



Paperid:225
Authors:Ruien Li, Yanhao Wang
Abstract:
Abstract:We consider the problem of diversity maximization from the perspective of individual fairness: given a set $P$ of $n$ points in a metric space, we aim to extract a subset $S$ of size $k$ from $P$ so that (1) the diversity of $S$ is maximized and (2) $S$ is \emph{individually fair} in the sense that every point in $P$ has at least one of its $\frac{n}{k}$-nearest neighbors as its ``representative'' in $S$. We propose $\left(O(1), 3\right)$-bicriteria approximation algorithms for the individually fair variants of the three most common diversity maximization problems, namely, max-min diversification, max-sum diversification, and sum-min diversification. Specifically, the proposed algorithms provide a set of points where every point in the dataset finds a point within a distance at most $3$ times its distance to its $\frac{n}{k}$-nearest neighbor while achieving a diversity value at most $O(1)$ times lower than the optimal solution. Numerical experiments on real-world and synthetic datasets demonstrate that the proposed algorithms generate individually fairer solutions than non-private ones and incur only a modest utility loss for diversity maximization.



Paperid:226
Authors:Dan Andrei Calian, Greg Farquhar, Iurii Kemaev, Luisa Zintgraf, Matteo Hessel, Jeremy Shar, Junhyuk Oh, András György, Tom Schaul, Jeff Dean, Hado van Hasselt, David Silver
Abstract:
The quality of foundation models depends heavily on their training data.Consequently, great efforts have been put into dataset curation.Yet most approaches rely on manual tuning of coarse-grained mixtures of large buckets of data, or filtering by hand-crafted heuristics.An approach that is ultimately more scalable (let alone more satisfying) is to \emph{learn} which data is actually valuable for training.This type of meta-learning could allow more sophisticated, fine-grained, and effective curation.Our proposed \emph{DataRater} is an instance of this idea. It estimates the value of training on any particular data point. This is done by meta-learning using `meta-gradients', with the objective of improving training efficiency on held out data.In extensive experiments across a range of model scales and datasets, we find that using our DataRater to filter data is highly effective, resulting in significantly improved compute efficiency.



Paperid:227
Authors:Harin Lee, Min-hwan Oh
Abstract:
We study the problem of infrequent exploration in linear bandits, addressing a significant yet overlooked gap between fully adaptive exploratory methods (e.g., UCB and Thompson Sampling), which explore potentially at every time step, and purely greedy approaches, which require stringent diversity assumptions to succeed. Continuous exploration can be impractical or unethical in safety-critical or costly domains, while purely greedy strategies typically fail without adequate contextual diversity. To bridge these extremes, we introduce a novel and practical framework, INFEX, explicitly designed for infrequent exploration. INFEX executes a base exploratory policy according to a given schedule while predominantly choosing greedy actions in between. Despite its simplicity, our theoretical analysis demonstrates that INFEX achieves instance-dependent regret matching standard provably efficient algorithms, provided the exploration frequency exceeds a logarithmic threshold. Additionally, INFEX is highly generic, allowing seamless integration of any fully adaptive exploratory method, thus facilitating wide applicability and ease of adoption. By restricting intensive exploratory computations to infrequent intervals, our approach can also enhance computational efficiency. Empirical evaluations confirm our theoretical findings, showing state-of-the-art regret performance and runtime improvements over existing methods.



Paperid:228
Authors:Junwei Ma, Valentin Thomas, Rasa Hosseinzadeh, Alex Labach, Jesse Cresswell, Keyvan Golestan, Guangwei Yu, Anthony Caterini, Maks Volkovs
Abstract:
Tabular data is one of the most ubiquitous sources of information worldwide, spanning a wide variety of domains. This inherent heterogeneity has slowed the development of Tabular Foundation Models (TFMs) capable of fast generalization to unseen datasets. In-Context Learning (ICL) has recently emerged as a promising solution for TFMs, enabling dynamic adaptation to new tasks without additional tuning. While many studies have attempted to re-purpose large language models for tabular ICL, they have had limited success, so recent works have focused on developing tabular-specific foundation models. In this work, we propose an approach to combine ICL-based retrieval with self supervised learning to train tabular foundation models. We also investigate the utility of real vs. synthetic data for model pre-training, and show that real data can contain useful signal not easily captured in synthetic training. Specifically, we show that incorporating real data during the pre-training phase can lead to significantly faster training and better downstream generalization to unseen data. Our resulting model, TabDPT, achieves top performance on both regression (CTR23) and classification (CC18) benchmarks. Importantly, we also demonstrate that with our pre-training procedure, scaling both model and data size leads to consistent performance improvements that follow power laws. This echoes scaling laws in LLMs and other foundation models, and suggests that Internet-scale TFMs can be achievable. We open-source our full pipeline: inference code including trained model weights can be found \href{https://anonymous.4open.science/r/TabDPT-inference-6D8C/README.md}{here}, and the training code to reproduce experiments can be found \href{https://anonymous.4open.science/r/TabDPT-8586/README.md}{here}.



Paperid:229
Authors:Ehsan Variani, Georg Heigold, Tom Bagby, Ji Ma, Cyril Allauzen, Michael D Riley, Shankar Kumar
Abstract:
Although sound information extraction appear distinct across spectrum of sound classes and technologies, all inherently involve creating some form of "embedding"—be it discrete as in textual tokens or continuous vectors—to encapsulate relevant information from the audio signal for downstream utilization. This unifying framework allows us to re-evaluate sound information extraction by researching the optimality of current task-specific representations, the quality headroom and the potential for a single, robust sound embedding to generalize across diverse applications and sound types. To expedite research in these directions, a standardized evaluation benchmark is indispensable, mirroring the established benchmarks in text and image domains. We present the Massive Sound Embedding Benchmark (MSEB) to serve this purpose. MSEB encompasses realistic tasks and datasets that reflect practical applications across diverse technologies and sound categories. Initial experimental findings indicate substantial headroom for enhancing prevalent information extraction methodologies. We encourage the sound processing community to contribute data and tasks to MSEB and employ it to assess their algorithms for improved overall sound encoding. The MSEB library code is publicly hosted at \url{https://github.com/google-research/mseb}.



Paperid:230
Authors:Joel Ye, Fabio Rizzoglio, Xuan Ma, Adam Smoulder, Hongwei Mao, Gary Blumenthal, William Hockeimer, Nicolas Kunigk, Dalton Moore, Patrick Marino, Raeed Chowdhury, J. Patrick Mayo, Aaron Batista, Steven Chase, Michael Boninger, Charles Greenspon, Andrew B Schwartz, Nicholas Hatsopoulos, Lee Miller, Kristofer Bouchard, Jennifer Collinger, Leila Wehbe, Robert Gaunt
Abstract:
Mapping the relationship between neural activity and motor behavior is a central aim of sensorimotor neuroscience and neurotechnology. While most progress to this end has relied on restricting complexity, the advent of foundation models instead proposes integrating a breadth of data as an alternate avenue for broadly advancing downstream modeling. We quantify this premise for motor decoding from intracortical microelectrode data, pretraining an autoregressive Transformer on 2000 hours of neural population spiking activity paired with diverse motor covariates from over 30 monkeys and humans. The resulting model is broadly useful, benefiting decoding on 8 downstream decoding tasks and generalizing to a variety of neural distribution shifts. However, we also highlight that scaling autoregressive Transformers seems unlikely to resolve limitations stemming from sensor variability and output stereotypy in neural datasets.



Paperid:231
Authors:Chung Nguen, Arash Amini, OSCAR HERNAN MADRID PADILLA
Abstract:
We consider the two-sample testing problem for networks, where the goal is to determine whether two sets of networks originated from the same stochastic model. Assuming no vertex correspondence and allowing for different numbers of nodes, we address a fundamental network testing problem that goes beyond simple adjacency matrix comparisons. We adopt the stochastic block model (SBM) for network distributions, due to their interpretability and the potential to approximate more general models. The lack of meaningful node labels and vertex correspondence translate to a graph matching challenge when developing a test for SBMs.We introduce an efficient algorithm to match estimated network parameters, allowing us to properly combine and contrast information within and across samples, leading to a powerful test. We show that the matching algorithm, and the overall test are consistent, under mild conditions on the sparsity of the networks and the sample sizes, and derive a chi-squared asymptotic null distribution for the test.Through a mixture of theoretical insights and empirical validations, including experiments with both synthetic and real-world data, this study advances robust statistical inference for complex network data.



Paperid:232
Authors:Wenjie Wei, Malu Zhang, Jieyuan (Eric) Zhang, Ammar Belatreche, Shuai Wang, Yimeng Shan, Hanwen Liu, Honglin Cao, Guoqing Wang, Yang Yang, Haizhou Li
Abstract:
Abstract:Spiking Neural Networks (SNNs) offer an energy-efficient paradigm for machine intelligence, but their continued scaling poses challenges for resource-limited deployment. Despite recent advances in binary SNNs, the storage and computational demands remain substantial for large-scale networks. To further explore the compression and acceleration potential of SNNs, we propose Sub-bit Spiking Neural Networks (S$^2$NNs) that represent weights with less than one bit. Specifically, we first establish an S$^2$NN baseline by leveraging the clustering patterns of kernels in well-trained binary SNNs. This baseline is highly efficient but suffers from \textit{outlier-induced codeword selection bias} during training. To mitigate this issue, we propose an \textit{outlier-aware sub-bit weight quantization} (OS-Quant) method, which optimizes codeword selection by identifying and adaptively scaling outliers. Furthermore, we propose a \textit{membrane potential-based feature distillation} (MPFD) method, improving the performance of highly compressed S$^2$NN via more precise guidance from a teacher model. Extensive results on vision and non-vision tasks reveal that S$^2$NN outperforms existing quantized SNNs in both performance and efficiency, making it promising for edge computing applications.



Paperid:233
Authors:Rong Li, Yuhao Dong, Tianshuai Hu, Alan Liang, Youquan Liu, Dongyue Lu, Liang Pan, Lingdong Kong, Junwei Liang, Ziwei Liu
Abstract:
Visual grounding in 3D is the key for embodied agents to localize language-referred objects in open-world environments. However, existing benchmarks are limited to indoor focus, single-platform constraints, and small scale. We introduce 3EED, a multi-platform, multi-modal 3D grounding benchmark featuring RGB and LiDAR data from vehicle, drone, and quadruped platforms. We provide over 134,000 objects and 25,000 validated referring expressions across diverse outdoor scenes -- 10x larger than existing datasets. We develop a scalable annotation pipeline combining vision-language model prompting with human verification to ensure high-quality spatial grounding. To support cross-platform learning, we propose platform-aware normalization and cross-modal alignment techniques, and establish benchmark protocols for in-domain and cross-platform evaluations. Our findings reveal significant performance gaps, highlighting the challenges and opportunities of generalizable 3D grounding. The 3EED dataset and benchmark toolkit are released to advance future research in language-driven 3D embodied perception.



Paperid:234
Authors:Yeahoon Kwon, Yesong Choe, Soungmin Park, Neil Dhir, Sanghack Lee
Abstract:
Abstract:We study the problem of sequential decision-making in environments governed by evolving causal mechanisms. Prior work on structural causal bandits--formulations that integrate causal graphs into multi-armed bandit problems to guide intervention selection--has shown that leveraging the causal structure can reduce unnecessary interventions by identifying possibly-optimal minimal intervention sets (POMISs). However, such formulations fall short in dynamic settings where reward distributions may vary over time, as their static, hence myopic, nature focuses on immediate rewards and overlooks the long-term effects of interventions. In this work, we propose a non-stationary structural causal bandit framework that leverages temporal structural causal models to capture evolving dynamics over time. We characterize how interventions propagate over time by developing graphical tools and assumptions, which form the basis for identifying non-myopic intervention strategies. Within this framework, we devise POMIS$^+$, which captures the existence of variables that contribute to maximizing both immediate and long-term rewards. Our framework provides a principled way to reason about temporally-aware interventions by explicitly modeling information propagation across time. Empirical results validate the effectiveness of our approach, demonstrating improved performance over myopic baselines.



Paperid:235
Authors:Pramith Devulapalli, Changlong Wu, Ananth Grama, Wojciech Szpankowski
Abstract:
Abstract:We study agnostic online learning from continuous-time data streams, a setting that naturally arises in applications such as environmental monitoring, personalized recommendation, and high-frequency trading. Unlike classical discrete-time models, learners in this setting must interact with a continually evolving data stream while making queries and updating models only at sparse, strategically selected times. We develop a general theoretical framework for learning from both *oblivious* and *adaptive* data streams, which may be noisy and non-stationary. For oblivious streams, we present a black-box reduction to classical online learning that yields a regret bound of $T \cdot R(S)/S$ for any class with discrete-time regret $R(S)$, where $T$ is the time horizon and $S$ is the *query budget*. For adaptive streams, which can evolve in response to learner actions, we design a dynamic query strategy in conjunction with a novel importance weighting scheme that enables unbiased loss estimation. In particular, for hypothesis class $\mathcal{H}$ with a finite Littlestone dimension, we establish a tight regret bound of $\tilde{\Theta}(T \cdot \sqrt{\mathsf{Ldim}(\mathcal{H})/S})$ that holds in both settings. Our results provide the first *quantitative* characterization of agnostic learning in continuous-time online environments with limited interaction.



Paperid:236
Authors:Stephen Chung, Wenyu Du, Jie Fu
Abstract:
Recent studies show that the reasoning capabilities of Large Language Models (LLMs) can be improved by applying Reinforcement Learning (RL) to question-answering (QA) tasks in areas such as math and coding. With a long context length, LLMs may learn to perform search, as indicated by the self-correction behavior observed in DeepSeek R1. However, this search behavior is often imprecise and lacks confidence, resulting in long, redundant responses and highlighting deficiencies in intuition and verification. Inspired by the Dual Process Theory in psychology, we introduce a simple modification to the QA task that includes four stages:Fast Thinking, where the LLM must answer within a strict token budget;Verification, where the model evaluates its initial response;Slow Thinking, where it refines the initial response with more deliberation; andSummarization, where it distills the refinement from the previous stage into precise steps. Our proposed task improves average accuracy from 24.9\% to 27.9\% for Qwen2.5-1.5B, and from 45.9\% to 49.8\% for DeepSeek-R1-Qwen-1.5B. Notably, for Qwen2.5-1.5B, the Fast Thinking mode alone achieves 26.8\% accuracy using fewer than 1000 tokens, demonstrating substantial inference efficiency gains. These findings suggest that intuition and deliberative reasoning are distinct, complementary systems benefiting from targeted training.



Paperid:237
Authors:Xiaoqing Sun, Alessandro Stolfo, Joshua Engels, Ben Wu, Senthooran Rajamanoharan, Mrinmaya Sachan, Max Tegmark
Abstract:
Sparse autoencoders (SAEs) are designed to extract interpretable features from language models by enforcing a sparsity constraint. Ideally, training an SAE would yield latents that are both sparse and semantically meaningful. However, many SAE latents activate frequently (i.e., aredense), raising concerns that they may be undesirable artifacts of the training procedure. In this work, we systematically investigate the geometry, function, and origin of dense latents and show that they are not only persistent but often reflect meaningful model representations. We first demonstrate that dense latents tend to form antipodal pairs that reconstruct specific directions in the residual stream, and that ablating their subspace suppresses the emergence of new dense features in retrained SAEs---suggesting that high density features are an intrinsic property of the residual space. We then introduce a taxonomy of dense latents, identifying classes tied to position tracking, context binding, entropy regulation, letter-specific output signals, part-of-speech, and principal component reconstruction. Finally, we analyze how these features evolve across layers, revealing a shift from structural features in early layers, to semantic features in mid layers, and final to output-oriented signals in the last layers of the model. Our findings indicate that dense latents serve functional roles in language model computation and should not be dismissed as training noise.



Paperid:238
Authors:Yifan Zhang, Yifeng Liu, Huizhuo Yuan, Zhen Qin, Yang Yuan, Quanquan Gu, Andrew Yao
Abstract:
Scaling language models to handle longer input sequences typically necessitates large key-value (KV) caches, resulting in substantial memory overhead during inference. In this paper, we propose Tensor Product Attention (TPA), a novel attention mechanism that uses tensor decompositions to represent queries, keys, and values compactly, substantially shrinking the KV cache size at inference time. By factorizing these representations into contextual low-rank components and seamlessly integrating with Rotary Position Embedding (RoPE), TPA achieves improved model quality alongside memory efficiency. Based on TPA, we introduce the Tensor ProducT ATTenTion Transformer (T6), a new model architecture for sequence modeling. Through extensive empirical evaluation on language modeling tasks, we demonstrate that T6 surpasses or matches the performance of standard Transformer baselines including Multi-Head Attention (MHA), Multi-Query Attention (MQA), Grouped-Query Attention (GQA), and Multi-Head Latent Attention (MLA) across various metrics, including perplexity and a range of established evaluation benchmarks. Notably, TPA's memory efficiency and computational efficiency at decoding stage enables processing longer sequences under fixed resource constraints, addressing a critical scalability challenge in modern language models.



Paperid:239
Authors:Sunpill Kim, Seunghun Paik, Chanwoo Hwang, Minsu Kim, Jae Hong Seo
Abstract:
Adversarial attacks on face recognition systems (FRSs) pose serious security and privacy threats, especially when these systems are used for identity verification. In this paper, we propose a novel method for generating adversarial faces—synthetic facial images that are visually distinct yet recognized as a target identity by the FRS. Unlike iterative optimization-based approaches (e.g., gradient descent or other iterative solvers), our method leverages the structural characteristics of the FRS feature space. We figure out that individuals sharing the same attribute (e.g., gender or race) form an attributed subsphere. By utilizing such subspheres, our method achieves both non-adaptiveness and a remarkably small number of queries. This eliminates the need for relying on transferability and open-source surrogate models, which have been a typical strategy when repeated adaptive queries to commercial FRSs are impossible. Despite requiring only a single non-adaptive query consisting of 100 face images, our method achieves a high success rate of over 93% against AWS’s CompareFaces API at its default threshold. Furthermore, unlike many existing attacks that perturb a given image, our method can deliberately produce adversarial faces that impersonate the target identity while exhibiting high-level attributes chosen by the adversary.



Paperid:240
Authors:Wenxin Zhang, Yueying Li, Tianyi Peng, Ciamac C Moallemi
Abstract:
Prompt caching is critical for reducing latency and cost in LLM inference---OpenAI and Anthropic report up to 50–90\% cost savings through prompt reuse. Despite its widespread success, little is known about what constitutes an optimal prompt caching policy, particularly when optimizing tail latency—a metric of central importance to practitioners. The widely used Least Recently Used (LRU) policy can perform arbitrarily poor on this metric, as it is oblivious to the heterogeneity of conversation lengths. To address this gap, we propose Tail-Optimized LRU, a simple two-line modification that reallocates KV cache capacity to prioritize high-latency conversations by evicting cache entries unlikely to affect future turns. Though the implementation is simple, we prove its optimality under a natural stochastic model of conversation dynamics, providing the first theoretical justification for LRU in this setting---a result that may be of independent interest to the caching community. Experimentally, on real conversation data WildChat~\cite{zhao2024wildchat}, Tail-Optimized LRU achieves up to 27.5\% reduction in P90 tail Time to First Token latency and 23.9\% in P95 tail latency compared to LRU, along with up to 40\% decrease in SLO violations of 200ms. We believe this provides a practical and theoretically grounded option for practitioners seeking to optimize tail latency in real-world LLM deployments.



Paperid:241
Authors:Howe Tissue, Venus Wang, Lu Wang
Abstract:
Abstract:We find that the cross-entropy loss curves of neural language models empirically adhere to a scaling law with learning rate (LR) annealing over training steps: $$L(s) = L_0 + A\cdot S_1^{-\alpha} - C\cdot S_2,$$ where $L(s)$ is the validation loss at step $s$, $S_1$ is the area under the LR curve, $S_2$ is the LR annealing area, and $L_0$, $A$, $C$, $\alpha$ are constant parameters. This formulation accounts for two main effects: (1) power-law scaling over data size, and (2) the additional loss reduction during LR annealing. Unlike previous studies that only fit losses at final steps, our formulation captures the entire training curve, allowing for parameter fitting using losses from any training step. Applying the scaling law with LR annealing and fitting only one or two training curves, we can accurately predict the loss at any given step under any learning rate scheduler (LRS). This approach significantly reduces computational cost in formulating scaling laws while providing more accuracy and expressiveness. Extensive experiments demonstrate that our findings hold across a range of hyper-parameters and model architectures and can extend to scaling effect of model sizes. Moreover, our formulation provides accurate theoretical insights into empirical results observed in numerous previous studies, particularly those focusing on LR schedule and annealing. We believe that this work is promising to enhance the understanding of LLM training dynamics while democratizing scaling laws, and it is helpful to guide both research and industrial participants in refining training strategies for further LLMs.



Paperid:242
Authors:Mona Schirmer, Metod Jazbec, Christian Andersson Naesseth, Eric Nalisnick
Abstract:
Encountering shifted data at test time is a ubiquitous challenge when deploying predictive machine learning models. Test-time adaptation (TTA) methods aim to address this issue by continuously adapting a deployed model using only unlabeled test data. While TTA can help extend the model's deployment lifespan, there are scenarios where, despite adaptation, the drop in the model's performance remains significant enough to warrant taking the model offline and retraining. To detect such failure cases, we propose pairing TTA with risk monitoring frameworks that track predictive performance and raise alerts when predefined performance criteria are violated. Specifically, we extend existing monitoring tools based on sequential testing with confidence sequences to accommodate scenarios where the model is updated at test time and no test labels are available to estimate the performance metrics of interest. Our extensions unlock the application of rigorous statistical risk monitoring in TTA and we demonstrate applicability of our proposed TTA monitoring framework across a representative set of TTA methods, datasets and distribution shift types.



Paperid:243
Authors:Bruno Mlodozeniec, Isaac Reid, Sam Power, David Krueger, Murat Erdogdu, Richard Turner, Roger Grosse
Abstract:
Randomness is an unavoidable part of training deep learning models, yet something that traditional training data attribution algorithms fail to rigorously account for. They ignore the fact that, due to stochasticity in the initialisation and batching, training on the same dataset can yield different models. In this paper, we address this shortcoming through introducingdistributionaltraining data attribution (d-TDA), the goal of which is to predict how the distribution of model outputs (over training runs) depends upon the dataset. We demonstrate the practical significance of d-TDA in experiments, e.g. by identifying training examples that drastically change the distribution of some target measurement without necessarily changing the mean. Intriguingly, we also find thatinfluence functions(IFs), a popular but poorly-understood data attribution tool, emerge naturally from our distributional framework as the limit to unrolled differentiation – without requiring restrictive convexity assumptions. This provides a new mathematical motivation for their efficacy in deep learning, and helps to characterise their limitations.



Paperid:244
Authors:Zhenghan Fang, Mateo Diaz, Sam Buchanan, Jeremias Sulam
Abstract:
Abstract:Diffusion models have quickly become some of the most popular and powerful generative models for high-dimensional data. The key insight that enabled their development was the realization that access to the score---the gradient of the log-density at different noise levels---allows for sampling from data distributions by solving a reverse-time stochastic differential equation (SDE) via forward discretization, and that popular denoisers allow for unbiased estimators of this score. In this paper, we demonstrate that an alternative, backward discretization of these SDEs, using proximal maps in place of the score, leads to theoretical and practical benefits. We leverage recent results in _proximal matching_ to learn proximal operators of the log-density and, with them, develop Proximal Diffusion Models (`ProxDM`). Theoretically, we prove that $\widetilde{\mathcal O}(d/\sqrt{\varepsilon})$ steps suffice for the resulting discretization to generate an $\varepsilon$-accurate distribution w.r.t. the KL divergence.Empirically, we show that two variants of `ProxDM` achieve significantly faster convergence within just a few sampling steps compared to conventional score-matching methods.



Paperid:245
Authors:Moongyu Jeon, Sangwoo Shin, Dongjae Jeon, Albert No
Abstract:
Abstract:We present an information-theoretic framework for discrete diffusion models, introducing principled estimators of log-likelihood based on score-matching losses.Inspired by the I-MMSE identity for Gaussian denoising models, we derive analogous results for discrete diffusion.Specifically, we establish the Information–Minimum Denoising Score Entropy (I-MDSE) relation,which links the mutual information between data and its diffused counterpart to the minimum denoising score entropy (DSE) loss.We extend this theory to the masked diffusion setting and introduce the Information–Minimum Denoising Cross-Entropy (I-MDCE) relation,which connects cross-entropy-based losses to information flow in absorbing diffusion processes.These results yield closed-form decompositions of the negative log-likelihood in both time-integral and time-free forms,enabling efficient estimation of data likelihood.We further show that these decompositions naturally extend to conditional likelihoods, allowing direct estimation of $\log p_0(\mathbf{x}^\text{target}|\mathbf{x}^\text{context})$ in structured tasks such as prompt–response modeling.Empirical results on synthetic and real-world data confirm that the proposed approach accurately estimates log-probabilities of discrete inputs.



Paperid:246
Authors:Mouxiang Chen, Binyuan Hui, Zeyu Cui, Jiaxi Yang, Dayiheng Liu, Jianling Sun, Junyang Lin, Zhongxin Liu
Abstract:
Abstract:It is commonly believed that scaling language models should commit a significant space or time cost, by increasing the parameters (parameter scaling) or output tokens (inference-time scaling). We introduce another and more inference-efficient scaling paradigm: increasing the model's parallel computation during both training and inference time. We apply $P$ diverse and learnable transformations to the input, execute forward passes of the model in parallel, and dynamically aggregate the $P$ outputs. This method, namely parallel scaling (ParScale), scales parallel computation by reusing existing parameters and can be applied to any model structure, optimization procedure, data, or task. We theoretically propose a new scaling law and validate it through large-scale pre-training (among $10^6$ GPU hours), which shows that a model with $P$ parallel streams is similar to scaling the parameters by $\mathcal O(\log P)$ while showing superior inference efficiency. For example, ParScale can use up to 22$\times$ less memory increase and 6$\times$ less latency increase compared to parameter scaling that achieves the same performance improvement. It can also recycle an off-the-shelf pre-trained model into a parallelly scaled one by post-training on a small amount of tokens, further reducing the training budget. The new scaling law we discovered potentially facilitates the deployment of more powerful models in low-resource scenarios, and provides an alternative perspective for the role of computation in machine learning. Code is available in the supplementary material.



Paperid:247
Authors:Junhong Shen, Kushal Tirumala, Michihiro Yasunaga, Ishan Misra, Luke Zettlemoyer, LILI YU, Chunting Zhou
Abstract:
Most existing image tokenizers encode images into a fixed number of tokens or patches, overlooking the inherent variability in image complexity and introducing unnecessary computate overhead for simpler images. To address this, we propose Content-Adaptive Tokenizer (CAT), which dynamically adjusts representation capacity based on the image content and encodes simpler images into fewer tokens. We design (1) a caption-based evaluation system that leverages LLMs to predict content complexity and determine the optimal compression ratio for an image, and (2) a novel nested VAE architecture that performs variable-rate compression in a single model.Trained on images with varying complexity, CAT achieves an average of 15% reduction in rFID across seven detail-rich datasets containing text, humans, and complex textures. On natural image datasets like ImageNet and COCO, it reduces token usage by 18% while maintaining high-fidelity reconstructions. We further evaluate CAT on two downstream tasks. For image classification, CAT consistently improves top-1 accuracy across five datasets spanning diverse domains. For image generation, it boosts training throughput by 23% on ImageNet, leading to more efficient learning and improved FIDs over fixed-token baselines.



Paperid:248
Authors:Damien Ferbach, Katie Everett, Gauthier Gidel, Elliot Paquette, Courtney Paquette
Abstract:
We investigate scaling laws for stochastic momentum algorithms on the power law random features model, parameterized by data complexity, target complexity, and model size. When trained with a stochastic momentum algorithm, our analysis reveals four distinct loss curve shapes determined by varying data-target complexities. While traditional stochastic gradient descent with momentum (SGD-M) yields identical scaling law exponents to SGD, dimension-adapted Nesterov acceleration (DANA) improves these exponents by scaling momentum hyperparameters based on model size and data complexity. This outscaling phenomenon, which also improves compute-optimal scaling behavior, is achieved by DANA across a broad range of data and target complexities, while traditional methods fall short. Extensive experiments on high-dimensional synthetic quadratics validate our theoretical predictions and large-scale text experiments with LSTMs show DANA's improved loss exponents over SGD hold in a practical setting.



Paperid:249
Authors:Jaemoo Choi, Yongxin Chen, Molei Tao, Guan-Horng Liu
Abstract:
Recently, there has been significant progress in learning-based diffusion samplers, which aim to sample from given unnormalized density. These methods typically follow one of two paradigms: (i) formulating sampling as an unbiased stochastic optimal control (SOC) problem using a canonical reference process, or (ii) refining annealed path measures through importance-weighted sampling (IWS). Although annealing approaches have advantages in guiding samples toward high-density regions, reliance on importance sampling leads to high variance and limited scalability in practice. In this paper, we introduce theNon-equilibrium Annealed Adjoint Sampler (NAAS), a novel SOC-based diffusion sampler that leverages annealed reference dynamics without resorting to importance sampling. NAAS employs a lean adjoint system inspired by adjoint matching, enabling efficient and scalable training. We demonstrate the effectiveness of our approach across a range of tasks, including sampling from classical energy landscapes and molecular Boltzmann distribution.



Paperid:250
Authors:Iasonas Nikolaou, Miltiadis Stouras, Efstratios Ioannidis, Evimaria Terzi
Abstract:
Abstract:Given a collection of monotone submodular functions, the goal of Two-Stage Submodular Maximization (2SSM) (Balkanski et al. 2016) is to restrict the ground set so an objective selected u.a.r. from the collection attains a high maximal value, on average, when optimized over the restricted ground set. We introduce the Online Two-Stage Submodular Maximization (O2SSM) problem, in which the submodular objectives are revealed in an online fashion. We study this problem for weighted threshold potential functions, a large and important subclass of monotone submodular functions that includes influence maximization, data summarization, and facility location, to name a few. We design an algorithm that achieves sublinear $(1 - 1/e)^2$-regret under general matroid constraints and $(1 - 1/e)(1-e^{-k}k^k/k!)$-regret in the case of uniform matroids of rank $k$; the latter also yields a state-of-the-art bound for the (offline) 2SSM problem. We empirically validate the performance of our online algorithm with experiments on real datasets.



Paperid:251
Authors:Merlijn Krale, Eline Bovy, Maris F. L. Galesloot, Thiago Simão, Nils Jansen
Abstract:
Robust partially observable Markov decision processes (RPOMDPs) model partially observable sequential decision-making problems where an agent must be \emph{robust} against a range of dynamics. RPOMDPs can be viewed as two-player games between an agent, who picks actions, and nature, who adversarially picks dynamics. Evaluating an agent policy requires finding an adversarial nature policy, which is computationally challenging. In this paper, we advance the evaluation of agent policies for RPOMDPs in three ways. First, we discuss suitable benchmarks. We observe that for some RPOMDPs, we can find an optimal agent policy by considering only subsets of nature policies, making them easier to solve. We formalize this concept of \emph{solvability} and construct three benchmarks that are only solvable for expressive sets of nature policies. Second, we describe a provably sound method to evaluate agent policies for RPOMDPs by solving an equivalent MDP. Third, we lift two well-known POMDP upper value bounds to RPOMDPs, which can be used to efficiently approximate the optimality gap of a policy and serve as baselines. Our experimental evaluation shows that (1) our proposed benchmarks cannot be solved by assuming naive nature policies, (2) our method of evaluating policies is accurate, and (3) the approximations provide solid baselines for evaluation.



Paperid:252
Authors:Yujia Zheng, Zhuokai Zhao, Zijian Li, Yaqi Xie, Mingze Gao, Lizhu Zhang, Kun Zhang
Abstract:
Natural language has long enabled human cooperation, but its lossy, ambiguous, and indirect nature limits the potential of collective intelligence. While machines are not subject to these constraints, most LLM-based multi-agent systems still rely solely on natural language, exchanging tokens or their embeddings. To go beyond language, we introduce a new paradigm,thought communication, which enables LLM agents to interact directly mind-to-mind, akin to telepathy. To uncover these latent thoughts in a principled way, we formalize the process as a general latent variable model, where agent states are generated by an unknown function of underlying thoughts. We prove that, in a nonparametric setting without auxiliary information, both shared and private latent thoughts between any pair of agents can be identified. Moreover, the global structure of thought sharing, including which agents share which thoughts and how these relationships are structured, can also be recovered with theoretical guarantees. Guided by the established theory, we develop a framework that extracts latent thoughts from all agents prior to communication and assigns each agent the relevant thoughts, along with their sharing patterns. Experiments on both synthetic and real-world benchmarks validate the theory and demonstrate the collaborative advantages of communication beyond language. We hope this work offers insight into the potential of thought communication, as superhuman intelligence will ultimately depend on superhuman communication.



Paperid:253
Authors:Maria Marrium, Arif Mahmood, Muhammad Haris Khan, M. Shakeel, Wenxiong Kang
Abstract:
Graph Neural Networks (GNNs) have achieved remarkable success in a wide range of applications. However, when trained on limited or low-diversity datasets, GNNs are prone to overfitting and memorization, which impacts their generalization. To address this, graph data augmentation (GDA) has become a crucial task to enhance the performance and generalization of GNNs. Traditional GDA methods employ simple transformations that result in limited performance gains. Although recent diffusion-based augmentation methods offer improved results, they are sparse, task-specific, and constrained by class labels. In this work, we propose a more general and effective diffusion-based GDA framework that is task-agnostic and label-free.For better training stability and reduced computational cost, we employ a graph variational auto-encoder (GVAE) to learn a compact latent graph representation. A diffusion model is used in the learned latent space to generate both consistent and diverse augmentations. For a fixed augmentation budget, our algorithm selects a subset of samples that would benefit the most from the augmentation.To further improve performance, we also perform test-time augmentation, leveraged by the label-free nature of our method.Thanks to the efficient utilization of GVAE and latent diffusion, our algorithm significantly enhances machine learning safety measures, including calibration, robustness to corruptions, and prediction consistency. Moreover, our method has shown improved robustness against four types of adversarial attacks and achieves better generalization performance. To demonstrate the effectiveness of the proposed method, we compare it with 30 existing methods on 12 benchmark datasets across node classification, link prediction, and graph classification in various learning settings, including semi-supervised, supervised, and long-tailed data distributions. The code will soon be made publicly available.



Paperid:254
Authors:Samuel Pfisterer, Florian Grötschla, Luca Lanzendörfer, Florian Yan, Roger Wattenhofer
Abstract:
Recent progress in speech processing has highlighted that high-quality performance across languages requires substantial training data for each individual language. While existing multilingual datasets cover many languages, they often contain insufficient data for each language, leading to models trained on these datasets to exhibit poor performance on most supported languages. Our work addresses this challenge by introducing a scalable pipeline for constructing speech datasets from parliamentary recordings. The proposed pipeline includes robust components for media retrieval and a two-stage alignment algorithm designed to handle non-verbatim transcripts and long-form audio. Applying this pipeline to recordings from 22 European parliaments, we extract over 61k hours of aligned speech segments, achieving substantial per-language coverage with 19 languages exceeding 1k hours and 22 languages exceeding 500 hours of high-quality speech data. We obtain an average 41.8\% reduction in word error rates over baselines when finetuning an existing ASR model on our dataset, demonstrating the usefulness of our approach.



Paperid:255
Authors:Neehar Kondapaneni, Oisin Mac Aodha, Pietro Perona
Abstract:
We propose a method for discovering and visualizing the differences between two learned representations, enabling more direct and interpretable model comparisons. We validate our method, which we call Representational Differences Explanations (RDX), by using it to compare models with known conceptual differences and demonstrate that it recovers meaningful distinctions where existing explainable AI (XAI) techniques fail. Applied to state-of-the-art models on challenging subsets of the ImageNet and iNaturalist datasets, RDX reveals both insightful representational differences and subtle patterns in the data. Although comparison is a cornerstone of scientific analysis, current tools in machine learning, namely post hoc XAI methods, struggle to support model comparison effectively. Our work addresses this gap by introducing an effective and explainable tool for contrasting model representations.



Paperid:256
Authors:Cheng Shi, Yizhou Yu, Sibei Yang
Abstract:
This study identifies that visual-related functional decoding is distributed across different decoder layers in Multimodal Large Language Models (MLLMs). Typically, each function, such as counting, grounding, or OCR recognition, narrows down to two or three layers, which we define as Vision Function Layers (VFL). Additionally, the depth and its order of different VFLs exhibits a consistent pattern across different MLLMs, which is well-aligned with human behaviors (e.g., recognition occurs first, followed by counting, and then grounding). These findings are derived from Visual Token Swapping, our novel analytical framework that modifies targeted KV cache entries to precisely elucidate layer-specific functions during decoding. Furthermore, these insights offer substantial utility in tailoring MLLMs for real-world downstream applications. For instance, when LoRA training is selectively applied to VFLs whose functions align with the training data, VFL-LoRA not only outperform full-LoRA but also prevent out-of-domain function forgetting. Moreover, by analyzing the performance differential on training data when particular VFLs are ablated, VFL-select automatically classifies data by function, enabling highly efficient data selection to directly bolster corresponding capabilities. Consequently, VFL-select surpasses human experts in data selection, and achieves 98% of full-data performance with only 20% of the original dataset. This study delivers deeper comprehension of MLLM visual processing, fostering the creation of more efficient, interpretable, and robust models. Code will be released.



Paperid:257
Authors:Imad Aouali
Abstract:
Efficient decision-making in contextual bandits with large action spaces is challenging, as methods lacking additional prior information may suffer from computational and statistical inefficiencies. In this work, we leverage pre-trained diffusion models as priors to capture complex action distributions and introduce a diffusion-based decision framework for contextual bandits. We develop practical algorithms to efficiently approximate posteriors under diffusion priors, enabling flexible decision-making strategies. Empirical evaluations demonstrate the effectiveness and versatility of our approach across diverse contextual bandit settings.



Paperid:258
Authors:Sean McLeish, John Kirchenbauer, David Miller, Siddharth Singh, Abhinav Bhatele, Micah Goldblum, Ashwinee Panda, Tom Goldstein
Abstract:
Scaling laws are typically fit using a family of models with a narrow range of frozen hyper-parameter choices. In this work we study scaling laws using multiple architectural shapes and hyperparameter choices, highlighting their impact on resulting prescriptions.As a primary artifact of our research, we release the Gemstones: an open-source scaling law dataset, consisting of over 4000 checkpoints from transformers with up to 2 billion parameters and diverse architectural shapes; including ablations over learning rate and cooldown.Our checkpoints enable more complex studies of scaling, such as analyzing the relationship between width and depth.By examining our model suite, we find that the prescriptions of scaling laws can be highly sensitive to the experimental design process and the specific model checkpoints used during fitting.



Paperid:259
Authors:Ali Behrouz, Peilin Zhong, Vahab Mirrokni
Abstract:
Over more than a decade there has been an extensive research effort on how to effectively utilize recurrent models and attention. While recurrent models aim to compress the data into a fixed-size memory (called hidden state), attention allows attending to the entire context window, capturing the direct dependencies of all tokens. This more accurate modeling of dependencies, however, comes with a quadratic cost, limiting the model to a fixed-length context. We present a neural long-term memory module that learns to memorize historical context and helps attention to attend to the current context while utilizing long-past information. We show that this neural memory has the advantage of fast parallelizable training. From a memory perspective, we argue that attention due to its limited context but accurate dependency modeling performs as a short-term memory, while neural memory due to its ability to memorize the data, acts as a long-term, more persistent, memory. Based on these two modules, we introduce a new family of architectures, called Titans, and present three variants to address how one can effectively incorporate memory into this architecture. Our experimental results on language modeling, common-sense reasoning, and time series tasks show that Titans are effective compared to baselines, while they can effectively scale to larger context window in needle-in-haystack tasks.



Paperid:260
Authors:Yuchuan Tian, Hanting Chen, Mengyu Zheng, Yuchen Liang, Chao Xu, Yunhe Wang
Abstract:
Abstract:Representation Alignment (REPA) that aligns Diffusion Transformer (DiT) hidden-states with ViT visual encoders has proven highly effective in DiT training, demonstrating superior convergence properties, but it has not been validated on the canonical diffusion U-Net architecture that shows faster convergence compared to DiTs. However, adapting REPA to U-Net architectures presents unique challenges: (1) different block functionalities necessitate revised alignment strategies; (2) spatial-dimension inconsistencies emerge from U-Net's spatial downsampling operations; (3) space gaps between U-Net and ViT hinder the effectiveness of tokenwise alignment. To encounter these challenges, we propose U-REPA, a representation alignment paradigm that bridges U-Net hidden states and ViT features as follows: Firstly, we propose via observation that due to skip connection, the middle stage of U-Net is the best alignment option. Secondly, we propose upsampling of U-Net features after passing them through MLPs. Thirdly, we observe difficulty when performing tokenwise similarity alignment, and further introduces a manifold loss that regularizes the relative similarity between samples. Experiments indicate that the resulting U-REPA could achieve excellent generation quality and greatly accelerates the convergence speed. With CFG guidance interval, U-REPA could reach FID<1.5 in 200 epochs or 1M iterations on ImageNet 256 $\times$ 256, and needs only half the total epochs to perform better than REPA under \textit{sd-vae-ft-ema}.



Paperid:261
Authors:Mila Gorecki, Moritz Hardt
Abstract:
Two narratives about machine learning ecosystems grew out of recent algorithmic fairness discourse. In one, dubbed \emph{monoculture}, algorithmic ecosystems tend toward homogeneity akin to a single model making all decisions. Individuals then face the risk of systematic exclusion with no recourse. In the other, \emph{model multiplicity}, many models solve the same task with similar accuracy, causing excessive variation in outcomes. Both narratives are compelling, yet, seemingly at odds: model multiplicity can’t exist in a strict monoculture. In this work, we conduct a comprehensive empirical evaluation to test both claims. We work from the premise that increasingly decision makers will use large language models for consequential prediction tasks. We therefore examine 50 language models, open source models ranging in size from 1B to 141B parameters and state-of-the-art commercial models, under 4 different prompt variations, and across 6 different prediction tasks. Evaluating both new and old quantitative measures of monoculture and multiplicity, we find the empirical landscape sits between the two extremes. Each narrative finds some empirical support, but neither is dominant. Systematic exclusion with no recourse is rare, but model similarity is real. Even when starting from a single model, prompt variation induces some diversity in predictions. Our results contribute critical empirical grounding to ongoing debates and point toward a middle ground between monoculture and multiplicity as the most realistic outcome.



Paperid:262
Authors:Manuel Iglesias-Alonso, Felix Schur, Julius von Kügelgen, Jonas Peters
Abstract:
We consider the problem of estimating a causal effect in a multi-domain setting. The causal effect of interest is confounded by an unobserved confounder and can change between the different domains. We assume that we have access to a proxy of the hidden confounder and that all variables are discrete or categorical. We propose methodology to estimate the causal effect in the target domain, where we assume to observe only the proxy variable. Under these conditions, we prove identifiability (even when treatment and response variables are continuous). We introduce two estimation techniques, prove consistency, and derive confidence intervals. The theoretical results are supported by simulation studies and a real-world example studying the causal effect of website rankings on consumer choices.



Paperid:263
Authors:Amber Hu, Henry Smith, Scott Linderman
Abstract:
Latent stochastic differential equation (SDE) models are important tools for the unsupervised discovery of dynamical systems from data, with applications ranging from engineering to neuroscience. In these complex domains, exact posterior inference of the latent state path is typically intractable, motivating the use of approximate methods such as variational inference (VI). However, existing VI methods for inference in latent SDEs often suffer from slow convergence and numerical instability. Here, we propose SDE Inference via Natural Gradients (SING), a method that leverages natural gradient VI to efficiently exploit the underlying geometry of the model and variational posterior. SING enables fast and reliable inference in latent SDE models by approximating intractable integrals and parallelizing computations in time. We provide theoretical guarantees that SING will approximately optimize the intractable, continuous-time objective of interest. Moreover, we demonstrate that better state inference enables more accurate estimation of nonlinear drift functions using, for example, Gaussian process SDE models. SING outperforms prior methods in state inference and drift estimation on a variety of datasets, including a challenging application to modeling neural dynamics in freely behaving animals. Altogether, our results illustrate the potential of SING as a tool for accurate inference in complex dynamical systems, especially those characterized by limited prior knowledge and non-conjugate structure.



Paperid:264
Authors:Maitreyi Swaroop, Tamar Krishnamurti, Bryan Wilder
Abstract:
We study the problem of selecting a limited number of features to observe such that models trained on those features can perform well simultaneously across multiple subpopulations. This problem has applications in settings where collecting each feature is costly, e.g. requiring adding survey questions or physical sensors, and we must be able to use the selected features to create high-quality downstream models for different populations. Our method frames the problem as a continuous relaxation of traditional variable selection using a noising mechanism, without requiring backpropagation through model training processes. By optimizing over the variance of a Bayes-optimal predictor, we develop a model-agnostic framework that balances overall performance of downstream prediction across populations. We validate our approach through experiments on both synthetic datasets and real-world data.



Paperid:265
Authors:Jinheng Xie, Zhenheng Yang, Mike Zheng Shou
Abstract:
This paper presents improved native unified multimodal models that leverage autoregressive modeling and flow matching. Built upon a 3D causal variational autoencoder space, unified visual representations are constructed through a dual-path of spatial (-temporal) fusion, enabling scalability across image and video modalities while ensuring effective multimodal understanding and generation. Based on a language model, autoregressive modeling and flow matching are natively applied to the language head and flow head, respectively, to facilitate text token prediction and image/video generation. A two-stage training recipe is designed to effectively learn and scale to larger models. The resulting model demonstrates versatility in handling a wide range of multimodal understanding and generation tasks across diverse modalities, including text, images, and videos. The training code and pre-trained models will be fully open-sourced.



Paperid:266
Authors:Luca Zanella, Massimiliano Mancini, Yiming Wang, Alessio Tonioni, Elisa Ricci
Abstract:
Given a task and a set of steps composing it, Video Step Grounding (VSG) aims to detect which steps are performed in a video. Standard approaches for this task require a labeled training set (e.g., with step-level annotations or narrations), which may be costly to collect. Moreover, they process the full video offline, limiting their applications for scenarios requiring online decisions. Thus, in this work, we explore how to perform VSG online and without training. We achieve this by exploiting the zero-shot capabilities of recent Large Multimodal Models (LMMs). In particular, we use LMMs to predict the step associated with a restricted set of frames, without access to the whole video. We show that this online strategy without task-specific tuning outperforms offline and training-based models. Motivated by this finding, we develop Bayesian Grounding with Large Multimodal Models (BAGLM), further injecting knowledge of past frames into the LMM-based predictions. BAGLM exploits Bayesian filtering principles, modeling step transitions via (i) a dependency matrix extracted through large language models and (ii) an estimation of step progress. Experiments on three datasets show superior performance of BAGLM over state-of-the-art training-based offline methods.



Paperid:267
Authors:Rees Chang, Angela Pak, Alex Guerra, Ni Zhan, Nick Richardson, Elif Ertekin, Ryan Adams
Abstract:
Accelerating inverse design of crystalline materials with generative models has significant implications for a range of technologies. Unlike other atomic systems, 3D crystals are invariant to discrete groups of isometries called the space groups. Crucially, these space group symmetries are known to heavily influence materials properties. We propose SGEquiDiff, a crystal generative model which naturally handles space group constraints with space group invariant likelihoods. SGEquiDiff consists of an SE(3)-invariant, telescoping discrete sampler of crystal lattices; permutation-invariant, transformer-based autoregressive sampling of Wyckoff positions, elements, and numbers of symmetrically unique atoms; and space group equivariant diffusion of atomic coordinates. We show that space group equivariant vector fields automatically live in the tangent spaces of the Wyckoff positions. SGEquiDiff achieves state-of-the-art performance on standard benchmark datasets as assessed by quantitative proxy metrics and quantum mechanical calculations.



Paperid:268
Authors:Fan-Ming Luo, Lei Yuan, Yang Yu
Abstract:
Privileged learning efficiently tackles high-dimensional, partially observable decision-making problems by first training a privileged policy (PP) on low-dimensional privileged observations, and then deriving a deployment policy (DP) either by imitating the PP or coupling it with an observation encoder. However, since the DP relies on local and partial observations, a behavioral divergence (BD) often emerges between the DP and the PP, ultimately degrading deployment performance. A promising strategy is to train a PP to learn the optimal behaviors attainable under the DP’s observation space by applying reward penalties in regions with large BD. However, producing these behaviors is challenging for the PP because they rely on the DP’s information-gathering progress, which is invisible to the PP. In this paper, we quantify the DP’s information-gathering progress by estimating the prediction uncertainty of privileged observations reconstructed from partial observations, and accordingly propose the framework of Uncertainty-Sensitive Privileged Learning (USPL). USPL feeds this uncertainty estimation to the PP and combines reward transformation with privileged-observation blurring, driving the PP to choose actions that actively reduce uncertainty and thus gather the necessary information. Experiments across nine tasks demonstrate that USPL significantly reduces the behavioral discrepancies, achieving superior deployment performance compared to baselines. Additional visualization results show that the DP accurately quantifies its uncertainty, and the PP effectively adapts to uncertainty variations.



Paperid:269
Authors:Xiaohao Liu, Xiaobo Xia, See-Kiong Ng, Tat-Seng Chua
Abstract:
Multimodal Contrastive Learning (MCL) advances in aligning different modalities and generating multimodal representations in a joint space. By leveraging contrastive learning across diverse modalities, large-scale multimodal data enhances representational quality. However, a critical yet often overlooked challenge remains: multimodal data is rarely collected in a single process, and training from scratch is computationally expensive. Instead, emergent multimodal data can be used to optimize existing models gradually, i.e., models are trained on a sequence of modality pair data. We define this problem as Continual Multimodal Contrastive Learning (CMCL), an underexplored yet crucial research direction at the intersection of multimodal and continual learning. In this paper, we formulate CMCL through two specialized principles of stability and plasticity. We theoretically derive a novel optimization-based method, which projects updated gradients from dual sides onto subspaces where any gradient is prevented from interfering with the previously learned knowledge. Two upper bounds provide theoretical insights on both stability and plasticity in our solution. Beyond our theoretical contributions, we conduct experiments on multiple datasets by comparing our method against advanced continual learning baselines. The empirical results further support our claims and demonstrate the efficacy of our method. The code will be publicly available.



Paperid:270
Authors:Travis Pence, Daisuke Yamada, Vikas Singh
Abstract:
Abstract:Contemporary large models often exhibit behaviors suggesting the presence of low-level primitives that compose into modules with richer functionality, but these fundamental building blocks remain poorly understood. We investigate this compositional structure in linear layers by asking: \textit{can we identify/synthesize linear transformations from a minimal set of geometric primitives?} Using Clifford algebra, we show that linear layers can be expressed as compositions of bivectors---geometric objects encoding oriented planes---and introduce a differentiable algorithm that decomposes them into products of rotors. This construction uses only $\mathcal{O}(\log^2 d)$ parameters, versus $\mathcal{O}(d^2)$ required by dense matrices. Applied to the key, query, and value projections in LLM attention layers, our rotor-based layers match the performance of strong baselines such as block-Hadamard and low-rank approximations. Our findings provide an algebraic perspective on how these geometric primitives can compose into higher-level functions within deep models.



Paperid:271
Authors:Declan Kutscher, David Chan, Yutong Bai, Trevor Darrell, Ritwik Gupta
Abstract:
Abstract:Transformers require inputs to be represented as one-dimensional sequences, and in vision, this typically involves flattening images using a fixed row-major (raster-scan) order. While full self-attention is permutation-equivariant, modern long-sequence transformers increasingly rely on architectural approximations that break this invariance and introduce sensitivity to patch ordering. We show that patch order significantly affects model performance in such settings, with simple alternatives like column-major or Hilbert curves yielding notable accuracy shifts. Motivated by this, we propose _REOrder_, a two-stage framework for discovering task-optimal patch orderings. First, we derive an information-theoretic prior by evaluating the compressibility of various patch sequences. Then, we learn a policy over permutations by optimizing a Plackett-Luce policy using REINFORCE. This approach enables efficient learning in a combinatorial permutation space. _REOrder_ improves top-1 accuracy over row-major ordering on ImageNet-1K by up to $3.01\%$ and Functional Map of the World by $13.35\%$. Code is available at [this link](https://anonymous.4open.science/r/patch-order-8C3D/).



Paperid:272
Authors:Zirui Wang, Jiayi Zhang, Tianwei Guan, Yuhan Zhou, Xingyuan Li, Minjing Dong, Jinyuan Liu
Abstract:
Image fusion is a fundamental and important task in computer vision, aiming to combine complementary information from different modalities to fuse images. In recent years, diffusion models have made significant developments in the field of image fusion. However, diffusion models often require complex computations and redundant inference time, which reduces the applicability of these methods. To address this issue, we propose RFfusion, an efficient one-step diffusion model for image fusion based on Rectified Flow. We incorporate Rectified Flow into the image fusion task to straighten the sampling path in the diffusion model, achieving one-step sampling without the need for additional training, while still maintaining high-quality fusion results. Furthermore, we propose a task-specific variational autoencoder (VAE) architecture tailored for image fusion, where the fusion operation is embedded within the latent space to further reduce computational complexity. To address the inherent discrepancy between conventional reconstruction-oriented VAE objectives and the requirements of image fusion, we introduce a two-stage training strategy. This approach facilitates the effective learning and integration of complementary information from multi-modal source images, thereby enabling the model to retain fine-grained structural details while significantly enhancing inference efficiency. Extensive experiments demonstrate that our method outperforms other state-of-the-art methods in terms of both inference speed and fusion quality.



Paperid:273
Authors:Pranjal Awasthi, Anupam Gupta, Ravi Kumar
Abstract:
Length generalization, the ability of sequence models to generalize to sequences longer than those encountered during training, remains a fundamental challenge for Transformers, especially in tasks requiring algorithmic reasoning. Existing theoretical understanding of length generalization is limited, often providing asymptotic results or focusing on specific problem classes or architectural variants; empirical solutions frequently rely on ad-hoc and often fragile techniques. In this work we introduce a novel framework for analyzing and proving length generalization bounds under specified, verifiable assumptions. A key outcome of the theory is the identification of a natural set of auxiliary tasks, intricately related to the primary task structure, such that strong performance on these auxiliary tasks, alongside the primary task, provably guarantees length generalization within the framework. This motivates a multi-task training procedure that explicitly optimizes performance on both the primary and the identified auxiliary tasks. Empirical evaluations on a variety of synthetic benchmarks known to be challenging for length generalization, including sequence sorting, and reversal, demonstrate that the proposed method yields significant improvements in generalization to substantially longer sequences.



Paperid:274
Authors:Timur Mudarisov, Mikhail Burtsev, Tatiana Petrova, Radu State
Abstract:
This paper investigates the limitations of the normalization in attention mechanisms. We begin with a theoretical framework that enables the identification of the model's selective ability and the geometric separation involved in token selection. Our analysis includes explicit bounds on distances and separation criteria for token vectors under softmax scaling. Through experiments with pre-trained GPT-2 model, we empirically validate our theoretical results and analyze key behaviors of the attention mechanism. Notably, we demonstrate that as the number of selected tokens increases, the model's ability to distinguish informative tokens declines, often converging toward a uniform selection pattern. We also show that gradient sensitivity under softmax normalization presents challenges during training, especially at low temperature settings. These findings advance current understanding of softmax-based attention mechanism and motivate the need for more robust normalization and selection strategies in future attention architectures.



Paperid:275
Authors:Jie Shao, Jianxin Wu
Abstract:
Despite the impressive performance of large language models (LLMs), the process of endowing them with new capabilities---such as mathematical reasoning---remains largely empirical and opaque. A critical open question is whether reasoning abilities stem from the entire model, specific modules, or are merely artifacts of overfitting. In this work, we hypothesize that the reasoning capabilities in well-trained LLMs are primarily attributed to the output projection module (oproj) in the Transformer’s multi-head self-attention (MHSA) module. To support this hypothesis, we introduce Stethoscope for Networks (SfN), a suite of diagnostic tools designed to probe and analyze the internal behaviors of LLMs. Using SfN, we provide both circumstantial and empirical evidence suggesting that oproj plays a central role in enabling reasoning, whereas other modules contribute more to fluent dialogue. These findings offer a new perspective on LLM interpretability and open avenues for more targeted training strategies, potentially enabling more efficient and specialized LLMs.



Paperid:276
Authors:Haoxuan Li, Chuan Zhou, Lina Yao, Mingming Gong
Abstract:
In recommendation systems, implicit feedback data can be automatically recorded and is more common than explicit feedback data. However, implicit feedback poses two challenges for relevance prediction, namely (a)positive-unlabeled: negative feedback does not necessarily imply low relevance and (b)missing not at random(MNAR): items that are popular or frequently recommended tend to receive more clicks than other items, even if the user does not have a significant interest in them. Existing methods either ignore the MNAR problem or neglect the intrinsic positive-unlabeled causal mechanism of implicit feedback, which can face inaccurate relevance predictions or inflated biases and variance when modeling the exposure status. In this paper, we formulate the implicit feedback problem as a counterfactual estimation problem with missing treatment variables. Prediction of the relevance in implicit feedback is equivalent to answering the counterfactual question that "whether a user would click a specific item if exposed to it?". To solve the counterfactual question, we propose Counterfactual Implicit Feedback (Counter-IF) prediction approach that divides the user-item pairs into four disjoint groups, namely definitely positive (DP), highly exposed (HE), highly unexposed (HU) and unknown (UN) groups. Specifically, Counter-IF first performs missing treatment imputation with different confidence level from raw implicit feedback, then estimate the counterfactual outcomes via causal representation learning that combines pointwise loss and pairwise loss based on the user-item pairs stratification. We also theoretically derive the generalization bound of the learned model in terms of estimating the user true relevance. Extensive experiments are conducted on publicly available real-world datasets to demonstrate the effectiveness of our approach.



Paperid:277
Authors:Zeqian Ju, Dongchao Yang, Jianwei Yu, Shen Kai, YICHONG LENG, Zhengtao Wang, Songxiang Liu, Xinyu Zhou, Tao Qin, Xiangyang Li, Xu Tan
Abstract:
Recent advances in text-to-speech synthesis have achieved notable success in generating high-quality short utterances for individual speakers. However, these systems still face challenges when extending their capabilities to long, multi-speaker, and spontaneous dialogues, typical of real-world scenarios such as podcasts. These limitations arise from two primary challenges: 1) long speech: podcasts typically span several minutes, exceeding the upper limit of most existing work; 2) spontaneity: podcasts are marked by their spontaneous, oral nature, which sharply contrasts with formal, written contexts; existing works often fall short in capturing this spontaneity. In this paper, we propose a solution for high-quality zero-shot podcast generation, aiming to synthesize spontaneous podcast-style speech from text-only sources (e.g., stories, technical reports, news in TXT, PDF, or Web URL formats) using the voices of unseen speakers. To enable long audio generation, we employ a language model with parameter, data, and context scaling to process sequences in an innovative format designed for modeling entire multi-speaker, multi-turn speech interactions. To enhance spontaneity, we observe that ASR transcripts capture spontaneous speech details (e.g., filler words indicating hesitations, and specific punctuation and spaces reflecting breathing pauses), suggesting that these transcripts can serve as a partial indicator of speech spontaneity. Building upon this assumption, we utilize a script generation module to generate scripts incorporating these spontaneous elements. Experiments show our system outperforms baselines, with notable improvements in contextual coherence and spontaneity.



Paperid:278
Authors:Hanshi Wang, Yuhao Xu, Zekun Xu, Jin Gao, Yufan Liu, Weiming Hu, Ke Wang, Zhipeng Zhang
Abstract:
The established redundancy in visual tokens within large vision–language models (LVLMs) allows for pruning to effectively reduce their substantial computational demands. Empirical evidence from previous works indicates that visual tokens in later decoder stages receive less attention than shallow layers. Then, previous methods typically employ heuristics layer-specific pruning strategies where, although the number of tokens removed may differ across decoder layers, the overall pruning schedule is fixed and applied uniformly to all input samples and tasks, failing to align token elimination with the model’s holistic reasoning trajectory. Cognitive science indicates that human visual processing often begins with broad exploration to accumulate evidence before narrowing focus as the target becomes distinct. Our experiments reveal an analogous pattern in LVLMs. This observation strongly suggests that neither a fixed pruning schedule nor a heuristics layer-wise strategy can optimally accommodate the diverse complexities inherent in different inputs. To overcome this limitation, we introduce Complexity-Adaptive Pruning (AdaPrune), which is a training-free, plug-and-play framework that tailors pruning policies to varying sample and task complexities. Specifically, AdaPrune quantifies the mutual information between visual and textual tokens, and then projects this signal to a budget-constrained logistic retention curve. Each such logistic curve, defined by its unique shape, is shown to effectively correspond with the specific complexity of different tasks, and can easily guarantee adherence to a pre-defined computational constraints. We evaluate AdaPrune not only on standard vision-language tasks but also on Vision-Language-Action (VLA) models for autonomous driving. Notably, when applied to LLaVA-1.5-7B, our method prunes 89\% of visual tokens and reduces inference FLOPs by 76.8\%, but still retaining 96.7\% of the original accuracy averaged over all tasks. This corresponds to a 9.1\% improvement over the recent work PDrop (CVPR'2025), demonstrating the effectivenes.



Paperid:279
Authors:Seyed Kamyar Seyed Ghasemipour, Ayzaan Wahid, Jonathan Tompson, Pannag R Sanketi, Igor Mordatch
Abstract:
Foundation models trained on web-scale data have revolutionized robotics, but their application to low-level control remains largely limited to behavioral cloning. Drawing inspiration from the success of the reinforcement learning stage in fine-tuning large language models, we propose a two-stage post-training approach for robotics. The first stage, Supervised Fine-Tuning (SFT), fine-tunes pretrained foundation models using both: a) behavioral cloning, and b) steps-to-go prediction objectives. In the second stage, Self-Improvement, steps-to-go prediction enables the extraction of a well-shaped reward function and a robust success detector, enabling a fleet of robots to autonomously practice downstream tasks with minimal human supervision. Through extensive experiments on real-world and simulated robot embodiments, our novel post-training recipe unveils significant results on Embodied Foundation Models. First, we demonstrate that the combination of SFT and Self-Improvement is significantly more sample-efficient than scaling imitation data collection for supervised learning, and that it leads to policies with significantly higher success rates. Further ablations highlight that the combination of web-scale pretraining and Self-Improvement is the key to this sample-efficiency. Next, we demonstrate that our proposed combination uniquely unlocks a capability not possible by current methods: autonomously practicing and acquiring novel skills that generalize far beyond the behaviors observed in the imitation learning datasets used during training. These findings highlight the transformative potential of combining pretrained foundation models with online Self-Improvement to enable autonomous skill acquisition in robotics.



Paperid:280
Authors:Rishi Sonthalia, Michael Murray, Guido Montufar
Abstract:
This paper investigates low-rank structure in the gradients of the training loss for two-layer neural networks while relaxing the usual isotropy assumptions on the training data and parameters. We consider a spiked data model in which the bulk can be anisotropic and ill-conditioned, we do not require independent data and weight matrices and we also analyze both the mean-field and neural-tangent-kernel scalings. We show that the gradient with respect to the input weights is approximately low rank and is dominated by two rank-one terms: one aligned with the bulk data–residue, and another aligned with the rank one spike in the input data. We characterize how properties of the training data, the scaling regime and the activation function govern the balance between these two components. Additionally, we also demonstrate that standard regularizers, such as weight decay, input noise and Jacobian penalties, also selectively modulate these components. Experiments on synthetic and real data corroborate our theoretical predictions.



Paperid:281
Authors:Su Hyeong Lee, Sidharth Sharma, Manzil Zaheer, Tian Li
Abstract:
Abstract:Adaptive optimization is critical in federated learning, where enabling adaptivity on both the server and client sides has proven essential for achieving optimal performance. However, the scalability of such jointly adaptive systems is often hindered by resource limitations in communication and memory. In this paper, we introduce a class of efficient adaptive algorithms, named $FedAda^2$ and its enhanced version $FedAda^2$++, designed specifically for large-scale, cross-device federated environments. $FedAda^2$ optimizes communication efficiency by avoiding the transfer of preconditioners between the server and clients. Additionally, $FedAda^2$++ extends this approach by incorporating memory-efficient adaptive optimizers on the client side, further reducing on-device memory usage. Theoretically, we demonstrate that $FedAda^2$ and $FedAda^2$++ achieve the same convergence rates for general, non-convex objectives as its more resource-intensive counterparts that directly integrate joint adaptivity. Extensive empirical evaluations on image and text datasets demonstrate both the advantages of joint adaptivity and the effectiveness and efficiency of $FedAda^2$/$FedAda^2$++.



Paperid:282
Authors:Themistoklis Haris, Fabian Spaeh, Spyridon Konstantinos Dragazis, Charalampos Tsourakakis
Abstract:
Abstract:Hitting times provide a fundamental measure of distance in random processes, quantifying the expected number of steps for a random walk starting at node $u$ to reach node $v$. They have broad applications across domains such as network centrality analysis, ranking and recommendation systems, and epidemiology. In this work, we develop local algorithms for estimating hitting times between a pair of vertices $u,v$ without accessing the full graph, overcoming scalability issues of prior global methods. Our first algorithm uses the key insight that hitting time computations can be truncated at the meeting time of two independent random walks from $u$ and $v$. This leads to an efficient estimator analyzed via the Kronecker product graph and Markov Chain Chernoff bounds. We also present an algorithm extending the work of Peng et al. [2021] that introduces a novel adaptation of the spectral cutoff technique to account for the asymmetry of hitting times. This adaptation captures the directionality of the underlying random walk and requires non-trivial modifications to ensure accuracy and efficiency. In addition to the algorithmic upper bounds, we also provide tight asymptotic lower bounds. Finally, we reveal a connection between hitting time estimation and distribution testing, and we validate our algorithms using experiments on both real and synthetic data.



Paperid:283
Authors:Xinsong Ma, Jie Wu, Weiwei Liu
Abstract:
Visual anomaly detection is significant in safety-critical and reliability-sensitive scenarios. Prior studies mainly emphasize the design and training of scoring functions, while little effort has been devoted to constructing decision rules based on these score functions. A recent work Ma et al. (2025) is the first to highlight this issue and proposes the SAC-BL algorithm to address it.This method consists of a strong anomaly constraint (SAC) network and a betting-like (BL) algorithm serving as the decision rule. The SAC-BL algorithm can control the false discovery rate (FDR). However the performance of SAC-BL algorithm on anomalous examples, or its false positive rate (FPR), has not been thoroughly investigated. This paper provides a deeper analysis of this problem and explores how to theoretically reduce its FPR. First, we show that as the number of testing examples tends to infinity, the SAC-BL algorithm performs well on abnormal data if the scores follow the generalized Gaussian-like distribution family. But such conditions about the number of testing examples and the distribution of scores are overly restrictive for the real-world applications. So, we attempt to decrease the FPR of the SAC-BL algorithm under the condition of finite samples for practical anomaly detection. To this end, we redesign the BL algorithm by incorporating a randomization strategy and propose a novel stochastic BL (SBL) algorithm. The combination of the SAC network and the SBL algorithm yields our method, SAC-SBL. Theoretical results show that the SAC-SBL algorithm can achieve smaller FPR than SAC-BL algorithm while controlling its FDR. Finally, extensive experimental results demonstrate the superiority of our method over SAC-BL algorithm on multiple visual anomaly detection benchmarks.



Paperid:284
Authors:Cheng Qian, Emre Can Acikgoz, Qi He, Hongru WANG, Xiusi Chen, Dilek Tur, Gokhan Tur, Heng Ji
Abstract:
Current Large Language Models (LLMs) often undergo supervised fine-tuning (SFT) to acquire tool use capabilities. However, SFT struggles to generalize to unfamiliar or complex tool use scenarios. Recent advancements in reinforcement learning (RL), particularly with R1-like models, have demonstrated promising reasoning and generalization abilities. Yet, reward design for tool use presents unique challenges: multiple tools may be invoked with diverse parameters, and coarse-grained reward signals, such as answer matching, fail to offer the finegrained feedback required for effective learning.In this work, we present the first comprehensive study on reward design for tool selection and application tasks within the RL paradigm. We systematically explore a wide range of reward strategies, analyzing their types, scales, granularity, and temporal dynamics. Building on these insights, we propose a principled reward design tailored for tool use tasks and apply it to train LLMs using RL methods.Empirical evaluations across diverse benchmarks demonstrate that our approach yields robust, scalable, and stable training, achieving a 17\% improvement over base models and a 15\% gain over SFT models. These results highlight the critical role of thoughtful reward design in enhancing the tool use capabilities and generalization performance of LLMs.



Paperid:285
Authors:Letian Peng, Jingbo Shang
Abstract:
This paper introduces Codified Profiles for role-playing, a novel approach that represents character logic as structured, executable functions for behavioral decision-making. Converted by large language model (LLM) from textual profiles, each codified profile defines a set of functions parsebyscene(scene) that output multiple logic-grounded assertions according to scene, using both explicit control structures (e.g., if-then-else) and flexible check_condition(scene, question) functions where each question is a semantically meaningful prompt about the scene (e.g., "Is the character in danger?") discriminated by the role-playing LLM as true, false, or unknown.This explicit representation offers three key advantages over traditional prompt-based textual profiles, which append character descriptions directly into text prompts:(1) Persistence, by enforcing complete and consistent execution of character logic, rather than relying on the model's implicit reasoning; (2) Updatability, through systematic inspection and revision of behavioral logic, which is difficult to track or debug in prompt-only approaches; (3) Controllable Randomness, by supporting stochastic behavior directly within the logic, enabling fine-grained variability that prompting alone struggles to achieve. To validate these advantages, we introduce a new benchmark constructed from 83 characters and 5,141 scenes curated from Fandom, using natural language inference (NLI)-based scoring to compare character responses against ground-truths. Our experiments demonstrate the significant benefits of codified profiles in improving persistence, updatability, and behavioral diversity. Notably, by offloading a significant portion of reasoning to preprocessing, codified profiles enable even 1B-parameter models to perform high-quality role-playing, providing an efficient, lightweight foundation for local deployment of role-play agents.



Paperid:286
Authors:Yue Fan, Xuehai He, Diji Yang, Kaizhi Zheng, Ching-Chen Kuo, Yuting Zheng, Xinze Guan, Xin Wang
Abstract:
Recent studies have demonstrated the efficacy of using Reinforcement Learning (RL) in building reasoning models that articulate chains of thoughts prior to producing final answers. However, despite ongoing advances that aim at enabling reasoning for vision-language tasks, existing open-source visual reasoning models typically generate reasoning content with pure natural language, lacking explicit integration of visual information. This limits their ability to produce clearly articulated and visually grounded reasoning chains. To this end, we propose Grounded Reasoning with Images and Texts (GRIT), a novel method for training MLLMs to think with images. GRIT introduces a grounded reasoning paradigm, in which models generate reasoning chains that interleave natural language and explicit bounding box coordinates. These coordinates point to regions of the input image that the model consults during its reasoning process. Additionally, GRIT is equipped with a reinforcement learning approach, GRPO-GR, built upon the GRPO algorithm. GRPO-GR employs robust rewards focused on the final answer accuracy and format of the grounded reasoning output, which eliminates the need for data with reasoning chain annotations or explicit bounding box labels. As a result, GRIT achieves exceptional data efficiency, requiring as few as 20 image-question-answer triplets from existing datasets. Comprehensive evaluations demonstrate that GRIT effectively trains MLLMs to produce coherent and visually grounded reasoning chains, showing a successful unification of reasoning and grounding abilities. All code, data, and checkpoints will be released.



Paperid:287
Authors:Chaoyue Song, Xiu Li, Fan Yang, Zhongcong XU, Jiacheng Wei, Fayao Liu, Jiashi Feng, Guosheng Lin, Jianfeng Zhang
Abstract:
Modern interactive applications increasingly demand dynamic 3D content, yet the transformation of static 3D models into animated assets constitutes a significant bottleneck in content creation pipelines. While recent advances in generative AI have revolutionized static 3D model creation, rigging and animation continue to depend heavily on expert intervention. We present Puppeteer, a comprehensive framework that addresses both automatic rigging and animation for diverse 3D objects. Our system first predicts plausible skeletal structures via an auto-regressive transformer that introduces a joint-based tokenization strategy for compact representation and a hierarchical ordering methodology with stochastic perturbation that enhances bidirectional learning capabilities. It then infers skinning weights via an attention-based architecture incorporating topology-aware joint attention that explicitly encodes skeletal hierarchical relationships. Finally, we complement these rigging advances with a differentiable optimization-based animation pipeline that generates stable, high-fidelity animations while requiring only a fraction of the computational resources demanded by existing approaches. Extensive evaluations across multiple benchmarks demonstrate that our method significantly outperforms state-of-the-art techniques in both skeletal prediction accuracy and skinning quality. The system robustly processes diverse 3D content, ranging from professionally designed game assets to AI-generated shapes, producing temporally coherent animations devoid of jittering prevalent in existing methods.



Paperid:288
Authors:Carlos Misael Madrid Padilla, OSCAR HERNAN MADRID PADILLA, Sabyasachi Chatterjee
Abstract:
This work examines risk bounds for nonparametric distributional regression estimators. For convex-constrained distributional regression, general upper bounds are established for the continuous ranked probability score (CRPS) and the worst-case mean squared error (MSE) across the domain. These theoretical results are applied to isotonic and trend filtering distributional regression, yielding convergence rates consistent with those for mean estimation. Furthermore, a general upper bound is derived for distributional regression under non-convex constraints, with a specific application to neural network-based estimators. Comprehensive experiments on both simulated and real data validate the theoretical contributions, demonstrating their practical effectiveness.



Paperid:289
Authors:A Sağtekin, Colin Bredenberg, Cristina Savin
Abstract:
How should feedback influence recurrent neural network (RNN) learning? One way to address the known limitations of backpropagation through time is to directly adjust neural activities during the learning process. However, it remains unclear how to effectively use feedback to shape RNN dynamics. Here, we introduce error forcing (EF), where the network activity is guided orthogonally toward the zero-error manifold during learning. This method contrasts with alternatives like teaching forcing, which impose stronger constraints on neural activity and thus induce larger feedback influence on circuit dynamics. Furthermore, EF can be understood from a Bayesian perspective as a form of approximate dynamic inference. Empirically, EF consistently outperforms other learning algorithms across several tasks and its benefits persist when additional biological constraints are taken into account. Overall, EF is a powerful temporal credit assignment mechanism and a promising candidate model for learning in biological systems.



Paperid:290
Authors:Sara Klein, Xiangyuan Zhang, Tamer Basar, Simon Weissmann, Leif Döring
Abstract:
Abstract:In this work, we study $\gamma$-discounted infinite-horizon tabular Markov decision processes (MDPs) and introduce a framework called dynamic policy gradient (DynPG). The framework directly integrates dynamic programming with (any) policy gradient method, explicitly leveraging the Markovian property of the environment. DynPG dynamically adjusts the problem horizon during training, decomposing the original infinite-horizon MDP into a sequence of contextual bandit problems. By iteratively solving these contextual bandits, DynPG converges to the stationary optimal policy of the infinite-horizon MDP. To demonstrate the power of DynPG, we establish its non-asymptotic global convergence rate under the tabular softmax parametrization, focusing on the dependencies on salient but essential parameters of the MDP. By combining classical arguments from dynamic programming with more recent convergence arguments of policy gradient schemes, we prove that softmax DynPG scales polynomially in the effective horizon $(1-\gamma)^{-1}$. Our findings contrast recent exponential lower bound examples for vanilla policy gradient.



Paperid:291
Authors:Daniel Steinberg, Asiri Wijesinghe, Rafael Oliveira, Piotr Koniusz, Cheng Soon Ong, Edwin Bonilla
Abstract:
We propose a new framework called active generation of Pareto sets (A-GPS) for online discrete black-box multi objective optimization (MOO) that learns a generative model of the Pareto set and supports a-posteriori preference conditioning. Our method actively learns a generative model conditioned on high-performance regions (active generation) using amortized variational inference. It uses a class probability estimator (CPE) for predicting Pareto-optimality and conditioning the generative model. Furthermore, motivated by discrete/mixed design problems where we must balance multiple competing objectives, it introduces preference direction vectors to capture subjective trade-offs. Thus, at each iteration, we update a generative model conditioned on Pareto set membershipandalignment with preference directions. Our method yields high-quality Pareto set approximations using only simple CPE guidance, avoids hyper-volume computation, and supports sampling at arbitrary trade-off points without retraining. Empirical results on synthetic functions and protein design benchmarks demonstrate strong sample efficiency and effective incorporation of users' preferences.



Paperid:292
Authors:Masahiro Kohjima
Abstract:
Shuffled regression is the problem of learning regression functions from shuffled data where the correspondence between the input features and target response is unknown. This paper proposes a probabilistic model for shuffled regression called Gaussian Process Shuffled Regression (GPSR). By introducing Gaussian processes as a prior of regression functions in function space via the kernel function, GPSR can express a wide variety of functions in a nonparametric manner while quantifying the uncertainty of the prediction. By adopting the Bayesian evidence maximization framework and a theoretical analysis of the connection between the marginal likelihood/predictive distribution of GPSR and that of standard Gaussian process regression (GPR), we derive an easy-to-implement inference algorithm for GPSR that iteratively applies GPR and updates the input-output correspondence. To reduce computation costs and obtain closed-form solutions for correspondence updates, we also develop a sparse approximate variant of GPSR using its weight space formulation, which can be seen as Bayesian shuffled linear regression with random Fourier features. Experiments on benchmark datasets confirm the effectiveness of our GPSR proposal.



Paperid:293
Authors:In-Jae Lee, Mungyeom Kim, Kwonyoung Ryu, Pierre Musacchio, Jaesik Park
Abstract:
Unsupervised and open-vocabulary 3D object detection has recently gained attention, particularly in autonomous driving, where reducing annotation costs and recognizing unseen objects are critical for both safety and scalability. However, most existing approaches apply designated 3D object detectors and require multiple iterations of self-training for pseudo-ground truth refinement, leading to substantial computational overhead. To address these challenges, we propose OpenBox, a two-stage automatic annotation pipeline that leverages a 2D visual foundation model. In the first stage, OpenBox associates instance-level information from 2D images processed by a visual foundation model with the corresponding 3D point cloud. In the second stage, it categorizes the instances by rigidity and motion state, and generates adaptive bounding boxes using class-specific size statistics. As a result, OpenBox can generate high-quality annotations without requiring self-training. Experiments on the Waymo and Lyft datasets demonstrate improved accuracy and efficiency compared to baselines.



Paperid:294
Authors:Jia Peng Lim, Shawn Tan, XianJun, Davin Choo, Hady Lauw
Abstract:
Abstract:Tokenization is the process of encoding strings into tokens of a fixed vocabulary size, and is widely utilized in Natural Language Processing applications.The leading tokenization algorithm today is Byte Pair Encoding (BPE), which formulates the tokenization problem as a compression problem and tackles it by performing sequences of merges.In this work, we formulate tokenization as an optimization objective, show that it is NP-hard via a simple reduction from vertex cover, and propose a polynomial-time greedy algorithm GreedTok.Our formulation naturally relaxes to the well-studied weighted maximum coverage problem which has a simple $(1 - 1/e)$-approximation algorithm GreedWMC.Through empirical evaluations on real-world corpora, we show that GreedTok outperforms BPE and Unigram on compression and achieves a covering score comparable to GreedWMC. Finally, our extensive pre-training for two transformer-based language models with 1 billion parameters, comparing the choices of BPE and GreedTok as the tokenizer, shows that GreedTok achieves a lower bit per byte even when we control for either the total dataset proportion or total training tokens.



Paperid:295
Authors:Yotam Erel, Olaf Dünkel, Rishabh Dabral, Vladislav Golyanik, Christian Theobalt, Amit Bermano
Abstract:
We introduce a new interpretation of the attention matrix as a discrete-time Markov chain. Our interpretation sheds light on common operations involving attention scores such as selection, summation, and averaging in a unified framework. It further extends them by considering indirect attention, propagated through the Markov chain, as opposed to previous studies that only model immediate effects. Our main observation is that tokens corresponding to semantically similar regions form a set of metastable states, where the attention clusters, while noisy attention scores tend to disperse. Metastable states and their prevalence can be easily computed through simple matrix multiplication and eigenanalysis, respectively. Using these lightweight tools, we demonstrate state-of-the-art zero-shot segmentation. Lastly, we define TokenRank---the steady state vector of the Markov chain, which measures global token importance. We demonstrate that using it brings improvements in unconditional image generation. We believe our framework offers a fresh view of how tokens are being attended in modern visual transformers.



Paperid:296
Authors:Jinda Jia, Cong Xie, Hanlin Lu, Fanjiang Ye, Hao Feng, Daoce Wang, Haibin Lin, Zhi Zhang, Xin Liu
Abstract:
Distributed training often suffers from high communication overhead due to large-scale gradient synchronization. Although gradient compression—particularly at 4-bit or even lower precision—significantly reduces transfer volume, it typically results in sacrifice in precision and degradation of the final model accuracy. In this work, we introduce DUO, a distributed training framework designed to mitigate accuracy degradation incurred by gradient compression without involving additional overhead. DUO achieves this by inserting an additional high-precision gradient synchronization step into a previously computation-only phase, so that its communication is fully hidden by computation.We provide a comprehensive theoretical proof of convergence for DUO and validate its effectiveness through extensive pre-training experiments on GPT models. Our results indicate that DUO effectively restores accuracy when using 4-bit gradient compression, achieving performance comparable to uncompressed training. Remarkably, DUO maintains minimal accuracy degradation even under extreme compression scenarios, including 1-bit gradients or complete omission of the low-precision gradient communication step (0-bit transmission).



Paperid:297
Authors:Yaoyu Zhu, Di Huang, Hanqi Lyu, Xiaoyun Zhang, Chongxiao Li, Wenxuan Shi, Yutong Wu, Jianan Mu, Jinghua Wang, Yang zhao, Pengwei Jin, Shuyao Cheng, shengwen Liang, xishan zhang, Rui Zhang, Zidong Du, Qi Guo, Xing Hu, Yunji Chen
Abstract:
Abstract:Large language models (LLMs) trained via reinforcement learning with verifiable reward (RLVR) have achieved breakthrough performance on tasks with explicit, automatable verification, such as software programming and mathematical problems. Extending RLVR to electronic design automation (EDA), especially automatically generating register-transfer level (RTL) code from natural-language (NL) specifications, however, poses three key challenges: the lack of automated and accurate verification environments, the scarcity of high‐quality NL–code pairs, and the prohibitive computation cost of RLVR. To this end, we introduce ReasoningRTL (R$^2$TL), an RLVR framework for training RTL generation LLMs. First, we develop a rule-based testbench generator that performs robust equivalence checking against golden references. Second, we propose a round-trip data synthesis method that pairs open-source RTL snippets with LLM‐generated NL descriptions, verifies code–NL–code consistency via the generated testbench, and filters out inequivalent examples to yield a high-quality dataset. Third, we employ a two-stage ``distill-then-RL'' training pipeline: distillation for the cold start of reasoning abilities, followed by adaptive DAPO, our novel RLVR algorithm that can reduce training cost by adaptively adjusting sampling rate. The resulting model, ReasoningRTL-7B (\modelFinal), achieves 68.6 \% and 72.9 \% pass@1 on VerilogEval v2 and RTLLM v2, respectively, surpassing prior state-of-the-art by 12$\sim$21 \%, while matching or even exceeding the performance of 671B DeepSeek-R1. We will release our model, training pipeline, and dataset to facilitate research in EDA and LLM communities.



Paperid:298
Authors:Yizhou Liu, Ziming Liu, Jeff Gore
Abstract:
The success of today's large language models (LLMs) depends on the observation that larger models perform better. However, the origin of this neural scaling law --- the finding that loss decreases as a power law with model size --- remains unclear. Starting from two empirical principles --- that LLMs represent more things than the model dimensions (widths) they have (i.e., representations are superposed), and that words or concepts in language occur with varying frequencies --- we constructed a toy model to study the loss scaling with model size. We found that when superposition is weak, meaning only the most frequent features are represented without interference, the scaling of loss with model size depends on the underlying feature frequency; if feature frequencies follow a power law, so does the loss. In contrast, under strong superposition, where all features are represented but overlap with each other, the loss becomes inversely proportional to the model dimension across a wide range of feature frequency distributions. This robust scaling behavior is explained geometrically: when many more vectors are packed into a lower dimensional space, the interference (squared overlaps) between vectors scales inversely with that dimension. We then analyzed four families of open-sourced LLMs and found that they exhibit strong superposition and quantitatively match the predictions of our toy model. The Chinchilla scaling law turned out to also agree with our results. We conclude that representation superposition is an important mechanism underlying the observed neural scaling laws. We anticipate that these insights will inspire new training strategies and model architectures to achieve better performance with less computation and fewer parameters.



Paperid:299
Authors:Haritz Puerto, Martin Gubri, Tommaso Green, Seong Joon Oh, Sangdoo Yun
Abstract:
Large Language Models (LLMs) are transforming search engines into Conversational Search Engines (CSE). Consequently, Search Engine Optimization (SEO) is being shifted into Conversational Search Engine Optimization (C-SEO). We are beginning to see dedicated C-SEO methods for modifying web documents to increase their visibility in CSE responses. However, they are often tested only for a limited breadth of application domains; we do not understand whether certain C-SEO methods would be effective for a broad range of domains. Moreover, existing evaluations consider only a single-actor scenario where only one web document adopts a C-SEO method; in reality, multiple players are likely to competitively adopt the cutting-edge C-SEO techniques, drawing an analogy from the dynamics we have seen in SEO. We present C-SEO Bench, the first benchmark designed to evaluate C-SEO methods across multiple tasks, domains, and number of actors. We consider two search tasks, question answering and product recommendation, with three domains each. We also formalize a new evaluation protocol with varying adoption rates among involved actors. Our experiments reveal that most current C-SEO methods are largely ineffective, contrary to reported results in the literature. Instead, traditional SEO strategies, those aiming to improve the ranking of the source in the LLM context, are significantly more effective. We also observe that as we increase the number of C-SEO adopters, the overall gains decrease, depicting a congested and zero-sum nature of the problem. Our code and data are available at https://anonymous.4open.science/r/cseo-bench-C247/README.md and https://huggingface.co/datasets/cseo/cseo-bench.



Paperid:300
Authors:Ranajoy Sadhukhan, Zhuoming Chen, Haizhong Zheng, Beidi Chen
Abstract:
We rethink test-time scaling laws from a practical efficiency perspective, revealing that the effectiveness of smaller models is significantly overestimated. Prior work, grounded in compute-optimality, overlooks critical memory access bottlenecks introduced by inference-time strategies (e.g., Best-of-N, long CoTs). Our holistic analysis, spanning models from 0.6B to 32B parameters, reveals a new Kinetics Scaling Law that better guides resource allocation by incorporating both computation and memory access costs. The Kinetics Scaling Law suggests that test-time compute is more effective when used on models above a threshold (14B) than on smaller ones. A key reason is that in test-time scaling, attention—rather than parameter count—emerges as the dominant cost factor. Motivated by this, we propose a new scaling paradigm centered on sparse attention, which lowers per-token cost and enables longer generations and more parallel samples within the same resource budget. Empirically, we show that sparse attention models consistently outperform dense counterparts, achieving over 60-point gains in low-cost regimes and over 5-point gains in high-cost regimes for problem-solving accuracy on AIME and LiveCodeBench. These results suggest that sparse attention is essential for realizing the full potential of test-time scaling because, unlike training where parameter scaling saturates, test-time accuracy continues to improve through increased generation.



Paperid:301
Authors:Vahid Jalili
Abstract:
Since its 2009 genesis block, the Bitcoin network has processed \num{>1.08} billion (B) transactions representing \num{>8.72}B BTC, offering rich potential for machine learning (ML); yet, its pseudonymity and obscured flow of funds inherent in its \utxo-based design, have rendered this data largely inaccessible for ML research. Addressing this gap, we present an ML-compatible graph modeling the Bitcoin's economic topology by reconstructing the provenance of funds. This temporal, heterogeneous graph encompasses complete transaction history up to block \num{863000}, consisting of \num{>2.4}B nodes and \num{>39.72}B edges. Additionally, we provide custom sampling methods yielding node and edge feature vectors of sampled communities, tools to load and analyze the Bitcoin graph data within specialized graph databases, and ready-to-use database snapshots. This comprehensive dataset and toolkit empower the ML community to tackle Bitcoin's intricate ecosystem at scale, driving progress in applications such as anomaly detection, address classification, market analysis, and large-scale graph ML benchmarking.Data and code available at \url{https://bit.ly/bc2g-neurips2025}{}



Paperid:302
Authors:Oumar Kaba, Kusha Sareen, Daniel Levy, Siamak Ravanbakhsh
Abstract:
Effectively leveraging prior knowledge of a system’s physics is crucial for applications of machine learning to scientific domains. Previous approaches mostly focused on incorporating physical insights at the architectural level. In this paper, we propose a framework to leverage physical information directly into the loss function for prediction and generative modeling tasks on systems like molecules and spins. We derive energy loss functions assuming that each data sample is in thermal equilibrium with respect to an approximate energy landscape. By using the reverse KL divergence with a Boltzmann distribution around the data, we obtain the loss as an energy difference between the data and the model predictions. This perspective also recasts traditional objectives like MSE as energy-based, but with a physically meaningless energy. In contrast, our formulation yields physically grounded loss functions with gradients that better align with valid configurations, while being architecture-agnostic and computationally efficient. The energy loss functions also inherently respect physical symmetries. We demonstrate our approach on molecular generation and spin ground-state prediction and report significant improvements over baselines.



Paperid:303
Authors:Yuqi Feng, Jiahao Fan, Yanan Sun
Abstract:
Fast adversarial training (FAT) has been considered as one of the most effective alternatives to the computationally-intensive adversarial training. Generally, FAT methods pay equal attention to each sample of the target task. However, the distance between each sample and the decision boundary is different, learning samples which are far from the decision boundary (i.e., less important to adversarial robustness) brings additional training cost and leads to sub-optimal results. To tackle this issue, we present vulnerable data-aware adversarial training (VDAT) in this study. Specifically, we first propose a margin-based vulnerability calculation method to measure the vulnerability of data samples. Moreover, we propose a vulnerability-aware data filtering method to reduce the training data for adversarial training thus improve the training efficiency, and we theoretically show the effectiveness of VDAT based on the upper bound of adversarial loss. The experiments are conducted in terms of adversarial training and robust neural architecture search on CIFAR-10, CIFAR-100, and ImageNet-1K. The results demonstrate that VDAT is up to 76\% more efficient than state-of-the-art FAT methods, while achieving improvements regarding the natural accuracy and adversarial accuracy in both scenarios. Furthermore, the visualizations and ablation studies show the effectiveness of both core components designed in VDAT.



Paperid:304
Authors:Hangting Ye, Jinmeng Li, He Zhao, Dandan Guo, Yi Chang
Abstract:
Tabular data have been playing a vital role in diverse real-world fields, including healthcare, finance, etc. With the recent success of Large Language Models (LLMs), early explorations of extending LLMs to the domain of tabular data have been developed. Most of these LLM-based methods typically first serialize tabular data into natural language descriptions, and then tune LLMs or directly infer on these serialized data. However, these methods suffer from two key inherent issues: (i) data perspective: existing data serialization methods lack universal applicability for structured tabular data, and may pose privacy risks through direct textual exposure, and (ii) model perspective: LLM fine-tuning methods struggle with tabular data, and in-context learning scalability is bottle-necked by input length constraints (suitable for few-shot learning). This work explores a novel direction of integrating LLMs into tabular data throughough logical decision tree rules as intermediaries, proposes a decision tree enhancer with LLM-derived rule for tabular prediction, DeLTa. The proposed DeLTa avoids tabular data serialization, and can be applied to full data learning setting without LLM fine-tuning. Specifically, we leverage the reasoning ability of LLMs to redesign an improved rule given a set of decision tree rules. Furthermore, we provide a calibration method for original decision trees via new generated rule by LLM, which approximates the error correction vector to steer the original decision tree predictions in the direction of ``errors'' reducing.Finally, extensive experiments on diverse tabular benchmarks show that our method achieves state-of-the-art performance.



Paperid:305
Authors:Licheng Liu, Mingda Qiao
Abstract:
Selective prediction [Dru13, QV19] models the scenario where a forecaster freely decides on the prediction window that their forecast spans. Many data statistics can be predicted to a non-trivial error ratewithout anydistributional assumptions or expert advice, yet these results rely on that the forecaster may predict at any time. We introduce a model of Prediction with Limited Selectivity (PLS) where the forecaster can start the prediction only on a subset of the time horizon. We study the optimal prediction error both on an instance-by-instance basis and via an average-case analysis. We introduce a complexity measure that gives instance-dependent bounds on the optimal error. For a randomly-generated PLS instance, these bounds match with high probability.



Paperid:306
Authors:Gautam Chandrasekaran, Adam Klivans
Abstract:
Abstract:We give an algorithm for learning $O(\log n)$ juntas in polynomial-time with respect to Markov Random Fields (MRFs) in a smoothed analysis framework, where only the external field has been randomly perturbed. This is a broad generalization of the work of Kalai and Teng, who gave an algorithm that succeeded with respect to smoothed *product* distributions (i.e., MRFs whose dependency graph has no edges). Our algorithm has two phases: (1) an unsupervised structure learning phase and (2) a greedy supervised learning algorithm. This is the first example where algorithms for learning the structure of undirected graphical models have downstream applications to supervised learning.



Paperid:307
Authors:Xi Zhang, Hanwei Zhu, Yan Zhong, Jiamang Wang, Weisi Lin
Abstract:
In this work, we propose a novel framework to enable diffusion models to adapt their generation quality based on real-time network bandwidth constraints. Traditional diffusion models produce high-fidelity images by performing a fixed number of denoising steps, regardless of downstream transmission limitations. However, in practical cloud-to-device scenarios, limited bandwidth often necessitates heavy compression, leading to loss of fine textures and wasted computation. To address this, we introduce a joint end-to-end training strategy where the diffusion model is conditioned on a target quality level derived from the available bandwidth. During training, the model learns to adaptively modulate the denoising process, enabling early-stop sampling that maintains perceptual quality appropriate to the target transmission condition. Our method requires minimal architectural changes and leverages a lightweight quality embedding to guide the denoising trajectory. Experimental results demonstrate that our approach significantly improves the visual fidelity of bandwidth-adapted generations compared to naive early-stopping, offering a promising solution for efficient image delivery in bandwidth-constrained environments.



Paperid:308
Authors:Moshe Shenfeld, Vitaly Feldman
Abstract:
Abstract:We consider the privacy amplification properties of a sampling scheme in which a user's data is used in $k$ steps chosen randomly and uniformly from a sequence (or set) of $t$ steps. This sampling scheme has been recently applied in the context of differentially private optimization [Chua et al., 2024a, Choquette-Choo et al., 2024] and is also motivated by communication-efficient high-dimensional private aggregation [Asi et al., 2025]. Existing analyses of this scheme either rely on privacy amplification by shuffling which leads to overly conservative bounds or require Monte Carlo simulations that are computationally prohibitive in most practical scenarios.We give the first theoretical guarantees and numerical estimation algorithms for this sampling scheme. In particular, we demonstrate that the privacy guarantees of random $k$-out-of-$t$ allocation can be upper bounded by the privacy guarantees of the well-studied independent (or Poisson) subsampling in which each step uses the user's data with probability $(1+o(1))k/t$. Further, we provide two additional analysis techniques that lead to numerical improvements in several parameter regimes. Altogether, our bounds give efficiently-computable and nearly tight numerical results for random allocation applied to Gaussian noise addition.



Paperid:309
Authors:Andrej Leban
Abstract:
The Distributional Principal Autoencoder (DPA) combines distributionally correct reconstruction with principal-component-like interpretability of the encodings. In this work, we provide exact theoretical guarantees on both fronts. First, we derive a closed-form relation linking each optimal level-set geometry to the data-distribution score. This result explains DPA's empirical ability to disentangle factors of variation of the data as well as allows the score to be recovered directly from samples. When the data follows the Boltzmann distribution, we demonstrate that this relation yields an approximation of the minimum free-energy path for the Müller–Brown potential in a single fit. Second, we prove that if the data lies on a manifold that can be approximated by the encoder, latent components beyond the manifold dimension are conditionally independent of the data distribution - carrying no additional information - and thus reveal the intrinsic dimension. Together, these results show that a single model can learn the data distribution and its intrinsic dimension with exact guarantees, unifying two longstanding goals of unsupervised learning.



Paperid:310
Authors:Xiaohua Wang, Kaitao Song, Xu Tan, Huiqiang Jiang, Chengruidong Zhang, Yongliang Shen, Cen Lu, Zihao Li, Zifan Song, Caihua Shan, Yansen Wang, Kan Ren, Xiaoqing Zheng, Tao Qin, Yuqing Yang, Dongsheng Li, Lili Qiu
Abstract:
In this paper, we propose a novel learning paradigm, termedChain-of-Model(CoM), which incorporates the causal relationship into the hidden states of each layer as a chain style. thereby introducing great scaling efficiency in model training and inference flexibility in deployment.We introduce the concept ofChain-of-Representation(CoR), which formulates the hidden states at each layer as a combination of multiple sub-representations (i.e., chains). In each layer, each chain from the output representations can only view all of its preceding chains in the input representations. Consequently, the model built upon CoM framework can progressively scale up the model size by increasing the chains based on the previous models (i.e., chains), and offer multiple sub-models at varying sizes for elastic inference by using different chain numbers. Based on this principle, we deviseChain-of-Language-Model(CoLM), which incorporates the idea of CoM into each layer of Transformer architecture. Based on CoLM, we further introduce CoLM-Air by introducing aKV sharingmechanism, that computes all keys and values within the first chain and then shares across all chains. This design demonstrates additional extensibility, such as enabling seamless LM switching, prefilling acceleration and so on. Experimental results demonstrate our CoLM family can achieve comparable performance to the standard Transformer, while simultaneously enabling greater flexiblity, such as progressive scaling to improve training efficiency and offer multiple varying model sizes for elastic inference, paving a a new way toward building language models.



Paperid:311
Authors:Matteo Ninniri, Marco Podda, Davide Bacciu
Abstract:
Generative models of graphs based on discrete Denoising Diffusion Probabilistic Models (DDPMs) offer a principled approach to molecular generation by systematically removing structural noise through iterative atom and bond adjustments. However, existing formulations are fundamentally limited by their inability to adapt the graph size (that is, the number of atoms) during the diffusion process, severely restricting their effectiveness in conditional generation scenarios such as property-driven molecular design, where the targeted property often correlates with the molecular size. In this paper, we reformulate the noising and denoising processes to support monotonic insertion and deletion of nodes. The resulting model, which we call GrIDDD, dynamically grows or shrinks the chemical graph during generation. GrIDDD matches or exceeds the performance of existing graph diffusion models on molecular property targeting despite being trained on a more difficult problem. Furthermore, when applied to molecular optimization, GrIDDD exhibits competitive performance compared to specialized optimization models. This work paves the way for size-adaptive molecular generation with graph diffusion.



Paperid:312
Authors:Wei Liu, Jiangtao Feng, Hongli Yu, Yuxuan Song, Yuqiang Li, Shufei Zhang, LEI BAI, Wei-Ying Ma, Hao Zhou
Abstract:
Multi-step retrosynthesis planning is paramount in chemical discovery, and is extremely accelerated by advanced artificial intelligence~(AI) methods. Recent research on Large Language Models (LLMs) has demonstrated potent reasoning and planning capabilities with reinforcement learning techniques, especially math and code problem solving. A natural question is whether an LLM can learn a good policy on multi-step retrosynthesis planning. In this study, we introduce Retro-R1, an innovative LLM-based retrosynthesis agent learned with reinforcement learning, to design molecular synthesis pathways. Unlike previous reasoning tasks, e.g., math and code, trained purely in a single-turn question-answering style, Retro-R1 is designed in an agent-action way and is trained with single-step retrosynthesis tools, which are plug-in. The experimental results indicate that Retro-R1 does boost the multi-step retrosynthesis policy with 55.79\% pass@1 success rate, surpassing the previous state-of-the-art method by 8.95\%. Retro-R1 exhibits impressive out-of-domain generalization performance, while previous methods lag behind despite their high performance in the in-domain testset. We hope this work would signify a crucial step toward imbuing LLMs with advanced, chemist-like reasoning, underscoring RL's potential to unlock data-efficient, generalizable, and sophisticated problem-solving in LLM agents for scientific discovery.



Paperid:313
Authors:Tomás González, Mateo Dulce Rubio, Aaditya Ramdas, Mónica Ribero
Abstract:
We propose a practical sequential test for auditing differential privacy guarantees of black-box mechanisms. The test processes streams of mechanisms' outputs providing anytime-valid inference while controlling Type I error, overcoming the fixed sample size limitation of previous batch auditing methods. Experiments show this test detects violations with sample sizes that are orders of magnitude smaller than existing methods, reducing this number from 50K to a few hundred examples, across diverse realistic mechanisms. Notably, it identifies DP-SGD privacy violations in under one training run, unlike prior methods needing full model training.



Paperid:314
Authors:Andy Keller
Abstract:
Data arrives at our senses as a continuous stream, smoothly transforming from one instant to the next. These smooth transformations can be viewed as continuous symmetries of the environment that we inhabit, defining equivalence relations between stimuli over time. In machine learning, neural network architectures that respect symmetries of their data are called equivariant and have provable benefits in terms of generalization ability and sample efficiency. To date, however, equivariance has been considered only for static transformations and feed-forward networks, limiting its applicability to sequence models, such as recurrent neural networks (RNNs), and corresponding time-parameterized sequence transformations. In this work, we extend equivariant network theory to this regime of `flows' -- one-parameter Lie subgroups capturing natural transformations over time, such as visual motion. We begin by showing that standard RNNs are generally not flow-equivariant: their hidden states fail to transform in a geometrically structured manner for moving stimuli. We then show how flow equivariance can be introduced, and demonstrate that these models significantly outperform their non-equivariant counterparts in terms of training speed, length generalization, and velocity generalization, on both next step prediction and sequence classification. We present this work as a first step towards building sequence models that respect the time-parameterized symmetries which govern the world around us.



Paperid:315
Authors:Jinwoo Kim, Max Beier, Petar Bevanda, Nayun Kim, Seunghoon Hong
Abstract:
A fundamental question in sequence modeling with neural networks is how to represent and learn highly nonlinear and probabilistic state dynamics. Operator theory views such dynamics as linear maps on Hilbert spaces containing mean embedding vectors of distributions, offering a useful yet currently overlooked perspective. We propose a new approach to sequence modeling based on an operator-theoretic view of a hidden Markov model (HMM). Instead of materializing stochastic recurrence, we embed the whole sequence distribution as a tensor in the product Hilbert space. A generative process is then defined as a maximum mean discrepancy (MMD) gradient flow in the space of sequences. To overcome challenges with large tensors and slow sampling convergence, we introduce spectral mean flows, a novel tractable algorithm integrating two core concepts. First, we propose a new neural architecture that leverages the spectral decomposition of linear operators to derive a scalable tensor network decomposition of the sequence mean embedding. Second, we extend MMD gradient flows to time-dependent Hilbert spaces, which might be of independent interest. The resulting technique is connected to flow matching via the continuity equation, enabling end-to-end, simulation-free learning via double backpropagation. We demonstrate state-of-the-art results on time-series modeling.



Paperid:316
Authors:Steve Hanneke, Amin Karbasi, Anay Mehrotra, Grigoris Velegkas
Abstract:
We investigate language generation in the limit – a model by Kleinberg and Mullainathan and extended by Li, Raman, and Tewari. While Kleinberg and Mullainathan proved generation is possible for all countable collections, Li, Raman, and Tewari defined a hierarchy of generation notions (uniform, non-uniform, and generatable) and explored their feasibility for uncountable collections. Our first set of results resolve two open questions of Li et al. by proving finite unions of generatable or non-uniformly generatable classes need not be generatable. These follow from a stronger result: there is non-uniformly generatable class and a uniformly generatable class whose union is non-generatable. This adds to the aspects along which language generation in the limit is different from traditional tasks in statistical learning theory like classification, which are closed under finite unions. In particular, it implies that given two generators for different collections, one cannot combine them to obtain a single "more powerful" generator, prohibiting this notion of boosting. Our construction also addresses a third of Li et al.'s open questions on whether there are uncountable classes that are non-uniformly generatable and do not satisfy the eventually unbounded closure (EUC) condition introduced by Li et al. Our approach utilizes carefully constructed classes along with a novel diagonalization argument that could be of independent interest in the growing area of language generation.



Paperid:317
Authors:Sebastian Gutierrez Hernandez, Peng Chen, Hao-Min Zhou
Abstract:
We introduce Parametric Density Path Optimization (PDPO), a novel method for computing action-minimizing paths between probability densities. The core idea is to represent the target probability path as the pushforward of a reference density through a parametric map, transforming the original infinite-dimensional optimization over densities to a finite-dimensional one over the parameters of the map. We derive a static formulation of the dynamic problem of action minimization and propose cubic spline interpolation of the path in parameter space to solve the static problem. Theoretically, we establish an error bound of the action under proper assumptions on the regularity of the parameter path. Empirically, we find that using 3–5 control points of the spline interpolation suffices to accurately resolve both multimodal and high-dimensional problems. We demonstrate thatPDPO can flexibly accommodate a wide range of potential terms, including those modeling obstacles, mean-field interactions, stochastic control, and higher-order dynamics. Our method outperforms existing state-of-the-art approaches in benchmark tasks, demonstrating superior computational efficiency and solution quality.



Paperid:318
Authors:Tianrun Zhao, Xiaojie Mao, Yong Liang
Abstract:
Abstract:In this paper, we study an online strategic classification problem, where a principal aims to learn an accurate binary linear classifier from sequentially arriving agents. For each agent, the principal announces a classifier. The agent can strategically exercise costly manipulations on his features to be classified as the favorable positive class. The principal is unaware of the true feature-label distribution, but observes all reported features and only labels of positively classified agents. We assume that the true feature-label distribution is given by a halfspace model subject to arbitrary feature-dependent bounded noise (i.e., Massart Noise). This problem faces the combined challenges of agents' strategic feature manipulations, partial label observations, and label noises. We tackle these challenges by a novel learning algorithm. We show that the proposed algorithm yields classifiers that converge to the clairvoyant optimal one and attains a regret rate of $ O(\sqrt{T})$ up to poly-logarithmic and constant factors over $T$ cycles.



Paperid:319
Authors:Xiang Li, Yunai Li, Huiying Zhong, Lihua Lei, Zhun Deng
Abstract:
Performativity of predictions refers to the phenomena that prediction-informed decisions may influence the target they aim to predict, which is widely observed in policy-making in social sciences and economics. In this paper, weinitiatethe study of statistical inference under performativity. Our contribution is two-fold. First, we build a central limit theorem for estimation and inference under performativity, which enables inferential purposes in policy-making such as constructing confidence intervals or testing hypotheses. Second, we further leverage the derived central limit theorem to investigate prediction-powered inference (PPI) under performativity, which is based on a small labeled dataset and a much larger dataset of machine-learning predictions. This enables us to obtain more precise estimation and improved confidence regions for the model parameter (i.e., policy) of interest in performative prediction. We demonstrate the power of our framework by numerical experiments. To the best of our knowledge, this paper is the first one to establish statistical inference under performativity, which brings up new challenges and inference settings that we believe will add significant values to policy-making, statistics, and machine learning.



Paperid:320
Authors:Jin Yao, Radowan Mahmud Redoy, Sebastian Elbaum, Matthew Dwyer, Zezhou Cheng
Abstract:
Detecting objects in 3D space from monocular input is crucial for applications ranging from robotics to scene understanding. Despite advanced performance in the indoor and autonomous driving domains, existing monocular 3D detection models struggle with in-the-wild images due to the lack of 3D in-the-wild datasets and the challenges of 3D annotation. We introduce LabelAny3D, an analysis-by-synthesis framework that reconstructs holistic 3D scenes from 2D images to efficiently produce high-quality 3D bounding box annotations.Built on this pipeline, we present COCO3D, a new benchmark for open-vocabulary monocular 3D detection, derived from the MS-COCO dataset and covering a wide range of object categories absent from existing 3D datasets. Experiments show that annotations generated by LabelAny3D improve monocular 3D detection performance across multiple benchmarks, outperforming prior auto-labeling approaches in quality. These results demonstrate the promise of foundation-model-driven annotation for scaling up 3D recognition in realistic, open-world settings.



Paperid:321
Authors:Federico Malato, Ville Hautamäki
Abstract:
World Models have vastly permeated the field of Reinforcement Learning. Their ability to model the transition dynamics of an environment have led to tremendous improvements in sample efficiency for online RL. Among them, the most notorious example is Dreamer, a model that learns to act in a diverse set of image-based environments. In this paper, we leverage similarity search and stochastic representations to approximate a world model without a training procedure. We establish a comparison with PlaNet, a well-established world model of the Dreamer family. We evaluate the models on the quality of latent reconstruction and on the perceived similarity of the reconstructed image, on both next-step and long horizon dynamics prediction. The results of our study demonstrate that a search-based world model is comparable to a training based one in both cases. Notably, our model shows stronger performance in long-horizon prediction with respect to the baseline on a range of visually different environments.



Paperid:322
Authors:Raj Ghugare, Benjamin Eysenbach
Abstract:
Modern reinforcement learning (RL) algorithms have found success by using powerful probabilistic models, such as transformers, energy-based models, and diffusion/flow-based models. To this end, RL researchers often choose to pay the price of accommodating these models into their algorithms -- diffusion models are expressive, but are computationally intensive due to their reliance on solving differential equations, while autoregressive transformer models are scalable but typically require learning discrete representations. Normalizing flows (NFs), by contrast, seem to provide an appealing alternative, as they enable likelihoods and sampling without solving differential equations or autoregressive architectures. However, their potential in RL has received limited attention, partly due to the prevailing belief that normalizing flows lack sufficient expressivity. We show that this is not the case. Building on recent work in NFs, we propose a single NF architecture which integrates seamlessly into RL algorithms, serving as a policy, Q-function, and occupancy measure. Our approach leads to much simpler algorithms, and achieves higher performance in imitation learning, offline, goal conditioned RL and unsupervised RL.



Paperid:323
Authors:Da Yu, Edith Cohen, Badih Ghazi, Yangsibo Huang, Pritish Kamath, Ravi Kumar, Daogao Liu, Chiyuan Zhang
Abstract:
Abstract:We propose SCONE (**S**calable, **C**ontextualized, **O**ffloaded, **N**-gram **E**mbedding), a new method for extending input embedding layers to enhance language model performance. To avoid increased decoding costs, SCONE retains the original vocabulary while introducing embeddings for a set of frequent $n$-grams. These embeddings provide contextualized representation for each input token and are learned with a separate model during training. After training, embeddings are precomputed and stored in off-accelerator memory; during inference, querying them has minimal impact on latency due to the low complexity of embedding lookups. SCONE enables two new scaling strategies: increasing the number of $n$-gram embeddings and scaling the model used to learn them, both while maintaining fixed accelerator usage during inference (in terms of FLOPS and memory). We show that scaling both aspects enables a model with 1B accelerator-resident parameters to outperform a 1.9B-parameter baseline across diverse corpora, while using only about half the FLOPS and accelerator memory during inference.



Paperid:324
Authors:Taeyoun Kim, Fahim Tajwar, Aditi Raghunathan, Aviral Kumar
Abstract:
Abstract:Reasoning methods that adaptively allocate test-time compute have advanced LLM performance in math and code. We study how we can utilize this framework to train models that are robust to safety vulnerabilities. We build a recipe called $\textit{\textbf{TARS}}$ (Training Adaptive Reasoners for Safety), a reinforcement learning (RL) approach that trains models to reason about safety using Chain-of-Thought traces and a reward signal that balances safety with task completion. When building TARS, we identify three critical design choices: (1) a premature SFT training stage, (2) a mix of harmful, harmless, and ambiguous prompts to prevent shortcut behaviors such as over-refusal, and (3) a reward function to prevent an absence of reasoning. Models trained with TARS exhibit adaptive behaviors and spend more compute on ambiguous queries, achieve better safety-refusal trade-offs, internally learn to better distinguish between safe and unsafe prompts, and attain greater robustness to both white-box (e.g., GCG) and black-box attacks (e.g., PAIR). We find that TARS also preserves general reasoning capabilities (e.g., on math benchmarks) despite minimal exposure on such data. Overall, our work provides a principled and open recipe for LLMs for safety through adaptive reasoning.



Paperid:325
Authors:Florian Redhardt, Yassir Akram, Simon Schug
Abstract:
Can neural networks systematically capture discrete, compositional task structure despite their continuous, distributed nature?The impressive capabilities of large-scale neural networks suggest that the answer to this question is yes.However, even for the most capable models, there are still frequent failure cases that raise doubts about their systematicity.Here, we seek to understand what it takes for a standard neural network to generalize over tasks that share compositional structure.We find that simply scaling data and model size leads to compositional generalization.We show that this holds across different task encodings as long as the training distribution sufficiently covers the task space.In line with this finding, we prove that standard multilayer perceptrons can approximate a general class of compositional task families to arbitrary precision using only a linear number of neurons with respect to the number of task modules.Finally, we uncover that if networks successfully compositionally generalize, the constituents of a task can be linearly decoded from their hidden activations.We show that this simple metric correlates with instances where text-to-image generation models fail to compose known concepts.



Paperid:326
Authors:Bowen Deng, Lele Fu, Sheng Huang, Tianchi Liao, Jialong Chen, Zhang Tao, Chuan Chen
Abstract:
Graph classification has long assumed a closed-world setting, limiting its applicability to real-world scenarios where new categories often emerge. To address this limitation, we introduce Graph-Level Novel Category Discovery (GLNCD), a new task aimed at identifying unseen graph categories without supervision from novel classes. We first adapt classical Novel Category Discovery (NCD) methods for images to the graph domain and evaluate these baseline methods on four diverse graph datasets curated for the GLNCD task. Our analysis reveals that these methods suffer a notable performance degradation compared to their image-based counterparts, due to two key challenges: (1) insufficient utilization of structural information in graph self-supervised learning (SSL), and (2) ineffective pseudo-labeling strategies based on ranking statistics (RS) that neglect graph structure. To alleviate these issues, we propose ProtoFGW-NCD, a framework consisting of two core components: ProtoFGW-CL, a novel graph SSL framework, and FGW-RS, a structure-aware pseudo-labeling method. Both components employ a differentiable Fused Gromov-Wasserstein (FGW) distance to effectively compare graphs by incorporating structural information. These components are built upon learnable prototype graphs, which enable efficient, parallel FGW-based graph comparisons and capture representative patterns within graph datasets. Experiments on four GLNCD benchmark datasets demonstrate the effectiveness of ProtoFGW-NCD.



Paperid:327
Authors:Roey Magen, Shuning Shang, Zhiwei Xu, Spencer Frei, Wei Hu, Gal Vardi
Abstract:
The phenomenon of benign overfitting, where a trained neural network perfectly fits noisy training data but still achieves near-optimal test performance, has been extensively studied in recent years for linear models and fully-connected/convolutional networks. In this work, we study benign overfitting in a single-head softmax attention model, which is the fundamental building block of Transformers. We prove that under appropriate conditions, the model exhibits benign overfitting in a classification setting already after two steps of gradient descent. Moreover, we show conditions where a minimum-norm/maximum-margin interpolator exhibits benign overfitting. We study how the overfitting behavior depends on the signal-to-noise ratio (SNR) of the data distribution, namely, the ratio between norms of signal and noise tokens, and prove that a sufficiently large SNR is both necessary and sufficient for benign overfitting.



Paperid:328
Authors:Giovanni De Felice, Arianna Casanova, Francesco De Santis, Silvia Santini, Johannes Schneider, Pietro Barbiero, Alberto Termine
Abstract:
Abstract:Concept-based models are an emerging paradigm in deep learning that constrains the inference process to operate through human-interpretable variables, facilitating explainability and human interaction. However, these architectures, on par with popular opaque neural models, fail to account for the true causal mechanisms underlying the target phenomena represented in the data. This hampers their ability to support causal reasoning tasks, limits out-of-distribution generalization, and hinders the implementation of fairness constraints. To overcome these issues, we propose Causally reliable Concept Bottleneck Models (C$^2$BMs), a class of concept-based architectures that enforce reasoning through a bottleneck of concepts structured according to a model of the real-world causal mechanisms. We also introduce a pipeline to automatically learn this structure from observational data and unstructured background knowledge (e.g., scientific literature). Experimental evidence suggests that C$^2$BMs are more interpretable, causally reliable, and improve responsiveness to interventions w.r.t. standard opaque and concept-based models, while maintaining their accuracy.



Paperid:329
Authors:Zhanpeng Luo, Haoxi Ran, Li Lu
Abstract:
Dynamic view synthesis has seen significant advances, yet reconstructing scenes from uncalibrated, casual video remains challenging due to slow optimization and complex parameter estimation. In this work, we presentInstant4D, a monocular reconstruction system that leverages native 4D representation to efficiently process casual video sequences within minutes, without calibrated cameras or depth sensors.Our method begins with geometric recovery through deep visual SLAM, followed by grid pruning to optimize scene representation. Our design significantly reduces redundancy while maintaining geometric integrity, cutting model size to under10%of its original footprint. To handle temporal dynamics efficiently, we introduce a streamlined 4D Gaussian representation, achieving a30×speed-up and reducing training time to within two minutes, while maintaining competitive performance across several benchmarks. We further apply our model to in-the-wild videos, showcasing its generalizability. Our project website will be published at https://instant4d.github.io/Instant4D/



Paperid:330
Authors:Siavash Ameli, Chris van der Heide, Liam Hodgkinson, Michael Mahoney
Abstract:
Computing eigenvalues of very large matrices is a critical task in many machine learning applications, including the evaluation of log-determinants, the trace of matrix functions, and other important metrics. As datasets continue to grow in scale, the corresponding covariance and kernel matrices become increasingly large, often reaching magnitudes that make their direct formation impractical or impossible. Existing techniques typically rely on matrix-vector products, which can provide efficient approximations, if the matrix spectrum behaves well. However, in settings like distributed learning, or when the matrix is defined only indirectly, access to the full data set can be restricted to only very small sub-matrices of the original matrix. In these cases, the matrix of nominal interest is not even available as an implicit operator, meaning that even matrix-vector products may not be available. In such settings, the matrix is "impalpable", in the sense that we have access to only masked snapshots of it. We draw on principles from free probability theory to introduce a novel method of "free decompression" to estimate the spectrum of such matrices. Our method can be used to extrapolate from the empirical spectral densities of small submatrices to infer the eigenspectrum of extremely large (impalpable) matrices (that we cannot form or even evaluate with full matrix-vector products). We demonstrate the effectiveness of this approach through a series of examples, comparing its performance against known limiting distributions from random matrix theory in synthetic settings, as well as applying it to submatrices of real-world datasets, matching them with their full empirical eigenspectra.



Paperid:331
Authors:Yukang Chen, Wei Huang, Baifeng Shi, Qinghao Hu, Hanrong Ye, Ligeng Zhu, Zhijian Liu, Pavlo Molchanov, Jan Kautz, Xiaojuan Qi, Sifei Liu, Hongxu Yin, Yao Lu, Song Han
Abstract:
We introduce LongVILA-R1, a full-stack framework that scales up reasoning in vision-language models (VLMs) to long videos, leveraging reinforcement learning. LongVILA-R1 addresses the unique challenges of long video reasoning by integrating three critical components: (1) a large-scale dataset, LongVideo-Reason, comprising 52K long video QA pairs with labeled high-quality reasoning annotations across diverse domains such as sports, games, and vlogs; (2) a two-stage training pipeline that extends VLMs with chain-of-thought supervised fine-tuning (CoT-SFT) and reinforcement learning (RL); and (3) a training infrastructure for long video RL, named Multi-modal Reinforcement Sequence Parallelism (MR-SP), which incorporates sequence parallelism and a vLLM-based engine tailored for long video, using cached video embeddings for efficient rollout and prefilling. In experiments, LongVILA-R1-7B achieves strong performance on long video QA benchmarks such as VideoMME. It also outperforms Video-R1-7B by 5.2% and even matches Gemini-1.5-Pro across temporal reasoning, goal and purpose reasoning, spatial reasoning, and plot reasoning on our LongVideo-Reason-eval benchmark. Notably, our MR-SP system achieves up to 9.2x speedup on long video RL training. LongVILA-R1 demonstrates consistent performance gains as the number of input video frames scales from 16 to 512. LongVILA-R1 marks a firm step forward in VLMs for long video reasoning.



Paperid:332
Authors:Fangyuan Sun, Ilyas Fatkhullin, Niao He
Abstract:
Stochastic Natural Gradient Variational Inference (NGVI) is a widely used method for approximating posterior distribution in probabilistic models. Despite its empirical success and foundational role in variational inference, its theoretical underpinnings remain limited---particularly in the case of non-conjugate likelihoods. While NGVI has been shown to be a special instance of Stochastic Mirror Descent, and recent work has provided convergence guarantees using relative smoothness and strong convexity for conjugate models, these results do not extend to the non-conjugate setting, where the variational loss becomes non-convex and harder to analyze. In this work, we focus on mean-field parameterization and advance the theoretical understanding of NGVI in three key directions. First, we derive sufficient conditions under which the variational loss satisfies relative smoothness with respect to a suitable mirror map. Second, leveraging this structure, we propose a modified NGVI algorithm incorporating non-Euclidean projections and prove its global non-asymptotic convergence to a stationary point. Finally, under additional structural assumptions about the likelihood, we uncover hidden convexity properties of the variational loss and establish fast global convergence of NGVI to a global optimum. These results provide new insights into the geometry and convergence behavior of NGVI in challenging inference settings.



Paperid:333
Authors:Changyu Hu, Yanke Qu, Qiuan Yang, Xiaoyu Xiong, Kui Wu, Wei Li, Tao Du
Abstract:
Swimming in nature achieves remarkable performance through diverse morphological adaptations and intricate solid-fluid interaction, yet exploring this capability in artificial soft swimmers remains challenging due to high-dimensional control complexity and the computational cost of resolving hydrodynamic details. Traditional approaches rely on morphology-dependent heuristics—typically handcrafted by human experts through predefined muscle layouts, which constrain the control space to human-biased priors and limit exploration of novel swimming strategies. Additionally, robot learning frameworks often compromise hydrodynamic fidelity for computational efficiency, eliminating fluid dynamics essential for advanced strategies like vortex exploitation. Inspired by biological swimmers' ability to achieve complex motion through low-dimensional actuation, we propose an automated framework featuring a unified reduced mode control space that naturally captures typical deformation patterns observed across diverse swimmers. Our approach requires only sampling to generate controllers for diverse morphologies, minimizing human design effort while balancing kinematic expressiveness and dynamic plausibility. We also develop a GPU-accelerated simulator that efficiently resolves fluid-structure interactions, specifically tailored for learning swimming controllers. We evaluate our method in various morphologies, demonstrating 50\% higher success rates in swimming task than state-of-the-art baselines, which particularly struggle with unconventional forms (e.g. torus) where our method discovers novel, effective swimming strategies. We showcase how our automated approach discovers high-performance swimming patterns similar to those found in nature over years of evolution. Our work unlocks future opportunities in automated co-design of morphology and control of soft robots in complex hydrodynamic environments.



Paperid:334
Authors:Jung-hun Kim, Milan Vojnovic, Min-hwan Oh
Abstract:
Abstract:We study the combinatorial semi-bandit problem where an agent selects a subset of base arms and receives individual feedback. While this generalizes the classical multi-armed bandit and has broad applicability, its scalability is limited by the high cost of combinatorial optimization, requiring oracle queries at *every* round. To tackle this, we propose oracle-efficient frameworks that significantly reduce oracle calls while maintaining tight regret guarantees. For worst-case linear rewards, our algorithms achieve $\tilde{O}(\sqrt{T})$ regret using only $O(\log\log T)$ oracle queries. We also propose covariance-adaptive algorithms that leverage noise structure for improved regret, and extend our approach to general (non-linear) rewards. Overall, our methods reduce oracle usage from linear to (doubly) logarithmic in time, with strong theoretical guarantees.



Paperid:335
Authors:Lingzhi Shen, Xiaohao Cai, Yunfei Long, Imran Razzak, Guanming Chen, Shoaib Jameel
Abstract:
Cultural awareness in language models refers to the ability to understand norms, values, and perspectives embedded in diverse cultural contexts. However, existing approaches often treat culture as static background knowledge, failing to capture the evolving nature of cultural context, which limits their reliability in dynamic downstream tasks that require cultural sensitivity. In this work, we introduce CALM, a novel framework designed to endow language models with cultural self-awareness. CALM simultaneously extracts explicit cultural concepts and latent cultural signals beyond task semantics, and structures them through contrastive learning to induce culturally coherent representations. These features are then aligned via cross-attention and routed through a dimension-specific Mixture-of-Experts mechanism, resulting in a unified representation that is fused with the model’s internal knowledge to form a culturally grounded identity profile. To enable continual cultural adaptation, CALM incorporates self-prompted reflective learning, allowing the model to adaptively self-correct its cultural understanding across contexts. Experiments on the benchmark datasets demonstrate that CALM outperforms state-of-the-art methods.



Paperid:336
Authors:Daniel Pfrommer, Zehao Dou, Christopher Scarvelis, Max Simchowitz, Ali Jadbabaie
Abstract:
We study the inductive biases of diffusion models with a conditioning-variable, which have seen widespread application as both text-conditioned generative image models and observation-conditioned continuous control policies. We observe that when these models are queried conditionally, their generations consistently deviate from the idealized "denoising" process upon which diffusion models are formulated, inducing disagreement between popular sampling algorithms (e.g. DDPM, DDIM). We introduceSchedule Deviation, a rigorous measure which captures the rate of deviation from a standard denoising process, and provide a methodology to compute it. Crucially, we demonstrate that the deviation from an idealized denoising process occurs irrespective of the model capacity or amount of training data. We posit that this phenomena occurs due to the difficulty of bridging distinct denoising flows across different parts of the conditioning space and show theoretically how such a phenomena can arise through an inductive bias towards smoothness.



Paperid:337
Authors:Wenfang Sun, Xinyuan Song, Pengxiang Li, Lu Yin, Yefeng Zheng, Shiwei Liu
Abstract:
In this paper, we re-introduce the Curse of Depth, a concept that re-introduces, explains, and addresses the recent observation in modern Large Language Models (LLMs) where deeper layers are much less effective than expected. We first confirm the wide existence of this phenomenon across the most popular families of LLMs, such as Llama, Mistral, DeepSeek, and Qwen. Our analysis, theoretically and empirically, identifies that the underlying reason for the ineffectiveness of deep layers in LLMs is the widespread usage of Pre-Layer Normalization (Pre-LN). While Pre-LN stabilizes the training of Transformer LLMs, its output variance exponentially grows with the model depth, which undesirably causes the derivative of the deep Transformer blocks to be an identity matrix, and therefore barely contributes to the training. To resolve this training pitfall, we propose LayerNorm Scaling, which scales the variance of output of the layer normalization inversely by the square root of its depth. This simple modification mitigates the output variance explosion of deeper Transformer layers, improving their contribution. Our experimental results, spanning model sizes from 130M to 7B, demonstrate that \ours significantly enhances LLM pre-training performance compared to Pre-LN. Moreover, this improvement seamlessly carries over to supervised fine-tuning. All these gains can be attributed to the fact that LayerNorm Scaling enables deeper layers to contribute more effectively during training.



Paperid:338
Authors:Hongming Piao, Hao Wang, Dapeng Wu, Ying Wei
Abstract:
Abstract:Model editing has become a *de-facto* practice to address hallucinations and outdated knowledge of large language models (LLMs). However, existing methods are predominantly evaluated in isolation, i.e., one edit at a time, failing to consider a critical scenario of compositional model editing, where multiple edits must be integrated and jointly utilized to answer real-world multifaceted questions. For instance, in medical domains, if one edit informs LLMs that COVID-19 causes "fever" and another that it causes "loss of taste", a qualified compositional editor should enable LLMs to answer the question "What are the symptoms of COVID-19?" with both "fever" and "loss of taste" (and potentially more). In this work, we define and systematically benchmark this compositional model editing (CME) task, identifying three key undesirable issues that existing methods struggle with: *knowledge loss*, *incorrect preceding* and *knowledge sinking*. To overcome these issues, we propose A$^3$E, a novel compositional editor that (1) ***a**daptively combines and **a**daptively regularizes* pre-trained foundation knowledge in LLMs in the stage of edit training and (2) ***a**daptively merges* multiple edits to better meet compositional needs in the stage of edit composing. Extensive experiments demonstrate that A$^3$E improves the composability by at least 22.45\% without sacrificing the performance of non-compositional model editing.



Paperid:339
Authors:Tengjie Li, Shikui Tu, Lei Xu
Abstract:
Recent advances in vision-language models have facilitated progress in sketch generation. However, existing specialized methods primarily focus on generic synthesis and lack mechanisms for precise control over sketch styles. In this work, we propose a training-free framework based on diffusion models that enables explicit style guidance via textual prompts and referenced style sketches. Unlike previous style transfer methods that overwrite key and value matrices in self-attention, we incorporate the reference features as auxiliary information with linear smoothing and leverage a style-content guidance mechanism. This design effectively reduces content leakage from reference sketches and enhances synthesis quality, especially in cases with low structural similarity between reference and target sketches. Furthermore, we extend our framework to support controllable multi-style generation by integrating features from multiple reference sketches, coordinated via a joint AdaIN module. Extensive experiments demonstrate that our approach achieves high-quality sketch generation with accurate style alignment and improved flexibility in style control.



Paperid:340
Authors:Canbin Huang, Tianyuan Shi, Yuhua Zhu, Ruijun Chen, Xiaojun Quan
Abstract:
Large language model (LLM) routers improve the efficiency of multi-model systems by directing each query to the most appropriate model while leveraging the diverse strengths of heterogeneous LLMs. Most existing approaches frame routing as a classification problem based solely on the input query. While this reduces overhead by avoiding inference across all models, it overlooks valuable information that could be gleaned from potential outputs and fails to capture implicit intent or contextual nuances that often emerge only during response generation. These limitations can result in suboptimal routing decisions, particularly for complex or ambiguous queries that require deeper semantic understanding. To address this challenge, we propose Lookahead, a routing framework that "foresees" potential model outputs by predicting their latent representations and uses these predictions to guide model selection, thus enabling more informed routing without full inference. Within this framework, we implement two approaches based on causal and masked language models. Empirical evaluations across seven public benchmarks—spanning instruction following, mathematical reasoning, and code generation—show that Lookahead consistently outperforms existing routing baselines, achieving an average performance gain of 7.7\% over the state-of-the-art.



Paperid:341
Authors:Maya Martirossyan, Thomas Egg, Philipp Höllmer, George Karypis, Mark Transtrum, Adrian Roitberg, Mingjie Liu, Richard Hennig, Ellad Tadmor, Stefano Martiniani
Abstract:
Abstract:Generative models for materials, especially inorganic crystals, have attracted significant interest for their potential to transform the theoretical prediction of novel compounds and structures. Progress in this area depends critically on robust benchmarks and minimal, information-rich datasets that enable efficient and meaningful model evaluation. This paper critically examines commonly used datasets, methodologies, and evaluation metrics for the crystal structure prediction task—predicting the most likely structures given a chemical composition—and offers concrete solutions. We focus on three key issues: First, materials datasets should contain diverse and unique crystal structures; for example, we show that the widely-utilized carbon-24 dataset only contains $\approx 40$% unique structures, with duplicates differing only by the choice of unit cell representation. Second, materials datasets need to be split with care, rather than randomly, if polymorphs of many different compositions are numerous—which we discover to be the case for the perov-5 dataset. Third, benchmarks for evaluation of generative models can be misleading if used uncritically—for example, the reporting of a 'match rate' metric without consideration of the structural complexity that can be exhibited by identical building blocks (atoms). To address these oft-overlooked issues, we introduce several fixes. We provide revised versions of the carbon-24 dataset: one with duplicates removed, one deduplicated and split by number of atoms $N$, and two containing only duplicates. We also propose a new split for the perov-5 dataset that ensures polymorphs are grouped within the same training, validation, or test set in order to set a more sensible standard for benchmarking model performance. Finally, we present METRe and cRMSE, new model evaluation metrics that can correctly handle materials datasets with polymorphs.



Paperid:342
Authors:Jesimon Barreto, Carlos Caetano, Andre Araujo, William Schwartz
Abstract:
Foundation models have advanced computer vision by enabling strong performance across diverse tasks through large-scale pretraining and supervised fine-tuning. However, they may underperform in domains with distribution shifts and scarce labels, where supervised fine-tuning may be infeasible. While continued self-supervised learning for model adaptation is common for generative language models, this strategy has not proven effective for vision-centric encoder models. To address this challenge, we introduce a novel formulation of self-supervised fine-tuning for vision foundation models, where the model is adapted to a new domain without requiring new annotations. Our method, referred to as VESSA (Video-based Efficient SSL Adaptation), leverages two simple and effective insights. First, self-supervised fine-tuning is an important strategy that can be applied efficiently and simply, offering significant benefits. Second, to achieve meaningful improvements in the learning process, it is essential to use videos rather than still images during training. Through comprehensive experiments with 3 vision foundation models on 2 datasets, VESSA demonstrates consistent improvements in downstream tasks, compared to the base models and previous adaptation methods. Code and models will be released.



Paperid:343
Authors:Sungyoon Lee
Abstract:
In-context learning (ICL) is one of the key capabilities contributing to the great success of LLMs. At test time, ICL is known to operate in the two modes: task recognition and task learning. In this paper, we investigate the emergence and dynamics of the two modes of ICL during pretraining. To provide an analytical understanding of the learning dynamics of the ICL abilities, we investigate the in-context random linear regression problem with a simple linear-attention-based transformer, and define and disentangle the strengths of the task recognition and task learning abilities stored in the transformer model’s parameters. We show that, during the pretraining phase, the model first learns the task learning and the task recognition abilities together in the beginning, but it (a) gradually forgets the task recognition ability to recall the priorly learned tasks and (b) relies more on the given context in the later phase, which we call (a) \textit{prior forgetting} and (b) \textit{in-context overfitting}, respectively.



Paperid:344
Authors:Jiachen Zhao, Jing Huang, Zhengxuan Wu, David Bau, Weiyan Shi
Abstract:
LLMs are trained to refuse harmful instructions, but do they truly understand harmfulness beyond just refusing? Prior work has shown that LLMs' refusal behavior is mediated by a one-dimensional subspace, i.e., a refusal direction, in the latent space. In this work, we identify a new dimension to analyze safety mechanisms in LLMs: we have found that harmfulness is encoded as a distinct concept from refusal in their latent representations. To disentangle these two safety signals, we construct a prompting task that inverts models' refusal response to acceptance and vice versa. We find that steering along the harmfulness direction causes LLMs to interpret harmless instructions as harmful, and reason about that increased harmfulness internally to fulfill the inversion task, while applying the refusal direction tends to directly amplify refusal signals without reversing the model’s harmfulness judgment. Furthermore, our clustering analysis of hidden states reveals that some jailbreak methods may not radically suppress the model's internal harmfulness judgment, but directly reduce refusal signals. We also observe that narrow fine-tuning that reverses models' behaviors has minimal impact on the model's underlying beliefs about harmfulness and refusal. These insights lead to a practical application that latent harmfulness representations can serve as an intrinsic safeguard for detecting unsafe inputs and reducing misrefusals, which is also robust to fine-tuning attacks. For example, on Llama2, the latent harmfulness classifier can detect all the tested harmful prompts that jailbreak the model with adversarial suffixes.



Paperid:345
Authors:Brandon Wood, Misko Dzamba, Xiang Fu, Meng Gao, Muhammed Shuaibi, Luis Barroso-Luque, Kareem Abdelmaqsoud, Vahe Gharakhanyan, John Kitchin, Daniel Levine, Kyle Michel, Anuroop Sriram, Taco Cohen, Abhishek Das, Sushree Sahoo, Ammar Rizvi, Zachary Ulissi, Larry Zitnick
Abstract:
Abstract:The ability to quickly and accurately compute properties from atomic simulations is critical for advancing a large number of applications from drug discovery to energy storage to semiconductor manufacturing. We introduce a family of Universal Models for Atoms (UMA), designed to push the frontier of speed, accuracy, and generalization across chemistry and materials science. UMA models are trained on half a billion unique 3D atomic structures (the largest training runs to date) by compiling data across multiple chemical domains, e.g. molecules, materials, and catalysis. We develop empirical scaling laws to help understand how to increase model capacity alongside dataset size to achieve the best accuracy. The UMA small and medium models utilize a novel architectural design we refer to as mixture of linear experts that enables increasing model capacity without sacrificing speed. For example, UMA-medium has 1.4B parameters but only $\sim$50M active parameters per atomic structure. We evaluate UMA models on a diverse set of applications across multiple domains and find that, remarkably, a single model without any fine-tuning can perform similarly or better than specialized models. All code, weights, and associated data are released with permissive licenses.



Paperid:346
Authors:Jiajun Hong, Jianan Wei, Wenguan Wang
Abstract:
Human-Object Interaction (HOI) detection aims to localize human-object pairs and identify interactive relationships between them. To aggregate the contextual cues of the scene, current HOI methods either propagate information across all the detected entities via self-attention mechanisms, or establish message passing between humans and objects using a bipartite graph architecture. However, theyall neglect inherent social attributes within human-centric scenarios, which are collective and beyond pairwise. In light of this, we revisit the relation modeling in HOI from a group view, and propose GroupHOI, a novel framework that propagates contextual information in terms of geometric proximity and semantic similarity. To exploit geometric proximity, humans and objects are grouped into distinct clusters using a learnable proximity estimator based on spatial features derived from bounding boxes. In each group, a soft correspondence is computed via self-attention to aggregate and dispatch contextual cues. To incorporate semantic similarity, we enhance the vanilla transformer-based interaction decoder with semantic-aware local cues derived from HO-pair features. Extensive experiments on HICO-DET and V-COCO benchmarks demonstrate the superiority of GroupHOI over the state-of-the-art methods. The source code will be released.



Paperid:347
Authors:Seohong Park, Kevin Frans, Deepinder Mann, Benjamin Eysenbach, Aviral Kumar, Sergey Levine
Abstract:
In this work, we study the scalability of offline reinforcement learning (RL) algorithms. In principle, a truly scalable offline RL algorithm should be able to solve any given problem, regardless of its complexity, given sufficient data, compute, and model capacity. We investigate if and how current offline RL algorithms match up to this promise on diverse, challenging, previously unsolved tasks, using datasets up to 1000× larger than typical offline RL datasets. We observe that despite scaling up data, many existing offline RL algorithms exhibit poor scaling behavior, saturating well below the maximum performance. We hypothesize that the horizon is the main cause behind the poor scaling of offline RL. Through several analysis experiments, we show that even when controlling for other factors, long horizons can present a fundamental barrier to scaling up offline RL. We then show that various horizon reduction techniques can substantially enhance scalability on challenging tasks. Finally, we propose SHARSA, a minimalistic yet scalable method that effectively reduces the horizon, and show that it achieves the best asymptotic performance and scaling behavior among our evaluation methods.



Paperid:348
Authors:Yana Veitsman, Mayank Jobanputra, Yash Sarrof, Aleksandra Bakalova, Vera Demberg, Ellie Pavlick, Michael Hahn
Abstract:
Transformers have theoretical limitations in modeling certain sequence-to-sequence tasks, yet it remains largely unclear if these limitations play a role in large-scale pretrained LLMs, or whether LLMs might effectively overcome these constraints in practice due to the scale of both the models themselves and their pretraining data. We explore how these architectural constraints manifest after pretraining, by studying a family ofretrievalandcopyingtasks inspired by Liu et al. [2024]. We use the recently proposed C-RASP framework for studying length generalization [Huang et al., 2025b] to provide guarantees for each of our settings. Empirically, we observe aninduction-versus-anti-induction asymmetry, where pretrained models are better at retrieving tokens to the right (induction) rather than the left (anti-induction) of a query token. This asymmetry disappears upon targeted fine-tuning if length-generalization is guaranteed by theory. Mechanistic analysis reveals that this asymmetry is connected to the differences in the strength of induction versus anti-induction circuits within pretrained Transformers. We validate our findings through practical experiments on real-world tasks demonstrating reliability risks. Our results highlight that pretraining selectively enhances certain Transformer capabilities, but does not overcome fundamental length-generalization limits.



Paperid:349
Authors:Liang Zhang, Bingcong Li, Kiran Thekumparampil, Sewoong Oh, Michael Muehlebach, Niao He
Abstract:
Zeroth-order methods are extensively used in machine learning applications where gradients are infeasible or expensive to compute, such as black-box attacks, reinforcement learning, and language model fine-tuning. Existing optimization theory focuses on convergence to an arbitrary stationary point, but less is known on the implicit regularization that provides a fine-grained characterization on which particular solutions are finally reached. We show that zeroth-order optimization with the standard two-point estimator favors solutions with small trace of Hessian, which is widely used in previous work to distinguish between sharp and flat minima. We further provide convergence rates of zeroth-order optimization to approximate flat minima for convex and sufficiently smooth functions, where flat minima are defined as the minimizers that achieve the smallest trace of Hessian among all optimal solutions. Experiments on binary classification tasks with convex losses and language model fine-tuning support our theoretical findings.



Paperid:350
Authors:Heeju Ko, Sung June Kim, Gyeongrok Oh, Jeongyoon Yoon, Honglak Lee, Sujin Jang, Seungryong Kim, Sangpil Kim
Abstract:
Vision-Language Navigation (VLN) policies trained on offline datasets often exhibit degraded task performance when deployed in unfamiliar navigation environments at test time, where agents are typically evaluated without access to external interaction or feedback. Entropy minimization has emerged as a practical solution for reducing prediction uncertainty at test time; however, it can suffer from accumulated errors, as agents may become overconfident in incorrect actions without sufficient contextual grounding. To tackle these challenges, we introduce ATENA (Active TEst-time Navigation Agent), a test-time active learning framework that enables a practical human-robot interaction via episodic feedback on uncertain navigation outcomes. In particular, ATENA learns to increase certainty in successful episodes and decrease it in failed ones, improving uncertainty calibration. Here, we propose mixture entropy optimization, where entropy is obtained from a combination of the action and pseudo-expert distributions—a hypothetical action distribution assuming the agent's selected action to be optimal—controlling both prediction confidence and action preference. In addition, we propose a self-active learning strategy that enables an agent to evaluate its navigation outcomes based on confident predictions. As a result, the agent stays actively engaged throughout all iterations, leading to well-grounded and adaptive decision-making. Extensive evaluations on challenging VLN benchmarks—REVERIE, R2R, and R2R-CE—demonstrate that ATENA successfully overcomes distributional shifts at test time, outperforming the compared baseline methods across various settings.



Paperid:351
Authors:Dipkamal Bhusal, Michael Clifford, Sara Rampazzi, Nidhi Rastogi
Abstract:
Interpreting deep neural networks through concept-based explanations offers a bridge between low-level features and high-level human-understandable semantics. However, existing automatic concept discovery methods often fail to align these extracted concepts with the model’s true decision-making process, thereby compromising explanation faithfulness. In this work, we propose FACE (Faithful Automatic Concept Extraction), a novel framework that combines Non-negative Matrix Factorization (NMF) with a Kullback-Leibler (KL) divergence regularization term to ensure alignment between the model’s original and concept-based predictions. Unlike prior methods that operate solely on encoder activations, FACE incorporates classifier supervision during concept learning, enforcing predictive consistency and enabling faithful explanations. We provide theoretical guarantees showing that minimizing the KL divergence bounds the deviation in predictive distributions, thereby promoting faithful local linearity in the learned concept space. Systematic evaluations on ImageNet, COCO, and CelebA datasets demonstrate that FACE outperforms existing methods across faithfulness and sparsity metrics.



Paperid:352
Authors:Kshipra Bhawalkar, Marios Mertzanidis, Divyarthi Mohan, Alexandros Psomas
Abstract:
Abstract:We study revenue maximization for agents with additive preferences, subject to downward-closed constraints on the set of feasible allocations. In seminal work,~\citet{alaei2014bayesian} introduced a powerful multi-to-single agent reduction based on an ex-ante relaxation of the multi-agent problem. This reduction employs a rounding procedure which is an online contention resolution scheme (OCRS) in disguise, a now widely-used method for rounding fractional solutions in online Bayesian and stochastic optimization problems. In this paper, we leverage our vantage point, 10 years after the work of Alaei, with a rich OCRS toolkit and modern approaches to analyzing multi-agent mechanisms; we introduce a general framework for designing non-sequential and sequential multi-agent, revenue-maximizing mechanisms, capturing a wide variety of problems Alaei's framework could not address. Our framework uses an \emph{interim} relaxation, that is rounded to a feasible mechanism using what we call a two-level OCRS, which allows for some structured dependence between the activation of its input elements. For a wide family of constraints, we can construct such schemes using existing OCRSs as a black box; for other constraints, such as knapsack, we construct such schemes from scratch. We demonstrate numerous applications of our framework, including a sequential mechanism that guarantees a $\frac{2e}{e-1} \approx 3.16$ approximation to the optimal revenue for the case of additive agents subject to matroid feasibility constraints. Finally, we show how our framework can be easily extended to multi-parameter procurement auctions, where we provide an OCRS for Stochastic Knapsack that might be of independent interest.



Paperid:353
Authors:Stephen Pasteris, Chris Hicks, Vasilios Mavroudis
Abstract:
Statistical parity is one of the most foundational constraints in algorithmic fairness and privacy. In this paper, we show that statistical parity can be enforced efficiently in the contextual bandit setting while retaining strong performance guarantees. Specifically, we present a meta-algorithm that transforms any efficient implementation of Hedge (or, equivalently, any discrete Bayesian inference algorithm) into an efficient contextual bandit algorithm that guarantees exact statistical parity on every trial. Compared to any comparator that satisfies the same statistical parity constraint, the algorithm achieves the same asymptotic regret bound as running the equivalent instance of Exp4 for each group. We also address the scenario where the target parity distribution is unknown and must be estimated online. Finally, using online-to-batch conversion, we extend our approach to the batch classification setting - achieving exact statistical parity there as well, whilst attaining excellent generalisation bounds. We believe these batch bounds to be a significant contribution to the literature in their own right.



Paperid:354
Authors:Zhiqiang Yan, Jianhao Jiao, Zhengxue Wang, Gim Hee Lee
Abstract:
Depth completion in dynamic scenes poses significant challenges due to rapid ego-motion and object motion, which can severely degrade the quality of input modalities such as RGB images and LiDAR measurements. Conventional RGB-D sensors often struggle to align precisely and capture reliable depth under such conditions. In contrast, event cameras, with their high temporal resolution and sensitivity to motion at the pixel level, provide complementary cues that are particularly beneficial in dynamic environments. To this end, we propose EventDC, the first event-driven depth completion framework. It consists of two key components: Event-Modulated Alignment (EMA) and Local Depth Filtering (LDF). Both modules adaptively learn the two fundamental components of convolution operations: offsets and weights conditioned on motion-sensitive event streams. In the encoder, EMA leverages events to modulate the sampling positions of RGB-D features to achieve pixel redistribution for improved alignment and fusion. In the decoder, LDF refines depth estimations around moving objects by learning motion-aware masks from events. Additionally, EventDC incorporates two loss terms to further benefit global alignment and enhance local depth recovery. Moreover, we establish the first benchmark for event-based depth completion, comprising one real-world and two synthetic datasets, to facilitate future research. Extensive experiments on this benchmark demonstrate the superiority of EventDC. Our code and dataset will be released on paper acceptance.



Paperid:355
Authors:Wu Ran, Weijia Zhang, ShuYang Pang, Zhu, Jinfan Liu, JingSheng Liu, Xin Cao, Qiang Li, Yichao Yan, Chao Ma
Abstract:
Low-Rank Adaptation (LoRA) methods have demonstrated considerable success in achieving parameter-efficient fine-tuning (PEFT) for Transformer-based foundation models. These methods typically fine-tune individual Transformer layers using independent LoRA adaptations. However, directly applying existing LoRA techniques to convolutional networks (ConvNets) yields unsatisfactory results due to the high correlation between the stacked sequential layers of ConvNets. To overcome this challenge, we introduce a novel framework called Correlated Low-Rank Adaptation (CoLoRA), which explicitly utilizes correlated low-rank matrices to model the inter-layer dependencies among convolutional layers. Additionally, to enhance tuning efficiency, we propose a parameter-free filtering method that enlarges the receptive field of LoRA, thus minimizing interference from non-informative local regions. Comprehensive experiments conducted across various mainstream vision tasks, including image classification, semantic segmentation, and object detection, illustrate that CoLoRA significantly advances the state-of-the-art PEFT approaches. Notably, our CoLoRA achieves superior performance with only 5\% of trainable parameters, surpassing full fine-tuning in the image classification task on the VTAB-1k dataset using ConvNeXt-S.



Paperid:356
Authors:Zitong Lan, Yiduo Hao, Mingmin Zhao
Abstract:
Achieving immersive auditory experiences in virtual environments requires flexible sound modeling that supports dynamic source positions.In this paper, we introduce a task called resounding, which aims to estimate impulse responses at arbitrary emitter locations from a sparse set of positions, analogous to the relighting problem in vision.We leverage the reciprocity property and introduce Versa, a physics-inspired approach to facilitating acoustic field learning.Our method creates physically valid samples with dense virtual emitter positions by exchanging emitter and listener poses.We also identify challenges in deploying reciprocity due to emitter/listener gain patterns and propose a self-supervised learning approach to address them.Results show that Versa substantially improve the performance of acoustic field learning on both simulated and real-world datasets in different metrics.A perceptual user study shows Versa can greatly improve the immersive spatial sound experience.We will release the code and dataset upon acceptance.



Paperid:357
Authors:Guanxiong Luo, Shoujin Huang
Abstract:
We proposeself-diffusion, a novel framework for solving inverse problems without relying on pretrained generative models. Traditional diffusion-based approaches require training a model on a clean dataset to learn to reverse the forward noising process. This model is then used to sample clean solutions---corresponding to posterior sampling from a Bayesian perspective---that are consistent with the observed data under a task. In contrast, self-diffusion introduces a self-contained iterative process that alternates between noising and denoising steps to progressively refine its estimate of the solution. At each step of self-diffusion, noise is added to the current estimate, and a self-denoiser, which is a single untrained convolutional network randomly initialized from scratch, is continuously trained for certain iterations via a data fidelity loss to predict the solution from the noisy estimate. Essentially, self-diffusion exploits the spectral bias of neural networks and modulates it through a scheduled noise process. Without relying on pretrained score functions or external denoisers on a clean dataset, this approach still remains adaptive to arbitrary forward operators and noisy observations, making it highly flexible and broadly applicable. We demonstrate the effectiveness of our approach on a variety of linear inverse problems, showing that self-diffusion achieves competitive or superior performance compared to other methods.



Paperid:358
Authors:Jiaxuan Peng, Mengshi Qi, Jie Zhang, Juan Zhu, Yong Li, Huadong Ma
Abstract:
The hybrid model parallelism combining tensor parallelism (TP) and pipeline parallelism (PP) has become the dominant solution for distributed training of Large Language Models (LLMs) and Multimodal LLMs (MLLMs). However, TP introduces significant collective communication overheads, while PP suffers from synchronization inefficiencies such as pipeline bubbles. Existing works primarily address these challenges from isolated perspectives, focusing either on overlapping TP communication or on flexible PP scheduling to mitigate pipeline bubbles. In this paper, we propose a new synergistic tensor and pipeline parallelism schedule that simultaneously reduces both types of bubbles. Our proposed schedule decouples the forward and backward passes in PP into fine-grained computation units, which are then braided to form a composite computation sequence. This compositional structure enables near-complete elimination of TP-related bubbles. Building upon this structure, we further design the PP schedule to minimize PP bubbles. Experimental results demonstrate that our approach improves training throughput by up to 12\% for LLMs and 16\% for MLLMs compared to existing scheduling methods. We will release our source code soon.



Paperid:359
Authors:Ziwen Wang, Jiajun Fan, Thao Nguyen, Heng Ji, Ge Liu
Abstract:
Contrastive learning has proven to be highly efficient and adaptable in shaping representation spaces across diverse modalities by pulling similar samples together and pushing dissimilar ones apart. However, two key limitations persist: (1) Without explicit regulation of the embedding distribution, semantically related instances can inadvertently be pushed apart unless complementary signals guide pair selection, and (2) excessive reliance on large in-batch negatives and tailored augmentations hinders generalization. To address these limitations, we propose Variational Supervised Contrastive Learning (VarCon), which reformulates supervised contrastive learning as variational inference over latent class variables and maximizes a posterior-weighted evidence lower bound (ELBO) that replaces exhaustive pair-wise comparisons for efficient class-aware matching and grants fine-grained control over intra-class dispersion in the embedding space. Trained exclusively on image data, our experiments on CIFAR-10, CIFAR-100, ImageNet-100, and ImageNet-1K show that VarCon (1) achieves state-of-the-art performance for contrastive learning frameworks, reaching 79.36% Top-1 accuracy on ImageNet-1K and 78.29% on CIFAR-100 with a ResNet-50 encoder while converging in just 200 epochs; (2) yields substantially clearer decision boundaries and semantic organization in the embedding space, as evidenced by KNN classification, hierarchical clustering results, and transfer-learning assessments; and (3) demonstrates superior performance in few-shot learning than supervised baseline and superior robustness across various augmentation strategies.



Paperid:360
Authors:Zhengyang Geng, Mingyang Deng, Xingjian Bai, J. Zico Kolter, Kaiming He
Abstract:
Abstract:We propose a principled and effective framework for one-step generative modeling. We introduce the notion of average velocity to characterize flow fields, in contrast to instantaneous velocity modeled by Flow Matching methods. A well-defined identity between average and instantaneous velocities is derived and used to guide neural network training. Our method, termed the \textit{MeanFlow} model, is self-contained and requires no pre-training, distillation, or curriculum learning. MeanFlow demonstrates strong empirical performance: it achieves an FID of 3.43 with a single function evaluation (1-NFE) on ImageNet 256$\times$256 trained from scratch, significantly outperforming previous state-of-the-art one-step diffusion/flow models. Our study substantially narrows the gap between one-step diffusion/flow models and their multi-step predecessors, and we hope it will motivate future research to revisit the foundations of these powerful models.



Paperid:361
Authors:Eric Xing, Abby Stylianou, Robert Pless, Nathan Jacobs
Abstract:
Massive-scale pretraining has made vision-language models increasingly popular for image-to-image and text-to-image retrieval across a broad collection of domains. However, these models do not perform well when used for challenging retrieval tasks, such as instance retrieval in very large-scale image collections. Recent work has shown that linear transformations of VLM features trained for instance retrieval can improve performance by emphasizing subspaces that relate to the domain of interest. In this paper, we explore a more extreme version of this specialization by learning to map a given query to a query-specific feature space transformation. Because this transformation is linear, it can be applied with minimal computational cost to millions of image embeddings, making it effective for large-scale retrieval or re-ranking. Results show that this method consistently outperforms state-of-the-art alternatives, including those that require many orders of magnitude more computation at query time.



Paperid:362
Authors:Pierre Musacchio, Hyunmin Lee, Jaesik Park
Abstract:
Even in controlled settings, understanding instance-wise geometries is a challenging task for a wide range of visual models. Although expert systems exist, modern arts still rely on expensive input formats (category labels, binary segmentation masks) and inference costs (quadratic amount of forward passes). We mitigate these limitations by proposing InstaFormer, a network capable of holistic order prediction. That is, solely given an input RGB image, InstaFormer returns the full segmentation masks along with occlusion and depth orderings for all the instances in the scene in a single forward pass. At its core, InstaFormer relies on interactions between object queries and latent mask descriptors that semantically represent the same objects while carrying complementary information. We comprehensively benchmark and ablate our approach to highlight its effectiveness. The link to our repository will be shared upon publication.



Paperid:363
Authors:Siyuan Zhou, Yilun Du, Yuncong Yang, Lei Han, Peihao Chen, Dit-Yan Yeung, Chuang Gan
Abstract:
The ability to simulate the effects of future actions on the world is a crucial ability of intelligent embodied agents, enabling agents to anticipate the effects of their actions and make plans accordingly. While a large body of existing work has explored how to construct such world models using video models, they are often myopic in nature, without any memory of a scene not captured by currently observed images, preventing agents from making consistent long-horizon plans in complex environments where many parts of the scene are partially observed. We introduce a new persistent embodied world model with an explicit memory of previously generated content, enabling much more consistent long-horizon simulation. During generation time, our video diffusion model predicts RGB-D video of the future observations of the agent. This generation is then aggregated into a persistent 3D map of the environment. By conditioning the video model on this 3D spatial map, we illustrate how this enables video world models to faithfully simulate both seen and unseen parts of the world. Finally, we illustrate the efficacy of such a world model in downstream embodied applications, enabling effective planning and policy learning.



Paperid:364
Authors:Paul Couairon, Loïck Chambon, Louis Serrano, Jean-Emmanuel HAUGEARD, Matthieu Cord, Nicolas THOME
Abstract:
Foundation Vision Encoders have become indispensable across a wide range of dense vision tasks. However, their operation at low spatial feature resolutions necessitates subsequent feature decompression to enable full-resolution processing. To address this limitation, we introduce JAFAR, a lightweight and flexible feature upsampler designed to enhance the spatial resolution of visual features from any Foundation Vision Encoder to any target resolution. JAFAR features an attention-based upsampling module that aligns the spatial representations of high-resolution queries with semantically enriched low-resolution keys via Spatial Feature Transform modulation. Despite the absence of high-resolution feature ground truth; we find that learning at low upsampling ratios and resolutions generalizes surprisingly well to much higher scales. Extensive experiments demonstrate that JAFAR recovers intricate pixel-level details and consistently outperforms existing feature upsampling techniques across a diverse set of dense downstream applications.



Paperid:365
Authors:William Powell, Jeongyeol Kwon, Qiaomin Xie, Hanbaek Lyu
Abstract:
We study offline policy optimization for infinite-horizon average-reward Markov decision processes (MDPs) with large or infinite state spaces. Specifically, we propose a pessimistic actor-critic algorithm that uses a computationally efficient linear function class for value function estimation. At the core of our method is a critic that computes a pessimistic estimate of the average reward under the current policy, as well as the corresponding policy gradient, by solving a fixed-point Bellman equation, rather than solving a successive sequence of regression problems as in finite horizon settings. This procedure reduces to solving a second-order cone program, which is computationally tractable. Our theoretical analysis is based on a weak partial data coverage assumption, which requires only that the offline data aligns well with the expected feature vector of a comparator policy. Under this condition, we show that our algorithm achieves the optimal sample complexity of O(\varepsilon^{-2}) for learning a near-optimal policy, up to model misspecification errors.



Paperid:366
Authors:Nikola Jovanović, Ismail Labiad, Tomas Soucek, Martin Vechev, Pierre Fernandez
Abstract:
Watermarking the outputs of generative models has emerged as a promising approach for tracking their provenance. Despite significant interest in autoregressive image generation models and their potential for misuse, no prior work has attempted to watermark their outputs at the token level. In this work, we present the first such approach by adapting language model watermarking techniques to this setting. We identify a key challenge: the lack of reverse cycle-consistency (RCC), wherein re-tokenizing generated image tokens significantly alters the token sequence, effectively erasing the watermark. To address this and to make our method robust to common image transformations and removal attacks, we introduce a custom tokenizer-detokenizer finetuning procedure that improves RCC and a watermark synchronization step. As our experiments demonstrate, our approach enables robust watermark detection and yields theoretically grounded p-values.



Paperid:367
Authors:Yiwen Kou, Raghu Meka
Abstract:
Abstract:Agnostic learning of Boolean halfspaces is a fundamental problem in computational learning theory, but it is known to be computationally hard even for weak learning. Recent work \citep{chandrasekaran2024smoothed} proposed smoothed analysis as a way to bypass such hardness, but existing frameworks rely on additive Gaussian perturbations, making them unsuitable for discrete domains. We introduce a new smoothed agnostic learning framework for Boolean inputs, where perturbations are modeled via random bit flips. This defines a natural discrete analogue of smoothed optimality generalizing the Gaussian case. Under strictly subexponential assumptions on the input distribution, we give an efficient algorithm for learning halfspaces in this model, with runtime and sample complexity $\tilde{O}(n^{\mathrm{poly}(\frac{1}{\sigma\epsilon})})$. Previously, such algorithms were known only with strong structural assumptions for the discrete hypercube—for example, independent coordinates or symmetric distributions. Our result provides the first computationally efficient guarantee for smoothed agnostic learning of halfspaces over the Boolean hypercube, bridging the gap between worst-case intractability and practical learnability in discrete settings.



Paperid:368
Authors:Sílvia Casacuberta, Varun Kanade
Abstract:
We propose new learning algorithms for building selective classifiers, which are predictors that are allowed to abstain on some fraction of the domain. We study the model where a classifier may abstain from predicting at a fixed cost. Building on the recent framework on multigroup fairness and omniprediction, given a pre-specified class of loss functions, we provide an algorithm for building a single classifier that learns abstentions and predictions optimally for every loss in the entire class, where the abstentions are decided efficiently for each specific loss function by applying a fixed post-processing function. Our algorithm and theoretical guarantees generalize the previously-known algorithms for learning selective classifiers in formal learning-theoretic models.We then extend the traditional multicalibration algorithms to propose a new notion of "multicalibration with abstention", which we use to efficiently build accurate selective classifiers that abstain optimally not only globally but also locally within each of the groups in any pre-specified collection of possibly intersecting subgroups of the domain, and are also accurate when they do not abstain. We show how our abstention algorithms can be used as conformal prediction methods in the binary classification setting to achieve both marginal and group-conditional coverage guarantees for an intersecting collection of groups. We provide empirical evaluations for all of our theoretical results, demonstrating the practicality of our learning algorithms for the goal of abstaining optimally and fairly.



Paperid:369
Authors:Reece Shuttleworth, Jacob Andreas, Antonio Torralba, Pratyusha Sharma
Abstract:
Fine-tuning is a crucial paradigm for adapting pre-trained large language models to downstream tasks. Recently, methods like Low-Rank Adaptation (LoRA) have been shown to effectively fine-tune LLMs with an extreme reduction in trainable parameters. But, \emph{are their learned solutions really equivalent?} We study how LoRA and full-finetuning change pre-trained models by analyzing the model's weight matrices through the lens of their spectral properties. We find that LoRA and full fine-tuning yield weight matrices whose singular value decompositions exhibit very different structure: weight matrices trained with LoRA have new, high-ranking singular vectors, which we call \emph{intruder dimensions}, while those trained with full fine-tuning do not. Further, we extend the finding that LoRA forgets less than full fine-tuning and find its forgetting is vastly localized to the intruder dimension -- by causally intervening on the intruder dimensions by changing their associated singular values post-fine-tuning, we show that they cause forgetting. Moreover, scaling them down significantly improves modeling of the pre-training distribution with a minimal drop in downstream task performance. Given this, we should expect accumulating intruder dimensions to be harmful and lead to more forgetting. This will be amplified during continual learning because of sequentially fine-tuning, and we show that LoRA models do accumulate intruder dimensions here tend to perform worse in this setting, emphasizing the practicality of our findings.



Paperid:370
Authors:Chuheng Zhang, Tim Pearce, Pushi Zhang, Kaixin Wang, Xiaoyu Chen, Wei Shen, Li Zhao, Jiang Bian
Abstract:
Latent action models (LAMs) aim to learn action-relevant changes from unlabeled videos by compressing changes between frames as latents. However, differences between video frames can be caused by \textit{controllable changes} as well as exogenous noise, leading to an important concern -- do latents capture the changes caused by actions or irrelevant noise? This paper studies this issue analytically, presenting a linear model that encapsulates the essence of LAM learning, while being tractable. This provides several insights, including connections between LAM and principal component analysis (PCA), desiderata of the data-generating policy, and justification of strategies to encourage learning controllable changes using data augmentation, data cleaning, and auxiliary action-prediction. We also provide illustrative results based on numerical simulation, shedding light on the specific structure of observations, actions, and noise in data that influence LAM learning.



Paperid:371
Authors:Gaole Dai, Chenghao Zhou, Yu Zhou, Rongyu Zhang, Yuan Zhang, Chengkai Hou, Tiejun Huang, Jianxu Chen, Shanghang Zhang
Abstract:
Deep learning has emerged as a pivotal tool for accelerating research in the life sciences, with the low-level processing of biomedical images (e.g., registration, fusion, restoration, super-resolution) being one of its most critical applications. Platforms such as ImageJ (Fiji) and napari have enabled the development of customized plugins for various models. However, these plugins are typically based on models that are limited to specific tasks and datasets, making them less practical for biologists. To address this challenge, we introduceOrochi, the first application-oriented, efficient, and versatile image processor designed to overcome these limitations. Orochi is pre-trained on patches/volumes extracted from the raw data of over 100 publicly available studies using our Random Multi-scale Sampling strategy. We further propose Task-related Joint-embedding Pre-Training (TJP), which employs biomedical task-related degradation for self-supervision rather than relying on Masked Image Modelling (MIM), which performs poorly in downstream tasks such as registration. To ensure computational efficiency, we leverage Mamba's linear computational complexity and construct Multi-head Hierarchy Mamba. Additionally, we provide a three-tier fine-tuning framework (Full, Normal, and Light) and demonstrate that Orochi achieves comparable or superior performance to current state-of-the-art specialist models, even with lightweight parameter-efficient options. We hope that our study contributes to the development of an all-in-one workflow, thereby relieving biologists from the overwhelming task of selecting among numerous models. Our pre-trained weights and code will be released.



Paperid:372
Authors:Q.G. Duan, Benyun ZHAO, Mingqiao Han, Yijun Huang, Ben Chen
Abstract:
Scene understanding based on 3D Gaussian Splatting (3DGS) has recently achieved notable advances. Although 3DGS related methods have efficient rendering capability, they fail to address the inherent contradiction between the anisotropic color representation of GS primitives and the isotropic requirements of semantic features, leading to insufficient cross-view feature consistency.To overcome the limitation, this work proposes FHGS (Feature-Homogenized Gaussian Splatting), a novel 3D feature fusion framework inspired by physical models, which can achieve high-precision mapping of arbitrary 2D features from pre-trained models to 3D scenes while preserving the real-time rendering efficiency of 3DGS.Specifically, our FHGS introduces the following innovations: Firstly, a universal feature fusion architecture is proposed, enabling robust embedding of large-scale pre-trained models' semantic features (e.g., SAM, CLIP) into sparse 3D structures.Secondly, a non-differentiable feature fusion mechanism is introduced, which enables semantic features to exhibit viewpoint independent isotropic distribution. This fundamentally balances the anisotropic rendering of gaussian primitives and the isotropic expression of features; Thirdly, a dual-driven optimization strategy inspired by electric potential fields is proposed, which combines external supervision from semantic feature fields with internal primitive clustering guidance. This mechanism enables synergistic optimization of global semantic alignment and local structural consistency.Extensive comparison experiments with other state-of-the-art methods on benchmark datasets demonstrate that our FHGS exhibits superior reconstruction performance in feature fusion, noise suppression, geometric precision. This work establishes a novel Gaussian Splatting (GS) data structure, offering practical advancements for real-time semantic mapping, 3D stylization, and interactive tasks in unmanned systems.



Paperid:373
Authors:Jiaru Zou, Ling Yang, Jingwen Gu, Jiahao Qiu, Ke Shen, Jingrui He, Mengdi Wang
Abstract:
Process Reward Models have recently emerged as a powerful framework for supervising intermediate reasoning steps in large language models. Existing PRMs are primarily trained on model final responses and struggle to robustly evaluate intermediate thinking traces, especially in the emerging setting of trajectory–response data generated by frontier models like OpenAI-o1 and DeepSeek-R1. These silver-standard traces are increasingly used in reasoning distillation and test-time inference, yet remain noisy and difficult to supervise. In this work, we introduce TAP (Trajectory-Aware PRM), a novel reward modeling framework explicitly designed to evaluate multi-step reasoning processes. TAP incorporates both step-level and trajectory-level supervision, enabling fine-grained reward assignment aligned with structured chain-of-thought data. We further adapt TAP into both offline and online reward supervision: select high-quality data for model distillation, provide dense process-level rewards for reinforcement learning, and support reward-guided test-time scaling via a best-of-N strategy. Empirical results on challenging benchmarks such as AIME and GPQA show that TAP outperforms other strong reward model baselines, leading to over 22.7% performance improvement.



Paperid:374
Authors:Zhen Li, Chuanhao Li, Xiaofeng Mao, Shaoheng Lin, Ming Li, Shitian Zhao, Zhaopan Xu, Xinyue Li, Yukang Feng, Jianwen Sun, Zizhen Li, Fanrui Zhang, Jiaxin Ai, Zhixiang Wang, Yuwei Wu, Tong He, Yunde Jia, Kaipeng Zhang
Abstract:
World exploration is an important human activity to forms the foundation of humankind's odyssey, while video generation techniques have made remarkable progress, promising to be the foundation of interactive world exploration.However, existing video generation datasets are not well-suited for world exploration training as they suffer from some limitations: limited duration, static scenes, and lack of exploratory and world annotations.In this paper, we introduce Sekai せかい, meaning "world'' in Japanese), a high-quality first-person view worldwide video dataset with rich annotations for world exploration. It consists of over 6,000 hours of walking or drone view (FPV and UVA) videos from 65 countries, over 1000 cities. We develop an efficient toolchain to pre-process videos and an effective toolbox to annotate them. For each video, we annotate location, scene type, weather, crowd density, captions, and camera trajectories.Experiments demonstrate the quality and effectiveness of the dataset. We believe Sekai will benefit the area of video generation and world exploration, and raise valuable applications. We make an introductory video about the dataset in the supplemental material.



Paperid:375
Authors:Huiri Tan, Juyong Jiang, Jiasi Shen
Abstract:
The landscape of compiler optimization has evolved significantly with the advent of profile-guided optimization (PGO), a technique that leverages program runtime information to generate highly optimized binaries. While instrumentation-based profiling method provides accurate profile data, it often suffers from heavy runtime overhead. A more efficient alternative is sampling-based profiling, which offers a rapid and scalable method for collecting profile data while incurring minimal runtime overhead and avoiding intrusive source code modifications. However, accurately collecting execution profiles via sampling remains challenging, especially when applied to fully optimized binaries. Such inaccurate profile data can diminish the potential benefits of PGO. This paper addresses the problem of profile inference, aiming to correct inaccuracies in the profiles collected by sampling. Our contributions include a novel Graph Neural Network (GNN)-based model to enhance profile accuracy. Extensive experiments on the SPEC 2017 benchmarks demonstrate that our GNN-based model achieves up to a 9.15\% performance improvement compared to the state-of-the-art traditional algorithm and an average 6.87\% improvement over the baseline machine learning models, showcasing the effectiveness of our approach in optimizing real-world application profiles.



Paperid:376
Authors:Hongjun Wang, yitong jiang, David Wehr, Hanrong Ye, Xinhao Li, Ka Chun Cheung, Kai Han, Hongxu Yin, Pavlo Molchanov, Sifei Liu, Wonmin Byeon, Collin McCarthy, Jinwei Gu, Jan Kautz, Ke Chen
Abstract:
Efficient vision transformer remains a bottleneck for high-resolution images and long-video related real-world applications. Generalized Spatial Propagation Network (GSPN) \cite{wang2025parallel} addresses this by replacing quadratic self-attention with a line-scan propagation scheme, bringing the cost close to linear in the number of rows or columns, while retaining accuracy. Despite this advancement, the existing GSPN implementation still suffers from (i) heavy overhead due to repeatedly launching GPU kernels, (ii) excessive data transfers from global GPU memory, and (iii) redundant computations caused by maintaining separate propagation weights for each channel. We introduce GSPN-2, a joint algorithm–system redesign. In particular, we eliminate thousands of micro-launches from the previous implementation into one single 2D kernel, explicitly pin one warp to each channel slice, and stage the previous column's activations in shared memory. On the model side, we introduce a set of channel-shared propagation weights that replace per-channel matrices, trimming parameters, and align naturally with the affinity map used in transformer attention. Experiments demonstrate GSPN-2's effectiveness across image classification and text-to-image synthesis tasks, matching transformer-level accuracy with significantly lower computational cost. GSPN-2 establishes a new efficiency frontier for modeling global spatial context in vision applications through its unique combination of structured matrix transformations and GPU-optimized implementation.



Paperid:377
Authors:Benyu Wu, Yue Liu, Qiaoyu Tan, Xinwang Liu, Wei Du, Jun Wang, Guoxian Yu
Abstract:
Abstract:Deep graph clustering (DGC) aims to partition nodes in a graph into distinct clusters in an unsupervised manner. Despite rapid advancements in this field, DGC remains inherently challenging due to the absence of ground truth, which complicates the design of effective algorithms and impedes the establishment of standardized benchmarks. The lack of unified datasets, evaluation protocols, and metrics further exacerbates these challenges, making it difficult to systematically assess and compare DGC methods. To address these limitations, we introduce $\texttt{DGCBench}$, the first comprehensive and unified benchmark for DGC methods. It evaluates 12 state-of-the-art DGC methods across 12 datasets from diverse domains and scales, spanning 6 critical dimensions: $\textbf{discriminability}$, $\textbf{effectiveness}$, $\textbf{scalability}$, $\textbf{efficiency}$, $\textbf{stability}$, and $\textbf{robustness}$. Additionally, we develop $\texttt{PyDGC}$, an open-source Python library that standardizes the DGC training and evaluation paradigm. Through systematic experiments, we reveal persistent limitations in existing methods-specifically regarding homophily assumptions, training instability, vulnerability to perturbations, efficiency plateau, scalability challenges, and poor discriminability—thereby offering actionable insights for future research. We hope that $\texttt{DGCBench}$, $\texttt{PyDGC}$, and our analyses will collectively accelerate the progress in the DGC community. The code is available at https://github.com/Marigoldwu/PyDGC.



Paperid:378
Authors:Guohao Sun, Hang Hua, Jian Wang, Jiebo Luo, Sohail Dianat, MAJID RABBANI, Raghuveer Rao, Zhiqiang Tao
Abstract:
Chain-of-thought (CoT) reasoning is critical for improving the interpretability and reliability of Large Vision-Language Models (LVLMs). However, existing training algorithms such as SFT, PPO, and GRPO may not generalize well across unseen reasoning tasks and heavily rely on a biased reward model. To address this challenge, we reformulate reasoning in LVLMs as posterior inference and propose a scalable training algorithm based on amortized variational inference. By leveraging diversity-seeking reinforcement learning algorithms, we introduce a novel sparse reward function for token-level learning signals that encourage diverse, high-likelihood latent CoT, overcoming deterministic sampling limitations and avoiding reward hacking. Additionally, we implement a Bayesian inference-scaling strategy that replaces costly Best-of-N and Beam Search with a marginal likelihood to efficiently rank optimal rationales and answers. We empirically demonstrate that the proposed method enhances the state-of-the-art LVLMs on four reasoning benchmarks, in terms of effectiveness, generalization, and interpretability.



Paperid:379
Authors:Hidir Yesiltepe, Pinar Yanardag Delul
Abstract:
In this work, we address dynamic view synthesis from monocular videos as an inverse problem in a training-free setting. By redesigning the noise initialization phase of a pre-trained video diffusion model, we enable high-fidelity dynamic view synthesis without any weight updates or auxiliary modules. We begin by identifying a fundamental obstacle to deterministic inversion arising from zero-terminal signal-to-noise ratio (SNR) schedules and resolve it by introducing a novel noise representation, termed K-order Recursive Noise Representation. We derive a closed form expression for this representation, enabling precise and efficient alignment between the VAE-encoded and the DDIM inverted latents. To synthesize newly visible regions resulting from camera motion, we introduce Stochastic Latent Modulation, which performs visibility aware sampling over the latent space to complete occluded regions. Comprehensive experiments demonstrate that dynamic view synthesis can be effectively performed through structured latent manipulation in the noise initialization phase.



Paperid:380
Authors:Morris Yau, Ekin Akyürek, Jiayuan Mao, Josh Tenenbaum, Stefanie Jegelka, Jacob Andreas
Abstract:
Previous research has explored the expressivity of Transformer models in simulating Boolean circuits or Turing machines. However, the efficient learnability of Transformers from data has remained an open question. Our study addresses this gap by providing the first polynomial-time learnability results (specifically strong, agnostic PAC learning) for single-layer Transformers with linear attention. We show that learning the optimal multi head linear attention can be recast as finding the optimal kernel predictor in a suitably defined RKHS. Moving to generalization, we construct an algorithm that, given a dataset, checks in polynomial time whether the set of best fit multi head linear attention networks on this data all perform an identical computation--a powerful notion for out of distribution generalization. We empirically validate our theoretical findings on several canonical tasks: learning random linear attention networks, key--value associations, and learning to execute finite automata. Our findings bridge a critical gap between theoretical expressivity and learnability of Transformer models.



Paperid:381
Authors:Songsong Ouyang, Yingying Zhu
Abstract:
Cross-view geo-localization remains a challenging task due to drastic viewpoint and appearance discrepancies between ground-level and aerial (BEV) images. In this work, we propose a novel framework that fuses street-view and BEV representations to leverage their complementary information for robust localization. Our method introduces a feature fusion module that integrates BEV and ground-level features via residual-aware and spatial-attentive mechanisms. To further enhance the robustness of the street-view features, we incorporate a causal learning strategy that distills stable semantic cues through frequency-domain augmentation. The enhanced features are supervised by a lightweight Dice Loss branch while being detached from the main backpropagation to avoid disrupting the primary localization objective. Experimental results on benchmark datasets such as VIGOR and CVACT demonstrate that our approach significantly improves performance compared to baseline methods, particularly under challenging viewpoint variations.



Paperid:382
Authors:Hyunjin Kim, Kunho Kim, Wonkwang Lee, Adam Lee
Abstract:
We present GOATex, a diffusion-based method for 3D mesh texturing that generates high-quality textures for both exterior and interior surfaces. While existing methods perform well on visible regions, they inherently lack mechanisms to handle occluded interiors, resulting in incomplete textures and visible seams. To address this, we introduce an occlusion-aware texturing framework based on the concept of hit levels, which quantify the relative depth of mesh faces via multi-view ray casting. This allows us to partition mesh faces into ordered visibility layers, from outermost to innermost. We then apply a two-stage visibility control strategy that progressively reveals interior regions with structural coherence, followed by texturing each layer using a pretrained diffusion model. To seamlessly merge textures obtained across layers, we propose a soft UV-space blending technique that weighs each texture’s contribution based on view-dependent visibility confidence. Empirical results demonstrate that GOATex consistently outperforms existing methods, producing seamless, high-fidelity textures across both visible and occluded surfaces. Unlike prior works, GOATex operates entirely without costly fine-tuning of a pretrained diffusion model and allows separate prompting for exterior and interior mesh regions, enabling fine-grained control over layered appearances.



Paperid:383
Authors:Henry Pinkard, Leyla Kabuli, Eric Markley, Tiffany Chien, Jiantao Jiao, Laura Waller
Abstract:
In modern imaging systems that computationally process raw measurements before or instead of human viewing, information content matters more than visual appearance. However, developing information estimators that can handle the complexity of real-world measurements yet remain practical enough for widespread use has proven challenging. We introduce a data-driven approach for estimating mutual information between unknown objects and their noisy measurements. Our technique fits probabilistic models to measurements and their noise processes, quantifying information content without requiring ground truth data or making assumptions about object structure. We validate our approach across diverse applications—color photography, radio astronomy, lensless imaging, and microscopy—demonstrating that information estimates reliably predict system performance. Finally, we introduce Information-Driven Encoder Analysis Learning (IDEAL), which optimizes imaging systems to maximize information capture. Our work unlocks information theory as a powerful, practical tool for analyzing and designing imaging systems across a broad range of applications.



Paperid:384
Authors:Ziang Cao, Zhaoxi Chen, Liang Pan, Ziwei Liu
Abstract:
3D modeling is moving from virtual to physical. Existing 3D generation primarily emphasizes geometries and textures while neglecting physical-grounded modeling. Consequently, despite the rapid development of 3D generative models, the synthesized 3D assets often overlook rich and important physical properties, hampering their real-world application in physical domains like simulation and embodied AI. As an initial attempt to address this challenge, we propose \textbf{PhysX}, an end-to-end paradigm for physical-grounded 3D asset generation.\textbf{1)} To bridge the critical gap in physics-annotated 3D datasets, we present \textbf{\ourname}\ - the first physics-grounded 3D dataset systematically annotated across five foundational dimensions:\textbf{\textcolor{color2}{absolute scale}}, \textbf{\textcolor{color3}{material}}, \textbf{\textcolor{color1}{affordance}}, \textbf{\textcolor{color4}{kinematics}}, and \textbf{\textcolor{color5}{function description}}. In particular, we devise a scalable human-in-the-loop annotation pipeline based on vision-language models, which enables efficient creation of physics-first assets from raw 3D assets. \textbf{2)} Furthermore, we propose \textbf{PhysXGen}, a feed-forward framework for physics-grounded 3D asset generation, injecting physical knowledge into the pre-trained 3D structural space.Specifically, PhysXGen employs a dual-branch architecture to explicitly model the latent correlations between 3D structures and physical properties, thereby producing 3D assets with plausible physical predictions while preserving the native geometry quality. Extensive experiments validate the superior performance and promising generalization capability of our framework. All the code, data, and models will be released to facilitate future research in generative physical AI.



Paperid:385
Authors:Jianhao Yan, Yafu Li, Zican Hu, Zhi Wang, Ganqu Cui, Xiaoye Qu, Yu Cheng, Yue Zhang
Abstract:
Recent advances in large reasoning models (LRMs) demonstrate that sophisticated behaviors such as multi-step reasoning and self-reflection can emerge via reinforcement learning with verifiable rewards~(RLVR). However, existing RLVR approaches are inherently ``on-policy'', limiting learning to a model's own outputs and failing to acquire reasoning abilities beyond its initial capabilities. To address this issue, we introduce LUFFY (Learning to reason Under oFF-policY guidance), a framework that augments RLVR with off-policy reasoning traces. LUFFY dynamically balances imitation and exploration by combining off-policy demonstrations with on-policy rollouts during training. Specifically, LUFFY combines the Mixed-Policy GRPO framework, which has a theoretically guaranteed convergence rate, alongside policy shaping via regularized importance sampling to avoid superficial and rigid imitation during mixed-policy training. Compared with previous RLVR methods, LUFFY achieves an over +6.4 average gain across six math benchmarks and an advantage of over +6.2 points in out-of-distribution tasks.Most significantly, we show that LUFFY successfully trains weak models in scenarios where on-policy RLVR completely fails. These results provide compelling evidence that LUFFY transcends the fundamental limitations of on-policy RLVR and demonstrates the great potential of utilizing off-policy guidance in RLVR.



Paperid:386
Authors:Davide Murari, Takashi Furuya, Carola-Bibiane Schönlieb
Abstract:
Abstract:$1$-Lipschitz neural networks are fundamental for generative modelling, inverse problems, and robust classifiers. In this paper, we focus on $1$-Lipschitz residual networks (ResNets) based on explicit Euler steps of negative gradient flows and study their approximation capabilities. Leveraging the Restricted Stone–Weierstrass Theorem, we first show that these $1$-Lipschitz ResNets are dense in the set of scalar $1$-Lipschitz functions on any compact domain when width and depth are allowed to grow. We also show that these networks can exactly represent scalar piecewise affine $1$-Lipschitz functions. We then prove a stronger statement: by inserting norm-constrained linear maps between the residual blocks, the same density holds when the hidden width is fixed. Because every layer obeys simple norm constraints, the resulting models can be trained with off-the-shelf optimisers. This paper provides the first universal approximation guarantees for $1$-Lipschitz ResNets, laying a rigorous foundation for their practical use.



Paperid:387
Authors:Yonatan Slutzky, ‪Yotam Alexander‬‏, Noam Razin, Nadav Cohen
Abstract:
Neural networks are powered by an implicit bias: a tendency of gradient descent to fit training data in a way that generalizes to unseen data. A recent class of neural network models gaining increasing popularity is structured state space models (SSMs). Prior work argued that the implicit bias of SSMs leads to generalization in a setting where data is generated by a low dimensional teacher. In this paper, we revisit the latter setting, and formally establish a phenomenon entirely undetected by prior work on the implicit bias of SSMs. Namely, we prove that while implicit bias leads to generalization under many choices of training data, there exist special examples whose inclusion in training completely distorts the implicit bias, to a point where generalization fails. This failure occurs despite the special training examples being labeled by the teacher, i.e., having clean labels! We empirically demonstrate the phenomenon, with SSMs trained independently and as part of non-linear neural networks. In the area of adversarial machine learning, disrupting generalization with cleanly labeled training examples is known as clean-label poisoning. Given the proliferation of SSMs, we believe significant efforts should be invested in delineating their susceptibility to clean-label poisoning, and in developing methods for overcoming this susceptibility.



Paperid:388
Authors:Yuga Hikida, Ayush Bharti, Niall Jeffrey, Francois-Xavier Briol
Abstract:
Neural simulation-based inference (SBI) is a popular set of methods for Bayesian inference when models are only available in the form of a simulator. These methods are widely used in the sciences and engineering, where writing down a likelihood can be significantly more challenging than constructing a simulator. However, the performance of neural SBI can suffer when simulators are computationally expensive, thereby limiting the number of simulations that can be performed. In this paper, we propose a novel approach to neural SBI which leverages multilevel Monte Carlo techniques for settings where several simulators of varying cost and fidelity are available. We demonstrate through both theoretical analysis and extensive experiments that our method can significantly enhance the accuracy of SBI methods given a fixed computational budget.



Paperid:389
Authors:Alejandro Hernández-Cano, Dhia Garbaya, Imanol Schlag, Martin Jaggi
Abstract:
Despite the significant potential of FP8 data formats for large language model (LLM) pre-training, their adoption has been limited due to challenges in maintaining stability at scale. Existing approaches often rely on suboptimal fine-grained FP8 kernels or fall back to higher-precision matrix multiplications (GEMMs) in sensitive components, such as attention projections, compromising potential throughput gains.We introduce a new class of LLM architectures that, for the first time, support FP8 computation for all GEMMs within transformer blocks during both forward and backward passes. This enables unprecedented throughput gains, particularly at scale, while matching the downstream performance of standard BF16 training. Our architecture design reduces large outlier activations, promoting stable long-term FP8 training. Additionally, we identify key metrics for monitoring low-precision training and predicting potential future divergences.



Paperid:390
Authors:Yige Yuan, Teng Xiao, Li Yunfan, Bingbing Xu, Shuchang Tao, Yunqi Qiu, Huawei Shen, Xueqi Cheng
Abstract:
Abstract:Aligning large language models with human feedback at inference time has received increasing attention due to its flexibility. Existing methods rely on generating multiple responses from the base policy for search using a reward model, which can be considered as searching in a discrete response space. However, these methods struggle to explore informative candidates when the base policy is weak or the candidate set is small, resulting in limited effectiveness. In this paper, to address this problem, we propose Simple Energy Adaptation ($\textbf{SEA}$), a simple yet effective algorithm for inference-time alignment. In contrast to expensive search over the discrete space, SEA directly adapts original responses from the base policy toward the optimal one via gradient-based sampling in continuous latent space. Specifically, SEA formulates inference as an iterative optimization procedure on an energy function over actions in the continuous space defined by the optimal policy, enabling simple and effective alignment. For instance, despite its simplicity, SEA outperforms the second-best baseline with a relative improvement of up to $ \textbf{77.51\%}$ on AdvBench and $\textbf{16.36\%}$ on MATH. Code is publicly available at [this link](https://github.com/SEA-NeurIPS/Code).



Paperid:391
Authors:Russell Tsuchida, Jiawei Liu, Cheng Soon Ong, Dino Sejdinovic
Abstract:
Squared families of probability distributions have been studied and applied in numerous machine learning contexts. Typically, they appear as likelihoods, where their advantageous computational, geometric and statistical properties are exploited for fast estimation algorithms, representational properties and statistical guarantees. Here, we investigate the use of squared families as prior beliefs in Bayesian inference.We find that they can form helpful conjugate families, often allowing for closed-form and tractable Bayesian inference and marginal likelihoods. We apply such conjugate families to Bayesian regression in feature space using end-to-end learnable neural network features. Such a setting allows for a rich multi-modal alternative to Gaussian processes with neural network features, often called deep kernel learning. We demonstrate our method on few shot learning, outperforming existing neural methods based on Gaussian processes and normalising flows.



Paperid:392
Authors:Yuxiang Yan, Zhiyuan Zhou, Xin Gao, Guanghao Li, Shenglin Li, Jiaqi Chen, Qunyan Pu, Jian Pu
Abstract:
Manipulation in cluttered environments is challenging due to spatial dependencies among objects, where an improper manipulation order can cause collisions or blocked access. Existing approaches often overlook these spatial relationships, limiting their flexibility and scalability. To address these limitations, we propose OrderMind, a unified spatial-aware manipulation ordering framework that directly learns object manipulation priorities based on spatial context. Our architecture integrates a spatial context encoder with a temporal priority structuring module. We construct a spatial graph using k-Nearest Neighbors to aggregate geometric information from the local layout and encode both object-object and object-manipulator interactions to support accurate manipulation ordering in real-time. To generate physically and semantically plausible supervision signals, we introduce a spatial prior labeling method that guides a vision-language model to produce reasonable manipulation orders for distillation. We evaluate OrderMind on our new Manipulation Ordering Benchmark, comprising 163,222 samples of varying difficulty. Extensive experiments in both simulation and real-world environments demonstrate that our method significantly outperforms prior approaches in effectiveness and efficiency, enabling robust manipulation in cluttered scenes. We will release our code upon paper acceptance.



Paperid:393
Authors:Linjiang Zhou, Chao Ma, Zepeng Wang, Libing Wu, XIAOCHUAN SHI
Abstract:
Abstract:Gradient smoothing is an efficient approach to reducing noise in gradient-based model explanation methods. SmoothGrad adds Gaussian noise to mitigate much of this noise. However, the crucial hyperparameter in this method, the variance $\sigma$ of the Gaussian noise, is often set manually or determined using a heuristic approach. This results in the smoothed gradients containing extra noise introduced by the smoothing process. In this paper, we aim to analyze the noise and its connection to the out-of-range sampling in the smoothing process of SmoothGrad. Based on this insight, we propose AdaptGrad, an adaptive gradient smoothing method that controls out-of-range sampling to minimize noise. Comprehensive experiments, both qualitative and quantitative, demonstrate that AdaptGrad could effectively reduce almost all the noise in vanilla gradients compared to baseline methods. AdaptGrad is simple and universal, making it a practical solution to enhance gradient-based interpretability methods to achieve clearer visualization.



Paperid:394
Authors:Yuan Qiu, Ke Yi
Abstract:
A fundamental problem in differential privacy is to release privatized answers to a class of linear queries with small error. This problem has been well studied in the static case. In this paper, we consider the fully dynamic setting where items may be inserted into or deleted from the dataset over time, and we need to continually release query answers at every time instance. We present efficient black-box constructions of such dynamic differentially private mechanisms from static ones with only a polylogarithmic degradation in the utility.



Paperid:395
Authors:Maxwell Fishelson, Noah Golowich, Mehryar Mohri, Jon Schneider
Abstract:
Abstract:We study the online calibration of multi-dimensional forecasts over an arbitrary convex set $\mathcal{P} \subset \mathbb{R}^d$ relative to an arbitrary norm $\Vert\cdot\Vert$. We connect this with the problem of external regret minimization for online linear optimization, showing that if it is possible to guarantee $O(\sqrt{\rho T})$ worst-case regret after $T$ rounds when actions are drawn from $\mathcal{P}$ and losses are drawn from the dual $\Vert \cdot \Vert_*$ unit norm ball, then it is also possible to obtain $\epsilon$-calibrated forecasts after $T = \exp(O(\rho /\epsilon^2))$ rounds. When $\mathcal{P}$ is the $d$-dimensional simplex and $\Vert \cdot \Vert$ is the $\ell_1$-norm, the existence of $O(\sqrt{T\log d})$ algorithms for learning with experts implies that it is possible to obtain $\epsilon$-calibrated forecasts after $T = \exp(O(\log{d}/\epsilon^2)) = d^{O(1/\epsilon^2)}$ rounds, recovering a recent result of Peng 2025.Interestingly, our algorithm obtains this guarantee without requiring access to any online linear optimization subroutine or knowledge of the optimal rate $\rho$ -- in fact, our algorithm is identical for every setting of $\mathcal{P}$ and $\Vert \cdot \Vert$. Instead, we show that the optimal regularizer for the above OLO problem can be used to upper bound the above calibration error by a swap regret, which we then minimize by running the recent TreeSwap algorithm with Follow-The-Leader as a subroutine. The resulting algorithm is highly efficient and plays a distribution over simple averages of past observations in each round.Finally, we prove that any online calibration algorithm that guarantees $\epsilon T$ $\ell_1$-calibration error over the $d$-dimensional simplex requires $T \geq \exp(\mathrm{poly}(1/\epsilon))$ (assuming $d \geq \mathrm{poly}(1/\epsilon)$). This strengthens the corresponding $d^{\Omega(\log{1/\epsilon})}$ lower bound of Peng 2025, and shows that an exponential dependence on $1/\epsilon$ is necessary.



Paperid:396
Authors:Xiaotian Zhang, Yue Shang, Entao Yang, Ge Zhang
Abstract:
Understanding neural network' (NN) generalizability remains a central question in deep learning research.The special phenomenon of grokking, where NNs abruptly generalize long after the training performance reaches near-perfect level, offers a unique window to investigate the underlying mechanisms of NNs' generalizability.Here we propose an interpretation for grokking by framing it as a computational glass relaxation: viewing NNs as a physical system where parameters are the degrees of freedom and train loss is the system energy, we find memorization process resembles a rapid cooling of liquid into non-equilibrium glassy state at low temperature and the later generalization is like a slow relaxation towards a more stable configuration. This mapping enables us to sample NNs' Boltzmann entropy (states of density) landscape as a function of training loss and test accuracy. Our experiments in transformers on arithmetic tasks suggests that there is NO entropy barrier in the memorization-to-generalization transition of grokking, challenging previous theory that defines grokking as a first-order phase transition.We identify a high-entropy advantage under grokking, an extension of prior work linking entropy to generalizability but much more significant. Inspired by grokking's far-from-equilibrium nature, we develop a toy optimizer WanD based on Wang-landau molecular dynamics, which can eliminate grokking without any constraints and find high-norm generalizing solutions. This provides strictly-defined counterexamples to theory attributing grokking solely to weight norm evolution towards the Goldilocks zone and also suggests new potential ways for optimizer design.



Paperid:397
Authors:Mattie Ji, Amauri Souza, Vikas Garg
Abstract:
Topological descriptors have been increasingly used for capturing multiscale structural information in relational data. In this work, we consider various filtrations on (Cartesian) graph products and the effect on their outputs on the topological descriptors - the Euler characteristics (EC) and persistent homology (PH). In particular, we establish a complete characterization of the expressive power of EC on general color-based filtrations. We also show that the PH descriptors of (virtual) graph products contain strictly more information than the computation on individual graphs while EC does not. Additionally, we provide algorithms to compute the PH diagrams of the product of vertex- and edge-level filtrations on the graph product. Overall, this work paves way for powerful graph persistent descriptors via graph product filtrations.



Paperid:398
Authors:Parikshit Ram, Kenneth Clarkson, Tim Klinger, Shashanka Ubaru, Alexander Gray
Abstract:
Various forms of sparse attention have been explored to mitigate the quadratic computational and memory cost of the attention mechanism in transformers. We study sparse transformers not through a lens of efficiency but rather in terms of learnability and generalization. Empirically studying a range of attention mechanisms, we find that input-dependent sparse attention models appear to converge faster and generalize better than standard attention models, while input-agnostic sparse attention models show no such benefits -- a phenomenon that is robust across architectural and optimization hyperparameter choices. This can be interpreted as demonstrating that concentrating a model's "semantic focus" with respect to the tokens currently being considered (in the form of input-dependent sparse attention) accelerates learning. We develop a theoretical characterization of the conditions that explain this behavior. We establish a connection between the stability of the standard softmax and the loss function's Lipschitz properties, then show how sparsity affects the stability of the softmax and the subsequent convergence and generalization guarantees resulting from the attention mechanism. This allows us to theoretically establish that input-agnostic sparse attention does not provide any benefits. We also characterize conditions when semantic focus (input-dependent sparse attention) can provide improved guarantees, and we validate that these conditions are in fact met in our empirical evaluations.



Paperid:399
Authors:Rui Huang, Guangyao Zhai, Zuria Bauer, Marc Pollefeys, Federico Tombari, Leonidas Guibas, Gao Huang, Francis Engelmann
Abstract:
Abstract:3D scene synthesis has traditionally required expert knowledge and considerable manual effort; automating this process holds the potential to greatly advance applications in architectural design, robotics simulation, virtual reality, and gaming.Most recent 3D scene synthesis models tap into the commonsense knowledge encoded in Large Language Models (LLMs) or leverage strong appearance priors of modern image generation models.However, current LLMs exhibit limited 3D spatial reasoning ability,which hinders their ability to generate realistic and coherent 3D scenes.Meanwhile, image generation-based methods often suffer from multi-view inconsistencies.In this work, we introduce ${\textbf{Vi}}$deo ${\textbf{P}}$erception Models for 3D $\textbf{Scene}$ synthesis ($\textbf{VIPScene}$), a novel framework that uses video generation models to leverage the encoded commonsense knowledge of the 3D physical world. VIPScene accepts both text and image prompts and seamlessly integrates video generation, feedforward 3D reconstruction, and open-vocabulary perception models to semantically and geometrically analyze each object in a scene. This enables flexible scene synthesis with high realism and structural consistency.We further introduce $\textbf{F}$irst-$\textbf{P}$erson $\textbf{V}$iew $\textbf{Score}$ $\textbf{(FPVScore)}$ for consistency and reality evaluation, utilizing continuous first-person perspective to capitalize on the reasoning ability of Multimodal Large Language Models.Extensive experiments show that VIPScene outperforms existing methods on indoor scene generation, and generalizes well to diverse room configurations. The code will be released.



Paperid:400
Authors:Jacob Block, Aryan Mokhtari, Sanjay Shakkottai
Abstract:
Machine unlearning algorithms aim to remove the influence of specific training samples, ideally recovering the model that would have resulted from training on the remaining data alone. We study unlearning in the overparameterized setting, where many models interpolate the data, and defining the solution as any loss minimizer over the retained set—as in prior work in the underparameterized setting—is inadequate, since the original model may already interpolate the retained data and satisfy this condition. In this regime, loss gradients vanish, rendering prior methods based on gradient perturbations ineffective, motivating both new unlearning definitions and algorithms. For this setting, we define the unlearning solution as the minimum-complexity interpolator over the retained data and propose a new algorithmic framework that only requires access to model gradients on the retained set at the original solution. We minimize a regularized objective over perturbations constrained to be orthogonal to these model gradients, a first-order relaxation of the interpolation condition. For different model classes, we provide exact and approximate unlearning guarantees, and we demonstrate that an implementation of our framework outperforms existing baselines across various unlearning experiments.



Paperid:401
Authors:Steven Cao, Gregory Valiant, Percy Liang
Abstract:
We study the problem of entropy calibration, which asks whether a language model's entropy over generations matches its log loss on human text. Past work found that models are miscalibrated, with entropy per step increasing as generations grow longer. This entropy growth is accompanied by a degradation in text quality, and fixing error accumulation has been the subject of many papers, which propose distribution truncation methods to trade diversity for quality. In this paper, we ask: is miscalibration likely to improve with scale, and is it theoretically possible to calibrate without tradeoffs? To build intuition, we first study a simplified theoretical setting to characterize the scaling behavior of miscalibration with respect to dataset size. We find that the scaling behavior depends on the power law exponent of the data distribution — in particular, for a power law exponent close to 1, the scaling exponent is close to 0, meaning that miscalibration improves very slowly with scale. Next, we measure miscalibration empirically in language models ranging from 0.5B to 70B parameters. We find that the observed scaling behavior is similar to what is predicted by the simplified setting: our fitted scaling exponents for text are close to 0, meaning that larger models accumulate error at a similar rate as smaller ones. This scaling (or, lack thereof) provides one explanation of why we sample from larger models with similar amounts of truncation as smaller models, even though the larger models are of higher quality. However, truncation is not a satisfying solution because it comes at the cost of increased log loss. In theory, is it even possible to reduce entropy while preserving log loss? We prove that it is possible, if we assume access to a black box which can fit models on the future entropy of text prefixes and attain low test error.



Paperid:402
Authors:Ziyad Benomar, Romain Cosson, Alexander Lindermayr, Jens Schlöter
Abstract:
In non-clairvoyant scheduling, the goal is to minimize the total job completion time without prior knowledge of individual job processing times. This classical online optimization problemhas recently gained attention through the framework of learning-augmented algorithms. We introduce a natural setting in which the scheduler receives continuous feedback in the form of progress bars—estimates of the fraction of each job completed over time.We design new algorithms for both adversarial and stochastic progress bars and prove strong competitive bounds. Our results in the adversarial case surprisingly induce improved guarantees for learning-augmented scheduling with job size predictions.We also introduce a general method for combining scheduling algorithms, yielding further insights in scheduling with predictions. Finally, we propose a stochastic model of progress bars as a more optimistic alternative to conventional worst-casemodels, and present an asymptotically optimal scheduling algorithm in this setting.



Paperid:403
Authors:Ali Gorji, Andisheh Amrollahi, Andreas Krause
Abstract:
SHAP (SHapley Additive exPlanations) values are a widely used method for local feature attribution in interpretable and explainable AI. We propose an efficient two-stage algorithm for computing SHAP values in both black-box setting and tree-based models. Motivated by spectral bias in real-world predictors, we first approximate models using compact Fourier representations, exactly for trees and approximately for black-box models. In the second stage, we introduce a closed-form formula for {\em exactly} computing SHAP values using the Fourier representation, that ``linearizes'' the computation into a simple summation and is amenable to parallelization. As the Fourier approximation is computed only once, our method enables amortized SHAP value computation, achieving significant speedups over existing methods and a tunable trade-off between efficiency and precision.



Paperid:404
Authors:Li-Chung Lin, Yaxu Liu, Chih-Jen Lin
Abstract:
Models requiring probabilistic outputs are ubiquitous and used in fields such as natural language processing, contrastive learning, and recommendation systems. The standard method of designing such a model is to output unconstrained logits, which are normalized into probabilities with the softmax function. The normalization involves computing a summation across all classes, which becomes prohibitively expensive for problems with a large number of classes. An important strategy to reduce the cost is to sum over a sampled subset of classes in the softmax function, known as sampled softmax. It was known that the sampled softmax is biased; the expectation taken over the sampled classes is not equal to the softmax function. Many works focused on reducing the bias by using a better way of sampling the subset. However, while sampled softmax is biased, it is unclear whether an unbiased function different from sampled softmax exists. In this paper, we show that all functions that only access a sampled subset of classes must be biased.With this result, we prevent efforts in finding unbiased loss functions and validate that past efforts devoted to reducing bias are the best we can do.



Paperid:405
Authors:Antonio Ferrara, Francesco Cozzi, Alan Perotti, André Panisson, Francesco Bonchi
Abstract:
Abstract:Determining whether an algorithmic decision-making system discriminates against a specific demographic typically involves comparing a single point estimate of a fairness metric against a predefined threshold. This practice is statistically brittle: it ignores sampling error and treats small demographic subgroups the same as large ones. The problem intensifies in intersectional analyses, where multiple sensitive attributes are considered jointly, giving rise to a larger number of smaller groups. As these groups become more granular, the data representing them becomes too sparse for reliable estimation, and fairness metrics yield excessively wide confidence intervals, precluding meaningful conclusions about potential unfair treatments.In this paper, we introduce a unified, size-adaptive, hypothesis‑testing framework that turns fairness assessment into an evidence‑based statistical decision.Our contribution is twofold. (i) For sufficiently large subgroups, we prove a Central‑Limit result for the statistical parity difference, leading to analytic confidence intervals and a Wald test whose type‑I (false positive) error is guaranteed at level $\alpha$. (ii) For the long tail of small intersectional groups, we derive a fully Bayesian Dirichlet–multinomial estimator; Monte-Carlo credible intervals are calibrated for any sample size and naturally converge to Wald intervals as more data becomes available. We validate our approach empirically on benchmark datasets, demonstrating how our tests provide interpretable, statistically rigorous decisions under varying degrees of data availability and intersectionality.



Paperid:406
Authors:Alper Ahmetoglu, Steven James, Cameron Allen, Sam Lobel, David Abel, George Konidaris
Abstract:
We consider how to construct state abstractions compatible with a given set of abstract actions, to obtain a well-formed abstract Markov decision process (MDP). We show that the Bellman equation suggests that abstract states should represent distributions over states in the ground MDP; we characterize the conditions under which the resulting process is Markov and approximately model-preserving, derive algorithms for constructing and planning with the abstract MDP, and apply them to a visual maze task. We generalize these results to the factored actions case, characterizing the conditions that result in factored abstract states and apply the resulting algorithm to Montezuma's Revenge. These results provide a powerful and principled framework for constructing neurosymbolic abstract Markov decision processes.



Paperid:407
Authors:Gramoz Goranci, Shaofeng Jiang, Peter Kiss, Eva Szilagyi, Qiaoyuan Yang
Abstract:
Abstract:We study the problem of computing Chamfer distance in the fully dynamic setting, where two set of points $A, B \subset \mathbb{R}^{d}$, each of size up to $n$, dynamically evolve through point insertions or deletions and the goal is to efficiently maintain an approximation to $dist_{\mathrm{CH}}(A,B) = \sum_{a \in A} \min_{b \in B} dist(a,b)$, where $dist$ is a distance measure. Chamfer distance is a widely used dissimilarity metric for point clouds, with many practical applications that require repeated evaluation on dynamically changing datasets, e.g., when used as a loss function in machine learning. In this paper, we present the first dynamic algorithm for maintaining an approximation of the Chamfer distance under the $\ell_p$ norm for $p \in$ {$1,2$}.Our algorithm reduces to approximate nearest neighbor (ANN) search with little overhead. Plugging in standard ANN bounds, we obtain $(1+\epsilon)$-approximation in $\tilde{O}(\epsilon^{-d})$ update time and $O(1/\epsilon)$-approximation in $\tilde{O}(d n^{\epsilon^2} \epsilon^{-4})$ update time.We evaluate our method on real-world datasets and demonstrate that it performs competitively against natural baselines.



Paperid:408
Authors:Haoyu Wang, Zeyu Qin, Yifei Zhao, Chao Du, Min Lin, Xueqian Wang, Tianyu Pang
Abstract:
LLMs have made impressive progress, but their growing capabilities also expose them to highly flexible jailbreaking attacks designed to bypass safety alignment. While many existing defenses focus on known types of attacks, it is more critical to prepare LLMs forunseenattacks that may arise during deployment. To address this, we propose alifelong safety alignmentframework that enables LLMs to continuously adapt to new and evolving jailbreaking strategies. Our framework introduces a competitive setup between two components: aMeta-Attacker, trained to actively discover novel jailbreaking strategies, and aDefender, trained to resist them. To effectively warm up the Meta-Attacker, we first leverage the GPT-4o API to extract key insights from a large collection of jailbreak-related research papers. Through iterative training, the first iteration Meta-Attacker achieves a 73% attack success rate (ASR) on RR and a 57% transfer ASR on LAT using onlysingle-turnattacks. Meanwhile, the Defender progressively improves its robustness and ultimately reduces the Meta-Attacker's success rate to just 7%, enabling safer and more reliable deployment of LLMs in open-ended environments.



Paperid:409
Authors:Silin Cheng, Kai Han
Abstract:
Vision-language models (VLMs), such as CLIP, have shown strong generalization under zero-shot settings, yet adapting them to downstream tasks with limited supervision remains a significant challenge. Existing multi-modal prompt learning methods typically rely on fixed, shared prompts and deterministic parameters, which limits their ability to capture instance-level variation or model uncertainty across diverse tasks and domains. We propose VaMP, a variational framework for prompt adaptation that enables sample-specific, uncertainty-aware tuning in multi-modal representation learning. VaMP generates instance-conditioned prompts by sampling from a learned posterior distribution, allowing the model to personalize its behavior based on input content. To guide this process with both local and global semantics, we introduce a class-aware prior constructed from the instance representation and class prototype. We formulate prompt tuning as variational inference over latent prompt representations and train the entire framework end-to-end via reparameterized sampling. Experiments on few-shot and domain generalization benchmarks show that VaMP achieves state-of-the-art performance among prompt tuning methods, highlighting the benefits of modeling both uncertainty and task structure in multi-modal prompt adaptation.



Paperid:410
Authors:Matias Romero, Frederik Mallmann-Trenn, Jose Correa
Abstract:
Abstract:A classic statistical problem is to study the asymptotic behavior of the order statistics of a large number of independent samples taken from a distribution with finite expectation. This behavior has implications for several core problems in machine learning and economics—including robust learning under adversarial noise, best-arm identification in bandit algorithms, revenue estimation in second-price auctions, and the analysis of tail-sensitive statistics used in out-of-distribution detection.The research question we tackle in this paper is: How large can the expectation of the $\ell$-th maximum of the $n$ samples be? For $\ell=1$, i.e., the maximum, this expectation is known to grow as $o(n)$, which can be shown to be tight. We show that there is a sharp contrast when considering any fixed $\ell>1$. Surprisingly, in this case, the largest possible growth rate for all fixed $\ell>1$ is $O(\frac{n}{\log(n)\log\log(n)})$ and $\Omega(\frac{n}{\log(n)(\log\log(n))^{1.01}})$. Our result is actually finer than the latter and provides a sharp characterization of the largest achievable growth rate for the expectation of the $\ell$-th maximum of $n$ i.i.d. samples. Beyond the theoretical analysis, we support our findings with extensive simulations. These empirical results highlight a notable phenomenon: although the multiplicative gap between the maximum and the second maximum grows quickly with $n$, the ratio remains approximately constant in 99\% of trials. This suggests that while worst-case growth is sharp and meaningful, typical-case behavior may be significantly more stable.



Paperid:411
Authors:Yumin Choi, Jinheon Baek, Sung Ju Hwang
Abstract:
Large Language Models (LLMs) have shown remarkable capabilities, with optimizing their input prompts playing a pivotal role in maximizing their performance. However, while LLM prompts consist of both the task-agnostic system prompts and task-specific user prompts, existing work on prompt optimization has focused on user prompts specific to individual queries or tasks, and largely overlooked the system prompt that is, once optimized, applicable across different tasks and domains. Motivated by this, we introduce the novel problem of bilevel system prompt optimization, whose objective is to design system prompts that are robust to diverse user prompts and transferable to unseen tasks. To tackle this problem, we then propose a meta-learning framework, which meta-learns the system prompt by optimizing it over various user prompts across multiple datasets, while simultaneously updating the user prompts in an iterative manner to ensure synergy between them. We conduct experiments on 14 unseen datasets spanning 5 different domains, on which we show that our approach produces system prompts that generalize effectively to diverse user prompts. Also, our findings reveal that the optimized system prompt enables rapid adaptation even to unseen tasks, requiring fewer optimization steps for test-time user prompts while achieving improved performance.



Paperid:412
Authors:Kiril Bangachev, Guy Bresler, Iliyas Noman, Yury Polyanskiy
Abstract:
Abstract:The meta-task of synchronizing representations is steadily gaining importance. Common representations to be synchronized include image and text data, student and teacher models, or multiple augmentations of the same image. In this paper we theoretically explain the advantages of synchronizing with trainable inverse temperature and bias under the sigmoid loss, as implemented in the recent SigLIP models of Google DeepMind. Temperature and bias can drive the loss function to zero for a rich class of configurations that we call $(\mathsf{m}, \mathsf{rb})$-Constellations. $(\mathsf{m}, \mathsf{rb})$-Constellations are a novel combinatorial object related to spherical codes and are parametrized by a margin $\mathsf{m}$ and relative bias $\mathsf{rb}.$ We use our characterization of constellations to theoretically justify the success of SigLIP on retrieval,to explainthe modality gap present in SigLIP,and to identify the necessary dimension for producing high-quality representations. We also propose a re-parameterization of the sigmoid loss with explicit relative bias. This parameterization yields several practical recommendations, including in the contexts of synchronizing with a locked encoder and synchronizing more than two modalities.



Paperid:413
Authors:Anurag Arnab, Ahmet Iscen, Mathilde Caron, Alireza Fathi, Cordelia Schmid
Abstract:
Despite recent advances in Vision-Language Models (VLMs), long-video understanding remains a challenging problem. Although state-of-the-art long-context VLMs can process around 1000 input frames, they still struggle to effectively leverage this sequence length, and succumb to irrelevant distractors within the context window. We present Dynamic Context Aggregation, an inference strategy for video question-answering that curates the model's input context. We use the VLM itself to iteratively identify and extract the most relevant frames from the video, which are then used for answering. We demonstrate how leveraging more computation at inference-time to select the most relevant context leads to improvements in accuracy, in agreement with recent work on inference-time scaling of LLMs. Moreover, we achieve state-of-the-art results on 4 diverse video question-answering datasets, showing consistent improvements with 3 different VLMs. In particular, our method shines on longer videos which would not otherwise fit in the model's context window: On longer videos of more than 1 hour on LVBench, our approach using a context window of 32K outperforms the same VLM using standard inference with a 700K context window by 2.8 points.



Paperid:414
Authors:David Zoltowski, Skyler Wu, Xavier Gonzalez, Leo Kozachkov, Scott Linderman
Abstract:
Markov chain Monte Carlo (MCMC) methods are foundational algorithms for Bayesian inference and probabilistic modeling. However, most MCMC algorithms are inherently sequential and incur computational costs that scale linearly with sample size. Previous work on adapting MCMC to modern hardware has therefore focused on running many independent chains in parallel. Here, we take an alternative approach: we propose algorithms to evaluate MCMC samplersin parallel across the chain length.We build on recent methods for parallel evaluation of nonlinear recursions that formulate the state sequence as a solution to a fixed-point problem and solve for the fixed-point using a parallel form of Newton's method. We show how the widely-used Gibbs sampling, Metropolis-adjusted Langevin, and Hamiltonian Monte Carlo algorithms can be evaluated in parallel using this approach. We showcase several examples in which we can generate hundreds of thousands of MCMC samples with only a few dozen parallel Newton iterations. Additionally, we develop two new quasi-Newton methods for the evaluation of nonlinear recursions that lower memory costs and reduce runtime. Across multiple challenging benchmarks, we find that the proposed parallel algorithms can accelerate MCMC chains by orders of magnitude compared to sequential evaluation.



Paperid:415
Authors:Manolis Zampetakis, Felix Zhou
Abstract:
We introduce a novel framework for differentially private (DP) statistical estimation via data truncation, addressing a key challenge in DP estimation when the data support is unbounded. Traditional approaches rely on problem-specific sensitivity analysis, limiting their applicability. By leveraging techniques from truncated statistics, we develop computationally efficient DP estimators for exponential family distributions, including Gaussian mean and covariance estimation, achieving near-optimal sample complexity. Previous works on exponential families only consider bounded or one-dimensional families. Our approach mitigates sensitivity through truncation while carefully correcting for the introduced bias using maximum likelihood estimation and DP stochastic gradient descent. Along the way, we establish improved uniform convergence guarantees for the log-likelihood function of exponential families, which may be of independent interest. Our results provide a general blueprint for DP algorithm design via truncated statistics.



Paperid:416
Authors:Liyao Li, Chao Ye, Wentao Ye, Yifei Sun, Zhe Jiang, Haobo Wang, Jiaming Tian, Yiming Zhang, NINGTAO WANG, Xing Fu, Gang Chen, Junbo Zhao
Abstract:
To migrate the remarkable successes of Large Language Models (LLMs), the community has made numerous efforts to generalize them to the table reasoning tasks for the widely deployed tabular data. Despite that, in this work, by showing a probing experiment on our proposed StructQA benchmark, we postulate that even the most advanced LLMs (such as GPTs) may still fall short of coping with tabular data. More specifically, the current scheme often simply relies on serializing the tabular data, together with the meta information, then inputting them through the LLMs. We argue that the loss of structural information is the root of this shortcoming. In this work, we further propose TAMO, which bears an ideology to treat the tables as an independent modality integrated with the text tokens. The resulting model in TAMO is a multimodal framework consisting of a hypergraph neural network as the global table encoder seamlessly integrated with the mainstream LLM. Empirical results on various benchmarking datasets, including HiTab, WikiTQ, WikiSQL, FeTaQA, and StructQA, have demonstrated significant improvements on generalization with an average relative gain of42.65%.



Paperid:417
Authors:mohit sharma, Amit Deshpande, Chiranjib Bhattacharyya, Rajiv Ratn Shah
Abstract:
To fix the `bias in, bias out' problem in fair machine learning, it is important to steer feature distributions of data or internal representations of Large Language Models (LLMs) to \emph{ideal} ones that guarantee group-fair outcomes. Previous work on fair generative models and representation steering could greatly benefit from provable fairness guarantees on the model output. We define a distribution as \emph{ideal} if the minimizer of any cost-sensitive risk on it is guaranteed to have exact group-fair outcomes (e.g., demographic parity, equal opportunity)---in other words, it has no fairness-utility trade-off. We formulate an optimization program for optimal steering by finding the nearest \emph{ideal} distribution in KL-divergence, and provide efficient algorithms for it when the underlying distributions come from well-known parametric families (e.g., normal, log-normal).Empirically, our optimal steering techniques on both synthetic and real-world datasets improve fairness without diminishing utility (and sometimes even improve utility). We demonstrate affine steering of LLM representations to reduce bias in multi-class classification, e.g., occupation prediction from a short biography in Bios dataset (De-Arteaga et al.). Furthermore, we steer internal representations of LLMs towards desired outputs so that it works equally well across different groups.



Paperid:418
Authors:Ming Tang, Yuxuan Zhou, Chao Huang
Abstract:
In dueling bandits, an agent explores and exploits choices (i.e., arms) by learning from their stochastic feedback in the form of relative preferences. Prior related studies focused on unbiased feedback. In practice, however, the feedback provided by evaluators can be biased. For example, human users are likely to provide biased evaluation towards large language models due to their heterogeneous background. In this work, we aim to minimize the regret in dueling bandits considering evaluators’ biased feedback. We begin with a benchmark case where evaluators’ bias information is known. Solving the known-bias case is nontrivial, because the bias cannot be easily decoupled from the feedback. We overcome this challenge and propose an unbiased arm performance estimator and a bias-sensitive dueling bandits algorithm. We manage to analyze the regret, dealing with the complex form of the estimator, and show that the feedback either matching or opposing the ground-truth reduces the regret. Then, we study the case where evaluators’ bias information is unknown. The associated estimator can hardly be solved in closed-form due to the non-convexity of the estimator solving problem. We address this challenge and propose an extended bias-sensitive algorithm by incorporating block coordinate descent. This algorithm is proven to achieve the same order of regret (as in the known bias case) with a bounded error. Experiments show that when compared with baselines, our algorithms reduces the regret by up to 86.9%.



Paperid:419
Authors:Eduardo Pignatelli, Jarek Liesen, Robert Lange, Chris Lu, Pablo Samuel Castro, Laura Toni
Abstract:
Abstract:As Deep Reinforcement Learning (Deep RL) research moves towards solving large-scale worlds, efficient environment simulations become crucial for rapid experimentation. However, most existing environments struggle to scale to high throughput, setting back meaningful progress. Interactions are typically computed on the CPU, limiting training speed and throughput, due to slower computation and communication overhead when distributing the task across multiple machines. Ultimately, Deep RL training is CPU-bound, and developing batched, fast, and scalable environments has become a frontier for progress. Among the most used Reinforcement Learning (RL) environments, MiniGrid is at the foundation of several studies on exploration, curriculum learning, representation learning, diversity, meta-learning, credit assignment, and language-conditioned RL, and still suffers from the limitations described above. In this work, we introduce NAVIX, a re-implementation of MiniGrid in JAX. NAVIX achieves over $160\,000\times$ speed improvements in batch mode, supporting up to 2048 agents in parallel on a single Nvidia A100 80 GB. This reduces experiment times from one week to 15 minutes, promoting faster design iterations and more scalable RL model development.



Paperid:420
Authors:Hongjian Wang, Aaditya Ramdas
Abstract:
Abstract:We present two sharp, closed-form empirical Bernstein inequalities for symmetric random matrices with bounded eigenvalues. By sharp, we mean that both inequalities adapt to the unknown variance in a tight manner: the deviation captured by the first-order $1/\sqrt{n}$ term asymptotically matches the matrix Bernstein inequality exactly, including constants, the latter requiring knowledge of the variance. Our first inequality holds for the sample mean of independent matrices, and our second inequality holds for a mean estimator under martingale dependence at stopping times.



Paperid:421
Authors:Zhanhui Zhou, Lingjie Chen, Chao Yang, Chaochao Lu
Abstract:
Abstract:One way to mitigate risks in vision-language models (VLMs) is to censor dangerous samples in their training data.However, data moderation can be easily bypassed when harmful images are split into small, benign-looking patches, scattered across many training samples. VLMs may then learn to piece these fragments together and generate harmful responses at inference, either from full images or text references.For instance, if trained on image patches from a bloody scene paired with the descriptions "safe," VLMs may later describe, the full image or a text reference to the scene, as "safe."We define the core ability of VLMs enabling this attack as $\textit{visual stitching}$—the ability to integrate visual information spread across multiple training samples that share the same textual descriptions.In our work, we first demonstrate visual stitching abilities in most open-source VLMs on three datasets where each image is labeled with a unique synthetic ID. We split each $(\texttt{image}, \texttt{ID})$ pair into $\{(\texttt{patch}, \texttt{ID})\}$ pairs at different granularity for finetuning, and we find that models can verbalize the correct IDs from full images or text reference.Building on this, we simulate the adversarial data poisoning scenario mentioned above by replacing IDs with text descriptions like "safe" or "unsafe" and using patches from dangerous images, demonstrating how harmful content can evade moderation in patches and later be reconstructed through visual stitching, posing serious VLM safety risks.



Paperid:422
Authors:Honglin Lin, Qizhi Pei, Zhuoshi Pan, Yu Li, Xin Gao, Juntao Li, Lijun Wu, Conghui He
Abstract:
Reasoning capability is pivotal for large language models (LLMs) to solve complex tasks, yet achieving reliable and scalable reasoning remains challenging. While chain-of-thought (CoT) prompting has become a mainstream approach, existing methods often suffer from uncontrolled generation, insufficient quality, and limited diversity in reasoning paths. Recent efforts leverage code to enhance CoT by grounding reasoning in executable steps, but such methods typically rely on static code templates or human-curated rules, hindering scalability and generalizability. In this work, we propose Caco (Code-Assisted Chain-of-ThOught), a novel framework that automates the generation of high-quality, verifiable, and diverse reasoning data through code-driven augmentation. Unlike prior work, Caco first fine-tunes a code-based CoT generator on existing math and programming solutions in a unified code format, then scales the data generation to a large amount of diverse reasoning traces. Crucially, we introduce automated validation via code execution and rule-based filtering to ensure logical correctness and structural diversity, while reverse-engineering filtered outputs into natural language instructions to enrich task adaptability. This closed-loop process enables fully automated, scalable synthesis of reasoning data with guaranteed executability. Experiments on our created Caco-597 dataset demonstrate that Caco-trained models achieve state-of-the-art (SOTA) performance on mathematical reasoning benchmarks, outperforming existing strong baselines. Further analysis reveals that Caco’s code-anchored verification and instruction diversity contribute to superior generalization across unseen tasks. Our work establishes a paradigm for building self-sustaining, trustworthy reasoning systems without human intervention.



Paperid:423
Authors:Boaz Lavon, Shahar Katz, Lior Wolf
Abstract:
We present a novel approach to neural code generation that incorporates real-time execution signals into the language model generation process. While large language models (LLMs) have demonstrated impressive code generation capabilities, they typically do not utilize execution feedback during inference - a critical signal that human programmers regularly leverage. Our method, Execution-Guided Classifier-Free Guidance (EG-CFG), dynamically incorporates execution signals as the model generates code, providing line-by-line feedback that guides the generation process toward executable solutions. EG-CFG employs a multi-stage process: first, we conduct beam search to sample candidate program completions for each line; second, we extract execution signals by executing these candidates against test cases; and finally, we incorporate these signals into the prompt during generation. By maintaining consistent signals across tokens within the same line and refreshing signals at line boundaries, our approach provides coherent guidance while preserving syntactic structure. Our experiments across diverse coding tasks demonstrate that EG-CFG significantly improves code generation performance compared to standard approaches.



Paperid:424
Authors:Valentina Pyatkin, Saumya Malik, Victoria Graf, Hamish Ivison, Shengyi Huang, Pradeep Dasigi, Nathan Lambert, Hanna Hajishirzi
Abstract:
A crucial factor for successful human and AI interaction is the ability of language models or chatbots to follow human instructions precisely. A common feature of instructions are output constraints likeonly answer with yes or no" ormention the word `abrakadabra' at least 3 times".Even today's strongest models struggle with fulfilling such constraints. We find that most models strongly overfit on a small set of verifiable constraints from the benchmarks that test these abilities, called precise instruction following, and are not able to generalize well to unseen output constraints. We introduce a new benchmark, IFBench, to evaluate verifiable instruction following generalization on 58 new, diverse, and challenging verifiable out-of-domain constraints. In addition, we perform an extensive analysis of how and on what data models can be trained to improve precise instruction following generalization. Specifically, we carefully design constraint verification modules and show that reinforcement learning with verifiable rewards (RLVR) significantly improves instruction following. In addition to IFBench, we release 28 additional new hand-annotated training constraints and verification functions, RLVR training prompts, and code.



Paperid:425
Authors:Travis Dick, Haim Kaplan, Alex Kulesza, Uri Stemmer, Ziteng Sun, Ananda Theertha Suresh
Abstract:
Abstract:In the private set union problem each user owns a bag of at most $k$ items (from some large universe of items), and we are interested in computing the union of the items in the bags of all of the users. This is trivial without privacy, but a differentially private algorithm must be careful about reporting items contained in only a small number of bags. We consider differentially private algorithms that always report a subset of the union, and define the utility of an algorithm to be the expected size of the subset that it reports. Because the achievable utility varies significantly with the dataset, we introduce the *utility ratio*, which normalizes utility by a dataset-specific upper bound and characterizes a mechanism by its lowest normalized utility across all datasets. We then develop algorithms with guaranteed utility ratios and complement them with bounds on the best possible utility ratio. Prior work has shown that a single algorithm can be simultaneously optimal for all datasets when $k=1$, but we show that instance-optimal algorithms do not exist when $k>1$, and characterize how performance degrades as $k$ grows. At the same time, we design a private algorithm that achieves the maximum possible utility, regardless of $k$, when the item histogram matches a prior prediction (for instance, from a previous data release) and degrades gracefully with the $L_\infty$ distance between the prediction and the actual histogram when the prediction is imperfect.



Paperid:426
Authors:Luc Brogat-Motte, Alessandro Rudi, Riccardo Bonalli
Abstract:
We address the problem of safely learning controlled stochastic dynamics from discrete-time trajectory observations, ensuring system trajectories remain within predefined safe regions during both training and deployment. Safety-critical constraints of this kind are crucial in applications such as autonomous robotics, finance, and biomedicine. We introduce a method that ensures safe exploration and efficient estimation of system dynamics by iteratively expanding an initial known safe control set using kernel-based confidence bounds. After training, the learned model enables predictions of the system's dynamics and permits safety verification of any given control. Our approach requires only mild smoothness assumptions and access to an initial safe control set, enabling broad applicability to complex real-world systems. We provide theoretical guarantees for safety and derive adaptive learning rates that improve with increasing Sobolev regularity of the true dynamics. Experimental evaluations demonstrate the practical effectiveness of our method in terms of safety, estimation accuracy, and computational efficiency.



Paperid:427
Authors:Vivek Jayaram, Ira Kemelmacher-Shlizerman, Steve Seitz, John Thickstun
Abstract:
We introduce linearly constrained diffusion implicit models (CDIM), a method for solving noisy linear inverse problems 10-50x faster than existing diffusion-based methods. CDIM accelerates conditional inference through a novel projection algorithm that dynamically updates the number of projection steps and step size based on the expected denoising trajectory. This leads to far more efficient conditional update steps that match the efficiency of the unconditional steps from denoising diffusion implicit models (DDIM). Furthermore, for linear inverse problems without measurement noise, CDIM exactly satisfies the constraints, even when existing methods fail to do so. We demonstrate CDIM’s effectiveness across super-resolution, denoising, inpainting, deblurring, and 3D point cloud reprojection.



Paperid:428
Authors:Ameya Anjarlekar, S. Rasoul Etesami, R. Srikant
Abstract:
The continuous nature of belief states in POMDPs presents significant computational challenges in learning the optimal policy. In this paper, we consider an approach that solves a Partially Observable Reinforcement Learning (PORL) problem by approximating the corresponding POMDP model into a finite-state Markov Decision Process (MDP) (called Superstate MDP). We first derive theoretical guarantees that improve upon prior work that relate the optimal value function of the transformed Superstate MDP to the optimal value function of the original POMDP. Next, we propose a policy-based learning approach with linear function approximation to learn the optimal policy for the Superstate MDP. Consequently, our approach shows that a POMDP can be approximately solved using TD-learning followed by Policy Optimization by treating it as an MDP, where the MDP state corresponds to a finite history. We show that the approximation error decreases exponentially with the length of this history. To the best of our knowledge, our finite-time bounds are the first to explicitly quantify the error introduced when applying standard TD learning to a setting where the true dynamics are not Markovian.



Paperid:429
Authors:Di Wu, Chengshuai Shi, Jing Yang, Cong Shen
Abstract:
Reinforcement Learning from Human Feedback (RLHF) has emerged as a key technique for post‑training large language models. Despite its empirical success, the theoretical understanding of RLHF is still limited, as learning the KL-regularized target with only preference feedback poses additional challenges compared with canonical RL. Existing works mostly study the reward-based Bradley-Terry (BT) preference model, and extend classical designs utilizing optimism or pessimism. This work, instead, considers the general preference model (whose practical relevance has been observed recently) and obtains performance guarantees with major, order-wise improvements over existing ones. Surprisingly, these results are derived from algorithms that directly use empirical estimates (i.e., greedy sampling), as opposed to constructing optimistic or pessimistic estimates in previous works. This insight has a deep root in the unique structural property of the optimal policy class under the KL-regularized target, and we further specialize it to the BT model, highlighting the surprising sufficiency of greedy sampling in RLHF.



Paperid:430
Authors:Yunxiao Shi, Yinhao Zhu, Hong Cai, Shizhong Han, Jisoo Jeong, Amin Ansari, Fatih Porikli
Abstract:
3D occupancy provides fine-grained 3D geometry and semantics for scene understanding which is critical for autonomous driving. Most existing methods, however, carry high compute costs, requiring dense 3D feature volume and cross-attention to effectively aggregate information. More recent works have adopted Bird's Eye View (BEV) or sparse points as scene representation with much reduced cost, but still suffer from their respective shortcomings. More concretely, BEV struggles with small objects that often experience significant information loss after being projected to the ground plane. On the other hand, points can flexibly model little objects in 3D, but is inefficient at capturing flat surfaces or large objects. To address these challenges, in this paper, we present a novel 3D occupancy prediction approach, ODG, which combines BEV and sparse points based representations. We propose a dual-branch design: a query-based sparse points branch and a BEV branch. The 3D information learned in the sparse points branch is shared with the BEV stream via cross-attention, which enriches the weakened signals of difficult objects on the BEV plane. The outputs of both branches are finally fused to generate predicted 3D occupancy. We conduct extensive experiments on the Occ3D-nuScenes and Occ3D-Waymo benchmarks that demonstrate the superiority of our proposed ODG. Moreover, ODG also delivers competitive inference speed when compared to the latest efficient approaches.



Paperid:431
Authors:Shuo Sun, Yimin Zhao, Christina Lee, JIAWEI SUN, Chengran Yuan, Zefan Huang, Dongen Li, Justin Yeoh, Alok Prakash, Thomas Malone, Marcelo Ang Jr
Abstract:
As the field progresses toward Artificial General Intelligence (AGI), there is a pressing need for more comprehensive and insightful evaluation frameworks that go beyond aggregate performance metrics. This paper introduces a unified rating system that jointly models the difficulty of individual test cases and the competency of AI models (or humans) across vision, language, and action domains. Unlike existing metrics that focus solely on models, our approach allows for fine-grained, difficulty-aware evaluations through competitive interactions between models and tasks, capturing both the long-tail distribution of real-world challenges and the competency gap between current models and full task mastery. We validate the generalizability and robustness of our system through extensive experiments on multiple established datasets and models across distinct AGI domains. The resulting rating distributions offer novel perspectives and interpretable insights into task difficulty, model progression, and the outstanding challenges that remain on the path to achieving full AGI task mastery. We have made our code and results publicly available at https://ss47816.github.io/AGI-Elo/.



Paperid:432
Authors:Zhengkai Lin, Zhihang Fu, Ze Chen, Chao Chen, Liang Xie, Wenxiao Wang, Deng Cai, Zheng Wang, Jieping Ye
Abstract:
Human cognition is theorized to operate in two modes: fast, intuitive System 1 thinking and slow, deliberate System 2 thinking.While current Large Reasoning Models (LRMs) excel at System 2 thinking, their inability to perform fast thinking leads to high computational overhead and latency.In this work, we enable LRMs to approximate human intelligence through dynamic thinking speed adjustment, optimizing accuracy-efficiency trade-offs.Our approach addresses two key questions: (1) how to control thinking speed in LRMs, and (2) when to adjust it for optimal performance.For the first question, we identify the steering vector that governs slow-fast thinking transitions in LRMs' representation space.Using this vector, we achieve the first representation editing-based test-time scaling effect, outperforming existing prompt-based scaling methods.For the second question, we apply real-time difficulty estimation to signal reasoning segments of varying complexity.Combining these techniques, we propose the first reasoning strategy that enables fast processing of easy steps and deeper analysis for complex reasoning. Without any training or additional cost, our plug-and-play method yields an average +1.3\% accuracy with -8.6\% token usage across leading LRMs and advanced reasoning benchmarks.All of our algorithms are implemented based on vLLM and are expected to support broader applications and inspire future research.



Paperid:433
Authors:Yake Wei, Yu Miao, Dongzhan Zhou, Di Hu
Abstract:
In this paper, we reveal that most current efficient multimodal fine-tuning methods are hindered by a key limitation: they are directly borrowed from LLMs, often neglecting the intrinsic differences of multimodal scenarios and even affecting the full utilization of all modalities. Inspired by our empirical observation, we argue that unimodal adaptation and cross-modal adaptation are two essential parts for the effective fine-tuning of MLLMs. From this perspective, we propose Multimodal Low-rank Adaptation (MokA), a multimodal-aware efficient fine-tuning strategy that takes multimodal characteristics into consideration. It compresses unimodal information by modality-specific parameters while explicitly enhancing cross-modal interaction, ensuring both unimodal and cross-modal adaptation. Extensive experiments cover three representative multimodal scenarios (audio-visual-text, visual-text, and speech-text), and multiple LLM backbones (LLaMA2, Qwen2, Qwen2.5-VL, etc). Consistent improvements indicate the efficacy and versatility of the proposed method. Ablation studies and efficiency evaluation are also conducted to fully asses our method. Overall, we think MokA provides a more targeted solution for efficient adaptation of MLLMs, paving the way for further exploration.



Paperid:434
Authors:Kaituo Feng, Kaixiong Gong, Bohao Li, Zonghao Guo, Yibing Wang, Tianshuo Peng, Junfei Wu, Xiaoying Zhang, Benyou Wang, Xiangyu Yue
Abstract:
Inspired by DeepSeek-R1's success in eliciting reasoning abilities through rule-based reinforcement learning (RL), we introduce Video-R1 as the first attempt to systematically explore the R1 paradigm for incentivizing video reasoning within multimodal large language models (MLLMs). However, directly applying RL training with the GRPO algorithm to video reasoning presents two primary challenges: (i) a lack of temporal modeling for video reasoning, and (ii) the scarcity of high-quality video-reasoning data. To address these issues, we first propose the T-GRPO algorithm, which encourages models to utilize temporal information in videos for reasoning. Additionally, instead of relying solely on video data, we incorporate high-quality image-reasoning data into the training process. We have constructed two datasets: Video-R1-CoT-165k for SFT cold start and Video-R1-260k for RL training, both comprising image and video data. Experimental results demonstrate that Video-R1 achieves significant improvements on video reasoning benchmarks such as VideoMMMU and VSI-Bench, as well as on general video benchmarks including MVBench and TempCompass, etc. Notably, Video-R1-7B attains a 37.1\% accuracy on video spatial reasoning benchmark VSI-bench, surpassing the commercial proprietary model GPT-4o. All code, models, and data will be released.



Paperid:435
Authors:Xu Cao, Pranav Virupaksha, Sangmin Lee, Bolin Lai, Wenqi Jia, Jintai Chen, James Rehg
Abstract:
Humans have a remarkable ability to use co-speech deictic gestures, such as pointing and showing, to enrich verbal communication and support social interaction. These gestures are so fundamental that infants begin to use them even before they acquire spoken language, which highlights their central role in human communication. Understanding the intended targets of another individual's deictic gestures enables inference of their intentions, comprehension of their current actions, and prediction of upcoming behaviors. Despite its significance, gesture target estimation remains an underexplored task within the computer vision community. In this paper, we introduceGestureTarget, a novel task designed specifically for comprehensive evaluation of social deictic gesture semantic target estimation. To address this task, we proposeTransGesture, a set of Transformer-based gesture target prediction models. Given an input image and the spatial location of a person, our models predict the intended target of their gesture within the scene. Critically, our gaze-aware joint cross attention fusion model demonstrates how incorporating gaze-following cues significantly improves gesture target mask prediction IoU by 6% and gesture existence prediction accuracy by 10%. Our results underscore the complexity and importance of integrating gaze cues into deictic gesture intention understanding, advocating for increased research attention to this emerging area. All data, code will be made publicly available upon acceptance.



Paperid:436
Authors:Tianrong Chen, Huangjie Zheng, David Berthelot, Jiatao Gu, Joshua Susskind, Shuangfei Zhai
Abstract:
Diffusion models have demonstrated exceptional capabilities in generating high-fidelity images but typically suffer from inefficient sampling. Many solver designs and noise scheduling strategies have been proposed to dramatically improve sampling speeds. In this paper, we introduce a new sampling method that is up to 186\% faster than the current state of the art solver for comparative FID on ImageNet512. This new sampling method is training-free and uses an ordinary differential equation (ODE) solver.The key to our method resides in using higher-dimensional initial noise, allowing to produce more detailed samples with less function evaluations from existing pretrained diffusion models. In addition, by design our solver allows to control the level of detail through a simple hyper-parameter at no extra computational cost. We present how our approach leverages momentum dynamics by establishing a fundamental equivalence between momentum diffusion models and conventional diffusion models with respect to their training paradigms. Moreover, we observe the use of higher-dimensional noise naturally exhibits characteristics similar to stochastic differential equations (SDEs). Finally, we demonstrate strong performances on a set of representative pretrained diffusion models, including EDM, EDM2, and Stable-Diffusion 3, which cover models in both pixel and latent spaces, as well as class and text conditional settings.



Paperid:437
Authors:Harsha Vardhan simhadri, Martin Aumüller, Matthijs Douze, Dmitry Baranchuk, Amir Ingber, Edo Liberty, George Williams, Ben Landrum, Magdalen Manohar, Mazin Karjikar, Laxman Dhulipala, Meng Chen, Yue Chen, Rui Ma, Kai Zhang, Yuzheng Cai, Jiayang Shi, Weiguo Zheng, Yizhuo Chen, Jie Yin, Ben Huang
Abstract:
The 2023 Big ANN Challenge, held at NeurIPS 2023, focused on advancing the state-of-the-art in indexing data structures and search algorithms for practical variants of Approximate Nearest Neighbor (ANN) search that reflect its the growing complexity and diversity of workloads. Unlike prior challenges that emphasized scaling up classical ANN search (Simhadri et al., NeurIPS 2021), this competition addressed sparse, filtered, out-of-distribution, and streaming variants of ANNS. Participants developed and submitted innovative solutions that were evaluated on new standard datasets with constrained computational resources. The results showcased significant improvements in search accuracy and efficiency, with notable contributions from both academic and industrial teams. This paper summarizes the competition tracks, datasets, evaluation metrics, and the innovative approaches of the top-performing submissions, providing insights into the current advancements and future directions in the field of approximate nearest neighbor search.



Paperid:438
Authors:Ziliang Zhang, Tianming Zhao, Albert Zomaya
Abstract:
Online portfolio selection seeks to determine a sequence of allocations to maximize capital growth. Classical universal strategies asymptotically match the best constant-rebalanced portfolio but ignore potential forecasts, whereas heuristic methods often collapse when belief fails. We formalize this tension in a learning-augmented setting in which an investor observes (possibly erroneous) predictions prior to each decision moment, and we introduce the Rebalanced Arithmetic Mean portfolio with predictions (RAM). Under arbitrary return sequences, we prove that RAM captures at least a constant fraction of the hindsight-optimal wealth when forecasts are perfect while still exceeding the geometric mean of the sequence even when the predictions are adversarial. Comprehensive experiments on large-scale equity data strengthen our theory, spanning both synthetic prediction streams and production-grade machine-learning models. RAM advantages over universal-portfolio variants equipped with side information across various regimes. These results demonstrate that modest predictive power can be reliably converted into tangible gains without sacrificing worst-case guarantees.



Paperid:439
Authors:Tim G. Zhou, Evan Shelhamer, Geoff Pleiss
Abstract:
Abstract:The go-to strategy to apply deep networks in settings where uncertainty informs decisions—ensembling multiple training runs with random initializations—is ill-suited for the extremely large-scale models and practical fine-tuning workflows of today. We introduce a new cost-effective strategy for improving the uncertainty quantification and downstream decisions of a large model (e.g. a fine-tuned ViT-B): coupling it with a less accurate but much smaller "sidekick" (e.g. a fine-tuned ResNet-34) with a fraction of the computational cost. We propose aggregating the predictions of this \emph{Asymmetric Duo} by simple learned weighted averaging. Surprisingly, despite their inherent asymmetry, the sidekick model almost never harms the performance of the larger model. In fact, across five image classification benchmarks, and a variety of model architectures and training schemes (including soups), Asymmetric Duos significantly improve accuracy, uncertainty quantification, and selective classification metrics with only ${\approx}10-20$% more computation.



Paperid:440
Authors:Jiahao Wang, Hualian Sheng, Sijia Cai, Weizhan Zhang, Caixia Yan, Yachuang Feng, Bing Deng, Jieping Ye
Abstract:
Video diffusion models substantially boost the productivity of artistic workflows with high-quality portrait video generative capacity. However, prevailing pipelines are primarily constrained to single-shot creation, while real-world applications urge for multiple shots with identity consistency and flexible content controllability. In this work, we propose EchoShot, a native and scalable multi-shot framework for portrait customization built upon a foundation video diffusion model. To start with, we propose shot-aware position embedding mechanisms within video diffusion transformer architecture to model inter-shot variations and establish intricate correspondence between multi-shot visual content and their textual descriptions. This simple yet effective design enables direct training on multi-shot video data without introducing additional computational overhead. To facilitate model training within multi-shot scenario, we construct PortraitGala, a large-scale and high-fidelity human-centric video dataset featuring cross-shot identity consistency and fine-grained captions such as facial attributes, outfits, and dynamic motions. To further enhance applicability, we extend EchoShot to perform reference image-based personalized multi-shot generation and long video synthesis with infinite shot counts. Extensive evaluations demonstrate that EchoShot achieves superior identity consistency as well as attribute-level controllability in multi-shot portrait video generation. Notably, the proposed framework demonstrates potential as a foundational paradigm for general multi-shot video modeling. All the models and the dataset will be made open-source upon acceptance.



Paperid:441
Authors:Jisang Han, Honggyu An, Jaewoo Jung, Takuya Narihira, Junyoung Seo, Kazumi Fukuda, Chaehyun Kim, Sunghwan Hong, Yuki Mitsufuji, Seungryong Kim
Abstract:
Abstract:In this work, we address the task of 3D reconstruction in dynamic scenes, where object motions frequently degrade the quality of previous 3D pointmap regression methods, such as DUSt3R, that are originally designed for static 3D scene reconstruction. Although these methods provide an elegant and powerful solution in static settings, they struggle in the presence of dynamic motions that disrupt alignment based solely on camera poses. To overcome this, we propose D$^2$USt3R that directly regresses Static-Dynamic Aligned Pointmaps (SDAP) that simultaneiously capture both static and dynamic 3D scene geometry. By explicitly incorporating both spatial and temporal aspects, our approach successfully encapsulates 3D dense correspondence to the proposed pointmaps, enhancing downstream tasks. Extensive experimental evaluations demonstrate that our proposed approach consistently achieves superior 3D reconstruction performance across various datasets featuring complex motions.



Paperid:442
Authors:Jiayu Bai, Zhanbo Feng, Zhijie Deng, TianQi Hou, Robert Qiu, Zenan Ling
Abstract:
Abstract:Consistency Models (CMs) have shown promise for efficient one-step generation. However, most existing CMs rely on manually designed discretization schemes, which can cause repeated adjustments for different noise schedules and datasets. To address this, we propose a unified framework for the automatic and adaptive discretization of CMs, formulating it as an optimization problem with respect to the discretization step. Concretely, during the consistency training process, we propose using local consistency as the optimization objective to ensure trainability by avoiding excessive discretization, and taking global consistency as a constraint to ensure stability by controlling the accumulated error across discretization intervals. We establish the trade-off between local and global consistency with a Lagrange multiplier. Building on this framework, we achieve adaptive discretization for CMs using the Gauss-Newton method. We refer to our approach as ADCMs. Experiments demonstrate that ADCMs significantly improve the training efficiency of CMs, achieving superior generative performance with minimal training overhead on both CIFAR-10 and ImageNet $64 \times 64$. Moreover, ADCMs exhibit strong adaptability to more advanced DM variants.



Paperid:443
Authors:Jiaran Ye, Zijun Yao, Zhidian Huang, Liangming Pan, Jinxin Liu, Yushi Bai, Amy Xin, Liu Weichuan, Xiaoyin Che, Lei Hou, Juanzi Li
Abstract:
Recent work suggests that large language models (LLMs) can perform multi-hop reasoning implicitly---producing correct answers without explicitly verbalizing intermediate steps---but the underlying mechanisms remain poorly understood.In this paper, we study how such implicit reasoning emerges by training transformers from scratch in a controlled symbolic environment.Our analysis reveals a three-stage developmental trajectory: early memorization, followed by in-distribution generalization, and eventually cross-distribution generalization.We find that training with atomic triples is not necessary but accelerates learning, and that second-hop generalization relies on query-level exposure to specific compositional structures.To interpret these behaviors, we introduce two diagnostic tools: cross-query semantic patching, which identifies semantically reusable intermediate representations, and a cosine-based representational lens, which reveals that successful reasoning correlates with the cosine-base clustering in hidden space.This clustering phenomenon in turn provides a coherent explanation for the behavioral dynamics observed across training, linking representational structure to reasoning capability. These findings provide new insights into the interpretability of implicit multi-hop reasoning in LLMs, helping to clarify how complex reasoning processes unfold internally and offering pathways to enhance the transparency of such models.



Paperid:444
Authors:Zirui Yan, Emre Acartürk, Ali Tajer
Abstract:
Abstract:Causal representation learning (CRL) is the process of disentangling the **latent** low-dimensional causally-related generating factors underlying high-dimensional observable data. Extensive recent studies have characterized CRL identifiability and **perfect** recovery of the latent variables and their attendant causal graph. This paper introduces the notion of **reward-oriented** CRL, the purpose of which is moving away from perfectly learning the latent representation and instead learning it the extended needed for optimizing a desired downstream task (reward). In reward-oriented CRL, perfectly learning the latent representation can be excessive, and instead it has to be learned at the **coarsest** level that is sufficient for optimizing the desired task. Reward-oriented CRL is formalized as optimizing a desired function of the observable data over the space of all possible interventions and focuses on linear causal and transformation models. With the objective of sequentially identifying the optimal subset of interventions, an algorithm with adaptive exploration is designed that learns the causal graph $\mathcal{G}$ and variables needed to identify the best intervention. It is shown that for an $n$-dimensional latent space and a $d$-dimensional observation space, over a horizon $T$ the regret upper bound scales as $\tilde O(d^{\frac{1}{2}}n^{\frac{1}{3}}u^{\frac{2}{3}}T^{\frac{2}{3}} + u\sqrt{T})$. Furthermore, an almost-matching lower bound is shown to scale as $\Omega(d^{\frac{1}{2}}n^{\frac{1}{3}}p^{\frac{2}{3}}T^{\frac{2}{3}} + p\sqrt{T})$, where $u$ measures total uncertainty in the graph estimates and $p$ counts the number of causal paths in the graph.



Paperid:445
Authors:Puqian Wang, Nikos Zarifis, Ilias Diakonikolas, Jelena Diakonikolas
Abstract:
Abstract:We consider the basic problem of learning Single-Index Models with respect to the square loss under the Gaussian distribution in the presence of adversarial label noise. Our main contribution is the first computationally efficient algorithm for this learning task, achieving a constant factor approximation,that succeeds for the class of {\em all} monotone activations with bounded moment of order $2 + \zeta,$ for $\zeta > 0.$ This class in particular includes all monotone Lipschitz functions and even discontinuous functions like (possibly biased) halfspaces. Prior work for the case of unknown activation either does not attain constant factor approximation or succeeds for a substantially smaller family of activations. The main conceptual novelty of our approach lies in developing an optimization framework that steps outside the boundaries of usual gradient methods and instead identifies a useful vector field to guide the algorithm updates by directly leveraging the problem structure, properties of Gaussian spaces, and regularity of monotone functions.



Paperid:446
Authors:Kristjan Greenewald, Luis Lastras, Thomas Parnell, Vraj Shah, Lucian Popa, Giulio Zizzo, Chulaka Gunasekara, Ambrish Rawat, David Cox
Abstract:
Low-Rank Adaptation (LoRA) has emerged as a highly efficient framework for finetuning the weights of large foundation models, and has become the go-to method for data-driven customization of LLMs. Despite the promise of highly customized behaviors and capabilities, switching between relevant LoRAs in a multiturn setting is inefficient, as the key-value (KV) cache of the entire turn history must be recomputed with the LoRA weights before generation can begin. To address this problem, we propose Activated LoRA (aLoRA), an adapter architecture which modifies the LoRA framework to only adapt weights for the tokens in the sequence \emph{after} the aLoRA is invoked. This change crucially allows aLoRA to accept the base model's KV cache of the input string, meaning that aLoRA can be instantly activated whenever needed in a chain without recomputing the cache. This enables building what we call \emph{intrinsics}, i.e. specialized models invoked to perform well-defined operations on portions of an input chain or conversation that otherwise uses the base model by default. We train a set of aLoRA-based intrinsics models, demonstrating competitive accuracy with standard LoRA while achieving significant inference benefits. We include a codebase implementing aLoRA in the supplementary material.



Paperid:447
Authors:Yijun Quan, Zushu Li, Giovanni Montana
Abstract:
Growing data privacy demands, driven by regulations like GDPR and CCPA, require machine unlearning methods capable of swiftly removing the influence of specific training points. Although verified approaches like SISA, using data slicing and checkpointing, achieve efficient unlearning for single models by reverting to intermediate states, these methods struggle in teacher-student knowledge distillation settings. Unlearning in the teacher typically forces costly, complete student retraining due to pervasive information propagation during distillation. Our primary contribution is PURGE (Partitioned Unlearning with Retraining Guarantee for Ensembles), a novel framework integrating verified unlearning with distillation. We introduce constituent mapping and an incremental multi-teacher strategy that partitions the distillation process, confines each teacher constituent’s impact to distinct student data subsets, and crucially maintains data isolation. The PURGE framework substantially reduces retraining overhead—requiring only partial student updates—when teacher-side unlearning occurs. We provide both theoretical analysis, quantifying significant speed-ups in the unlearning process, and empirical validation on multiple datasets, demonstrating that PURGE achieves these efficiency gains while maintaining student accuracy comparable to standard baselines.



Paperid:448
Authors:Yufan Zhuang, Liyuan Liu, Chandan Singh, Jingbo Shang, Jianfeng Gao
Abstract:
In standard autoregressive generation, an LLM predicts the next-token distribution, samples a discrete token, and then discards the distribution, passing only the sampled token as new input. To preserve this distribution’s rich information, we propose Mixture of Inputs (MoI), a training-free method for autoregressive generation. After generating a token following the standard paradigm, we construct a new input that blends the generated discrete token with the previously discarded token distribution. Specifically, we employ a Bayesian estimation method that treats the token distribution as the prior, the sampled token as the observation, and replaces the conventional one-hot vector with the continuous posterior expectation as the new model input. MoI allows the model to maintain a richer internal representation throughout the generation process, resulting in improved text quality and reasoning capabilities. On mathematical reasoning, code generation, and PhD-level QA tasks, MoI consistently improves performance across multiple models including QwQ-32B, Nemotron-Super-49B, Gemma-3-27B, and DAPO-Qwen-32B, with no additional training and negligible computational overhead.



Paperid:449
Authors:Aniello Panariello, Daniel Marczak, Simone Magistri, Angelo Porrello, Bartłomiej Twardowski, Andrew Bagdanov, SIMONE CALDERARA, Joost van de Weijer
Abstract:
In this paper we address the challenges associated with merging low-rank adaptations of large neural networks. With the rise of parameter-efficient adaptation techniques, such as Low-Rank Adaptation (LoRA), model fine-tuning has become more accessible. While fine-tuning models with LoRA is highly efficient, existing merging methods often sacrifice this efficiency by merging fully-sized weight matrices. We propose the Core Space merging framework, which enables the merging of LoRA-adapted models within a common alignment basis to preserve the efficiency of low-rank adaptation and improve performance. We further provide a formal proof that projection into Core Space ensures no loss of information and provide a complexity analysis showing the efficiency gains. Extensive empirical results demonstrate that Core Space significantly improves existing merging techniques and achieves state-of-the-art results on both vision and language tasks while utilizing a fraction of the computational resources.



Paperid:450
Authors:Arman Zarei, Samyadeep Basu, Keivan Rezaei, Zihao Lin, Sayan Nag, Soheil Feizi
Abstract:
Abstract:Understanding how knowledge is distributed across the layers of generative models is crucial for improving interpretability, controllability, and adaptation. While prior work has explored knowledge localization in UNet-based architectures, Diffusion Transformer (DiT)-based models remain underexplored in this context. In this paper, we propose a model- and knowledge-agnostic method to localize where specific types of knowledge are encoded within the DiT blocks. We evaluate our method on state-of-the-art DiT-based models, including PixArt-$\alpha$, FLUX, and SANA, across six diverse knowledge categories. We show that the identified blocks are both interpretable and causally linked to the expression of knowledge in generated outputs.Building on these insights, we apply our localization framework to two key applications: *model personalization* and *knowledge unlearning*. In both settings, our localized fine-tuning approach enables efficient and targeted updates, reducing computational cost, improving task-specific performance, and better preserving general model behavior with minimal interference to unrelated or surrounding content.Overall, our findings offer new insights into the internal structure of DiTs and introduce a practical pathway for more interpretable, efficient, and controllable model editing.



Paperid:451
Authors:Hongcan Guo, Haolang Lu, Guoshun Nan, Bolun Chu, Jialin Zhuang, Yuan Yang, Wenhao Che, Xinye Cao, Sicong Leng, Qimei Cui, Xudong Jiang
Abstract:
Mixture-of-Experts (MoE) models enable efficient scaling of large language models (LLMs) by activating only a subset of experts per input. However, we observe that the commonly used auxiliary load balancing loss often leads to expert overlap and overly uniform routing, which hinders expert specialization and degrades overall performance during post-training.To address this, we propose a simple yet effective solution that introduces two complementary objectives: (1) an orthogonality loss to encourage experts to process distinct types of tokens, and (2) a variance loss to encourage more discriminative routing decisions.Gradient-level analysis demonstrates that these objectives are compatible with the existing auxiliary loss and contribute to optimizing the training process.Experimental results over various model architectures and across multiple benchmarks show that our method significantly enhances expert specialization. Notably, our method improves classic MoE baselines with auxiliary loss by up to 23.79\%, while also maintaining load balancing in downstream tasks, without any architectural modifications or additional components. We will release our code to contribute to the community.



Paperid:452
Authors:Ritika Jha, Aanisha Bhattacharyya, Yaman Singla, Rajiv Ratn Shah, Changyou Chen, Balaji Krishnamurthy
Abstract:
Recent advances in diffusion models have led to impressive improvements in image fidelity and diversity. However, aligning these models with nuanced human preferences such as aesthetic quality, engagement, and subjective user appeal remains a major challenge in the absence of large-scale human annotations. Collecting data that captures such preferences is both expensive and limited in diversity. To address this, we leverage the reasoning capabilities of Vision-language models (VLMs) and introduce a scalable, human-annotation-free training paradigm based on self-play. We present Self-Play Reward Optimization (SPRO), a unified, multimodal, multi-model framework that aligns both prompt generation and image synthesis with human preferences using synthetic reward signals. SPRO operates in three stages: SPRO-Prompt trains a Guider-VLM through self-play to generate diverse, high-reward prompts tailored to objectives such as PickScore (user preference), LAION-Aesthetics, and EngagetNet (engagement); SPRO-Image fine-tunes the diffusion model using high-reward images produced in the previous stage; and SPRO-Multimodal (SPRO-MM) combines both components, using optimized prompts and a fine-tuned generator for full-capacity alignment. Without relying on human-annotated data, SPRO achieves an average 30\% improvement across objectives and generalizes across both open-source and proprietary diffusion models. Additionally, SPRO uncovers novel prompting strategies rarely authored by humans, such as visual harmony for aesthetics and background or shadow-based cues for engagement. Across a range of diffusion pipelines, SPRO significantly outperforms prior methods and commercial baselines, demonstrating a scalable path toward aligning generative models with diverse human preferences.



Paperid:453
Authors:Gunnar König, Hidde Fokkema, Timo Freiesleben, Celestine Mendler-Dünner, Ulrike Luxburg
Abstract:
When applicants get rejected by a high-stakes algorithmic decision system, recourse explanations provide actionable suggestions for applicants on how to change their input features to get a positive evaluation. A crucial yet overlooked phenomenon is that recourse explanations areperformative: When many applicants act according to their recommendations, their collective behavior may shift the data distribution and, once the model is refitted, also the decision boundary. Consequently, the recourse algorithm may render its own recommendationsinvalid, such that applicants who make the effort of implementing their recommendations may be rejected again when they reapply.In this work, we formally characterize the conditions under which recourse explanations remain valid under their own performative effects. In particular, we prove that recourse actions may become invalid if they are influenced by or if they intervene on non-causal variables. Based on this analysis, we caution against the use of standard counterfactual explanations and causal recourse methods, and instead advocate for recourse methods that recommend actions exclusively on causal variables.



Paperid:454
Authors:Haowen Lai, Zitong Lan, Mingmin Zhao
Abstract:
Seeing hidden structures and objects around corners is critical for robots operating in complex, cluttered environments. We present HoloRadar, a practical system that reconstructs both line-of-sight (LOS) and non-line-of-sight (NLOS) 3D scenes using a single mmWave radar. HoloRadar uses a two-stage pipeline: the first stage generates high-resolution multi-return range images that capture both LOS and NLOS reflections, and the second stage reconstructs the physical scene by mapping mirrored observations to their true locations using a physics-guided architecture that models ray interactions.We deploy HoloRadar on a mobile robot and evaluate it across diverse real-world environments.Our evaluation results demonstrate accurate and robust reconstruction in both LOS and NLOS regions.



Paperid:455
Authors:JIAAN LUO, Feng Hong, Qiang Hu, Xiaofeng Cao, Feng Liu, Jiangchao Yao
Abstract:
Long-tailed recognition is ubiquitous and challenging in deep learning and even in the downstream finetuning of foundation models, since the skew class distribution generally prevents the model generalization to the tail classes. Despite the promise of previous methods from the perspectives of data augmentation, loss rebalancing and decoupled training etc., consistent improvement in the broad scenarios like multi-label long-tailed recognition is difficult. In this study, we dive into the essential model capacity impact under long-tailed context, and propose a novel framework, Model Rebalancing (MORE), which mitigates imbalance by directly rebalancing the model's parameter space. Specifically, MORE introduces a low-rank parameter component to mediate the parameter space allocation guided by a tailored loss and sinusoidal reweighting schedule, but without increasing the overall model complexity or inference costs. Extensive experiments on diverse long-tailed benchmarks, spanning multi-class and multi-label tasks, demonstrate that MORE significantly improves generalization, particularly for tail classes, and effectively complements existing imbalance mitigation methods. These results highlight MORE's potential as a robust plug-and-play module in long-tailed settings.



Paperid:456
Authors:Sitong Chen, Shen Nie, Jiacheng Sun, Zijin Feng, Zhenguo Li, Ji-Rong Wen, Chongxuan LI
Abstract:
We present a systematic theoretical framework that interprets masked diffusion models (MDMs) as solutions to energy minimization problems in discrete optimal transport. Specifically, we prove that three distinct energy formulations—kinetic, conditional kinetic, and geodesic energy—are mathematically equivalent under the structure of MDMs, and that MDMs minimize all three when the mask schedule satisfies a closed-form optimality condition. This unification not only clarifies the theoretical foundations of MDMs, but also motivates practical improvements in sampling. By parameterizing interpolation schedules via Beta distributions, we reduce the schedule design space to a tractable 2D search, enabling efficient post-training tuning without model modification. Experiments on synthetic and real-world benchmarks demonstrate that our energy-inspired schedules outperform hand-crafted baselines, particularly in low-step sampling settings.



Paperid:457
Authors:Yutong Feng, Linlin Zhang, Hengyuan Cao, Yiming Chen, Xiaoduan Feng, Jian Cao, Yuxiong Wu, Bin Wang
Abstract:
Virtual Try-ON (VTON) is a practical and widely-applied task, for which most of existing works focus on clothes. This paper presents OmniTry, a unified framework that extends VTON beyond garment to encompass any wearable objects, e.g., jewelries and accessories, with mask-free setting for more practical application. When extending to various types of objects, data curation is challenging for obtaining paired images, i.e., the object image and the corresponding try-on result. To tackle this problem, we propose a two-staged pipeline: For the first stage, we leverage large-scale unpaired images, i.e., portraits with any wearable items, to train the model for mask-free localization. Specifically, we repurpose the inpainting model to automatically draw objects in suitable positions given an empty mask. For the second stage, the model is further fine-tuned with paired images to transfer the consistency of object appearance. We observed that the model after the first stage shows quick convergence even with few paired samples. OmniTry is evaluated on a comprehensive benchmark consisting of 12 common classes of wearable objects, with both in-shop and in-the-wild images. Experimental results suggest that OmniTry shows better performance on both object localization and ID-preservation compared with existing methods. The code, model weights, and evaluation benchmark of OmniTry will be made publicly available.



Paperid:458
Authors:Wang Yanzheng, Xin Yang, Yujun Wang, Shizhe Hu, Mingliang Xu
Abstract:
Deep multi-modal clustering (DMC) aims to explore the correlated information from different modalities to improve the clustering performance. Most existing DMCs attempt to investigate the consistency or/and complementarity information by fusing all modalities, but this will lead to the following challenges: 1) Information conflicts between modalities emerge. 2) Information-rich modalities may be weakened. To address the above challenges, we propose a diversity-oriented deep multi-modal clustering (DDMC) method, where the core is dominant modality enhancement instead of multi-modal fusion. Specifically, we select the modality with the highest average silhouette coefficient as the dominant modality, then learn the diversity information between the dominant madality and the remaining ones with diversity learning, and finally enhance the dominant modality for clustering. Extensive experiments show the superiority of the proposed method over several compared DMC methods. To our knowledge, this is the first work to perform multi-modal clustering by enhancing the dominant modality instead of fusion.



Paperid:459
Authors:Kevin H. Lam, Thang Bui, George Deligiannidis, Yee Whye Teh
Abstract:
Latent Gaussian variables have been popularised in probabilistic machine learning. In turn, gradient estimators are the machinery that facilitates gradient-based optimisation for models with latent Gaussian variables. The reparameterisation trick is often used as the default estimator as it is simple to implement and yields low-variance gradients for variational inference. In this work, we propose the R2-G2 estimator as the Rao-Blackwellisation of the reparameterisation gradient estimator. Interestingly, we show that the local reparameterisation gradient estimator for Bayesian MLPs is an instance of the R2-G2 estimator and Rao-Blackwellisation. This lets us extend benefits of Rao-Blackwellised gradients to a suite of probabilistic models. We show that initial training with R2-G2 consistently yields better performance in models with multiple applications of the reparameterisation trick.



Paperid:460
Authors:Sheng-Yu Wang, Aaron Hertzmann, Alexei Efros, Richard Zhang, Jun-Yan Zhu
Abstract:
Data attribution for text-to-image models aims to identify the training images that most significantly influenced a generated output. Existing attribution methods involve considerable computational resources for each query, making them impractical for real-world applications. We propose a novel approach for scalable and efficient data attribution. Our key idea is to distill a slow, unlearning-based attribution method to a feature embedding space for efficient retrieval of highly influential training images. During deployment, combined with efficient indexing and search methods, our method successfully finds highly influential images without running expensive attribution algorithms. We show extensive results on both medium-scale models trained on MSCOCO and large-scale Stable Diffusion models trained on LAION, demonstrating that our method can achieve better or competitive performance in a few seconds, faster than existing methods by 2,500x - 400,000x. Our work represents a meaningful step towards the large-scale application of data attribution methods on real-world models such as Stable Diffusion. Our code, models, and datasets will be made publicly available.



Paperid:461
Authors:Eray Can Elumar, Cem Tekin, Osman Yagan
Abstract:
Recent advances in large language models (LLMs) have enabled automated dataset labeling with minimal human supervision. While majority voting across multiple LLMs can improve label reliability by mitigating individual model biases, it incurs high computational costs due to repeated querying. In this work, we propose a novel online framework, Cost-aware Majority Voting (CaMVo), for efficient and accurate LLM-based dataset annotation. CaMVo adaptively selects a subset of LLMs for each data instance based on contextual embeddings, balancing confidence and cost without requiring pre-training or ground-truth labels. Leveraging a LinUCB-based selection mechanism and a Bayesian estimator over confidence scores, CaMVo estimates a lower bound on labeling accuracy for each LLM and aggregates responses through weighted majority voting. Our empirical evaluation on the MMLU and IMDB Movie Review datasets demonstrates that CaMVo achieves comparable or superior accuracy to full majority voting while significantly reducing labeling costs. This establishes CaMVo as a practical and robust solution for cost-efficient annotation in dynamic labeling environments.



Paperid:462
Authors:Syamantak Kumar, Daogao Liu, Kevin Tian, Chutong Yang
Abstract:
Abstract:Estimating the geometric median of a dataset is a robust counterpart to mean estimation, and is a fundamental problem in computational geometry. Recently, [HSU24] gave an $(\epsilon, \delta)$-differentially private algorithm obtaining an $\alpha$-multiplicative approximation to the geometric median objective, $\frac 1 n \sum_{i \in [n]} \|\cdot - \mathbf{x}_i\|$, given a dataset $D$ of $x_i$ for $i \in [n]$. Their algorithm requires $n \gtrsim \sqrt d \cdot \frac 1 {\alpha\epsilon}$ samples, which they prove is information-theoretically optimal. This result is surprising because its error scales with the effective radius of $D$ (i.e., of a ball capturing most points), rather than the worst-case radius. We give an improved algorithm that obtains the same approximation quality, also using $n \gtrsim \sqrt d \cdot \frac 1 {\alpha\epsilon}$ samples, but in time $\widetilde{O}(nd + \frac d {\alpha^2})$. Our runtime is nearly-linear, plus the cost of the cheapest non-private first-order method due to [CLMPS16]. To achieve our results, we use subsampling and geometric aggregation tools inspired by FriendlyCore [TCKMS22] to speed up the "warm start" component of the [HSU24] algorithm, combined with a careful custom analysis of DP-SGD's sensitivity for the geometric median objective.



Paperid:463
Authors:Timo Löhr, Paul Hofman, Felix Mohr, Eyke Hüllermeier
Abstract:
Predictions in the form of sets of probability distributions, so-called credal sets, provide a suitable means to represent a learner's epistemic uncertainty. In this paper, we propose a theoretically grounded approach to credal prediction based on the statistical notion of relative likelihood: The target of prediction is the set of all (conditional) probability distributions produced by the collection of plausible models, namely those models whose relative likelihood exceeds a specified threshold. This threshold has an intuitive interpretation and allows for controlling the trade-off between correctness and precision of credal predictions. We tackle the problem of approximating credal sets defined in this way by means of suitably modified ensemble learning techniques. To validate our approach, we illustrate its effectiveness by experiments on benchmark datasets demonstrating superior uncertainty representation without compromising predictive performance. We also compare our method against several state-of-the-art baselines in credal prediction.



Paperid:464
Authors:Abhinav Joshi, Divyanshu Bhatt, Ashutosh Modi
Abstract:
Abstract:Large Language Models (LLMs) show strong generalization across diverse tasks, yet the internal decision-making processes behind their predictions remains opaque. In this work, we study the geometry of hidden representations in LLMs through the lens of $\textit{intrinsic dimension}$ (ID), focusing specifically on decision-making dynamics in a multiple-choice question answering (MCQA) setting. We perform a large-scale study, with 28 open-weight transformer models and estimate ID across layers using multiple estimators, while also quantifying per-layer performance on MCQA tasks. Our findings reveal a consistent ID pattern across models: early layers operate on low-dimensional manifolds, middle layers expand this space, and later layers compress it again, converging to decision-relevant representations. Together, these results suggest LLMs implicitly learn to project linguistic inputs onto structured, low-dimensional manifolds aligned with task-specific decisions, providing new geometric insights into how generalization and reasoning emerge in language models.



Paperid:465
Authors:Daman Arora, Andrea Zanette
Abstract:
Scaling model size and training data has led to great advances in the performance of Large Language Models (LLMs). However, the diminishing returns of this approach necessitate alternative methods to improve model capabilities, particularly in tasks requiring advanced reasoning. Large reasoning models, which leverage long chain-of-thoughts, bring unprecedented breakthroughs in problem-solving capabilities but at a substantial deployment cost associated to longer generations. Reducing inference costs is crucial for the economic feasibility, user experience, and environmental sustainability of these models.In this work, we propose to train large reasoning models to reason efficiently. Our method incentivizes models to minimize unnecessary computational overhead while largely maintaining accuracy, thereby achieving substantial deployment efficiency gains. It enables the derivation of a family of reasoning models with varying efficiency levels, controlled via a single hyperparameter. Experiments on two open-weight large reasoning models demonstrate significant reductions in inference cost while preserving most of the accuracy.



Paperid:466
Authors:Nuowei Liu, Jiahao Kuang, Yanting Liu, Changzhi Sun, Tao Ji, Yuanbin Wu, Man Lan
Abstract:
Protein design is a fundamental challenge in biotechnology, aiming to design novel sequences with specific functions within the vast space of possible proteins. Recent advances in deep generative models have enabled function-based protein design from textual descriptions, yet struggle with structural plausibility. Inspired by classical protein design methods that leverage natural protein structures, we explore whether incorporating fragments from natural proteins can enhance foldability in generative models. Our empirical results show that even random incorporation of fragments improves foldability. Building on this insight, we introduce ProDVa, a novel protein design approach that integrates a text encoder for functional descriptions, a protein language model for designing proteins, and a fragment encoder to dynamically retrieve protein fragments based on textual functional descriptions. Experimental results demonstrate that our approach effectively designs protein sequences that are both functionally aligned and structurally plausible. Compared to state-of-the-art models, ProDVa achieves comparable function alignment using less than 0.04% of the training data, while designing significantly more well-folded proteins, with the proportion of proteins having pLDDT above 70 increasing by 7.38% and those with PAE below 10 increasing by 9.6%. Datasets and codes will be made publicly available in the final version.



Paperid:467
Authors:Jiongli Zhu, Parjanya Prashant, Alex Cloninger, Babak Salimi
Abstract:
Abstract:Training data increasingly shapes not only model accuracy but also regulatory compliance and market valuation of AI assets. Yet existing valuation methods remain inadequate: model-based techniques depend on a single fitted model and inherit its biases, while algorithm-based approaches such as Data Shapley require costly retrainings at web scale. Recent Wasserstein-based model-agnostic methods rely on approximations that misrank examples relative to their true leave-one-out (LOO) utility. We introduce KAIROS, a scalable, model-agnostic valuation framework that assigns each example a distributional influence score: its contribution to the Maximum Mean Discrepancy (MMD) between the empirical training distribution and a clean reference set. Unlike Wasserstein surrogates, our MMD-based influence admits a closed-form solution that faithfully approximates the exact LOO ranking within O($\frac{1}{N^2}$) error, requires no retraining, and naturally extends to conditional kernels for unified label- and feature-error detection. Moreover, KAIROS supports efficient online updates: when a new batch of size m arrives, all scores can be updated in O(mN) time, delivering up to 50x speedup without compromising ranking quality. Empirical evaluations on noise, mislabeling, and poisoning benchmarks show that KAIROS consistently outperforms state-of-the-art model-, Shapley-, and Wasserstein-based baselines in both accuracy and runtime. We provide rigorous theoretical guarantees, including symmetry for reproducible rankings and density-separation for interpretable thresholds.



Paperid:468
Authors:Bhavya, Lenart Treven, Carmelo Sferrazza, Florian Dorfler, Pieter Abbeel, Andreas Krause
Abstract:
We address the challenge of efficient exploration in model-based reinforcement learning (MBRL), where the system dynamics are unknown and the RL agent must learn directly from online interactions. We proposeScalable andOptimisticMBRL(SOMBRL), an approach based on the principle of optimism in the face of uncertainty. SOMBRL learns an uncertainty-aware dynamics model andgreedilymaximizes a weighted sum of the extrinsic reward and the agent's epistemic uncertainty. SOMBRL is compatible with any policy optimizers or planners, and under common regularity assumptions on the system, we show that SOMBRL has sublinear regret for nonlinear dynamics in the (i) finite-horizon, (ii) discounted infinite-horizon, and (iii) non-episodic setting. Additionally, SOMBRL offers a flexible and scalable solution for principled exploration. We evaluate SOMBRL on state-based and visual-control environments, where it displays strong performance across all tasks and baselines. We also evaluate SOMBRL on a dynamic RC car hardware and show SOMBRL outperforms the state-of-the-art, illustrating the benefits of principled exploration for MBRL.



Paperid:469
Authors:Yunan Lu, Bowen Xue, Xiuyi Jia, Lei Yang
Abstract:
Label Distribution Learning (LDL) has emerged as a powerful framework for estimating complete conditional label distributions, providing crucial reliability for risk-sensitive decision-making tasks. While existing LDL algorithms exhibit competent performance under the conventional LDL performance evaluation methods, two key limitations remain: (1) current algorithms systematically underperform on the samples with low-entropy label distributions, which can be particularly valuable for decision making, and (2) the conventional performance evaluation methods are inherently biased due to the numerical imbalance of samples. In this paper, through empirical and theoretical analyses, we find that excessive cohesion between anchor vectors contributes significantly to the observed entropy bias phenomenon in LDL algorithms. Accordingly, we propose an inter-anchor angular regularization term that mitigates cohesion among anchor vectors by penalizing over-small angles. Besides, to alleviate the numerical imbalance of high-entropy samples in test set, we propose an entropy-calibrated aggregation strategy that obtains the overall model performance by evaluating performance on the low-entropy and high-entropy subsets of the overall test set separately. Finally, we conduct extensive experiments on various real-world datasets to demonstrate the effectiveness of our proposal.



Paperid:470
Authors:Lingxiao Huang, Zhize Li, Nisheeth K. Vishnoi, Runkai Yang, Haoyu Zhao
Abstract:
Abstract:We study the problem of constructing coresets for $(k, z)$-clustering when the input dataset is corrupted by stochastic noise drawn from a known distribution. In this setting, evaluating the quality of a coreset is inherently challenging, as the true underlying dataset is unobserved. To address this, we investigate coreset construction using surrogate error metrics that are tractable and provably related to the true clustering cost. We analyze a traditional metric from prior work and introduce a new error metric that more closely aligns with the true cost. Although our metric is defined independent of the noise distribution, it enables approximation guarantees that scale with the noise level. We design a coreset construction algorithm based on this metric and show that, under mild assumptions on the data and noise, enforcing an $\varepsilon$-bound under our metric yields smaller coresets and tighter guarantees on the true clustering cost than those obtained via classical metrics. In particular, we prove that the coreset size can improve by a factor of up to $\mathrm{poly}(k)$, where $n$ is the dataset size. Experiments on real-world datasets support our theoretical findings and demonstrate the practical advantages of our approach.



Paperid:471
Authors:Matthew Zurek, Yudong Chen
Abstract:
We study value-iteration (VI) algorithms for solving general (a.k.a. multichain) Markov decision processes (MDPs) under the average-reward criterion, a fundamental but theoretically challenging setting. Beyond the difficulties inherent to all average-reward problems posed by the lack of contractivity and non-uniqueness of solutions to the Bellman operator, in the multichain setting an optimal policy must solve the navigation subproblem of steering towards the best connected component, in addition to optimizing long-run performance within each component. We develop algorithms which better solve this navigational subproblem in order to achieve faster convergence for multichain MDPs, obtaining improved rates of convergence and sharper measures of complexity relative to prior work. Many key components of our results are of potential independent interest, including novel connections between average-reward and discounted problems, optimal fixed-point methods for discounted VI which extend to general Banach spaces, new sublinear convergence rates for the discounted value error, and refined suboptimality decompositions for multichain MDPs. Overall our results yield faster convergence rates for discounted and average-reward problems and expand the theoretical foundations of VI approaches.



Paperid:472
Authors:Dhananjay Ashok, Jonathan May
Abstract:
The text produced by language models (LMs) can exhibit specific `behaviors,' such as a failure to follow alignment training, that we hope to detect and react to during deployment. Identifying these behaviors can often only be done post facto, i.e., after the entire text of the output has been generated. We provide evidence that there are times when we can predict how an LM will behave early in computation, before even a single token is generated. We show that probes trained on the internal representation of input tokens alone can predict a wide range of eventual behaviors over the entire output sequence. Using methods from conformal prediction, we provide provable bounds on the estimation error of our probes, creating precise early warning systems for these behaviors. The conformal probes can identify instances that will trigger alignment failures (jailbreaking) and instruction-following failures, without requiring a single token to be generated. An early warning system built on the probes reduces jailbreaking by 91%. Our probes also show promise in pre-emptively estimating how confident the model will be in its response, a behavior that cannot be detected using the output text alone. Conformal probes can preemptively estimate the final prediction of an LM that uses Chain-of-Thought (CoT) prompting, hence accelerating inference. When applied to an LM that uses CoT to perform text classification, the probes drastically reduce inference costs (65% on average across 27 datasets), with negligible accuracy loss. Encouragingly, probes generalize to unseen datasets and perform better on larger models, suggesting applicability to the largest of models in real-world settings.



Paperid:473
Authors:Jie Liu, Gongye Liu, Jiajun Liang, Ziyang Yuan, Xiaokun Liu, Mingwu Zheng, Xiele Wu, Qiulin Wang, Menghan Xia, Xintao Wang, Xiaohong Liu, Fei Yang, Pengfei Wan, Di ZHANG, Kun Gai, Yujiu Yang, Wanli Ouyang
Abstract:
Video generation has achieved significant advances through rectified flow techniques, but issues like unsmooth motion and misalignment between videos and prompts persist. In this work, we develop a systematic pipeline that harnesses human feedback to mitigate these problems and refine the video generation model. Specifically, we begin by constructing a large-scale human preference dataset focused on modern video generation models, incorporating pairwise annotations across multi-dimensions. We then introduce VideoReward, a multi-dimensional video reward model, and examine how annotations and various design choices impact its rewarding efficacy. From a unified reinforcement learning perspective aimed at maximizing reward with KL regularization, we introduce three alignment algorithms for flow-based models. These include two training-time strategies: direct preference optimization for flow (Flow-DPO) and reward weighted regression for flow (Flow-RWR), and an inference-time technique, Flow-NRG, which applies reward guidance directly to noisy videos. Experimental results indicate that VideoReward significantly outperforms existing reward models, and Flow-DPO demonstrates superior performance compared to both Flow-RWR and supervised fine-tuning methods. Additionally, Flow-NRG lets users assign custom weights to multiple objectives during inference, meeting personalized video quality needs.



Paperid:474
Authors:Uri Gadot, Rinon Gal, Yftah Ziser, Gal Chechik, Shie Mannor
Abstract:
Text-to-image generation has evolved beyond single monolithic models to complex multi-component pipelines that combine various enhancement tools. While these pipelines significantly improve image quality, their effective design requires substantial expertise. Recent approaches automating this process through large language models (LLMs) have shown promise but suffer from two critical limitations: extensive computational requirements from generating images with hundreds of predefined pipelines, and poor generalization beyond memorized training examples. We introduce a novel reinforcement learning-based framework that addresses these inefficiencies. Our approach first trains an ensemble of reward models capable of predicting image quality scores directly from prompt-workflow combinations, eliminating the need for costly image generation during training. We then implement a two-phase training strategy: initial workflow prediction training followed by GRPO-based optimization that guides the model toward higher-performing regions of the workflow space. Additionally, we incorporate a classifier-free guidance based enhancement technique that extrapolates along the path between the initial and GRPO-tuned models, further improving output quality. We validate our approach through a set of comparisons, showing that it can successfully create new flows with greater diversity and lead to superior image quality compared to existing baselines.



Paperid:475
Authors:Lin Xu, Xinyun Yuan, Yuxuan Liang, Suwan Yin, Yuankai Wu
Abstract:
We introduce Aeolus, a large-scale Multi-modal Flight Delay Dataset designed to advance research on flight delay prediction and support the development of foundation models for tabular data. Existing datasets in this domain are typically limited to flat tabular structures and fail to capture the spatiotemporal dynamics inherent in delay propagation. Aeolus addresses this limitation by providing three aligned modalities: (i) a tabular dataset with rich operational, meteorological, and airport-level features for over 50 million flights; (ii) a flight chain module that models delay propagation along sequential flight legs, capturing upstream and downstream dependencies; and (iii) a flight network graph that encodes shared aircraft, crew, and airport resource connections, enabling cross-flight relational reasoning. The dataset is carefully constructed with temporal splits, comprehensive features, and strict leakage prevention to support realistic and reproducible machine learning evaluation. Aeolus supports a broad range of tasks, including regression, classification, temporal structure modeling, and graph learning, serving as a unified benchmark across tabular, sequential, and graph modalities. We release baseline experiments and preprocessing tools to facilitate adoption. Aeolus fills a key gap for both domain-specific modeling and general-purpose structured data research.



Paperid:476
Authors:Alberto Fernandez-Hernandez, Jose Mestre, Manuel F. Dolz, José Duato, Enrique Quintana-Orti
Abstract:
Initialization plays a critical role in Deep Neural Network training, directly influencing convergence, stability, and generalization. Common approaches such as Glorot and He initializations rely on randomness, which can produce uneven weight distributions across layer connections. In this paper, we introduce the Sinusoidal initialization, a novel deterministic method that employs sinusoidal functions to construct structured weight matrices expressly to improve the spread and balance of weights throughout the network while simultaneously fostering a more uniform, well‑conditioned distribution of neuron activation states from the very first forward pass. Because Sinusoidal initialization begins with weights and activations that are already evenly and efficiently utilized, it delivers consistently faster convergence, greater training stability, and higher final accuracy across a wide range of models, including convolutional neural networks, vision transformers, and large language models. On average, our experiments show an increase of 4.8 % in final validation accuracy and 20.9 % in convergence speed. By replacing randomness with structure, this initialization provides a stronger and more reliable foundation for Deep Learning systems.



Paperid:477
Authors:Marco Pacini, Gabriele Santin, Bruno Lepri, Shubhendu Trivedi
Abstract:
Equivariant neural networks provide a principled framework for incorporating symmetry into learning architectures and have been extensively analyzed through the lens of theirseparation power, that is, the ability to distinguish inputs modulo symmetry. This notion plays a central role in settings such as graph learning, where it is often formalized via the Weisfeilern–Leman hierarchy. In contrast, theuniversalityof equivariant models—their capacity to approximate target functions—remains comparatively underexplored. In this work, we investigate the approximation power of equivariant neural networks beyond separation constraints.We show that separation power does not fully capture expressivity: models with identical separation power may differ in their approximation ability.To demonstrate this, we characterize the universality classes of shallow invariant networks, providing a general framework for understanding which functions these architectures can approximate. Since equivariant models reduce to invariant ones under projection, this analysis yields sufficient conditions under which shallow equivariant networks fail to be universal.Conversely, we identify settings where shallow models do achieve separation-constrained universality.These positive results, however, depend critically on structural properties of the symmetry group, such as the existence of adequate normal subgroups, which may not hold in important cases like permutation symmetry.



Paperid:478
Authors:Tom Yan, Chicheng Zhang
Abstract:
With businesses starting to deploy agents to act on their behalf, an emerging challenge that businesses have to contend with is how to incentivize other agents with differing interests to work alongside its own agent. In present day commerce, payment is a common way that different parties use to align their business interests. In this paper, we study how one could analogously learn such payment schemes for aligning agents in the decentralized multi-agent setting. We model this problem as a Stackelberg Markov game, in which the leader can commit to a policy and also designate a set of outcome-based payments. We are interested in answering the question: when do efficient learning algorithms exist? To this end, we study the computational and statistical complexity of planning and learning in general-sum and cooperative games. In general-sum games, we find that planning is computationally intractable. In cooperative games, we show that learning can be statistically hard without payment and efficient with payment, showing that payment is necessary for learning even with aligned rewards. Altogether, our work aims to consolidate our theoretical understanding of algorithms that can compute optimal payment schemes for aligning decentralized agents.



Paperid:479
Authors:Jaeik Kim, Woojin Kim, Woohyeon Park, Jaeyoung Do
Abstract:
Visual personalization is essential in user-facing AI systems such as smart homes and healthcare, where aligning model behavior with user-centric concepts is critical. However, recent large Vision-Language Models (VLMs), despite their broad applicability, remain underexplored in their ability to adapt to individual users. In this paper, we introduce MMPB, the first extensive benchmark for evaluating VLMs on personalization. MMPB comprises 10k image-query pairs and includes 111 personalizable concepts across four categories: humans, animals, objects, and characters, with the human category enriched with preference-grounded queries. We structure personalization into three main task types, each highlighting a different key property of VLMs. Using 23 widely used VLMs including both open- and closed-source models, we evaluate personalization performance via a three-stage protocol: concept injection, multi-turn dialogue, and personalized querying. Our findings indicate that most VLMs (including some closed-source models) struggle with personalization, particularly in maintaining consistency over dialogue, handling user preferences, and adapting to visual cues. Our analysis reveals that the challenges in VLM personalization (such as refusal behaviors and long-context forgetting) highlight substantial room for improvement. By identifying these limitations and offering a scalable benchmark, MMPB offers valuable insights and a solid foundation for future research toward truly personalized multi-modal AI.



Paperid:480
Authors:Daniel Wesego, Pedram Rooshenas
Abstract:
Deep neural networks remain highly susceptible to adversarial attacks, where small, subtle perturbations to input images may induce misclassification. We propose a novel optimization-based purification framework that directly removes these perturbations by maximizing a Bayesian-inspired objective combining a pretrained diffusion prior with a likelihood term tailored to the adversarial perturbation space. Our method iteratively refines a given input through gradient-based updates of a combined score-based loss to guide the purification process. Unlike existing optimization-based defenses that treat adversarial noise as generic corruption, our approach explicitly integrates the adversarial landscape into the objective. Experiments performed on CIFAR-10 and CIFAR-100 demonstrate strong robust accuracy against a range of common adversarial attacks. Our work offers a principled test-time defense grounded in probabilistic inference using score-based generative models.



Paperid:481
Authors:Ahmed Khaled, Satyen Kale, Arthur Douillard, Chi Jin, Rob Fergus, Manzil Zaheer
Abstract:
Abstract:Modern machine learning often requires training with large batch size, distributed data, and massively parallel compute hardware (like mobile and other edge devices or distributed data centers). Communication becomes a major bottleneck in such settings but methods like Local Stochastic Gradient Descent (Local SGD) show great promise to reduce the global communication need. Local SGD consists of three parts: a local optimization processes, an aggregation mechanism, and an outer optimizer that uses the aggregated updates from the nodes to produce a new model. While there exists an extensive literature on understanding the impact of hyperparameters in the local optimization process, the choice of outer optimizer and its hyperparameters is less clear. We study the role of the outer learning in Local SGD, and prove new convergence guarantees for the algorithm. In particular, we show that tuning the outer learning rate allows us to (a) trade off between optimization error and stochastic gradient noise variance, and (b) make up for ill-tuning of the inner learning rate. Our theory suggests that the outer learning rate should sometimes be set to values greater than $1$. We extend our results to apply to when we use momentum in the outer optimizer, and also introduce a novel data-dependent analysis of Local SGD that yields further insights on outer learning rate tuning. We conduct comprehensive experiments with standard language models and various outer optimizers to validate our theory.



Paperid:482
Authors:Haiqing Zhu, Tijana Zrnic, Celestine Mendler-Dünner
Abstract:
Abstract:Modern predictive models can learn complex statistical patterns and personalize predictions to individual users. While minimizing a notion of risk for the learner, the consequences of these predictions might not always equally align with the interests of the users they concern.As a way to contest predictive systems, users might act strategically in order to achieve favorable outcomes. While past work has studied strategic user behavior in learning systems, the focus has largely been on individual strategies: users respond to the system without consideration of the other users' utilities and actions.In this paper, we study the consequences of collective reasoning in predictive systems. First, we formalize level-$k$ thinking, a concept from behavioral economics, and analyze how more sophisticated levels of strategic thought influence learning dynamics and equilibrium outcomes. Then, by contrasting the resulting equilibria with those achievable through coordinated actions we characterize the benefits and trade-offs of collective reasoning. A key quantity in our analysis is a new notion of alignment between the learner's and the population's utilities. Together with extensive simulations, our novel concepts and results, bridge between strategic classification and collective action, two widely studied, but so far disjoint subfields of machine learning.



Paperid:483
Authors:Marius Aasan, Martine Hjelkrem-Tan, Nico Catalano, Changkyu Choi, Adín Ramírez Rivera
Abstract:
Vision Transformers rely on fixed patch tokens that ignore the spatial and semantic structure of images. In this work, we introduce an end-to-end differentiable tokenizer that adapts to image content with pixel-level granularity while remaining backward-compatible with existing architectures for retrofitting pretrained models. Our method uses hierarchical model selection with information criteria to provide competitive performance in both image-level classification and dense-prediction tasks, and even supports out-of-the-box raster-to-vector conversion.



Paperid:484
Authors:Sergio Calo, Anders Jonsson, Gergely Neu, Ludovic Schwartz, Javier Segovia-Aguas
Abstract:
Bisimulation metrics are powerful tools for measuring similarities between stochastic processes, and specifically Markov chains. Recent advances have uncovered that bisimulation metrics are, in fact, optimal-transport distances, which has enabled the development of fast algorithms for computing such metrics with provable accuracy and runtime guarantees. However, these recent methods, as well as all previously known methods, assume full knowledge of the transition dynamics. This is often an impractical assumption in most real-world scenarios, where typically only sample trajectories are available.In this work, we propose a stochastic optimization method that addresses this limitationand estimates bisimulation metrics based on sample access, without requiring explicit transition models.Our approach is derived from a new linear programming (LP) formulation of bisimulation metrics, which we solve using a stochastic primal-dual optimization method. We provide theoretical guarantees on the sample complexity of the algorithm and validate its effectiveness through a series of empirical evaluations.



Paperid:485
Authors:Gyubin Lee, Bao Truong, Jaesik Yoon, Dongwoo Lee, Minsu Kim, Yoshua Bengio, Sungjin Ahn
Abstract:
Diffusion models have demonstrated strong generative capabilities across domains ranging from image synthesis to complex reasoning tasks. However, most inference-time scaling methods rely on fixed denoising schedules, limiting their ability to allocate computation based on instance difficulty or task-specific demands adaptively. We introduce the challenge of adaptive inference-time scaling—dynamically adjusting computational effort during inference—and propose Adaptive Bi-directional Cyclic Diffusion (ABCD), a flexible, search-based inference framework. ABCD refines outputs through bi-directional diffusion cycles while adaptively controlling exploration depth and termination. It comprises three components: Cyclic Diffusion Search, Automatic Exploration-Exploitation Balancing, and Adaptive Thinking Time. Experiments show that ABCD improves performance across diverse tasks while maintaining computational efficiency.



Paperid:486
Authors:Taichi Liu, Zhenyu Wang, Ruofeng Liu, Guang Wang, Desheng Zhang
Abstract:
Recent advancements in 3D object detection and novel category detection have made significant progress, yet research on learning generalized 3D objectness remains insufficient. In this paper, we delve into learning open-world 3D objectness, which focuses on detecting all objects in a 3D scene, including novel objects unseen during training. Traditional closed-set 3D detectors struggle to generalize to open-world scenarios, while directly incorporating 3D open-vocabulary models for open-world ability struggles with vocabulary expansion and semantic overlap. To achieve generalized 3D object discovery, We propose OP3Det, a class-agnostic Open-World Prompt-free 3D Detector to detect any objects within 3D scenes without relying on hand-crafted text prompts. We introduce the strong generalization and zero-shot capabilities of 2D foundation models, utilizing both 2D semantic priors and 3D geometric priors for class-agnostic proposals to broaden 3D object discovery. Then, by integrating complementary information from point cloud and RGB image in the cross-modal mixture of experts, OP3Det dynamically routes uni-modal and multi-modal features to learn generalized 3D objectness. Extensive experiments demonstrate the extraordinary performance of OP3Det, which significantly surpasses existing open-world 3D detectors by up to 16.0% in AR and achieves a 13.5% improvement compared to closed-world 3D detectors.



Paperid:487
Authors:Ali Bekar, Siddhant Agarwal, Christian Hüttig, Nicola Tosi, David Greenberg
Abstract:
For classical PDE solvers, adjusting the spatial resolution and time step offers a trade-off between speed and accuracy. Neural emulators often achieve better speed-accuracy trade-offs by operating accurately on a compact representation of the PDE system. Coarsened PDE fields are a simple and effective representation, but cannot exploit fine spatial scales in the high-fidelity numerical solutions. Alternatively, unstructured latent representations provide efficient autoregressive rollouts, but cannot enforce local interactions or physical laws as inductive biases. To overcome these limitations, we introduce hybrid representations that augment coarsened PDE fields with spatially structured latent variables extracted from high-resolution inputs. Hybrid representations provide efficient rollouts, can be trained on a simple loss defined on coarsened PDE fields, and support hard physical constraints. When predicting fine- and coarse-scale features across multiple PDE emulation tasks, they outperform or match the speed-accuracy trade-offs of the best convolutional, attentional, Fourier operator-based and autoencoding baselines.



Paperid:488
Authors:Anh Nguyen, Viet Nguyen, Duc Vu, Trung Dao, Chi Tran, Toan Tran, Anh Tran
Abstract:
Abstract:Shortcut models represent a promising pathway for generative modeling, supporting one-step, few-step, and many-step sampling without adversarial training. Yet their practical adoption has been hindered by several key issues: the need to fix its classifier-free guidance scale at training time, which limits flexibility at inference; high variance from joint training on random noise–data pairs, which decelerates and destabilizes convergence; and reliance on low-level distance on direct domain that bias reconstructions toward low frequencies and degrade sample quality. Moreover, we uncover a previously overlooked problem of accumulation scale in classifier-free guidance and a subtle conflict between EMA updates and the self-consistency objective. To address these challenges, we introduce a unified training framework for shortcut models that (1) parameterizes guidance scales to support dynamic guidance sampling at inference, (2) mitigates frequency bias with a Multi-level Wavelet Loss, (3) incorporates interval guidance directly into the loss, (4) reduces training variance via Scaling Optimal Transport Matching, and (5) preserves self-consistency alongside training stability through a Twin EMA strategy. Extensive experiments on ImageNet $256\times256$ demonstrate that our approach yields substantial FID improvements over baseline shortcut models in one-step, few-step, and multi-step generation.



Paperid:489
Authors:Yudong Han, Yan Yang, Hao Yang, Liyuan Pan
Abstract:
In photography, an All-in-Focus (AiF) image may not always effectively convey the creator’s intent. Professional photographers manipulate Depth of Field (DoF) to control which regions appear sharp or blurred, achieving compelling artistic effects. For general users, the ability to flexibly adjust DoF enhances creative expression and image quality. In this paper, we propose UiD, a User-Instructed DoF control framework, that allows users to specify refocusing regions using text, box, or point prompts, and our UiD automatically simulates in-focus and out-of-focus (OoF) regions in the given images.However, controlling defocus blur in a single-lens camera remains challenging due to the difficulty in estimating depth-aware aberrations and the suboptimal quality of reconstructed AiF images.%To address this, we leverage dual-pixel (DP) sensors, commonly found in DSLR-style and mobile cameras. DP sensors provide a small-baseline stereo pair in a single snapshot, enabling depth-aware aberration estimation. %Our approach first establishes an invertible mapping between OoF and AiF images to learn spatially varying defocus kernels and the disparity features. These depth-aware kernels enable both deblurring of OoF images into AiF representations and reblurring AiF images into OoF representations. %For user-guided refocusing, we first generate masks based on user prompts using SAM, which modulates disparity features in closed form, allowing dynamic kernel re-estimation for reblurring. This achieves user-controlled refocusing effects. Extensive experiments on both common datasets and the self-collected dataset demonstrate that UiD offers superior flexibility and quality in DoF manipulation imaging.



Paperid:490
Authors:Chengpeng Li, Zhengyang Tang, Ziniu Li, Mingfeng Xue, Keqin Bao, Tian Ding, Ruoyu Sun, Benyou Wang, Xiang Wang, Junyang Lin, Dayiheng Liu
Abstract:
Large Reasoning Models (LRMs) like o1 and DeepSeek-R1 have shown remarkable progress in natural language reasoning with long chain-of-thought (CoT), yet they remain inefficient or inaccurate when handling complex mathematical operations.Addressing these limitations through computational tools (e.g., computation libraries and symbolic solvers) is promising, but it introduces a technical challenge: Code Interpreter (CI) brings external knowledge beyond the model's internal text representations, thus the direct combination is not efficient. This paper introduces CoRT, a post-training framework for teaching LRMs to leverage CI effectively and efficiently. As a first step, we address the data scarcity issue by synthesizing code-integrated reasoning data through Hint-Engineering, which strategically inserts different hints at appropriate positions to optimize LRM-CI interaction. We manually create 30 high-quality samples, upon which we post-train models ranging from 1.5B to 32B parameters, with supervised fine-tuning, rejection fine-tuning and reinforcement learning. Our experimental results demonstrate that Hint-Engineering models achieve 4\% and 8\% absolute improvements on DeepSeek-R1-Distill-Qwen-32B and DeepSeek-R1-Distill-Qwen-1.5B respectively, across five challenging mathematical reasoning datasets. Furthermore, Hint-Engineering models use about 30\% fewer tokens for the 32B model and 50\% fewer tokens for the 1.5B model compared with the natural language models.



Paperid:491
Authors:Zitao Fang, Chenxuan Li, Hongting Zhou, Shuyang Yu, Guodong DU, Ashwaq Qasem, Yang Lu, Jing Li, Junsong Zhang, Sim Kuan Goh
Abstract:
Electroencephalography (EEG) has wide-ranging applications, from clinical diagnosis to brain-computer interfaces (BCIs). With the increasing volume and variety of EEG data, there has been growing interest in establishing foundation models (FMs) to scale up and generalize neural decoding. Despite showing early potential, applying FMs to EEG remains challenging due to substantial inter-subject, inter-task, and inter-condition variability, as well as diverse electrode configurations across recording setups. To tackle these open challenges, we proposeNeurIPT, a foundation model tailored for diverse EEG-basedNeuralInterfaces with aPre-trainedTransformer by capturing both homogeneous and heterogeneous spatio-temporal characteristics inherent in EEG signals. Temporally, we introduce Amplitude-Aware Masked Pretraining (AAMP), masking based on signal amplitude rather than random intervals, to learn robust representations across varying signal intensities beyond local interpolation. Moreover, this temporal representation is enhanced by a progressive Mixture-of-Experts (MoE) architecture, where specialized expert subnetworks are progressively introduced at deeper layers, adapting effectively to the diverse temporal characteristics of EEG signals. Spatially, NeurIPT leverages the 3D physical coordinates of electrodes, enabling effective transfer across varying EEG settings, and develops Intra-Inter Lobe Pooling (IILP) during fine-tuning to efficiently exploit regional brain features. Empirical evaluations across nine downstream BCI datasets, via fine-tuning and training from scratch, demonstrated NeurIPT consistently achieved state-of-the-art performance, highlighting its broad applicability and robust generalization. Our work pushes forward the state of FMs in EEG and offers insights into scalable and generalizable neural information processing systems.



Paperid:492
Authors:Yuheng Yuan, Qiuhong Shen, Shizun Wang, Xingyi Yang, Xinchao Wang
Abstract:
Abstract:Dense matching methods like DUST3R regress pairwise pointmaps for 3D reconstruction. However, the reliance on pairwise prediction and the limited generalization capability inherently restrict the global geometric consistency. In this work, we introduce \textbf{Test3R}, a surprisingly simple test-time learning technique that significantly boosts geometric accuracy. Using image triplets ($I_1,I_2,I_3$), Test3R generates reconstructions from pairs ($I_1,I_2$) and ($I_1,I_3$). The core idea is to optimize the network at test time via a self-supervised objective: maximizing the geometric consistency between these two reconstructions relative to the common image $I_1$. This ensures the model produces cross-pair consistent outputs, regardless of the inputs. Extensive experiments demonstrate that our technique significantly outperforms previous state-of-the-art methods on the 3D reconstruction and multi-view depth estimation tasks. Moreover, it is universally applicable and nearly cost-free, making it easily applied to other models and implemented with minimal test-time training overhead and parameter footprint.



Paperid:493
Authors:Kang An, Yuxing Liu, Rui Pan, Yi Ren, Shiqian Ma, Donald Goldfarb, Tong Zhang
Abstract:
Training deep neural networks (DNNs) is a structured optimization problem, because the parameters are naturally represented by matrices and tensors rather than simple vectors. Under this structural representation, it has been widely observed that gradients are low-rank and Hessians are approximately block-wise diagonal. These structured properties are crucial for designing efficient optimization algorithms but may not be utilized by current popular optimizers like Adam. In this paper, we present a novel optimization algorithm ASGO that capitalizes on these properties by employing a preconditioner that is adaptively updated using structured gradients. By fine-grained theoretical analysis, ASGO is proven to achieve superior convergence rates compared to existing structured gradient methods. Based on the convergence theory, we further demonstrate that ASGO can benefit from the low-rank and block-wise diagonal properties. We also discuss practical modifications of ASGO and empirically verify the effectiveness of the algorithm on language model tasks.



Paperid:494
Authors:Adarsh Jamadandi, Jing Xu, Adam Dziedzic, Franziska Boenisch
Abstract:
Deep neural networks (DNNs) have been shown to memorize their training data, yet similar analyses for graph neural networks (GNNs) remain largely under-explored. We introduce NCMemo (Node Classification Memorization), the first framework to quantify label memorization in semi-supervised node classification. We first establish an inverse relationship between memorization and graph homophily, i.e the property that connected nodes share similar labels/features. We find that lower homophily significantly increases memorization, indicating that GNNs rely on memorization to learn less homophilic graphs. Secondly, we analyze GNN training dynamics. We find that the increased memorization in low homophily graphs is tightly coupled to the GNNs' implicit bias on using graph structure during learning. In low homophily regimes, this structure is less informative, hence inducing memorization of the node labels to minimize training loss. Finally, we show that nodes with higher label inconsistency in their feature-space neighborhood are significantly more prone to memorization. Building on our insights into the link between graph homophily and memorization, we investigate graph rewiring as a means to mitigate memorization. Our results demonstrate that this approach effectively reduces memorization without compromising model performance. Moreover, we show that it lowers the privacy risk for previously memorized data points in practice. Thus, our work not only advances understanding of GNN learning but also supports more privacy-preserving GNN deployment.



Paperid:495
Authors:Junde Xu, Zijun Gao, Xinyi Zhou, hujie, Xingyi Cheng, Le Song, Guangyong Chen, Pheng-Ann Heng, Jiezhong Qiu
Abstract:
The inverse folding problem, aiming to design amino acid sequences that fold into desired three-dimensional structures, is pivotal for various biotechnological applications. Here, we introduce a novel approach leveraging Direct Preference Optimization (DPO) to fine-tune an inverse folding model using feedback from a protein folding model.Given a target protein structure, we begin by sampling candidate sequences from the inverse‐folding model, then predict the three‐dimensional structure of each sequence with the folding model to generate pairwise structural‐preference labels. These labels are used to fine‐tune the inverse‐folding model under the DPO objective. Our results on the CATH 4.2 test set demonstrate that DPO fine-tuning not only improves sequence recovery of baseline models but also leads to a significant improvement in average TM-Score from 0.77 to 0.81, indicating enhanced structure similarity. Furthermore, iterative application of our DPO-based method on challenging protein structures yields substantial gains, with an average TM-Score increase of 79.5\% with regard to the baseline model.This work establishes a promising direction for enhancing protein sequence design ability from structure feedback by effectively utilizing preference optimization.



Paperid:496
Authors:Kexin ZHENG, Lauriane Teyssier, Yinan Zheng, Yu Luo, Xianyuan Zhan
Abstract:
The recent development of zero-shot reinforcement learning (RL) has opened a new avenue for learning pre-trained generalist policies that can adapt to arbitrary new tasks in a zero-shot manner. While the popular Forward-Backward (FB) representations and related methods have shown promise in zero-shot RL, we found empirically that their non-expressive modeling architecture and extrapolation errors caused by out-of-the-distribution (OOD) actions during offline learning often lead to potentially biased and non-robust representations, eventually causing suboptimal performance. To address these issues, we propose Behavior-Rgularized Zero-shot RL with Expressivity enhancement (BREEZE), an upgraded FB-based framework that simultaneously enhances offline learning stability, FB representation learning quality, and policy extraction capability. Specifically, BREEZE introduces behavioral regularization in zero-shot RL policy learning, transforming policy optimization into a stable in-sample learning paradigm. Additionally, BREEZE employs the expressive self-attention based architectures for forward and backward representations, providing more accurate successor measure estimates that capture the complicated environment-task relationships. Moreover, BREEZE extracts the policy with a guided task-conditioned diffusion model, enabling optimal action synthesis while capturing the highly multi-modal action distributions in zero-shot RL settings. Extensive experiments on ExORL and FrankaKitchen demonstrate that BREEZE achieves state-of-the-art performance while exhibiting superior robustness compared to prior offline zero-shot RL methods.



Paperid:497
Authors:Sebastian Sanokowski, Lukas Gruber, Christoph Bartmann, Sepp Hochreiter, Sebastian Lehner
Abstract:
Diffusion bridges are a promising class of deep-learning methods for sampling from unnormalized distributions. Recent works show that the Log Variance (LV) loss consistently outperforms the reverse Kullback-Leibler (rKL) loss when using the reparametrization trick to compute rKL-gradients. While the on-policy LV loss yields identical gradients to the rKL loss when combined with the log-derivative trick for diffusion samplers with non-learnable forward processes, this equivalence does not hold for diffusion bridges or when diffusion coefficients are learned.Based on this insight we argue that for diffusion bridges the LV loss does not represent an optimization objective that can be motivated like the rKL loss via the data processing inequality. Our analysis shows that employing the rKL loss with the log-derivative trick (rKL-LD) does not only avoid these conceptual problems but also consistently outperforms the LV loss. Experimental results with different types of diffusion bridges on challenging benchmarks show that samplers trained with the rKL-LD loss achieve better performance. From a practical perspective we find that rKL-LD requires significantly less hyperparameter optimization and yields more stable training behavior.



Paperid:498
Authors:Ali Shirali, Arash Nasr-Esfahany, Abdullah Alomar, Parsa Mirtaheri, Rediet Abebe, Ariel Procaccia
Abstract:
Alignment with human preferences is commonly framed using a universal reward function, even though human preferences are inherently heterogeneous. We formalize this heterogeneity by introducing user types and examine the limits of the homogeneity assumption.We show that aligning to heterogeneous preferences with a single policy is best achieved using the average reward across user types. However, this requires additional information about annotators. We examine improvements under different information settings, focusing on direct alignment methods. We find that minimal information can yield first-order improvements, while full feedback from each user type leads to consistent learning of the optimal policy. Surprisingly, however, no sample-efficient consistent direct loss exists in this latter setting. These results reveal a fundamental tension between consistency and sample efficiency in direct policy alignment.



Paperid:499
Authors:Nilesh Gupta, Chong You, Srinadh Bhojanapalli, Sanjiv Kumar, Inderjit Dhillon, Felix Yu
Abstract:
Information retrieval (IR)—finding relevant documents for a given query from a corpus—is a problem of fundamental importance in machine learning. An emerging paradigm for IR is In-context Retrieval (ICR), which leverages large language models (LLMs) by directly incorporating candidate documents, the query, and task description into the model's input prompt for relevance identification. While this method harnesses LLM capabilities, achieving efficient and effective ICR presents a significant challenge, especially as the candidate list expands. To this end, this paper first identifies inherent structures in LLMs finetuned for ICR: (1) document tokens attend primarily to tokens within the same document and in the instructions when prompted properly; (2) query tokens, particularly the last ones and certain special ones such as ":", attend strongly towards the relevant documents. Motivated by these observations, we introduce ReTuning (Retrieval-Tuning), a novel method that adapts an LLM for ICR by (1) enforcing a structured sparse attention, which reduces attention complexity from quadratic to linear, and (2) directly optimizing attention to the relevant documents from the signal-carrying query tokens using a contrastive training objective, which improves retrieval effectiveness. Experiments on MSMarco and NQ with Mistral-7B demonstrate that ReTuned Mistral matches or outperforms standard fine-tuned Mistral while being significantly more efficient at inference (4.7x for 100 MSMarco documents in context) and 2−3x during training. Our approach presents a scalable and effective solution for ICR.



Paperid:500
Authors:Preston Fu, Oleh Rybkin, Zhiyuan Zhou, Michal Nauman, Pieter Abbeel, Sergey Levine, Aviral Kumar
Abstract:
As models become bigger and more powerful, it becomes increasingly essential to scale up training recipes to not just be bigger along every dimension, but to be compute optimal, in the sense that maximal benefit is extracted per unit of compute as a function of model size, gradient steps, batch size, and other hyperparameters. While compute optimal scaling has been explored extensively for supervised and unsupervised pre-training, it has received comparatively less attention for reinforcement learning (RL). In this paper, we study compute scaling for online, value-based deep RL. These methods present two axes for compute allocation: the capacity of the model and update-to-data (UTD) ratio. Given a total compute budget, we investigate how to best partition resources between these two axes to maximize performance or sample efficiency. In the process, we observe an interesting interplay between model size, UTD, and batch size, leading us to a phenomenon we call ``TD-overfitting'', where large batch size training results in worse Q-functions when smaller networks are used. Interestingly, we find that the presence of this sort of TD-overfitting reverses with larger models, enabling the use of greater batch sizes for training. We develop a mechanistic understanding of these interactions, allowing us to identify effective batch sizes when operating at scale. Our results lead to empirically validated prescriptions for compute-optimal scaling in deep RL, analogous to scaling laws for compute allocation in supervised learning but tailored to the unique dynamics of TD-learning.



Paperid:501
Authors:Rishi Jha, Collin Zhang, Vitaly Shmatikov, John Morris
Abstract:
We introduce the first method for translating text embeddings from one vector space to another without any paired data, encoders, or predefined sets of matches. Our unsupervised approach translates any embedding to and from a universal latent representation (i.e., a universal semantic structure conjectured by the Platonic Representation Hypothesis). Our translations achieve high cosine similarity across model pairs with different architectures, parameter counts, and training datasets.The ability to translate unknown embeddings into a different space while preserving their geometry has serious implications for the security of vector databases. An adversary with access only to embedding vectors can extract sensitive information about the underlying documents, sufficient for classification and attribute inference.



Paperid:502
Authors:Jiabin Tang, Lianghao Xia, Zhonghang Li, Chao Huang
Abstract:
The powerful reasoning capabilities of Large Language Models (LLMs) in mathematics and coding, combined with their ability to automate complex tasks through agentic frameworks, present unprecedented opportunities for accelerating scientific innovation. In this paper, we introduce AI-Researcher, a fully autonomous research system that transforms how AI-driven scientific discovery is conducted and evaluated. Our framework seamlessly orchestrates the complete research pipeline--from literature review and hypothesis generation to algorithm implementation and publication-ready manuscript preparation--with minimal human intervention. To rigorously assess autonomous research capabilities, we develop Scientist-Bench, a comprehensive benchmark comprising state-of-the-art papers across diverse AI research domains, featuring both guided innovation and open-ended exploration tasks. Through extensive experiments, we demonstrate that AI-Researcher achieves remarkable implementation success rates and produces research papers that approach human-level quality. This work establishes new foundations for autonomous scientific innovation that can complement human researchers by systematically exploring solution spaces beyond cognitive limitations.



Paperid:503
Authors:Arjun Patrawala, Jiahai Feng, Erik Jones, Jacob Steinhardt
Abstract:
Recent methods in language model interpretability employ sparse autoencoders and other techniques to decompose residual stream contributions into linear, semantically-meaningful features. Our work demonstrates that the underlying assumption of these methods—that residual stream contributions build additively upon each other—is fundamentally incomplete. We identify the Transformer Layer Correction Mechanism (TLCM), wherein adjacent transformer layers systematically counteract each other's contributions to the residual stream. TLCM appears in 5 out of 7 major open-source model families and operates across diverse contexts. We establish that TLCM emerges during pretraining, operates most strongly on tokens with high contextual dependency, and adaptively calibrates its correction strength based on the preceding layer's output. We further demonstrate that TLCM functions by selectively promoting and rejecting specific subspaces. We propose a "propose-and-reject" hypothesis: layers may suggest multiple potential features, while subsequent layers selectively filter inappropriate ones. This framework explains three persistent challenges in feature-based interpretability: why extracted features descriptions often suffer from low specificity; why feature-based interventions for model steering are unsuccessful at low magnitude; why recent work finds cross-layer transcoders outperform SAEs.



Paperid:504
Authors:Kevin Lu, Nicky Kriplani, Rohit Gandikota, Minh Pham, David Bau, Chinmay Hegde, Niv Cohen
Abstract:
Concept erasure, the ability to selectively prevent a model from generating specific concepts, has attracted growing interest, with various approaches emerging to address the challenge. However, it remains unclear how thoroughly these methods erase the target concept. We begin by proposing two conceptual models for the erasure mechanism in diffusion models: (i) reducing the likelihood of generating the target concept, and (ii) interfering with the model’s internal guidance mechanisms. To thoroughly assess whether a concept has been truly erased from the model, we introduce a suite of independent evaluations. Our evaluation framework includes adversarial attacks, novel probing techniques, and analysis of the model's alternative generations in place of the erased concept. Our results shed light on the tension between minimizing side effects and maintaining robustness to adversarial prompts. Broadly, our work underlines the importance of comprehensive evaluation for erasure in diffusion models.



Paperid:505
Authors:Zhiyuan Liang, Dongwen Tang, Yuhao Zhou, Xuanlei Zhao, Mingjia Shi, Wangbo Zhao, Zekai Li, Peihao Wang, Konstantin Schürholt, Damian Borth, Michael Bronstein, Yang You, Zhangyang "Atlas" Wang, Kai Wang
Abstract:
Abstract:Modern Parameter-Efficient Fine-Tuning (PEFT) methods such as low-rank adaptation (LoRA) reduce the cost of customizing large language models (LLMs), yet still require a separate optimization run for every downstream dataset. We introduce \textbf{Drag-and-Drop LLMs (\textit{DnD})}, a prompt-conditioned parameter generator that eliminates per-task training by mapping a handful of unlabeled task prompts directly to LoRA weight updates. A lightweight text encoder distills each prompt batch into condition embeddings, which are then transformed by a cascaded hyper-convolutional decoder into the full set of LoRA matrices. Once trained in a diverse collection of prompt–checkpoint pairs, DnD produces task-specific parameters in seconds, yielding (i) up to \textbf{12,000$\times$} lower overhead than full fine-tuning, (ii) average gains up to \textbf{30\%} in performance over the strongest training LoRAs on unseen common-sense reasoning, math, coding, and multimodal benchmarks, and (iii) robust cross-domain generalization despite never seeing the target data or labels. Our results demonstrate that prompt-conditioned parameter generation is a viable alternative to gradient-based adaptation for rapidly specializing LLMs.



Paperid:506
Authors:Jiaxin Wang, Wenxuan Tu, Jieren Cheng
Abstract:
Abstract:Graph kernels have emerged as a fundamental and widely adopted technique in graph machine learning. However, most existing graph kernel methods rely on fixed graph similarity estimation that cannot be directly optimized for task-specific objectives, leading to sub-optimal performance. To address this limitation, we propose a kernel-based learning framework called Hierarchical Shortest-Path Graph Kernel Network HSP-GKN, which seamlessly integrates graph similarity estimation with downstream tasks within a unified optimization framework. Specifically, we design a hierarchical shortest-path graph kernel that efficiently preserves both the semantic and structural information of a given graph by transforming it into hierarchical features used for subsequent neural network learning. Building upon this kernel, we develop a novel end-to-end learning framework that matches hierarchical graph features with learnable $hidden$ graph features to produce a similarity vector. This similarity vector subsequently serves as the graph embedding for end-to-end training, enabling the neural network to learn task-specific representations. Extensive experimental results demonstrate the effectiveness and superiority of the designed kernel and its corresponding learning framework compared to current competitors.



Paperid:507
Authors:Mihir Prabhudesai, Mengning Wu, Amir Zadeh, Katerina Fragkiadaki, Deepak Pathak
Abstract:
Autoregressive Models (ARMs) have dominated the landscape of large language models (LLMs), resulting in significant advancements across various tasks. Recently, diffusion-based language models have emerged, demonstrating promising results, yet their advantages over established ARMs remain unclear. In this paper, we investigate the strengths of Masked Diffusion Models (MDMs), showing that while they are computationally less efficient than ARMs, they excel significantly under data-constrained conditions. Through systematic empirical evaluations and detailed scaling law analyses, we illustrate how MDMs leverage repeated data more effectively, achieving superior performance across varying model sizes, unique data volumes, and computational budgets. In comprehensive experiments, including validation loss contours and diverse downstream benchmarks we validate that MDMs consistently yields lower validation losses and tangible practical benefits in data-constrainted settings. Our findings underscore the substantial advantage of diffusion-based approaches in scenarios where data availability is limited, marking them as particularly beneficial in resource-constrained environments.



Paperid:508
Authors:Sara Atito, Josef Kittler, Imran Razzak, Muhammad Awais
Abstract:
Abstract:Humans understand visual scenes by first capturing a global impression and then refining this understanding into distinct, object-like components. Inspired by this process, we introduce \textbf{C}oncept-\textbf{G}uided \textbf{S}elf-\textbf{S}upervised \textbf{L}earning (CG-SSL), a novel framework that brings structure and interpretability to representation learning through a curriculum of three training phases: (1) global scene encoding, (2) discovery of visual concepts via tokenised cross-attention, and (3) alignment of these concepts across views.Unlike traditional SSL methods, which simply enforce similarity between multiple augmented views of the same image, CG-SSL accounts for the fact that these views may highlight different parts of an object or scene. To address this, our method establishes explicit correspondences between views and aligns the representations of meaningful image regions. At its core, CG-SSL augments standard SSL with a lightweight decoder that learns and refines concept tokens via cross-attention with patch features. The concept tokens are trained using masked concept distillation and a feature-space reconstruction objective. A final alignment stage enforces view consistency by geometrically matching concept regions under heavy augmentation, enabling more compact, robust, and disentangled representations of scene regions. Across multiple backbone sizes, CG-SSL achieves state-of-the-art results on image segmentation benchmarks using $k$-NN and linear probes, substantially outperforming prior methods and approaching, or even surpassing, the performance of leading SSL models trained on over $100\times$ more data. Code and pretrained models will be released.



Paperid:509
Authors:Hamish Flynn, Gergely Neu, Ludovic Schwartz
Abstract:
Many high-dimensional online decision-making problems can be modeled as stochastic sparse linear bandits. Most existing algorithms are designed to achieve optimal worst-case regret in either the data-rich regime, where polynomial dependence on the ambient dimension is unavoidable, or the data-poor regime, where dimension-independence is possible at the cost of worse dependence on the number of rounds. In contrast, the Bayesian approach of Information Directed Sampling (IDS) achieves the best of both worlds: a Bayesian regret bound that has the optimal rate in both regimes simultaneously. In this work, we explore the use of Sparse Optimistic Information Directed Sampling (SOIDS) to achieve the best of both worlds in the worst-case setting, without Bayesian assumptions. Through a novel analysis that enables the use of a time-dependent learning rate, we show that OIDS can be tuned without prior knowledge to optimally balance information and regret. Our results extend the theoretical guarantees of IDS, providing the first algorithm that simultaneously achieves optimal worst-case regret in both the data-rich and data-poor regimes. We empirically demonstrate the good performance of SOIDS.



Paperid:510
Authors:Sloan Nietert, Ziv Goldfeld
Abstract:
The optimal transport (OT) map is a geometry-driven transformation between high-dimensional probability distributions which underpins a wide range of tasks in statistics, applied probability, and machine learning. However, existing statistical theory for OT map estimation is quite restricted, hinging on Brenier's theorem (quadratic cost, absolutely continuous source) to guarantee existence and uniqueness of a deterministic OT map, on which various additional regularity assumptions are imposed to obtain quantitative error bounds. In many real‐world problems these conditions fail or cannot be certified, in which case optimal transportation is possible only via stochastic maps that can split mass. To broaden the scope of map estimation theory to such settings, this work introduces a novel metric for evaluating the transportation quality of stochastic maps. Under this metric, we develop computationally efficient map estimators with near-optimal finite-sample risk bounds, subject to easy-to-verify minimal assumptions. Our analysis further accommodates common forms of adversarial sample contamination, yielding estimators with robust estimation guarantees. Empirical experiments are provided which validate our theory and demonstrate the utility of the proposed framework in settings where existing theory fails. These contributions constitute the first general-purpose theory for map estimation, compatible with a wide spectrum of real-world applications where optimal transport may be intrinsically stochastic.



Paperid:511
Authors:Aleksandar Petrov, Shruti Agarwal, Philip Torr, Adel Bibi, John Collomosse
Abstract:
Watermarking, the practice of embedding imperceptible information into media such as images, videos, audio, and text, is essential for intellectual property protection, content provenance and attribution. The growing complexity of digital ecosystems necessitates watermarks for different uses to be embedded in the same media. However, to detect and decode all watermarks, they need to coexist well with one another. We perform the first study of coexistence of deep image watermarking methods and, contrary to intuition, we find that various open-source watermarks can coexist with only minor impacts on image quality and decoding robustness. The coexistence of watermarks also opens the avenue for ensembling watermarking methods. We show how ensembling can increase the overall message capacity and enable new trade-offs between capacity, accuracy, robustness and image quality, without needing to retrain the base models.



Paperid:512
Authors:Marshal Sinaga, Julien Martinelli, Samuel Kaski
Abstract:
Gaussian processes (GPs) are widely used for regression and optimization tasks such as Bayesian optimization (BO) due to their expressiveness and principled uncertainty estimates. However, in settings with large datasets corrupted by outliers, standard GPs and their sparse approximations struggle with computational tractability and robustness.We introduce Robust Computation-aware Gaussian Process (RCaGP), a novel GP model that jointly addresses these challenges by combining a principled treatment of approximation-induced uncertainty with robust generalized Bayesian updating. The key insight is that robustness and approximation-awareness are not orthogonal but intertwined: approximations can exacerbate the impact of outliers, and mitigating one without the other is insufficient.Unlike previous work that focuses narrowly on either robustness or approximation quality, RCaGP combines both in a principled and scalable framework, thus effectively managing both outliers and computational uncertainties introduced by approximations such as low-rank matrix multiplications.Our model ensures more conservative and reliable uncertainty estimates, a property we rigorously demonstrate. Additionally, we establish a robustness property and show that the mean function is key to preserving it, motivating a tailored model selection scheme for robust mean functions. Empirical results confirm that solving these challenges jointly leads to superior performance across both clean and outlier-contaminated settings, both on regression and high-throughput Bayesian optimization benchmarks.



Paperid:513
Authors:Fuxiang Zhang, Jiacheng Xu, Chaojie Wang, Ce Cui, Yang Liu, Bo An
Abstract:
Large Language Models (LLMs) have demonstrated remarkable progress in complex reasoning tasks through both post-training and test-time scaling laws. While prevalent test-time scaling approaches are often realized by using external reward models to guide the model generation process, we find only marginal gains can be acquired when scaling a model post-trained on specific reasoning tasks. We identify that the limited improvement stems from distribution discrepancies between the specific post-trained generator and the general reward model. To address this, we propose a framework that incentivizes LLMs to self-verify their own answers. By unifying answer generation and verification within a single reinforcement learning (RL) process, we train models that can effectively assess the correctness of their own solutions. The trained model can further scale its performance during inference time by verifying its generations, without the need for external verifiers. We train our self-verification models based on Qwen2.5-Math-7B and DeepSeek-R1-Distill-Qwen-1.5B, demonstrating its capabilities across varying reasoning context lengths. Experiments on multiple mathematical reasoning benchmarks show that our models can not only improve post-training performance but also enable effective test-time scaling.



Paperid:514
Authors:Ilias Diakonikolas, Mingchen Ma, Daniel Kane
Abstract:
Abstract:We study the task of agnostically learning general (as opposed to homogeneous) ReLUs under the Gaussian distribution with respect to the squared loss. In the passive learning setting, recent work gave a computationally efficient algorithm that uses $poly(d,1/\epsilon)$ labeled examples and outputs a hypothesis with error $O(opt)+\epsilon$, where $opt$ is the squared loss of the best fit ReLU.Here we focus on the interactive setting, where the learner has some form of query access to the labels of unlabeled examples. Our main result is the first computationally efficient learner that uses $d polylog(1/\epsilon)+\tilde{O}(\min\{1/p, 1/\epsilon\})$ black-box label queries, where $p$ is the bias of the target function, and achieves error $O(opt)+\epsilon$.We complement our algorithmic result by showing that its query complexity bound is qualitatively near-optimal, even ignoring computational constraints. Finally, we establish that query access is essentially necessary to improve on the label complexity of passive learning. Specifically, for pool-based active learning, any active learner requires $\tilde{\Omega}(d/\epsilon)$ labels, unless it draws a super-polynomial number of unlabeled examples.



Paperid:515
Authors:Zhiqun Zuo, Mahdi Khalili
Abstract:
Strategic classification, where individuals modify their features to influence machine learning (ML) decisions, presents critical fairness challenges. While group fairness in this setting has been widely studied, individual fairness remains underexplored. We analyze threshold-based classifiers and prove that deterministic thresholds violate individual fairness. Then, we investigate the possibility of using a randomized classifier to achieve individual fairness. We introduce conditions under which a randomized classifier ensures individual fairness and leverage these conditions to find an optimal and individually fair randomized classifier through a linear programming problem. Additionally, we demonstrate that our approach can be extended to group fairness notions. Experiments on real-world datasets confirm that our method effectively mitigates unfairness and improves the fairness-accuracy trade-off.



Paperid:516
Authors:Zicheng Zhang, Haoran Li, Yifeng Zhang, Guoqiang Gong, Jiaxing Wang, Pengzhang Liu, Qixia Jiang, Junxing Hu
Abstract:
Low-Rank Adaptation (LoRA) offers a parameter-efficient paradigm for tuning large models. While recent spectral initialization methods improve convergence and performance over the naive “Noise \& Zeros” scheme, their extra computational and storage overhead undermines efficiency. In this paper, we establish update magnitude as the fundamental driver of LoRA performance and propose LoRAM, a magnitude-driven “Basis \& Basis” initialization scheme that matches spectral methods without their inefficiencies. Our key contributions are threefold: (i) Magnitude of weight updates determines convergence.We prove low-rank structures intrinsically bound update magnitudes, unifying hyperparameter tuning in learning rate, scaling factor, and initialization as mechanisms to optimize magnitude regulation. (ii) Spectral initialization succeeds via magnitude amplification. We demystify that the presumed knowledge-driven benefit of spectral component essentially arises from the boost in the weight update magnitude.(iii) A novel and compact initialization strategy, LoRAM, scales deterministic orthogonal bases using pretrained weight magnitudes to simulate spectral gains. Extensive experiments show that LoRAM serves as a strong baseline, retaining the full efficiency of LoRA while matching or outperforming spectral initialization across benchmarks.



Paperid:517
Authors:Róbert Csordás, Christopher D Manning, Christopher Potts
Abstract:
Modern LLMs are increasingly deep, and depth correlates with performance, albeit with diminishing returns. However, do these models use their depth efficiently? Do they compose more features to create higher-order computations that are impossible in shallow models, or do they merely spread the same kinds of computation out over more layers? To address these questions, we analyze the residual stream of the Llama 3.1 family of models. We find: First, comparing the output of the sublayers to the residual stream reveals that layers in the second half contribute much less than those in the first half, with a clear phase transition between the two halves. Second, skipping layers in the second half has a much smaller effect on future computations and output predictions. Third, for multihop tasks, we are unable to find evidence that models are using increased depth to compose subresults in examples involving many hops. Fourth, we seek to directly address whether deeper models are using their additional layers to perform new kinds of computation. To do this, we train linear maps from the residual stream of a shallow model to a deeper one. We find that layers with the same relative depth map best to each other, suggesting that the larger model simply spreads the same computations out over its many layers. All this evidence suggests that deeper models are not using their depth to learn new kinds of computation, but only using the greater depth to perform more fine-grained adjustments to the residual. This may help explain why increasing scale leads to diminishing returns for stacked Transformer architectures.



Paperid:518
Authors:Fares El Khoury, Edouard Pauwels, Samuel Vaiter, Michael Arbel
Abstract:
Bilevel optimization has emerged as a technique for addressing a wide range of machine learning problems that involve an outer objective implicitly determined by the minimizer of an inner problem. While prior works have primarily focused on the parametric setting, a learning-theoretic foundation for bilevel optimization in the nonparametric case remains relatively unexplored. In this paper, we take a first step toward bridging this gap by studying Kernel Bilevel Optimization (KBO), where the inner objective is optimized over a reproducing kernel Hilbert space. This setting enables rich function approximation while providing a foundation for rigorous theoretical analysis. In this context, we derive novel finite-sample generalization bounds for KBO, leveraging tools from empirical process theory. These bounds further allow us to assess the statistical accuracy of gradient-based methods applied to the empirical discretization of KBO. We numerically illustrate our theoretical findings on a synthetic instrumental variable regression task.



Paperid:519
Authors:Tarun Suresh, Debangshu Banerjee, Shubham Ugare, Sasa Misailovic, Gagandeep Singh
Abstract:
Abstract:Diffusion LLMs have emerged as a promising alternative to conventional autoregressive LLMs, offering substantial potential for improving runtime efficiency. However, existing diffusion models fail to provably enforce user-specified formal constraints, such as regular expressions, which makes them unreliable for tasks that require structured outputs, such as fixed-schema JSON generation. Unlike autoregressive models, which generate tokens sequentially, diffusion LLMs predict a block of tokens in parallel. This parallelism makes traditional constrained decoding algorithms, designed to enforce constraints with sequential token prediction, ineffective at preserving the true output distribution. To address this limitation, we propose DINGO, a dynamic programming-based constrained decoding strategy that is both efficient and provably distribution-preserving. DINGO enables sampling of output strings with the highest probability under the model’s predicted distribution while strictly adhering to any user-specified regular expression. On standard symbolic math and JSON generation benchmarks, DINGO achieves up to $68$% points of improvement over unconstrained inference.



Paperid:520
Authors:Sam Buchanan, Druv Pai, Yi Ma, Valentin De Bortoli
Abstract:
When do diffusion models reproduce their training data, and when are they able to generate samples beyond it? A practically relevant theoretical understanding of this interplay between memorization and generalization may significantly impact real-world deployments of diffusion models with respect to issues such as copyright infringement and data privacy. In this paper, we propose a scientific andmathematical “laboratory” for investigating memorization and generalization in diffusion models trained on fully synthetic or natural image-like structured data. Within this setting, we theoretically characterize a crossover point wherein the weighted training loss of a fully generalizing model becomes greater than that of an underparameterized memorizing model at a critical value of model (under)parameterization. We then demonstrate via carefully-designed experiments that the location of this crossover predicts a phase transition in diffusion models trained via gradient descent, as our theory enables us to analytically predict the model size at which memorization becomes predominant. Our work provides an analytically tractable and practically meaningful setting for future theoretical and empirical investigations. Code for our experiments will be made available online.



Paperid:521
Authors:Hai Duong, ThanhVu Nguyen, Matthew Dwyer
Abstract:
Deep Neural Networks (DNN) have emerged as an effective approach to implementing challenging sub-problems. They are increasingly being used as components in critical transportation, medical, and military systems. However, like human-written software, DNNs may have flaws that can lead to unsafe system performance. To confidently deploy DNNs in such systems, strong evidence is needed that they do not contain such flaws. This has led researchers to explore the adaptation and customization of software verification approaches to the problem of neural network verification (NNV).Many dozens of NNV tools have been developed in recent years and as a field these techniques have matured to the point where realistic networks can be analyzed to detect flaws and to prove conformance with specifications. NNV tools are highly-engineered and complex may harbor flaws that cause them to produce unsound results.We identify commonalities in the algorithmic approaches taken by NNV tools to define a verifier independent proof format -- activation pattern tree proofs (APTP) -- and design an algorithm for checking those proofs that is proven correct and optimized to enable scalable checking. We demonstrate that existing verifiers can efficiently generate APTP proofs, and that an APTPchecker significantly outperforms prior work on a benchmark of 16 neural networks and 400 NNV problems, and that it is robust to variation in APTP proof structure arising from different NNV tools.APTPchecker is available at: https://anonymous.4open.science/r/APTPchecker-0482/.



Paperid:522
Authors:Xihong Su, Jia Lin Hau, Gersi Doko, Kishan Panaganti, Marek Petrik
Abstract:
Risk-averse total-reward Markov Decision Processes (MDPs) offer a promising framework for modeling and solving undiscounted infinite-horizon objectives. Existing model-based algorithms for risk measures like the entropic risk measure (ERM) and entropic value-at-risk (EVaR) are effective in small problems, but require full access to transition probabilities. We propose a Q-learning algorithm to compute the optimal stationary policy for total-reward ERM and EVaR objectives with strong convergence and performance guarantees. The algorithm and its optimality are made possible by ERM's dynamic consistency and elicitability. Our numerical results on tabular domains demonstrate quick and reliable convergence of the proposed Q-learning algorithm to the optimal risk-averse value function.



Paperid:523
Authors:Mingyang Fu, Yuyang Peng, Dongping Chen, Zetong Zhou, Benlin Liu, Yao Wan, Zhou Zhao, Philip S Yu, Ranjay Krishna
Abstract:
The visual world around us constantly evolves, from real-time news and social media trends to global infrastructure changes visible through satellite imagery and augmented reality enhancements. However, Multimodal Large Language Models (MLLMs), which automate many tasks, struggle to stay current, limited by the cutoff dates in their fixed training datasets.To quantify this stagnation, we introduce LiveVQA, the first-of-its-kind dataset featuring 107,143 samples and 12 categories data specifically designed to support research in both seeking and updating with live visual knowledge.Drawing from recent news articles, video platforms, and academic publications in April 2024-May 2025, LiveVQA enables evaluation of how models handle latest visual information beyond their knowledge boundaries and how current methods help to update them. Our comprehensive benchmarking of 17 state-of-the-art MLLMs reveals significant performance gaps on content beyond knowledge cutoff, and tool-use or agentic visual seeking framework drastically gain an average of 327% improvement. Furthermore, we explore parameter-efficient fine-tuning methods to update MLLMs with new visual knowledge.We dive deeply to the critical balance between adapter capacity and model capability when updating MLLMs with new visual knowledge. All the experimental dataset and source code are publicly available at: https://livevqa.github.io.



Paperid:524
Authors:Ben Jacobsen, Kassem Fawaz
Abstract:
Abstract:We study the problem of differentially private continual counting in the unbounded setting where the input size $n$ is not known in advance. Current state-of-the-art algorithms based on optimal instantiations of the matrix mechanism cannot be directly applied here because their privacy guarantees only hold when key parameters are tuned to $n$. Using the common `doubling trick' avoids knowledge of $n$ but leads to suboptimal and non-smooth error. We solve this problem by introducing novel matrix factorizations based on logarithmic perturbations of the function $\frac{1}{\sqrt{1-z}}$ studied in prior works, which may be of independent interest. The resulting algorithm has smooth error, and for any $\alpha > 0$ and $t\leq n$ it is able to privately estimate the sum of the first $t$ data points with $O(\log^{2+2\alpha}(t))$ variance. It requires $O(t)$ space and amortized $O(\log t)$ time per round, compared to $O(\log(n)\log(t))$ variance, $O(n)$ space and $O(n \log n)$ pre-processing time for the nearly-optimal bounded-input algorithm of Henzinger et al. Empirically, we find that our algorithm's performance is also comparable to theirs in absolute terms: our variance is less than $1.5\times$ theirs for $t$ as large as $2^{24}$.



Paperid:525
Authors:Yujia Hu, Songhua Liu, Zhenxiong Tan, Xingyi Yang, Xinchao Wang
Abstract:
While diffusion models have achieved remarkable success in text-to-image generation, they encounter significant challenges with instruction-driven image editing. Our research highlights a key challenge: these models particularly struggle with structurally-inconsistent edits that involve substantial layout changes. To address this gap, we introduce Image Editing As Programs (IEAP), a unified image editing framework built upon the Diffusion Transformer (DiT) architecture. Specifically, IEAP deals with complex instructions by decomposing them into a sequence of programmable atomic operations. Each atomic operation manages a specific type of structurally consistent edit; when sequentially combined, IEAP enables the execution of arbitrary, structurally-inconsistent transformations. This reductionist approach enables IEAP to robustly handle a wide spectrum of edits, encompassing both structurally-consistent and -inconsistent changes. Extensive experiments demonstrate that IEAP significantly outperforms state-of-the-art methods on standard benchmarks across various editing scenarios. In these evaluations, our framework delivers superior accuracy and semantic fidelity, particularly for complex, multi-step instructions.



Paperid:526
Authors:Yasha Ektefaie, Andrew Shen, Lavik Jain, Maha Farhat, Marinka Zitnik
Abstract:
Protein language models (PLMs) are commonly assumed to capture evolutionary information by training on large protein sequence datasets. However, it remains unclear whether PLMs can reason about evolution—that is, infer evolutionary relationships between protein sequences. To test this capability, we introduce a benchmark for evolutionary reasoning and find that existing PLMs consistently fail to recover phylogenetic structure, despite strong performance on standard tasks such as masked token prediction and contact prediction. To address this limitation, we present Phyla. Phyla introduces a hybrid state-space and transformer architecture that jointly process multiple sequences and is trained using a tree-based objective over 3,000 phylogenies spanning diverse protein families. Phyla achieves state-of-the-art performance in evolutionary reasoning, outperforming the next-best model by 13\% on tree reconstruction and 10\% on taxonomic clustering. Beyond synthetic benchmarks, Phyla applies to real-world settings: it reconstructs biologically accurate branches of the tree of life and infers whole-genome evolutionary relationships from Mycobacterium tuberculosis genomes. These findings suggest that evolutionary reasoning is not an emergent property of large-scale sequence modeling. Instead, Phyla shows that models trained with phylogenetic supervision can reason about evolution more effectively, offering a biologically grounded path toward evolutionary foundation models.



Paperid:527
Authors:Shuai Wang, Malu Zhang, Jingya Wang, Dehao Zhang, Yimeng Shan, Jieyuan (Eric) Zhang, Yichen Xiao, Honglin Cao, Haonan Zhang, Zeyu Ma, Yang Yang, Haizhou Li
Abstract:
Harnessing the event-driven computation paradigm, Spiking Neural Networks present a promising avenue toward energy-efficient Transformer architectures. However, existing Spiking Transformers still suffer significant performance gaps compared to their artificial neural network counterparts. Through comprehensive analysis, we attribute this gap to these two factors. First, the binary nature of spike trains limits Spiking Self-attention (SSA)’s capacity to capture negative–negative and positive–negative membrane potential interactions on Querys and Keys. Second, SSA typically omits Softmax functions to avoid energy-intensive multiply-accumulate operations, thereby failing to maintain row-stochasticity constraints on attention scores.To address these issues, we propose a Bipolar Self-Attention (BSA) paradigm, effectively modeling multi-polar membrane potential interactions with a fully spike-driven characteristic. Specifically, we demonstrate that ternary matrix multiplication provides a closer approximation to real-valued computation on both distribution and local correlation, enabling clear differentiation between homopolar and heteropolar interactions. Moreover, we propose a shift-based Softmax approximation named Shiftmax, which efficiently achieves low-entropy activation and partly maintains row-stochasticity without non-linear operation, enabling precise attention allocation. Extensive experiments show that BSA achieves substantial performance improvements across various tasks, including image classification, semantic segmentation, and event-based tracking. These results establish its potential as a fundamental building block for energy-efficient Spiking Transformers.



Paperid:528
Authors:Elliot Epstein, Rajat Vadiraj Dwaraknath, Thanawat Sornwanee, John Winnicki, Jerry Liu
Abstract:
We propose a method for density estimation that leverages an estimated score function to debias kernel density estimation (SD-KDE).In our approach, each data point is adjusted by taking a single step along the score function with a specific choice of step size, followed by standard KDE with a modified bandwidth.The step size and modified bandwidth are chosen to remove the leading order bias in the KDE, improving the asymptotic convergence rate.Our experiments on synthetic tasks in 1D, 2D and on MNIST, demonstrate that our proposed SD-KDE method significantly reduces the mean integrated squared error compared to the standard Silverman KDE, even with noisy estimates in the score function.These results underscore the potential of integrating score-based corrections into nonparametric density estimation.



Paperid:529
Authors:Kealan Dunnett, Reza Arablouei, Volkan Dedeoglu, Dimity Miller, Raja Jurdak
Abstract:
Model pruning has gained traction as a promising defense strategy against backdoor attacks in deep learning. However, existing pruning-based approaches often fall short in accurately identifying and removing the specific parameters responsible for inducing backdoor behaviors. Despite the dominance of fine-tuning-based defenses in recent literature, largely due to their superior performance, pruning remains a compelling alternative, offering greater interpretability and improved robustness in low-data regimes. In this paper, we propose a novel pruning approach featuring a learned \emph{selection} mechanism to identify parameters critical to both main and backdoor tasks, along with an \emph{invertible} pruning mask designed to simultaneously achieve two complementary goals: eliminating the backdoor task while preserving it through the inverse mask. We formulate this as a bi-level optimization problem that jointly learns selection variables, a sparse invertible mask, and sample-specific backdoor perturbations derived from clean data. The inner problem synthesizes candidate triggers using the inverse mask, while the outer problem refines the mask to suppress backdoor behavior without impairing clean-task accuracy. Extensive experiments demonstrate that our approach outperforms existing pruning-based backdoor mitigation approaches, maintains strong performance under limited data conditions, and achieves competitive results compared to state-of-the-art fine-tuning approaches. Notably, the proposed approach is particularly effective in restoring correct predictions for compromised samples after successful backdoor mitigation.



Paperid:530
Authors:Benedict Clark, Rick Wilming, Hjalmar Schulz, Rustam Zhumagambetov, Danny Panknin, Stefan Haufe
Abstract:
Correct model interpretation in high-stakes settings is critical, yet both post-hoc explanation methods and so-called intrinsically interpretable models can systematically attribute false-positive importance to non-informative features such as suppressor variables (Wilming et al., 2023).Specifically, both linear models and their powerful non-linear generalisation such as Neural Additive Models (NAMs, Agarwal et al., 2021) are susceptible to spurious attributions to suppressors. We present a principled generalisation of activation patterns -- originally developed to make linear models truly interpretable -- to additive models, introducing a covariance-aware explanation framework that correctly disentangles suppressor effects.This yields PatternGAM, an activation-based explanation method for the broad range of additive models that also includes NAMs, and PatternQLR for polynomial logistic regression models. These provide meaningful and reliable attributions to truly informative features in the presence of complex input dependencies, and especially in scenarios where pairwise interaction terms are required. Empirical evaluations on the XAI-TRIS benchmark (Clark et al., 2024) with a novel false-negative invariant formulation of the earth mover's distance metric demonstrates significant improvements over popular XAI methods and the traditional interpretation of additive models. Finally, real-world case studies on the COMPAS (ProPublica, 2016) and MIMIC-IV (Johnson et al., 2023b) datasets confirm that our approach aligns with domain knowledge and provides new insights into the role of specific features by disentangling genuine target-related information from suppressor effects.



Paperid:531
Authors:Yifan Zhu, Lijia Yu, Xiao-Shan Gao
Abstract:
Abstract:In recent years, data poisoning attacks have been increasingly designed to appear harmless and even beneficial, often with the intention of verifying dataset ownership or safeguarding private data from unauthorized use. However, these developments have the potential to cause misunderstandings and conflicts, as data poisoning has traditionally been regarded as a security threat to machine learning systems. To address this issue, it is imperative for harmless poisoning generators to claim ownership of their generated datasets, enabling users to identify potential poisoning to prevent misuse. In this paper, we propose the deployment of watermarking schemes as a solution to this challenge. We introduce two provable and practical watermarking approaches for data poisoning: post-poisoning watermarking and poisoning-concurrent watermarking. Our analyses demonstrate that when the watermarking length is $\Theta(\sqrt{d}/\epsilon_w)$ for post-poisoning watermarking, and falls within the range of $\Theta(1/\epsilon_w^2)$ to $O(\sqrt{d}/\epsilon_p)$ for poisoning-concurrent watermarking, the watermarked poisoning dataset provably ensures both watermarking detectability and poisoning utility, certifying the practicality of watermarking under data poisoning attacks. We validate our theoretical findings through experiments on several attacks, models and datasets.



Paperid:532
Authors:Myeongseob Ko, Hoang Anh Just, Charles Fleming, Ming Jin, Ruoxi Jia
Abstract:
Machine unlearning has emerged as a prevalent technical solution for selectively removing unwanted knowledge absorbed during pre-training, without requiring full retraining. While recent unlearning techniques can effectively remove undesirable content without severely compromising performance on standard benchmarks, we find that they may inadvertently create ``knowledge holes''---unintended losses of benign knowledge that standard benchmarks fail to capture. To probe where unlearned models reveal knowledge holes, we propose a test case generation framework that explores both immediate neighbors of unlearned content and broader areas of potential failures.Our evaluation demonstrates significant hidden costs of unlearning: up to 98.7\% of the test cases yield irrelevant or nonsensical responses from unlearned models, despite being answerable by the pretrained model. These findings necessitate rethinking the conventional approach to evaluating knowledge preservation in unlearning, moving beyond standard, static benchmarks.



Paperid:533
Authors:Zhiqin Zhang, Yining Ma, Zhiguang Cao, Hoong Chuin Lau
Abstract:
Neural combinatorial optimization (NCO) has achieved remarkable performance, yet its learned model representations and decision rationale remain a black box. This impedes both academic research and practical deployment, since researchers and stakeholders require deeper insights into NCO models. In this paper, we take the first critical step towards interpreting NCO models by investigating their representations through various probing tasks. Moreover, we introduce a novel probing tool named Coefficient Significance Probing (CS-Probing) to enable deeper analysis of NCO representations by examining the coefficients and statistical significance during probing. Extensive experiments and analysis reveal that NCO models encode low-level information essential for solution construction, while capturing high-level knowledge to facilitate better decisions. Using CS-Probing, we find that prevalent NCO models impose varying inductive biases on their learned representations, uncover direct evidence related to model generalization, and identify key embedding dimensions associated with specific knowledge. Our work represents the first systematic effort to interpret black-box NCO models, showcasing probing as a valuable tool for analyzing their internal mechanisms and revealing impactful insights for the NCO community.



Paperid:534
Authors:Yi Fan, Yu-Bin Yang
Abstract:
The Mamba-type neural networks have gained significant popularity recently. To effectively and efficiently establish model architectures of Mamba, it is natural to introduce Neural Architecture Search (NAS) methods into Mamba. However, existing NAS methods tailored for Mamba are training-based, leading to substantial time and computational resource expenditure. To address this issue, and considering that Mamba2 is an improved version of the original Mamba, we propose a training-free NAS method specifically designed for Mamba2. Based on rank collapse in stacked State Space Duality (SSD) blocks, we design a proxy that only requires the computation of the transformation matrix and its gradient between two tensors within the network. Additionally, we develop a corresponding search space and introduce a novel approach for determining adjustable hyperparameter ranges. Experimental results show that our method outperforms all existing training-free NAS approaches in terms of both ranking correlation and the performance of search results for Mamba2 architecture. To the best of our knowledge, this is the first training-free NAS method designed for Mamba-type architectures.



Paperid:535
Authors:Chenxuan Miao, Yutong Feng, Jianshu Zeng, Zixiang Gao, Liu Hantang, Yunfeng Yan, Donglian Qi, Xi Chen, Bin Wang, Hengshuang Zhao
Abstract:
Video object removal has achieved advanced performance due to the recent success of video generative models. However, when addressing the side effects of objects, \textit{e.g.,} their shadows and reflections, existing works struggle to eliminate these effects for the scarcity of paired video data as supervision. This paper presents \method, termed \textbf{R}emove \textbf{O}bjects with \textbf{S}ide \textbf{E}ffects, a framework that systematically studies the object's effects on environment, which can be categorized into five common cases: shadows, reflections, light, translucency and mirror. Given the challenges of curating paired videos exhibiting the aforementioned effects, we leverage a 3D rendering engine for synthetic data generation. We carefully construct a fully-automatic pipeline for data preparation, which simulates a large-scale paired dataset with diverse scenes, objects, shooting angles, and camera trajectories. ROSE is implemented as an video inpainting model built on diffusion transformer. To localize all object-correlated areas, the entire video is fed into the model for reference-based erasing. Moreover, additional supervision is introduced to explicitly predict the areas affected by side effects, which can be revealed through the differential mask between the paired videos. To fully investigate the model performance on various side effect removal, we presents a new benchmark, dubbed ROSE-Bench, incorporating both common scenarios and the five special side effects for comprehensive evaluation. Experimental results demonstrate that \method achieves superior performance compared to existing video object erasing models and generalizes well to real-world video scenarios.



Paperid:536
Authors:Taekyung Kim, Dongyoon Han, Byeongho Heo, Jeongeun Park, Sangdoo Yun
Abstract:
Deriving compact and temporally aware visual representations from dynamic scenes is essential for successful execution of sequential scene understanding tasks such as visual tracking and robotic manipulation. In this paper, we introduce Token Bottleneck (ToBo), a simple yet intuitive self-supervised learning pipeline that squeezes a scene into a bottleneck token and predicts the subsequent scene using minimal patches as hints. The ToBo pipeline facilitates the learning of sequential scene representations by conservatively encoding the reference scene into a compact bottleneck token during the squeeze step. In the expansion step, we guide the model to capture temporal dynamics by predicting the target scene using the bottleneck token along with few target patches as hints. This design encourages the vision backbone to embed temporal dependencies, thereby enabling understanding of dynamic transitions across scenes. Extensive experiments in diverse sequential tasks, including video label propagation and robot manipulation in simulated environments demonstrate the superiority of ToBo over baselines. Moreover, deploying our pre-trained model on physical robots confirms its robustness and effectiveness in real-world environments. We further validate the scalability of ToBo across different model scales.



Paperid:537
Authors:Weikai Wang, Erick Delage
Abstract:
For continuing tasks, average cost Markov decision processes have well-documented value and can be solved using efficient algorithms. However, it explicitly assumes that the agent is risk-neutral. In this work, we extend risk-neutral algorithms to accommodate the more general class of dynamic risk measures. Specifically, we propose a relative value iteration (RVI) algorithm for planning and design two model-free Q-learning algorithms, namely a generic algorithm based on the multi-level Monte Carlo (MLMC) method, and an off-policy algorithm dedicated to utility-base shortfall risk measures. Both the RVI and MLMC-based Q-learning algorithms are proven to converge to optimality. Numerical experiments validate our analysis, confirm empirically the convergence of the off-policy algorithm, and demonstrate that our approach enables the identification of policies that are finely tuned to the intricate risk-awareness of the agent that they serve.



Paperid:538
Authors:Jincheng Zhou, Mengbo Wang, Anqi He, Yumeng Zhou, Hessam Olya, Murat Kocaoglu, Bruno Ribeiro
Abstract:
Abstract:Causal discovery from observational data is a fundamental task in artificial intelligence, with far-reaching implications for decision-making, predictions, and interventions. Despite significant advances, existing methods can be broadly categorized as constraint-based or score-based approaches. Constraint-based methods offer rigorous causal discovery but are often hindered by small sample sizes, while gradient-based methods provide flexible optimization but typically forgo explicit conditional independence testing. This work explores a third avenue: develop differentiable $d$-separation scores, obtained through a percolation theory using soft logic. This enables the implementation of a new type of causal discovery method: gradient-based optimization of conditional independence constraints. Empirical evaluations demonstrate the robust performance of our approach in low-sample regimes, surpassing traditional constraint-based and score-based baselines on a real-world dataset.



Paperid:539
Authors:Jacob Block, Sundararajan Srinivasan, Liam Collins, Aryan Mokhtari, Sanjay Shakkottai
Abstract:
The power of foundation models (FMs) lies in their capacity to learn highly expressive representations that can be adapted to a broad spectrum of tasks. However, these pretrained models require additional training stages to become effective for downstream applications. In the multi-task setting, prior works have shown empirically that specific meta-learning approaches for preparing a model for future adaption through parameter-efficient fine-tuning (PEFT) can outperform standard retraining methods, but the mechanism of the benefits of meta-learning has been largely unexplored. We introduce a framework for generic PEFT-based meta-learning to learn a model that can easily adapt to unseen tasks. For linear models using LoRA, we show that standard retraining is provably suboptimal for finding an adaptable set of parameters and provide strict performance guarantees for our proposed method. We verify these theoretical insights through experiments on synthetic data as well as real-data vision and language tasks. We observe significant performance benefits using a simple implementation of our proposed meta-learning scheme during retraining relative to the conventional approach.



Paperid:540
Authors:Jiawei Liu, Nirav Diwan, Zhe Wang, Haoyu Zhai, Xiaona Zhou, Kiet Nguyen, Tianjiao Yu, Muntasir Wahed, Yinlin Deng, Hadjer Benkraouda, Yuxiang Wei, LINGMING ZHANG, Ismini Lourentzou, Gang Wang
Abstract:
We introduce PurpCode-R1, a novel post-training recipe that aligns coding assistants to perform safety reasoning to defend against malicious cyber activities and provide secure and functional code. PurpCode-R1 trains a reasoning model in two stages: (i) Rule Learning, which explicitly teaches the model to reference cyber safety rules to avoid facilitating malicious cyber events and to generate vulnerability-free code; and (ii) Reinforcement Learning, which optimizes model utility and safety through multi-objective diverse reward mechanisms. To empower the training pipelines with safety data, we perform internal red-teaming to synthesize comprehensive and high-coverage prompts based on real-world tasks for inducing unsafe cyber activities in the model. Based on PurpCode-R1, we develop PurpCode-R1-32B, which outperforms existing SOTA models in cyber safety, while preserving utilities in code generation and basic security knowledge.



Paperid:541
Authors:Yu Shang, Xin Zhang, Yinzhou Tang, Lei Jin, Chen Gao, Wei Wu, Yong Li
Abstract:
World models have become indispensable tools for embodied intelligence, serving as powerful simulators capable of generating realistic robotic videos while addressing critical data scarcity challenges. However, current embodied world models exhibit limited physical awareness, particularly in modeling 3D geometry and motion dynamics, resulting in unrealistic video generation for contact-rich robotic scenarios. In this paper, we present RoboScape, a unified physics-informed world model that jointly learns RGB video generation and physics knowledge within an integrated framework. We introduce two key physics-informed joint training tasks: temporal depth prediction that enhances 3D geometric consistency in video rendering, and keypoint dynamics learning that implicitly encodes physical properties (e.g., object shape and material characteristics) while improving complex motion modeling. Extensive experiments demonstrate that RoboScape generates videos with superior visual fidelity and physical plausibility across diverse robotic scenarios. We further validate its practical utility through downstream applications including robotic policy training with synthetic data and policy evaluation. We hope this work provides new insights for building efficient physics-informed world models to advance embodied intelligence research. Our code is available at: https://anonymous.4open.science/r/RoboScape-3652.



Paperid:542
Authors:Hugo Negrel, Florentin Coeurdoux, Michael Albergo, Eric Vanden-Eijnden
Abstract:
We propose a framework for learning maps between probability distributions that broadly generalizes the time dynamics of flow and diffusion models. To enable this, we generalize stochastic interpolants by replacing the scalar time variable with vectors, matrices, or linear operators, allowing us to bridge probability distributions across multiple dimensional spaces. This approach enables the construction of versatile generative models capable of fulfilling multiple tasks without task-specific training. Our operator-based interpolants not only provide a unifying theoretical perspective for existing generative models but also extend their capabilities. Through numerical experiments, we demonstrate the zero-shot efficacy of our method on conditional generation and inpainting, fine-tuning and posterior sampling, and multiscale modeling, suggesting its potential as a generic task-agnostic alternative to specialized models.



Paperid:543
Authors:Khiem HUYNH, Malcolm Egan, Giovanni Neglia, Jean-Marie Gorce
Abstract:
Federated learning (FL) is now recognized as a key framework for communication-efficient collaborative learning. Most theoretical and empirical studies, however, rely on the assumption that clients have access to pre-collected data sets, with limited investigation into scenarios where clients continuously collect data. In many real-world applications, particularly when data is generated by physical or biological processes, client data streams are often modeled by non-stationary Markov processes. Unlike standard i.i.d. sampling, the performance of FL with Markovian data streams remains poorly understood due to the statistical dependencies between client samples over time. In this paper, we investigate whether FL can still support collaborative learning with Markovian data streams. Specifically, we analyze the performance of Minibatch SGD, Local SGD, and a variant of Local SGD with momentum. We answer affirmatively under standard assumptions and smooth non-convex client objectives: the sample complexity is proportional to the inverse of the number of clients with a communication complexity comparable to the i.i.d. scenario. However, the sample complexity for Markovian data streams remains higher than for i.i.d. sampling. Our analysis is validated via experiments with real pollution monitoring time series data.



Paperid:544
Authors:Javad Rajabi, Soroush Mehraban, Seyedmorteza Sadat, Babak Taati
Abstract:
Classifier-free guidance (CFG) has become an essential component of modern diffusion models to enhance both generation quality and alignment with input conditions. However, CFG requires specific training procedures and is limited to conditional generation. To address these limitations, we propose Token Perturbation Guidance (TPG), a novel method that applies perturbation matrices directly to intermediate token representations within the diffusion network. TPG employs a norm-preserving shuffling operation to provide effective and stable guidance signals that improve generation quality without architectural changes. As a result, TPG is training-free and agnostic to input conditions, making it readily applicable to both conditional and unconditional generation. We also analyze the guidance term provided by TPG and show that its effect on sampling more closely resembles CFG compared to existing training-free guidance techniques. We extensively evaluate TPG on SDXL and Stable Diffusion 2.1, demonstrating nearly a 2x improvement in FID for unconditional generation over the SDXL baseline and showing that TPG closely matches CFG in prompt alignment. Thus, TPG represents a general, condition-agnostic guidance method that extends CFG-like benefits to a broader class of diffusion models.



Paperid:545
Authors:Andrew Draganov, Pascal Weber, Rasmus Jørgensen, Anna Beer, Claudia Plant, Ira Assent
Abstract:
Abstract:Hierarchical clustering is a powerful tool for exploratory data analysis, organizing data into a tree of clusterings from which a partition can be chosen. This paper generalizes these ideas by proving that, for any reasonable hierarchy, one can optimally solve any center-based clustering objective over it (such as $k$-means). Moreover, these solutions can be found exceedingly quickly and are *themselves* necessarily hierarchical. Thus, given a cluster tree, we show that one can quickly access a plethora of new, equally meaningful hierarchies.Just as in standard hierarchical clustering, one can then choose any desired partition from these new hierarchies. We conclude by verifying the utility of our proposed techniques across datasets, hierarchies, and partitioning schemes.



Paperid:546
Authors:Jingwen Liu, Hantao Yu, Clayton Sanford, Alexandr Andoni, Daniel Hsu
Abstract:
Abstract:Transformers have the representational capacity to simulate Massively Parallel Computation (MPC) algorithms, but they suffer from quadratic time complexity, which severely limits their scalability. We introduce an efficient attention mechanism called Approximate Nearest Neighbor Attention (ANNA) with sub-quadratic time complexity. We prove that ANNA-transformers (1) retain the expressive power previously established for standard attention in terms of matching the capabilities of MPC algorithms, and (2) can solve key reasoning tasks such as Match2 and $k$-hop with near-optimal depth. Using the MPC framework, we further prove that constant-depth ANNA-transformers can simulate constant-depth low-rank transformers, thereby providing a unified way to reason about a broad class of efficient attention approximations.



Paperid:547
Authors:Nika Haghtalab, Omar Montasser, Mingda Qiao
Abstract:
Abstract:We study the tradeoff between sample complexity and round complexity in *on-demand sampling*, where the learning algorithm adaptively samples from $k$ distributions over a limited number of rounds. In the realizable setting of Multi-Distribution Learning (MDL), we show that the optimal sample complexity of an $r$-round algorithm scales approximately as $dk^{\Theta(1/r)} / \epsilon$. For the general agnostic case, we present an algorithm that achieves near-optimal sample complexity of $\widetilde O((d + k) / \epsilon^2)$ within $\widetilde O(\sqrt{k})$ rounds. Of independent interest, we introduce a new framework, Optimization via On-Demand Sampling (OODS), which abstracts the sample-adaptivity tradeoff and captures most existing MDL algorithms. We establish nearly tight bounds on the round complexity in the OODS setting. The upper bounds directly yield the $\widetilde O(\sqrt{k})$-round algorithm for agnostic MDL, while the lower bounds imply that achieving sub-polynomial round complexity would require fundamentally new techniques that bypass the inherent hardness of OODS.



Paperid:548
Authors:Felix Dangel, Tim Siebert, Marius Zeinhofer, Andrea Walther
Abstract:
Computing partial differential equation (PDE) operators via nested backpropagation is expensive, yet popular, and severely restricts their utility for scientific machine learning. Recent advances, like the forward Laplacian and randomizing Taylor mode automatic differentiation (AD), propose forward schemes to address this. We introduce an optimization technique for Taylor mode that 'collapses' derivatives by rewriting the computational graph, and demonstrate how to apply it to general linear PDE operators, and randomized Taylor mode. The modifications simply require propagating a sum up the computational graph, which could---or should---be done by a machine learning compiler, without exposing complexity to users. We implement our collapsing procedure and evaluate it on popular PDE operators, confirming it accelerates Taylor mode and outperforms nested backpropagation.



Paperid:549
Authors:Matthew T Jackson, Uljad Berdica, Jarek Liesen, Shimon Whiteson, Jakob Foerster
Abstract:
Progress in offline reinforcement learning (RL) has been impeded by ambiguous problem definitions and entangled algorithmic designs, resulting in inconsistent implementations, insufficient ablations, and unfair evaluations. Although offline RL explicitly avoids environment interaction, prior methods frequently employ extensive, undocumented online evaluation for hyperparameter tuning, complicating method comparisons. Moreover, existing reference implementations differ significantly in boilerplate code, obscuring their core algorithmic contributions. We address these challenges by first introducing a rigorous taxonomy and a transparent evaluation protocol that explicitly quantifies online tuning budgets. To resolve opaque algorithmic design, we provide clean, minimalistic, single-file implementations of various model-free and model-based offline RL methods, significantly enhancing clarity and achieving substantial speed-ups. Leveraging these streamlined implementations, we propose Unifloral, a unified algorithm that encapsulates diverse prior approaches and enables development within a single, comprehensive hyperparameter space. Using Unifloral with our rigorous evaluation protocol, we develop two novel algorithms - TD3-AWR (model-free) and MoBRAC (model-based) - which substantially outperform established baselines. Our implementation is publicly available at https://anonymous.4open.science/r/Unifloral-NeurIPS-D1A2.



Paperid:550
Authors:Clemens Damke, Eyke Hüllermeier
Abstract:
Quantification learningis the task of predicting the label distribution of a set of instances.We study this problem in the context of graph-structured data, where the instances are vertices.Previously, this problem has only been addressed via node clustering methods.In this paper, we extend the popularAdjusted Classify & Count(ACC) method to graphs.We show that the prior probability shift assumption upon which ACC relies is often not applicable to graph quantification problems.To address this issue, we propose structural importance sampling (SIS), the first graph quantification method that is applicable under (structural) covariate shift.Additionally, we propose Neighborhood-aware ACC, which improves quantification in the presence of non-homophilic edges.We show the effectiveness of our techniques on multiple graph quantification tasks.



Paperid:551
Authors:Sebastian H. M. Hickman, Ilija Trajković, Julia Kaltenborn, Francis Pelletier, Alex Archibald, Yaniv Gurwicz, Peer Nowack, David Rolnick, Julien Boussard
Abstract:
Traditional models of climate change use complex systems of coupled equations to simulate physical processes across the Earth system. These simulations are highly computationally expensive, limiting our predictions of climate change and analyses of its causes and effects. Machine learning has the potential to quickly emulate data from climate models, but current approaches are not able to incorporate physics-informed causal relationships. Here, we develop an interpretable climate model emulator based on causal representation learning. We derive a physics-informed approach including a Bayesian filter for stable long-term autoregressive emulation. We demonstrate that our emulator learns accurate climate dynamics, and we show the importance of each one of its components on a realistic synthetic dataset and data from two widely deployed climate models.



Paperid:552
Authors:Javad Aliakbari, Johan Östman, Ashkan Panahi, Alexandre Graell i Amat
Abstract:
We consider the problem of federated learning (FL) with graph-structured data distributed across multiple clients. In particular, we address the common scenario of interconnected subgraphs, where interconnections between clients significantly influence the learning process. Existing approaches suffer from critical limitations, either requiring the exchange of sensitive node embeddings, thereby posing privacy risks, or relying on computationally-intensive steps, which hinders scalability.To tackle these challenges, we propose FedLap, a novel framework that leverages global structure information via Laplacian smoothing in the spectral domain to effectively capture inter-node dependencies while ensuring privacy and scalability. We provide a formal analysis of the privacy of FedLap, demonstrating that it preserves privacy. Notably, FedLap is the first subgraph FL scheme with strong privacy guarantees. Extensive experiments on benchmark datasets demonstrate that the proposed method achieves competitive or superior utility compared to existing techniques.



Paperid:553
Authors:Tong Wu, Shuai Yang, Ryan Po, Yinghao Xu, Ziwei Liu, Dahua Lin, Gordon Wetzstein
Abstract:
Emerging world models autoregressively generate video frames in response to actions, such as camera movements and text prompts, among other control signals. Due to limited temporal context window sizes, these models often struggle to maintain scene consistency during revisits, leading to severe forgetting of previously generated environments. Inspired by the mechanisms of human memory, we introduce a novel framework to enhancing long-term consistency of video world models through a geometry-grounded long-term spatial memory. Our framework includes mechanisms to store and retrieve information from the long-term spatial memory and we curate custom datasets to train and evaluate world models with explicitly stored 3D memory mechanisms. Our evaluations show improved quality, consistency, and context length compared to relevant baselines, paving the way towards long-term consistent world generation.



Paperid:554
Authors:Anshul Nasery, Jonathan Hayase, Creston Brooks, Peiyao Sheng, Himanshu Tyagi, Pramod Viswanath, Sewoong Oh
Abstract:
Model fingerprinting has emerged as a powerful tool for model owners to identify their shared model given API access. In order to lower false discovery rate, fight fingerprint leakage, and defend against coalitions of model users attempting to bypass detection, we argue that scaling up the number of fingerprints one can embed into a model, i.e.Scalabilityof fingerprints, is critical. Hence, we pose scalability as a crucial requirement for fingerprinting schemes. We experiment with fingerprint design at a scale significantly larger than previously considered,and introduce a new method, dubbed Perinucleus sampling, to generate scalable, persistent, and harmless fingerprints. We demonstrate that this scheme can add 24,576 fingerprints to a Llama-3.1-8B model---two orders of magnitude more than existing schemes---without degrading the model's utility. Our inserted fingerprints persist even after supervised fine-tuning on standard post-training data. We further address security risks for fingerprinting, and theoretically and empirically show how a scalable fingerprinting scheme like ours can mitigate these risks.



Paperid:555
Authors:Jiayang Liu, Siyuan Liang, Shiqian Zhao, Rong-Cheng Tu, Wenbo Zhou, Aishan Liu, Dacheng Tao, Siew Kei Lam
Abstract:
In recent years, fueled by the rapid advancement of diffusion models, text-to-video (T2V) generation models have achieved remarkable progress, with notable examples including Pika, Luma, Kling, and Open-Sora. Although these models exhibit impressive generative capabilities, they also expose significant security risks due to their vulnerability to jailbreak attacks, where the models are manipulated to produce unsafe content such as pornography, violence, or discrimination. Existing works such as T2VSafetyBench provide preliminary benchmarks for safety evaluation, but lack systematic methods for thoroughly exploring model vulnerabilities.To address this gap, we are the first to formalize the T2V jailbreak attack as a discrete optimization problem and propose a joint objective-based optimization framework, called \emph{T2V-OptJail}. This framework consists of two key optimization goals: bypassing the built-in safety filtering mechanisms to increase the attack success rate, preserving semantic consistency between the adversarial prompt and the unsafe input prompt, as well as between the generated video and the unsafe input prompt, to enhance content controllability. In addition, we introduce an iterative optimization strategy guided by prompt variants, where multiple semantically equivalent candidates are generated in each round, and their scores are aggregated to robustly guide the search toward optimal adversarial prompts.We conduct large-scale experiments on several T2V models, covering both open-source models (\textit{e.g.}, Open-Sora) and real commercial closed-source models (\textit{e.g.}, Pika, Luma, Kling). The experimental results show that the proposed method improves 11.4\% and 10.0\% over the existing state-of-the-art method (SoTA) in terms of attack success rate assessed by GPT-4, attack success rate assessed by human accessors, respectively, verifying the significant advantages of the method in terms of attack effectiveness and content control. This study reveals the potential abuse risk of the semantic alignment mechanism in the current T2V model and provides a basis for the design of subsequent jailbreak defense methods. Our code is available at \url{https://anonymous.4open.science/r/NeruIPS25t2v-CE60}.



Paperid:556
Authors:Miaoyu Li, Qin Chao, Boyang Li
Abstract:
Evaluating the video understanding capabilities of Video-Language Models (VLMs) remains a significant challenge. We propose a long-context video understanding benchmark, Causal2Needles, that assesses two crucial abilities insufficiently evaluated by existing benchmarks: (1) the ability to extract information from two separate locations in a long video and understand them jointly, and (2) the ability to model the world in terms of cause and effect in human behaviors. Specifically, Causal2Needles introduces 2-needle questions, which require extracting information from both the cause and effect human-behavior events in a long video and the associated narration text. To prevent textual bias, these questions comprise two complementary formats: one asking to identify the video clip containing the answer, and one asking for the textual description of an unrelated visual detail from that video clip. Our experiments reveal that models excelling in pre-existing benchmarks struggle with 2-needle visual grounding, and the model performance is negatively correlated with the distance between the two needles. These findings highlight critical limitations in current VLMs.



Paperid:557
Authors:Jiazhi Yang, Kashyap Chitta, Shenyuan Gao, Long Chen, Yuqian Shao, Xiaosong Jia, Hongyang Li, Andreas Geiger, Xiangyu Yue, Li Chen
Abstract:
How can we reliably simulate future driving scenarios under a wide range of ego driving behaviors? Recent driving world models, developed exclusively on real-world driving data composed mainly of safe expert trajectories, struggle to follow hazardous or non-expert behaviors, which are rare in such data. This limitation restricts their applicability to tasks such as policy evaluation. In this work, we address this challenge by enriching real-world human demonstrations with diverse non-expert data collected from a driving simulator (e.g., CARLA), and building a controllable world model trained on this heterogeneous corpus. Starting with a video generator featuring diffusion transformer architecture, we devise several strategies to effectively integrate conditioning signals and improve prediction controllability and fidelity. The resulting model, ReSim, enables Reliable Simulation of diverse open-world driving scenarios under various actions, including hazardous non-expert ones. To close the gap between high-fidelity simulation and applications that require reward signals to judge different actions, we introduce a Video2Reward module that estimates reward from ReSim’s simulated future. Our ReSim paradigm achieves up to 44% higher visual fidelity, improves controllability for both expert and non-expert actions by over 50%, and boosts planning and policy selection performance on NAVSIM by 2% and 25%, respectively. Our code, model, and dataset will be released.



Paperid:558
Authors:Michael Robinson, Sourya Dey, Tony Chiang
Abstract:
Abstract:A full understanding of the behavior of a large language model (LLM) requires our understanding of its input token space. If this space differs from our assumptions, our understanding of and conclusions about the LLM will likely be flawed. We elucidate the structure of the token embeddings both empirically and theoretically. We present a novel statistical test assuming that the neighborhood around each token has a relatively flat and smooth structure as the null hypothesis. Failing to reject the null is uninformative, but rejecting it at a specific token $\psi$ implies an irregularity in the token subspace in a $\psi$-neighborhood, $B(\psi)$. The structure assumed in the null is a generalization of a manifold with boundary called a \emph{smooth fiber bundle} (which can be split into two spatial regimes -- small and large radius), so we denote our new hypothesis test as the ``fiber bundle hypothesis.'' Failure to reject the null hypothesis is uninformative, but rejecting it at $\psi$ indicates a statistically significant irregularity at $B(\psi)$. By running our test over several open-source LLMs, each with unique token embeddings, we find that the null is frequently rejected, and so the evidence suggests that the token subspace is not a fiber bundle and hence also not a manifold. As a consequence of our findings, when an LLM is presented with two semantically equivalent prompts, if one prompt contains a token implicated by our test, the response to that prompt will likely exhibit less stability than the other.



Paperid:559
Authors:Jiaxi Hu, Yongqi Pan, Jusen Du, Disen Lan, Tang, Qingsong Wen, Yuxuan Liang, Weigao Sun
Abstract:
Recent efficient sequence modeling methods, such as Gated DeltaNet, TTT, and RWKV-7, have achieved performance improvements by supervising the recurrent memory management through the Delta learning rule. Unlike previous state-space models (e.g., Mamba) and gated linear attentions (e.g., GLA), these models introduce interactions between recurrent state and the key vector, resulting in a nonlinear recursive structure. In this paper, we first introduce the concept of Nonlinear RNNs with a comprehensive analysis on the advantages and limitations of these models. Then based on the closed-loop control theory, we propose a novel Nonlinear RNN variant named Comba, which adopts a scalar-plus-low-rank state transition, with both state feedback and output feedback corrections. We also implement a hardware-efficient chunk-wise parallel kernel in Triton and train models with 340M/1.3B parameters on large-scale corpus. Comba demonstrates its superior performance and computation efficiency on both language modeling and vision tasks.



Paperid:560
Authors:Jaehyeong Jo, Sung Ju Hwang
Abstract:
Diffusion models have emerged as a promising alternative to autoregressive models in modeling discrete categorical data. However, diffusion models that directly work on discrete data space fail to fully exploit the power of iterative refinement, as the signals are lost during transitions between discrete states. Existing continuous diffusion models for discrete data underperform compared to discrete methods, and the lack of a clear connection between the two approaches hinders the development of effective diffusion models for discrete data. In this work, we propose a continuous diffusion model for language modeling that incorporates the geometry of the underlying categorical distribution. We establish a connection between the discrete diffusion and continuous flow on the statistical manifold, and building on this analogy, introduce a simple diffusion process that generalizes existing discrete diffusion models. We further propose a simulation-free training framework based on radial symmetry, along with a simple technique to address the high dimensionality of the manifold. Comprehensive experiments on language modeling benchmarks and other modalities show that our method outperforms existing discrete diffusion models and approaches the performance of autoregressive models.



Paperid:561
Authors:William Shen, Xinchi Qiu, Meghdad Kurmanji, Alexandru-Andrei Iacob, Lorenzo Sani, Yihong Chen, Nicola Cancedda, Nicholas Lane
Abstract:
The ability to selectively remove knowledge from LLMs is highly desirable. However, existing methods often struggle with balancing unlearning efficacy and retain model utility, and lack controllability at inference time to emulate base model behavior as if it had never seen the unlearned data. In this paper, we propose LUNAR, a novel unlearning method grounded in the Linear Representation Hypothesis and operates by redirecting the representations of unlearned data to activation regions that expresses its inability to answer. We show that contrastive features are not a prerequisite for effective activation redirection, and LUNAR achieves state-of-the-art unlearning performance and superior controllability. Specifically, LUNAR achieves between 2.9x and 11.7x improvement in the combined unlearning efficacy and model utility score (Deviation Score) across various base models and generates coherent, contextually appropriate responses post-unlearning. Moreover, LUNAR effectively reduces parameter updates to a single down-projection matrix, a novel design that significantly enhances efficiency by 20x and robustness. Finally, we demonstrate that LUNAR is robust to white-box adversarial attacks and versatile in real-world scenarios, including handling sequential unlearning requests.



Paperid:562
Authors:Haiyu Wang, Yuanyuan Lin
Abstract:
Abstract:One research direction for understanding Transformers' representational capacity involves quantifying their data memorization ability. However, the memorization capacity of Transformers with prompts is not yet well understood. In this work, we extend the existing results on prompt tuning Transformers to simulate ReLU neural networks with the Autoregressive algorithm in \citep{nakada2025theoretical} to a more general case where the width of the ReLU neural network can be a small constant. Based on this, we prove that a constant-size Transformer with prompts whose length is $\tilde{O}(\sqrt{nN})$ can memorize any $N$ input sequence of length $n$. Our results discard the reliance on large feed-forward layers. Besides, we also theoretically prove that a trade-off exists between the prompt length and computational efficiency. Our findings, supported by experiments, demonstrate that a single-layer randomly initialized Transformer with prompts can possess competitive data memorization ability compared with models trained from scratch. In addition, we validate that the prompt length is potentially reduced if the low-rank structure exists.



Paperid:563
Authors:Xiaojun Guo, Ang Li, Yifei Wang, Stefanie Jegelka, Yisen Wang
Abstract:
Although Large Language Models (LLMs) have demonstrated remarkable progress, their proficiency in graph-related tasks remains notably limited, hindering the development of truly general-purpose models. Previous attempts, including pretraining graph foundation models or employing supervised fine-tuning, often face challenges such as the scarcity of large-scale, universally represented graph data. We introduce G1, a simple yet effective approach demonstrating that Reinforcement Learning (RL) on synthetic graph-theoretic tasks can significantly scale LLMs' graph reasoning abilities. To enable RL training, we curate Erdos, the largest graph reasoning dataset to date comprising 50 diverse graph-theoretic tasks of varying difficulty levels, 100k training data and 5k test data, all drived from real-world graphs. With RL on Erdos, G1 obtains substantial improvements in graph reasoning, where our finetuned 3B model even outperforms Qwen2.5-72B-Instruct (24x size). RL-trained models also show strong zero-shot generalization to unseen tasks, domains, and graph encoding schemes, including other graph-theoretic benchmarks as well as real-world node classification and link prediction tasks, without compromising general reasoning abilities.Our findings offer an efficient, scalable path for building strong graph reasoners by finetuning LLMs with RL on graph-theoretic tasks, which combines the strengths of pretrained LLM capabilities with abundant, automatically generated synthetic data, suggesting that LLMs possess graph understanding abilities that RL can elicit successfully.



Paperid:564
Authors:Chang Li, Zehua Chen, Liyuan Wang, Jun Zhu
Abstract:
Audio super-resolution (SR), i.e., upsampling the low-resolution (LR) waveform to the high-resolution (HR) version, has recently been explored with diffusion and bridge models, while previous methods often suffer from sub-optimal upsampling quality due to their uninformative generation prior. Towards high-quality audio super-resolution, we present a new system with latent bridge models (LBMs), where we compress the audio waveform into a continuous latent space and design an LBM to enable a latent-to-latent generation process that naturally matches the LR-to-HR upsampling process, thereby fully exploiting the instructive prior information contained in the LR waveform. To further enhance the training results despite the limited availability of HR samples, we introduce frequency-aware LBMs, where the prior and target frequency are taken as model input, enabling LBMs to explicitly learn an any-to-any upsampling process at the training stage. Furthermore, we design cascaded LBMs and present two prior augmentation strategies, where we make the first attempt to unlock the audio upsampling beyond 48 kHz and empower a seamless cascaded SR process, providing higher flexibility for audio post-production. Comprehensive experimental results evaluated on the VCTK, ESC-50, Song-Describer benchmark datasets and two internal testsets demonstrate that we achieve state-of-the-art objective and perceptual quality for any-to-48 kHz SR across speech, audio, and music signals, as well as setting the first record for any-to-192kHz audio SR. Demo at https://AudioLBM.github.io/.



Paperid:565
Authors:Laurence Davies, Daniel MacKinlay, Rafael Oliveira, Scott SIsson
Abstract:
The expressiveness of flow-based models combined with stochastic variational inference (SVI) has, in recent years, expanded the application of optimization-based Bayesian inference to include problems with complex data relationships. However, until now, SVI using flow-based models has been limited to problems of fixed dimension. We introduce CoSMIC normalizing flows (COntextually-Specified Masking for Identity-mapped Components), an extension to neural autoregressive conditional normalizing flow architectures that enables using a single amortized variational density for inference over a transdimensional target distribution. We propose a combined stochastic variational transdimensional inference approach to training CoSMIC flows using techniques from Bayesian optimization and Monte Carlo gradient estimation. Numerical examples are provided to demonstrate the proposed methodology on challenging problems that scale to high-cardinality model spaces.



Paperid:566
Authors:John Zhou, Jonathan Kao
Abstract:
Offline goal-conditioned reinforcement learning (GCRL) is a promising approach for pretraining generalist policies on large datasets of reward-free trajectories, akin to the self-supervised objectives used to train foundation models for computer vision and natural language processing. However, scaling GCRL to longer horizons remains challenging due to the combination of sparse rewards and discounting, which obscures the comparative advantages of primitive actions with respect to distant goals. Hierarchical RL methods achieve strong empirical results on long-horizon goal-reaching tasks, but their reliance on modular, timescale-specific policies and subgoal generation introduces significant additional complexity and hinders scaling to high-dimensional goal spaces. In this work, we introduce an algorithm to train a flat (non-hierarchical) goal-conditioned policy by bootstrapping on subgoal-conditioned policies with advantage-weighted importance sampling. Our approach eliminates the need for a generative model over the (sub)goal space, which we find is key for scaling to high-dimensional control in large state spaces. We further show that existing hierarchical and bootstrapping-based approaches correspond to specific design choices within our derivation. Across a comprehensive suite of state- and pixel-based locomotion and manipulation benchmarks, our method matches or surpasses state-of-the-art offline GCRL algorithms and scales to complex, long-horizon tasks where prior approaches fail.



Paperid:567
Authors:Dongkeun Yoon, Seungone Kim, Sohee Yang, Sunkyoung Kim, Soyeon Kim, Yongil Kim, Eunbi Choi, Yireun Kim, Minjoon Seo
Abstract:
Despite their strengths, large language models (LLMs) often fail to communicate their confidence accurately, making it difficult to assess when they might be wrong and limiting their reliability. In this work, we demonstrate that reasoning models—LLMs that engage in extended chain-of-thought (CoT) reasoning—exhibit superior performance not only in problem-solving but also in accurately expressing their confidence. Specifically, we benchmark six reasoning models across six datasets and find that they achieve strictly better confidence calibration than their non-reasoning counterparts in 33 out of the 36 settings. Our detailed analysis reveal that these gains in calibration stem from the slow thinking behaviors of reasoning models—such as exploring alternative approaches and backtracking—which enable them to adjust their confidence dynamically throughout their CoT, making it progressively more accurate. In particular, we find that reasoning models become increasingly better calibrated as their CoT unfolds, a trend not observed in non-reasoning models. Moreover, removing slow thinking behaviors from the CoT leads to a significant drop in calibration. Lastly, we show that these gains are not exclusive to reasoning models—non-reasoning models also benefit when guided to perform slow thinking via in-context learning.



Paperid:568
Authors:Nicholas Jiang, Amil Dravid, Alexei Efros, Yossi Gandelsman
Abstract:
Abstract:We investigate the mechanism underlying a previously identified phenomenon in Vision Transformers: the emergence of high-norm tokens that lead to noisy attention maps (Darcet et al., 2024). We observe that in multiple models (e.g., CLIP, DINOv2), a sparse set of neurons is responsible for concentrating high-norm activations on outlier tokens, leading to anomalous attention patterns and degrading downstream visual processing. While the existing solution for removing these outliers involves retraining models from scratch with additional learned register tokens, we use our findings to create a training-free approach to mitigate these artifacts. By shifting the high norm activations from our discovered $\textit{register neurons}$ into an additional untrained token, we can mimic the effect of register tokens on a model already trained without registers. Our method produces cleaner attention and feature maps, enhances performance over base models across multiple downstream tasks, and achieves results comparable to models explicitly trained with register tokens. This suggests that our approach effectively takes on the role of register tokens at test-time, offering a training-free solution for any pre-trained off-the-shelf model released without them.



Paperid:569
Authors:Qingyun Chen, Sungjin Im, Ben Moseley, Ryan Milstrey, Chenyang Xu, Ruilong Zhang
Abstract:
Abstract:$k$-means++ and the related greedy $k$-means++ algorithm are celebrated algorithms that efficiently compute seeds for Lloyd's algorithm. Greedy $k$-means++ is a generalization of $k$-means++ where, in each iteration, a new seed is greedily chosen among multiple $\ell \geq 2$ points sampled, as opposed to a single seed being sampled in $k$-means++. While empirical studies consistently show the superior performance of greedy $k$-means++, making it a preferred method in practice, a discrepancy exists between theory and practice. No theoretical justification currently explains this improved performance. Indeed, the prevailing theory suggests that greedy $k$-means++ exhibits worse performance than $k$-means++ in worst-case scenarios.This paper presents an analysis demonstrating the outperformance of the greedy algorithm compared to $k$-means++ for a natural class of well-separated instances with exponentially decaying distributions, such as Gaussian, specifically when $\ell = \Theta(\log k)$, a common parameter setting in practical applications.



Paperid:570
Authors:Omri Hirsch, Ron Shapira Weber, Shira Ifergane, Oren Freifeld
Abstract:
Joint Alignment (JA) of images aims to align a collection of images into a unified coordinate frame, such that semantically similar features appear at corresponding spatial locations. Most existing approaches often require extensive training times, large-capacity models,and extensive hyperparameter tuning. We introduce FastJAM, a rapid, graph-based method that drastically reduces the computational complexity of joint alignment tasks. FastJAM leverages pairwise matches computed by off-the-shelf image matchers to construct a graph representing intra- and inter-image keypoint relations. A graph neural network propagates and aggregates these correspondences, efficiently predicting per-image homography parameters via image-level pooling. Utilizing an inverse-compositional warping strategy, FastJAM performs image JA quickly and effectively. Experimental results on several benchmarks demonstrate that FastJAM achieves results better than existing modern JA methods in terms of alignment quality, while reducing computation time from hours or minutes to mere seconds. Our code will be made public upon acceptance.



Paperid:571
Authors:Daiqing Qi, Dongliang Guo, Hanzhang Yuan, Handong Zhao, Mengxuan Hu, Lehan Yang, Sheng Li
Abstract:
A major distinction between video and image understanding is that the former requires reasoning over time.Existing Video Large Language Models (VLLMs) demonstrate promising performance in general video understanding, such as brief captioning or object recognition within individual frames. However, they often struggle with temporal reasoning such as understanding continuous actions or tracking object transformations over time—which typically demands the integration of multiple frames in a temporally coherent manner.We first explore and explain such failures in Video LLMs from the perspective of \textit{language and ``image'' priors.}While existing research has attempted to enhance the temporal understanding of VLLMs through various training strategies, the demand for expensive computational resources and training data often presents significant barriers.To this end, we further propose a simple yet novel idea for improving temporal reasoning in videos at no additional training cost.Specifically, to better capture the temporal structure across multiple frames—the key to effective temporal reasoning—we distort the temporal consistency in key frames \textit{during the decoding phase}. Such corruption induces time-insensitive wrong responses from the model, which are then contrastively avoided when generating the final correct output. In this way, the model is encouraged to perform more temporally coherent reasoning.Our method yields consistent improvements across both temporal-specific and general video understanding benchmarks, demonstrating its effectiveness and generalizability.



Paperid:572
Authors:Hanshi Wang, Cai Zijian, Jin Gao, Yiwei Zhang, Weiming Hu, Ke Wang, Zhipeng Zhang
Abstract:
Online, real-time, and fine-grained 3D segmentation constitutes a fundamental capability for embodied intelligent agents to perceive and comprehend their operational environments. Recent advancements employ predefined object queries to aggregate semantic information from Vision Foundation Models (VFMs) outputs that are lifted into 3D point clouds, facilitating spatial information propagation through inter-query interactions.Nevertheless, perception, whether human or robotic, is an inherently dynamic process, rendering temporal understanding a critical yet overlooked dimension within these prevailing query-based pipelines. This deficiency in temporal reasoning can exacerbate issues such as the over-segmentation commonly produced by VFMs, necessitating more handcrafted post-processing. Therefore, to further unlock the temporal environmental perception capabilities of embodied agents, our work reconceptualizes online 3D segmentation as an instance tracking problem. Our core strategy involves utilizing object queries for temporal information propagation, where long-term instance association promotes the coherence of features and object identities, while short-term instance update enriches instant observations. Given that viewpoint variations in embodied robotics often lead to partial object visibility across frames, this mechanism aids the model in developing a holistic object understanding beyond incomplete instantaneous views. Furthermore, we introduce spatial consistency learning to mitigate the fragmentation problem inherent in VFMs, yielding more comprehensive instance information for enhancing the efficacy of both long-term and short-term temporal learning. The temporal information exchange and consistency learning facilitated by these sparse object queries not only enhance spatial comprehension but also circumvent the computational burden associated with dense temporal point cloud interactions. Our method establishes a new state-of-the-art, surpassing ESAM by 2.8 AP on ScanNet200 and delivering consistent gains on ScanNet, SceneNN, and 3RScan datasets, corroborating that identity-aware temporal reasoning is a crucial, previously underemphasized component for robust 3D segmentation in real-time embodied intelligence. Code and models will be made available.



Paperid:573
Authors:Wenxuan Li, Xinze Zhou, Qi Chen, Tianyu Lin, Pedro R. A. S. Bassi, Xiaoxi Chen, Chen Ye, Zheren Zhu, Kai Ding, Heng Li, Kang Wang, Yang Yang, Yucheng Tang, Daguang Xu, Alan Yuille, Zongwei Zhou
Abstract:
PanTS is a large-scale, multi-institutional dataset curated to advance research in pancreatic CT analysis. It contains 36,390 CT scans from 145 medical centers, with expert-validated, voxel-wise annotations of over 993,000 anatomical structures, covering pancreatic tumors, pancreas head, body, and tail, and 24 surrounding anatomical structures such as vascular/skeletal structures and abdominal/thoracic organs. Each scan includes metadata such as patient age, sex, scan phase, diagnosis, voxel spacing, etc. AI models trained on PanTS achieve significantly better performance in pancreatic tumor detection, localization, and segmentation compared to those trained on existing public datasets. Our analysis indicates that these gains are directly attributable to the 12x larger-scale tumor annotations and indirectly supported by the 24 additional surrounding anatomical structures. As the largest and most comprehensive resource of its kind, PanTS offers a new benchmark for developing and evaluating AI models in pancreatic CT analysis.



Paperid:574
Authors:Yohan Jung, Hyungi Lee, Wenlong Chen, Thomas Möllenhoff, Yingzhen Li, Juho Lee, Mohammad Emtiyaz Khan
Abstract:
Abstract:Despite recent progress, continual lifelong learning algorithms still do not match the performance of batch training. A major challenge is to represent old knowledge and reuse it to learn while avoiding forgetting. It remains unclear how to compactly represent past memories, and there is no clear answer even in shallow neural networks with just one or two layers.In this paper, we take a step toward building compact memory. We build upon a result by Khan and Swaroop [2021], who show the existence of optimal memory for logistic regression. We show that finding the memory can be formulated as Hessian matching and propose a Probabilistic principal component analysis (PPCA)-based method to obtain the memory. Our approach can drastically reduce memory size without sacrificing performance, for instance, we can obtain $30\%$ improvement in accuracy over Experience Replay that uses $1\%$ of the data for memory when adapting it with the the pre-trained Vision transformer as a feature extractor. Our work takes the first step to build compact memory which can be useful for deeper models in the future.



Paperid:575
Authors:Chunyu Wei, Haozhe Lin, Yueguo Chen, Yunhai Wang
Abstract:
Graph anomaly detection (GAD) has become a critical research area, with successful applications in financial fraud and telecommunications. Traditional Graph Neural Networks (GNNs) face significant challenges: at the topology level, they suffer from over-smoothing that averages out anomalous signals; at the feature level, discriminative models struggle when fraudulent nodes obfuscate their features to evade detection. In this paper, we propose a Conditional Graph Anomaly Diffusion Model (CGADM) that addresses these issues through the iterative refinement and denoising reconstruction properties of diffusion models. Our approach incorporates a prior-guided diffusion process that injects a pre-trained conditional anomaly estimator into both forward and reverse diffusion chains, enabling more accurate anomaly detection. For computational efficiency on large-scale graphs, we introduce a prior confidence-aware mechanism that adaptively determines the number of reverse denoising steps based on prior confidence. Experimental results on benchmark datasets demonstrate that CGADM achieves state-of-the-art performance while maintaining significant computational advantages for large-scale graph applications.



Paperid:576
Authors:Yue Wu, Jun Jiang, Yongzhe Yuan, Maoguo Gong, Qiguang Miao, Hao Li, Mingyang Zhang, wenping ma
Abstract:
Point cloud registration is a fundamental task in computer vision and remote sensing. Recent advances have shown that graph-based approaches are effective for outlier removal in this context. However, existing clique-based methods impose overly strict constraints and are NP-hard, making it difficult to achieve both robustness and efficiency. While the k-core reduces computational complexity, it only considers node degree and ignores higher-order topological structures such as triangles, limiting its effectiveness in complex scenarios. To overcome these limitations, we introduce the k-truss from graph theory into point cloud registration, leveraging triangle support as a constraint for inlier selection. We further propose a consensus voting-based low-scale sampling strategy to efficiently extract the structural skeleton of the point cloud prior to k-truss decomposition. Additionally, we design a spatial distribution score to jointly evaluate the coverage and uniformity of inliers, thus avoiding solutions where inliers are spatially clustered. Extensive experiments on KITTI, 3DMatch, and 3DLoMatch demonstrate that our method consistently outperforms both traditional and learning-based approaches in various indoor and outdoor scenarios, achieving state-of-the-art results.



Paperid:577
Authors:Puheng Li, Tijana Zrnic, Emmanuel Candes
Abstract:
Active statistical inference is a new method for inference with AI-assisted data collection. Given a budget on the number of labeled data points that can be collected and assuming access to an AI predictive model, the basic idea is to improve estimation accuracy by prioritizing the collection of labels where the model is most uncertain. The drawback, however, is that inaccurate uncertainty estimates can make active sampling produce highly noisy results, potentially worse than those from naive uniform sampling.In this work, we present robust sampling strategies for active statistical inference. Robust sampling ensures that the resulting estimator is never worse than the estimator using uniform sampling. Furthermore, with reliable uncertainty estimates, the estimator can even surpass standard active inference. This is achieved by optimally interpolating between uniform and active sampling, depending on the quality of the uncertainty scores, and by using ideas from robust optimization. We demonstrate the utility of the method on a series of real datasets from computational social science and survey research.



Paperid:578
Authors:Yuxiang Bao, Huijie Liu, xun gao, Huan Fu, Guoliang Kang
Abstract:
Naive DDIM inversion process usually suffers from a trajectory deviation issue, i.e., the latent trajectory during reconstruction deviates from the one during inversion. To alleviate this issue, previous methods either learn to mitigate the deviation or design cumbersome compensation strategy to reduce the mismatch error, exhibiting substantial time and computation cost. In this work, we present a nearly free-lunch method (named FreeInv) to address the issue more effectively and efficiently. In FreeInv, we randomly transform the latent representation and keep the transformation the same between the corresponding inversion and reconstruction time-step. It is motivated from a statistical perspective that an ensemble of DDIM inversion processes for multiple trajectories yields a smaller trajectory mismatch error on expectation.Moreover, through theoretical analysis and empirical study, we show that FreeInv performs an efficient ensemble of multiple trajectories. FreeInv can be freely integrated into existing inversion-based image and video editing techniques. Especially for inverting video sequences, it brings more significant fidelity and efficiency improvements. Comprehensive quantitative and qualitative evaluation on PIE benchmark and DAVIS dataset shows that FreeInv remarkably outperforms conventional DDIM inversion, and is competitive among previous state-of-the-art inversion methods, with superior computation efficiency.



Paperid:579
Authors:Aly Hakie, Yiren Lu, Yu Yin, Michael Jenkins, Yehe Liu
Abstract:
3D Gaussian Splatting (3DGS) has demonstrated impressive capabilities in 3D reconstruction. However, its α-blending can induce ’false transparency’ artifacts, particularly where low point cloud density in sparse or low-texture regions causes foreground objects to appear improperly transparent. This issue stems from an ill-posed optimization. During training, background Gaussians blend with foreground ones, making them difficult to differentiate using only photometric loss, which leads to the observed transparency in these regions. This view-inconsistency issue is hard to detect in static renderings during training and validation, but becomes evident in object-centric reconstruction during interactive rotation. Although other causes of view-inconsistency (e.g., popping artifacts) have been explored recently, false transparency has not been explicitly identified. This paper proposes a novel explanation to the problem and a solution to remedy it by injecting opaque noise Gaussians in the object volume during training. Our strategy, Noise Guided Splatting ( NGS), encourages surface Gaussians to adopt higher opacity while minimally modifying the existing splatting process. To quantitatively evaluate the false transparency in static renderings, we propose a transmittance-based metric to characterize the extent of the false transparency problem. We also introduce a customized high-quality object-centric scan dataset exhibiting prominent transparency issues and supplement popular existing datasets (e.g., DTU) with new complementary infill noise specifically designed to evaluate false transparency handling in 3D reconstruction methods. Across various datasets, NGS substantially reduces surface transmittance while maintaining performance on standard rendering metrics (e.g., PSNR), demonstrating its effectiveness



Paperid:580
Authors:Nathan Gavenski, Odinaldo Rodrigues
Abstract:
Imitation learning benchmarks often lack sufficient variation between training and evaluation, limiting meaningful generalisation assessment. We introduce Labyrinth, a benchmarking environment designed to test generalisation with precise control over structure, start and goal positions, and task complexity.It enables verifiably distinct training, evaluation, and test settings.Labyrinth provides a discrete, fully observable state space and known optimal actions, supporting interpretability and fine-grained evaluation.Its flexible setup allows targeted testing of generalisation factors and includes variants like partial observability, key-and-door tasks, and ice-floor hazards.By enabling controlled, reproducible experiments, Labyrinth advances the evaluation of generalisation in imitation learning and provides a valuable tool for developing more robust agents.



Paperid:581
Authors:Alejandro Almodóvar, Adrián Javaloy, Juan Parras, Santiago Zazo, Isabel Valera
Abstract:
We introduce DeCaFlow, a deconfounding causal generative model. Training once per dataset using just observational data and the underlying causal graph, DeCaFlow enables accurate causal inference on continuous variables under the presence of hidden confounders. Specifically, we extend previous results on causal estimation under hidden confounding to show that a single instance of DeCaFlow provides correct estimates for all causal queries identifiable with do-calculus, leveraging proxy variables to adjust for the causal effects when do-calculus alone is insufficient. Moreover, we show that counterfactual queries are identifiable as long as their interventional counterparts are identifiable, and thus are also correctly estimated by DeCaFlow. Our empirical results on diverse settings—including the Ecoli70 dataset, with 3 independent hidden confounders, tens of observed variables and hundreds of causal queries—show that DeCaFlow outperforms existing approaches, while demonstrating its out-of-the-box applicability to any given causal graph.



Paperid:582
Authors:Benjamin Hoover, Zhaoyang Shi, Krishnakumar Balasubramanian, Dmitry Krotov, Parikshit Ram
Abstract:
We propose a novel energy function for Dense Associative Memory (DenseAM) networks, the log-sum-ReLU (LSR), inspired by optimal kernel density estimation. Unlike the common log-sum-exponential (LSE) function, LSR is based on the Epanechnikov kernel and enables exact memory retrieval with exponential capacity without requiring exponential separation functions. Uniquely, it introduces abundant additional emergent local minima while preserving perfect pattern recovery --- a characteristic previously unseen in DenseAM literature. Empirical results show that LSR energy has significantly more local minima (memories) that have comparable log-likelihood to LSE-based models. Analysis of LSR's emergent memories on image datasets reveals a degree of creativity and novelty, hinting at this method's potential for both large-scale memory storage and generative tasks.



Paperid:583
Authors:Zipeng Yan, Yinjie Chen, Chong Zhou, Bo Dai, Andrew Luo
Abstract:
Vision Transformers (ViTs) have emerged as the dominant architecture for visual processing tasks, demonstrating excellent scalability with increased training data and model size.However, recent work has identified the emergence of artifact tokens in ViTs that are incongruous with the local semantics.These anomalous tokens degrade ViT performance in tasks that require fine-grained localization or structural coherence. An effective mitigation of this issue is to the addition of register tokens to ViTs, which implicitly "absorb" the artifact term during training.Given the availability of various large-scale pre-trained ViTs,in this paper we aim at equipping them with such register tokens without the need of re-training them from scratch, which are infeasible considering their size.Specifically, we propose Post Hoc Registers (PH-Reg), an efficient self-distillationmethod that integrates registers into an existing ViT without requiring additional labeled data and full retraining. PH-Reg initializes both teacher and student networks from the same pre-trained ViT. The teacher remains frozen and unmodified, while the student is augmented with randomly initialized register tokens. By applying test-time augmentation to the teacher’s inputs, we generate denoised dense embeddings free of artifacts, which are then used to optimize only a small subset of unlocked student weights. We show that our approach can effectively reduce the number of artifact tokens, improving the segmentation and depth prediction of the student ViT under zero-shot and linear probing.



Paperid:584
Authors:Efstathios Karypidis, Ioannis Kakogeorgiou, Spyridon Gidaris, Nikos Komodakis
Abstract:
Predicting future dynamics is crucial for applications like autonomous driving and robotics, where understanding the environment is key. Existing pixel-level methods are computationally expensive and often focus on irrelevant details. To address these challenges, we introduce DINO-Foresight, a novel framework that operates in the semantic feature space of pretrained Vision Foundation Models (VFMs). Our approach trains a masked feature transformer in a self-supervised manner to predict the evolution of VFM features over time. By forecasting these features, we can apply off-the-shelf, task-specific heads for various scene understanding tasks. In this framework, VFM features are treated as a latent space, to which different heads attach to perform specific tasks for future-frame analysis. Extensive experiments show the very strong performance, robustness and scalability of our framework.



Paperid:585
Authors:Wenda Chu, Zihui Wu, Yifan Chen, Yang Song, Yisong Yue
Abstract:
We study the problem of posterior sampling in discrete-state spaces using discrete diffusion models. While posterior sampling methods for continuous diffusion models have achieved remarkable progress, analogous methods for discrete diffusion models remain challenging. In this work, we introduce a principled plug-and-play discrete diffusion posterior sampling algorithm based on split Gibbs sampling, which we call SGDD. Our algorithm enables reward-guided generation and solving inverse problems in discrete-state spaces. We demonstrate the convergence of SGDD to the target posterior distribution and verify this through controlled experiments on synthetic benchmarks. Our method enjoys state-of-the-art posterior sampling performance on a range of benchmarks for discrete data, including DNA sequence design, discrete image inverse problems, and music infilling, achieving more than 30% improved performance compared to existing baselines.



Paperid:586
Authors:jusheng zhang, Kaitong Cai, Yijia Fan, Jian Wang, Keze Wang
Abstract:
Recent advances in vision-language models (VLMs) have greatly improved cross-modal semantic understanding, yet significant limitations remain in fine-grained discrimination and deep causal reasoning tasks. Existing VLMs often rely on superficial statistical correlations, lacking the ability to capture the underlying causal logic between visual and textual content. To address this, we propose theCounterFactual Vision-Language Fine-tuning Model (CF-VLM), a novel framework that enhances the causal reasoning capabilities of VLMs through the targeted use of counterfactual samples. CF-VLM introduces three complementary training objectives: maintaining foundational cross-modal alignment, reinforcing the uniqueness, and stability of factual scene representations against coherent counterfactuals, and sharpening the model’s sensitivity to minimal but critical causal edits. Extensive experiments demonstrate that CF-VLM consistently outperforms strong baselines and state-of-the-art methods on compositional reasoning and generalization benchmarks. Furthermore, it shows promise in mitigating visual hallucinations, indicating improved factual consistency. Our CF-VLM provides a robust foundation for deploying VLMs in high-stakes, real-world scenarios requiring reliable reasoning and interpretability.



Paperid:587
Authors:Yuchen Lin, Chenguo Lin, Panwang Pan, Honglei Yan, Feng Yiqiang, Yadong Mu, Katerina Fragkiadaki
Abstract:
We introduce PartCrafter, the first structured 3D generative model that jointly synthesizes multiple semantically meaningful and geometrically distinct 3D meshes from a single RGB image. Unlike existing methods that either produce monolithic 3D shapes or follow two-stage pipelines, i.e. first segmenting an image and then reconstructing each segment, PartCrafter adopts a unified, compositional generation architecture that does not rely on pre-segmented inputs. Conditioned on a single image, it simultaneously denoises multiple 3D parts, enabling end-to-end part-aware generation of both individual objects and complex multi-object scenes. PartCrafter builds upon a pretrained 3D mesh diffusion transformer (DiT) trained on whole objects, inheriting the pretrained weights, encoder, and decoder, and introduces two key innovations: (1) A compositional latent space, where each 3D part is represented by a set of disentangled latent tokens; (2) A hierarchical attention mechanism that enables structured information flow both within individual parts and across all parts, ensuring global coherence while preserving part-level detail during generation. To support part-level supervision, we curate a new dataset by mining part-level annotations from large-scale 3D object datasets. Experiments show that PartCrafter outperforms existing approaches in generating decomposable 3D meshes, including parts that are not directly visible in input images, demonstrating the strength of part-aware generative priors for 3D understanding and synthesis. Code and training data will be released.



Paperid:588
Authors:Tongle Wu, Ying Sun
Abstract:
In this paper, we propose a novel tube-wise local tensor compressed sensing (CS) model, where sensing operators are independently applied to each tube of a third-order tensor. To recover the low-rank ground truth tensor, we minimize a non-convex objective via Burer–Monteiro factorization and solve it using gradient descent with spectral initialization. We prove that this approach achieves exact recovery with a linear convergence rate. Notably, our method attains provably lower sample complexity than existing TCS methods. Our proof leverages the leave-one-out technique to show that gradient descent generates iterates implicitly biased towards solutions with bounded incoherence, which ensures contraction of optimization error in consecutive iterates. Empirical results validate the effectiveness of GD in solving the proposed local TCS model.



Paperid:589
Authors:Bilgehan Sel, Vaishakh Keshava, Phillip Wallis, Lukas Rutishauser, Ming Jin, Dingcheng Li
Abstract:
Addressing the critical need for robust safety in Large Language Models (LLMs), particularly against adversarial attacks and in-distribution errors, we introduce Reinforcement Learning with Backtracking Feedback (RLBF). This framework advances upon prior methods, such as BSAFE, by primarily leveraging a Reinforcement Learning (RL) stage where models learn to dynamically correct their own generation errors. Through RL with critic feedback on the model's live outputs, LLMs are trained to identify and recover from their actual, emergent safety violations by emitting an efficient "backtrack by x tokens" signal, then continuing generation autoregressively. This RL process is crucial for instilling resilience against sophisticated adversarial strategies, including middle filling, Greedy Coordinate Gradient (GCG) attacks, and decoding parameter manipulations. To further support the acquisition of this backtracking capability, we also propose an enhanced Supervised Fine-Tuning (SFT) data generation strategy (BSAFE+). This method improves upon previous data creation techniques by injecting violations into coherent, originally safe text, providing more effective initial training for the backtracking mechanism. Comprehensive empirical evaluations demonstrate that RLBF significantly reduces attack success rates across diverse benchmarks and model scales, achieving superior safety outcomes while critically preserving foundational model utility.



Paperid:590
Authors:Zekun Qian, Ruize Han, Zhixiang Wang, Junhui Hou, Wei Feng
Abstract:
Open-Vocabulary Multi-Object Tracking (OVMOT) aims to detect and track multi-category objects that have both been seen and not seen during training.Currently, a significant challenge in this domain is the lack of large-scale annotated video data for training.To address this challenge, this work aims to effectively train the OV tracker using only the existing limited and sparse-annotated video data.Specifically, on the one hand, for the association task, we apply the diffusion models to complement the missing objects at the gaps between annotated frames, which is achieved directly in the implicit embedding space and enhances the coherence and accuracy of association features.On the other hand, for the object detection task, we develop a dynamic group contrastive learning strategy for object recognition and a simple yet effective loss, namely adaptive localization loss.Extensive results verify that the proposed method achieves promising results on the OVMOT benchmark, without additional data or annotations.



Paperid:591
Authors:Omar Salemohamed, Laurent Charlin, Shivam Garg, Vatsal Sharan, Gregory Valiant
Abstract:
We explore if it is possible to learn data structures end-to-end with neural networks, with a focus on the problem of nearest-neighbor (NN) search. We introduce a framework for data structure discovery, which adapts to the underlying data distribution and provides fine-grained control over query and space complexity. Crucially, the data structure is learned from scratch, and does not require careful initialization or seeding with candidate data structures. In several settings, we are able to reverse-engineer the learned data structures and query algorithms. For 1D nearest neighbor search, the model discovers optimal distribution (in)dependent algorithms such as binary search and variants of interpolation search. In higher dimensions, the model learns solutions that resemble k-d trees in some regimes, while in others, elements of locality-sensitive hashing emerge. Additionally, the model learns useful representations of high-dimensional data such as images and exploits them to design effective data structures. Beyond NN search, we believe the framework could be a powerful tool for data structure discovery for other problems and adapt our framework to the problem of estimating frequencies over a data stream. To encourage future work in this direction, we conclude with a discussion on some of the opportunities and remaining challenges of learning data structures end-to-end.



Paperid:592
Authors:Gianmarco Genalti, Francesco Emanuele Stradi, Matteo Castiglioni, Alberto Marchesi, Nicola Gatti
Abstract:
Abstract:This paper initiates the study of data-dependent regret bounds in constrained MAB settings. These are bounds that depend on the sequence of losses that characterize the problem instance. Thus, in principle they can be much smaller than classical $\widetilde{\mathcal{O}}(\sqrt{T})$ regret bounds, while being equivalent to them in the worst case. Despite this, data-dependent regret bounds have been completely overlooked in constrained MABs. The goal of this paper is to answer the question: Can data-dependent regret bounds be derived in the presence of constraints? We provide an affirmative answer in constrained MABs with adversarial losses and stochastic constraints. Specifically, our main focus is on the most challenging and natural settings with hard constraints, where the learner must ensure that the constraints are always satisfied with high probability. We design an algorithm with a regret bound consisting of two data-dependent terms. The first one captures the difficulty of satisfying the constraints, while the second one encodes the complexity of learning independently of their presence. We also prove a lower bound showing that these two terms are not artifacts of our specific approach and analysis, but rather the fundamental components that inherently characterize the problem complexity. Finally, in designing our algorithm, we also derive some novel results in the related (and easier) soft constraints settings, which may be of independent interest.



Paperid:593
Authors:Jonah Brown-Cohen, Geoffrey Irving, Georgios Piliouras
Abstract:
Training powerful AI systems to exhibit desired behaviors hinges on the ability to provide accurate human supervision on increasingly complex tasks. A promising approach to this problem is to amplify human judgement by leveraging the power of two competing AIs in a debate about the correct solution to a given problem. Prior theoretical work has provided a complexity-theoretic formalization of AI debate, and posed the problem of designing protocols for AI debate that guarantee the correctness of human judgements for as complex a class of problems as possible. Recursive debates, in which debaters decompose a complex problem into simpler subproblems, hold promise for growing the class of problems that can be accurately judged in a debate. However, existing protocols for recursive debate run into theobfuscated arguments problem: a dishonest debater can employ a computationally efficient strategy that forces an honest opponent to solve a computationally intractable problem to win.We mitigate this problem with a new recursive debate protocol that, under certain stability assumptions, ensures that an honest debater can win with a strategy requiring computational efficiency comparable to their opponent.



Paperid:594
Authors:Zhe Tao, Aditya V Thakur
Abstract:
Abstract:In explainable AI, DNN gradients are used to interpret the prediction; in safety-critical control systems, gradients could encode safety constraints; in scientific-computing applications, gradients could encode physical invariants. While recent work on provable editing of DNNs has focused on input-output constraints, the problem of enforcing hard constraints on DNN gradients remains unaddressed. We present ProGrad, the first efficient approach for editing the parameters of a DNN to provably enforce hard constraints on the DNN gradients. Given a DNN $\mathcal{N}$ with parameters $\theta$, an input $\mathrm{x}$, and convex polytopes $\mathrm{Q}_j$ denoting hard constraints on the gradients $\frac{\partial}{\partial \mathrm{x}}\mathcal{N}_j(\mathrm{x}; \theta)$ of the $j$-th output with respect to the input $\mathrm{x}$, ProGrad finds new parameters $\theta'$ such that $\bigwedge_j \frac{\partial}{\partial \mathrm{x}}\mathcal{N}_j(\mathrm{x}; \theta') \in \mathrm{Q}_j$ while minimizing the changes $\lVert\theta' - \theta\rVert$. The key contribution is a novel *conditional variable gradient* of DNN, which relaxes the NP-hard provable gradient editing problem to a linear program (LP), enabling ProGrad to use an LP solver to efficiently and effectively enforce the gradient constraints. We experimentally evaluated ProGrad via enforcing i) hard Grad-CAM constraints on an ImageNet ResNet152 DNN; ii) hard Integrated Gradients constraints on a Llama 3 LLM; iii) hard gradient constraints in training a DNN to approximate a target function as a proxy for safety constraints in control systems and physical invariants in scientific applications. The results highlight the unique capability of ProGrad in enforcing hard constraints on DNN gradients.



Paperid:595
Authors:Fengwei Teng, Quan Shi, Zhaoyang Yu, Jiayi Zhang, Yuyu Luo, Chenglin Wu, Zhijiang Guo
Abstract:
Large Language Models (LLMs) achieve superior performance through training-time scaling, and test-time scaling further enhances their capabilities by conducting effective reasoning during inference. However, as the scale of reasoning increases, existing test-time scaling methods suffer from accumulated historical information, which not only wastes computational resources but also interferes with effective reasoning. To address this issue, we observe that complex reasoning can be achieved by solving a series of independent and self-contained subquestions. These subquestions are essentially \textit{atomic questions}, exhibiting the memoryless property similar to Markov processes.Based on this observation, we propose Atom of Thoughts (\our), where each state transition consists of decomposing the current question into a dependency-based directed acyclic graph and contracting its subquestions, forming a simplified question that maintains answer equivalence with the original problem. This answer preservation enables the iterative \textit{decomposition-contraction} process to naturally form a meaningful Markov reasoning process.Furthermore, these atomic states can be seamlessly integrated into existing test-time scaling methods, enabling \our to serve as a plug-in enhancement for improving reasoning capabilities.



Paperid:596
Authors:Joachim Tesch, Giorgio Becherini, Prerana Achar, Anastasios Yiannakidis, Muhammed Kocabas, Priyanka Patel, Michael Black
Abstract:
Inferring 3D human motion from video remains a challenging problem with many applications. While traditional methods estimate the human in image coordinates, many applications require human motion to be estimated in world coordinates. This is particularly challenging when there is both human and camera motion. Progress on this topic has been limited by the lack of rich video data with ground truth human and camera movement. We address this with BEDLAM2.0, a new dataset that goes beyond the popular BEDLAM dataset in important ways. In addition to introducing more diverse and realistic cameras and camera motions, BEDLAM2.0 increases diversity and realism of body shape, motions, clothing, hair, and 3D environments. Additionally, it adds shoes, which were missing in BEDLAM. BEDLAM has become a key resource for training 3D human pose and motion regressors today and we show that BEDLAM2.0 is significantlybetter, particularly for training methods that estimate humans in world coordinates. We compare state-of-the art methods trained on BEDLAM and BEDLAM2.0, and find that BEDLAM2.0 significantly improves accuracy over BEDLAM. For research purposes, we provide the rendered videos, ground truth body parameters, and camera motions. We also provide the 3D assets to which we have rights and links to those from third parties. The review link is here and will be replaced by the public link upon acceptance: https://bedlam2.is.tuebingen.mpg.de/b2neurips2025review.html



Paperid:597
Authors:Yihong Sun, Xinyu Yang, Jennifer Sun, Bharath Hariharan
Abstract:
Real-world objects frequently undergo state transformations. From an apple being cut into pieces to a butterfly emerging from its cocoon, tracking through these changes is important for understanding real-world objects and dynamics. However, existing methods often lose track of the target object after transformation, due to significant changes to object appearance. To address this limitation, we introduce TubeletGraph, a zero-shot system that recovers missing objects after transformation and maps out how the object states are evolving over time. TubeletGraph first identifies potentially overlooked tracks, and determines whether they should be integrated based on semantic and proximity priors. Then, it reasons about the added tracks and generates a state graph describing each observed transformation. TubeletGraph achieves state-of-the-art tracking performance under transformations, while demonstrating deeper understandings of object transformations and promising capabilities in temporal grounding and semantic reasoning for complex object transformations.



Paperid:598
Authors:Yushen Zuo, Qi Zheng, Mingyang Wu, Xinrui Jiang, Renjie Li, Jian Wang, Yide Zhang, Gengchen Mai, Lihong Wang, James Zou, Xiaoyu Wang, Ming-Hsuan Yang, Zhengzhong Tu
Abstract:
Abstract:We present 4KAgent, a unified agentic super-resolution generalist system designed to universally upscale any image to 4K resolution. Our system can transform images from extremely low resolutions with severe degradations, for example, highly distorted inputs at $256\times 256$, into crystal clear, high-quality 4K outputs. 4KAgent comprises three core components: (1) Profiling, a module that customizes the 4KAgent pipeline based on bespoke use cases; (2) A Perception Agent, which leverages vision-language models alongside image quality assessment experts to analyze the input image and make a tailored restoration plan; and (3) A Restoration Agent, which executes the plan, following a recursive execution-reflection paradigm, guided by a quality-driven mixture-of-expert policy to select the optimal output for each step. Additionally, 4KAgent embeds a specialized face restoration pipeline, significantly enhancing facial details in portrait and selfie photos. We rigorously evaluate our 4KAgent across 12 distinct task categories encompassing a total of 26 diverse benchmarks, setting new state-of-the-art on a broad spectrum of imaging domains. Our evaluations cover natural images, portrait photos, AI-generated content, satellite imagery, fluorescence microscopy, and medical imaging, demonstrating superior performance in terms of both perceptual (e.g., NIQE, MUSIQ) and fidelity (e.g., PSNR) metrics. By establishing a novel agentic paradigm for low-level vision tasks, we aim to catalyze broader interest and innovation within vision-centric autonomous agents across diverse research communities.



Paperid:599
Authors:Zhiqiu Lin, Siyuan Cen, Daniel Jiang, Jay Karhade, Hewei Wang, Chancharik Mitra, Yu Tong Tiffany Ling, Yuhan Huang, Rushikesh Zawar, Xue Bai, Yilun Du, Chuang Gan, Deva Ramanan
Abstract:
We introduce CameraBench, a large-scale dataset and benchmark designed to assess and improve camera motion understanding. CameraBench consists of ~3,000 diverse internet videos, annotated by experts through a rigorous multi-stage quality control process. One of our core contributions is a taxonomy or "language" of camera motion primitives, designed in collaboration with cinematographers. We find, for example, that some motions like "follow" (or tracking) require understanding scene content like moving subjects. We conduct a large-scale human study to quantify human performance, revealing that domain expertise and tutorial-based training can significantly enhance accuracy. For example, a novice may confuse zoom-in (a change of intrinsics) with translating forward (a change of extrinsics), but can be trained to differentiate the two. Using CameraBench, we evaluate Structure-from-Motion (SfM) and Video-Language Models (VLMs), finding that SfM models struggle to capture semantic primitives that depend on scene content, while generative VLMs struggle to capture geometric primitives that require precise estimation of trajectories. We then fine-tune a generative VLM on CameraBench to achieve the best of both worlds and showcase its applications, including motion-augmented captioning, video question answering, and video-text retrieval. We hope our taxonomy, benchmark, and tutorials will drive future efforts towards the ultimate goal of understanding camera motions in any video.



Paperid:600
Authors:Jianqiao Zheng, Xueqian Li, Hemanth Saratchandran, Simon Lucey
Abstract:
Convolutional Neural Networks (CNNs) inherently encode strong inductive biases, enabling effective generalization on small-scale datasets. In this paper, we propose integrating this inductive bias into ViTs, not through an architectural intervention but solely through initialization.The motivation here is to have a ViT that can enjoy strong CNN-like performance when data assets are small, but can still scale to ViT-like performance as the data expands. Our approach is motivated by our empirical results that random impulse filters can achieve commensurate performance to learned filters within a CNN. We improve upon current ViT initialization strategies, which typically rely on empirical heuristics such as using attention weights from pretrained models or focusing on the distribution of attention weights without enforcing structures. Empirical results demonstrate that our method significantly outperforms standard ViT initialization across numerous small and medium-scale benchmarks, including Food-101, CIFAR-10, CIFAR-100, STL-10, Flowers, and Pets, while maintaining comparative performance on large-scale datasets such as ImageNet-1K. Moreover, our initialization strategy can be easily integrated into various transformer-based architectures such as Swin Transformer and MLP-Mixer with consistent improvements in performance.



Paperid:601
Authors:Erhan Xu, Kai Ye, Hongyi Zhou, Luhan Zhu, Francesco Quinzan, Chengchun Shi
Abstract:
This paper studies reinforcement learning from human feedback (RLHF) for aligning large language models with human preferences. While RLHF has demonstrated promising results, many algorithms are highly sensitive to misspecifications in the underlying preference model (e.g., the Bradley-Terry model), the reference policy, or the reward function, resulting in undesirable fine-tuning. To address model misspecification, we propose a doubly robust preference optimization algorithm that remains consistent when either the preference model or the reference policy is correctly specified (without requiring both). Our proposal demonstrates superior and more robust performance than state-of-the-art algorithms, both in theory and in practice.



Paperid:602
Authors:Xu Jie, Xuesong Zhang, Jing Jiang, Qinghua Cui
Abstract:
Particle Image Velocimetry (PIV) aims to infer underlying velocity fields from time-separated particle images, forming a PDE-constrained inverse problem governed by advection dynamics. Traditional cross-correlation methods and deep learning-based feature matching approaches often struggle with ambiguity, limited resolution, and generalization to real-world conditions. To address these challenges, we propose a PIV Neural Operator (PIVNO) framework that directly approximates the inverse mapping from paired particle images to flow fields within a function space. Leveraging a position informed Galerkin-style attention operator, PIVNO captures global flow structures while supporting resolution-adaptive inference across arbitrary subdomains. Moreover, to enhance real-world adaptability, we introduce a self-supervised fine-tuning scheme based on physical divergence constraints, enabling the model to generalize from synthetic to real experiments without requiring labeled data. Extensive evaluations demonstrate the accuracy, flexibility, and robustness of our approach across both simulated and experimental PIV datasets. We will release our code upon acceptance.



Paperid:603
Authors:Ari Karchmer, Seth Neel, Martin Pawelczyk
Abstract:
Abstract:Data attribution methods aim to answer useful counterfactual questions like "what would a ML model's prediction be if it were trained on a different dataset?'' However, estimation of data attribution models through techniques like empirical influence or "datamodeling" remains very computationally expensive.This causes a critical trust issue: if only a few computationally rich parties can obtain data attributions, how can resource-constrained parties trust that the provided attributions are indeed "good," especially when they are used for important downstream applications (e.g., data pricing)?In this paper, we address this trust issue by proposing an interactive verification paradigm for data attribution. An untrusted and computationally powerful Prover learns data attributions, and then engages in an interactive proof with a resource-constrained Verifier. Our main result is a protocol that provides formal completeness, soundness, and efficiency guarantees in the sense of Probably-Approximately-Correct (PAC) verification Goldwasser et al (2021). Specifically, if both Prover and Verifier follow the protocol, the Verifier accepts data attributions that are $\varepsilon$-close to the optimal data attributions (in terms of the Mean Squared Error) with probability $1-\delta$. Conversely, if the Prover arbitrarily deviates from the protocol, even with infinite compute, then this is detected (or it still yields data attributions to the Verifier) except with probability $\delta$. Importantly, our protocol ensures the Verifier’s workload scales only as $O(1/\varepsilon^2)$; i.e., independently of the dataset size.At a technical level, our results apply to efficiently verifying any linear function over the boolean hypercube computed by the Prover, making them broadly applicable to various attribution tasks.



Paperid:604
Authors:Yixiu Zhao, Jiaxin Shi, Feng Chen, Shaul Druckmann, Lester Mackey, Scott Linderman
Abstract:
Discrete diffusion has emerged as a powerful framework for generative modeling in discrete domains, yet efficiently sampling from these models remains challenging. Existing sampling strategies often struggle to balance computation and sample quality when the number of sampling steps is reduced, even when the model has learned the data distribution well. To address these limitations, we propose a predictor-corrector sampling scheme where the corrector is informed by the diffusion model to more reliably counter the accumulating approximation errors. To further enhance the effectiveness of our informed corrector, we introduce complementary architectural modifications based on hollow transformers and a simple tailored training objective that leverages more training signal. We use a synthetic example to illustrate the failure modes of existing samplers and show how informed correctors alleviate these problems. On the Text8 dataset, the informed corrector improves sample quality by generating text with significantly fewer errors than the baselines. On tokenized ImageNet 256x256, this approach consistently produces superior samples with fewer steps, achieving improved FID scores for discrete diffusion models. These results underscore the potential of informed correctors for fast and high-fidelity generation using discrete diffusion.



Paperid:605
Authors:Limeng Qiao, Yiyang Gan, Bairui Wang, Jie Qin, Shuang Xu, Siqi Yang, Lin Ma
Abstract:
Conventional Vision Transformer streamlines visual modeling by employing a uniform input resolution, which underestimates the inherent variability of natural visual data and incurs a cost in spatial-contextual fidelity. While preliminary explorations have superficially investigated native resolution modeling, existing works still lack systematic training recipe from the visual representation perspective. To bridge this gap, we introduce Unified Vision Transformer with Native Resolution, i.e. UniViTAR, a family of homogeneous vision foundation models tailored for unified visual modality and native resolution scenario in the era of multimodal. Our framework first conducts architectural upgrades to the vanilla paradigm by integrating multiple advanced components. Building upon these improvements, a progressive training paradigm is introduced, which strategically combines two core mechanisms: (1) resolution curriculum learning, transitioning from fixed-resolution pretraining to native resolution tuning, thereby leveraging ViT’s inherent adaptability to variable-length sequences, and (2) visual modality adaptation via inter-batch image-video switching, which balances computational efficiency with enhanced temporal reasoning. In parallel, a hybrid training framework further synergizes sigmoid-based contrastive loss with feature distillation from a frozen teacher model, thereby accelerating early-stage convergence. Finally, trained exclusively on public accessible image-caption data, our UniViTAR family across multiple model scales from 0.3B to 1B achieves state-of-the-art performance on a wide variety of visual-related tasks. The code and models will be available soon.



Paperid:606
Authors:Paul Krzakala, Gabriel Melo, Charlotte Laclau, Florence d'Alché-Buc, Rémi Flamary
Abstract:
Although graph-based learning has attracted a lot of attention, graph representation learning is still a challenging task whose resolution may impact key application fields such as chemistry or biology. To this end, we introduce GRALE, a novel graph autoencoder that encodes and decodes graphs of varying sizes into a shared embedding space. GRALE is trained using an Optimal Transport-inspired loss that compares the source and reconstructed graphs and leverages a differentiable matching module, which is trained jointly with the encoder and decoder. The proposed attention-based architecture relies on Evoformer, the core component of AlphaFold, which we extend to support both graph encoding and decoding. We show, in numerical experiments on simulated and molecular data, that GRALE enables a highly general form of pre-training, applicable to a wide range of downstream tasks, from classification and regression to more complex tasks such as graph interpolation, editing, matching, and prediction.



Paperid:607
Authors:Zhuangzhuang Jia, Yijie Wang, Roy Dong, Grani A. Hanasusanto
Abstract:
In performative stochastic optimization, decisions can influence the distribution of random parameters, rendering the data-generating process itself decision-dependent. In practice, decision-makers rarely have access to the true distribution map and must instead rely on imperfect surrogate models, which can lead to severely suboptimal solutions under misspecification. To address this challenge, we propose a distributionally robust framework for performative optimization that explicitly accounts for ambiguity in the decision-dependent distribution. Our framework introduces two modeling paradigms that capture a broad range of applications in machine learning and decision-making under uncertainty. This latter setting has not previously been explored in the performative optimization literature. To tackle the intractability of the resulting nonconvex objectives, we develop an iterative algorithm named repeated robust risk minimization, which alternates between solving a decision-independent distributionally robust optimization problem and updating the ambiguity set based on the previous decision. This decoupling ensures computational tractability at each iteration while enhancing robustness to model uncertainty. We provide reformulations compatible with off-the-shelf solvers and establish theoretical guarantees on convergence and suboptimality. Extensive numerical experiments in strategic classification, revenue management, and portfolio optimization demonstrate significant performance gains over state-of-the-art baselines, highlighting the practical value of our approach.



Paperid:608
Authors:Philipp Lindenberger, Eduard Trulls, Paul-Edouard Sarlin, Jan Hosang, Marc Pollefeys, Simon Lynen
Abstract:
Determining the precise geographic location of an image at a global scaleremains an unsolved challenge. Standard image retrieval techniques are simply not applicable, due to the sheer volume of images (>100M).Scalable solutions, however, involve a trade-off:global classification typically yields coarse results (10+ kilometers), while cross-view retrieval between ground and aerial imagery suffers from a domain gap and has been primarily studied on smaller geographic regions.This paper introduces a novel hybrid approach that achieves fine-grained geo-localization across a large geographic expanse at the scale of a continent.We leverage a proxy classification task during training to learn rich feature representations that implicitly encode precise location information.We combine these learned prototypes with embeddings of aerial imagery to increase robustness to the sparsity of ground-level data.This enables direct, fine-grained retrieval over areas spanning multiple countries.Our extensive evaluationdemonstrates that our approach can achieve >50% top-1 recall at 100 meters over a substantial part of Europe.



Paperid:609
Authors:Luhuan Wu, Yi Han, Christian Andersson Naesseth, John Cunningham
Abstract:
We propose a novel sequential Monte Carlo (SMC) method for sampling from unnormalized target distributions based on a reverse denoising diffusion process. While recent diffusion-based samplers simulate the reverse diffusion using approximate score functions, they can suffer from accumulating errors due to time discretization and imperfect score estimation. In this work, we introduce a principled SMC framework that formalizes diffusion-based samplers as proposals while systematically correcting for their biases. The core idea is to construct informative intermediate target distributions that progressively steer the sampling trajectory toward the final target distribution. Although ideal intermediate targets are intractable, we develop \emph{exact} approximations using quantities from the score estimation-based proposal, without requiring additional model training or inference overhead. The resulting sampler, termed \textit{\ourmethodfull}, enables consistent sampling and unbiased estimation of the target's normalization constant under mild conditions. We demonstrate the effectiveness of our method on a range of synthetic targets and real-world Bayesian inference problems.



Paperid:610
Authors:Amit Kumar, Nisheeth K. Vishnoi
Abstract:
Abstract:We study a two-institution stable matching model in which candidates from two distinct groups are evaluated using partially correlated signals that are group-biased. This extends prior work (which assumes institutions evaluate candidates in an identical manner) to a more realistic setting in which institutions rely on overlapping, but independently processed, criteria. These evaluations could consist of a variety of informative tools such as standardized tests, shared recommendation systems, or AI-based assessments with local noise. Evaluations are governed by two key parameters: the bias parameter $\beta \in (0,1]$, which models systematic disadvantage faced by one group, and the correlation parameter $\gamma \in [0,1]$, which captures the alignment between institutional rankings. We study the representation ratio $\mathcal{R}(\beta, \gamma)$, i.e., the ratio of disadvantaged to advantaged candidates selected by the matching process in this setting. Focusing on a regime in which all candidates prefer the same institution, we characterize the large-market equilibrium and derive a closed-form expression for the resulting representation ratio. Prior work shows that when $\gamma = 1$, this ratio scales linearly with $\beta$. In contrast, we show that $\mathcal{R}(\beta, \gamma)$ increases nonlinearly with $\gamma$ and even modest losses in correlation can cause sharp drops in the representation ratio. Our analysis identifies critical $\gamma$-thresholds where institutional selection behavior undergoes discrete transitions, and reveals structural conditions under which evaluator alignment or bias mitigation are most effective. Finally, we show how this framework and results enable interventions for fairness-aware design in decentralized selection systems.



Paperid:611
Authors:Jiangwei Chen, Kieu Thao Nguyen Pham, Rachael Sim, Arun Verma, Zhaoxuan Wu, Chuan Sheng Foo, Bryan Kian Hsiang Low
Abstract:
In collaborative data sharing and machine learning, multiple parties aggregate their data resources to train a machine learning model with better model performance. However, as the parties incur data collection costs, they are only willing to do so when guaranteed incentives, such as fairness and individual rationality. Existing frameworks assume that all parties join the collaboration simultaneously, which does not hold in many real-world scenarios. Due to the long processing time for data cleaning, difficulty in overcoming legal barriers, or unawareness, the parties may join the collaboration at different times. In this work, we propose the following perspective: As a party who joins earlier incurs higher risk and encourages the contribution from other wait-and-see parties, that party should receive a reward of higher value for sharing data earlier. To this end, we propose a fair and time-aware data sharing framework, including novel time-aware incentives.We develop new methods for deciding reward values to satisfy these incentives. We further illustrate how to generate model rewards that realize the reward values and empirically demonstrate the properties of our methods on synthetic and real-world datasets.



Paperid:612
Authors:Shiguang Wu, Yaqing Wang, Yatao Bian, Quanming Yao
Abstract:
Meta-learning enables learning systems to adapt quickly to new tasks, similar to humans.Different meta-learning approaches all work under/with the mini-batch episodic training framework. Such framework naturally gives the information about task identity, which can serve as additional supervision for meta-training to improve generalizability. We propose to exploit task identity as additional supervision in meta-training, inspired by the alignment and discrimination ability which is is intrinsic in human's fast learning.This is achieved by contrasting what meta-learners learn, i.e., model representations.The proposed ConML is evaluating and optimizing the contrastive meta-objective under a problem- and learner-agnostic meta-training framework.We demonstrate that ConML integrates seamlessly with existing meta-learners, as well as in-context learning models, and brings significant boost in performance with small implementation cost.Code is provided at https://anonymous.4open.science/r/conml_ano-3372.



Paperid:613
Authors:Jean-Baptiste Fermanian, Pierre Humbert, Gilles Blanchard
Abstract:
Abstract:We introduce a method based on Conformal Prediction (CP) to quantify the uncertainty of full ranking algorithms. We focus on a specific scenario where $n+m$ items are to be ranked by some ``black box'' algorithm. It is assumed that the relative (ground truth) ranking of $n$ of them is known. The objective is then to quantify the error made by the algorithm on the ranks of the $m$ new items among the total $(n+m)$.In such a setting, the true ranks of the $n$ original items in the total $(n+m)$ depend on the (unknown) true ranks of the $n$ new ones. Consequently, we have no direct access to a calibration set to apply a classical CP method. To address this challenge, we propose to construct distribution-free bounds of the unknown conformity scores using recent results on the distribution of conformal p-values. Using these scores upper bounds, we provide valid prediction sets for the rank of any item. We also control the false coverage proportion, a crucial quantity when dealing with multiple prediction sets. Finally, we empirically show on both synthetic and real data the efficiency of our CP method for state-of-the-art algorithms such as RankNet or LambdaMart.



Paperid:614
Authors:Kevin Du, Yash Nair, Lucas Janson
Abstract:
Uncertainty quantification (UQ) for adaptively collected data, such as that coming from adaptive experiments, bandits, or reinforcement learning, is necessary for critical elements of data collection such as ensuring safety and conducting after-study inference. The data's adaptivity creates significant challenges for frequentist UQ, yet Bayesian UQ remains the same as if the data were independent and identically distributed (i.i.d.), making it an appealing and commonly used approach. Bayesian UQ requires the (correct) specification of a prior distribution while frequentist UQ does not, but for i.i.d. data the celebrated Bernstein–von Mises theorem shows that as the sample size grows, the prior `washes out' and Bayesian UQ becomes frequentist-valid, implying that the choice of prior need not be a major impediment to Bayesian UQ as it makes no difference asymptotically. This paper for the first time extends the Bernstein–von Mises theorem to adaptively collected data, proving asymptotic equivalence between Bayesian UQ and Wald-type frequentist UQ in this challenging setting. Our results do not require the standard stability condition for validity of Wald-type frequentist UQ, and thus provide positive results on frequentist validity of Bayesian UQ under stability. Counterintuitively however, they also provide a negative result that Bayesian UQ is not asymptotically frequentist valid when stability fails, despite the fact that the prior washes out and Bayesian UQ asymptotically matches standard Wald-type frequentist UQ. We empirically validate our theory (positive and negative) via a range of simulations.



Paperid:615
Authors:Salman Rahman, Issaka, Ashima Suvarna, Genglin Liu, James Shiffer, jaeyoung lee, Md Rizwan Parvez, Hamid Palangi, Shi Feng, Nanyun Peng, Yejin Choi, Julian Michael, Liwei Jiang, Saadia Gabriel
Abstract:
As AI grows more powerful, it will increasingly shape how we understand the world. But with this influence comes the risk of amplifying misinformation and deepening social divides—especially on consequential topics like public health where factual accuracy directly impacts well-being. Scalable Oversight aims to ensure AI truthfulness by enabling humans to supervise systems that may exceed human capabilities---yet humans themselves hold different beliefs and biases that impair their judgment. We study whether AI debate can guide biased judges toward the truth by having two AI systems debate opposing sides of controversial COVID-19 factuality claims where people hold strong prior beliefs. We conduct two studies: one with human judges holding either mainstream or skeptical beliefs evaluating factuality claims through AI-assisted debate or consultancy protocols, and a second examining the same problem with personalized AI judges designed to mimic these different human belief systems. In our human study, we find that debate—where two AI advisor systems present opposing evidence-based arguments—consistently improves judgment accuracy and confidence calibration, outperforming consultancy with a single-advisor system by 10\% overall. The improvement is most significant for judges with mainstream beliefs (+15.2\% accuracy), though debate also helps skeptical judges who initially misjudge claims move toward accurate views (+4.7\% accuracy). In our AI judge study, we find that AI judges with human-like personas achieve even higher accuracy (78.5\%) than human judges (70.1\%) and default AI judges without personas (69.8\%), suggesting their potential for supervising frontier AI models. These findings highlight AI debate as a promising path toward scalable, bias-resilient oversight---leveraging both diverse human and AI judgments to move closer to truth in contested domains.



Paperid:616
Authors:Nguyen Hung-Quang, Hoang Phan, Khoa D Doan
Abstract:
Diffusion models have shown remarkable abilities in generating realistic and high-quality images from text prompts. However, a trained model remains largely black-box; little do we know about the roles of its components in exhibiting a concept such as objects or styles. Recent works employ causal tracing to localize knowledge-storing layers in generative models without showing how other layers contribute to the target concept. In this work, we approach diffusion models' interpretability problem from a more general perspective and pose a question: \textit{``How do model components work jointly to demonstrate knowledge?''}. To answer this question, we decompose diffusion models using component attribution, systematically unveiling the importance of each component (specifically the model parameter) in generating a concept. The proposed framework, called \textbf{C}omponent \textbf{A}ttribution for \textbf{D}iffusion Model (CAD), discovers the localization of concept-inducing (positive) components, while interestingly uncovers another type of components that contribute negatively to generating a concept, which is missing in the previous knowledge localization work. Based on this holistic understanding of diffusion models, we present and empirically evaluate one utility of component attribution in controlling the generation process. Specifically, we introduce two fast, inference-time model editing algorithms, CAD-Erase and CAD-Amplify; in particular, CAD-Erase enables erasure and CAD-Amplify allows amplification of a generated concept by ablating the positive and negative components, respectively, while retaining knowledge of other concepts. Extensive experimental results validate the significance of both positive and negative components pinpointed by our framework, demonstrating the potential of providing a complete view of interpreting generative models.



Paperid:617
Authors:Rohit Gandikota, Sheridan Feucht, Samuel Marks, David Bau
Abstract:
In this work, we introduce Erasure of Language Memory (ELM), a principled approach to concept-level unlearning that operates by matching distributions defined by the model's own introspective classification capabilities. Our key insight is that effective unlearning should leverage the model's ability to evaluate its own knowledge, using the language model itself as a classifier to identify and reduce the likelihood of generating content related to undesired concepts. ELM applies this framework to create targeted low-rank updates that reduce generation probabilities for concept-specific content while preserving the model's broader capabilities. We demonstrate ELM's efficacy on biosecurity, cybersecurity, and literary domain erasure tasks. Comparative evaluation reveals that ELM-modified models achieve near-random performance on assessments targeting erased concepts, while simultaneously preserving generation coherence, maintaining benchmark performance on unrelated tasks, and exhibiting strong robustness to adversarial attacks.



Paperid:618
Authors:Huaisheng Zhu, Teng Xiao, Shijie Zhou, Zhimeng Guo, Hangfan Zhang, Siyuan Xu, Vasant Honavar
Abstract:
Diffusion models have established themselves as leading techniques for image generation. However, their reliance on an iterative denoising process results in slow sampling speeds, which limits their applicability to interactive and creative applications. An approach to overcoming this limitation involves distilling multistep diffusion models into efficient one-step generators. However, existing distillation methods typically suffer performance degradation or require complex iterative training procedures which increase their complexity and computational cost. In this paper, we propose Contrastive Energy Distillation (CED), a simple yet effective approach to distill multistep diffusion models into effective one-step generators. Our key innovation is the introduction of an unnormalized joint energy-based model (EBM) that represents the generator and an auxiliary score model. CED optimizes a Noise Contrastive Estimation (NCE) objective to efficiently transfers knowledge from a multistep teacher diffusion model without additional modules or iterative training complexity. We further show that CED implicitly optimizes the KL divergence between the distributions modeled by the multistep diffusion model and the one-step generator. We present results of experiments which demonstrate that CED achieves competitive performance with the representative baselines for distilling multistep diffusion models while maintaining excellent memory efficiency.



Paperid:619
Authors:Xiangwen Wang, Yibo Jacky Zhang, Zhoujie Ding, Katherine Tsai, Haolun Wu, Sanmi Koyejo
Abstract:
Compound AI systems, comprising multiple interacting components such as LLMs, foundation models, and external tools, have demonstrated remarkable improvements compared to single models in various tasks. To ensure their effective deployment in real-world applications, aligning these systems with human preferences is crucial. However, aligning the compound system via policy optimization, unlike the alignment of a single model, is challenging for two main reasons: (i) non-differentiable interactions between components make end-to-end gradient-based optimization method inapplicable, and (ii) system-level preferences cannot be directly transformed into component-level preferences. To address these challenges, we first formulate compound AI systems as Directed Acyclic Graphs (DAGs), explicitly modeling both component interactions and the associated data flows. Building on this formulation, we introduce SysDPO, a framework that extends Direct Preference Optimization (DPO) to enable joint system-level alignment. We propose two variants, SysDPO-Direct and SysDPO-Sampling, tailored for scenarios depending on whether we construct a system-specific preference dataset. We empirically demonstrate the effectiveness of our approach across two applications: the joint alignment of a language model and a diffusion model, and the joint alignment of an LLM collaboration system.



Paperid:620
Authors:Hangoo Kang, Jehyeok Yeon, Gagandeep Singh
Abstract:
Autonomous agentic AI systems powered by vision-language models (VLMs) are rapidly advancing toward real-world deployment, yet their cross-modal reasoning capabilities introduce new attack surfaces for adversarial manipulation that exploit semantic reasoning across modalities. Existing adversarial attacks typically rely on visible pixel perturbations or require privileged model or environment access, making them impractical for stealthy, real-world exploitation. We introduce TRAP, a generative adversarial framework that manipulates the agent's decision-making using diffusion-based semantic injections. Our method combines negative prompt–based degradation with positive semantic optimization, guided by a Siamese semantic network and layout-aware spatial masking. Without requiring access to model internals, TRAP produces visually natural images yet induces consistent selection biases in agentic AI systems. We evaluate TRAP on the Microsoft Common Objects in Context(COCO) dataset, building multi-candidate decision scenarios. Across these scenarios, TRAP achieves a 100\% attack success rate on leading models, including LLaVA-34B, Gemma3, and Mistral-3.1, significantly outperforming baselines such as SPSA, Bandit, and standard diffusion approaches. These results expose a critical vulnerability: Autonomous agents can be consistently misled through human-imperceptible cross-modal manipulations. These findings highlight the need for defense strategies beyond pixel-level robustness to address semantic vulnerabilities in cross-modal decision-making.



Paperid:621
Authors:Samira Goudarzi, Kiarash Banihashem, MohammadTaghi Hajiaghayi, Peyman Jabbarzade, Morteza Monemizadeh
Abstract:
Abstract:Submodular maximization subject to a $p$-matchoid constraint has various applications in machine learning, particularly in tasks such as feature selection, video and text summarization, movie recommendation, graph-based learning, and constraint-based optimization. We study this problem in the dynamic setting, where a sequence of insertions and deletions of elements to a $p$-matchoid $\mathcal{M}(\mathcal{V},\mathcal{I})$ occurs over time and the goal is to efficiently maintain an approximate solution.We propose a dynamic algorithm for non-monotone submodular maximization under a $p$-matchoid constraint. For a $p$-matchoid $\mathcal{M}(\mathcal{V},\mathcal{I})$ of rank $k$, defined by a collection of $m$ matroids, our algorithm guarantees a $(2p + 2\sqrt{p(p+1)} + 1 + \epsilon)$-approximate solution at any time $t$ in the update sequence, with an expected amortized query complexity of $O(\epsilon^{-3} pk^4 \log^2(k))$ per update.



Paperid:622
Authors:Uros Zivanovic, Serafina Di Gioia, Andre Scaffidi, Martín de los Rios, Gabriella Contardo, Roberto Trotta
Abstract:
Applying Transformers to irregular time-series typically requires specializations to their baseline architecture, which can result in additional computational overhead and increased method complexity. We present the Rotary Masked Autoencoder (RoMAE), utilizing the popular Rotary Positional Embedding (RoPE) method for continuous positions. RoMAE is an extension to the Masked Autoencoder (MAE) that enables representation learning with continuous positional information while avoiding any time-series-specific architectural specializations. We showcase RoMAE's performance on a variety of modalities including irregular and multi-variate time-series, images, and audio, showing that RoMAE surpasses specialized time-series architectures on difficult datasets such as the DESC ELAsTiCC Challenge while maintaining MAE's usual performance across other modalities. We additionally investigate RoMAE's ability to reconstruct the embedded continuous positions. We also demonstrate that including learned embeddings in the input sequence breaks RoPE's relative position property.



Paperid:623
Authors:Ge Zheng, Jiajin Tang, Jiaye Qian, Hanzhuo Huang, Cheng Shi, Sibei Yang
Abstract:
Large language models (LLMs) and vision-language models (LVLMs) have driven the paradigm shift towards general-purpose foundation models. However, both of them are prone to hallucinations, which compromise their factual accuracy and reliability. While existing research primarily focuses on isolated textual- or visual-centric errors, a critical yet underexplored phenomenon persists in LVLMs: Even neither of textual- or visual centric errors occur, LVLMs often struggle with a new and subtle hallucination mode that arising from composition of them. In this paper, we define this issue as Simple Compositional Hallucination (SCHall). Through an preliminary analysis, we present two key findings: (1) visual abstraction fails under compositional questioning, and (2) visual inputs induce degradation in language processing, leading to hallucinations. To facilitate future research on this phenomenon, we introduce a customized benchmark, VLComp, and propose a novel VLR-distillation method, which serves as the first baseline to effectively mitigate SCHall. Furthermore, experiment results on publicly available benchmarks, including both hallucination-specific and general-purpose ones, demonstrate the effectiveness of our VLR-distillation method.



Paperid:624
Authors:Yuwen Ji, Luodan Zhang, Ambyer han, Yue Zhang, Haoran Que, Lei Shi, Wang Chao
Abstract:
Recent advances in retrieval-based in-context learning (ICL) train the retriever as a classification model, which categorizes in-context examples (ICEs) into the most useful and the rest based on absolute scores.However, during inference, ICEs areretrieved by score rankingrather than classification --- The classification training objective deviates from this test scenario.Hence, in this paper, we proposeSE-DPO--- a novel algorithm that trains a retrieval model by ranking formulation, where the preference rankings between ICEs are given by comparing the likelihood of the LLM generating the correct answer conditioned on each exemplar.By learning to rank, we motivate the retriever to automatically learn diverse rationales why specific examples are more useful for ICL decisions.This addresses the issue that classification models poorly capture broader utility.Experimental results demonstrate the top-1 performance of our proposal across 9 NLP tasks, with ablation studies and case studies further validating the effectiveness of our design.



Paperid:625
Authors:Vedant Gupta, Haotian Fu, Calvin Luo, Yiding Jiang, George Konidaris
Abstract:
We present DEPS, an end-to-end algorithm for discovering parameterized skills from expert demonstrations. Our method learns parameterized skill policies jointly with a meta-policy that selects the appropriate discrete skill and continuous parameters at each timestep. Using a combination of temporal variational inference and information-theoretic regularization methods, we address the challenge of degeneracy common in latent variable models, ensuring that the learned skills are temporally extended, semantically meaningful, and adaptable. Our empirical results show that learning parameterized skills from multitask expert demonstrations significantly improves generalization to unseen tasks. Our method outperforms multitask as well as skill learning baselines on both LIBERO and MetaWorld benchmarks. We also demonstrate that our approach discovers interpretable parameterized skills, such as an object grasping skill whose continuous arguments define the grasp location.



Paperid:626
Authors:Max Hamilton, Jinlin Lai, Wenlong Zhao, Subhransu Maji, Daniel Sheldon
Abstract:
AI has the potential to transform scientific discovery by analyzing vast datasets with little human effort. However, current workflows often do not provide the accuracy or statistical guarantees that are needed. We introduce \emph{active measurement}, a human-in-the-loop AI framework for scientific measurement. An AI model is used to predict measurements for individual units, which are then sampled for human labeling using importance sampling. With each new set of human labels, the AI model is improved and an unbiased Monte Carlo estimate of the total measurement is refined. Active measurement can provide precise estimates even with an imperfect AI model, and requires little human effort when the AI model is very accurate. We derive novel estimators, weighting schemes, and confidence intervals, and show that active measurement reduces estimation error compared to alternatives in several measurement tasks.



Paperid:627
Authors:Jingkang Wang, Henry Che, Yun Chen, Ze Yang, Lily Goli, Sivabalan Manivasagam, Raquel Urtasun
Abstract:
Reconstructing large-scale dynamic scenes from visual observations is a fundamental challenge in computer vision, with critical implications for robotics and autonomous systems. While recent differentiable rendering methods such as Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) have achieved impressive photorealistic reconstruction, they suffer from scalability limitations and require annotations to decouple actor motion. Existing self-supervised methods attempt to eliminate explicit annotations by leveraging motion cues and geometric priors, yet they remain constrained by per-scene optimization and sensitivity to hyperparameter tuning. In this paper, we introduce Flux4D, a simple and scalable framework for 4D reconstruction of large-scale dynamic scenes. Flux4D directly predicts 3D Gaussians and their motion dynamics to reconstruct sensor observations, in a fully unsupervised manner. By adopting only photometric losses and enforcing an "as static as possible" regularization, Flux4D learns to decompose dynamic elements directly from raw data without requiring pre-trained supervised models or foundational priors simply by training across many scenes. Our approach enables efficient reconstruction of dynamic scenes within seconds, scales effectively to large datasets, and generalizes well to unseen environments, including rare and unknown objects. Experiments on outdoor driving datasets show Flux4D significantly outperforms existing methods in scalability, generalization, and reconstruction quality.



Paperid:628
Authors:Dongsheng Yang, Austin Li, Kai Li, Wyatt Lloyd
Abstract:
Prefix caching is a key technique for reducing Large Language Model (LLM) inference costs. However, the prevalent least-recently-used (LRU) eviction algorithm has a large gap to the optimal algorithm. This paper introduces LPC, the first learned method to perform LLM prefix cache eviction. LPC leverages conversational content analysis to provide predictive guidance for eviction, determining which conversations are likely to continue. These insights, combined with last access timestamps, inform more effective cache management. Extensive evaluations across three real-world datasets demonstrate that LPC achieves 18-47% reductions in required cache sizes for equivalent hit ratios and has an 11% improvement in LLM prefilling throughput in an emulated environment.



Paperid:629
Authors:Jiacheng Cen, Anyi Li, Ning Lin, Tingyang Xu, Yu Rong, Deli Zhao, Zihe Wang, Wenbing Huang
Abstract:
Equivariant Graph Neural Networks (GNNs) have demonstrated significant success across various applications. To achieve completeness---that is, the universal approximation property over the space of equivariant functions---the network must effectively capture the intricate multi-body interactions among different nodes. Prior methods attain this via deeper architectures, augmented body orders, or increased degrees of steerable features, often at high computational cost and without polynomial-time solutions. In this work, we present a theoretically grounded framework for constructing complete equivariant GNNs that is both efficient and practical. We prove that a complete equivariant GNN can be achieved through two key components: 1) a complete scalar function, referred to as the canonical form of the geometric graph; and 2) a full-rank steerable basis set. Leveraging this finding, we propose an efficient algorithm for constructing complete equivariant GNNs based on two common models: EGNN and TFN. Empirical results demonstrate that our model demonstrates superior completeness and excellent performance with only a few layers, thereby significantly reducing computational overhead while maintaining strong practical efficacy.



Paperid:630
Authors:Shahrzad Haddadan, Sara Ahmadian
Abstract:
Abstract:The classic Mallows model is a foundational tool for modeling user preferences. However, it has limitations in capturing real-world scenarios, where users often focus only on a limited set of top choices and are indifferent to the rest. To address this, extensions such as the top-$k$ Mallows model have been proposed, aligning better with practical applications.In this paper, we address several challenges related to generalized top-$k$ Mallows models, with a focus on analyzing buyer preferences and choices. Our key contributions are: (1) a novel sampling scheme tailored to generalized top-$k$ Mallows models, (2) an efficient algorithm for computing choice probabilities under this model, and (3) a method for learning the model parameters from observed choice data. These contributions provide new tools for analysis and prediction in critical decision-making scenarios.We present rigorous mathematical analysis for the performance of our algorithms. Furthermore, through extensive experiments on synthetic data and real world data we demonstrate the scalability and accuracy of our proposed methods and we compare the predictive power of Mallows model for top-$k$ lists compared to the simpler Multinomial Logit model.



Paperid:631
Authors:Andre Kassis, Urs Hengartner, Yaoliang Yu
Abstract:
Diffusion-based purification (DBP) has become a cornerstone defense against adversarial examples (AEs), regarded as robust due to its use of diffusion models (DMs) that project AEs onto the natural data manifold. We refute this core claim, theoretically proving that gradient-based attacks effectively target the DM rather than the classifier, causing DBP's outputs to align with adversarial distributions. This prompts a reassessment of DBP's robustness, accrediting it two critical flaws: incorrect gradients and inappropriate evaluation protocols that test only a single random purification of the AE. We show that with proper accounting for stochasticity and resubmission risk, DBP collapses. To support this, we introduce DiffBreak, the first reliable toolkit for differentiation through DBP, eliminating gradient flaws that previously further inflated robustness estimates. We also analyze the current defense scheme used for DBP where classification relies on a single purification, pinpointing its inherent invalidity. We provide a statistically grounded majority-vote (MV) alternative that aggregates predictions across multiple purified copies, showing partial but meaningful robustness gain. We then propose a novel adaptation of an optimization method against deepfake watermarking, crafting systemic perturbations that defeat DBP even under MV, challenging DBP's viability.



Paperid:632
Authors:Peter Chen, Xi Chen, Wotao Yin, Tianyi Lin
Abstract:
Direct alignment methods are increasingly used for aligning large language models (LLMs) with human preferences. However, these methods suffer from the issues of \textit{verbosity} and \textit{likelihood displacement}, which can be driven by the noisy preference pairs that induce similar likelihood for preferred and dispreferred responses. The contributions of this paper are two-fold. First, we propose a new preference alignment method based on comparison oracles and provide the convergence guarantee for its basic scheme. Second, we improve our method using some heuristics and conduct the experiments to demonstrate the flexibility and compatibility of practical scheme in improving the performance of LLMs using noisy preference pairs. Evaluations are conducted across multiple base and instruction-tuned models (Mistral-7B, Llama-3-8B and Gemma-2-9B) with benchmarks (AlpacaEval 2, MT-Bench and Arena-Hard). Experimental results show the effectiveness of our method as an alternative to addressing the limitations of existing direct alignment methods. A highlight of our work is that we evidence the importance of designing specialized methods for preference pairs with distinct likelihood margin, which complements the recent findings in Razin et al., [71].



Paperid:633
Authors:Qian Li, Yuyi Wang
Abstract:
Abstract:We prove that any Turing machine running on inputs of arbitrary length can be simulated by a constant bit-size transformer, as long as the context window is sufficiently long. This improves previous works, which require scaling up either the model's precision or the number of parameters on longer inputs. Furthermore, we prove that the complexity class SPACE$[s(n)]$ exactly characterizes the expressive power of a constant bit-size transformer with a context window of length $s(n)$. Our approach relies on simulating Post machines, a Turing-complete computational model. Post machines can be modeled as automata equipped with a queue, exhibiting computational behaviors naturally aligned with those of transformers. The behavioral similarity between transformers and Post machines may offer new insights into the mechanisms underlying the reasoning abilities of transformers.



Paperid:634
Authors:Tianchen Zhao, Xuanbai Chen, Zhihua Li, Jun Fang, DONGSHENG An, Xiang Xu, Zhuowen Tu, Yifan Xing
Abstract:
Recent generative data augmentation methods conditioned on both image and text prompts struggle to balance between fidelity and diversity, as it is challenging to preserve essential image details while aligning with varied text prompts. This challenge arises because representations in the synthesis process often become entangled with non-essential input image attributes such as environmental contexts, creating conflicts with text prompts intended to modify these elements.To address this, we propose a personalized image generation framework that uses a salient concept-aware image embedding model to reduce the influence of irrelevant visual details during the synthesis process, thereby maintaining intuitive alignment between image and text inputs.By generating images that better preserve class-discriminative features with additional controlled variations, our framework effectively enhances the diversity of training datasets and thereby improves the robustness of downstream models.Our approach demonstrates superior performance across eight fine-grained vision datasets, outperforming state-of-the-art augmentation methods with averaged classification accuracy improvements by 0.73\% and 6.5\% under conventional and long-tail settings, respectively.



Paperid:635
Authors:Xinyuan Wang, Bowen Wang, Dunjie Lu, Junlin Yang, Tianbao Xie, Junli Wang, Jiaqi Deng, Xiaole Guo, Zhennan Shen, Zhuokai Li, Ryan Li, Xiaochuan Li, Junda Chen, Boyuan Zheng, LI PEIHANG, Fangyu Lei, Chen Wu, Ruisheng Cao, Yeqiao Fu, Dongchan Shin, Martin Shin, Hu Jiarui, Yuyan Wang, Jixuan Chen, Yuxiao Ye, Yiheng Xu, Danyang Zhang, Yipu Wang, Heng Wang, Diyi Yang, Victor Zhong, Y.Charles, Zhilin Yang, Tao Yu
Abstract:
Vision-language models have demonstrated impressive capabilities as computer-use agents (CUAs) capable of automating diverse computer tasks. As their commercial potential grows, critical details of the most capable CUA systems remain closed and proprietary. As these agents will increasingly mediate digital interactions and execute consequential decisions on our behalf, the research community needs access to truly open CUA frameworks to study their capabilities, limitations, and risks. To bridge this gap, we propose AgentNet, a comprehensive open-source framework for scaling CUA data and foundation models. Our framework consists of: (1) an annotation infrastructure that seamlessly captures human computer-use demonstrations; (2) AgentNet dataset, a dataset of 27K computer-use data samples spanning various operating systems, applications, and websites; (3) a pipeline that discretizes continuous actions into state-action pairs and synthesizes reflective long chain-of-thought (CoT) reasoning; (4) a training recipe for scalable CUA modeling; and (5) AgentNetBench, a multi-dimensional offline benchmark for faster CUA evaluation. Our AgentNet-7B, fine-tuned on AgentNet dataset, demonstrates strong performance on several CUA benchmarks, achieving a success rate of 20.1% on OSWorld and 21.1% on WindowsAgentArena. Our training recipe, particularly its advanced reasoning mechanisms and strategic data mixture, enables robust performance scaling with increased data size. Further in-depth analysis of our models also demonstrate strong cross-domain generalization and performance scaling with test-time compute. We will release the annotation tool, datasets, code, and models to build open foundations for further CUA research.



Paperid:636
Authors:Junke Wang, Xun Wang, Qiushan Guo, Peize Sun, Weilin Huang, Zuxuan Wu, Yu-Gang Jiang
Abstract:
Autoregressive (AR) models have demonstrated strong potential in visual generation, offering competitive performance with simple architectures and optimization objectives. However, existing methods are typically limited to single-modality conditions, \eg, text or category labels, restricting their applicability in real-world scenarios that demand image synthesis from diverse forms of controls. In this work, we present \system, the first unified autoregressive framework for Any-to-Image generation. By discretizing various visual conditions through a shared visual tokenizer and text prompts with a text tokenizer, \system supports a broad spectrum of conditional inputs within a single model, including text (text-to-image generation), spatial signals (segmentation-to-image and depth-to-image), and visual context (image editing, frame prediction, and text-to-video generation). To mitigate the risk of information leakage from condition tokens to content tokens, we introduce Disentangled Causal Attention (DCA), which separates the full-sequence causal mask into condition causal attention and content causal attention. It serves as a training-time regularizer without affecting the standard next-token prediction during inference. With this design, \system achieves new state-of-the-art results across a range of benchmark, \eg, 0.63 on GenEval and 80.02 on VBench, demonstrating its effectiveness in flexible and high-fidelity visual generation.



Paperid:637
Authors:Zhanwei Zhang, kaiyuan liu, Junjie Liu, Wenxiao Wang, Binbin Lin, Liang Xie, Chen Shen, Deng Cai
Abstract:
Abstract:Local geometry-controllable computer-aided design (CAD) generation aims to modify local parts of CAD models automatically, enhancing design efficiency. It also ensures that the shapes of newly generated local parts follow user-specific geometric instructions (e.g., an isosceles right triangle or a rectangle with one corner cut off).However, existing methods encounter challenges in achieving this goal.Specifically, they either lack the ability to follow textual instructions or are unable to focus on the local parts.To address this limitation, we introduce GeoCAD, a user-friendly and local geometry-controllable CAD generation method. Specifically, we first propose a complementary captioning strategy to generate geometric instructions for local parts.This strategy involves vertex-based and VLLM-based captioning for systematically annotating simple and complex parts, respectively.In this way, we caption $\sim$221k different local parts in total.In the training stage, given a CAD model, we randomly mask a local part.Then, using its geometric instruction and the remaining parts as input, we prompt large language models (LLMs) to predict the masked part.During inference, users can specify any local part for modification while adhering to a variety of predefined geometric instructions.Extensive experiments demonstrate the effectiveness of GeoCAD in generation quality, validity and text-to-CAD consistency.



Paperid:638
Authors:Luckeciano Carvalho Melo, Alessandro Abate, Yarin Gal
Abstract:
Machine Learning models in real-world applications must continuously learn new tasks to adapt to shifts in the data-generating distribution. Yet, for Continual Learning (CL), models often struggle to balance learning new tasks (plasticity) with retaining previous knowledge (memory stability). Consequently, they are susceptible to Catastrophic Forgetting, which degrades performance and undermines the reliability of deployed systems. In the Bayesian CL literature, variational methods tackle this challenge by employing a learning objective that recursively updates the posterior distribution while constraining it to stay close to its previous estimate. Nonetheless, we argue that these methods may be ineffective due to compounding approximation errors over successive recursions. To mitigate this, we propose new learning objectives that integrate the regularization effects of multiple previous posterior estimations, preventing individual errors from dominating future posterior updates and compounding over time. We reveal insightful connections between these objectives and Temporal-Difference methods, a popular learning mechanism in Reinforcement Learning and Neuroscience. Experiments on challenging CL benchmarks show that our approach effectively mitigates Catastrophic Forgetting, outperforming strong Variational CL methods.



Paperid:639
Authors:Fabrizio Boncoraglio, Emanuele Troiani, Vittorio Erba, Lenka Zdeborová
Abstract:
We introduce the Attention-Indexed Model (AIM), a theoretical framework for analyzing learning in deep attention layers. Inspired by multi-index models, AIM captures how token-level outputs emerge from layered bilinear interactions over high-dimensional embeddings. Unlike prior tractable attention models, AIM allows full-rank key and query matrices, aligning more closely with practical transformers. Using tools from statistical mechanics and random matrix theory, we derive closed-form predictions for Bayes-optimal generalization error and identify sharp phase transitions as a function of sample complexity, model width, and sequence length. We propose a matching Approximate Message Passing algorithm and show that gradient descent can reach optimal performance. AIM offers a solvable playground for understanding learning in modern attention architectures.



Paperid:640
Authors:Dong Liang, Jinyuan Jia, Yuhao LIU, Rynson Lau
Abstract:
While existing image‑guided composition methods may help insert a foreground object onto a user-specified region of a background image, achieving natural blending inside the region with the rest of the image unchanged, we observe that these existing methods often struggle in synthesizing seamless interaction-aware compositions when the task involves human-object interactions.In this paper, we first propose HOComp, a novel approach for compositing a foreground object onto a human-centric background image, while ensuring harmonious interactions between the foreground object and the background person and their consistent appearances. Our approach includes two key designs: (1) MLLMs-driven Region-based Pose Guidance (MRPG), which utilizes MLLMs to identify the interaction region as well as the interaction type (e.g., holding and lefting) to provide coarse-to-fine constraints to the generated pose for the interaction while incorporating human pose landmarks to track action variations and enforcing fine-grained pose constraints; and (2) Detail-Consistent Appearance Preservation (DCAP), which unifies a shape-aware attention modulation mechanism, a multi-view appearance loss, and a background consistency loss to ensure consistent shapes/textures of the foreground and faithful reproduction of the background human. We then propose the first dataset, named Interaction-aware Human-Object Composition (IHOC), for the task.Experimental results on our dataset show that HOComp effectively generates harmonious human-object interactions with consistent appearances, and outperforms relevant methods qualitatively and quantitatively.



Paperid:641
Authors:Nikita Araslanov, Anna Sonnweber, Daniel Cremers
Abstract:
Dense and versatile image representations underpin the success of virtually any computer vision application. However, state-of-the-art networks, such as transformers, produce low-resolution feature grids, which are suboptimal for dense prediction tasks. Addressing this limitation, our work presents FlowFeat, a high-resolution and multi-task feature representation. The key ingredient behind FlowFeat is a novel distillation technique embedding a distribution of plausible apparent motion, ormotion profiles. Leveraging optical flow networks and diverse video data, we develop an effective self-supervised training framework, which statistically approximates the apparent motion. With a remarkable level of spatial detail, FlowFeat encodes a compelling degree of geometric and semantic cues while exhibiting high temporal consistency. Empirically, FlowFeat significantly enhances the representation power of five state-of-the-art encoders and alternative upsampling strategies across three dense tasks: video object segmentation, monocular depth estimation and semantic segmentation.Training FlowFeat is robust to inaccurate flow estimation and remains highly effective also with unsupervised flow networks.Our work is a confident step forward towards reliable dense and versatile image representations.



Paperid:642
Authors:Xiangkun Wu, Ting Li, Gholamali Aminian, Armin Behnamnia, Hamid Rabiee, Chengchun Shi
Abstract:
This paper studies how to integrate historical control data with experimental data to enhance A/B testing, while addressing the distributional shift between historical and experimental datasets. We propose a pessimistic data integration method that combines two causal effect estimators constructed based on experimental and historical datasets. Our main idea is to conceptualize the weight function for this combination as a policy so that existing pessimistic policy learning algorithms are applicable to learn the optimal weight that minimizes the resulting weighted estimator's mean squared error. Additionally, we conduct comprehensive theoretical and empirical analyses to compare our method against various baseline estimators across five scenarios. Both our theoretical and numerical findings demonstrate that the proposed estimator achieves near-optimal performance across all scenarios.



Paperid:643
Authors:Jianfei Jiang, Qiankun Liu, Hongyuan Liu, Haochen Yu, Liyong Wang, Jiansheng Chen, Huimin Ma
Abstract:
Robust feature representations are essential for learning-based Multi-View Stereo (MVS), which relies on accurate feature matching. Recent MVS methods leverage Transformers to capture long-range dependencies based on local features extracted by conventional feature pyramid networks. However, the quadratic complexity of Transformer-based MVS methods poses challenges to balance performance and efficiency. Motivated by the global modeling capability and linear complexity of the Mamba architecture, we propose MVSMamba, the first Mamba-based MVS network. MVSMamba enables efficient global feature aggregation with minimal computational overhead. To fully exploit Mamba's potential in MVS, we propose a Dynamic Mamba module (DM-module) based on a novel reference-centered dynamic scanning strategy, which enables: (1) Efficient intra- and inter-view feature interaction from the reference to source views, (2) Omnidirectional multi-view feature representations, and (3) Multi-scale global feature aggregation. Extensive experimental results demonstrate MVSMamba outperforms state-of-the-art MVS methods on the DTU dataset and the Tanks-and-Temples benchmark with both superior performance and efficiency.



Paperid:644
Authors:Gabrielle Berrada, Jannik Kossen, Muhammed Razzak, Freddie Bickford Smith, Yarin Gal, Thomas Rainforth
Abstract:
Active testing enables label-efficient evaluation of models through careful data acquisition. However, its significant computational costs have previously undermined its use for large models. We show how it can be successfully scaled up to the evaluation of large language models (LLMs). In particular we show that the surrogate model used to guide data acquisition can be constructed cheaply using in-context learning, does not require updating within an active-testing loop, and can be smaller than the target model. We even find we can make good data-acquisition decisions without computing predictions with the target model and further introduce a single-run error estimator to asses how well active testing is working on the fly. We find that our approach is able to more effectively evaluate LLM performance with less data than current standard practices.



Paperid:645
Authors:Sicheng Zhu, Brandon Amos, Yuandong Tian, Chuan Guo, Ivan Evtimov
Abstract:
Many jailbreak attacks on large language models (LLMs) rely on a common objective: making the model respond with the prefix ``Sure, here is (harmful request)''. While straightforward, this objective has two limitations: limited control over model behaviors, yielding incomplete or unrealistic jailbroken responses, and a rigid format that hinders optimization. We introduce AdvPrefix, a plug-and-play prefix-forcing objective that selects one or more model-dependent prefixes by combining two criteria: high prefilling attack success rates and low negative log-likelihood. AdvPrefix integrates seamlessly into existing jailbreak attacks to mitigate the previous limitations for free. For example, replacing GCG's default prefixes on Llama-3 improves nuanced attack success rates from 14\% to 80\%, revealing that current safety alignment fails to generalize to new prefixes. Code and selected prefixes are released.



Paperid:646
Authors:Soham Deshmukh, Satvik Dixit, Rita Singh, Bhiksha Raj
Abstract:
Multimodal Audio-Language Models (ALMs) can understand and reason over both audio and text. Typically, reasoning performance correlates with model size, with the best results achieved by models exceeding 8 billion parameters. However, no prior work has explored enabling small audio-language models to perform reasoning tasks, despite the potential applications for edge devices. To address this gap, we introduce Mellow, a small Audio-Language Model specifically designed for reasoning. Mellow achieves state-of-the-art performance among existing small audio-language models and surpasses several larger models in reasoning capabilities. For instance, Mellow scores 52.11 on MMAU, comparable to SoTA Qwen2 Audio (which scores 52.5) while using 50 times fewer parameters and being trained on 60 times less data (audio hrs). To train Mellow, we introduce ReasonAQA, a dataset designed to enhance audio-grounded reasoning in models. It consists of a mixture of existing datasets (30\% of the data) and synthetically generated data (70\%). The synthetic dataset is derived from audio captioning datasets, where Large Language Models (LLMs) generate detailed and multiple-choice questions focusing on audio events, objects, acoustic scenes, signal properties, semantics, and listener emotions. To evaluate Mellow’s reasoning ability, we benchmark it on a diverse set of tasks, assessing on both in-distribution and out-of-distribution data, including audio understanding, deductive reasoning, and comparative reasoning. Finally, we conduct extensive ablation studies to explore the impact of projection layer choices, synthetic data generation methods, and language model pretraining on reasoning performance. Our training dataset, findings, and baseline pave the way for developing small ALMs capable of reasoning.



Paperid:647
Authors:Jean Feng, Avni Kothari, Lucas Zier, Chandan Singh, Yan Shuo Tan
Abstract:
Concept Bottleneck Models (CBMs) have been proposed as a compromise between white-box and black-box models, aiming to achieve interpretability without sacrificing accuracy.The standard training procedure for CBMs is to predefine a candidate set of human-interpretable concepts, extract their values from the training data, and identify a sparse subset as inputs to a transparent prediction model.However, such approaches are often hampered by the tradeoff between exploring a sufficiently large set of concepts versus controlling the cost of obtaining concept extractions, resulting in a large interpretability-accuracy tradeoff.This work investigates a novel approach that sidesteps these challenges: BC-LLM iteratively searches over a potentially infinite set of concepts within a Bayesian framework, in which Large Language Models (LLMs) serve as both a concept extraction mechanism and prior.Even though LLMs can be miscalibrated and hallucinate, we prove that BC-LLM can provide rigorous statistical inference and uncertainty quantification.Across image, text, and tabular datasets, BC-LLM outperforms interpretable baselines and even black-box models in certain settings, converges more rapidly towards relevant concepts, and is more robust to out-of-distribution samples.



Paperid:648
Authors:Pritish Chakraborty, Indradyumna Roy, Soumen Chakrabarti, Abir De
Abstract:
Retrieving graphs from a large corpus, that contain a subgraph isomorphic to a given query graph, is a core operation in many real-world applications. While recent multi-vector graph representations and scores based on set alignment and containment can provide accurate subgraph isomorphism tests, their use in retrieval remains limited by their need to score corpus graphs exhaustively.We introduce CoRGII (COntextual Representation of Graphs for Inverted Indexing), a graph indexing framework in which, starting with a contextual dense graph representation, a differentiable discretization module computes sparse binary codes over a learned latent vocabulary. This text document-like representation allows us to leverage classic, highly optimized inverted indexes, while supporting soft (vector) set containment scores. Improving on this paradigm further, we replace the classical impact score of a `word' on a graph (such as defined by TFIDF or BM25) with a data-driven, trainable impact score.Crucially, CoRGII is trained end-to-end using only binary relevance labels, without fine-grained supervision of query-to-document set alignments. Extensive experiments show that CoRGII provides better trade-offs between efficiency and accuracy, compared to several baselines.



Paperid:649
Authors:Weixiang Zhang, Boxi Li, Shuzhao Xie, Chengwei Ren, Yuan Xue, Zhi Wang
Abstract:
In this paper, we examine the impact and significance of bias terms in Implicit Neural Representations (INRs). While bias terms are known to enhance nonlinear capacity by shifting activations in typical neural networks, we discover their functionality differs markedly in neural representation networks.Our analysis reveals that INR performance neither scales with increased number of bias terms nor shows substantial improvement through bias term gradient propagation. We demonstrate that bias terms in INRs primarily serve to eliminate \textit{spatial aliasing} caused by symmetry from both coordinates and activation functions, with input-layer bias terms yielding the most significant benefits.These findings challenge the conventional practice of implementing full-bias INR architecture. We propose using freezing bias terms exclusively in input layers, which consistently outperforms fully biased networks in signal fitting tasks.Furthermore, we introduce Feature-Biased INRs~(Feat-Bias), which initialize input-layer bias with high-level features extracted from pre-trained models. This feature-biasing approach effectively addresses the limited performance in INR post-processing tasks due to neural parameter uninterpretability, achieving superior accuracy while reducing parameter count and improving reconstruction quality.We will release the code.



Paperid:650
Authors:Andy Yang, Michaël Cadilhac, David Chiang
Abstract:
Abstract:It has been observed that transformers with greater depth (that is, more layers) have more capabilities, but can we explain precisely which capabilities are gained with greater depth? We answer this question with a theoretical proof followed by an empirical study. First, we consider transformers that round to fixed precision except inside attention. We show that this subclass of transformers is expressively equivalent to the programming language $\textsf{C}$-$\textsf{RASP}$, and this equivalence preserves depth. Second, we prove that deeper $\textsf{C}$-$\textsf{RASP}$ programs are more expressive than shallower $\textsf{C}$-$\textsf{RASP}$ programs, implying that deeper transformers are more expressive than shallower transformers (within the subclass mentioned above). Finally, we provide empirical evidence that our theory predicts the depth required for transformers without positional encodings to length-generalize on a family of sequential dependency tasks.



Paperid:651
Authors:Edoardo Pona, Milad Kazemi, Yali Du, David Watson, Nicola Paoletti
Abstract:
Counterfactual inference is a powerful tool for analysing and evaluating autonomous agents, but its application to language model (LM) agents remains challenging. Existing work on counterfactuals in LMs has primarily focused on token-level counterfactuals, which are often inadequate for LM agents due to their open-ended action spaces. Unlike traditional agents with fixed, clearly defined action spaces, the actions of LM agents are often implicit in the strings they output, making their action spaces difficult to define and interpret. Furthermore, the meanings of individual tokens can shift depending on the context, adding complexity to token-level reasoning and sometimes leading to biased or meaningless counterfactuals.We introduce \emph{Abstract Counterfactuals}, a framework that emphasises high-level characteristics of actions and interactions within an environment, enabling counterfactual reasoning tailored to user-relevant features. Our experiments demonstrate that the approach produces consistent and meaningful counterfactuals while minimising the undesired side effects of token-level methods.We conduct experiments on text-based games and counterfactual text generation, while considering both token-level and latent-space interventions.



Paperid:652
Authors:Xinyin Ma, Runpeng Yu, Gongfan Fang, Xinchao Wang
Abstract:
Abstract:Diffusion Language Models (DLMs) have been seen as a promising competitor for autoregressive language models (ARs). However, diffusion language models have long been constrained by slow inference. A core challenge is that their non‑autoregressive architecture and bidirectional attention preclude the key–value cache that accelerates decoding. We address this bottleneck by proposing a KV-cache-like mechanism, **d**elayed **KV-Cache**, for the denoising process of DLMs. Our approach is motivated by the observation that different tokens have distinct representation dynamics throughout the diffusion process. Accordingly, we propose a delayed and conditioned caching strategy for key and value states. We design two complementary variants to cache key and value step‑by‑step: (1) dKV-Cache-Decode, which provides almost lossless acceleration, and even improves performance on long sequences, suggesting that existing DLMs may under‑utilise contextual information during inference. (2) dKV-Cache‑Greedy, which has aggressive caching with reduced lifespan, achieving higher speed-ups with quadratic time complexity at the cost of some performance degradation. dKV-Cache, in final, achieves from 2-10$\times$ speedup in inference, largely narrowing the gap between ARs and DLMs. We evaluate our dKV-Cache on several benchmarks, delivering acceleration across general language understanding, mathematical, and code‑generation benchmarks. Experiments demonstrate that cache can also be used in DLMs, even in a training-free manner from current DLMs.



Paperid:653
Authors:Haixu Wu, Minghao Guo, Yuezhou Ma, Yuanxu Sun, Jianmin Wang, Wojciech Matusik, Mingsheng Long
Abstract:
Abstract:Attention mechanism has emerged as a foundation module of modern deep learning models and has also empowered many milestones in various domains. Moreover, FlashAttention with IO-aware speedup resolves the efficiency issue of standard attention, further promoting its practicality. Beyond canonical attention, attention with bias also widely exists, such as relative position bias in vision and language models and pair representation bias in AlphaFold. In these works, prior knowledge is introduced as an additive bias term of attention weights to guide the learning process, which has been proven essential for model performance. Surprisingly, despite the common usage of attention with bias, its targeted efficiency optimization is still absent, which seriously hinders its wide applications in complex tasks. Diving into the computation of FlashAttention, we prove that its optimal efficiency is determined by the rank of the attention weight matrix. Inspired by this theoretical result, this paper presents FlashBias based on the low-rank compressed sensing theory, which can provide fast-exact computation for many widely used attention biases and a fast-accurate approximation for biases in general formalization. FlashBias can fully take advantage of the extremely optimized matrix multiplication operation in modern GPUs, achieving 1.5$\times$ speedup for AlphaFold, and over 2$\times$ speedup for attention with bias in vision and language models without loss of accuracy.



Paperid:654
Authors:Michael Gastpar, Ido Nachum, Jonathan Shafer, Thomas Weinberger
Abstract:
We study which machine learning algorithms have tight generalization bounds with respect to a given collection of population distributions. Our results build on and extend the recent work of Gastpar et al. (2023). First, we present conditions that preclude the existence of tight generalization bounds. Specifically, we show that algorithms that have certain inductive biases that cause them to be unstable do not admit tight generalization bounds. Next, we show that algorithms that are sufficiently stable do have tight generalization bounds. We conclude with a simple characterization that relates the existence of tight generalization bounds to the conditional variance of the algorithm's loss.



Paperid:655
Authors:Yaroslav Mukhin
Abstract:
The influence function (also known as the first variation and Fisher-Rao gradient)of a statistical functional is the Riesz representer of its derivative operator.It is the key analytic object in both the theory and implementation ofestimators in semiparametric statistical and machine learning models,e.g., one-step estimator, targeted learning, debiased machine learning.It is also essential for inference about their statistical, robustness andinterpretability properties,e.g., for finding confidence intervals, partial identification andmisspecification bounds, adversarial perturbations, influential data points.However, the analytic derivation of the influence function is often anobstruction to the broader adoption of these methods by practitioners.Toward automating this task,we derive a regularized representation of the influence function usingspectral theory of positive semidefinite kernels.Based on this representation we construct an estimator that:(i) is a nonparametric functional RKHS estimator;(ii) admits theoretical guarantees in function norms relevant for downstream tasks;(iii) can be computed via automatic differentiation or finite differences,without requiring analytic derivation by the user.



Paperid:656
Authors:Jose Blanchet, Yassine Hamoudi, Mario Szegedy, Guanyang Wang
Abstract:
The mean of a random variable can be understood as alinearfunctional on the space of probability distributions. Quantum computing is known to provide a quadratic speedup over classical Monte Carlo methods for mean estimation. In this paper, we investigate whether a similar quadratic speedup is achievable for estimatingnon-linearfunctionals of probability distributions. We propose aquantum-inside-quantumalgorithm that achieves this speedup for the broad class of nonlinear estimation problems known as nested expectations. Our algorithm improves upon the direct application of the quantum-accelerated multilevel Monte Carlo algorithm introduced by An et al.. The existing lower bound indicates that our algorithm is optimal up polylogarithmic factors. A key innovation of our approach is a new sequence of multilevel Monte Carlo approximations specifically designed for quantum computing, which is central to the algorithm's improved performance.



Paperid:657
Authors:Saba Ahmadi, Rabiul Awal, Ankur Sikarwar, Amirhossein Kazemnejad, Ge Ya Luo, Juan Rodriguez, Sai Rajeswar Mudumba, Siva Reddy, Chris Pal, Benno Krojer, Aishwarya Agrawal
Abstract:
While image generation techniques are now capable of producing high quality images that respect prompts which span multiple sentences, the task of text-guided image editing remains a challenge. Even edit requests that consist of only a few words often fail to be executed correctly. We explore three strategies to enhance performance on a wide range of image editing tasks: supervised fine-tuning (SFT), reinforcement learning (RL), and Chain-of-Thought (CoT) reasoning. In order to study all these components in one consistent framework we adopt an autoregressive multimodal model that processes textual and visual tokens in a unified manner.We find RL combined with a large multi-modal LLM verifier to be the most effective of these strategies.As a result, we release EARL:Editing withAutoregression andRL, a strong RL-based image editing model that performs competitively on a diverse range of edits compared to strong baselines with much more training data. Thus, EARL pushes the frontier of autoregressive multimodal models on image editing.



Paperid:658
Authors:Xuchen Gong, Tian Li
Abstract:
Abstract:One of the major bottlenecks for deploying popular first-order differentially private (DP) machine learning algorithms (e.g., DP-SGD) lies in their high computation and memory cost, despite the existence of optimized implementations. Zeroth-order methods have promise in mitigating the overhead, as they leverage function evaluations to approximate the gradients, hence significantly easier to privatize. While recent works have explored zeroth-order approaches in both private and non-private settings, they still suffer from relatively low utilities compared with DP-SGD and limited application domains. In this work, we propose to leverage public information to guide and improve gradient approximation of private zeroth-order algorithms. We explore a suite of \underline{p}ublic data \underline{a}ssisted \underline{z}eroth-\underline{o}rder optimizers (PAZO) with minimal overhead. We provide theoretical analyses of the PAZO framework under an assumption of the similarity between public and private data. Empirically, we demonstrate that PAZO achieves stronger privacy/utility tradeoffs across vision and text tasks in both pre-training and fine-tuning regimes, outperforming the best first-order baselines (with public gradients) especially in highly private regimes, while offering up to $16\times$ runtime speedup.



Paperid:659
Authors:Ryo Fujii, Hideo Saito, Ryo Hachiuma
Abstract:
Predicting accurate future trajectories of pedestrians is essential for autonomous systems but remains a challenging task due to the need for adaptability in different environments and domains. A common approach involves collecting scenario-specific data and performing fine-tuning via backpropagation. However, this process is often impractical on edge devices due to constrained computational resources. To address this challenge, we introduce TrajICL an In-Context Learning (ICL) framework for pedestrian trajectory prediction that enables rapid adaptation without fine-tuning on the scenario-specific data. We propose a spatio-temporal similarity-based example selection (STES) method that selects relevant examples from previously observed trajectories within the same scene by identifying similar motion patterns at corresponding locations. To further refine this selection, we introduce prediction-guided example selection (PG-ES), which selects examples based on both the past trajectory and the predicted future trajectory, rather than relying solely on the past trajectory. This approach allows the model to account for long-term dynamics when selecting examples. Finally, instead of relying on small real-world datasets with limited scenario diversity, we train our model on a large-scale synthetic dataset to enhance its prediction ability by leveraging in-context examples. Extensive experiments demonstrate that TrajICL achieves remarkable adaptation across both in-domain and cross-domain scenarios, outperforming even fine-tuned approaches across multiple public benchmarks. We will release the code and pretrained model upon acceptance.



Paperid:660
Authors:Ivica Nikolic, Teodora Baluta, Prateek Saxena
Abstract:
Large language models are increasingly customized through fine-tuning and other adaptations, creating challenges in enforcing licensing terms and managing downstream impacts such as protecting intellectual property or identifying vulnerabilities. We address this challenge by developing a framework for testing model provenance. Our approach is based on the key observation that real-world model derivations preserve significant similarities in model outputs that can be detected through statistical analysis. Using only black-box access to models, we employ multiple hypothesis testing to compare model similarities against a baseline established by unrelated models.On two comprehensive real-world benchmarks spanning models from 30M to 4B parameters and comprising over 600 models, our tester achieves 90-95% precision and 80-90% recall in identifying derived models.These results demonstrate the viability of systematic provenance verification in production environments even when only API access is available.



Paperid:661
Authors:Amitai Yacobi, Nir Ben-Ari, Ronen Talmon, Uri Shaham
Abstract:
Learning shared representations is a primary area of multimodal representation learning. The current approaches to achieve a shared embedding space rely heavily on paired samples from each modality, which are significantly harder to obtain than unpaired ones. In this work, we demonstrate that shared representations can be learned almost exclusively from unpaired data. Our arguments are grounded in the spectral embeddings of the random walk matrices constructed independently from each unimodal representation. Empirical results in computer vision and natural language processing domains support its potential, revealing the effectiveness of unpaired data in capturing meaningful cross-modal relations, demonstrating high capabilities in retrieval tasks, generation, arithmetics, zero-shot, and cross-domain classification. This work, to the best of our knowledge, is the first to demonstrate these capabilities almost exclusively from unpaired samples, giving rise to a cross-modal embedding that could be viewed as universal, i.e., independent of the specific modalities of the data.



Paperid:662
Authors:Qianshu Cai, Chao Wu, Yonggang Zhang, Jun Yu, Xinmei Tian
Abstract:
The effective detection of generated images is crucial to mitigate potential risks associated with their misuse. Despite significant progress, a fundamental challenge remains: ensuring the generalizability of detectors. To address this, we propose a novel perspective on understanding and improving generated image detection, inspired by the human cognitive process: Humans identify an image as unnatural based on specific patterns because these patterns lie outside the space spanned by those of natural images. This is intrinsically related to out-of-distribution (OOD) detection, which identifies samples whose semantic patterns (i.e., labels) lie outside the semantic pattern space of in-distribution (ID) samples. By treating patterns of generated images as OOD samples, we demonstrate that models trained merely over natural images bring guaranteed generalization ability under mild assumptions. This transforms the generalization challenge of generated image detection into the problem of fitting natural image patterns. Based on this insight, we propose a generalizable detection method through the lens of ID energy. Theoretical results capture the generalization risk of the proposed method. Experimental results across multiple benchmarks demonstrate the effectiveness of our approach.



Paperid:663
Authors:Xinyang Liu, Junming Hou, Chenxu Wu, Xiaofeng Cong, zihao chen, Shangqi Deng, Junling Li, Liang-Jian Deng, Bo Liu
Abstract:
Abstract:Over the past decades, pansharpening has contributed greatly to numerous remote sensing applications, with methods evolving from theoretically grounded models to deep learning approaches and their hybrids. Though promising, existing methods rarely address pansharpening through the lens of underlying physical imaging processes. In this work, we revisit the spectral imaging mechanism and propose a novel physics‐informed neural operator framework for pansharpening, termed PINO, which faithfully models the end‐to‐end electro‐optical sensor process. Specifically, PINO operates as: (1) First, a spatial-spectral encoder pair is introduced to aggregate multi-granularity high-resolution panchromatic (PAN) and low-resolution multispectral (LRMS) features.(2) Subsequently, an iterative neural integral process utilizes these fused spatial-spectral characteristics to learn a continuous radiance field $L_i(x, y, \lambda)$ over spatial coordinates and wavelength, effectively emulating band-wise spectral integration. (3) Finally, the learned radiance field is modulated by the sensor’s spectral responsivity $R_b(\lambda)$ to produce physically consistent spatial–spectral fusion products. This physics-grounded fusion paradigm offers a principled solution for reconstructing high-resolution multispectral and hyperspectral images in accordance with sensor imaging physics, effectively harnessing the unique advantages of spectral data to better uncover real-world characteristics. Experiments on multiple benchmark datasets show that our method surpasses state-of-the-art fusion algorithms, achieving reduced spectral aberrations and finer spatial textures. Furthermore, extension to hyperspectral (HS) data demonstrates its generalizability and universality. The code will be available upon potential acceptance.



Paperid:664
Authors:haonan he, Peng Ye, Yuchen Ren, yuan yuan, LuyangZhou, ShucunJu, lei chen
Abstract:
Low-Rank Adaptation (LoRA) is a crucial method for efficiently fine-tuning large language models (LLMs), with its effectiveness influenced by two key factors: rank selection and weight initialization. While numerous LoRA variants have been proposed to improve performance by addressing one of these aspects, they often compromise usability or computational efficiency. In this paper, we analyze and identify the core limitations of existing approaches and propose a novel framework—GoRA (Gradient-driven Adaptive Low Rank Adaptation)—that simultaneously adapts both the rank and initialization strategy within a unified framework. GoRA leverages gradient information during training to dynamically assign optimal ranks and initialize low-rank adapter weights in an adaptive manner. To our knowledge, GoRA is the first method that not only addresses the limitations of prior approaches—which often focus on either rank selection or initialization in isolation—but also unifies both aspects within a single framework, enabling more effective and efficient adaptation. Extensive experiments across various architectures and modalities show that GoRA consistently outperforms existing LoRA-based methods while preserving the efficiency of vanilla LoRA. For example, when fine-tuning Llama3.1-8B-Base for mathematical reasoning, GoRA achieves a 5.13-point improvement over standard LoRA and even outperforms full fine-tuning by 2.05 points under high-rank settings



Paperid:665
Authors:Zining Fan, He Zhu
Abstract:
Designing reinforcement learning agents to satisfy complex temporal objectives expressed in Linear Temporal Logic (LTL), presents significant challenges, particularly in ensuring sample efficiency and task alignment over infinite horizons. Recent works have shown that by leveraging the corresponding Limit Deterministic Büchi Automaton (LDBA) representation, LTL formulas can be translated into variable discounting schemes over LDBA-accepting states to maximize a lower bound on the probability of formula satisfaction. However, the resulting reward signals are inherently sparse, making exploration of LDBA-accepting states increasingly difficult as task horizons lengthen to infinity. In this work, we address these challenges by leveraging finite-length demonstrations to overcome the exploration bottleneck for LTL objectives over infinite horizons. We segment demonstrations and agent exploratory trajectories at LDBA-accepting states and iteratively guide the agent within each segment to learn to reach these accepting states. By incentivizing the agent to visit LDBA-accepting states from arbitrary states, our approach increases the probability of LTL formula satisfaction without the need for extensive or lengthy demonstrations. We demonstrate the applicability of our method across a variety of high-dimensional continuous control domains. It achieves faster convergence and consistently outperforms baseline approaches.



Paperid:666
Authors:Andrew Jacobsen, Alessandro Rudi, Francesco Orabona, Nicolò Cesa-Bianchi
Abstract:
Abstract:We study dynamic regret in online convex optimization, where the objective is to achieve low cumulative loss relative to an arbitrary benchmark sequence. By observing that competing with an arbitrary sequence of comparators $u_{1},\ldots,u_{T}$ in $\mathcal{W}\subseteq\mathbb{R}^{d}$ is equivalent to competing with a *fixed* comparator *function* $u:[1,T]\to \mathcal{W}$, we frame dynamic regret minimization as a *static* regret problem in a *function space*. By carefully constructing a suitable function space in the form of a Reproducing Kernel Hilbert Space (RKHS), our reduction enables us to recover the optimal $R_{T}(u_{1},\ldots,u_{T}) = \mathcal{O}(\sqrt{\sum_{t}\\|u_{t}-u_{t-1}\\|T})$ dynamic regret guarantee in the setting of linear losses, and yields new scale-free and directionally-adaptive dynamic regret guarantees. Moreover, unlike prior dynamic-to-static reductions---which are valid only for linear losses---our reduction holds for *any* sequence of losses, allowing us to recover $\mathcal{O}\big(\\|u\\|^2+d_{\mathrm{eff}}(\lambda)\ln T\big)$ bounds in exp-concave and improper linear regression settings, where $d_{\mathrm{eff}}(\lambda)$ is a measure of complexity of the RKHS. Despite working in an infinite-dimensional space, the resulting reduction leads to algorithms that are computable in practice, due to the reproducing property of RKHSs.



Paperid:667
Authors:Zhenrui Yue, Bowen Jin, Huimin Zeng, Honglei Zhuang, Zhen Qin, Jinsung Yoon, Lanyu Shang, Jiawei Han, Dong Wang
Abstract:
Recent advances in large language models (LLMs) have introduced latent reasoning as a promising alternative to autoregressive reasoning. By performing internal computation with hidden states from previous steps, latent reasoning benefit from more informative features rather than sampling a discrete chain-of-thought (CoT) path. Yet latent reasoning approaches are often incompatible with LLMs, as their continuous paradigm conflicts with the discrete nature of autoregressive generation. Moreover, these methods rely on CoT traces for training and thus fail to exploit the inherent reasoning patterns of LLMs. In this work, we explore latent reasoning by leveraging the intrinsic capabilities of LLMs via reinforcement learning (RL). To this end, we introduce hybrid reasoning policy optimization (HRPO), an RL-based hybrid latent reasoning approach that (1) integrates prior hidden states into sampled tokens with a learnable gating mechanism, and (2) initializes training with predominantly token embeddings while progressively incorporating more hidden features. This design maintains LLMs' inherent generative capabilities and incentivizes hybrid reasoning using both discrete and continuous representations. In addition, HRPO introduces stochasticity into latent reasoning via token sampling, thereby enabling RL-based optimization without requiring CoT trajectories. Extensive evaluations across diverse benchmarks show that HRPO outperforms prior methods in both knowledge- and reasoning-intensive tasks. Furthermore, HRPO-trained LLMs remain interpretable and exhibit intriguing behaviors like cross-lingual patterns and shorter completion lengths, highlighting the potential of our RL-based approach and offer insights for future work in latent reasoning.



Paperid:668
Authors:Min Woo Park, Andy Arditi, Elias Bareinboim, Sanghack Lee
Abstract:
In a decision-making process, intelligent agents with causal knowledge can optimize action spaces to avoid unnecessary exploration. Astructural causal banditframework provides guidance on how to prune actions that are unable to maximize reward by leveraging prior knowledge of the underlying causal structure among actions. A key assumption of this framework is that the agent has access to a fully-specified causal diagram representing the target system. However, this assumption is often violated, making it difficult to apply in practical contexts. In this paper, we extend the structural causal bandits to scenarios where the agent leverages a Markov equivalent class. In such cases, the causal structure is provided to the agent in the form of apartial ancestral graph(PAG). We propose a generalized framework for identifying potentially optimal actions within this graph structure, thereby broadening the applicability of structural causal bandits.



Paperid:669
Authors:Qilong Guo, Tianjing Zhang, Zhiyuan Ma, Hui Ji
Abstract:
Blind-spot networks offer a powerful paradigm for zero-shot image denoising by training models to predict masked target pixels from their neighbors. However, they struggle with real-world noise exhibiting strong local correlations, where efforts to suppress noise correlation often weaken essential pixel dependencies, adversely affecting denoising performance. This paper presents a theoretical analysis quantifying the impact of replacing masked pixels with observations exhibiting weaker noise correlation but potentially reduced similarity, revealing a trade-off that impacts the statistical risk of the estimation. Guided by this insight, we propose a computational scheme that replaces masked pixels with distant ones of similar appearance and lower noise correlation. This strategy improves the prediction by balancing noise suppression and structural consistency. Experiments confirm the effectiveness of our method, outperforming existing zero-shot BSNs.



Paperid:670
Authors:Hanyang Yuan, Ning Tang, Tongya Zheng, Jiarong Xu, Xintong Hu, Renhong Huang, Shunyu Liu, Jiacong Hu, Jiawei Chen, Mingli Song
Abstract:
Diversified Recommendation has attracted increasing attention from both researchers and practitioners, which can effectively address the homogeneity of recommended items.Existing approaches predominantly aim to infer the diversity of user preferences from observed user feedback.Nonetheless, due to inherent data biases, the observed data may not fully reflect user interests, where underexplored preferences can be overwhelmed or remain unmanifested. Failing to capture these preferences can lead to suboptimal diversity in recommendations. To fill this gap, this work aims to study diversified recommendation from a data-bias perspective.Inspired by the outstanding performance of large language models (LLMs) in zero-shot inference leveraging world knowledge, we propose a novel approach that utilizes LLMs' expertise to uncover underexplored user preferences from observed behavior, ultimately providing diverse and relevant recommendations.To achieve this, we first introduce the Tree of Preference (ToP), an innovative structure constructed to model user preferences from coarse to fine. ToP enables LLMs to systematically reason over the user's rationale behind their behavior, thereby uncovering their underexplored preferences.To guide diversified recommendations using uncovered preferences, we adopt a data-centric approach, identifying candidate items that match user preferences and generating synthetic interactions that reflect underexplored preferences. These interactions are integrated to train a general recommender for diversification.Moreover, we scale up overall efficiency by dynamically selecting influential users during optimization.Extensive evaluations of both diversity and relevance show that our approach outperforms existing methods in most cases and achieves near-optimal performance in others, with reasonable inference latency.Our code is publicly available at https://anonymous.4open.science/r/TPRec-7047/.



Paperid:671
Authors:Jiankang Wang, Jianjun Xu, Xiaorui Wang, Yuxin Wang, Mengting Xing, Shancheng Fang, Hongtao Xie
Abstract:
Large-scale, high-quality data is essential for advancing the reasoning capabilities of large language models (LLMs). As publicly available Internet data becomes increasingly scarce, synthetic data has emerged as a crucial research direction. However, existing data synthesis methods often suffer from limited scalability, insufficient sample diversity, and a tendency to overfit to seed data, which constrains their practical utility. In this paper, we present \textit{\textbf{GRIP}}, a \textbf{G}raph-based \textbf{R}easoning \textbf{I}nstruction \textbf{P}roducer that efficiently synthesizes high-quality and diverse reasoning instructions. \textit{GRIP} constructs a knowledge graph by extracting high-level concepts from seed data, and uniquely leverages both explicit and implicit relationships within the graph to drive large-scale and diverse instruction data synthesis, while employing open-source multi-model supervision to ensure data quality. We apply \textit{GRIP} to the critical and challenging domain of mathematical reasoning. Starting from a seed set of 7.5K math reasoning samples, we construct \textbf{GRIP-MATH}, a dataset containing 2.1 million synthesized question-answer pairs. Compared to similar synthetic data methods, \textit{GRIP} achieves greater scalability and diversity while also significantly reducing costs. On mathematical reasoning benchmarks, models trained with GRIP-MATH demonstrate substantial improvements over their base models and significantly outperform previous data synthesis methods.



Paperid:672
Authors:Jiexuan Zhang, Yiheng Du, Qian Wang, Weiqi Li, Yu Gu, Jian Zhang
Abstract:
Abstract:Diffusion-based generative models struggle to maintain high style consistency across generated images via text description. Although several style-aligned image generation methods have been proposed to address this issue, they exhibit suboptimal performance and are primarily built upon the U-Net architecture, limiting their compatibility with MM-DiT diffusion models like Flux that has emerged as a predominant model in the field of image generation. To address these limitations, we propose $\textit{\textbf{AlignedGen}}$, a novel training-free style-aligned image generation method for Flux to significantly enhance style consistency across generated images. Specifically, AlignedGen incorporates two key components to achieve this: Shifted Position Embedding (ShiftPE) and Selective Shared Attention (SSA) layer. ShiftPE alleviates the text controllability degradation observed in prior methods when applied to Flux through its non-overlapping position indices design, while SSA further enhances style consistency across images. In addition, our method can be seamlessly integrated with various controllable generation technologies (e.g., subject-driven generation, depth control), demonstrating broad applicability across diverse scenarios. Extensive experimental results validate that our method effectively enhances style consistency across generated images while maintaining favorable text controllability.



Paperid:673
Authors:Alankrita Bhatt, Victoria Kostina
Abstract:
Prediction with expert advice serves as a fundamental model in online learning and sequential decision-making. However, in many real-world settings, this classical model proves insufficient as the feedback available to the decision-maker is often subject to noise, errors, or communication constraints. This paper provides fundamental limits on performance, quantified by the regret, in the case when the feedback is corrupted by an additive noise. Our general analysis achieves sharp regret bounds for canonical examples of such additive noise as the Gaussian distribution, the uniform distribution, and a general noise with a log-concave density. This analysis demonstrates how different noise characteristics affect regret bounds and identifies how the regret fundamentally scales as a function of the properties of the noise distribution.



Paperid:674
Authors:Parsa Moradi, Hanzaleh Nodehi, Mohammad Maddah-Ali
Abstract:
Abstract:In this paper, we investigate the adversarial robustness of regression, a fundamental problem in machine learning, under the setting where an adversary can arbitrarily corrupt a subset of the input data. While the robustness of parametric regression has been extensively studied, its nonparametric counterpart remains largely unexplored. We characterize the adversarial robustness in nonparametric regression, assuming the regression function belongs to the second-order Sobolev space (i.e., it is square integrable up to its second derivative). The contribution of this paper is two-fold: (i) we establish a minimax lower bound on the estimation error, revealing a fundamental limit that no estimator can overcome, and (ii) we show that, perhaps surprisingly, the classical smoothing spline estimator, when properly regularized, exhibits robustness against adversarial corruption. These results imply that if $o(n)$ out of $n$ samples are corrupted, the estimation error of the smoothing spline vanishes as $n \to \infty$. On the other hand, when a constant fraction of the data is corrupted, no estimator can guarantee vanishing estimation error, implying the optimality of the smoothing spline in terms of maximum tolerable number of corrupted samples.



Paperid:675
Authors:Konstantin Usevich, Ricardo Borsoi, Clara Dérand, Marianne Clausel
Abstract:
Polynomial Neural Networks (PNNs) possess a rich algebraic and geometric structure. However, their identifiability---a key property for ensuring interpretability---remains poorly understood. In this work, we present a comprehensive analysis of the identifiability of deep PNNs, including architectures with and without bias terms. Our results reveal an intricate interplay between activation degrees and layer widths in achieving identifiability. As special cases, we show that architectures with non-increasing layer widths are generically identifiable under mild conditions, while encoder-decoder networks are identifiable when the decoder widths do not grow too rapidly. Our proofs are constructive and center on a connection between deep PNNs and low-rank tensor decompositions, and Kruskal-type uniqueness theorems. This yields both generic conditions determined by the architecture, and effective conditions that depend on the network’s parameters. We also settle an open conjecture on the expected dimension of PNN's neurovarieties, and provide new bounds on the activation degrees required for it to reach its maximum.



Paperid:676
Authors:Michael Niemeyer, Fabian Manhardt, Marie-Julie Rakotosaona, Michael Oechsle, Christina Tsalicoglou, Keisuke Tateno, Jonathan Barron, Federico Tombari
Abstract:
Recent advances in neural scene representations have led to unprecedented quality in 3D reconstruction and view synthesis. Despite achieving high-quality results for common benchmarks with curated data, outputs often degrade for data that contain per image variations such as strong exposure changes, present, e.g., in most scenes with indoor and outdoor areas or rooms with windows. In this paper, we introduce Neural Exposure Fields (NExF), a novel technique for robustly reconstructing 3D scenes with high quality and 3D-consistent appearance from challenging real-world captures. In the core, we propose to learn a neural field predicting an optimal exposure value per 3D point, enabling us to optimize exposure along with the neural scene representation. While capture devices such as cameras select optimal exposure per image/pixel, we generalize this concept and perform optimization in 3D instead. This enables accurate view synthesis in high dynamic range scenarios, bypassing the need of post-processing steps or multi-exposure captures. Our contributions include a novel neural representation for exposure prediction, a system for joint optimization of the scene representation and the exposure field via a novel neural conditioning mechanism, and demonstrated superior performance on challenging real-world data. We find that our approach trains faster than prior works and produces state-of-the-art results on several benchmarks improving by over 55% over best-performing baselines.



Paperid:677
Authors:Julian Asilis, Mikael Møller Høgsgaard, Grigoris Velegkas
Abstract:
Abstract:Binary classification in the classic PAC model exhibits a curious phenomenon: Empirical Risk Minimization (ERM) learners are suboptimal in the realizable case yet optimal in the agnostic case. Roughly speaking, this owes itself to the fact that non-realizable distributions $\mathcal{D}$ are more difficult to learn than realizable distributions -- even when one discounts a learner's error by $\mathrm{err}(h^*_\mathcal{D})$, the error of the best hypothesis in $\mathcal{H}$. Thus, optimal agnostic learners are permitted to incur excess error on (easier-to-learn) distributions $\mathcal{D}$ for which $\tau = \mathrm{err}(h^*_\mathcal{D})$ is small.Recent work of Hanneke, Larsen, and Zhivotovskiy (FOCS '24) addresses this shortcoming by including $\tau$ itself as a parameter in the agnostic error term. In this more fine-grained model, they demonstrate tightness of the error lower bound $\tau + \Omega \left(\sqrt{\frac{\tau (d + \log(1 / \delta))}{m}} + \frac{d + \log(1 / \delta)}{m} \right)$ in a regime where $\tau > d/m$, and leave open the question of whether there may be a higher lower bound when $\tau \approx d/m$, with $d$ denoting $\mathrm{VC}(\mathcal{H})$.In this work, we resolve this question by exhibiting a learner which achieves error $c \cdot \tau + O \left(\sqrt{\frac{\tau (d + \log(1 / \delta))}{m}} + \frac{d + \log(1 / \delta)}{m} \right)$ for a constant $c \leq 2.1$, matching the lower bound and demonstrating optimality when $\tau =O( d/m)$. Further, our learner is computationally efficient and is based upon careful aggregations of ERM classifiers, making progress on two other questions of Hanneke, Larsen, and Zhivotovskiy (FOCS '24). We leave open the interesting question of whether our approach can be refined to lower the constant from 2.1 to 1, which would completely settle the complexity of agnostic learning.



Paperid:678
Authors:Zhihan Zhou, Feng Hong, JIAAN LUO, Yushi Ye, Jiangchao Yao, Dongsheng Li, Bo Han, Ya Zhang, Yanfeng Wang
Abstract:
We propose LIT, an advancement of visual instruction tuning (VIT). While VIT equips Multimodal LLMs (MLLMs) with promising multimodal capabilities, the current design choices for VIT often result in overfitting and shortcut learning, potentially degrading performance. This gap arises from an overemphasis on instruction-following abilities, while neglecting the proactive understanding of visual information. Inspired by this, LIT adopts a simple yet effective approach by incorporating the loss function into both the instruction and response sequences. It seamlessly expands the training data, and regularizes the MLLMs from overly relying on language priors. Based on this merit, LIT achieves a significant relative improvement of up to 9% on comprehensive multimodal benchmarks, requiring no additional training data and incurring negligible computational overhead. Surprisingly, LIT attains exceptional fundamental visual capabilities, yielding up to an 18% improvement in captioning performance, while simultaneously alleviating hallucination in MLLMs. The model weights and source code will be publicly available.



Paperid:679
Authors:Mingyu Xu, Tenglong Ao, Jiaao He, Jianqiao Lu, Guang Shi, Shu Zhong
Abstract:
Abstract:In recent years, large language models with Transformer architecture as the core have made breakthrough progress in many fields. At the same time, there are also some weaknesses in the large language model that have prompted people to reflect, among which the most fundamental one is the reflection on the Transformer architecture. The Transformer architecture has high parallelism and can fully utilize the computing power of GPUs, thus replacing models such as LSTM in the past few years. However, high parallelism is not a free lunch, as it fundamentally limits the performance of models. Especially, the problems that logarithmic precision Transformer architecture can solve are strictly limited to the $TC^0$. And there are many important issues that are usually considered out of $TC^0$, such as Python code evaluation, entity tracking, chess, and other state tracking tasks. Meanwhile, some recent state space methods based on Delta Rule have been able to break through the $TC^0$ architecture, but they are limited by fixed size state spaces and perform poorly on many tasks. To this end, we have re-examined the Transformer from the perspective of a state space with kernel functions and propose an improved Transformer called DeltaFormer. We have theoretically and practically demonstrated that the proposed new architecture can break through the limitation of the inherent $TC^0$ expressivity of Transformers and verified that it is not weaker than standard Transformer in language modeling tasks. We hope our work can provide inspiration for designing more expressive models.



Paperid:680
Authors:Federico Stella, Nicolas Talabot, Hieu Le, Pascal Fua
Abstract:
Unsigned Distance Fields (UDFs) are a natural implicit representation for open surfaces but, unlike Signed Distance Fields (SDFs), are challenging to triangulate into explicit meshes. This is especially true at high resolutions where neural UDFs exhibit higher noise levels, which makes it hard to capture fine details.Most current techniques perform within single voxels without reference to their neighborhood, resulting in missing surface and holes where the UDF is ambiguous or noisy. We show that this can be remedied by performing several passes and by reasoning on previously extracted surface elements to incorporate neighborhood information. Our key contribution is an iterative neural network that does this and progressively improves surface recovery within each voxel by spatially propagating information from increasingly distant neighbors. Unlike single-pass methods, our approach integrates newly detected surfaces, distance values, and gradients across multiple iterations, effectively correcting errors and stabilizing extraction in challenging regions. Experiments on diverse 3D models demonstrate that our method produces significantly more accurate and complete meshes than existing approaches, particularly for complex geometries, enabling UDF surface extraction at higher resolutions where traditional methods fail.



Paperid:681
Authors:Hao-Yuan He, Ming LI
Abstract:
This paper presents a comprehensive theoretical analysis of the learnability of neuro-symbolic (NeSy) tasks within hybrid systems. We characterize the learnability of NeSy tasks by their derived constraint satisfaction problems (DCSPs), demonstrating that a task is learnable if and only if its corresponding DCSP admits a unique solution. Under mild assumptions, we establish the sample complexity for learnable tasks and show that, for general tasks, the asymptotic expected concept error is controlled by the degree of disagreement among DCSP solutions. Our findings unify the characterization of learnability and the phenomenon of reasoning shortcuts, providing theoretical guarantees and actionable guidance for the principled design of NeSy systems.



Paperid:682
Authors:Jeff Brown, Andrew Kirjner, Annika Vivekananthan, Edward Boyden
Abstract:
Connectomics—the mapping of neural connections in an organism's brain—currently requires extraordinary human effort to proofread the data collected from imaging and machine-learning assisted segmentation. With the growing excitement around using AI agents to automate important scientific tasks, we explore whether current AI systems can perform multiple tasks necessary for data proofreading. We introduce ConnectomeBench, a multimodal benchmark evaluating large language model (LLM) capabilities in three critical proofreading tasks: segment type identification, split error correction, and merge error detection. Using two large open-source datasets—a cubic millimeter of mouse visual cortex and the complete Drosophila brain—we systematically evaluate proprietary multimodal LLMs including Claude 3.7 Sonnet, o4-mini, GPT-4.1, GPT-4o, as well as open source models like InternVL-3 and NVLM. Our results demonstrate that while some current models achieve suprisingly high performance in segment identification (67-78\% accuracy) and pairwise split error correction (75-81\% accuracy), they struggle with single-shot split error correction, as well as single-shot and pairwise merge error identification. These findings suggest that while LLMs show promising capabilities, substantial improvements are needed before they can reliably augment and, eventually, replace human proofreading in connectomics.



Paperid:683
Authors:Arnab Bhattacharyya, XianJun, Davin Choo, Philips George John, Themis Gouleakis
Abstract:
Abstract:Given i.i.d.~samples from an unknown distribution $P$, the goal of distribution learning is to recover the parameters of a distribution that is close to $P$. When $P$ belongs to the class of product distributions on the Boolean hypercube $\{0,1\}^d$, it is known that $\Omega(d/\epsilon^2)$ samples are necessary to learn $P$ within total variation (TV) distance $\epsilon$. We revisit this problem when the learner is also given as advice the parameters of a product distribution $Q$. We show that there is an efficient algorithm to learn $P$ within TV distance $\epsilon$ that has sample complexity $\tilde{O}(d^{1-\eta}/\epsilon^2)$, if $\|\mathbf{p} - \mathbf{q}\|_1<\epsilon d^{0.5 - \Omega(\eta)}$. Here, $\mathbf{p}$ and $\mathbf{q}$ are the mean vectors of $P$ and $Q$ respectively, and no bound on $\|\mathbf{p} - \mathbf{q}\|_1$ is known to the algorithm a priori.



Paperid:684
Authors:Antonin Joly, Nicolas Keriven, Aline Roumy
Abstract:
Graph coarsening aims to diminish the size of a graph to lighten its memory footprint, and has numerous applications in graph signal processing and machine learning. It is usually defined using a reduction matrix and a lifting matrix, which, respectively, allows to project a graph signal from the original graph to the coarsened one and back. This results in a loss of information measured by the so-called Restricted Spectral Approximation (RSA). Most coarsening frameworks impose a fixed relationship between the reduction and lifting matrices, generally as pseudo-inverses of each other, and seek to define a coarsening that minimizes the RSA.In this paper, we remark that the roles of these two matrices are not entirely symmetric: indeed, putting constraints on thelifting matrix aloneensures the existence of important objects such as the coarsened graph's adjacency matrix or Laplacian.In light of this, in this paper, we introduce a more general notion of reduction matrix, that isnotnecessarily the pseudo-inverse of the lifting matrix. We establish a taxonomy of ``admissible'' families of reduction matrices, discuss the different properties that they must satisfy and whether they admit a closed-form description or not. We show that, for afixedcoarsening represented by a fixed lifting matrix, the RSA can befurtherreduced simply by modifying the reduction matrix. We explore different examples, including some based on a constrained optimization process of the RSA. Since this criterion has also been linked to the performance of Graph Neural Networks, we also illustrate the impact of this choices on different node classification tasks on coarsened graphs.



Paperid:685
Authors:Zhenting Qi, Fan Nie, Alexandre Alahi, James Zou, Himabindu Lakkaraju, Yilun Du, Eric Xing, Sham Kakade, Hanlin Zhang
Abstract:
Modern language model (LM) training has been divided into multiple stages, making it difficult for downstream developers to evaluate the impact of design choices made at each stage.We present a model suite that enables systematic analysis of LMs' training dynamics across their entire lifecycle, including pre-training, continued pre-training, supervised fine-tuning, and reinforcement learning. By training over 100 LMs with 1B and 4B parameters from scratch, we rigorously evaluate both upstream (language modeling) and downstream (problem-solving) reasoning capabilities, including considerations of both in-domain and out-of-domain generalization. Key insights highlight the diminishing returns from excessive pre-training and post-training, the importance and practices of mitigating forgetting during domain-specific continued pre-training, the crucial role of continued pre-training in bridging pretraining and post-training phases, and various intricate trade-offs when configuring supervised fine-tuning and reinforcement learning. To facilitate open research and reproducibility, we release all pretrained and post-trained models, training datasets, and our entire training and evaluation pipeline.



Paperid:686
Authors:Panagiotis Theodoropoulos, Augustinos Saravanos, Evangelos Theodorou, Guan-Horng Liu
Abstract:
Understanding complex systems by inferring trajectories from sparse sample snapshots is a fundamental challenge in a wide range of domains, e.g., single-cell biology, meteorology, and economics. Despite advancements in Bridge and Flow matching frameworks, current methodologies rely on pairwise interpolation between adjacent snapshots. This hinders their ability to capture long-range temporal dependencies and potentially affects the coherence of the inferred trajectories. To address these issues, we introduce Momentum Multi-Marginal Schrödinger Bridge Matching (3MSBM), a novel matching framework that learns smooth measure-valued splines for stochastic systems that satisfy multiple positional constraints. This is achieved by lifting the dynamics to phase space and generalizing stochastic bridges to be conditioned on several points, forming a multi-marginal conditional stochastic optimal control problem. The underlying dynamics are then learned by minimizing a variational objective, having fixed the path induced by the multi-marginal conditional bridge. As a matching approach, 3MSBM learns transport maps that preserve intermediate marginals throughout training, significantly improving convergence and scalability. Extensive experimentation in a series of real-world applications validates the superior performance of 3MSBM compared to existing methods in capturing complex dynamics with temporal dependencies, opening new avenues for training matching frameworks in multi-marginal settings.



Paperid:687
Authors:Shuchen Wu, Stephan Alaniz, Shyamgopal Karthik, Peter Dayan, Eric Schulz, Zeynep Akata
Abstract:
Neural networks are often described as black boxes, reflecting the significant challenge of understanding their internal workings and interactions. We propose a different perspective that challenges the prevailing view: rather than being inscrutable, neural networks exhibit patterns in their raw population activity that mirror regularities in the training data. We refer to this as the \textit{Reflection Hypothesis} and provide evidence for this phenomenon in both simple recurrent neural networks (RNNs) and complex large language models (LLMs).Building on this insight, we propose to leverage cognitively-inspired methods of \textit{chunking} to segment high-dimensional neural population dynamics into interpretable units that reflect underlying concepts.We propose three methods to extract these emerging entities, complementing each other based on label availability and neural data dimensionality. Discrete sequence chunking (DSC) creates a dictionary of entities in a lower-dimensional neural space; population averaging (PA) extracts recurring entities that correspond to known labels; and unsupervised chunk discovery (UCD) can be used when labels are absent. We demonstrate the effectiveness of these methods in extracting entities across varying model sizes, ranging from inducing compositionality in RNNs to uncovering recurring neural population states in large language models with diverse architectures, and illustrate their advantage to other interpretability methods. Throughout, we observe a robust correspondence between the extracted entities and concrete or abstract concepts in the sequence. Artificially inducing the extracted entities in neural populations effectively alters the network's generation of associated concepts.Our work points to a new direction for interpretability, one that harnesses both cognitive principles and the structure of naturalistic data to reveal the hidden computations of complex learning systems, gradually transforming them from black boxes into systems we can begin to understand.



Paperid:688
Authors:Tal Argov, Tal Wagner
Abstract:
Explainable clustering by axis-aligned decision trees was introduced by Moshkovitz et al. (2020) and has gained considerable interest. Prior work has focused on minimizing the price of explainability for specific clustering objectives, lacking a general method to fit an explanation tree to any given clustering, without restrictions. In this work, we propose a new and generic approach to explainable clustering, based on spectral graph partitioning. With it, we design an explainable clustering algorithm that can fit an explanation tree to any given non-explainable clustering, or directly to the dataset itself. Moreover, we show that prior algorithms can also be interpreted as graph partitioning, through a generalized framework due to Trevisan (2013) wherein cuts are optimized in two graphs simultaneously. Our experiments show the favorable performance of our method compared to baselines on a range of datasets.



Paperid:689
Authors:Will Merrill, Ashish Sabharwal
Abstract:
Abstract:Chain of thought is a natural inference-time method for increasing the computational power of transformer-based large language models (LLMs), but comes at the cost of sequential decoding. Are there more efficient alternatives to expand a transformer's expressive power without adding parameters? We consider transformers with *padding* tokens as a form of parallelizable test-time compute. We show that fixed-depth averaging-hard-attention transformers with polynomial padding converge to precisely the class $\mathsf{TC}^0$ of extremely parallelizable problems. While the $\mathsf{TC}^0$ upper bound was known, proving a matching lower bound had been elusive. Further, our novel analysis reveals the precise expanded power of padded transformers when coupled with another form of inference-time compute, namely dynamically increasing depth via *looping*. Our core technical contribution is to show how padding helps bring the notions of *complete problems* and *reductions*, which have been a cornerstone of classical complexity theory, to the formal study of transformers. Armed with this new tool, we prove that padded transformers with $O(\log^d n)$ looping recognize exactly the class $\mathsf{TC}^d$ of similarly deep threshold circuits. Thus, padding and looping together greatly and systematically expand the expressive power of transformers: with polylogarithmic looping, padded transformers converge to the class $\mathsf{NC}$, the best that could be expected without losing parallelism (unless $\mathsf{NC} = \mathsf{P}$). Our results thus motivate further exploration of padding and looping as parallelizable alternatives to chain of thought.



Paperid:690
Authors:Brian Cho, Aurelien Bibaut, Nathan Kallus
Abstract:
Multi-arm bandit experimental designs are increasingly being adopted over standard randomized trials due to their potential to improve outcomes for study participants, enable faster identification of the best-performing options, and/or enhance the precision of estimating key parameters. Current approaches for inference after adaptive sampling either rely on asymptotic normality under restricted experiment designs or underpowered martingale concentration inequalities that lead to weak power in practice. To bypass these limitations, we propose a simulation-based approach for conducting hypothesis tests and constructing confidence intervals for arm specific means and their differences. Our simulation-based approach uses positively biased nuisances to generate additional trajectories of the experiment, which we call \textit{simulation with optimism}. Using these simulations, we characterize the distribution potentially non-normal sample mean test statistic to conduct inference. We provide guarantees for (i) asymptotic type I error control, (ii) convergence of our confidence intervals, and (iii) asymptotic strong consistency of our estimator over a wide variety of common bandit designs. Our empirical results show that our approach achieves the desired coverage while reducing confidence interval widths by up to 50\%, with drastic improvements for arms not targeted by the design.



Paperid:691
Authors:Jin-Ting He, Fu-Jen Tsai, Yan-Tsung Peng, Min-Hung Chen, Chia-Wen Lin, Yen-Yu Lin
Abstract:
Diffusion models show promise for dynamic scene deblurring; however, existing studies often fail to leverage the intrinsic nature of the blurring process within diffusion models, limiting their full potential. To address it, we present a Blur Diffusion Model (BlurDM), which seamlessly integrates the blur formation process into diffusion for image deblurring. Observing that motion blur stems from continuous exposure, BlurDM implicitly models the blur formation process through a dual-diffusion forward scheme, diffusing both noise and blur onto a sharp image. During the reverse generation process, we derive a dual denoising and deblurring formulation, enabling BlurDM to recover the sharp image by simultaneously denoising and deblurring, given pure Gaussian noise conditioned on the blurred image as input. Additionally, to efficiently integrate BlurDM into deblurring networks, we perform BlurDM in the latent space, forming a flexible prior generation network for deblurring. Extensive experiments demonstrate that BlurDM significantly and consistently enhances existing deblurring methods on four benchmark datasets.



Paperid:692
Authors:Ngoc Bao Nguyen, Hieu Nguyen, Ruifeng She, Xiaojin Fu, Viet Anh Nguyen
Abstract:
Abstract:Selecting an appropriate reasoning method for a given query remains a key challenge in language model generation. Existing methods typically generate multiple candidate answers and use an aggregation strategy to select the output answer, often assuming that more candidate answers yield higher accuracy. We challenge this assumption through a rigorous theoretical analysis, deriving accuracy bounds for common aggregation methods under fixed generation distributions and candidate sizes. Building on these insights, we introduce EPIC, an $\textbf{E}$nsemble $\textbf{P}$lann$\textbf{I}$ng with $\textbf{C}$ontrastive learning framework to learn a shared representation space that captures both model reasoning abilities and query-method compatibility. EPIC incorporates our probability bounds as a regularizer in a utility-driven optimization that balances accuracy and computational cost. Experiments on diverse mathematical reasoning tasks show that EPIC consistently selects optimal reasoning methods, improving accuracy while reducing computational overhead.



Paperid:693
Authors:Roman Garipov, Fedor Velikonivtsev, Ivan Ermakov, Ruslan Svirschevski, Vage Egiazarian, Max Ryabinin
Abstract:
Abstract:We introduce AutoJudge, a framework that accelerates large language model (LLM) inference with task-specific lossy speculative decoding. Instead of matching the original model output distribution token-by-token, we identify which of the generated tokens affect the downstream quality of the generated response, relaxing the guarantee so that the "unimportant" tokens can be generated faster.Our approach relies on a semi‑greedy search algorithm to test which of the mismatches between target and draft models should be corrected to preserve quality and which ones may be skipped.We then train a lightweight classifier based on existing LLM embeddings to predict, at inference time, which mismatching tokens can be safely accepted without compromising the final answer quality.We evaluate the effectiveness of AutoJudge with multiple draft/target model pairs on mathematical reasoning and programming benchmarks, achieving significant speedups at the cost of minor accuracy drawdown. Notably, on GSM8k with Llama 3.1 70B target model, our approach achieves up to ${\approx}2{\times}$ speedup *over speculative decoding* at the cost of ${\le} 1\%$ drop in accuracy.When applied to the LiveCodeBench benchmark, AutoJudge automatically detects programming-specific important tokens, accepting ${\ge}25$ tokens per speculation cycle at ${\le} 1\%$ drop in Pass@1. Our approach requires no human annotation and is easy to integrate with modern LLM inference software.



Paperid:694
Authors:Yisong Fu, Zezhi Shao, Chengqing Yu, Yujie Li, Zhulin An, Qi Wang, Yongjun Xu, Fei Wang
Abstract:
Benefiting from high capacity for capturing complex temporal patterns, deep learning (DL) has significantly advanced time series forecasting (TSF). However, deep models tend to suffer from severe overfitting due to the inherent vulnerability of time series to noise and anomalies. The prevailing DL paradigm uniformly optimizes all timesteps through the MSE loss and learns those uncertain and anomalous timesteps without difference, ultimately resulting in overfitting. To address this, we propose a novel selective learning strategy for deep TSF. Specifically, selective learning screens a subset of the whole timesteps to calculate the MSE loss in optimization, guiding the model to focus on generalizable timesteps while disregarding non-generalizable ones. Our framework introduces a dual-mask mechanism to target timesteps: (1) an uncertainty mask leveraging residual entropy to filter uncertain timesteps, and (2) an anomaly mask employing residual lower bound estimation to exclude anomalous timesteps. Extensive experiments across eight real-world datasets demonstrate that selective learning can significantly improve the predictive performance for typical state-of-the-art deep models, including 37.4% MSE reduction for Informer, 8.4% for TimesNet, and 6.5% for iTransformer.



Paperid:695
Authors:ShengYun Peng, Pin-Yu Chen, Jianfeng Chi, Seongmin Lee, Duen Horng Chau
Abstract:
Finetuning large language models (LLMs) enables user-specific customization but introduces critical safety risks: even a few harmful examples can compromise safety alignment.A common mitigation strategy is to update the model more strongly on examples deemed safe, while downweighting or excluding those flagged as unsafe.However, because safety context can shift within a single example, updating the model equally on both harmful and harmless parts of a response is suboptimal—a coarse treatment we term static safety shaping.In contrast, we propose dynamic safety shaping (DSS), a framework that uses fine-grained safety signals to reinforce learning from safe segments of a response while suppressing unsafe content.To enable such fine-grained control during finetuning, we introduce a key insight: guardrail models, traditionally used for filtering, can be repurposed to evaluate partial responses, tracking how safety risk evolves throughout the response, segment by segment.This leads to the Safety Trajectory Assessment of Response (STAR), a token-level signal that enables shaping to operate dynamically over the training sequence.Building on this, we present ★DSS, guided by STAR scores, that robustly mitigates finetuning risks and delivers substantial safety improvements across diverse threats, datasets, and model families—all without compromising capability on intended tasks.We encourage future safety research to build on dynamic shaping principles for stronger mitigation against evolving finetuning risks.



Paperid:696
Authors:Piotr Faliszewski, Łukasz Janeczko, Andrzej Kaczmarczyk, Grzegorz Lisowski, Piotr Skowron, Stanisław Szufa, Mateusz Szwagierczak
Abstract:
We study strategic behavior of project proposers in the context of approval-basedparticipatory budgeting (PB). In our model we assume that the votes are fixed andknown and the proposers want to set as high project prices as possible, providedthat their projects get selected and the prices are not below the minimum costs oftheir delivery. We study the existence of pure Nash equilibria (NE) in such games,focusing on the AV/Cost, Phragmen, and Method of Equal Shares rules. We alsoprovide an experimental study of cost selection on real-life PB election data.



Paperid:697
Authors:Sai Tanmay Reddy Chakkera, Aggelina Chatziagapi, Md Moniruzzaman, Chen-Ping Yu, Yi-Hsuan Tsai, Dimitris Samaras
Abstract:
We introduce a novel method for low-rank personalization of a generic model for head avatar generation. Prior work proposes generic models that achieve high-quality face animation by leveraging large-scale datasets of multiple identities. However, such generic models usually fail to synthesize unique identity-specific details, since they learn a general domain prior. To adapt to specific subjects, we find that it is still challenging to capture high-frequency facial details via popular solutions like low-rank adaptation (LoRA). This motivates us to propose a specific architecture, a Register Module, that enhances the performance of LoRA, while requiring only a small number of parameters to adapt to an unseen identity. Our module is applied to intermediate features of a pre-trained model, storing and re-purposing information in a learnable 3D feature space. To demonstrate the efficacy of our personalization method, we collect a dataset of talking videos of individuals with distinctive facial details, such as wrinkles and tattoos. Our approach faithfully captures unseen faces, outperforming existing methods quantitatively and qualitatively. We will release the code, models, and dataset to the public.



Paperid:698
Authors:Laetitia Chapel, Romain Tavenard, Samuel Vaiter
Abstract:
Optimal Transport (OT) has attracted significant interest in the machine learning community, not only for its ability to define meaningful distances between probability distributions -- such as the Wasserstein distance -- but also for its formulation of OT plans. Its computational complexity remains a bottleneck, though, and slicing techniques have been developed to scale OT to large datasets. Recently, a novel slicing scheme, dubbed min-SWGG, lifts a single one-dimensional plan back to the original multidimensional space, finally selecting the slice that yields the lowest Wasserstein distance as an approximation of the full OT plan. Despite its computational and theoretical advantages, min-SWGG inherits typical limitations of slicing methods: (i) the number of required slices grows exponentially with the data dimension, and (ii) it is constrained to linear projections. Here, we reformulate min-SWGG as a bilevel optimization problem and propose a differentiable approximation scheme to efficiently identify the optimal slice, even in high-dimensional settings. We furthermore define its generalized extension for accommodating to data living on manifolds. Finally, we demonstrate the practical value of our approach in various applications, including gradient flows on manifolds and high-dimensional spaces, as well as a novel sliced OT-based conditional flow matching for image generation -- where fast computation of transport plans is essential.



Paperid:699
Authors:Peiman Mohseni, Nick Duffield
Abstract:
Conditional Neural Processes (CNPs) are a class of probabilistic models that leverage neural networks to parameterize conditional stochastic processes. Through simple maximum-likelihood training and the ability to produce well‐calibrated uncertainty estimates, CNPs have emerged as powerful tools for modeling data that naturally resides on a continuum—such as temporal signals, spatial fields, or functional trajectories. However, early variants of the CNP family rely on fixed finite-dimensional representations of observations, which can lead to information loss and a mismatch with the inherently infinite-dimensional nature of stochastic processes, ultimately degrading predictive performance. Convolutional CNPs (ConvCNPs) address this issue by mapping observations into an infinite-dimensional embedding and applying a convolutional neural network. Yet, their reliance on local discrete kernels introduces a trade-off: large kernels are needed to capture long-range dependencies, but they come at significant computational and memory costs.To overcome this limitation, we exploit the compact and global representation provided by the frequency domain, inspired by Fourier Neural Operators (FNOs) in approximating solution operators of parametric partial differential equations (PDEs). We introduce Spectral Convolutional Conditional Neural Processes (SConvCNPs), which directly parameterize global convolutional filters in the frequency domain. We validate SConvCNPs' effectiveness through several experiments on both synthetic and real-world datasets.



Paperid:700
Authors:Zihui (Sherry) Xue, Romy Luo, Kristen Grauman
Abstract:
The Arrow of Time (AoT)—time's irreversible flow shaping physical events—is fundamental to video comprehension, yet remains a significant challenge for modern large multimodal models (LMMs). Current LMMs struggle to perceive and utilize temporal directionality in video when responding to language queries, obstructing deeper temporal understanding. We tackle this deficiency by first providing a critical analysis of existing benchmarks and models. We then introduce ArrowRL, a reinforcement learning (RL)-based training strategy with an innovative reverse reward that instills AoT awareness by encouraging divergent video interpretations between forward and reversed visual frames. For rigorous evaluation, we additionally develop AoTBench, a new multi-faceted benchmark probing temporally challenging questions. Experiments show ArrowRL greatly advances temporal perception: it not only achieves substantial improvements on our challenging AoTBench but also demonstrably boosts performance on standard video question answering (VQA) benchmarks (with peak accuracy gains reaching over 20% and 10% respectively). This validates ArrowRL's effectiveness and highlights the critical need for dedicated AoT understanding in LMMs.



Paperid:701
Authors:Yijun Hu, Bing Fan, Xin Gu, 海青 任, Dongfang Liu, Heng Fan, Libo Zhang
Abstract:
Establishing object-level correspondence between egocentric and exocentric views is essential for AI assistants to deliver precise and intuitive visual guidance. However, this task presents numerous challenges, including extreme viewpoint variations, occlusions, and the presence of many small objects. Existing methods usually borrow solutions from video object segmentation, such as XSegTx and XView-XMem, which struggle to address aforementioned difficulties effectively. Recently, the Segment Anything Model 2 (SAM 2) has demonstrated strong generalization capabilities and shown excellent performance in video object segmentation. Nevertheless, SAM 2 faces significant challenges when directly applied to the ego-exo correspondence (EEC) task, including suboptimal feature fusion between views and inefficient memory management for long video sequences. To address these limitations, we propose a novel EEC framework based on SAM 2 with long-term memories by presenting a dual-memory system and an adaptive Mixture-of-Experts module. Specifically, our approach features (1) a Memory-View Mixture-of-Experts module which consists of a dual-branch routing mechanism to adaptively assign contribution weights to each expert feature along both channel and spatial dimensions, and (2) a dual-memory bank system with a dedicated compression strategy to retain critical long-term information while eliminating redundancy. In the extensive experiments on the challenging EgoExo4D benchmark, our method, dubbed LM-EEC, achieves new state-of-the-art results and significantly outperforms existing methods and the SAM 2 baseline, showcasing its strong generalization across diverse scenarios. Our code will be released.



Paperid:702
Authors:Boxuan Zhang, Runqing Wang, Wei Xiao, Weipu Zhang, Jian Sun, Gang Wang, Gao Huang, Jie Chen
Abstract:
Abstract:A critical bottleneck in deep reinforcement learning (DRL) is sample inefficiency, as training high-performance agents often demands extensive environmental interactions. Model-based reinforcement learning (MBRL) mitigates this by building world models that simulate environmental dynamics and generate synthetic experience, improving sample efficiency. However, conventional world models process observations holistically, failing to decouple dynamic objects and temporal features from static backgrounds. This approach is computationally inefficient, especially for visual tasks where dynamic objects significantly influence rewards and decision-making performance. To address this, we introduce DyMoDreamer, a novel MBRL algorithm that incorporates a dynamic modulation mechanism to improve the extraction of dynamic features and enrich the temporal information. DyMoDreamer employs differential observations derived from a novel inter-frame differencing mask, explicitly encoding object-level motion cues and temporal dynamics. Dynamic modulation is modeled as stochastic categorical distributions and integrated into a recurrent state-space model (RSSM), enhancing the model's focus on reward-relevant dynamics. Experiments demonstrate that DyMoDreamer sets a new state-of-the-art on the Atari $100$k benchmark with a $156.6$\% mean human-normalized score, establishes a new record of $832$ on the DeepMind Visual Control Suite, and gains a $9.5$\% performance improvement after $1$M steps on the Crafter benchmark.



Paperid:703
Authors:Gustavo Perez, Stella X. Yu
Abstract:
Classical image filters, such as those for averaging or differencing, are carefully normalized to ensure consistency, interpretability, and to avoid artifacts like intensity shifts, halos, or ringing. In contrast, convolutional filters learned end-to-end in deep networks lack such constraints. Although they may resemble wavelets and blob/edge detectors, they are not normalized in the same or any way. Consequently, when images undergo atmospheric transfer, their responses become distorted, leading to incorrect outcomes. We address this limitation by proposing filter normalization, followed by learnable scaling and shifting, akin to batch normalization. This simple yet effective modification ensures that the filters are atmosphere-equivariant, enabling co-domain symmetry. By integrating classical filtering principles into deep learning (applicable to both convolutional neural networks and convolution-dependent vision transformers), our method achieves significant improvements on artificial and natural intensity variation benchmarks. Our ResNet34 could even outperform CLIP by a large margin. Our analysis reveals that unnormalized filters degrade performance, whereas filter normalization regularizes learning, promotes diversity, and improves robustness and generalization.



Paperid:704
Authors:Zara Hall, Melanie Subbiah, Thomas Zollo, Kathleen McKeown, Richard Zemel
Abstract:
Large language models are increasingly used to support high-stakes decisions, potentially influencing who is granted bail or receives a loan. Naive chain-of-thought sampling can improve average decision accuracy, but has also been shown to amplify unfair bias. To address this challenge and enable the trustworthy use of reasoning models in high-stakes decision-making, we propose a framework for training a generalizable Fairness Reward Model (FRM). Our model assigns a fairness score to LLM reasoning, enabling the system to down-weight biased trajectories and favor equitable ones when aggregating decisions across reasoning chains. We show that a single Fairness Reward Model, trained on weakly supervised, LLM-annotated examples of biased versus unbiased reasoning, transfers across tasks, domains, and model families without additional fine-tuning. Applied to real-world decision-making tasks including recidivism prediction and social media moderation, we show that our approach consistently improves fairness while matching, or even surpassing, baseline accuracy. We explore design decisions in our approach, such as labeling methods and the choice to use process (step-level) vs. outcome (full reasoning chain) rewards.



Paperid:705
Authors:Lei Lv, Yunfei Li, Yu Luo, Fuchun Sun, Tao Kong, Jiafeng Xu, Xiao Ma
Abstract:
Abstract:We present $\textbf{FlowRL}$, a novel framework for online reinforcement learning that integrates flow-based policy representation with Wasserstein-2-regularized optimization. We argue that in addition to training signals, enhancing the expressiveness of the policy class is crucial for the performance gains in RL. Flow-based generative models offer such potential, excelling at capturing complex, multimodal action distributions. However, their direct application in online RL is challenging due to a fundamental objective mismatch: standard flow training optimizes for static data imitation, while RL requires value-based policy optimization through a dynamic buffer, leading to difficult optimization landscapes. FlowRL first models policies via a state-dependent velocity field, generating actions through deterministic ODE integration from noise. We derive a constrained policy search objective that jointly maximizes Q through the flow polciy while bounding the Wasserstein-2 distance to a behavior-optimal policy implicitly derived from the replay buffer. This formulation effectively aligns the flow optimization with the RL objective, enabling efficient and value-aware policy learning despite the complexity of the policy class. Empirical evaluations on DMControl and Humanoidbench demonstrate that FlowRL achieves competitive performance in online reinforcement learning benchmarks.



Paperid:706
Authors:Yuwei Zhang, Kumar Ayush, Siyuan Qiao, A. Ali Heydari, Girish Narayanswamy, Maxwell Xu, Ahmed Metwally, Jinhua Xu, Jake Garrison, Xuhai "Orson" Xu, Tim Althoff, Yun Liu, Pushmeet Kohli, Jiening Zhan, Mark Malhotra, Shwetak Patel, Cecilia Mascolo, Xin Liu, Daniel McDuff, Yuzhe Yang
Abstract:
We present SensorLM, a family of sensor-language foundation models that enable wearable sensor data understanding with natural language. Despite its pervasive nature, aligning and interpreting sensor data with language remains challenging due to the lack of paired, richly annotated sensor-text descriptions in uncurated, real-world wearable data. We introduce a hierarchical caption generation pipeline designed to capture statistical, structural, and semantic information from sensor data. This approach enabled the curation of the largest sensor-language dataset to date, comprising over 59.7 million hours of data from more than 103,000 people. Furthermore, SensorLM extends prominent multimodal pretraining architectures (e.g., CLIP, CoCa) and recovers them as specific variants within a generic architecture. Extensive experiments on real-world tasks in human activity analysis and healthcare verify the superior performance of SensorLM over state-of-the-art in zero-shot recognition, few-shot learning, and cross-modal retrieval. SensorLM also demonstrates intriguing capabilities including scaling behaviors, label efficiency, sensor captioning, and zero-shot generalization to unseen tasks.



Paperid:707
Authors:Navdeep Kumar, Priyank Agrawal, Giorgia Ramponi, Kfir Y. Levy, Shie Mannor
Abstract:
Abstract:We analyze the global convergence of the single-timescale actor-critic (AC) algorithm for the infinite-horizon discounted Markov Decision Processes (MDPs) with finite state spaces. To this end, we introduce an elegant analytical framework for handling complex, coupled recursions inherent in the algorithm. Leveraging this framework, we establish that the algorithm converges to an $\epsilon$-close \textbf{globally optimal} policy with a sample complexity of $ O(\epsilon^{-3}) $. This significantly improves upon the existing complexity of $O(\epsilon^{-2})$ to achieve $\epsilon$-close \textbf{stationary policy}, which is equivalent to the complexity of $O(\epsilon^{-4})$ to achieve $\epsilon$-close \textbf{globally optimal} policy using gradient domination lemma. Furthermore, we demonstrate that to achieve this improvement, the step sizes for both the actor and critic must decay as $ O(k^{-\frac{2}{3}}) $ with iteration $k$, diverging from the conventional $O(k^{-\frac{1}{2}}) $ rates commonly used in (non)convex optimization.



Paperid:708
Authors:L. Elisa Celis, Lingxiao Huang, Milind Sohoni, Nisheeth K. Vishnoi
Abstract:
Meritocratic systems, from admissions to hiring, aim to impartially reward skill and effort. Yet, observed disparities in outcomes across race, gender, or class challenge this ideal. Some attribute these gaps to structural inequality; others to individual choice. We introduce a game-theoretic model in which candidates from different socioeconomic groups differ in perceived post-selection value; such values are shaped by social context and (increasingly) by AI-powered tools that can provide personalized career or salary guidance. In our model, each candidate strategically chooses the amount of effort to expend, weighing the cost against the expected reward. Effort is converted into observable merit, and selection is based solely on merit. We characterize the unique Nash equilibrium in the large-agent limit and derive explicit expressions for how valuation disparities and institutional selectivity jointly shape effort, representation, social welfare, and utility. We also formulate a cost-sensitive optimization framework to guide interventions by quantifying how shifts in selectivity or perceived value can reduce disparities without sacrificing institutional goals. Our results uncover a behavioral feedback loop: algorithmic or social signals about post-selection value distort pre-selection effort, propagating disparities backward through otherwise ``fair'' processes. This reconciles competing explanations of inequality—rational choice vs. structural bias—by showing how techno-social environments shape individual incentives in meritocratic systems.



Paperid:709
Authors:Jacob Imola, Fabrizio Boninsegna, Hannah Keller, Anders Aamand, Amrita Roy Chowdhury, Rasmus Pagh
Abstract:
Abstract:In the approximate quantiles problem, the goal is to output $m$ quantile estimates, the ranks of which are as close as possible to $m$ given quantiles $q_1,\dots,q_m$. We present a mechanism for approximate quantiles that satisfies $\varepsilon$-differential privacy for a dataset of $n$ real numbers where the ratio between the closest pair of points and the size of the domain is bounded by $b$.As long as the minimum gap between quantiles is large enough, $|q_i-q_{i-1}|\geq \Omega\left(\frac{m\log(m)\log(b)}{n\varepsilon}\right)$ for all $i$, the maximum rank error of our mechanism is $O\left(\frac{\log(b) + \log^2(m)}{\varepsilon}\right)$ with high probability.Previously, the best known algorithm under pure DP was due to Kaplan, Schnapp, and Stemmer (ICML '22), who achieve a bound of $O\left(\log(b)\log^2(m)/\varepsilon\right)$, so we save a factor $\Omega(\min(\log(b),\log^2(m)))$.Our improvement stems from the use of continual counting techniques to randomize the quantiles in a correlated way.We also present an $(\varepsilon,\delta)$-differentially private mechanism that relaxes the gap assumption without affecting the error bound, improving on existing methods when $\delta$ is sufficiently close to zero.We provide experimental evaluation which confirms that our mechanism performs favorably compared to prior work in practice, in particular when the number of quantiles $m$ is large.



Paperid:710
Authors:Bo Guo, Sijia Wen, Ziwei Wang, Yifan Zhao
Abstract:
Recent advances in 3D Gaussian Splatting (3DGS) have improved the visual fidelity of dynamic avatar reconstruction. However, existing methods often overlook the inherent chromatic similarity of human skin tones, leading to poor capture of intricate facial geometry under subtle appearance changes. This is caused by the affine approximation of Gaussian projection, which fails to be perspective-aware to depth-induced shear effects. To this end, we propose True-to-Geometry Avatar Dynamic Reconstruction (TGA), a perspective-aware 4D Gaussian avatar framework that sensitively captures fine-grained facial variations for accurate 3D geometry reconstruction. Specifically, to enable color-sensitive and geometry-consistent Gaussian representations under dynamic conditions, we introduce Perspective-Aware Gaussian Transformation that jointly models temporal deformations and spatial projection by integrating Jacobian-guided adaptive deformation into the homogeneous formulation. Furthermore, we develop Incremental BVH Tree Pivoting to enable fast frame-by-frame mesh extraction for 4D Gaussian representations. A dynamic Gaussian Bounding Volume Hierarchy (BVH) tree is used to model the topological relationships among points, where active ones are filtered out by BVH pivoting and subsequently re-triangulated for surface reconstruction. Extensive experiments demonstrate that TGA achieves superior geometric accuracy.



Paperid:711
Authors:Felix Chalumeau, Refiloe Shabe, Noah De Nicola, Arnu Pretorius, Tom Barrett, Nathan Grinsztajn
Abstract:
Routing Problems are central to many real-world applications, yet remain challenging due to their (NP-)hard nature. Amongst existing approaches, heuristics often offer the best trade-off between quality and scalability, making them suitable for industrial use. While Reinforcement Learning (RL) offers a flexible framework for designing heuristics, its adoption over handcrafted heuristics remains incomplete. Existing learned methods still lack the ability to adapt to specific instances and fully leverage the available computational budget. Current best methods either rely on a collection of pre-trained policies, or on RL fine-tuning; hence failing to fully utilize newly available information within the constraints of the budget. In response, we present MEMENTO, an approach that leverages memory to improve the search of neural solvers at inference. MEMENTO updates the action distribution dynamically based on the outcome of previous decisions. We validate its effectiveness on Traveling Salesman and Capacitated Vehicle Routing problems, demonstrating its superiority over tree-search and policy-gradient fine-tuning; and showing that it can be zero-shot combined with diversity-based solvers. We successfully train all RL auto-regressive solvers on large instances, and verify MEMENTO's scalability and data-efficiency: pushing the state-of-the-art on 11 out of 12 evaluated tasks.



Paperid:712
Authors:Jun Nie, Yonggang Zhang, Tongliang Liu, Yiu-ming Cheung, Bo Han, Xinmei Tian
Abstract:
We introduce a novel framework for AI-generated image detection through epistemic uncertainty, aiming to address critical security concerns in the era of generative models. Our key insight stems from the observation that distributional discrepancies between training and testing data manifest distinctively in the epistemic uncertainty space of machine learning models. In this context, the distribution shift between natural and generated images leads to elevated epistemic uncertainty in models trained on natural images when evaluating generated ones. Hence, we exploit this phenomenon by using epistemic uncertainty as a proxy for detecting generated images. This converts the challenge of generated image detection into the problem of uncertainty estimation, underscoring the generalization performance of the model used for uncertainty estimation. Fortunately, advanced large-scale vision models pre-trained on extensive natural images have shown excellent generalization performance for various scenarios. Thus, we utilize these pre-trained models to estimate the epistemic uncertainty of images and flag those with high uncertainty as generated. Extensive experiments demonstrate the efficacy of our method.



Paperid:713
Authors:Fu Feng, Yucheng Xie, Ruixiao Shi, Jianlu Shen, Jingq Wang, Xin Geng
Abstract:
Knowledge in modern neural networks is often entangled and structurally opaque, making current transfer methods—typically based on reusing entire parameter sets—inefficient and inflexible. Efforts to improve flexibility by reusing partial parameters frequently depend on handcrafted heuristics or rigid structural assumptions, which constrain generalization. In contrast, biological evolution enables efficient knowledge transfer by encoding only essential information into genes through iterative refinement under environmental pressure. Inspired by this principle, we propose \textbf{ECO}, a framework that \textbf{E}volves \textbf{CO}re knowledge into modular, reusable neural components—termed \textit{learngenes}—through similar evolutionary dynamics. To this end, we redefine learngenes as neural circuits and introduce Genetic Transfer Learning (GTL), a biologically inspired paradigm that establishes a genetic mechanism within neural networks. GTL simulates evolutionary processes by generating diverse network populations, selecting high-performing individuals, and transferring their learngenes to subsequent generations. Through iterative refinement, GTL enables learngenes to accumulate transferable, task-agnostic knowledge. Extensive experiments show that ECO achieves efficient initialization and strong generalization across diverse models and tasks, while significantly reducing computational and memory costs compared to conventional methods.



Paperid:714
Authors:Jichen Zhang, Liqun Zhao, Antonis Papachristodoulou, Jack Umenberger
Abstract:
Diffusion models have shown remarkable potential in planning and control tasks due to their ability to represent multimodal distributions over actions and trajectories. However, ensuring safety under constraints remains a critical challenge for diffusion models. This paper proposes Constrained Diffusers, a novel framework that incorporates constraints into pre-trained diffusion models without retraining or architectural modifications. Inspired by constrained optimization, we apply a constrained Langevin sampling mechanism for the reverse diffusion process that jointly optimizes the trajectory and realizes constraint satisfaction through three iterative algorithms: projected method, primal-dual method and augmented Lagrangian approaches. In addition, we incorporate discrete control barrier functions as constraints for constrained diffusers to guarantee safety in online implementation. Experiments in Maze2D, locomotion, and pybullet ball running tasks demonstrate that our proposed methods achieve constraint satisfaction with less computation time, competitive to existing methods in the environment with static and time-varying constraints. Codes are available in supplemental material.



Paperid:715
Authors:Luciano Vinas, Arash Amini
Abstract:
Abstract:We consider the empirical distribution of the embeddings of a $k$-layer polynomial GNN on a semi-supervised node classification task and prove a central limit theorem for them. Assuming a community based model for the underlying graph, with growing average degree $\nu_n\to\infty$, we show that the empirical distribution of the centered features, when scaled by $\nu_{n}^{k-1/2}$ converge in 1-Wasserstein distance to a centered stable mixture of multivariate normal distributions. In addition, the joint empirical distribution of uncentered features and labels when normalized by $\nu_n^k$ approach that of mixture of multivariate normal distributions, with stable means and covariance matrices vanishing as $\nu_n^{-1}$. We explicitly identify the asymptotic means and covariances, showing that the mixture collapses towards a 1-D version as $k$ is increased. Our results provides a precise and nuanced lens on how oversmoothing presents itself in the large graph limit, in the sparse regime. In particular, we show that training with cross-entropy on these embeddings is asymptotically equivalent to training on these nearly collapsed Gaussian mixtures.



Paperid:716
Authors:Sang Min Kim, Byeongchan Kim, Arijit Sehanobish, Somnath Basu Roy Chowdhury, Rahul Kidambi, Dongseok Shim, Kumar Avinava Dubey, Snigdha Chaturvedi, Min-hwan Oh, Krzysztof M Choromanski
Abstract:
Abstract:Efficient neural networks are essential for scaling machine learning models to real-time applications and resource-constrained environments. Fully-connected feedforward layers (FFLs) introduce computation and parameter count bottlenecks within neural network architectures. To address this challenge, in this work, we propose a new class of dense layers that generalize standard fully-connected feedforward layers, $\textbf{E}$fficient, $\textbf{U}$nified and $\textbf{Gen}$eral dense layers (EUGens). EUGens leverage random features to approximate standard FFLs and go beyond them by incorporating a direct dependence on the input norms in their computations. The proposed layers unify existing efficient FFL extensions and improve efficiency by reducing inference complexity from quadratic to linear time. They also lead to $\textbf{the first}$ unbiased algorithms approximating FFLs with arbitrary polynomial activation functions. Furthermore, EuGens reduce the parameter count and computational overhead while preserving the expressive power and adaptability of FFLs. We also present a layer-wise knowledge transfer technique that bypasses backpropagation, enabling efficient adaptation of EUGens to pre-trained models. Empirically, we observe that integrating EUGens into Transformers and MLPs yields substantial improvements in inference speed (up to $\textbf{27}$\%) and memory efficiency (up to $\textbf{30}$\%) across a range of tasks, including image classification, language model pre-training, and 3D scene reconstruction. Overall, our results highlight the potential of EUGens for the scalable deployment of large-scale neural networks in real-world scenarios.



Paperid:717
Authors:Shizhuo Zhang, Qirun Dai, Hao Peng
Abstract:
High-quality supervised finetuning (SFT) data are essential for unlocking pretrained LLMs’ capabilities. Typically, instructions are paired with responses from various sources—by human annotators or other LMs—which are often out of the distribution of the target model to be finetuned. At scale, this mismatch can lead to diminishing returns and even hurt model performance and robustness. We hypothesize that SFT is most effective when the data is aligned with the model’s pretrained distribution, and proposeGRAPE—a novel SFT framework that tailors supervision to the target model. For each instruction, itgathersresponses from various sources and selects the one thataligns most closely to the model’spretrained distribution, as measured by the normalized probability. Standard SFT is then performed on these selected responses.We first evaluate GRAPE in a controlled experiment, sampling multiple responses per question in the UltraInteract dataset from diverse models. We finetune using GRAPE-selected data on LMs from different families, including LLaMA-1-8B, Mistral-7B, and Qwen2.5-7B. GRAPE significantly outperforms strong baselines—including distilling from the strongest model—with absolute gains up to13.8%averaged across benchmarks, and outperforms a 3× larger data baseline with improvements up to17.3%.GRAPE's benefits generalize to off-the-shelf SFT data. When used to subsample from the post-training data of Tulu3 and Olmo-2, GRAPE surpasses strong baselines trained on 4.5× more data by6.1%, and outperforms state-of-the-art selection methods by3.9%on average. Notably, with only1/3 the dataandhalf the training epochs, GRAPE enables LLaMA-1-8B toexceed Tulu3-SFT performance by 3.5%.Our findings highlight that aligning supervision with the pretrained distribution provides a simple yet powerful strategy to improve both theefficiencyandeffectivenessof SFT.



Paperid:718
Authors:Santiago Mazuelas, Veronica Alvarez
Abstract:
Boosting methods often achieve excellent classification accuracy, but can experience notable performance degradation in the presence of label noise. Existing robust methods for boosting provide theoretical robustness guarantees for certain types of label noise, and can exhibit only moderate performance degradation. However, previous theoretical results do not account for realistic types of noise and finite training sizes, and existing robust methods can provide unsatisfactory accuracies, even without noise. This paper presents methods for robust minimax boosting (RMBoost) that minimize worst-case error probabilities and are robust to general types of label noise. In addition, we provide finite-sample performance guarantees for RMBoost with respect to the error obtained without noise and with respect to the best possible error (Bayes risk). The experimental results corroborate that RMBoost is not only resilient to label noise but can also provide strong classification accuracy in practice.



Paperid:719
Authors:Hyeongyu Kim, GeonHui Han, Dosik Hwang
Abstract:
In recent advancements in Test Time Adaptation (TTA), most existing methodologies focus on updating normalization layers to adapt to the test domain. However, the reliance on normalization-based adaptation presents key challenges. First, normalization layers such as Batch Normalization (BN) are highly sensitive to small batch sizes, leading to unstable and inaccurate statistics. Moreover, normalization-based adaptation is inherently constrained by the structure of the pre-trained model, as it relies on training-time statistics that may not generalize well to unseen domains. These issues limit the effectiveness of normalization-based TTA approaches, especially under significant domain shift. In this paper, we introduce a novel paradigm based on the concept of a Buffer Layer, which addresses the fundamental limitations of normalization layer updates. Unlike existing methods that modify the core parameters of the model, our approach preserves the integrity of the pre-trained backbone, inherently mitigating the risk of catastrophic forgetting during online adaptation. Through comprehensive experimentation, we demonstrate that our approach not only outperforms traditional methods in mitigating domain shift and enhancing model robustness, but also exhibits strong resilience to forgetting. Furthermore, our Buffer Layer is modular and can be seamlessly integrated into nearly all existing TTA frameworks, resulting in consistent performance improvements across various architectures. These findings validate the effectiveness and versatility of the proposed solution in real-world domain adaptation scenarios. The code is available at \url{https://github.com/anonymous}.



Paperid:720
Authors:Samuel Stocksieker, Denys Pommeret
Abstract:
Traditional correlation measures, such as Pearson’s and Spearman’s coefficients, are limited in their ability to capture complex relationships, particularly nonlinear and multivariate dependencies. The Hirschfeld–Gebelein–Rényi (HGR) maximal correlation offers a powerful alternative by seeking the highest Pearson correlation attainable through nonlinear transformations of two random variables. However, estimating the HGR remains challenging due to the complexity of optimizing arbitrary nonlinear functions. We introduce a new coefficient inspired by the HGR but grounded in the Spearman rank correlation, which we call the Spearman HGR (SHGR). We propose a neural‐network–based estimator tailored to (i) bivariate correlation matrices, (ii) multivariate correlations between a set of predictors and a target, and (iii) full correlations between two sets of variables. The SHGR satisfies Rényi's axioms, effectively detects nonlinear dependencies, and demonstrates robustness to noise, outliers, and spurious correlations. Additionally, it achieves competitive computational efficiency through tailored neural architectures. Numerical experiments and feature selection tasks confirm that the SHGR outperforms existing state-of-the-art methods.



Paperid:721
Authors:Ziyad Sheebaelhamd, Michael Tschannen, Michael Muehlebach, Claire Vernade
Abstract:
Current transformer-based imitation learning approaches introduce discrete action representations and train an autoregressive transformer decoder on the resulting latent code. However, the initial quantization breaks the continuous structure of the action space thereby limiting the capabilities of the generative model. We propose a quantization-free method instead that leverages Generative Infinite-Vocabulary Transformers (GIVT) as a direct, continuous policy parametrization for autoregressive transformers. This simplifies the imitation learning pipeline while achieving state-of-the-art performance on a variety of popular simulated robotics tasks. We enhance our policy roll-outs by carefully studying sampling algorithms, further improving the results.



Paperid:722
Authors:Flavio Martinelli, Alexander van Meegen, Berfin Simsek, Wulfram Gerstner, Johanni Brea
Abstract:
Abstract:The loss landscapes of neural networks contain minima and saddle points that may be connected in flat regions or appear in isolation. We identify and characterize a special structure in the loss landscape: channels along which the loss decreases extremely slowly, while the output weights of at least two neurons, $a_i$ and $a_j$, diverge to infinity, and their input weight vectors, $\mathbf{w_i}$ and $\mathbf{w_j}$, become equal to each other. At convergence, the two neurons implement a gated linear unit: $a_i\sigma(\mathbf{w_i} \cdot \mathbf{x}) + a_j\sigma(\mathbf{w_j} \cdot \mathbf{x}) \rightarrow c \sigma(\mathbf{w} \cdot \mathbf{x}) + (\mathbf{v} \cdot \mathbf{x}) \sigma'(\mathbf{w} \cdot \mathbf{x})$. Geometrically, these channels to infinity are asymptotically parallel to symmetry-induced lines of critical points. Gradient flow solvers, and related optimization methods like SGD or ADAM, reach the channels with high probability in diverse regression settings, but without careful inspection they look like flat local minima with finite parameter values. Our characterization provides a comprehensive picture of this quasi-flat region in terms of gradient dynamics, geometry, and functional interpretation. The emergence of gated linear units at the end of the channels highlights a surprising aspect of the computational capabilities of fully connected layers.



Paperid:723
Authors:Juliusz Ziomek, George Whittle, Michael A Osborne
Abstract:
In spite of their prevalence, the behaviour of Neural Networks when extrapolating far from the training distribution remains poorly understood, with existing results limited to specific cases. In this work, we prove general results---the first of their kind---by applying Neural Tangent Kernel (NTK) theory to analyse infinitely-wide neural networks trained until convergence and prove that the inclusion of just one Layer Norm (LN) fundamentally alters the induced NTK, transforming it into a bounded-variance kernel. As a result, the output of an infinitely wide network with at least one LN remains bounded, even on inputs far from the training data. In contrast, we show that a broad class of networks without LN can produce pathologically large outputs for certain inputs. We support these theoretical findings with empirical experiments on finite-width networks, demonstrating that while standard NNs often exhibit uncontrolled growth outside the training domain, a single LN layer effectively mitigates this instability. Finally, we explore real-world implications of this extrapolatory stability, including applications to predicting residue sizes in proteins larger than those seen during training and estimating age from facial images of underrepresented ethnicities absent from the training set.



Paperid:724
Authors:Siddharth Narayanan, James Braza, Ryan-Rhys Griffiths, Albert Bou, Geemi Wellawatte, Mayk Caldas Ramos, Ludovico Mitchener, Michael Pieler, Sam Rodriques, Andrew White
Abstract:
Reasoning models are large language models that use extra "thought tokens" before answering, providing both higher accuracy and explicit reasoning for their response. A major question has been whether language model reasoning generalizes beyond mathematics, programming, and logic, where most previous work has focused. We demonstrate that reasoning models can be post-trained in scientific domains without additional domain pretraining, and require substantially less data compared to contemporary domain-specific models.We report Science0-c, a 24B parameter LLM (based on Mistral-Small-24B) that can reason in natural language and respond with chemical structures. This reasoning model was trained with reinforcement learning on 577,790 experimentally-grounded chemistry tasks involving synthesized organic molecules. Our model outperforms all previous general-purpose chemistry models, frontier models, and humans, and is more data efficient relative to specialized models. We anticipate that this method can be applied to train highly data-efficient language models specialized for predictive and generative tasks across a wide variety of scientific domains.



Paperid:725
Authors:Ziang Zhou, Tianyuan Jin, Jieming Shi, Qing Li
Abstract:
Large Language Models (LLMs) exhibit impressive performance across diverse domains but often suffer from overconfidence, limiting their reliability in critical applications. We propose SteerConf, a novel framework that systematically steers LLMs' confidence scores to improve their calibration and reliability. SteerConf introduces three key components: (1) a steering prompt strategy that guides LLMs to produce confidence scores in specified directions (e.g., conservative or optimistic) by leveraging prompts with varying steering levels; (2) a steered confidence consistency measure that quantifies alignment across multiple steered confidences to enhance calibration; and (3) a steered confidence calibration method that aggregates confidence scores using consistency measures and applies linear quantization for answer selection. SteerConf operates without additional training or fine-tuning, making it broadly applicable to existing LLMs. Experiments on seven benchmarks spanning professional knowledge, common sense, ethics, and reasoning tasks, using advanced LLM models (GPT-3.5, LLaMA 3, GPT-4), demonstrate that SteerConf significantly outperforms existing methods, often by a significant margin. Our findings highlight the potential of steering the confidence of LLMs to enhance their reliability for safer deployment in real-world applications. The implementation is at \url{https://anonymous.4open.science/r/SteerConf-00FB/}.



Paperid:726
Authors:Rana Shahout, Cong Liang, Shiji Xin, Qianru Lao, Yong Cui, Minlan Yu, Michael Mitzenmacher
Abstract:
Augmented Large Language Models (LLMs) enhance standalone LLMs by integrating external data sources through API calls. In interactive applications, efficient scheduling is crucial for maintaining low request completion times, directly impacting user engagement. However, these augmentations introduce new scheduling challenges: the size of augmented requests (in tokens) no longer correlates proportionally with execution time, making traditional size-based scheduling algorithms like Shortest Job First less effective. Additionally, requests may require different handling during API calls, which must be incorporated into scheduling.This paper presents MARS, a novel inference framework that optimizes augmented LLM latency by explicitly incorporating system- and application-level considerations into scheduling. MARS introduces a predictive, memory-aware scheduling approach that integrates API handling and request prioritization to minimize completion time. We implement MARS on top of vLLM and evaluate its performance against baseline LLM inference systems, demonstrating improvements in end-to-end latency by 27%-85% and reductions in TTFT by 4%-96% compared to the existing augmented-LLM system, with even greater gains over vLLM. Our implementation is available online.



Paperid:727
Authors:Zhen Liu, Tim Xiao, Carles Domingo i Enrich, Weiyang Liu, Dinghuai Zhang
Abstract:
While methods exist for aligning flow matching models -- a popular and effective class of generative models -- with human preferences, existing approaches fail to achieve both adaptation efficiency and probabilistically sound prior preservation. In this work, we leverage the theory of optimal control and propose VGG-Flow, a gradient matching–based method for finetuning pretrained flow matching models. The key idea in this algorithm is that the optimal difference between the finetuned velocity field and the pretrained one should be matched with the gradient field of a value function. This method not only incorporates first-order information from the reward model but also benefits from heuristic initialization of the value function to enable fast adaptation. Empirically, we show on a popular text-to-image flow matching model, Stable Diffusion 3, that our method can finetune flow matching models under limited computational budgets while achieving effective and prior-preserving alignment.



Paperid:728
Authors:Sheikh Md Shakeel Hassan, Xianwei Zou, Akash Dhruv, Aparna Chandramowlishwaran
Abstract:
Modeling boiling--an inherently chaotic, multiphase process central to energy and thermal systems--remains a significant challenge for neural PDE surrogates. Existing models rely on simulation inputs (e.g., bubble positions), limiting their ability to forecast future states autonomously. We introduce Bubbleformer, a transformer-based spatiotemporal model that forecasts stable and long-range boiling dynamics including nucleation, interface evolution, and heat transfer without dependence on simulation data during inference. Bubbleformer combines axial attention with physics-informed architectural modifications to mitigate spectral bias and conditions on thermophysical parameters to generalize across fluids, geometries, and operating conditions. To evaluate models of these chaotic systems, we introduce physically grounded metrics. We also release BubbleML 2.0, a high-fidelity dataset spanning diverse working fluids (cryogens, refrigerants, and dielectrics), boiling configurations (pool boiling and flow boiling), distinct flow boiling regimes (bubbly, slug, and annular), and operational and boundary conditions. Bubbleformer sets new benchmarks in both prediction and forecasting tasks of two-phase boiling flows.



Paperid:729
Authors:Shenghe Zheng, Qianjia Cheng, Junchi Yao, Mengsong Wu, haonan he, Ning Ding, Yu Cheng, Shuyue Hu, LEI BAI, Dongzhan Zhou, Ganqu Cui, Peng Ye
Abstract:
Large Language Models (LLMs) have achieved remarkable progress on advanced reasoning tasks such as mathematics and coding competitions. Meanwhile physics, despite being both reasoning-intensive and essential to real-world understanding, received limited academic and industrial attention. This paper introduces PHYSICS, a dataset containing 16,568 high-quality physics problems spanning subjects and difficulty levels, to facilitate this issue. Specifically, PHYSICS is curated with exercises from over 100 textbooks through a carefully designed pipeline for quality control. It covers five major physics domains: Mechanics, Electromagnetism, Thermodynamics, Optics, and Modern Physics. It also spans a wide range of difficulty levels, from high school to graduate-level physics courses. To utilize the data for improving and evaluating the model's physical reasoning capabilities, we split the dataset into training and test sets, and provide reasoning paths generated by powerful reasoning models for the training data to facilitate model training. In addition, for the evaluation part, we find that existing evaluation frameworks exhibit biases in aspects such as units, simplification, and precision in physics domain. To balance efficiency and accuracy, we introduce a Rule+Model evaluation framework tailored to physics problems. Our evaluations on current state-of-the-art open-source and proprietary models highlight the limitations of current models in handling physics-related tasks. We hope that our dataset and evaluation methodology will jointly advance the development of LLMs in the field of physics.



Paperid:730
Authors:Yutong Bai, Danny Tran, Amir Bar, Trevor Darrell, Yann LeCun, Jitendra Malik
Abstract:
We train models to predict ego-centric video from human actions (PEVA), given the past video and an action represented by the relative 3D body pose. By conditioning on kinematic pose trajectories, structured by the joint hierarchy of the body, our model learns to simulate how physical human actions shape the environment from a first-person point of view. We train an auto-regressive conditional diffusion transformer on Nymeria, a large-scale dataset of real-world egocentric video and body pose capture. We further design a hierarchical evaluation protocol with increasingly challenging tasks, enabling a comprehensive analysis of the model’s embodied prediction and control abilities. Our work represents an initial attempt to tackle the challenges of modeling complex real-world environments and embodied agent behaviors with video prediction from the perspective of a human.



Paperid:731
Authors:Georgy Noarov, Soham Mallick, Tao Wang, Sunay Joshi, Yan Sun, Yangxinyu Xie, Mengxin Yu, Edgar Dobriban
Abstract:
Abstract:Top-$k$ decoding is a widely used method for sampling from LLMs: at each token, only the largest $k$ next-token-probabilities are kept, and the next token is sampled after re-normalizing them to sum to unity. Top-$k$---as well as other---sampling methods are motivated by the intuition that true next-token distributions are sparse, and the noisy LLM estimates need to be truncated. However, to our knowledge, a precise theoretical motivation for the use of top-$k$ decoding is missing. In this work, we develop a theoretical framework that both explains and generalizes top-$k$ decoding. We view decoding for a fixed token as the estimation of a sparse probability distribution. We consider estimators obtained by minimizing a separable Bregman divergence with a sparsity-inducing $\ell_0$ regularization. Despite the combinatorial nature of the objective, we show how to optimize it efficiently for a large class of divergences (for both \emph{primal} and \emph{dual} minimization). We show that the optimal decoding strategies are greedy, and further that the loss function is discretely convex in $k$, so that binary search provably and efficiently finds the optimal $k$. We show that top-$k$ decoding arises as a special case for the KL divergence, and identify new decoding strategies that have distinct behavior (e.g., non-linearly up-weighting larger probabilities after re-normalization). Our work thus facilitates understanding and generalizing the widely popular top-$k$ decoding method for LLMs.



Paperid:732
Authors:Vedang Lad, Jin Hwa Lee, Wes Gurnee, Max Tegmark
Abstract:
We investigate the robustness of Large Language Models (LLMs) to structural interventions by deleting and swapping adjacent layers during inference. Surprisingly, models retain 72–95\% of their original top-1 prediction accuracy without any fine-tuning. We find that performance degradation is not uniform across layers: interventions to the early and final layers cause the most degradation, while the model is remarkably robust to dropping middle layers. This pattern of localized sensitivity motivates our hypothesis of four stages of inference, observed across diverse model families and sizes: (1) detokenization, where local context is integrated to lift raw token embeddings into higher-level representations; (2) feature engineering, where task- and entity-specific features are iteratively refined; (3) prediction ensembling, where hidden states are aggregated into plausible next-token predictions; and (4) residual sharpening, where irrelevant features are suppressed to finalize the output distribution. Synthesizing behavioral and mechanistic evidence, we provide a framework for interpreting depth-dependent computations in LLMs.



Paperid:733
Authors:Alex Nguyen, David Schwab, Vudtiwat Ngampruetikorn
Abstract:
Machine learning models are often brittle under distribution shift, i.e., when data distributions at test time differ from those during training. Understanding this failure mode is central to identifying and mitigating safety risks of mass adoption of machine learning. Here we analyze ridge regression under concept shift—a form of distribution shift in which the input-label relationship changes at test time. We derive an exact expression for prediction risk in the thermodynamic limit. Our results reveal nontrivial effects of concept shift on generalization performance, including a phase transition between weak and strong concept shift regimes and nonmonotonic data dependence of test performance even when double descent is absent. Our theoretical results are in good agreement with experiments based on transformers pretrained to solve linear regression; under concept shift, too long context length can be detrimental to generalization performance of next token prediction. Finally, our experiments on MNIST and FashionMNIST suggest that this intriguing behavior is present also in classification problems.



Paperid:734
Authors:Zihao Cheng, Binrui Wu, Zhiwei Li, Yuesen Liao, Su Zhao, Shuai Chen, Yuan Gao, Weizhong Zhang
Abstract:
Dynamic coreset selection is a promising approach for improving the training efficiency of deep neural networks by periodically selecting a small subset of the most representative or informative samples, thereby avoiding the need to train on the entire dataset. However, it remains inherently challenging due not only to the complex interdependencies among samples and the evolving nature of model training, but also to a criticalcoreset representativeness degradation issueidentified and explored in-depth in this paper, that is, the representativeness or information content of the coreset degrades over time as training progresses. Therefore, we argue that, in addition to designing accurate selection rules, it is equally important to endow the algorithms with the ability to assess the quality of the current coreset. Such awareness enables timely re-selection, mitigating the risk of overfitting to stale subsets—a limitation often overlooked by existing methods. To this end, this paper proposes anEfficientRepresentativeness-AwareCoresetSelection method for deep neural networks, a lightweight framework that enables dynamic tracking and maintenance of coreset quality during training. While the ideal criterion—gradient discrepancy between the coreset and the full dataset—is computationally prohibitive, we introduce a scalable surrogate based on the signal-to-noise ratio (SNR) of gradients within the coreset, which is the main technical contribution of this paper and is also supported by our theoretical analysis. Intuitively, a decline in SNR indicates overfitting to the subset and declining representativeness. Leveraging this observation, our method triggers coreset updates without requiring costly Hessian or full-batch gradient computations, maintaining minimal computational overhead. Experiments on multiple datasets confirm the effectiveness of our approach. Notably, compared with existing gradient-based dynamic coreset selection baselines, our method achieves up to a 5.4\% improvement in test accuracy across multiple datasets.



Paperid:735
Authors:Thomas Kwa, Ben West, Joel Becker, Amy Deng, Katharyn Garcia, Max Hasin, Sami Jawhar, Megan Kinniment, Nate Rush, Sydney Von Arx, Ryan Bloom, Thomas Broadley, Haoxing Du, Brian Goodrich, Nikola Jurkovic, Luke Miles, Seraphina Nix, Tao Lin, Neev Parikh, David Rein, Lucas Jun Koba Sato, Hjalmar Wijk, Daniel Ziegler, Elizabeth Barnes, Lawrence Chan
Abstract:
Despite rapid progress on AI benchmarks, the real-world meaning of benchmark performance remains unclear. To quantify the capabilities of AI systems in terms of human capabilities, we propose a new metric: 50%-task-completion time horizon. This is the time humans typically take to complete tasks that AI models can complete with 50% success rate. We first timed humans with relevant domain expertise on a combination of RE-Bench, HCAST, and 66 novel shorter tasks. On these tasks, current frontier AI models such as o3 have a 50% time horizon of around 110 minutes. Furthermore, frontier AI time horizon has been doubling approximately every seven months since 2019, though the trend may have accelerated since 2024. The increase in AI models’ time horizons seems to be primarily driven by greater reliability and ability to adapt to mistakes, combined with better logical reasoning and tool use capabilities. We discuss the limitations of our results—including their degree of external validity—and the implications of increased autonomy for dangerous capabilities. If these results generalize to real-world software tasks, extrapolation of this trend predicts that within 5 years, AI systems will be capable of automating many software tasks that currently take humans a month.



Paperid:736
Authors:Anirudh Chakravarthy, Shuai Zheng, Xin Huang, Sachithra Hemachandra, Xiao Zhang, Yuning Chai, Zhao Chen
Abstract:
Fine-tuning pre-trained models has become invaluable in computer vision and robotics. Recent fine-tuning approaches focus on improving efficiency rather than accuracy by using a mixture of smaller learning rates or frozen backbones. To return the spotlight to model accuracy, we present PROFIT (Proximally Restricted Optimizer For Iterative Training), one of the first optimizers specifically designed for incrementally fine-tuning converged models on new tasks or datasets. Unlike traditional optimizers such as SGD or Adam, which make minimal assumptions due to random initialization, PROFIT leverages the structure of a converged model to regularize the optimization process, leading to improved results. By employing a simple temporal gradient orthogonalization process, PROFIT outperforms traditional fine-tuning methods across various tasks: image classification, representation learning, and large-scale motion prediction. Moreover, PROFIT is encapsulated within the optimizer logic, making it easily integrated into any training pipeline with minimal engineering effort. A new class of fine-tuning optimizers like PROFIT can drive advancements as fine-tuning and incremental training become increasingly prevalent, reducing reliance on costly model training from scratch.



Paperid:737
Authors:Badih Ghazi, Pritish Kamath, Alexander Knop, Ravi Kumar, Pasin Manurangsi, Chiyuan Zhang
Abstract:
The conventional approach in differential privacy (DP) literature formulates the privacy-utility tradeoff with a "privacy-first" perspective: for a predetermined level of privacy, a certain utility is achievable. However, practitioners often operate under a "utility-first" paradigm, prioritizing a desired level of utility and then determining the corresponding privacy cost. Wu et al. [2019] initiated a formal study of this ``utility-first'' perspective by introducing ex-post DP. They demonstrated that by adding correlated Laplace noise and progressively reducing it on demand, a sequence of increasinglyaccurate estimates of a private parameter can be generated, with the privacy cost attributed only to the least noisy iterate released. This led to a Laplace mechanism variant that achieves a specified utility with minimal privacy loss.However, their work, and similar findings by Whitehouse et al. [2023], are primarily limited to simple mechanisms basedon Laplace or Gaussian noise. In this paper, we significantly generalize these results. In particular, we extend the findings of Wu et al. [2019] and Liu and Talwar [2019] to support any sequence of private estimators, incurring at most a doubling of the original privacy budget. Furthermore, we demonstrate that hyperparameter tuning for these estimators, including the selection of an optimal privacy budget, can be performed without additional privacy cost. Finally, we extend our results to ex-post R\'{e}nyi DP, further broadening the applicability of utility-first privacy mechanisms.



Paperid:738
Authors:Lee Cohen, Yishay Mansour, Shay Moran, Han Shao
Abstract:
PrecisionandRecallare fundamental metrics in machine learning tasks where both accurate predictions and comprehensive coverage are essential, such as in multi-label learning, language generation, medical studies, and recommender systems. A key challenge in these settings is the prevalence of one-sided feedback, where only positive examples are observed during training—e.g., in multi-label tasks like tagging people in Facebook photos, we may observe only a few tagged individuals, without knowing who else appears in the image. To address learning under such partial feedback, we introduce a Probably Approximately Correct (PAC) framework in which hypotheses are set functions that map each input to a set of labels, extending beyond single-label predictions and generalizing classical binary, multi-class, and multi-label models. Our results reveal sharp statistical and algorithmic separations from standard settings: classical methods such as Empirical Risk Minimization provably fail, even for simple hypothesis classes. We develop new algorithms that learn from positive data alone, achieving optimal sample complexity in the realizable case, and establishing multiplicative—rather than additive—approximation guarantees in the agnostic case, where achieving additive regret is impossible.



Paperid:739
Authors:Thodoris Lykouris, Sloan Nietert, Princewill Okoroafor, Chara Podimata, Julian Zimmert
Abstract:
Abstract:We initiate the study of contextual dynamic pricing with a heterogeneous population of buyers, where a seller repeatedly (over $T$ rounds) posts prices that depend on the observable $d$ dimensional context and receives binary purchase feedback. Unlike prior work assuming homogeneous buyer types, in our setting the buyer's valuation type is drawn from an unknown distribution with finite support $K_{\star}$. We develop a contextual pricing algorithm based on Optimistic Posterior Sampling with regret $\tilde{O}(K_{\star}\sqrt{dT})$, which we prove to be tight in $d, T$ up to logarithmic terms. Finally, we refine our analysis for the non-contextual pricing case, proposing a variance-aware Zooming algorithm that achieves the optimal dependence on $K_{\star}$.



Paperid:740
Authors:Lyuzhou Chen, Yijia Sun, Yanze Gao, Xiangyu Wang, Derui Lyu, Taiyu Ban, Xin Wang, Xiren Zhou, Huanhuan Chen
Abstract:
Learning the causality between variables, known as DAG structure learning, is critical yet challenging due to issues such as insufficient data and noise. While prior knowledge can improve the learning process and refine the DAG structure, incorporating prior knowledge is not without pitfalls. In particular, we find that the gap between the imprecise prior knowledge and the exact weights modeled by existing methods may result in deviation in edge weights. Such deviation can subsequently cause significant inaccuracies when learning the DAG structure. This paper addresses this challenge by providing a theoretical analysis of the impact of deviation in edge weights during the optimization process of structure learning. We identify two special graph patterns that arise due to the deviation and show that their occurrence increases as the degree of deviation grows. Building on this analysis, we propose the Pattern-Guided Adaptive Prior (PGAP) framework. PGAP detects these patterns as structural signals during optimization and adaptively adjusts the structure learning process to counteract the identified weight deviation, thereby improving the integration of prior knowledge. Experiments verify the effectiveness and robustness of the proposed method.



Paperid:741
Authors:Dorian Baudry, Emmeran Johnson, Simon Vary, Ciara Pike-Burke, Patrick Rebeschini
Abstract:
Abstract:Recent works of Mei et al. (2023, 2024) have deepened the theoretical understanding of the *Stochastic Gradient Bandit* (SGB) policy, showing that using a constant learning rate guarantees asymptotic convergence to the optimal policy, and that sufficiently *small* learning rates can yield logarithmic regret. However, whether logarithmic regret holds beyond small learning rates remains unclear. In this work, we take a step towards characterizing the regret *regimes* of SGB as a function of its learning rate. For two--armed bandits, we identify a sharp threshold, scaling with the sub-optimality gap $\Delta$, below which SGB achieves *logarithmic* regret on all instances, and above which it can incur *polynomial* regret on some instances. This result highlights the necessity of knowing (or estimating) $\Delta$ to ensure logarithmic regret with a constant learning rate.For general $K$-armed bandits, we further show the learning rate must scale inversely with $K$ to avoid polynomial regret. We introduce novel techniques to derive regret upper bounds for SGB, laying the groundwork for future advances in the theory of gradient-based bandit algorithms.



Paperid:742
Authors:Zhihao Li, Jiale Cai, Gezheng Xu, Hao Zheng, Qiuyue Li, Fan Zhou, Shichun Yang, Charles Ling, Boyu Wang
Abstract:
Abstract:The rapid growth of publicly available data has fueled deep learning advancements but also raises concerns about unauthorized data usage. Unlearnable Examples (UEs) have emerged as a data protection strategy that introduces imperceptible perturbations to prevent unauthorized learning. However, most existing UE methods produce perturbations strongly tied to specific training sets, leading to a significant drop in unlearnability when applied to unseen data or tasks. In this paper, we argue that for broad applicability, UEs should maintain their effectiveness across diverse application scenarios. To this end, we conduct the first comprehensive study on the transferability of UEs across diverse and practical yet demanding settings. Specifically, we identify key scenarios that pose significant challenges for existing UE methods, including varying styles, out-of-distribution classes, resolutions, and architectures. Moreover, we propose $\textbf{Versatile Transferable Generator}$ (VTG), a transferable generator designed to safeguard data across various conditions. Specifically, VTG integrates adversarial domain augmentation into the generator’s training process to synthesize out-of-distribution samples, thereby improving its generalizability to unseen scenarios. Furthermore, we propose a Perturbation-Label Coupling mechanism that leverages contrastive learning to directly align perturbations with class labels. This approach reduces the generator’s reliance on data semantics, allowing VTG to produce unlearnable perturbations in a distribution-agnostic manner. Extensive experiments demonstrate the effectiveness and broad applicability of our approach.



Paperid:743
Authors:Yuetong Fang, Deming Zhou, Ziqing Wang, Hongwei Ren, zeng zecui, Lusong Li, shibo zhou, Renjing Xu
Abstract:
Spiking Transformers offer an energy-efficient alternative to conventional deep learning by transmitting information solely through binary (0/1) spikes. However, there remains a substantial performance gap compared to artificial neural networks. A common belief is that their binary and sparse activation transmission leads to information loss, thus degrading feature representation and accuracy. In this work, however, we reveal for the first time that spiking neurons preferentially propagate low-frequency information. We hypothesize that the rapid dissipation of high-frequency components is the primary cause of performance degradation. For example, on Cifar-100, adopting Avg-Pooling (low-pass) for token mixing lowers performance to 76.73\%; interestingly, replacing it with Max-Pooling (high-pass) pushes the top-1 accuracy to 79.12\%, surpassing the well-tuned Spikformer baseline by 0.97\%. Accordingly, we introduce Max-Former that restores high-frequency signals through two frequency-enhancing operators: extra Max-Pooling in patch embedding and Depth-Wise Convolution in place of self-attention. Notably, our Max-Former (63.99 M) hits the top-1 accuracy of 82.39\% on ImageNet, showing a +7.58\% improvement over Spikformer with comparable model size (74.81\%, 66.34 M). We hope this simple yet effective solution inspires future research to explore the distinctive nature of spiking neural networks, beyond standard deep learning.



Paperid:744
Authors:Zhengxuan Wu, Qinan Yu, Aryaman Arora, Christopher D Manning, Christopher Potts
Abstract:
Steering methods for language models (LMs) seek to provide fine-grained and interpretable control over model generations by variously changing model inputs, weights, or representations to adjust behavior. Recent work has shown that adjusting weights or representations is often less effective than steering by prompting, for instance when wanting to introduce or suppress a particular concept. We demonstrate how to improve representation steering via our new Reference-free Preference Steering (RePS), a bidirectional preference-optimization objective that jointly does concept steering and suppression. We train three parameterizations of RePS and evaluate them on AxBench, a large-scale model steering benchmark. On Gemma models with sizes ranging from 2B to 27B, RePS outperforms all existing steering methods trained with a language modeling objective and substantially narrows the gap with prompting -- while promoting interpretability and minimizing parameter count. In suppression, RePS matches the language-modeling objective on Gemma-2 and outperforms it on the larger Gemma-3 variants while remaining resilient to prompt-based jailbreaking attacks that defeat prompting. Overall, our results suggest that RePS provides an interpretable and robust alternative to prompting for both steering and suppression.



Paperid:745
Authors:Jiayi Sheng, Luna Lyu, Jikai Jin, Tanglin Xia, Alex Gu, James Zou, Pan Lu
Abstract:
Inequality proving, crucial across diverse scientific and mathematical fields, tests advanced reasoning skills such as discovering tight bounds and strategic theorem application. This makes it a distinct, demanding frontier for large language models (LLMs), offering insights beyond general mathematical problem-solving. Progress in this area is hampered by existing datasets that are often scarce, synthetic, or rigidly formal. We address this by proposing aninformal yet verifiabletask formulation, recasting inequality proving into two automatically checkable subtasks: bound estimation and relation prediction. Building on this, we releaseIneqMath, an expert-curated dataset of Olympiad-level inequalities, including a test set and training corpus enriched with step-wise solutions and theorem annotations. We also develop a novel LLM-as-judge evaluation suite, combining afinal-answerjudge with four specializedstep-wisejudges designed to detect common reasoning flaws. A systematic evaluation of 29 leading LLMs onIneqMathreveals a surprising reality: even top models like o1 achieve less than 10% overall accuracy under step-wise scrutiny; this is a drop of up to 65.5% from their accuracy considering only final answer equivalence. This discrepancy exposes fragile deductive chains and a critical gap for current LLMs between merely finding an answer and constructing a rigorous proof. Scaling model size and increasing test-time computation yield limited gains in overall proof correctness. Instead, our findings highlight promising research directions such as theorem-guided reasoning and self-refinement.



Paperid:746
Authors:Zichun Yu, Fei Peng, Jie Lei, Arnold Overwijk, Scott Yih, Chenyan Xiong
Abstract:
In this paper, we introduceGroup-MATES, an efficient group-level data selection approach to optimize the speed-quality frontier of language model pretraining. Specifically, Group-MATES parameterizes costly group-level selection with a relational data influence model. To train this model, we sample training trajectories of the language model and collect oracle data influences alongside. The relational data influence model approximates the oracle data influence by weighting individual influence with relationships among training data. To enable efficient selection with our relational data influence model, we partition the dataset into small clusters using relationship weights and select data within each cluster independently. Experiments on DCLM 400M-4x, 1B-1x, and 3B-1x show that Group-MATES achieves 3.5\%-9.4\% relative performance gains over random selection across 22 downstream tasks, nearly doubling the improvements achieved by state-of-the-art individual data selection baselines. Furthermore, Group-MATES reduces the number of tokens required to reach a certain downstream performance by up to 1.75x, substantially elevating the speed-quality frontier. Further analyses highlight the critical role of relationship weights in the relational data influence model and the effectiveness of our cluster-based inference. We will open-source our code via GitHub.



Paperid:747
Authors:Anaïs Després, Mario Morawski, Remi Rehm
Abstract:
Abstract:Sequential data—ranging from financial time series to natural language—has driven the growing adoption of autoregressive models. However, these algorithms rely on the presence of underlying patterns in the data, and their identification often depends heavily on human expertise. Misinterpreting these patterns can lead to model misspecification, resulting in increased generalization error and degraded performance. The recently proposed \texttt{evolving pattern (EvoRate)} metric addresses this by using the mutual information between the next data point and its past to guide regression order estimation and feature selection. Building on this idea, we introduce a general framework based on predictive information—the mutual information between the past and the future, $\mathbf{I}(X_{\text{past}}; X_{\text{future}})$. This quantity naturally defines an information-theoretic learning curve, which quantifies the amount of predictive information available as the observation window grows. Using this formalism, we show that the presence or absence of temporal patterns fundamentally constrains the learnability of sequential models: even an optimal predictor cannot outperform the intrinsic information limit imposed by the data. We validate our framework through experiments on synthetic data, demonstrating its ability to assess model adequacy, quantify the inherent complexity of a dataset, and reveal interpretable structure in sequential data.



Paperid:748
Authors:Amin Heyrani Nobari, Kaveh Alimohammadi, Ali ArjomandBigdeli, Akash Srivastava, Faez Ahmed, Navid Azizan
Abstract:
Model merging, a method that combines the parameters and embeddings of multiple fine-tuned large language models (LLMs), offers a promising approach to enhance model performance across various tasks while maintaining computational efficiency. This paper introduces Activation-Informed Merging (AIM), a technique that integrates the information from the activation space of LLMs into the merging process to improve performance and robustness. AIM is designed as a flexible, complementary solution that is applicable to any existing merging method. It aims to preserve critical weights from the base model, drawing on principles from continual learning~(CL) and model compression. Utilizing a task-agnostic calibration set, AIM selectively prioritizes essential weights during merging. We empirically demonstrate that AIM significantly enhances the performance of merged models across multiple benchmarks. Our findings suggest that considering the activation-space information can provide substantial advancements in the model merging strategies for LLMs with up to 40\% increase in benchmark performance.



Paperid:749
Authors:Gabriel Sarch, Snigdha Saha, Naitik Khandelwal, Ayush Jain, Michael Tarr, Aviral Kumar, Katerina Fragkiadaki
Abstract:
While reinforcement learning (RL) over chains of thought has significantly advanced language models in tasks such as mathematics and coding, visual reasoning introduces added complexity by requiring models to direct visual attention, interpret perceptual inputs, and ground abstract reasoning in spatial evidence. We introduce ViGoRL (Visually Grounded Reinforcement Learning), a vision-language model trained with RL to explicitly anchor each reasoning step to specific visual coordinates. Inspired by human visual decision-making, ViGoRL learns to produce spatially grounded reasoning traces, guiding visual attention to task-relevant regions at each step. Across a diverse set of visual reasoning benchmarks—including SAT-2 and BLINK for spatial reasoning, and ScreenSpot and VisualWebArena for web-based grounding—ViGoRL consistently outperforms both supervised fine-tuning and conventional RL baselines that lack explicit grounding mechanisms. Incorporating multi-turn RL with visual feedback further improves ViGoRL’s performance on localizing small elements. Additionally, we find that grounding amplifies other visual behaviors such as region exploration, visual subgoal setting, and verification. Finally, human evaluations show that the model’s visual references are not only spatially accurate but also helpful for understanding model reasoning steps. Our results show that visually grounded RL is a strong paradigm for imbuing models with general-purpose visual reasoning.



Paperid:750
Authors:Arnab Bhattacharyya, Sutanu Gayen, Philips George John, Sayantan Sen, N. V. Vinodchandran
Abstract:
Abstract:This work establishes a novel link between the problem of PAC-learning high-dimensional graphical models and the task of (efficient) counting and sampling of graph structures, using an online learning framework. The problem of efficiently counting and sampling graphical structures, such as spanning trees and acyclic orientations, has been a vibrant area of research in algorithms. We show that this rich algorithmic foundation can be leveraged to develop new algorithms for learning high-dimensional graphical models.We present the first efficient algorithm for (both realizable and agnostic) learning of Bayes nets with a chordal skeleton. In particular, we present an algorithm that, given integers $k,d > 0$, error parameter $\varepsilon > 0$, an undirected chordal graph $G$ on $n$ vertices, and sample access to a distribution $P^*$ on $[k]^n$; (1) returns a Bayes net $\widehat{P}$ with skeleton $G$ and indegree $d$, whose KL-divergence from $P^*$ is at most $\varepsilon$ more than the optimal KL-divergence between $P^*$ and any Bayes net with skeleton $G$ and indegree $d$, (2) uses $\widetilde{O}(n^3k^{d+1}/\varepsilon^2)$ samples from $P^*$ and runs in time $\mathrm{poly}(n,k,\varepsilon^{-1})$ for constant $d$. Prior results in this spirit were for only for \adds{trees ($d=1$, tree skeleton) via Chow-Liu}, and in the realizable setting for polytrees (arbitrary $d$ but tree skeleton). Thus, our result significantly extends the state-of-the-art in learning Bayes net distributions. We also establish new results for learning tree and polytree distributions.



Paperid:751
Authors:Lu Yi, Runlin Lei, Fengran Mo, Yanping Zheng, Zhewei Wei, Yuhang Ye
Abstract:
Future link prediction on temporal graphs is a fundamental task with wide applicability in real-world dynamic systems. These scenarios often involve both recurring (seen) and novel (unseen) interactions, requiring models to generalize effectively across both types of edges. However, existing methods typically rely on complex memory and aggregation modules, yet struggle to handle unseen edges. In this paper, we revisit the architecture of existing temporal graph models and identify two essential but overlooked modeling requirements for future link prediction: representing nodes with unique identifiers and performing target-aware matching between source and destination nodes. To this end, we propose Cross-Attention based Future Link Predictor on Temporal Graphs (CRAFT), a simple yet effective architecture that discards memory and aggregation modules and instead builds on two components: learnable node embeddings and cross-attention between the destination and the source's recent interactions. This design provides strong expressive power and enables target-aware modeling of the compatibility between candidate destinations and the source's interaction patterns. Extensive experiments on diverse datasets demonstrate that CRAFT consistently achieves superior performance with high efficiency, making it well-suited for large-scale real-world applications.



Paperid:752
Authors:Ernst Röell, Bastian Rieck
Abstract:
Point-cloud synthesis, i.e. the generation of novel point clouds from an input distribution, remains a challenging task, for which numerous complex machine-learning models have been devised. We develop a novel method that encodes geometrical-topological characteristics of point clouds using inner products, leading to a highly-efficient point cloud representation with provable expressivity properties. Integrated into deep learning models, our encoding exhibits high quality in typical tasks like reconstruction, generation, and interpolation, with inference times orders of magnitude faster than existing methods.



Paperid:753
Authors:Ziteng Gao, Jay Zhangjie Wu, Mike Zheng Shou
Abstract:
Abstract:Natural images often exhibit underlying sparse structures, with information density varying significantly across different spatial locations. However, most generative models rely on dense grid-based pixels or latents, neglecting this inherent sparsity. In this paper, we explore modeling visual generation paradigm via sparse non-grid latent representations. Specifically, we design a sparse autoencoder that represent an image as a few number of latents along with their positional properties (i.e., regions of interest, RoIs) with high reconstruction quality. Then, we explore the flow matching approach jointly on modeling non-grid latents and RoIs with transformers. To the best knowledge, we are the first to address spatial sparsity using RoIs in diffusion process. Experimental results show that our sparse flow-based transformers have competitive performance compared with dense grid-based counterparts with significantly reduced lower compute, and reaches a competitive 2.76 FID with just 64 latents on class-conditional ImageNet $256\times 256$ generation.



Paperid:754
Authors:Chengchang Liu, Zongqi Wan, Institute of Computing Jialin Zhang, Institute of Computing Xiaoming Sun, John C. S. Lui
Abstract:
Abstract:This paper investigates convex-concave minimax optimization problems where only the function value access is allowed.We introduce a class of Hessian-aware quantum zeroth-order methods that can find the $\epsilon$-saddle point within $\tilde{\mathcal{O}}(d^{2/3}\epsilon^{-2/3})$ function value oracle calls.This represents an improvement of $d^{1/3}\epsilon^{-1/3}$ over the $\mathcal{O}(d\epsilon^{-1})$ upper bound of classical zeroth-order methods, where $d$ denotes the problem dimension.We also extend these results to $\mu$-strongly-convex $\mu$-strongly-concave minimax problems using a restart strategy, and show a speedup of $d^{1/3}\mu^{-1/3}$ compared to classical zeroth-order methods.The acceleration achieved by our methods stems from the construction of efficient quantum estimators for the Hessian, requiring only $\tilde{\mathcal{O}}(d)$ queries to function value oracle, and the subsequent design of efficient Hessian-aware algorithms.Furthermore, we show that such quantum estimators can also be applied to non-convex problems, leading to a reduction in the query complexity compared to classical methods.



Paperid:755
Authors:Feng Xiao, Xiaoying Tang, Jicong Fan
Abstract:
Unsupervised anomaly detection is a critical task in many high-social-impact applications such as finance, healthcare, social media, and cybersecurity, where demographics involving age, gender, race, disease, etc, are used frequently. In these scenarios, possible bias from anomaly detection systems can lead to unfair treatment for different groups and even exacerbate social bias. In this work, first, we thoroughly analyze the feasibility and necessary assumptions for ensuring group fairness in unsupervised anomaly detection. Second, we propose a novel fairness-aware anomaly detection method FairAD. From the normal training data, FairAD learns a projection to map data of different demographic groups to a common target distribution that is simple and compact, and hence provides a reliable base to estimate the density of the data. The density can be directly used to identify anomalies while the common target distribution ensures fairness between different groups. Furthermore, we propose a threshold-free fairness metric that provides a global view for model's fairness, eliminating dependence on manual threshold selection. Experiments on real-world benchmarks demonstrate that our method achieves an improved trade-off between detection accuracy and fairness under both balanced and skewed data across different groups.



Paperid:756
Authors:Tong Zhao, Jiacheng Li, Yuanchang Zhou, Guangming Tan, Weile Jia
Abstract:
Finding lower and better-generalized minima is crucial. However, most existing optimizers stop searching the parameter space once they reach a local minimum. Given the complex geometric properties of the loss landscape, it is difficult to guarantee that such a point is the lowest or provides the best generalization. To address this, we propose a novel optimization strategy: after reaching a local minimum, the optimizer inclines to continue exploring along valleys (areas with low gradient norms) in the landscape to search for other possible local minima. This approach increases the likelihood of finding a lower and flatter local minimum, which is often associated with better generalization. We analyze the relationship between the flat local minima and generalization and provide a proof of convergence for the proposed method ALTO, in both convex and non-convex scenarios. Finally, we demonstrate its effectiveness in a typical training scenario, large minibatch training. Our testing results show that the accuracy (generalization) of ALTO is 2.5% higher than the of the current state-of-the-art method when using a large batch size on a variety of tasks averagely. This work potentially opens a new research direction in the design of optimization algorithms.



Paperid:757
Authors:Zihan Weng, Lucas Gomez, Taylor Webb, Pouya Bashivan
Abstract:
Vision-Language Models (VLMs) have shown remarkable progress in visual understanding in recent years. Yet, they still lag behind human capabilities in specific visual tasks such as counting or relational reasoning. To understand the underlying limitations, we adopt methodologies from cognitive science, analyzing VLM performance along core cognitive axes: Perception, Attention, and Memory. Using a suite of tasks targeting these abilities, we evaluate state-of-the-art VLMs, including GPT-4o. Our analysis reveals distinct cognitive profiles: while advanced models approach ceiling performance on some tasks (e.g. category identification), a significant gap persists, particularly in tasks requiring spatial understanding or selective attention. Investigating the source of these failures and potential methods for improvement, we employ a vision-text decoupling analysis, finding that models struggling with direct visual reasoning show marked improvement when reasoning over their own generated text captions. These experiments reveal a strong need for improved VLM CoT abilities, even in models that consistently exceed human performance. Furthermore, we demonstrate the potential of targeted fine-tuning on composite visual reasoning tasks and show that fine-tuning smaller VLMs substantially improves core cognitive abilities. While this improvement does not translate to large enhancements on challenging, out-of-distribution benchmarks, we show broadly that VLM performance on our datasets strongly correlates with performance on these other benchmarks. Our work provides a detailed analysis of VLM cognitive strengths and weaknesses and identifies key bottlenecks in simultaneous perception and reasoning while also providing an effective and simple solution.



Paperid:758
Authors:Kevin Tirta Wijaya, Michael Sun, Minghao Guo, Hans-peter Seidel, Wojciech Matusik, Vahid Babaei
Abstract:
Accurate prediction of continuous properties is essential to many scientific and engineering tasks. Although deep-learning regressors excel with abundant labels, their accuracy deteriorates in data-scarce regimes. We introduce RankRefine, a model-agnostic, plug-and-play post-hoc refinement technique that injects expert knowledge through pairwise rankings. Given a query item and a small reference set with known properties, RankRefine combines the base regressor’s output with a rank-based estimate via inverse-variance weighting, requiring no retraining. In molecular property prediction task, RankRefine achieves up to 10\% relative reduction in mean absolute error using only 20 pairwise comparisons obtained through a general-purpose large language model (LLM) with no finetuning. As rankings provided by human experts or general-purpose LLMs are sufficient for improving regression across diverse domains, RankRefine offers practicality and broad applicability, especially in low-data settings.



Paperid:759
Authors:Yepeng Weng, Qiao Hu, Xujie Chen, Li Liu, Dianwen Mei, Huishi Qiu, Jiang Tian, zhongchao shi
Abstract:
Speculative decoding is a promising approach for accelerating large language models. The primary idea is to use a lightweight draft model to speculate the output of the target model for multiple subsequent timesteps, and then verify them in parallel to determine whether the drafted tokens should be accepted or rejected. To enhance acceptance rates, existing frameworks typically construct token trees containing multiple candidates in each timestep. However, their reliance on token-level verification mechanisms introduces two critical limitations: First, the probability distribution of a sequence differs from that of individual tokens, leading to suboptimal acceptance length. Second, current verification schemes begin from the root node and proceed layer by layer in a top-down manner. Once a parent node is rejected, all its child nodes should be discarded, resulting in inefficient utilization of speculative candidates. This paper introduces Traversal Verification, a novel speculative decoding algorithm that fundamentally rethinks the verification paradigm through leaf-to-root traversal. Our approach considers the acceptance of the entire token sequence from the current node to the root, and preserves potentially valid subsequences that would be prematurely discarded by existing methods. We theoretically prove that the probability distribution obtained through Traversal Verification is identical to that of the target model, guaranteeing lossless inference while achieving substantial acceleration gains. Experimental results on different large language models and multiple tasks show that our method consistently improves acceptance length and throughput over existing token-level verification methods.



Paperid:760
Authors:Chenyang Ma, Kai Lu, Ruta Desai, Xavier Puig, Andrew Markham, Niki Trigoni
Abstract:
To understand and collaborate with humans, robots must account for individual human traits, habits, and activities over time. However, most robotic assistants lack these abilities, as they primarily focus on predefined tasks in structured environments and lack a human model to learn from. This work introduces COOPERA, a novel framework for COntinual, OPen-Ended human-Robot Assistance, where simulated humans, driven by psychological traits and long-term intentions, interact with robots in complex environments. By integrating continuous human feedback, our framework, for the first time, enables the study of long-term, open-ended human-robot collaboration (HRC) in different collaborative tasks across various time-scales. Within COOPERA, we introduce a benchmark and an approach to personalize the robot's collaborative actions by learning human traits and context-dependent intents. Experiments validate the extent to which our simulated humans reflect realistic human behaviors and demonstrate the value of inferring and personalizing to human intents for open-ended and long-term HRC.



Paperid:761
Authors:Keshigeyan Chandrasegaran, Michael Poli, Daniel Fu, Dongjun Kim, Lea M. Hadzic, Manling Li, Agrim Gupta, Stefano Massaroli, Azalia Mirhoseini, Juan Carlos Niebles, Stefano Ermon, Fei-Fei Li
Abstract:
Abstract:Model architecture design requires decisions such as selecting operators (e.g., attention, convolution) and configurations (e.g., depth, width). However, understanding quality-efficiency tradeoffs of such decisions typically requires costly pretraining, limiting architectural exploration.We present *grafting*, a simple approach for editing pretrained diffusion transformers (DiTs) to materialize new architectures by replacing expensive operators (e.g., self-attention, MLPs) with efficient alternatives. Informed by our analysis of activation behavior and attention locality, we construct a DiT-XL/2-based testbed to study grafting’s impact on model quality. Using this testbed, we develop a family of hybrid designs via grafting:replacing softmax attention with gated convolution, local, and linear attention; and MLPs with variable-width and convolutional variants. Notably, many hybrid designs achieve competitive quality (FID: 2.38–2.64 vs. 2.27 for DiT-XL/2).Next, we graft a text-to-image model (PixArt-$\Sigma$), achieving a 43% speedup with less than a 2% drop in GenEval score. Finally, we present a case study that restructures sequential computation into parallel in DiT-XL/2 via grafting, reducing model depth by 2$\times$, achieving better quality (FID: 2.84) than other models of comparable depth.Our work shows that new diffusion model designs can be explored via grafting pretrained DiTs, with edits ranging from operator replacement to restructuring computation.Pytorch code and grafted models are provided.



Paperid:762
Authors:Benyamin Trachtenberg, Nir Rosenfeld
Abstract:
In strategic classification, the standard supervised learning setting is extended to support the notion of strategic user behavior in the form of costly feature manipulations made in response to a classifier. While standard learning supports a broad range of model classes, the study of strategic classification has, so far, been dedicated mostly to linear classifiers. This work aims to expand the horizon by exploring how strategic behavior manifests under non-linear classifiers and what this implies for learning. We take a bottom-up approach showing how non-linearity affects decision boundary points, classifier expressivity, and model classes complexity. A key finding is that universal approximators (e.g., neural nets) are no longer universal once the environment is strategic. We demonstrate empirically how this can create performance gaps even on an unrestricted model class.



Paperid:763
Authors:Yixuan Zhang, Ruihao Zhu, Qiaomin Xie
Abstract:
Abstract:We study contextual online pricing with biased offline data. For the scalar price elasticity case, we identify the instance-dependent quantity $\delta^2$ that measures how far the offline data lies from the (unknown) online optimum. We show that the time length $T$, bias bound $V$, size $N$ and dispersion $\lambda_{\min}(\hat{\Sigma})$ of the offline data, and $\delta^2$ jointly determine the statistical complexity. An Optimism‑in‑the‑Face‑of‑Uncertainty (OFU) policy achieves a minimax-optimal, instance-dependent regret bound $\tilde{\mathcal{O}}\big(d\sqrt{T} \wedge (V^2T + \frac{dT }{\lambda_{\min}(\hat{\Sigma}) + (N \wedge T) \delta^2})\big)$. For general price elasticity, we establish a worst‑case, minimax-optimal rate $\tilde{\mathcal{O}}\big(d\sqrt{T} \wedge (V^2T + \frac{dT }{\lambda_{\min}(\hat{\Sigma})})\big)$ and provide a generalized OFU algorithm that attains it. When the bias bound $V$ is unknown, we design a robust variant that always guarantees sub‑linear regret and strictly improves on purely online methods whenever the exact bias is small. These results deliver the first tight regret guarantees for contextual pricing in the presence of biased offline data. Our techniques also transfer verbatim to stochastic linear bandits with biased offline data, yielding analogous bounds.



Paperid:764
Authors:Kevin Black, Manuel Galliker, Sergey Levine
Abstract:
Modern AI systems, especially those interacting with the physical world, increasingly require real-time performance. However, the high latency of state-of-the-art generalist models, including recent vision-language-action models (VLAs), poses a significant challenge. While action chunking has enabled temporal consistency in high-frequency control tasks, it does not fully address the latency problem, leading to pauses or out-of-distribution jerky movements at chunk boundaries. This paper presents a novel inference-time algorithm that enables smooth asynchronous execution of action chunking policies. Our method, real-time chunking (RTC), is applicable to any diffusion- or flow-based VLA out of the box with no retraining. It generates the next action chunk while executing the current one, "freezing" actions guaranteed to execute and "inpainting" the rest. To test RTC, we introduce a new benchmark of 12 highly dynamic tasks in the Kinetix simulator, as well as evaluate 6 challenging real-world bimanual manipulation tasks. Results demonstrate that RTC is fast, performant, and uniquely robust to inference delay, significantly improving task throughput and enabling success in precise tasks --- such as lighting a match --- even in the presence of extreme latency.



Paperid:765
Authors:Weijia Shi, Akshita Bhagia, Kevin Farhat, Niklas Muennighoff, Jacob Morrison, Evan Walsh, Dustin Schwenk, Shayne Longpre, Jake Poznanski, Allyson Ettinger, Daogao Liu, Margaret Li, Mike Lewis, Scott Yih, Dirk Groeneveld, Luca Soldaini, Kyle Lo, Noah Smith, Luke Zettlemoyer, Pang Wei Koh, Hanna Hajishirzi, Ali Farhadi, Sewon Min
Abstract:
We introduce FlexOLMo, a new class of language models (LMs) that supports (1) distributed training without data sharing, where different model parameters are independently trained on private datasets, and (2) data-flexible inference, where these parameters along with their associated data can be easily included or excluded from model inferences with no further training. FlexOLMo employs a mixture-of-experts (MoE) architecture where each expert is trained independently on private datasets and later integrated through a new nonparametric routing without any joint training across datasets. FlexOLMo is trained on FLEXMIX, a corpus we curate comprising seven restricted sets, either real or realistic approximations, alongside publicly available datasets. We evaluate models with up to 37 billion parameters (20 billion active) on 31 diverse downstream tasks. We show that a general expert trained on public data can be effectively combined with independently trained experts from other data owners significantly benefiting from these restricted sets (an average 41% relative improvement) while allowing flexible opt-out at inference time (e.g., for users without appropriate licenses or permissions). Our approach also outperforms prior model merging methods by 10.1% on average and surpasses the standard MoE trained without data restrictions using the same training FLOPs. Altogether, FlexOLMo enables training on restricted data while keeping data local and supports fine-grained control of data access at inference.



Paperid:766
Authors:Marcus Lassila, Johan Östman, Khac-Hoang Ngo, Alexandre Graell i Amat
Abstract:
We develop practical and theoretically grounded membership inference attacks (MIAs) against both independent and identically distributed (i.i.d.) data and graph-structured data. Building on the Bayesian decision-theoretic framework of Sabrayolles et al., we derive the Bayes-optimal membership inference rule for node-level MIAs against graph neural networks, addressing key open questions about optimal query strategies in the graph setting. We introduce BASE and G-BASE, computationally efficient approximations of the Bayes-optimal attack. G-BASE achieves superior performance compared to previously proposed classifier-based node-level MIA attacks. BASE, which is also applicable to non-graph data, matches or exceeds the performance of prior state-of-the-art MIAs, such as LiRA and \textsc{RMIA}, at a significantly lower computational cost. Finally, we show that BASE and RMIA are equivalent under a specific hyperparameter setting, providing a principled, Bayes-optimal justification for the RMIA attack.



Paperid:767
Authors:Nianxin Li, Mao Ye, Lihua Zhou, Shuaifeng Li, Song Tang, Luping Ji, Ce Zhu
Abstract:
Unmanned Aerial Vehicle (UAV) object detection faces significant challenges due to complex environmental conditions and different imaging conditions. These factors introduce significant changes in scale and appearance, particularly for small objects that occupy limited pixels and exhibit limited information, complicating detection tasks. To address these challenges, we propose a Multimodel Causal Reasoning framework based on YOLO backbone for UAV Object Detection (MCR-UOD). The key idea is to use the backdoor adjustment to discover the condition-invariant object representation for easy detection. Specifically, the YOLO backbone is first adjusted to incorporate the pre-trained vision-language model. The original category labels are replaced with semantic text prompts, and the detection head is replaced with text-image contrastive learning. Based on this backbone, our method consists of two parts. The first part, named language guided region exploration, discovers the regions with high probability of object existence using text embeddings based on vision-language model such as CLIP. Another part is the backdoor adjustment casual reasoning module, which constructs a confounder dictionary tailored to different imaging conditions to capture global image semantics and derives a prior probability distribution of shooting conditions. During causal inference, we use the confounder dictionary and the prior to intervene on local instance features, disentangling condition variations, and obtaining condition-invariant representations. Experimental results on several public datasets confirm the state-of-the-art performance of our approach. The code, data and models will be released upon publication of this paper.



Paperid:768
Authors:Steve Hanneke, Amirreza Shaeiri, Hongao Wang
Abstract:
Abstract:We study the problem of multiclass learning with bandit feedback in both the i.i.d. batch and adversarial online models. In the *uniform* learning framework, it is well known that no hypothesis class $\mathcal{H}$ is learnable in either model when the effective number of labels is unbounded. In contrast, within the *universal* learning framework, recent works by (Hanneke et al., 2025b) and (Hanneke et al., 2025a) have established surprising exact equivalences between learnability under bandit feedback and full supervision in both the i.i.d. batch and adversarial online models, respectively. This raises a natural question: What happens in the *non-uniform* learning framework, which lies between the uniform and universal learning frameworks? Our contributions are twofold: (1) We provide a combinatorial characterization of learnable hypothesis classes in both models, in the realizable and agnostic settings, within the non-uniform learning framework. Notably, this includes elementary and natural hypothesis classes, such as a countably infinite collection of constant functions over some domain that is learnable in both models. (2) We construct a hypothesis class that is non-uniformly learnable under full supervision in the adversarial online model (and thus also in the i.i.d. batch model), but not non-uniformly learnable under bandit feedback in the i.i.d. batch model (and thus also not in the adversarial online model). This serves as our main novel technical contribution that reveals a fundamental distinction between the non-uniform and universal learning frameworks.



Paperid:769
Authors:Edoardo Cetin, Tianyu Zhao, Yujin Tang
Abstract:
Training reasoning language models (LMs) with reinforcement learning (RL) for one-hot correctness inherently relies on the LM being able to explore and solve its task with some chance at initialization. Furthermore, a key use case of reasoning LMs is to act as teachers for distilling new students and cold-starting future RL iterations rather than being deployed themselves. From these considerations, we introduce a new framework that avoids RL's exploration challenge by training a new class of Reinforcement-Learned Teachers (RLTs) focused on yielding the most effective downstream distillation. RLTs are prompted with both the question and solution to each problem, and tasked to simply "connect-the-dots" with detailed explanations tailored for their students. We train RLTs with dense rewards obtained by feeding each explanation to the student and testing its understanding of the problem's solution. In practice, the raw outputs of a 7B RLT provide higher final performance on competition and graduate-level tasks than existing distillation and cold-starting pipelines that collect and postprocess the reasoning traces of orders of magnitude larger LMs. Furthermore, RLTs maintain their effectiveness when training larger students and when applied zero-shot to out-of-distribution tasks, unlocking new levels of efficiency and re-usability for the RL reasoning framework.



Paperid:770
Authors:Maria-Florina Balcan, Avrim Blum, Zhiyuan Li, Dravyansh Sharma
Abstract:
Chain-of-Thought reasoning has emerged as a powerful approach for solving complex mathematical and logical problems. However, it can often veer off track through incorrect or unsubstantiated inferences. Formal mathematical reasoning, which can be checked with a formal verifier, is one approach to addressing this issue. However, currently LLMs are simply not good enough to solve complex problems in a formal way, and even just formalizing an informal problem statement can be challenging. Motivated by this fact, in this work we consider the problem of learning reliable verifiers for natural language Chain-of-Thought reasoning. That is, given a problem statement and step-by-step solution in natural language, the aim of the verifier is to output [Yes] if the reasoning steps in the solution are all valid, and [No] otherwise. In this work we give a formal PAC-learning framework for studying this problem. We propose and analyze several natural verification goals, at different levels of strength, in this framework. We provide sample complexity upper-bounds for learning verifiers satisfying these goals, as well as lower-bound and impossibility results for learning other natural verification objectives without additional assumptions.



Paperid:771
Authors:Kale-ab Tessera, Muhammad Arrasy Rahman, Amos Storkey, Stefano Albrecht
Abstract:
Adaptability to specialised or homogeneous behaviours is critical in cooperative multi-agent reinforcement learning (MARL). Parameter sharing (PS) techniques, common for efficient adaptation, often limit behavioural diversity due to cross-agent gradient interference, which we show can be exacerbated by the coupling of observations and agent IDs. Current remedies typically add complexity through altered objectives, manual preset diversity levels, or sequential updates. We ask: can shared policies adapt without these complexities? We proposeHyperMARL, a PS approach using hypernetworks for dynamic agent-specific parameters, without altering the RL objective or requiring preset diversity levels. HyperMARL's explicitdecouplingof observation- and agent-conditioned gradients empirically reduces policy gradient variance, facilitates shared-policy adaptation (including specialisation), and helps mitigate cross-agent interference. Across diverse MARL benchmarks (up to 20 agents), requiring homogeneous, heterogeneous, or mixed behaviours, HyperMARL achieves competitive performance against key baselines -- fully shared, non-parameter sharing, and three diversity-promoting methods -- while preserving behavioural diversity comparable to non-parameter sharing. These findings establish HyperMARL as a versatile approach for adaptive MARL.



Paperid:772
Authors:Wei Wang, Fuqing Bie, Junzhe Chen, Dan Zhang, Shiyu Huang, Evgeny Kharlamov, Jie Tang
Abstract:
Complex games have long been an important benchmark for testing the progress of artificial intelligence algorithms. AlphaGo, AlphaZero, and MuZero have defeated top human players in Go and Chess, garnering widespread societal attention towards artificial intelligence. Concurrently, large language models (LLMs) have exhibited remarkable capabilities across various tasks, raising the question of whether LLMs can achieve similar success in complex games. In this paper, we explore the potential of LLMs in mastering complex card games. We systematically assess the learning capabilities of LLMs across eight diverse card games, evaluating the impact of fine-tuning on high-quality gameplay data, and examining the models' ability to retain general capabilities while mastering these games. Our findings indicate that: (1) LLMs can approach the performance of strong game AIs through supervised fine-tuning on high-quality data, (2) LLMs can master multiple complex card games simultaneously, with performance augmentation for games with similar rules and conflicts for dissimilar ones, and (3) LLMs experience a decline in general capabilities when mastering complex games, but this decline can be mitigated by integrating a certain amount of general instruction data. The evaluation results demonstrate strong learning ability and versatility of LLMs. The code is available at https://anonymous.4open.science/r/LLM4CardGame-D834



Paperid:773
Authors:Brian Moser, Federico Raue, Sebastian Palacio, Stanislav Frolov, Andreas Dengel
Abstract:
Abstract:Dataset distillation seeks to condense datasets into smaller but highly representative synthetic samples. While diffusion models now lead all generative benchmarks, current distillation methods avoid them and rely instead on GANs or autoencoders, or, at best, sampling from a fixed diffusion prior. This trend arises because naive backpropagation through the long denoising chain leads to vanishing gradients, which prevents effective synthetic sample optimization. To address this limitation, we introduce LD3M, the first method to learn gradient-based distilled latents and class embeddings end-to-end through a pre-trained latent diffusion model. A linearly decaying skip connection, injected from the initial noisy state into every reverse step, preserves the gradient signal across dozens of timesteps without requiring diffusion weight fine-tuning. Across multiple ImageNet subsets at $128\times128$ and $256\times256$, LD3M improves downstream accuracy by up to 4.8 percentage points (1 IPC) and 4.2 points (10 IPC) over the prior state-of-the-art. The code for LD3M is provided at https://github.com/DOUBLE_BLIND/ld3m.



Paperid:774
Authors:Yufan Dang, Chen Qian, Xueheng Luo, Jingru Fan, Zihao Xie, Ruijie Shi, Weize Chen, Cheng Yang, Xiaoyin Che, Ye Tian, Xuantang Xiong, Lei Han, Zhiyuan Liu, Maosong Sun
Abstract:
Large language models (LLMs) have achieved remarkable results across diverse downstream tasks, but their monolithic nature restricts scalability and efficiency in complex problem-solving. While recent research explores multi-agent collaboration among LLMs, most approaches rely on static organizational structures that struggle to adapt as task complexity and agent numbers grow, resulting in coordination overhead and inefficiencies. To this end, we propose a puppeteer-style paradigm for LLM-based multi-agent collaboration, where a centralized orchestrator ("puppeteer") dynamically directs agents ("puppets") in response to evolving task states. This orchestrator is trained via reinforcement learning to adaptively sequence and prioritize agents, enabling flexible and evolvable collective reasoning. Experiments on closed- and open-domain scenarios show that this method achieves superior performance with reduced computational costs. Analyses further reveal that the key improvements consistently stem from the emergence of more compact, cyclic reasoning structures under the orchestrator’s evolution.



Paperid:775
Authors:Achraf Azize, Yulian Wu, Junya Honda, Francesco Orabona, Shinji Ito, Debabrota Basu
Abstract:
Abstract:As sequential learning algorithms are increasingly applied to real life, ensuring data privacy while maintaining their utilities emerges as a timely question. In this context, regret minimisation in stochastic bandits under $\epsilon$-global Differential Privacy (DP) has been widely studied. The present literature poses a significant gap between the best-known regret lower and upper bound in this setting, though they ``match in order''.Thus, we revisit the regret lower and upper bounds of $\epsilon$-global DP bandits and improve both. First, we prove a tighter regret lower bound involving a novel information-theoretic quantity characterising the hardness of $\epsilon$-global DP in stochastic bandits. This quantity smoothly interpolates between Kullback–Leibler divergence and Total Variation distance, depending on the privacy budget $\epsilon$.Then, we choose two asymptotically optimal bandit algorithms, i.e., KL-UCB and IMED, and propose their DP versions using a unified blueprint, i.e., (a) running in arm-dependent phases, and (b) adding Laplace noise to achieve privacy. For Bernoulli bandits, we analyse the regrets of these algorithms and show that their regrets asymptotically match our lower bound up to a constant arbitrary close to 1.At the core of our algorithms lies a new concentration inequality for sums of Bernoulli variables under Laplace mechanism, which is a new DP version of the Chernoff bound.Finally, our numerical experiments validate that DP-KLUCB and DP-IMED achieve lower regret than the existing $\epsilon$-global DP bandit algorithms.



Paperid:776
Authors:Carlo Alfano, Silvia Sapora, Jakob Foerster, Patrick Rebeschini, Yee Whye Teh
Abstract:
Evaluating preference optimization (PO) algorithms on LLM alignment is a challenging task that presents prohibitive costs, noise, and several variables like model size and hyper-parameters. In this work, we show that it is possible to gain insights on the efficacy of PO algorithm on much simpler benchmarks. We design a diagnostic suite of MuJoCo tasks and datasets, which we use to systematically evaluate PO algorithms, establishing a more controlled and cheaper benchmark. We then propose a novel family of PO algorithms based on mirror descent, which we call Mirror Preference Optimization (MPO). Through evolutionary strategies, we search this class to discover algorithms specialized to specific properties of preference datasets, such as mixed-quality or noisy data. We demonstrate that our discovered PO algorithms outperform all known algorithms in the targeted MuJoCo settings. Finally, based on the insights gained from our MuJoCo experiments, we design a novel PO algorithm that significantly outperforms existing baselines in an LLM alignment task.



Paperid:777
Authors:Brandon Feng, David Park, Xihaier Luo, Arantxa Urdangarin, Shinjae Yoo, Brian Reich
Abstract:
Fitting Gaussian Processes (GPs) provides interpretable aleatoric uncertainty quantification for estimation of spatio-temporal fields. Spatio-temporal deep learning models, while scalable, typically assume a simplistic independent covariance matrix for the response, failing to capture the underlying correlation structure. However, spatio-temporal GPs suffer from issues of scalability and various forms of approximation bias resulting from restrictive assumptions of the covariance kernel function. We propose STACI, a novel framework consisting of a variational Bayesian neural network approximation of non-stationary spatio-temporal GP along with a novel spatio-temporal conformal inference algorithm. STACI is highly scalable, taking advantage of GPU training capabilities for neural network models, and provides statistically valid prediction intervals for uncertainty quantification. STACI outperforms competing GPs and deep methods in accurately approximating spatio-temporal processes and we show it easily scales to datasets with millions of observations.



Paperid:778
Authors:Sina Tootoonian, Andreas Schaefer
Abstract:
A common view of sensory processing is as probabilistic inference of latent causes from receptor activations. Standard approaches often assume these causes are \textit{a priori} independent, yet real-world generative factors are typically correlated. Representing such structured priors in neural systems poses architectural challenges, particularly when direct interactions between units representing latent causes are biologically implausible or computationally expensive. Inspired by the architecture of the olfactory bulb, we propose a novel circuit motif that enables inference with correlated priors without requiring direct interactions among latent cause units. The key insight lies in using \textit{sister cells}: neurons receiving shared receptor input but connected differently to local interneurons. The required interactions among latent units are implemented indirectly through their connections to the sister cells, such that correlated connectivity implies anti-correlation in the prior and vice versa. We use geometric arguments to construct connectivity that implements a given prior and to bound the number of causes for which such priors can be constructed. Using simulations, we demonstrate the efficacy of such priors for inference in noisy environments and compare the inference dynamics to those experimentally observed. Finally, we show how, under certain assumptions on latent representations, the prior used can be inferred from sister cell activations. While biologically grounded in the olfactory system, our mechanism generalises to other natural and artificial sensory systems and may inform the design of architectures for efficient inference under correlated latent structure.



Paperid:779
Authors:Yuanhang Yang, Chaozheng Wang, Jing Li
Abstract:
Sparse Mixture of Experts (MoE) architectures have emerged as a promising approach for scaling Transformer models. While initial works primarily incorporated MoE into feed-forward network (FFN) layers, recent studies have explored extending the MoE paradigm to attention layers to enhance model performance. However, existing attention-based MoE layers require specialized implementations and demonstrate suboptimal performance compared to their FFN-based counterparts. In this paper, we aim to unify the MoE designs in attention and FFN layers by introducing a novel reformulation of the attention mechanism, revealing an underlying FFN-like structure within attention modules. Our proposed architecture, UMoE, achieves superior performance through attention-based MoE layers while enabling efficient parameter sharing between FFN and attention components.



Paperid:780
Authors:Runsong Zhu, Ka-Hei Hui, Zhengzhe Liu, Qianyi Wu, Weiliang Tang, Shi Qiu, Pheng-Ann Heng, Chi-Wing Fu
Abstract:
Open-vocabulary 3D segmentation is a fundamental yet challenging task, requiring a mutual understanding of both segmentation and language. However, existing Gaussian-splatting-based methods rely either on a single 3D language field, leading to inferior segmentation, or on pre-computed class-agnostic segmentations, suffering from error accumulation. To address these limitations, we present COS3D, a new collaborative prompt-segmentation framework that contributes to effectively integrating complementary language and segmentation cues throughout its entire pipeline. We first introduce the new concept of collaborative field, comprising an instance field and a language field, as the cornerstone for collaboration. During training, to effectively construct the collaborative field, our key idea is to capture the intrinsic relationship between the instance field and language field, through a novel instance-to-language feature mapping and designing an efficient two-stage training strategy. During inference, to bridge distinct characteristics of the two fields, we further design an adaptive language-to-instance prompt refinement, promoting high-quality prompt-segmentation inference. Extensive experiments not only demonstrate COS3D's leading performance over existing methods on two widely-used benchmarks but also show its high potential to various applications,~\ie, novel image-based 3D segmentation, hierarchical segmentation, and robotics.



Paperid:781
Authors:Yuanqi Du, Jiajun He, Francisco Vargas, Yuanqing Wang, Carla Gomes, José Miguel Hernández-Lobato, Eric Vanden-Eijnden
Abstract:
We present Free energy Estimators with Adaptive Transport (FEAT), a novel framework for free energy estimation -- a critical challenge across scientific domains. FEAT leverages learned transports implemented via stochastic interpolants and provides consistent, minimum-variance estimators based on escorted Jarzynski equality and controlled Crooks theorem, alongside variational upper and lower bounds on free energy differences. Unifying equilibrium and non-equilibrium methods under a single theoretical framework, FEAT establishes a principled foundation for neural free energy calculations. Experimental validation on toy examples, molecular simulations, and quantum field theory demonstrates promising improvements over existing learning-based methods.



Paperid:782
Authors:Reginald McLean, Evangelos Chatzaroulas, Luc McCutcheon, Frank Röder, Tianhe Yu, Zhanpeng He, K.R. Zentner, Ryan Julian, J Terry, Isaac Woungang, Nariman Farsad, Pablo Samuel Castro
Abstract:
Meta-World is widely used for evaluating multi-task and meta-reinforcement learning agents, which are challenged to master diverse skills simultaneously. Since its introduction however, there have been numerous undocumented changes which inhibit a fair comparison of algorithms. This work strives to disambiguate these results from the literature, while also leveraging the past versions of Meta-World to provide insights into multi-task and meta-reinforcement learning benchmark design. Through this process we release an open-source version of Meta-World that has full reproducibility of past results, is more technically ergonomic, and gives users more control over the tasks that are included in a task set.



Paperid:783
Authors:Sana Ayromlou, David B. Emerson
Abstract:
Federated learning (FL) has become an effective and widely used approach to training deep learning models on decentralized datasets held by distinct clients. FL also strengthens both security and privacy protections for training data. Common challenges associated with statistical heterogeneity between distributed datasets have spurred significant interest in personalized FL (pFL) methods, where models combine aspects of global learning with local modeling specific to each client's unique characteristics. In this work, the efficacy of theoretically supported, adaptive MMD measures within the Ditto framework, a state-of-the-art technique in pFL, are investigated. The use of such measures significantly improves model performance across a variety of tasks, especially those with pronounced feature heterogeneity. While the Ditto algorithm is specifically considered, such measures are directly applicable to a number of other pFL settings, and the results motivate the use of constraints tailored to the various kinds of heterogeneity expected in FL systems.



Paperid:784
Authors:Shuyang Xu, Zhiyang Dou, Mingyi Shi, Liang Pan, Leo Ho, Jingbo Wang, Yuan Liu, Cheng Lin, Yuexin Ma, Wenping Wang, Taku Komura
Abstract:
Enabling virtual humans to dynamically and realistically respond to diverse auditory stimuli remains a key challenge in character animation, demanding the integration of perceptual modeling and motion synthesis. Despite its significance, this task remains largely unexplored. Most previous works have primarily focused on mapping modalities such as language, audio, and music to human motion generation. As of yet, these models typically overlook the impact of spatial features encoded in spatial audio signals on human motion. To bridge this gap and enable high-quality modeling of human movements in response to spatial audio, we introduce the Spatial Audio-Driven Human Motion (SAM) dataset, which contains diverse and high-quality spatial audio and motion data. For benchmarking, we develop a simple yet effective diffusion-based generative framework for human MOtion generation driven by SPatial Audio, termed MOSPA, which captures the relationship between body motion and spatial audio through an effective fusion mechanism. Once trained, MOSPA could generate diverse and realistic human motions conditioned on varying spatial audio inputs. We conducted extensive experiments to validate our method, which achieves state-of-the-art performance on this task. Our model and dataset will be open-sourced upon acceptance. Please refer to our supplementary video for more details.



Paperid:785
Authors:Simon Ferreira, Charles Assaad
Abstract:
Abstract:The do-calculus is a sound and complete tool for identifying causal effects in acyclic directed mixed graphs (ADMGs) induced by structural causal models (SCMs). However, in many real-world applications, especially in high-dimensional setting, constructing a fully specified ADMG is often infeasible. This limitation has led to growing interest in partially specified causal representations, particularly through cluster-directed mixed graphs (C-DMGs), which group variables into clusters and offer a more abstract yet practical view of causal dependencies. While these representations can include cycles, recent work has shown that the do-calculus remains sound and complete for identifying macro-level causal effects in C-DMGs over ADMGs under the assumption that all clusters size are greater than $1$. Nevertheless, real-world systems often exhibit cyclic causal dynamics at the structural level. To account for this, input-output structural causal models (ioSCMs) have been introduced as a generalization of SCMs that allow for cycles. ioSCMs induce another type of graph structure known as a directed mixed graph (DMG). Analogous to the ADMG setting, one can define C-DMGs over DMGs as high-level representations of causal relations among clusters of variables. In this paper, we prove that, unlike in the ADMG setting, the do-calculus is unconditionally sound and complete for identifying macro causal effects in C-DMGs over DMGs. Furthermore, we show that the graphical criteria for non-identifiability of macro causal effects previously established C-DMGs over ADMGs naturally extend to a subset of C-DMGs over DMGs.



Paperid:786
Authors:Richard Cole, Anupam Gupta, Pranav Jangir
Abstract:
Abstract:We consider the problem of scheduling $m$ jobs on $n$ unrelated strategic machines to minimize the maximum load of any machine, but the machines are strategic and may misreport processing times to minimize their own load. The pioneering work of Nisan and Ronen gave an $n$-approximate deterministic strategyproof mechanism for this setting, and this was recently shown to be best possible by the breakthrough results of Christodoulou et al. This large approxation guarantee begs the question: how can we avoid these large worst-case results. In this work, we use the powerful framework of algorithms with (machine-learned) predictions to bypass these strong impossibility results. We show how we can predict $O(m+n)$ values to obtain a deterministic strategyproof algorithm whose makespan is within a constant factor of the optimal makespan when the predictions are correct, and $O(n)$ times the optimum no matter how poor the predictions are.



Paperid:787
Authors:Lei Hu, Yongjing Ye, Shihong Xia
Abstract:
The expansion of instruction-tuning data has enabled foundation language models to exhibit improved instruction adherence and superior performance across diverse downstream tasks. Semantically-rich 3D human motion is being progressively integrated with these foundation models to enhance multimodal understanding and cross-modal generation capabilities. However, the modality gap between human motion and text raises unresolved concerns about catastrophic forgetting during this integration. In addition, developing autoregressive-compatible pose representations that preserve generalizability across heterogeneous downstream tasks remains a critical technical barrier. To address these issues, we propose the Human Motion-Vision-Language Model (HMVLM), a unified framework based on the Mixture of Expert Low-Rank Adaption(MoE LoRA) strategy. The framework leverages the gating network to dynamically allocate LoRA expert weights based on the input prompt, enabling synchronized fine-tuning of multiple tasks. To mitigate catastrophic forgetting during instruction-tuning, we introduce a novel zero expert that preserves the pre-trained parameters for general linguistic tasks. For pose representation, we implement body-part-specific tokenization by partitioning the human body into different joint groups, enhancing the spatial resolution of the representation. Experiments show that our method effectively alleviates knowledge forgetting during instruction-tuning and achieves remarkable performance across diverse human motion downstream tasks.



Paperid:788
Authors:Yedi Zhang, Sun Emma, Annabelle En, Jin Song Dong
Abstract:
Large language models (LLMs) have emerged as a dominant AI paradigm due to their exceptional text understanding and generation capabilities. However, their tendency to generate inconsistent or erroneous outputs challenges their reliability, especially in high-stakes domains requiring accuracy and trustworthiness. Existing research primarily focuses on detecting and mitigating model misbehavior in general-purpose scenarios, often overlooking the potential of integrating domain-specific knowledge. In this work, we advance misbehavior detection by incorporating domain knowledge. The core idea is to design a general specification language that enables domain experts to customize domain-specific predicates in a lightweight and intuitive manner, supporting later runtime verification of LLM outputs. To achieve this, we design a novel specification language ESL and introduce a runtime verification framework RvLLM to validate LLM output against domain-specific constraints defined in ESL. RvLLM operates in two main stages: interpretation and reasoning. During interpretation, it derives interpretations of the specification based on context, which then guide the reasoning process to identify inconsistencies. When new knowledge is derived, RvLLM issues a follow-up query to the LLM to further verify the consistency. We evaluate RvLLM on three representative tasks: violation detection against Singapore Rapid Transit Systems Act, numerical comparison, and inequality solving. Experimental results demonstrate that RvLLM effectively detects erroneous outputs across various LLMs in a lightweight and flexible manner. The results reveal that despite their impressive capabilities, LLMs remain prone to low-level errors due to limited interpretability and a lack of formal guarantees during inference, and our framework offers a potential long-term solution by leveraging expert domain knowledge to rigorously and efficiently verify LLM outputs.



Paperid:789
Authors:Chenyu Zheng, Xinyu Zhang, Rongzhen Wang, Wei Huang, Zhi Tian, Weilin Huang, Jun Zhu, Chongxuan LI
Abstract:
Abstract:Diffusion Transformers have emerged as the foundation for vision generative models, but their scalability is limited by the high cost of hyperparameter (HP) tuning at large scales. Recently, Maximal Update Parametrization ($\mu$P) was proposed for vanilla Transformers, which enables stable HP transfer from small to large language models, and dramatically reduces tuning costs. However, it remains unclear whether $\mu$P of vanilla Transformers extends to diffusion Transformers, which differ architecturally and objectively. In this work, we generalize $\mu$P to diffusion Transformers and validate its effectiveness through large-scale experiments. First, we rigorously prove that $\mu$P of mainstream diffusion Transformers, including DiT, U-ViT, PixArt-$\alpha$, and MMDiT, aligns with that of the vanilla Transformer, enabling the direct application of existing $\mu$P methodologies. Leveraging this result, we systematically demonstrate that DiT-$\mu$P enjoys robust HP transferability. Notably, DiT-XL-2-$\mu$P with transferred learning rate achieves 2.9$\times$ faster convergence than the original DiT-XL-2. Finally, we validate the effectiveness of $\mu$P on text-to-image generation by scaling PixArt-$\alpha$ from 0.04B to 0.61B and MMDiT from 0.18B to 18B. In both cases, models under $\mu$P outperform their respective baselines while requiring small tuning cost—only 5.5% of one training run for PixArt-$\alpha$ and 3% of consumption by human experts for MMDiT-18B. \textit{These results establish $\mu$P as a principled and efficient framework for scaling diffusion Transformers}.



Paperid:790
Authors:Lingren Wang, Wenxuan Tu, Jiaxin Wang, Xiong Wang, Jieren Cheng, Jingxin Liu
Abstract:
Abstract:Federated Graph Learning (FGL) shows superiority in cross-domain graph training while preserving data privacy. Existing approaches usually assume shared generic knowledge (e.g., prototypes, spectral features) via aggregating local structures statistically to alleviate structural heterogeneity. However, imposing overly strict assumptions about the presumed correlation between structural features and the global objective often fails in generalizing to local tasks, leading to suboptimal performance. To tackle this issue, we propose a $\underline{\text{Fed}}$erated $\underline{\text{I}}$nvariant $\underline{\text{G}}$raph $\underline{\text{L}}$earning (FedIGL) framework based on invariant learning, which effectively disrupts spurious correlations and further mines the invariant factors across different distributions. Specifically, we design a global model for collaborative training across clients and client-specific models retained on the clients, which are developed to handle client-agnostic and client-specific subgraph patterns, respectively. Subsequently, we propose a novel Bi-Gradient Regularization strategy that introduces gradient constraints to guide the model in identifying client-agnostic and client-specific subgraph patterns for better graph representations without compromising privacy. Extensive experiments on graph-level clustering and classification tasks demonstrate the effectiveness and superiority of FedIGL against its competitors.



Paperid:791
Authors:Yunjie Tian, Qixiang Ye, DAVID DOERMANN
Abstract:
Enhancing the network architecture of the YOLO framework has been crucial for a long time. Still, it has focused on CNN-based improvements despite the proven superiority of attention mechanisms in modeling capabilities. This is because attention-based models cannot match the speed of CNN-based models. This paper proposes an attention-centric YOLO framework, namely YOLOv12, that matches the speed of previous CNN-based ones while harnessing the performance benefits of attention mechanisms. YOLOv12 surpasses popular real-time object detectors in accuracy with competitive speed. For example, YOLOv12-N achieves 40.6% mAP with an inference latency of 1.63 ms on a T4 GPU, outperforming advanced YOLOv10-N / YOLO11-N by 2.1%/1.2% mAP with a comparable speed. This advantage extends to other model scales. YOLOv12 also surpasses end-to-end real-time detectors that improve DETR, such as RT-DETRv2 / RT-DETRv3: YOLOv12-X beats RT-DETRv2-R101 / RT-DETRv3-R101 while running faster with fewer calculations and parameters. See more comparisons in Figure 1. The code and models will be open-sourced.



Paperid:792
Authors:Ahmed Rashwan, Keith Briggs, Chris Budd, Lisa Kreusser
Abstract:
Many machine learning applications require outputs that satisfy complex, dynamic constraints. This task is particularly challenging in Graph Neural Network models due to the variable output sizes of graph-structured data. In this paper, we introduce ProjNet, a Graph Neural Network framework which satisfies input-dependant constraints. ProjNet combines a sparse vector clipping method with the Component-Averaged Dykstra (CAD) algorithm, an iterative scheme for solving the best-approximation problem. We establish a convergence result for CAD and develop a GPU-accelerated implementation capable of handling large-scale inputs efficiently. To enable end-to-end training, we introduce a surrogate gradient for CAD that is both computationally efficient and better suited for optimization than the exact gradient. We validate ProjNet on four classes of constrained optimisation problems: linear programming, two classes of non-convex quadratic programs, and radio transmit power optimization, demonstrating its effectiveness across diverse problem settings.



Paperid:793
Authors:Dhruv Agarwal, Bodhisattwa Prasad Majumder, Reece Adamson, Megha Chakravorty, Satvika Gavireddy, Aditya Parashar, Harshit Surana, Bhavana Dalvi Mishra, Andrew McCallum, Ashish Sabharwal, Peter Clark
Abstract:
The promise of autonomous scientific discovery (ASD) hinges not only on answering questions, but also on knowing which questions to ask. Most recent works in ASD explore the use of large language models (LLMs) in goal-driven settings, relying on human-specified research questions to guide hypothesis generation. However, scientific discovery may be accelerated further by allowing the AI system to drive exploration by its own criteria. The few existing approaches in open-ended ASD select hypotheses based on diversity heuristics or subjective proxies for human interestingness, but the former struggles to meaningfully navigate the typically vast hypothesis space, and the latter suffers from imprecise definitions. This paper presents AutoDS—a method for open-ended ASD that instead drives scientific exploration using Bayesian surprise. Here, we quantify the epistemic shift from the LLM’s prior beliefs about a hypothesis to its posterior beliefs after gathering experimental results. To efficiently explore the space of nested hypotheses, our method employs a Monte Carlo tree search (MCTS) strategy with progressive widening using surprisal as the reward function. We evaluate AutoDS in the setting of data-driven discovery across 21 real-world datasets spanning domains such as biology, economics, finance, and behavioral science. Our results demonstrate that under a fixed budget, AutoDS substantially outperforms competitors by producing 5-29% more discoveries deemed surprising by the LLM. Our human evaluation further finds that two-thirds of AutoDS discoveries are surprising to the domain experts, suggesting this is an important step forward towards building open-ended ASD systems.



Paperid:794
Authors:Jiayou Zhang, Yifan Shen, Guangyi Chen, Le Song, Eric Xing
Abstract:
Vector-Quantized Variational Autoencoders (VQVAEs) have enabled strong performance in generative modeling by mapping continuous data to learnable codes.In this work, we identify a surprising yet consistent phenomenon that we term \emph{dimensional collapse}: despite using high-dimensional embeddings, VQVAEs tend to compress their representations into a much smaller subspace, typically only 4 to 10 dimensions.We provide an in-depth analysis of this phenomenon and reveal its relation to model performance and learning dynamics.Interestingly, VQVAEs naturally gravitate toward this low-dimensional regime, and enforcing higher-dimensional usage (e.g., via rank regularization) could lead to degraded performance.To overcome this low-dimensionality limitation, we propose \textbf{Divide-and-Conquer VQ (DCVQ)}, which partitions the latent space into multiple low-dimensional subspaces, each quantized independently.By design, each subspace respects the model’s preference for low dimensionality, while their combination expands the overall capacity.Our results show that DCVQ overcomes the inherent dimensional bottleneck and achieves improved reconstruction quality across image datasets.



Paperid:795
Authors:Antoine Ledent, Rodrigo Alves, Yunwen Lei
Abstract:
Abstract:It has been recently observed in much of the literature that neural networks exhibit a bottleneck rank property: for larger depths, the activation and weights of neural networks trained with gradient-based methods tend to be of approximately low rank. In fact, the rank of the activations of each layer converges to a fixed value referred to as the ``bottleneck rank", which is the minimum rank required to represent the training data. This perspective is in line with the observation that regularizing linear networks (without activations) with weight decay is equivalent to minimizing the Schatten $p$ quasi norm of the neural network. In this paper we investigate the implications of this phenomenon for generalization. More specifically, we prove generalization bounds for neural networks which exploit the approximate low rank structure of the weight matrices if present. The final results rely on the Schatten $p$ quasi norms of the weight matrices: for small p, the bounds exhibit a sample complexity $ \widetilde{O}(WrL^2)$ where $W$ and $L$ are the width and depth of the neural network respectively and where $r$ is the rank of the weight matrices. As $p$ increases, the bound behaves more like a norm-based bound instead. The proof techniques involve a careful interpolation between the parametric and norm based regimes. We also demonstrate in experiments that this bound outperforms both classic parameter counting and norm based bounds in the typical overparametrized regime.



Paperid:796
Authors:Zhongchen Zhao, Chaodong Xiao, Hui LIN, Qi Xie, Lei Zhang, Deyu Meng
Abstract:
Global dependency modeling and spatial position modeling are two core issues of the foundational architecture design in current deep learning frameworks. Recently, Vision Transformers (ViTs) have achieved remarkable success in computer vision, leveraging the powerful global dependency modeling capability of the self-attention mechanism. Furthermore, Mamba2 has demonstrated its significant potential in natural language processing tasks by explicitly modeling the spatial adjacency prior through the structured mask. In this paper, we propose Polyline Path Masked Attention (PPMA) that integrates the self-attention mechanism of ViTs with an enhanced structured mask of Mamba2, harnessing the complementary strengths of both architectures.Specifically, we first ameliorate the traditional structured mask of Mamba2 by introducing a 2D polyline path scanning strategy and derive its corresponding structured mask, polyline path mask, which better preserves the adjacency relationships among image tokens. Notably, we conduct a thorough theoretical analysis of the structural characteristics of the proposed polyline path mask and design an efficient algorithm for the computation of the polyline path mask. Next, we embed the polyline path mask into the self-attention mechanism of ViTs, enabling explicit modeling of spatial adjacency prior. Extensive experiments on standard benchmarks, including image classification, object detection, and segmentation, demonstrate that our model outperforms previous state-of-the-art approaches based on both state-space models and Transformers. For example, our proposed PPMA-T/S/B models achieve 48.7\%/51.1\%/52.3\% mIoU on the ADE20K semantic segmentation task, surpassing RMT-T/S/B by 0.7\%/1.3\%/0.3\%, respectively. Code is available at \url{https://anonymous.4open.science/r/PPMA-3948}.



Paperid:797
Authors:Saptarshi Saha, Dhruv Rathore, Utpal Garain
Abstract:
Most counterfactual inference frameworks traditionally assume acyclic structural causal models (SCMs), i.e. directed acyclic graphs (DAGs). However, many real-world systems (e.g. biological systems) contain feedback loops or cyclic dependencies that violate acyclicity. In this work, we study counterfactual inference under shift-scale interventions in cyclic SCMs.



Paperid:798
Authors:Haibo Wu, Marina Knight, Keiland Cooper, Norbert Fortin, Hernando Ombao
Abstract:
Understanding the evolving dependence between two clusters of multivariate signals is fundamental in neuroscience and other domains where sub-networks in a system interact dynamically over time. Despite the growing interest in multivariate time series analysis, existing methods for between-clusters dependence typically rely on the assumption of stationarity and lack the temporal resolution to capture transient, frequency-specific interactions. To overcome this limitation, we propose scale-specific wavelet canonical coherence (WaveCanCoh), a novel framework that extends canonical coherence analysis to the nonstationary setting by leveraging the multivariate locally stationary wavelet model. The proposed WaveCanCoh enables the estimation of time-varying canonical coherence between clusters, providing interpretable insight into scale-specific time-varying interactions between clusters. Through extensive simulation studies, we demonstrate that WaveCanCoh accurately recovers true coherence structures under both locally stationary and general nonstationary conditions. Application to local field potential (LFP) activity data recorded from the hippocampus reveals distinct dynamic coherence patterns between correct and incorrect memory-guided decisions, illustrating capacity of the method to detect behaviorally relevant neural coordination. These results highlight WaveCanCoh as a flexible and principled tool for modeling complex cross-group dependencies in nonstationary multivariate systems.



Paperid:799
Authors:Divya Shanmugam, Shuvom Sadhuka, Manish Raghavan, John Guttag, Bonnie Berger, Emma Pierson
Abstract:
It is difficult to evaluate machine learning classifiers without large labeled datasets, which are often unavailable. In contrast, unlabeled data is plentiful, but not easily used for evaluation. Here, we introduce Semi-Supervised Model Evaluation (SSME), a method that uses both labeled and unlabeled data to evaluate machine learning classifiers. The key idea is to estimate the joint distribution of ground truth labels and classifier scores using a semi-supervised mixture model. The semi-supervised mixture model allows SSME to learn from three sources of information: unlabeled data, multiple classifiers, and probabilistic classifier scores. Once fit, the mixture model enables estimation of any metric that is a function of classifier scores and ground truth labels (e.g., accuracy or AUC). We derive theoretical bounds on the error of these estimates, showing that estimation error decreases with the number of classifiers and the amount of unlabeled data. We present experiments in four domains where obtaining large labeled datasets is often impractical: healthcare, content moderation, molecular property prediction, and text classification. Our results demonstrate that SSME estimates performance more accurately than do competing methods, reducing error by 5.1x relative to using labeled data alone and 2.4x relative to the next best method.



Paperid:800
Authors:Haowei Wang, Jingyi Wang, Zhongxiang Dai, Nai-Yuan Chiang, Szu Hui Ng, Cosmin Petra
Abstract:
Abstract:Constrained Bayesian optimization (CBO) methods have seen significant success in black-box optimization with constraints, and one of the most commonly used CBO methods is the constrained expected improvement (CEI) algorithm. CEI is a natural extension of the expected improvement (EI) when constraints are incorporated. However, the theoretical convergence rate of CEI has not been established. In this work, we study the convergence rate of CEI by analyzing its simple regret upper bound. First, we show that when the objective function $f$ and constraint function $c$ are assumed to each lie in a reproducing kernel Hilbert space (RKHS), CEI achieves the convergence rates of $\mathcal{O} \left(t^{-\frac{1}{2}}\log^{\frac{d+1}{2}}(t) \right) \ \text{and }\ \mathcal{O}\left(t^{\frac{-\nu}{2\nu+d}} \log^{\frac{\nu}{2\nu+d}}(t)\right)$ for the commonly used squared exponential and Matérn kernels, respectively. Second, we show that when $f$ and $c$ are assumed to be sampled from Gaussian processes (GPs), CEI achieves the same convergence rates with a high probability. Numerical experiments are performed to validate the theoretical analysis.



Paperid:801
Authors:Yoav Ger, Omri Barak
Abstract:
Recurrent neural networks (RNNs) trained on neuroscience-inspired tasks offer powerful models of brain computation. However, typical training paradigms rely on open-loop, supervised settings, whereas real-world learning unfolds in closed-loop environments. Here, we develop a mathematical theory describing the learning dynamics of linear RNNs trained in closed-loop contexts. We first demonstrate that two otherwise identical RNNs, trained in either closed- or open-loop modes, follow markedly different learning trajectories. To probe this divergence, we analytically characterize the closed-loop case, revealing distinct stages aligned with the evolution of the training loss. Specifically, we show that the learning dynamics of closed-loop RNNs, in contrast to open-loop ones, are governed by an interplay between two competing objectives: short-term policy improvement and long-term stability of the agent-environment interaction. Finally, we apply our framework to a realistic motor control task, highlighting its broader applicability. Taken together, our results underscore the importance of modeling closed-loop dynamics in a biologically plausible setting.



Paperid:802
Authors:Haoran Sun, Fei Xiong, Yuanzhe Hu, Liang Wang
Abstract:
Collaborative filtering (CF) methods are now facing the challenge of data sparsity in recommender systems. In order to reduce the effect of data sparsity, researchers proposed contrastive learning methods to extract self-supervised signals from raw data. Contrastive learning methods address this problem by graph augmentation and maximizing the consistency of node representations between different augmented graphs. However, these methods tends to unintentionally distance the target node from its path nodes on the interaction path, thus limiting its effectiveness. In this regard, we propose a solution that uses paths as samples in the contrastive loss function. In order to obtain the path samples, we design a path sampling method. In addition to the contrast of the relationship between the target node and the nodes within the path (intra-path contrast), we also designed a method of contrasting the relationship between the paths (inter-path contrast) to better pull the target node and its path nodes closer to each other. We use Simplifying and Powering Graph Convolution Network (LightGCN) as the basis and combine with a new path-enhanced graph approach proposed for graph augmentation. It effectively improves the performance of recommendation models. Our proposed Path Enhanced Contrastive Loss (PECL) model replaces the common contrastive loss function with our novel loss function, showing significant performance improvement. Experiments on two real-world datasets demonstrate the effectiveness of our model.



Paperid:803
Authors:Sungee Hong, Jiayi Wang, Zhengling Qi, Raymond K. W. Wong
Abstract:
In reinforcement learning, distributional off-policy evaluation (OPE) focuses on estimating the return distribution of a target policy using offline data collected under a different policy. This work focuses on extending the widely used fitted-Q evaluation---developed for expectation-based reinforcement learning---to the distributional OPE setting. We refer to this extension as fitted distributional evaluation (FDE). While only a few related approaches exist, there remains no unified framework for designing FDE methods. To fill this gap, we present a set of guiding principles for constructing theoretically sound FDE methods. Building on these principles, we develop several new FDE methods with convergence analysis and provide theoretical justification for existing methods, even in non-tabular environments with infinitely large state-action spaces. Extensive experiments, including simulations on linear quadratic regulators and Atari games, demonstrate the superior performance of the FDE methods.



Paperid:804
Authors:Gil Goldman, Raja Giryes, Mahadev Satyanarayanan
Abstract:
We propose a smooth regularization technique that instillsa strong temporal inductive bias in video recognition models,particularly benefiting lightweight architectures.Our method encourages smoothness in the intermediate-layer embeddings ofconsecutive frames by modeling their changes as a Gaussian Random Walk (GRW).This penalizes abrupt representational shifts,thereby promoting low-acceleration solutions that better alignwith the natural temporal coherence inherent in videos.By leveraging this enforced smoothness, lightweight models can moreeffectively capture complex temporal dynamics.Applied to such models, our technique yields a 2.8%–6.0%accuracy improvement on Kinetics-600.Notably, the MoViNets model family trained with our smoothregularization improves the current state-of-the-art by 2.8%–3.4% withintheir respective FLOP constraints, while MobileNetV3 andthe MoViNets-Stream family achieve gains of 2.8%–6.0% overprior state-of-the-art models with comparable memory footprints.Code and models will be released upon acceptance.



Paperid:805
Authors:Bowei Zhang, Lei Ke, Adam Harley, Katerina Fragkiadaki
Abstract:
We introduce TAPIP3D, a novel approach for long-term 3D point tracking in monocular RGB and RGB-D videos. TAPIP3D represents videos as camera-stabilized spatio-temporal feature clouds, leveraging depth and camera motion information to lift 2D video features into a 3D world space where camera movement is effectively canceled out. Within this stabilized 3D representation, TAPIP3D iteratively refines multi-frame motion estimates, enabling robust point tracking over long time horizons. To handle the irregular structure of 3D point distributions, we propose a 3D Neighborhood-to-Neighborhood (N2N) attention mechanism—a 3D-aware contextualization strategy that builds informative, spatially coherent feature neighborhoods to support precise trajectory estimation. Our 3D-centric formulation significantly improves performance over existing 3D point tracking methods and even surpasses state-of-the-art 2D pixel trackers in accuracy when reliable depth is available. The model supports inference in both camera-centric (unstabilized) and world-centric (stabilized) coordinates, with experiments showing that compensating for camera motion leads to substantial gains in tracking robustness. By replacing the conventional 2D square correlation windows used in prior 2D and 3D trackers with a spatially grounded 3D attention mechanism, TAPIP3D achieves strong and consistent results across multiple 3D point tracking benchmarks. Our code and trained checkpoints will be public.



Paperid:806
Authors:Alessio Ragno, Marc Plantevit, Céline Robardet
Abstract:
Graphs are complex, non-Euclidean structures that require Graph Neural Networks (GNNs) to adapt dynamically to varying neighborhood sizes in order to capture the local structural information essential for learning tasks. This inherent complexity makes explaining GNN decisions particularly challenging. Most existing explainable AI (XAI) methods for GNNs focus on identifying influential nodes or extracting subgraphs that highlight relevant motifs. However, these approaches often fall short of clarifying how such elements contribute to the final prediction. To overcome this limitation, logic-based explanations aim to derive explicit logical rules that reflect the model's decision-making process. Current logic-based methods are limited to post-hoc analyses and are predominantly applied to graph classification, leaving a significant gap in intrinsically explainable GNN architectures. In this paper, we explore the potential of integrating logic reasoning directly into graph learning. We introduce LogiX-GIN, a novel, self-explainable GNN architecture that incorporates logic layers to produce interpretable logical rules as part of the learning process. Unlike post-hoc methods, LogiX-GIN provides faithful, transparent, and inherently interpretable explanations aligned with the model's internal computations. We evaluate LogiX-GIN across several graph-based tasks and show that it achieves competitive predictive performance while delivering clear, logic-based insights into its decision-making process.



Paperid:807
Authors:Maximilien Dreveton, Elaine Liu, Matthias Grossglauser, Patrick Thiran
Abstract:
Abstract:This paper establishes the theoretical limits of graph clustering under the Popularity-Adjusted Block Model (PABM), addressing limitations of existing models. In contrast to the Stochastic Block Model (SBM), which assumes uniform vertex degrees, and to the Degree-Corrected Block Model (DCBM), which applies uniform degree corrections across clusters, PABM introduces separate popularity parameters for intra- and inter-cluster connections. Our main contribution is the characterization of the optimal error rate for clustering under PABM, which provides novel insights on clustering hardness: we demonstrate that unlike SBM and DCBM, cluster recovery remains possible in PABM even when traditional edge-density signals vanish, provided intra- and inter-cluster popularity coefficients differ. This highlights a dimension of degree heterogeneity captured by PABM but overlooked by DCBM: differences in connectivity patterns can enhance cluster separability independently of overall edge densities. Finally, because PABM exhibits a richer structure, its expected adjacency matrix has rank between $k$ and $k^2$, where $k$ is the number of clusters. As a result, spectral embeddings based on only the top $k$ eigenvectors may fail to capture important structural information. Our numerical experiments on both synthetic and real datasets confirm that spectral clustering algorithms incorporating~$k^2$ eigenvectors outperform traditional spectral approaches.



Paperid:808
Authors:Kaiping Zheng, Horng-Ruey Chua, Beng Chin Ooi
Abstract:
Data deviation in electronic health records (EHR) refers to discrepancies between recorded entries and a patient’s actual physiological state, indicating a decline in EHR data fidelity. Such deviation can result from pre-analytical variability, documentation errors, or unvalidated data sources. Effectively detecting data deviation is clinically valuable for identifying erroneous records, excluding them from downstream clinical workflows, and informing corrective actions. Despite its importance and practical relevance, this problem remains largely under-explored in existing research. To bridge this gap, we propose a bi-level knowledge distillation approach centered on a task-agnostic formulation of EHR data fidelity as an intrinsic measure of data reliability. Our approach performs layered knowledge distillation in two stages: from a computation-intensive, task-specific data Shapley oracle to a neural oracle for individual tasks, and then to a unified EHR data fidelity predictor. This design enables the integration of task-specific insights into a holistic assessment of a patient’s EHR data fidelity from a multi-task perspective. By tracking the outputs of this learned predictor, we detect potential data deviation in EHR data. Experiments on both our hospital’s proprietary dataset and the public MIMIC-III dataset consistently validate the effectiveness of our approach in detecting data deviation in EHR data. Case studies further demonstrate its practical value in identifying clinically meaningful data deviation.



Paperid:809
Authors:Zahra Babaiee, Peyman M. Kiasari, Daniela Rus, Radu Grosu
Abstract:
A recent study has proposed the ``Master Key Filters Hypothesis" for convolutional neural network filters. This paper extends this hypothesis by radically constraining its scope to a single set of just 8 universal filters that depthwise separable convolutional networks inherently converge to. While conventional DS-CNNs employ thousands of distinct trained filters, our analysis reveals these filters are predominantly linear shifts (ax+b) of our discovered universal set. Through systematic unsupervised search, we extracted these fundamental patterns across different architectures and datasets. Remarkably, networks initialized with these 8 unique frozen filters achieve over 80\% ImageNet accuracy, and even outperform models with thousands of trainable parameters when applied to smaller datasets. The identified master key filters closely match Difference of Gaussians (DoGs), Gaussians, and their derivatives, structures that are not only fundamental to classical image processing but also strikingly similar to receptive fields in mammalian visual systems. Our findings provide compelling evidence that depthwise convolutional layers naturally gravitate toward this fundamental set of spatial operators regardless of task or architecture. This work offers new insights for understanding generalization and transfer learning through the universal language of these master key filters.



Paperid:810
Authors:Or Goldreich, Ziyang Wei, SOHAM BONNERJEE, Jiaqi Li, Wei Biao Wu
Abstract:
Stochastic gradient descent (SGD) with polynomially decaying step‐sizes has long underpinned theoretical analyses, yielding a broad spectrum of statistically attractive guarantees. Yet in practice, such schedules find rare use due to their prohibitively slow convergence, revealing a persistent gap between theory and empirical performance. In this paper, we introduce a unified framework that quantifies the uncertainty of online SGD under arbitrary learning‐rate choices. In particular, we provide the first comprehensive convergence characterizations for two widely used but theoretically under-examined schemes—cyclical learning rates and linear decay to zero. Our results not only explain the observed behavior of these schedules but also facilitate principled tools for statistical inference and algorithm design. All theoretical findings are corroborated by extensive simulations across diverse settings.



Paperid:811
Authors:Gilad Yehudai, Clayton Sanford, Maya Bechler-Speicher, Orr Fischer, Ran Gilad-Bachrach, Amir Globerson
Abstract:
Transformers have revolutionized the field of machine learning. In particular, they can be used to solve complex algorithmic problems, including graph-based tasks. In such algorithmic tasks a key question is what is the minimal size of a transformer that can implement the task. Recent work has begun to explore this problem for graph-based tasks, showing that for sub-linear embedding dimension (i.e., model width) logarithmic depth suffices. However, an open question, which we address here, is what happens if width is allowed to grow linearly, while depth is kept fixed. Here we analyze this setting, and provide the surprising result that with linear width, constant depth suffices for solving a host of graph-based problems. This suggests that a moderate increase in width can allow much shallower models, which are advantageous in terms of inference and train time. For other problems, we show that quadratic width is required. Our results demonstrate the complex and intriguing landscape of transformer implementations of graph-based algorithms. We empirically investigate these trade-offs between the relative powers of depth and width and find tasks where wider models have the same accuracy as deep models, while having much faster train and inference time due to parallelizable hardware.



Paperid:812
Authors:Sarah Mameche, Janis Kalofolias, Jilles Vreeken
Abstract:
Real-world datasets are often comprised of combinations of unobserved subpopulations with distinct underlying causal processes. In an observational study, for example, patients may fall into unobserved groups that either (a) respond effectively to a drug vs. (b) show no response due to drug resistance. If we do not account for this, we will obtain biased estimates of drug effectiveness. In this work, we formulate such settings as a causal mixture model, where the data-generating process of each variable depends on membership to a certain group (a or b). Specifically, we assume a mixture of structural causal equation models with latent categorical variables indexing subpopulation assignment. Unlike prior work, our framework allows for multiple such latent variables affecting distinct observed variable sets. To infer this model from mixed data sources, we propose a topological ordering-based approach that jointly discovers (i) the causal graph and (ii) the number of mixing variables, number of their components, and assignments. In empirical evaluations, we show that our approach effectively discovers these in practice and that mixed data sources can even enhance the identification of cause-effect relationships.



Paperid:813
Authors:Shaorong Zhang, Yuanbin Cheng, Greg Ver Steeg
Abstract:
Diffusion bridge models and stochastic interpolants enable high-quality image-to-image (I2I) translation by creating paths between distributions in pixel space. However, the proliferation of techniques based on incompatible mathematical assumptions have impeded progress. In this work, we unify and expand the space of bridge models by extending Stochastic Interpolants (SIs) with preconditioning, endpoint conditioning, and an optimized sampling algorithm.These enhancements expand the design space of diffusion bridge models, leading to state-of-the-art performance in both image quality and sampling efficiency across diverse I2I tasks. Furthermore, we identify and address a previously overlooked issue of low sample diversity under fixed conditions. We introduce a quantitative analysis for output diversity and demonstrate how we can modify the base distribution for further improvements.



Paperid:814
Authors:Jiayi Zhou, Jiaming Ji, Boyuan Chen, Jiapeng Sun, wenqi chen, Donghai Hong, Sirui Han, Yike Guo, Yaodong Yang
Abstract:
Training multi-modal large language models (MLLMs) that align with human intentions is a long-term challenge. Traditional score-only reward models for alignment suffer from low accuracy, weak generalization, and poor interpretability, blocking the progress of alignment methods, \textit{e.g.,} reinforcement learning from human feedback (RLHF). Generative reward models (GRMs) leverage MLLMs' intrinsic reasoning capabilities to discriminate pair-wise responses, but their pair-wise paradigm makes it hard to generalize to learnable rewards. We introduce Generative RLHF-V, a novel alignment framework that integrates GRMs with multi-modal RLHF. We propose a two-stage pipeline: \textbf{multi-modal generative reward modeling from RL}, where RL guides GRMs to actively capture human intention, then predict the correct pair-wise scores; and \textbf{RL optimization from grouped comparison}, which enhances multi-modal RL scoring precision by grouped responses comparison. Experimental results demonstrate that, besides out-of-distribution generalization of RM discrimination, our framework improves 4 MLLMs' performance across 7 benchmarks by 18.1\%, while the baseline RLHF is only 5.3\%. We further validate that Generative RLHF-V achieves a near-linear improvement with an increasing number of candidate responses.



Paperid:815
Authors:Arman Behnam, Binghui Wang
Abstract:
Causal representation learning for the anti-causal setting—labels cause features rather than the reverse—presents unique challenges that require specialized approaches. This paper introduces a novel measure-theoretic framework called anti-causal invariant abstractions (ACIA) that effectively handles both perfect and imperfect interventions. ACIA learns two-level representations where low-level representations capture causal dynamics from input data, while high-level abstractions learn environment-invariant features. ACIA addresses key limitations of existing approaches by: (1) accommodating imperfect interventions through interventional kernels, (2) eliminating dependency on explicit causal structures, (3) handling high-dimensional data effectively, and (4) providing theoretical guarantees for out-of-distribution generalization. Experiments on synthetic and real-world medical datasets demonstrate that our ACIA consistently outperforms state-of-the-art methods across accuracy and invariance metrics. Our theoretical results establish tight bounds on performance gaps between training and unseen environments, confirming the efficacy of our approach for robust anti-causal learning.



Paperid:816
Authors:Noboru Isobe, Kenshi Abe, Kaito Ariu
Abstract:
Abstract:In the Lasry--Lions framework, Mean-Field Games (MFGs) model interactions among an infinite number of agents.However, existing algorithms either require strict monotonicity or only guarantee the convergence of averaged iterates, as in Fictitious Play in continuous time.We address this gap with the following theoretical result.First, we prove that the last-iterated policy of a proximal-point (PP) update with KL regularization converges to an MFG equilibrium under non-strict monotonicity.Second, we see that each PP update is equivalent to finding the equilibria of a KL-regularized MFG.We then prove that this equilibrium can be found using mirror descent with an exponential last-iterate convergence rate.Building on these insights, we propose the Approximate Proximal-Point ($\mathtt{APP}$) algorithm, which approximately implements the PP update via a small number of Mirror Descent steps.Numerical experiments on standard benchmarks confirm that the $\mathtt{APP}$ algorithm reliably converges to the unregularized mean-field equilibrium without time-averaging.



Paperid:817
Authors:Katharina Limbeck, Lydia Mezrag, Guy Wolf, Bastian Rieck
Abstract:
Graph Neural Networks (GNNs) have shown significant success for graph-based tasks. Motivated by the prevalence of large datasets in real-world applications, pooling layers are crucial components of GNNs. By reducing the size of input graphs, pooling enables faster training and potentially better generalisation. However, existing pooling operations often optimise for the learning task at the expense of fundamental graph structures and interpretability. This leads to unreliable performance across varying dataset types, downstream tasks and pooling ratios. Addressing these concerns, we propose novel graph pooling layers for structure aware pooling via edge collapses. Our methods leverage diffusion geometry and iteratively reduce a graph’s size while preserving both its metric structure and structural diversity. We guide pooling usingmagnitude, an isometry-invariant diversity measure, which permits us to control the fidelity of the pooling process. Further, we use thespreadof a metric space as a faster and more stable alternative ensuring computational efficiency. Empirical results demonstrate that our methods (i) achieve superior performance compared to alternative pooling layers across a range of diverse graph classification tasks, (ii) preserve key spectral properties of the input graphs, and (iii) retain high accuracy across varying pooling ratios.



Paperid:818
Authors:Ionel-Alexandru Hosu, Traian Rebedea, Razvan Pascanu
Abstract:
One proposed mechanism to improve exploration in reinforcement learning is the use of macro-actions, a form of temporal abstractions over actions.Paradoxically though, in many scenarios the naive addition of macro-actions does not lead to better exploration, but rather the opposite. In this work, we argue that the difficulty stems from the trade-offs between reducing the average number of decisions per episode versus increasing the size of the action space. Namely, one typically treats each potential macro-action as independent and atomic, hence strictly increasing the search space and making typical exploration strategies inefficient. To address this problem we propose a novel regularization term that exploits the relationship between actions and macro-actions to improve the credit assignment mechanism reducing the effective dimension of the action space and therefore improving exploration. The term relies on a similarity matrix that is meta-learned jointly with learning the desired policy.We empirically validate our strategy looking at macro-actions in Atari games, and the StreetFighter II environment. Our results show significant improvements over the Rainbow-DQN baseline in all environments. Additionally, we show that the macro-action similarity is transferable to other environments with similar dynamics.We believe this work is a small but important step towards understanding how the similarity-imposed geometry on the action space can be exploited to improve credit assignment and exploration, therefore making learning more efficient.



Paperid:819
Authors:Zakaria Mhammedi
Abstract:
Abstract:We study the problem of Online Convex Optimization (OCO) over a convex set $\mathcal{K} \subset \mathbb{R}^d$, accessed via a separation oracle. While classical projection-based algorithms such as projected Online Gradient Descent (OGD) achieve the optimal $O(\sqrt{T})$ regret, they require computing Euclidean projections onto $\mathcal{K}$ whenever an iterate falls outside the feasible set. These projections can be computationally expensive, especially for complex or high-dimensional sets. Projection-free algorithms address this by replacing projections with alternative oracle-based procedures, such as separation or linear optimization oracles. However, the regret bounds of existing separation-based methods scale poorly with the set's \emph{asphericity} $\kappa$, defined as the ratio between the radii of the smallest enclosing ball and the largest inscribed ball in $\mathcal{K}$; for ill-conditioned sets, $\kappa$ can be arbitrarily large.We introduce a new separation-based algorithm for OCO that achieves a regret bound of $\tilde{O}(\sqrt{dT} + d^2)$, with only logarithmic dependence on $\kappa$. This removes a key limitation of prior work and eliminates the need for costly geometric pre-processing, such as transforming $\mathcal{K}$ into isotropic position. Our algorithm is based on a novel reduction to online optimization over a sequence of dynamically updated ellipsoids, inspired by the classical ellipsoid method for convex optimization. It requires only $\tilde{O}(1)$ separation oracle calls per round, on par with existing separation-based approaches. These advances make our method particularly well suited for online optimization over geometrically complex feasible sets.



Paperid:820
Authors:Alexander Theus, Alessandro Cabodi, Sotiris Anagnostidis, Antonio Orvieto, Sidak Pal Singh, Valentina Boeva
Abstract:
Abstract:Understanding the geometry of neural network loss landscapes is a central question in deep learning, with implications for generalization and optimization. A striking phenomenon is $\textit{linear mode connectivity}$ (LMC), where independently trained models can be connected by low- or zero-loss paths, despite appearing to lie in separate basins. However, this is often obscured by symmetries in parameter space—such as neuron permutations—which make functionally equivalent models appear dissimilar. Prior work has predominantly focused on neuron re-ordering through permutations, but such approaches are limited in scope and fail to capture the richer symmetries exhibited by modern architectures such as Transformers. In this work, we introduce a unified framework that captures four symmetry classes—permutations, semi-permutations, orthogonal transformations, and general invertible maps—broadening the set of valid reparameterizations and subsuming many previous approaches as special cases. Crucially, this generalization enables, for the first time, the discovery of low- and zero-barrier linear interpolation paths between independently trained Vision Transformers and GPT-2 models. These results reveal deeper structure in the loss landscape and underscore the importance of symmetry-aware analysis for understanding model space geometry.



Paperid:821
Authors:Yichao Fu, Rui Ge, Zelei Shao, Zhijie Deng, Hao Zhang
Abstract:
Abstract:Reasoning models excel by generating long chain-of-thoughts, but decoding the resulting thousands of tokens is slow. Token-level specualtive decoding (SD) helps, but its benefit is capped, because the chance that an entire $\gamma$-token guess is correct falls exponentially as $\gamma$ grows. This means allocating more compute for longer token drafts faces an algorithmic ceiling -- making the speedup modest and hardware-agnostic. We raise this ceiling with lookahead reasoning, which exploits a second, step-level layer of parallelism. Our key insight is that reasoning models generate step-by-step, and each step needs only to be semantically correct, not exact token matching. In lookahead reasoning, a lightweight draft model proposes several future steps; the target model expands each proposal in one batched pass, and a verifier keeps semantically correct steps while letting the target regenerate any that fail. Token-level SD still operates within each reasoning step, so the two layers of parallelism multiply. We show lookahead reasoning lifts the peak speedup of SD both theoretically and empirically. Across GSM8K, AIME, and other benchmarks, lookahead reasoning improves the speedup of SD from 1.4x to 2.1x while preserving answer quality, and its speedup scales better with additional GPU throughput.



Paperid:822
Authors:Annan Yu, N. Benjamin Erichson
Abstract:
Abstract:Mamba extends earlier state space models (SSMs) by introducing input-dependent dynamics, and has demonstrated strong empirical performance across a range of domains, including language modeling, computer vision, and foundation models. However, a surprising weakness remains: despite being built on architectures designed for long-range dependencies, Mamba performs poorly on long-range sequential tasks. Understanding and addressing this gap is important for improving Mamba's universality and versatility. In this work, we analyze Mamba’s limitations through three perspectives: expressiveness, inductive bias, and training stability. Our theoretical results show how Mamba falls short in each of these aspects compared to earlier SSMs such as S4D. To address these issues, we propose $\text{B}\_{2}\text{S}\_{6}$, a simple extension of Mamba's S6 unit that combines block-wise selective dynamics with a channel-specific bias. We prove that these changes equip the model with a better-suited inductive bias and improve its expressiveness and stability. Empirically, $\text{B}\_{2}\text{S}\_{6}$ outperforms S4 and S4D on Long-Range Arena (LRA) tasks while maintaining Mamba's performance on language modeling benchmarks.



Paperid:823
Authors:Daniel Halpern, Evi Micha, Ariel Procaccia, Itai Shapira
Abstract:
Current alignment pipelines presume a single, universal notion of desirable behavior. However, human preferences often diverge across users, contexts, and cultures. As a result, disagreement collapses into the majority signal and minority perspectives are discounted. To address this, we propose reflecting diverse human preferences through a distribution over multiple reward functions, each inducing a distinct aligned policy. The distribution is learned directly from pairwise preference without annotator identifiers or predefined groups. Instead, annotator disagreements are treated as informative soft labels. Our central criterion is \emph{pairwise calibration}: for every pair of candidate responses, the proportion of reward functions preferring one response matches the fraction of annotators with that preference. We prove that even a small outlier-free ensemble can accurately represent diverse preference distributions. Empirically, we introduce and validate a practical training heuristic to learn such ensembles, and demonstrate its effectiveness through improved calibration, implying a more faithful representation of pluralistic values.



Paperid:824
Authors:Julien Chhor, Parker Knight
Abstract:
Abstract:We consider the problem of detecting a community of densely connected vertices in a high-dimensional bipartite graph of size $n_1 \times n_2$. Under the null hypothesis, the observed graph is drawn from a bipartite Erdos-Renyi distribution with connection probability $p_0$. Under the alternative hypothesis, there exists an unknown bipartite subgraph of size $k_1 \times k_2$ in which edges appear with probability $p_1 = p_0 + \delta$ for some $\delta > 0$, while all other edges outside the subgraph appear with probability $p_0$. Specifically, we provide non-asymptotic upper and lower bounds on the smallest signal strength $\delta^*$ that is both necessary and sufficient to ensure the existence of a test with small enough type one and type two errors. We also derive novel minimax-optimal tests achieving these fundamental limits when the underlying graph is sufficiently dense. Our proposed tests involve a combination of hard-thresholded nonlinear statistics of the adjacency matrix, the analysis of which may be of independent interest. In contrast with previous work, our non-asymptotic upper and lower bounds match for any configuration of $n_1,n_2, k_1,k_2$.



Paperid:825
Authors:Chengyuan Deng, Jie Gao, Kevin Lu, Feng Luo, Cheng Xin
Abstract:
Abstract:The Johnson-Lindenstrauss (JL) lemma is a cornerstone of dimensionality reduction in Euclidean space, but its applicability to non-Euclidean data has remained limited. This paper extends the JL lemma beyond Euclidean geometry to handle general dissimilarity matrices that are prevalent in real-world applications. We present two complementary approaches: First, we show how the JL transform can be applied to vectors in pseudo-Euclidean space with signature $(p,q)$, providing theoretical guarantees that depend on the ratio of the $(p, q)$ norm and Euclidean norm of two vectors, measuring the deviation from Euclidean geometry. Second, we prove that any symmetric hollow dissimilarity matrix can be represented as a matrix of generalized power distances, with an additional parameter representing the uncertainty level within the data. In this representation, applying the JL transform yields multiplicative approximation with a controlled additive error term proportional to the deviation from Euclidean geometry. Our theoretical results provide fine-grained performance analysis based on the degree to which the input data deviates from Euclidean geometry, making practical and meaningful reduction in dimensionality accessible to a wider class of data. We validate our approaches on both synthetic and real-world datasets, demonstrating the effectiveness of extending the JL lemma to non-Euclidean settings.



Paperid:826
Authors:Sang Truong, Yuheng Tu, Michael Hardy, Anka Reuel-Lamparth, Zeyu Tang, Jirayu Burapacheep, Jonathan Perera, Chibuike Uwakwe, Benjamin Domingue, Nick Haber, Sanmi Koyejo
Abstract:
Benchmarks are pivotal in driving progress in large language models, yet ambiguous questions, incorrect answer keys, and grading issues frequently undermine their reliability. Manually identifying and fixing issues among thousands of benchmark questions is not only infeasible but also a critical bottleneck for reliable evaluation. In this work, we introduce a scalable, theory-driven framework for systematic benchmark revision that leverages psychometric tools to flag problematic questions requiring expert review. We demonstrate that the No Free Lunch theorem applies directly to benchmark quality assessment: no detector can excel across all anomaly patterns, and effective detection requires prior anomaly knowledge. Furthermore, recognizing the high cost of LLM evaluations and the limited diversity of available LLMs, we assess each tool’s sensitivity to the number of model responses. Finally, across nine widely used benchmarks, our signals guide expert review to identify flawed questions with up to 84\% precision, offering an efficient, scalable framework for systematic benchmark revision.



Paperid:827
Authors:Chao Huang, Yuesheng Ma, Junxuan Huang, Susan Liang, Yunlong Tang, Jing Bi, Wenqiang Liu, Nima Mesgarani, Chenliang Xu
Abstract:
Audio source separation is fundamental for machines to understand complex acoustic environments and underpins numerous audio applications. Current supervised deep learning approaches, while powerful, are limited by the need for extensive, task-specific labeled data and struggle to generalize to the immense variability and open-set nature of real-world acoustic scenes. Inspired by the success of generative foundation models, we investigate whether pre-trained text-guided audio diffusion models can overcome these limitations. We make a surprising discovery: zero-shot source separation can be achieved purely through a pre-trained text-guided audio diffusion model under the right configuration. Our method, named ZeroSep, works by inverting the mixed audio into the diffusion model's latent space and then using text conditioning to guide the denoising process to recover individual sources. Without any task-specific training or fine-tuning, ZeroSep repurposes the generative diffusion model for a discriminative separation task and inherently supports open-set scenarios through its rich textual priors. ZeroSep is compatible with a variety of pre-trained text-guided audio diffusion backbones and delivers strong separation performance on multiple separation benchmarks, surpassing even supervised methods.



Paperid:828
Authors:Jie Liu, Gongye Liu, Jiajun Liang, Yangguang Li, Jiaheng Liu, Xintao Wang, Pengfei Wan, Di ZHANG, Wanli Ouyang
Abstract:
Abstract:We propose Flow-GRPO, the first method to integrate online reinforcement learning (RL) into flow matching models. Our approach uses two key strategies: (1) an ODE-to-SDE conversion that transforms a deterministic Ordinary Differential Equation (ODE) into an equivalent Stochastic Differential Equation (SDE) that matches the original model's marginal distribution at all timesteps, enabling statistical sampling for RL exploration; and (2) a Denoising Reduction strategy that reduces training denoising steps while retaining the original number of inference steps, significantly improving sampling efficiency without sacrificing performance. Empirically, Flow-GRPO is effective across multiple text-to-image tasks. For compositional generation, RL-tuned SD3.5-M generates nearly perfect object counts, spatial relations, and fine-grained attributes, increasing GenEval accuracy from $63$\% to $95$\%. In visual text rendering, accuracy improves from $59$\% to $92$\%, greatly enhancing text generation. Flow-GRPO also achieves substantial gains in human preference alignment. Notably, very little reward hacking occurred, meaning rewards did not increase at the cost of appreciable image quality or diversity degradation.



Paperid:829
Authors:Bariscan Bozkurt, Houssam Zenati, Dimitri Meunier, Liyuan Xu, Arthur Gretton
Abstract:
We study the problem of causal function estimation in the Proxy Causal Learning (PCL) framework, where confounders are not observed but proxies for the confounders are available. Two main approaches have been proposed: outcome bridge-based and treatment bridge-based methods. In this work, we propose two kernel-based doubly robust estimators that combine the strengths of both approaches, and naturally handle continuous and high-dimensional variables. Our identification strategy builds on a recent density ratio-free method for treatment bridge-based PCL; furthermore, in contrast to previous approaches, it does not require indicator functions or kernel smoothing over the treatment variable. These properties make it especially well-suited for continuous or high-dimensional treatments. By using kernel mean embeddings, we have closed-form solutions and strong consistency guarantees. Our estimators outperform existing methods on PCL benchmarks, including a prior doubly robust method that requires both kernel smoothing and density ratio estimation.



Paperid:830
Authors:Shauli Ravfogel, Gilad Yehudai, Tal Linzen, Joan Bruna, Alberto Bietti
Abstract:
Recent probing studies reveal that large language models exhibit linear subspaces that separate true from false statements, yet the mechanism behind their emergence is unclear. We introduce a transparent, one-layer transformer toy model that reproduces such truth subspaces end-to-end and exposes one concrete route by which they can arise. We study one simple setting in which truth encoding can emerge: a data distribution where factual statements co-occur with other factual statements (and vice-versa), encouraging the model to learn this distinction in order to lower the LM loss on future tokens. We corroborate this pattern with experiments in pretrained language models. Finally, in the toy setting we observe a two-phase learning dynamic: networks first memorize individual factual associations in a few steps, then---over a longer horizon---learn to linearly separate true from false, which in turn lowers language-modeling loss. Together, these results provide both a mechanistic demonstration and an empirical motivation for how and why linear truth representations can emerge in language models.



Paperid:831
Authors:BOWEN FAN, Yuming Ai, Xunkai Li, Zhilin Guo, LEI ZHU, Guang Zeng, Rong-Hua Li, Guoren Wang
Abstract:
Abstract:Graph Machine Learning is essential for understanding and analyzing relational data. However, privacy-sensitive applications demand the ability to efficiently remove sensitive information from trained graph neural networks (GNNs), avoiding the unnecessary time and space overhead caused by retraining models from scratch.To address this issue, Graph Unlearning (GU) has emerged as a critical solution to support dynamic graph updates while ensuring privacy compliance. Unlike machine unlearning in computer vision or other fields, GU faces unique difficulties due to the non-Euclidean nature of graph data and the recursive message-passing mechanism of GNNs. Additionally, the diversity of downstream tasks and the complexity of unlearning requests further amplify these challenges. Despite the proliferation of diverse GU strategies, the absence of a benchmark providing fair comparisons for GU, and the limited flexibility in combining downstream tasks and unlearning requests, have yielded inconsistencies in evaluations, hindering the development of this domain. To fill this gap, we present OpenGU, the first GU benchmark, where 16 SOTA GU algorithms and 37 multi-domain datasets are integrated, enabling various downstream tasks with 13 GNN backbones when responding to flexible unlearning requests. Through extensive experimentation, we have drawn $10$ crucial conclusions about existing GU methods, while also gaining valuable insights into their limitations, shedding light on potential avenues for future research. Our code is available at \href{https://anonymous.4open.science/r/OpenGU-2025}{https://anonymous.4open.science/r/OpenGU-2025}.



Paperid:832
Authors:Zhongyi Zhou, Yichen Zhu, Xiaoyu Liu, Zhibin Tang, Junjie Wen, Yaxin Peng, Chaomin Shen, Yi Xu
Abstract:
Abstract:Vision-language-action (VLA) models have emerged as the next generation of models in robotics. However, despite leveraging powerful pre-trained Vision-Language Models (VLMs), existing end-to-end VLA systems often lose key capabilities during fine-tuning as the model adapts to specific robotic tasks. We argue that a generalizable VLA model should retain and expand upon the VLM's core competencies: 1) **Open-world reasoning** - the VLA should inherit the knowledge from VLM, i.e., recognize anything that the VLM can recognize, capable of solving math problems, possessing visual-spatial intelligence, 2) **Reasoning following** – effectively translating the open-world reasoning into actionable steps for the robot. In this work, we introduce **ChatVLA-2**, a novel mixture-of-expert VLA model coupled with a specialized three-stage training pipeline designed to preserve the VLM’s original strengths while enabling actionable reasoning. To validate our approach, we design a math-matching task wherein a robot interprets math problems written on a whiteboard and picks corresponding number cards from a table to solve equations. Remarkably, our method exhibits exceptional mathematical reasoning and OCR capabilities, despite these abilities not being explicitly trained within the VLA. Furthermore, we demonstrate that the VLA possesses strong spatial reasoning skills, enabling it to interpret novel directional instructions involving previously unseen objects. Overall, our method showcases reasoning and comprehension abilities that significantly surpass state-of-the-art imitation learning methods such as OpenVLA, DexVLA, and $\pi_0$. This work represents a substantial advancement toward developing truly generalizable robotic foundation models endowed with robust reasoning capacities.



Paperid:833
Authors:Miruna Oprescu, Brian Cho, Nathan Kallus
Abstract:
We study the problem of estimating the average treatment effect (ATE) in adaptive experiments where treatment can only be encouraged—rather than directly assigned—via a binary instrumental variable. Building on semiparametric efficiency theory, we derive the efficiency bound for ATE estimation under arbitrary, history-dependent instrument-assignment policies, and show it is minimized by a variance-aware allocation rule that balances outcome noise and compliance variability. Leveraging this insight, we introduce AMRIV—anAdaptive,Multiply-Robust estimator forInstrumental-Variable settings with variance-optimal assignment. AMRIV pairs (i) an online policy that adaptively approximates the optimal allocation with (ii) a sequential, influence-function–based estimator that attains the semiparametric efficiency bound while retaining multiply-robust consistency. We establish asymptotic normality, explicit convergence rates, and anytime-valid asymptotic confidence sequences that enable sequential inference. Finally, we demonstrate the practical effectiveness of our approach through empirical studies, showing that adaptive instrument assignment, when combined with the AMRIV estimator, yields improved efficiency and robustness compared to existing baselines.



Paperid:834
Authors:Ivan Marisca, Jacob Bamberger, Cesare Alippi, Michael Bronstein
Abstract:
Graph Neural Networks (GNNs) have achieved remarkable success across various domains. However, recent theoretical advances have identified fundamental limitations in their information propagation capabilities, such as over-squashing, where distant nodes fail to effectively exchange information. While extensively studied in static contexts, this issue remains unexplored in Spatiotemporal GNNs (STGNNs), which process sequences associated with graph nodes. Nonetheless, the temporal dimension amplifies this challenge by increasing the information that must be propagated. In this work, we formalize the spatiotemporal over-squashing problem and demonstrate its distinct characteristics compared to the static case. Our analysis reveals that counterintuitively, convolutional STGNNs favor information propagation from points temporally distant rather than close in time. Moreover, we prove that architectures that follow either time-and-space or time-then-space processing paradigms are equally affected by this phenomenon, providing theoretical justification for computationally efficient implementations. We validate our findings on synthetic and real-world datasets, providing deeper insights into their operational dynamics and principled guidance for more effective designs.



Paperid:835
Authors:Richa Rastogi, Yuta Saito, Thorsten Joachims
Abstract:
The feedback that AI systems (e.g., recommender systems, chatbots) collect from user interactions is a crucial source of training data. While short-term feedback (e.g., clicks, engagement) is widely used for training, there is ample evidence that optimizing short-term feedback does not necessarily achieve the desired long-term objectives. Unfortunately, directly optimizing for long-term objectives is challenging, and we identify the disconnect in the timescales of short-term interventions (e.g., rankings) and the long-term feedback (e.g., user retention) as one of the key obstacles. To overcome this disconnect, we introduce the framework of MultiScale Policy Learning to contextually reconcile that AI systems need to act and optimize feedback at multiple interdependent timescales. Following a PAC-Bayes motivation, we show how the lower timescales with more plentiful data can provide a data-dependent hierarchical prior for faster learning at higher scales, where data is more scarce. As a result, the policies at all levels effectively optimize for the long-term. We instantiate the framework with MultiScale Off-Policy Bandit Learning (MSBL) and demonstrate its effectiveness on three tasks relating to recommender and conversational systems.



Paperid:836
Authors:Namhoon Kim, Sara Fridovich-Keil
Abstract:
Implicit Neural Representations (INRs) have recently shown impressive results, but their fundamental capacity, implicit biases, and scaling behavior are poorly understood. We investigate the performance of diverse INRs across a suite of 2D and 3D real and synthetic signals with varying effective bandwidth, as well as both overfitting and generalization tasks including tomography, super-resolution, and denoising. By stratifying performance according to model size as well as signal type and bandwidth, our results shed light on how different INR and grid representations allocate their capacity. We find that, for most tasks and signals, a simple regularized grid with interpolation trains faster and to higher quality than any INR with the same number of parameters. We also find limited settings--namely fitting binary signals such as shape contours--where INRs outperform grids, to guide future use of INRs towards the most advantageous applications.



Paperid:837
Authors:Tong Bu, Xinyu Shi, Zhaofei Yu
Abstract:
While sparse coding plays an important role in promoting the efficiency of biological neural systems, it has not been fully utilized by artificial models as the activation sparsity is not well suited to the current structure of deep networks. Spiking Neural Networks (SNNs), with their event-driven characteristics, offer a more natural platform for leveraging activation sparsity. In this work, we specifically target the reduction of neuronal activity, which directly leads to lower computational cost and facilitates efficient SNN deployment on Neuromorphic hardware. We begin by analyzing the limitations of existing activity regularization methods and identifying critical challenges in training sparse SNNs. To address these issues, we propose a modified neuron model, AT-LIF, coupled with a threshold adaptation technique that stabilizes training and effectively suppresses spike activity. Through extensive experiments on multiple datasets, we demonstrate that our approach achieves significant reductions in average firing rates and synaptic operations without sacrificing much accuracy. Furthermore, we show that our method complements weight-based pruning techniques and successfully trains an SNN with only 0.06 average firing rate and 2.22M parameters on ImageNet, highlighting its potential for building highly efficient and scalable SNN models.



Paperid:838
Authors:Yunshui Li, Yiyuan Ma, Shen Yan, Chaoyi Zhang, Jing Liu, Jianqiao Lu, Ziwen Xu, Mengzhao Chen, Minrui Wang, Shiyi Zhan, Jin Ma, Xunhao Lai, Yao Luo, Xingyan Bin, Hongbin Ren, Mingji Han, Wenhao Hao, Bairen Yi, LingJun Liu, Bole Ma, Xiaoying Jia, zhou Xun, liang xiang, Yonghui Wu
Abstract:
Model merging has emerged as a promising technique for enhancing large language models, though its application in large-scale pre-training remains relatively unexplored. In this paper, we present a comprehensive investigation of model merging techniques during the pre-training process. Through extensive experiments with both dense and Mixture-of-Experts (MoE) architectures ranging from millions to over 100 billion parameters, we demonstrate that merging checkpoints trained with constant learning rates not only achieves significant performance improvements but also enables accurate prediction of annealing behavior. These improvements lead to both more efficient model development and significantly lower training costs. Our detailed ablation studies on merging strategies and hyperparameters provide new insights into the underlying mechanisms while uncovering novel applications. Through comprehensive experimental analysis, we offer the open-source community practical pre-training guidelines for effective model merging.



Paperid:839
Authors:Gorjan Radevski, Teodora Popordanoska, Matthew Blaschko, Tinne Tuytelaars
Abstract:
Audio-visual understanding is a rapidly evolving field that seeks to integrate and interpret information from both auditory and visual modalities. Despite recent advances in multi-modal learning, existing benchmarks often suffer from strong visual bias -- when answers can be inferred from visual data alone -- and provide only aggregate scores that conflate multiple sources of error. This makes it difficult to determine whether models struggle with visual understanding, audio interpretation, or audio-visual alignment. In this work, we introduce DAVE: Diagnostic Audio Visual Evaluation), a novel benchmark dataset designed to systematically evaluate audio-visual models across controlled settings. DAVE alleviates existing limitations by (i) ensuring both modalities are necessary to answer correctly and (ii) decoupling evaluation into atomic subcategories. Our detailed analysis of state-of-the-art models reveals specific failure modes and provides targeted insights for improvement. By offering this standardized diagnostic framework, we aim to facilitate more robust development of audio-visual models.Dataset: https://huggingface.co/datasets/gorjanradevski/daveCode: https://github.com/gorjanradevski/dave



Paperid:840
Authors:Rajesh Mangannavar, Stefan Lee, Alan Fern, Prasad Tadepalli
Abstract:
We propose a novel approach to learn relational policies for classical planning based on learning to rank actions. We introduce a new graph representation that explicitly captures action information and propose a Graph Neural Network (GNN) architecture augmented with Gated Recurrent Units (GRUs) to learn action rankings. Unlike value-function based approaches that must learn a globally consistent function, our action ranking method only needs to learn locally consistent ranking, which is more sample-efficient. Our model is trained on data generated from small problem instances that are easily solved by planners and is applied to significantly larger instances where planning is computationally prohibitive. Experimental results across standard planning benchmarks demonstrate that our action-ranking approach not only achieves better generalization to larger problems than those used in training but also outperforms multiple baseline (value function and action ranking) methods in terms of success rate and plan quality.



Paperid:841
Authors:Nakul Upadhya, Eldan Cohen
Abstract:
Abstract:Decision trees are prized for their interpretability and strong performance on tabular data. Yet, their reliance on simple axis‑aligned linear splits often forces deep, complex structures to capture non‑linear feature effects, undermining human comprehension of the constructed tree. To address this limitation, we propose a novel generalization of a decision tree, the Shape Generalized Tree (SGT), in which each internal node applies a learnable axis‑aligned shape function to a single feature, enabling rich, non‑linear partitioning in one split. As users can easily visualize each node's shape functions, SGTs are inherently interpretable and provide intuitive, visual explanations of the model's decision mechanisms. To learn SGTs from data, we propose ShapeCART, an efficient induction algorithm for SGTs. We further extend the SGT framework to bivariate shape functions (S$^2$GT) and multi‑way trees (SGT$_K$), and present Shape$^2$CART and ShapeCART$_K$, extensions to ShapeCART for learning S$^2$GTs and SGT$_K$s, respectively. Experiments on various datasets show that SGTs achieve superior performance with reduced model size compared to traditional axis-aligned linear trees.



Paperid:842
Authors:Chen Lu, Ke Xue, Lei Yuan, Yao Wang, Yaoyuan Wang, Sheng Fu, Chao Qian
Abstract:
The performance of an algorithm often critically depends on its hyperparameter configuration. Dynamic Algorithm Configuration (DAC) is a recent trend in automated machine learning, which can dynamically adjust the algorithm’s configuration during the execution process and relieves users from tedious trial-and-error tuning tasks. Recently, Multi-Agent Reinforcement Learning (MARL) approaches have improved the configuration of multiple heterogeneous hyperparameters, making various parameter configurations for complex algorithms possible. However, many complex algorithms have inherent inter-dependencies among multiple parameters (e.g., determining the operator type first and then the operator's parameter), which are, however, not considered in previous approaches, thus leading to sub-optimal results. In this paper, we propose the Sequential Multi-Agent DAC (Seq-MADAC) framework to address this issue by considering the inherent inter-dependencies of multiple parameters. Specifically, we propose a sequential advantage decomposition network, which can leverage action-order information through sequential advantage decomposition. Experiments from synthetic functions to the configuration of multi-objective optimization algorithms demonstrate Seq-MADAC's superior performance over state-of-the-art MARL methods and show strong generalization across problem classes. Seq-MADAC establishes a new paradigm for the widespread dependency-aware automated algorithm configuration. Our code is available in the supplemental file.



Paperid:843
Authors:Lingyi Wang, Rashed Shelim, Walid Saad, Naren Ramakrishnan
Abstract:
Imagination in world models is crucial for enabling agents to learn long-horizon policy in a sample-efficient manner. Existing recurrent state-space model (RSSM)-based world models depend on single-step statistical inference to capture the environment dynamics, and, hence, they are unable to perform long-term imagination tasks due to the accumulation of prediction errors. Inspired by the dual-process theory of human cognition, we propose a novel dual-mind world model (DMWM) framework that integrates logical reasoning to enable imagination with logical consistency. DMWM is composed of two components: an RSSM-based System 1 (RSSM-S1) component that handles state transitions in an intuitive manner and a logic-integrated neural network-based System 2 (LINN-S2) component that guides the imagination process through hierarchical deep logical reasoning. The inter-system feedback mechanism is designed to ensure that the imagination process follows the logical rules of the real environment. The proposed framework is evaluated on benchmark tasks that require long-term planning from the DMControl suite and robotic environment. Extensive experimental results demonstrate that the proposed framework yields significant improvements in terms of logical coherence, trial efficiency, data efficiency and long-term imagination over the state-of-the-art world models.



Paperid:844
Authors:Xin Li, Xiaotao Zheng, Zhihong Xia
Abstract:
We study a shallow variant of XNet, a neural architecture with activation functions derived from the Cauchy integral formula. Prior work applied these activations in deep models; we show that even a single-layer XNet achieves near-exponential approximation rates—surpassing the polynomial limits of MLPs and spline-based networks such as Kolmogorov–Arnold Networks (KANs).Empirically, XNet reduces approximation error by over 600× on discontinuous and high-dimensional functions, achieves 10,000× lower residuals in physics-informed PDEs, and improves policy accuracy and sample efficiency on PPO tasks.These results highlight that expressivity can emerge from principled activation design rather than depth. XNet offers a compact, theoretically grounded alternative for approximation, scientific computing, and control.



Paperid:845
Authors:Yaozhong Shi, Zachary Ross, Domniki Asimaki, Kamyar Azizzadenesheli
Abstract:
Expanding on neural operators, we propose a novel framework for stochastic process learning across arbitrary domains. In particular, we develop operator flow matching (OFM) for learning stochastic process priors on function spaces. OFM provides the probability density of the values of any collection of points and enables mathematically tractable functional regression at new points with mean and density estimation. Our method outperforms state-of-the-art models in stochastic process learning, functional regression, and prior learning.



Paperid:846
Authors:Walter Simoncini, Michael Dorkenwald, Tijmen Blankevoort, Cees Snoek, Yuki Asano
Abstract:
Vision foundation models achieve remarkable performance but are only available in a limited set of pre-determined sizes, forcing sub-optimal deployment choices under real-world constraints. We introduce a novel post-pretraining structured pruning method that enables elastic inference across a continuum of compute budgets. Our approach combines gradient information with cross-network structure correlations, efficiently approximated through an evolutionary algorithm, does not require labeled data, generalizes to models without a classification head, and is retraining free. Experiments on DINO and AugReg models demonstrate superior performance over state of the art methods across various sparsities, requiring less than five minutes on a A100 GPU to generate elastic models that can be adjusted to any computational budget. Our key contributions include an efficient pruning strategy for pretrained Vision Transformers, a novel evolutionary approximation of Hessian off-diagonal structures, and a self-supervised importance scoring mechanism that maintains strong performance without requiring retraining nor labels.



Paperid:847
Authors:Haimo Fang, Kevin Tan, Giles Hooker
Abstract:
Gradient boosting is widely popular due to its flexibility and predictive accuracy. However, statistical inference and uncertainty quantification for gradient boosting remain challenging and under-explored. We propose a unified framework for statistical inference in gradient boosting regression. Our framework integrates dropout or parallel training with a recently proposed regularization procedure that allows for a central limit theorem (CLT) for boosting. With these enhancements, we surprisingly find that \textit{increasing} the dropout rate and the number of trees grown in parallel at each iteration substantially enhances signal recovery and overall performance. Our resulting algorithms enjoy similar CLTs, which we use to construct built-in confidence intervals, prediction intervals, and rigorous hypothesis tests for assessing variable importance. Numerical experiments demonstrate that our algorithms perform well, interpolate between regularized boosting and random forests, and confirm the validity of their built-in statistical inference procedures.



Paperid:848
Authors:Peiyuan Zhang, Haofeng Huang, Yongqi Chen, Will Lin, Zhengzhong Liu, Ion Stoica, Eric Xing, Hao Zhang
Abstract:
Abstract:Scaling video diffusion transformers (DiTs) is limited by their quadratic 3D attention, even though most of the attention mass concentrates on a small subset of positions. We turn this observation into ViSA, a trainable, hardware-efficient sparse attention that replaces full attention at \emph{both} training and inference. In ViSA, a lightweight coarse stage pools tokens into tiles and identifies high-weight \emph{critical tokens}; a fine stage computes token-level attention only inside those tiles, both subject to block computing layout to ensure hard efficiency. This leads to a single differentiable kernel that trains end-to-end, requires no post-hoc profiling, and sustains 85\% of FlashAttention3 MFU. We perform a large sweep of ablation studies and scaling-law experiments by pretraining DiTs from 60M to 1.4B parameters. ViSA reaches a Pareto point that cuts training FLOPS by 2.53$\times$ with no drop in diffusion loss. Retrofitting the open-source Wan-2.1 model speeds up attention time by 6$\times$ and lowers end-to-end generation time from 31s to 18s with comparable quality.



Paperid:849
Authors:Mingyu Yang, Mehdi Rezagholizadeh, Guihong Li, Vikram Appia, Emad Barsoum
Abstract:
With the growing demand for deploying large language models (LLMs) across diverse applications, improving their inference efficiency is crucial for sustainable and democratized access. However, retraining LLMs to meet new user-specific requirements is prohibitively expensive and environmentally unsustainable. In this work, we propose a practical and scalable alternative: composing efficient hybrid language models from existing pre-trained models.Our approach, X-EcoMLA, introduces a family of 1B, 3B, and 8B hybrid models by combining State Space Models (SSMs) and Multi-head Latent Attention (MLA) layers, using a refined initialization and post-training pipeline to efficiently transfer knowledge from pre-trained Transformers.X-EcoMLA achieves Transformer-level accuracy with near-SSM efficiency using only 7–11 billion training tokens (compared to the trillions required for pre-training) and an 8B teacher. Moreover, it dramatically reduces KV cache size—down to 3.9%, 2%, and 2.73% of the original for the 1B, 3B, and 8B variants, respectively—while preserving 100%, 100%, and over 97% of average zero-shot performance on LM Harness tasks.Compared to models like MambaInLLaMA, X-EcoMLA, Minitron, and Llamba, our approach consistently delivers competitive or superior accuracy while using significantly fewer tokens, smaller teachers, and vastly reduced KV cache memory. Notably, X-EcoMLA-8B surpasses Minitron-8B in few-shot accuracy by 7%, while using 8× fewer training tokens, over 12× smaller KV cache, and a smaller teacher (8B vs. 15B). It also achieves 2.6× to 3.8× higher throughput (tokens per second) than MambaInLLaMA at up to 32k context length.We will release code and model checkpoints upon acceptance.



Paperid:850
Authors:Adrian Łańcucki, Konrad Staniszewski, Piotr Nawrot, Edoardo Maria Ponti
Abstract:
Inference-time scaling trades off efficiency for increased reasoning accuracy by generating longer or more parallel sequences. However, the cost of generation in Transformer LLMs is bottlenecked by the size of key-value (KV) cache, rather than the number of generated tokens. Hence, we explore inference-time hyper-scaling: by compressing the KV cache we can generate more tokens for the same compute budget, further improving the accuracy of scaled inference. The success of this approach, however, hinges on the ability of compression methods to maintain accuracy even under high compression ratios. To address this challenge, we introduce Dynamic Memory Sparsification (DMS), a novel and sample-efficient method for sparsifying KV caches. Instead of prematurely discarding cached tokens, DMS delays such decisions, thus implicitly merging representations and preserving critical information. We demonstrate the effectiveness of inference-time hyper-scaling on multiple families of LLMs, including DeepSeek R1-distilled models, showing that KV cache compression methods boost accuracy across various reasoning benchmarks for comparable inference runtime and memory load. Moreover, we find that DMS offers the best efficiency-accuracy trade-off among these methods.



Paperid:851
Authors:Nan Lu, Jian Shi, Xinyu Tian
Abstract:
Preference-based data often appear complex and noisy but may conceal underlying homogeneous structures. This paper introduces a novel framework of ranking structure recognition for preference-based data. We first develop an approach to identify dynamic ranking groups by incorporating temporal penalties into a spectral estimation for the celebrated Bradley-Terry model. To detect structural changes, we introduce an innovative objective function and present a practicable algorithm based on dynamic programming. Theoretically, we establish the consistency of ranking group recognition by exploiting properties of a random `design matrix' induced by a reversible Markov chain. We also tailor a group inverse technique to quantify the uncertainty in item ability estimates. Additionally, we prove the consistency of structure change recognition, ensuring the robustness of the proposed framework. Experiments on both synthetic and real-world datasets demonstrate the practical utility and interpretability of our approach.



Paperid:852
Authors:Mingyang Liu, Gabriele Farina, Asuman Ozdaglar
Abstract:
Post-training has demonstrated its importance in enhancing the reasoning capabilities of large language models (LLMs). The primary post-training methods can be categorized into supervised fine-tuning (SFT) and reinforcement fine-tuning (RFT). SFT is efficient and well-suited for small language models, but it may lead to overfitting and limit the reasoning abilities of larger models. In contrast, RFT generally yields better generalization but depends heavily on the strength of the base model. To address the limitations of SFT and RFT, we propose Unified Fine-Tuning (UFT), a novel post-training paradigm that unifies SFT and RFT into a single, integrated process. UFT enables the model to effectively explore solutions while incorporating informative supervision signals, bridging the gap between memorizing and thinking underlying existing methods. Notably, UFT outperforms both SFT and RFT in general, regardless of model sizes. Furthermore, we theoretically prove that UFT breaks RFT's inherent exponential sample complexity bottleneck, showing for the first time that unified training can exponentially accelerate convergence on long-horizon reasoning tasks.



Paperid:853
Authors:Mouadh Yagoubi, David Danan, Milad LEYLI ABADI, Jocelyn Mazari, Jean-Patrick Brunet, Abbas Kabalan, Fabien Casenave, Yuxin Ma, Giovanni Catalani, Jean Fesquet, Jacob Helwig, Xuan Zhang, Haiyang Yu, Xavier BERTRAND, Frédéric TOST, Michaël Bauerheim, Joseph Morlier, Shuiwang Ji
Abstract:
The integration of machine learning (ML) into the physical sciences is reshaping computational paradigms, offering the potential to accelerate demanding simulations such as computational fluid dynamics (CFD). Yet, persistent challenges in accuracy, generalization, and physical consistency hinder the practical deployment of ML models in scientific domains. To address these limitations and systematically benchmark progress, we organized the ML4CFD competition, centered on surrogate modeling for aerodynamic simulations over two-dimensional airfoils. The competition attracted over 240 teams, who were provided with a curated dataset generated via OpenFOAM and evaluated through a multi-criteria framework encompassing predictive accuracy, physical fidelity, computational efficiency, and out-of-distribution generalization. This retrospective analysis reviews the competition outcomes, highlighting several approaches that outperformed baselines under our global evaluation score. Notably, the top entry exceeded the performance of the original OpenFOAM solver on aggregate metrics, illustrating the promise of ML-based surrogates to outperform traditional solvers under tailored criteria. Drawing from these results, we analyze the key design principles of top submissions, assess the robustness of our evaluation framework, and offer guidance for future scientific ML challenges.



Paperid:854
Authors:Chong You, Kan Wu, Zhipeng Jia, Lin Chen, Srinadh Bhojanapalli, Jiaxian Guo, Utku Evci, Jan Wassenberg, Praneeth Netrapalli, Jeremiah Willcock, Suvinay Subramanian, Felix Chern, Alek Andreev, Shreya Pathak, Felix Yu, Prateek Jain, David Culler, Henry Levy, Sanjiv Kumar
Abstract:
Abstract:The discovery of the *lazy neuron phenomenon* (Li et al., 2022), where fewer than 10% of the feedforward networks (FFN) parameters in trained Transformers are activated per token, has spurred significant interests in *activation sparsity* for enhancing large model efficiency. While notable progress has been made in translating such sparsity to wall-time benefits across CPUs, GPUs, and TPUs, modern Transformers have moved away from the ReLU activation function crucial to this phenomenon. Existing efforts on re-introducing activation sparsity, e.g., by reverting to ReLU or applying top-k masking, often degrade model quality, increase parameter count, or complicate training. Sparse attention, the application of sparse activation to the attention mechanism, often face similar challenges. This paper introduces the Spark Transformer, a novel architecture that achieves high activation sparsity in both FFN and the attention mechanism while maintaining model quality, parameter count, and standard training procedures. Our method realizes sparsity via top-$k$ masking for explicit control over sparsity level. Crucially, we introduce *statistical top-k*, a hardware-accelerator-friendly, linear-time approximate algorithm that avoids costly sorting and mitigates significant training slowdown from standard top-k operators. Furthermore, Spark Transformer reallocates existing FFN parameters and attention key embeddings to form a low-cost predictor for identifying activated entries. This design not only mitigates quality loss from enforced sparsity, but also enhances wall-time benefit. Pretrained with the Gemma-2 recipe, Spark Transformer demonstrates competitive performance on standard benchmarks while exhibiting significant sparsity: only 8\% of FFN neurons are activated, and each token attends to a maximum of 256 tokens. This translates to a 2.5x reduction in FLOPs, leading to decoding wall-time speedups of up to 1.79x on CPU and 1.40xon GPU.



Paperid:855
Authors:Weiheng Zhao, Zilong Huang, Jiashi Feng, Xinggang Wang
Abstract:
Contrastive Language-Image Pretraining (CLIP) achieves strong generalization in vision-language tasks by aligning images and texts in a shared embedding space. However, recent findings show that CLIP-like models still underutilize fine-grained semantic signals in text, and this issue becomes even more pronounced when dealing with long and detailed captions.This stems from CLIP’s training objective, which optimizes only global image-text similarity and overlooks token-level supervision—limiting its ability to achieve fine-grained visual-text alignment. To address this, we propose SuperCLIP, a simple yet effective framework that augments contrastive learning with classification-based supervision. By adding only a lightweight linear layer to the vision encoder, SuperCLIP leverages token-level cues to enhance visual-textual alignment — with just a 0.077\% increase in total FLOPs, and no need for additional annotated data.Experiments show that SuperCLIP consistently improves zero-shot classification, image-text retrieval, and purely visual tasks. These gains hold regardless of whether the model is trained on original web data or rich re-captioned data, demonstrating SuperCLIP’s ability to recover textual supervision in both cases. Furthermore, SuperCLIP alleviates CLIP’s small-batch performance drop through classification-based supervision that avoids reliance on large batch sizes. Code and models will be made open source.



Paperid:856
Authors:Xingyu Zhu, Beier Zhu, Shuo Wang, Kesen Zhao, Hanwang Zhang
Abstract:
Vision-language models (VLMs) such as CLIP demonstrate strong generalization in zero-shot classification but remain highly vulnerable to adversarial perturbations. Existing methods primarily focus on adversarial fine-tuning or prompt optimization, they often overlook the gaps in CLIP’s encoded features, which is shown as the text and image features lie far apart from each other. This misalignment is significantly amplified under adversarial perturbations, leading to severe degradation in classification performance. To address this problem, we proposeCrOss-modaLityAlignment, dubbedCOLA, an optimal transport-based framework that explicitly addresses adversarial misalignment by restoring both global image-text alignment and local structural consistency in the feature space. (1) COLA first projects adversarial image embeddings onto a subspace spanned by class text features, effectively filtering out non-semantic distortions while preserving discriminative information. (2) It then models images and texts as discrete distributions over multiple augmented views and refines their alignment via OT, with the subspace projection seamlessly integrated into the cost computation. This design ensures stable cross-modal alignment even under adversarial conditions. COLA is training-free and compatible with existing fine-tuned models.Extensive evaluations across 14 zero-shot classification benchmarks demonstrate the effectiveness of COLA, especially with an average improvement of 6.7% on ImageNet and its variants under PGD adversarial attacks, while maintaining high accuracy on clean samples.



Paperid:857
Authors:Xiaoyu Hu, Gengyu Xue, Zhenhua Lin, Yi Yu
Abstract:
Algorithmic fairness has become a central topic in machine learning, and mitigating disparities across different subpopulations has emerged as a rapidly growing research area. In this paper, we systematically study the classification of functional data under fairness constraints, ensuring the disparity level of the classifier is controlled below a pre-specified threshold. We propose a unified framework for fairness-aware functional classification, tackling an infinite-dimensional functional space, addressing key challenges from the absence of density ratios and intractability of posterior probabilities, and discussing unique phenomena in functional classification. We further design a post-processing algorithm Fair Functional Linear Discriminant Analysis classifier (Fair-FLDA), which targets at homoscedastic Gaussian processes and achieves fairness via group-wise thresholding. Under weak structural assumptions on eigenspace, theoretical guarantees on fairness and excess risk controls are established. As a byproduct, our results cover the excess risk control of the standard FLDA as a special case, which, to the best of our knowledge, is first time seen. Our theoretical findings are complemented by extensive numerical experiments on synthetic and real datasets, highlighting the practicality of our designed algorithm.



Paperid:858
Authors:Sajad Movahedi, Felix Sarnthein, Nicola Muca Cirone, Antonio Orvieto
Abstract:
Abstract:Linear recurrent neural networks (RNNs) and state-space models (SSMs) such as Mamba have become promising alternatives to softmax-attention as sequence mixing layers in Transformer architectures. Current models, however, do not exhibit the full state-tracking expressivity of RNNs because they rely on channel-wise (i.e. diagonal) sequence mixing. In this paper, we investigate parameterizations of dense linear RNNs as fixed-points of parallelizable diagonal linear RNNs. The resulting models can naturally trade expressivity for efficiency at a fixed number of parameters and achieve state-of-the-art results on the commonly used toy tasks $A_5$, $S_5$, copying, and modular arithmetics. We hope our results will open new avenues to more expressive and efficient sequence mixers.



Paperid:859
Authors:Samuel G. Fadel, Hrittik Roy, Nicholas Krämer, Yevgen Zainchkovskyy, Stas Syrota, Alejandro Valverde Mahou, Carl Henrik Ek, Søren Hauberg
Abstract:
Variational mean field approximations tend to struggle with contemporary overparametrized deep neural networks. Where a Bayesian treatment is usually associated with high-quality predictions and uncertainties, the practical reality has been the opposite, with unstable training, poor predictive power, and subpar calibration. Building upon recent work on reparametrizations of neural networks, we propose a simple variational family that considers two independent linear subspaces of the parameter space. These represent functional changes inside and outside the support of training data. This allows us to build a fully-correlated approximate posterior reflecting the overparametrization that tunes easy-to-interpret hyperparameters. We develop scalable numerical routines that maximize the associated evidence lower bound (ELBO) and sample from the approximate posterior. Empirically, we observe state-of-the-art performance across tasks, models, and datasets compared to a wide array of baseline methods. Our results show that approximate Bayesian inference applied to deep neural networks is far from a lost cause when constructing inference mechanisms that reflect the geometry of reparametrizations.



Paperid:860
Authors:Emmeran Johnson, Alberto Rumi, Ciara Pike-Burke, Patrick Rebeschini
Abstract:
Abstract:We study the adversarial Stochastic Shortest Path (SSP) problem with sparse costs under full-information feedback. In the known transition setting, existing bounds based on Online Mirror Descent (OMD) with negative-entropy regularization scale with $\sqrt{\log S A}$, where $SA$ is the size of the state-action space. While we show that this is optimal in the worst-case, this bound fails to capture the benefits of sparsity when only a small number $M \ll SA$ of state-action pairs incur cost. In fact, we also show that the negative-entropy is inherently non-adaptive to sparsity: it provably incurs regret scaling with $\sqrt{\log S}$ on sparse problems. Instead, we propose a novel family of $\ell_r$-norm regularizers ($r \in (1,2)$) that adapts to the sparsity and achieves regret scaling with $\sqrt{\log M}$ instead of $\sqrt{\log SA}$. We show this is optimal via a matching lower bound, highlighting that $M$ captures the effective dimension of the problem instead of $SA$. Finally, in the unknown transition setting the benefits of sparsity are limited: we prove that even on sparse problems, the minimax regret for any learner scales polynomially with $SA$.



Paperid:861
Authors:Syomantak Chaudhuri, Jerry Li, Thomas Courtade
Abstract:
Abstract:We study the problem of robust estimation under heterogeneous corruption rates, where each sample may be independently corrupted with a known but non-identical probability. This setting arises naturally in distributed and federated learning, crowdsourcing, and sensor networks, yet existing robust estimators typically assume uniform or worst-case corruption, ignoring structural heterogeneity. For mean estimation for multivariate bounded distributions and univariate gaussian distributions, we give tight minimax rates for all heterogeneous corruption patterns. For multivariate gaussian mean estimation and linear regression, we establish the minimax rate for squared error up to a factor of $\sqrt{d}$, where $d$ is the dimension. Roughly, our findings suggest that samples beyond a certain corruption threshold may be discarded by the optimal estimators -- this threshold is determined by the empirical distribution of the corruption rates given.



Paperid:862
Authors:Connall Garrod, Jonathan Keating
Abstract:
Neural collapse (NC) and its multi-layer variant, deep neural collapse (DNC), describe a structured geometry that occurs in the features and weights of trained deep networks. Recent theoretical work by Sukenik et al. using a deep unconstrained feature model (UFM) suggests that DNC is suboptimal under mean squared error (MSE) loss. They heuristically argue that this is due to low-rank bias induced by L2 regularization. In this work, we extend this result to deep UFMs trained with cross-entropy loss, showing that high-rank structures—including DNC—are not generally optimal. We characterize the associated low-rank bias, proving a fixed bound on the number of non-negligible singular values at global minima as network depth increases. We further analyze the loss surface, demonstrating that DNC is more prevalent in the landscape than other critical configurations, which we argue explains its frequent empirical appearance. Our results are validated through experiments in deep UFMs and deep neural networks.



Paperid:863
Authors:Shaojie Zhang, Ruoceng Zhang, Pei Fu, Shaokang Wang, Jiahui Yang, Xin Du, ShiqiCui, Bin Qin, Ying Huang, Zhenbo Luo, Jian Luan
Abstract:
In the field of AI-driven human-GUI interaction automation, while rapid advances in multimodal large language models and reinforcement fine-tuning techniques have yielded remarkable progress, a fundamental challenge persists: their interaction logic significantly deviates from natural human-GUI communication patterns. To fill this gap, we propose ``Blink-Think-Link'' (BTL), a brain-inspired framework for human-GUI interaction that mimics the human cognitive process between users and graphical interfaces. The system decomposes interactions into three biologically plausible phases: (1) \textbf{Blink} - rapid detection and attention to relevant screen areas, analogous to saccadic eye movements; (2) \textbf{Think} - higher-level reasoning and decision-making, mirroring cognitive planning; and (3) \textbf{Link} - generation of executable commands for precise motor control, emulating human action selection mechanisms. Additionally, we introduce two key technical innovations for BTL framework:(1) Blink Data Generation - an automated annotation pipeline specifically optimized for blink data, and(2) {BTL Reward – the first rule-based reward mechanism that enables reinforcement learning driven by both process and outcome.}Building upon this framework, we develop a GUI agent model named BTL-UI, which demonstrates consistent state-of-the-art performance across both static GUI understanding and dynamic interaction tasks in comprehensive benchmarks. These results provide conclusive empirical validation of the framework's efficacy in developing advanced GUI Agents. We will soon release the relevant data and models.



Paperid:864
Authors:Enzhe Lu, Zhejun Jiang, Jingyuan Liu, Yulun Du, Tao Jiang, Chao Hong, Shaowei Liu, Weiran He, Enming Yuan, Yuzhi Wang, Zhiqi Huang, Huan Yuan, Suting Xu, Xinran Xu, Guokun Lai, Yanru Chen, Huabin Zheng, Junjie Yan, Jianlin Su, Yuxin Wu, Yutao Zhang, Zhilin Yang, Xinyu Zhou, Mingxing Zhang, Jiezhong Qiu
Abstract:
Scaling the effective context length is essential for advancing large language models (LLMs) toward artificial general intelligence (AGI). However, the quadratic increase in computational complexity inherent in traditional attention mechanisms presents a prohibitive overhead. Existing approaches either impose strongly biased structures, such as sink or window attention which are task-specific, or radically modify the attention mechanism into linear approximations, whose performance in complex reasoning tasks remains inadequately explored.In this work, we propose a solution that adheres to the ``less structure'' principle, allowing the model to determine where to attend autonomously, rather than introducing predefined biases. We introduce Mixture of Block Attention (MoBA), an innovative approach that applies the principles of Mixture of Experts (MoE) to the attention mechanism. This novel architecture demonstrates superior performance on long-context tasks while offering a key advantage: the ability to seamlessly transition between full and sparse attention, enhancing efficiency without the risk of compromising performance. MoBA has already been deployed to handle actual production workloads with long-context requirements, demonstrating significant advancements in efficient attention computation for LLMs.



Paperid:865
Authors:Sotirios Panagiotis Chytas, Rudrasis Chakraborty, Vikas Singh
Abstract:
Recent results show that modern Large Language Models (LLM) are indeed capable of understanding and answering questions about structured data such as graphs. This new paradigm can lead to solutions that require less supervision while, at the same time, providing a model that can generalize and answer questions beyond the training labels. Existing proposals often use some description of the graph to create an ``augmented'' prompt fed to the LLM. For a chosen class of graphs, if a well-tailored graph encoder is deployed to play together with a pre-trained LLM, the model can answer graph-related questions well. Existing solutions to graph-based prompts range from graph serialization to graph transformers. In this work, we show that the use of a parameter-free graph encoder based on Fock space representations, a concept borrowed from mathematical physics, is remarkably versatile in this problem setting. The simple construction, inherited directly from the theory with a few small adjustments, can provide rich and informative graph encodings, for a wide range of different graphs. We investigate the use of this idea for prefix-tuned prompts leveraging the capabilities of a pre-trained, frozen LLM. The modifications lead to a model that can answer graph-related questions -- from simple graphs to proteins to hypergraphs -- effectively and with minimal, if any, adjustments to the architecture. Our work significantly simplifies existing solutions and generalizes well to multiple different graph-based structures effortlessly.



Paperid:866
Authors:Yang Ye, Xianyi He, Zongjian Li, lin bin, Shenghai Yuan, Zhiyuan Yan, Bohan Hou, Li Yuan
Abstract:
Recent advancements in generative models have enabled high-fidelity text-to-image generation. However, open-source image-editing models still lag behind their proprietary counterparts, primarily due to limited high-quality data and insufficient benchmarks.To overcome these limitations, we introduceImgEdit, a large-scale, high-quality image-editing dataset comprising one million carefully curated edit pairs, which contain both novel and complex single-turn edits, as well as challenging multi-turn tasks.To ensure the data quality, we employ a multi-stage pipeline that integrates a cutting-edge vision-language model, a detection model, a segmentation model, alongside task-specific in-painting procedures and strict post-processing. ImgEdit surpasses existing datasets in both task novelty and data quality.Using ImgEdit, we trainImgEdit-E1, an editing model using Vision Language Model to process the reference image and editing prompt, which outperforms existing open-source models on multiple tasks, highlighting the value of ImgEdit and model design.For comprehensive evaluation, we introduceImgEdit-Bench, a benchmark designed to evaluate image editing performance in terms of instruction adherence, editing quality, and detail preservation.It includes a basic testsuite, a challenging single-turn suite, and a dedicated multi-turn suite. We evaluate both open-source and proprietary models, as well as ImgEdit-E1, providing deep analysis and actionable insights into the current behavior of image-editing models.



Paperid:867
Authors:Taiye Chen, Xun Hu, Zihan Ding, Chi Jin
Abstract:
Foundational world models must be both interactive and preserve spatialtemporal coherence to enable effective future planning with different action choices. However, present models for long video generation have limited inherent world modeling capabilities due to two main challenges: compounding errors and insufficient memory mechanisms. We enhance image-to-video models with interactive capabilities through additional action conditioning and autoregressive framework, and reveal that compounding error is inherently irreducible in autoregressive video generation, while insufficient memory mechanism leads to incoherence of world models. We propose video retrieval augmented generation (VRAG) with explicit global state conditioning, which significantly reduces long-term compounding errors and increases spatialtemporal consistency of video world models. In contrast, naive autoregressive generation with extended context windows and retrieval-augmented generation prove less effective for video generation, primarily due to the limited in-context learning capabilities of current video models. Our work illuminates the fundamental challenges in video world models and establishes a comprehensive benchmark for improving video generation models with internal world modeling capabilities.



Paperid:868
Authors:Orr Paradise, Liangyuan Chen, Pranav Muralikrishnan, Hugo Flores, Bryan Pardo, Roee Diamant, David Gruber, Shane Gero, Shafi Goldwasser
Abstract:
Sperm whales communicate in short sequences of clicks known as codas. We present WhAM (Whale Acoustics Model), the first transformer-based model capable of generating synthetic sperm whale codas from any audio prompt. WhAM is built by finetuning VampNet, a masked acoustic token model pretrained on musical audio, using 10k coda recordings collected over the past two decades. Through iterative masked token prediction, WhAM generates high-fidelity synthetic codas that preserve key acoustic features of the source recordings. We evaluate WhAM's synthetic codas using Fréchet Audio Distance and through perceptual studies with expert marine biologists. On downstream tasks including rhythm, social unit, and vowel classification, WhAM's learned representations achieve strong performance, despite being trained for generation rather than classification.



Paperid:869
Authors:Zachary Chase, Steve Hanneke, Shay Moran, Jonathan Shafer
Abstract:
Abstract:We resolve a 30-year-old open problem concerning the power of unlabeled data in online learning by tightly quantifying the gap between transductive and standard online learning. We prove that for every concept class $\mathcal{H}$ with Littlestone dimension $d$, the transductive mistake bound is at least $\Omega(\sqrt{d})$. This establishes an exponential improvement over previous lower bounds of $\Omega(\log \log d)$, $\Omega(\sqrt{\log d})$, and $\Omega(\log d)$, respectively due to Ben-David, Kushilevitz, and Mansour (1995, 1997) and Hanneke, Moran, and Shafer (2023). We also show that our bound is tight: for every $d$, there exists a class of Littlestone dimension $d$ with transductive mistake bound $O(\sqrt{d})$. Our upper bound also improves the previous best known upper bound of $(2/3) \cdot d$ from Ben-David et al. (1997). These results demonstrate a quadratic gap between transductive and standard online learning, thereby highlighting the benefit of advanced access to the unlabeled instance sequence. This stands in stark contrast to the PAC setting, where transductive and standard learning exhibit similar sample complexities.



Paperid:870
Authors:Haowen Cui, Shuo Chen, Jun Li, Jian Yang
Abstract:
Contrastive learning has achieved remarkable success in self-supervised learning by pretraining a generalizable feature representation based on the augmentation invariance. Most existing approaches assume that different augmented views of the same instance (i.e., thepositive pairs) remain semantically invariant. However, the augmentation results withvarying extentmay introduce semantic discrepancies or even content distortion, and thus the conventional (pseudo) supervision from augmentation invariance may lead to misguided learning objectives. In this paper, we propose a novel method called Contrastive Learning with Variable Similarity (CLVS) to accurately characterize the intrinsic similarity relationships between different augmented views. Our method dynamically adjusts the similarity based on the augmentation extent, and it ensures that strongly augmented views are always assigned lower similarity scores than weakly augmented ones. We provide a theoretical analysis to guarantee the effectiveness of the variable similarity in improving model generalizability. Extensive experiments demonstrate the superiority of our approach, achieving gains of 2.1\% on ImageNet-100 and 1.4\% on ImageNet-1k compared with the state-of-the-art methods.



Paperid:871
Authors:Daniel Korchinski, Dhruva Karkada, Yasaman Bahri, Matthieu Wyart
Abstract:
Abstract:Models such as Word2Vec and GloVe construct word embeddings based on the co-occurrence probability $P(i,j)$ of words $i$ and $j$ in text corpora. The resulting vectors $W_i$ not only group semantically similar words but also exhibit a striking linear analogy structure---for example, $W_{\text{king}} - W_{\text{man}} + W_{\text{woman}} \approx W_{\text{queen}}$---whose theoretical origin remains unclear.Previous observations indicate that this analogy structure:(i) already emerges in the top eigenvectors of the matrix $M(i,j) = P(i,j)/P(i)P(j)$,(ii) strengthens and then saturates as more eigenvectors of $M (i, j)$, which controls the dimension of the embeddings, are included,(iii) is enhanced when using $\log M(i,j)$ rather than $M(i,j)$, and(iv) persists even when all word pairs involved in a specific analogy relation (e.g., king--queen, man--woman) are removed from the corpus.To explain these phenomena, we introduce a theoretical generative model in which words are defined by binary semantic attributes, and co-occurrence probabilities are derived from attribute-based interactions. This model analytically reproduces the emergence of linear analogy structure and naturally accounts for properties (i)--(iv). It can be viewed as giving fine-grained resolution into the role of each additional embedding dimension. It is robust to various forms of noise and agrees well with co-occurrence statistics measured on Wikipedia and the analogy benchmark introduced by Mikolov et al.



Paperid:872
Authors:Daniele Foffano, Alessio Russo, Alexandre Proutiere
Abstract:
Robustness to modeling errors and uncertainties remains a central challenge in reinforcement learning (RL). In this work, we address this challenge by leveraging diffusion models to train robust RL policies. Diffusion models have recently gained popularity in model-based RL due to their ability to generate full trajectories "all at once", mitigating the compounding errors typical of step-by-step transition models. Moreover, they can be conditioned to sample from specific distributions, making them highly flexible. We leverage conditional sampling to learn policies that are robust to uncertainty in environment dynamics. Building on the established connection between Conditional Value at Risk (CVaR) optimization and robust RL, we introduce Adversarial Diffusion for Robust Reinforcement Learning (AD-RRL). AD-RRL guides the diffusion process to generate worst-case trajectories during training, effectively optimizing the CVaR of the cumulative return. Empirical results across standard benchmarks show that AD-RRL achieves superior robustness and performance compared to existing robust RL methods.



Paperid:873
Authors:Kai Lion, Liang Zhang, Bingcong Li, Niao He
Abstract:
We show that low-rank adaptation of large-scale models suffers from a low stable rank that is well below the linear algebraic rank of the subspace, degrading fine-tuning performance. To mitigate the underutilization of the allocated subspace, we propose PoLAR, a parameterization inspired by the polar decomposition that factorizes the low-rank update into two direction matrices constrained to Stiefel manifolds and an unconstrained scale matrix. Our theory shows that PoLAR yields an exponentially faster convergence rate on a canonical low-rank adaptation problem. Pairing the parameterization with Riemannian optimization leads to consistent gains on three different benchmarks testing general language understanding, commonsense reasoning, and mathematical problem solving with base model sizes ranging from 350M to 27B.



Paperid:874
Authors:Zehan Wang, Jiayang Xu, Ziang Zhang, Tianyu Pang, Chao Du, Hengshuang Zhao, Zhou Zhao
Abstract:
Humans can intuitively compose and arrange scenes in the 3D space for photography. However, can advanced AI image generators plan scenes with similar 3D spatial awareness when creating images from text or image prompts? We present GenSpace, a novel benchmark and evaluation pipeline to comprehensively assess the spatial awareness of current image generation models. Furthermore, standard evaluations using general Vision-Language Models (VLMs) frequently fail to capture the detailed spatial errors. To handle this challenge, we propose a specialized evaluation pipeline and metric, which reconstructs 3D scene geometry using multiple visual foundation models and provides a more accurate and human-aligned metric of spatial faithfulness. Our findings show that while AI models create visually appealing images and can follow general instructions, they struggle with specific 3D details like object placement, relationships, and measurements. We summarize three core limitations in the spatial perception of current state-of-the-art image generation models: 1) Object Perspective Understanding, 2) Egocentric-Allocentric Transformation, and 3) Metric Measurement Adherence, highlighting possible directions for improving spatial intelligence in image generation.



Paperid:875
Authors:Diego de Oliveira Hitzges, Suman Ghosh, Guillermo Gallego
Abstract:
Event cameras offer a promising avenue for multi-view stereo depth estimation and Simultaneous Localization And Mapping (SLAM) due to their ability to detect blur-free 3D edges at high-speed and over broad illumination conditions. However, traditional deep learning frameworks designed for conventional cameras struggle with the asynchronous, stream-like nature of event data, as their architectures are optimized for discrete, image-like inputs. We propose a scalable, flexible and adaptable framework for pixel-wise depth estimation with event cameras in both monocular and stereo setups. The 3D scene structure is encoded into disparity space images (DSIs), representing spatial densities of rays obtained by back-projecting events into space via known camera poses. Our neural network processes local subregions of the DSIs combining 3D convolutions and a recurrent structure to recognize valuable patterns for depth prediction. Local processing enables fast inference with full parallelization and ensures constant ultra-low model complexity and memory costs, regardless of camera resolution. Experiments on standard benchmarks (MVSEC and DSEC datasets) demonstrate unprecedented effectiveness:(i) using purely monocular data, our method achieves comparable results to existing stereo methods; (ii) when applied to stereo data, it strongly outperforms all state-of-the-art (SOTA) approaches, reducing the mean absolute error by at least 42\%; (iii) our method also allows for increases in depth completeness by more than 3-fold while still yielding a reduction in median absolute error of at least 30\%. Given its remarkable performance and effective processing of event-data, our framework holds strong potential to become a standard approach for using deep learning for event-based depth estimation and SLAM.



Paperid:876
Authors:Kiet Nguyen, Chanhyuk Lee, Donggyun Kim, Dong Hoon Lee, Seunghoon Hong
Abstract:
Spatial conditioning in pretrained text-to-image diffusion models has significantly improved fine-grained control over the structure of generated images. However, existing control adapters exhibit limited adaptability and incur high training costs when encountering novel spatial control conditions that differ substantially from the training tasks.To address this limitation, we propose Universal Few-Shot Control (UFC), a versatile few-shot control adapter capable of generalizing to novel spatial conditions.Given a few image-condition pairs of an unseen task and a query condition, UFC leverage the analogy between query and support conditions to construct task-specific control feature, instantiated by a matching mechanism and an update on a small set of task-specific parameters. Experiments on six spatial control tasks show that UFC, finetuned with only 30 annotated examples, achieves fine-grained control consistent with the spatial conditions.Notably, when finetuned with 0.1% of the full training data, UFC can even be on par with the fully supervised baseline, i.e., Uni-ControlNet, on the Normal, Depth, and Canny tasks.These results highlight UFC's effectiveness in controlling diffusion models when faced with novel spatial conditions.



Paperid:877
Authors:Matthieu Kirchmeyer, Pedro O. Pinheiro, Karolis Martinkus, Emma Willett, Joseph Kleinhenz, Emily Makowski, Andrew Watkins, Vladimir Gligorijevic, Richard Bonneau, Saeed Saremi
Abstract:
Generative models for structure-based drug design are often limited to a specific modality, restricting their broader applicability. To address this challenge, we introduce FuncBind, a framework based on computer vision to generate target-conditioned, all-atom molecules across atomic systems. FuncBind uses neural fields to represent molecules as continuous atomic densities and employs score-based generative models with modern architectures adapted from the computer vision literature. This modality-agnostic representation allows a single unified model to be trained on diverse atomic systems, from small to large molecules, and handle variable atom/residue counts, including non-canonical amino acids. FuncBind achieves competitive in silico performance in generating small molecules, macrocyclic peptides, and antibody complementarity-determining region loops, conditioned on target structures. As a final contribution, we introduce a new dataset and benchmark for structure-conditioned macrocyclic peptide generation.



Paperid:878
Authors:John Yang, Kilian Lieret, Carlos Jimenez, Alexander Wettig, Kabir Khandpur, Yanzhe Zhang, Binyuan Hui, Ofir Press, Ludwig Schmidt, Diyi Yang
Abstract:
Despite recent progress in Language Models (LMs) for software engineering, collecting training data remains a significant pain point.Existing datasets are small, with at most 1,000s of training instances from 11 or fewer GitHub repositories.The procedures to curate such datasets are often complex, necessitating hundreds of hours of human labor; companion execution environments also take up several terabytes of storage, severely limiting their scalability and usability.To address this pain point, we introduce SWE-smith, a novel pipeline for generating software engineering training data at scale.Given any Python codebase, SWE-smith constructs a corresponding execution environment, then automatically synthesizes 100s to 1,000s of task instances that break existing test(s) in the codebase.Using SWE-smith, we create a dataset of 50k instances sourced from 128 GitHub repositories, an order of magnitude larger than all previous works.We train SWE-agent-LM-32B, achieving 40.2% Pass@1 resolve rate on the SWE-bench Verified benchmark, state of the art among open source models.We open source SWE-smith (collection procedure, task instances, trajectories, models) to lower the barrier of entry for research in LM systems for automated software engineering.All assets available at \url{https://swesmith.com}.



Paperid:879
Authors:Haoyuan Cai, Zhenghao (Mark) Peng, Bolei Zhou
Abstract:
Abstract:Learning from human involvement aims to incorporate the human subject to monitor and correct agent behavior errors. While learning from corrective demonstrations addresses present mistakes, it fails to avert errors in the agent’s subsequent trajectory. To address this, we introduce Predictive Preference Learning from Human Interventions (PPL), which exploits the implicit preference signals contained in human interventions over predicted future rollouts. The key idea of PPL is to convert each human intervention into contrastive preference labels over the predicted future trajectory with the length $L$, referred to as the preference horizon. By applying preference optimization on these forecasted states, expert corrections are propagated into the regions where the agent is expected to explore, significantly improving learning efficiency and reducing human demonstrations needed. We evaluate our approach with experiments on both autonomous driving and robotic manipulation benchmarks and show its efficiency and generality. Our theoretical analysis further shows that selecting an appropriate preference horizon $L$ balances coverage of risky states with label fidelity, thereby bounding the algorithmic optimality gap.



Paperid:880
Authors:Miaomiao Huang, Yuhai Zhao, Daniel Zhengkui Wang, Fenglong Ma, Yejiang Wang, Meixia Wang, Xingwei Wang
Abstract:
Graph self-supervised learning aims to learn the intrinsic representations of graph structures from unlabeled data. Although graph contrastive learning (GCL) has achieved remarkable progress by generating perturbed views through data augmentation and optimizing sample similarity, its performance is poor in heterophilic graph scenarios (where connected nodes are likely to belong to different classes or exhibit dissimilar features). In heterophilic graphs, existing methods typically rely on random or carefully designed augmentation strategies (e.g., edge dropping) to create contrastive views. However, such graph structures exhibit intricate edge relationships, where topological perturbations may completely alter the semantics of neighborhoods. Moreover, most methods focus solely on local contrastive signals while neglecting global structural constraints. To address these limitations, inspired by graph coloring, we propose a novel Coloring learning for heterophilic graph Representation framework, CoRep, which: 1) Pioneers a coloring classifier to generate coloring labels, explicitly minimizing the discrepancy between homophilic nodes while maximizing the discrepancy of heterophilic nodes. A global positive sample set is constructed using multi-hop same-color nodes to capture global semantic consistency. 2) Introduces a learnable edge evaluator to guide the coloring learning process dynamically and utilizes the edges' triplet relationship to enhance its robustness. 3) Leverages Gumbel-Softmax to differentially discretize color distributions, suppressing noise via a redundancy constraint and enhancing intra-class compactness. Experimental results on 14 benchmark datasets demonstrate that CoRep significantly outperforms current state-of-the-art methods.



Paperid:881
Authors:Hugo Koubbi, Nirmit Joshi, Theodor Misiakiewicz, Nati Srebro
Abstract:
Abstract:We study the problem of learning single-index models, where the label $y \in \mathbb{R}$ depends on the input $\boldsymbol{x} \in \mathbb{R}^d$ only through an unknown one-dimensional projection $\langle \boldsymbol{w_*}, \boldsymbol{x} \rangle$. Prior work has shown that under Gaussian inputs, the statistical and computational complexity of recovering $\boldsymbol{w_*}$ is governed by the Hermite expansion of the link function. In this paper, we propose a new perspective: we argue that *spherical harmonics*---rather than *Hermite polynomials*---provide the natural basis for this problem, as they capture its intrinsic *rotational symmetry*. Building on this insight, we characterize the complexity of learning single-index models under arbitrary spherically symmetric input distributions. We introduce two families of estimators---based on tensor-unfolding and online SGD---that respectively achieve either optimal sample complexity or optimal runtime, and argue that estimators achieving both may not exist in general. When specialized to Gaussian inputs, our theory not only recovers and clarifies existing results but also reveals new phenomena that had previously been overlooked.



Paperid:882
Authors:Houyi Li, Wenzhen Zheng, Qiufeng Wang, Zhenyu Ding, Haoying Wang, Zili Wang, Shijie Xuyang, Ning DING, Shuigeng Zhou, Xiangyu Zhang, Daxin Jiang
Abstract:
Abstract:Training Large Language Models (LLMs) is prohibitively expensive, creating a critical scaling gap where insights from small-scale experiments often fail to transfer to resource-intensive production systems, thereby hindering efficient innovation.To bridge this, we introduce Farseer, a novel and refined scaling law offering enhanced predictive accuracy across scales.. By systematically constructing a model loss surface $L(N,D)$, Farseer achieves a significantly better fit to empirical data than prior laws (e.g., Chinchilla's law). Our methodology yields accurate, robust, and highly generalizable predictions, demonstrating excellent extrapolation capabilities. This allows for the reliable evaluation of competing training strategies across all $(N,D)$ settings, enabling conclusions from small-scale ablation studies to be confidently extrapolated to predict large-scale performance. Furthermore, Farseer provides new insights into optimal compute allocation, better reflecting the nuanced demands of modern LLM training. To validate our approach, we trained an extensive suite of approximately 1,000 LLMs across diverse scales and configurations, consuming roughly 1.5 million NVIDIA H100 GPU hours. We are comprehensively open-sourcing all models, data, results, and logs at https://anonymous.4open.science/r/Farseer/ to foster further research.



Paperid:883
Authors:Sota Moriyama, Katsumi Inoue
Abstract:
The use of deep learning to solve fundamental AI problems such as Boolean Satisfiability (SAT) has been explored recently to develop robust and scalable reasoning systems. This work advances such neural-based reasoning approaches by developing a new Graph Neural Network (GNN) to differentiably solve (weighted) Maximum Satisfiability (MaxSAT). To this end, we propose SAT-based Graph Attention Networks (SGATs) as novel GNNs that are built on t-norm based attention and message passing mechanisms, and structurally designed to approximate greedy distributed local search. To demonstrate the effectiveness of our model, we develop a local search solver that uses SGATs to continuously solve any given MaxSAT problem. Experiments on (weighted) MaxSAT benchmark datasets demonstrate that SGATs significantly outperform existing neural-based architectures, and achieve state-of-the-art performance among continuous approaches, highlighting the strength of the proposed model.



Paperid:884
Authors:Dennis Frauen, Valentyn Melnychuk, Jonas Schweisthal, Mihaela van der Schaar, Stefan Feuerriegel
Abstract:
Decision-making across various fields, such as medicine, heavily relies on conditional average treatment effects (CATEs). Practitioners commonly make decisions by checking whether the estimated CATE is positive, even though the decision-making performance of modern CATE estimators is poorly understood from a theoretical perspective. In this paper, we study optimal decision-making based on two-stage CATE estimators (e.g., DR-learner), which are considered state-of-the-art and widely used in practice. We prove that, while such estimators may be optimal for estimating CATE, they can be suboptimal when used for decision-making. Intuitively, this occurs because such estimators prioritize CATE accuracy in regions far away from the decision boundary, which is ultimately irrelevant to decision-making. As a remedy, we propose a novel two-stage learning objective that retargets the CATE to balance CATE estimation error and decision performance. We then propose a neural method that optimizes an adaptively-smoothed approximation of our learning objective. Finally, we confirm the effectiveness of our method both empirically and theoretically. In sum, our work is the first to show how state-of-the-art CATE estimators can be adapted for optimal decision-making.



Paperid:885
Authors:Peter Ochieng
Abstract:
Contrastive learning thrives—or fails—based on how we construct \emph{positive} and \emph{negative} pairs. In the absence of explicit labels, models must infer semantic structure from these proxy signals. Early work on Siamese networks \citep{chopra2005learning,hadsell2006dimensionality} already showed that pair construction directly shapes learned representations. In modern contrastive frameworks, poor pair selection remains a primary failure mode: it either causes collapse, where all embeddings converge to a point, or wastes the representational capacity of the space \citep{chen2020simple,tian2020makes,khosla2020supervised}.Contemporary methods typically generate positives via semantic-preserving augmentations (crop, jitter, view transform), while negatives are drawn from other elements in the mini-batch under the assumption that different images are semantically dissimilar. But this assumption breaks down in fine-grained, low-diversity, or high-resolution settings \citep{kalantidis2020hard,robinson2020contrastive,chuang2020debiased}, motivating techniques such as hard-negative mining and debiased losses \citep{bachman2019learning,tian2020makes}.\paragraph{Beyond pairs: batch-level diversity.} While most prior work focuses on \emph{which} individual negatives to select, we study the geometry of the entire batch. Our central observation is this: the overall \emph{diversity} of the batch embedding space strongly governs both training dynamics and representational quality. If diversity is too low, the model sees nearly identical negatives and gradients vanish—leading to collapse. If diversity is too high, negatives become almost orthogonal, but the resulting gradients shrink in magnitude, and learning slows. Optimal training thus occurs within a \emph{moderate diversity window}: high enough to avoid collapse, low enough to preserve update strength.



Paperid:886
Authors:Zhen Tian, Xin Zhao, Ji-Rong Wen
Abstract:
Abstract:In this paper, we propose a novel optimizer state compression algorithm, namely \textbf{$\pi$-Quant}, which leverages the properties of irrational numbers (\eg $\pi$) for memory-efficient training. The core idea is based on our mathematical findings, which show that a pair of parameters can be represented by a single rotation angle using the complex rotation scheme.Building on this insight, we map the parameters into a complex space and perform quantization using the corresponding rotation angles. To efficiently integrate it into optimization process, we develop an efficient system of geometric equations that computes the precise rotation angles with linear complexity.We evaluate $\pi$-Quant on a wide range of tasks. Our experiments show that it can reduce the bit-width of parameters to 3.32-bit, achieving a 41.8\% decrease in GPU memory usage, all while maintaining full accuracy. \textcolor{blue}{We have submitted the code in supplementary materials}.



Paperid:887
Authors:Junno Yun, Yasar Utku Alcalar, Mehmet Akcakaya
Abstract:
Algorithm unrolling methods have proven powerful for solving the regularized least squares problem in computational magnetic resonance imaging (MRI). These approaches unfold an iterative algorithm with a fixed number of iterations, typically alternating between a neural network-based proximal operator for regularization, a data fidelity operation and auxiliary updates with learnable parameters. While the connection to optimization methods dictate that the proximal operator network should be shared across unrolls, this can introduce artifacts or blurring. Heuristically, practitioners have shown that using distinct networks may be beneficial, but this significantly increases the number of learnable parameters, making it challenging to prevent overfitting. To address these shortcomings, by taking inspirations from proximal operators with varying thresholds in approximate message passing (AMP) and the success of time-embedding in diffusion models, we propose a time-embedded algorithm unrolling scheme for inverse problems. Specifically, we introduce a novel perspective on the iteration-dependent proximal operation in vector AMP (VAMP) and the subsequent Onsager correction in the context of algorithm unrolling, framing them as a time-embedded neural network. Similarly, the scalar weights in the data fidelity operation and its associated Onsager correction are cast as time-dependent learnable parameters. Our extensive experiments on the fastMRI dataset, spanning various acceleration rates and datasets, demonstrate that our method effectively reduces aliasing artifacts and mitigates noise amplification, achieving state-of-the-art performance. Furthermore, we show that our time-embedding strategy extends to existing algorithm unrolling approaches, enhancing reconstruction quality without increasing the computational complexity significantly.



Paperid:888
Authors:Ling Niu, Xiaoji Zheng, han wang, Ziyuan Yang, Chen Zheng, Bokui Chen, Jiangtao Gong
Abstract:
Abstract:In recent years, vision-based end-to-end autonomous driving has emerged as a new paradigm. However, mainstream end-to-end approaches typically rely on visual feature extraction networks trained under label supervision, and this limited supervision scheme restricts the generality and applicability of such models. In this paper, we propose a novel paradigm, termed $E^{3}AD$, which advocates for cross-modal self-supervised learning between visual feature extraction networks and a general EEG large model, in order to learn latent human driving cognition for enhancing end-to-end planning. In this work, self-collected cognitive data are utilized in the aforementioned cross-modal self-supervised learning framework. Subsequently, we investigate the approaches and potential mechanisms of augmenting end-to-end planning with human driving cognition on mainstream end-to-end autonomous driving benchmark models, leveraging well-established autonomous driving datasets. Both open-loop tests and closed-loop simulations are conducted for a comprehensive evaluation of planning performance. Experimental results demonstrate that the $E^{3}AD$ paradigm significantly enhances the end-to-end planning performance of baseline models. Ablation studies further validate the contribution of driving cognition and the effectiveness of self-supervised training. To the best of our knowledge, this is the first work to integrate human driving cognition for improving end-to-end autonomous driving planning. It represents an initial attempt to incorporate embodied cognitive data into end-to-end autonomous driving, providing valuable insights for future brain-inspired autonomous driving systems



Paperid:889
Authors:Cheng Tan, Zhenxiao Cao, Zhangyang Gao, Siyuan Li, Yufei Huang, Stan Z. Li
Abstract:
The AlphaFold Protein Structure Database (AFDB) offers unparalleled structural coverage at near-experimental accuracy, positioning it as a valuable resource for data-driven protein design. However, its direct use in training deep models that are sensitive to fine-grained atomic geometry—such as inverse folding—exposes a critical limitation. Comparative analysis of structural feature distributions reveals that AFDB structures exhibit distinct statistical regularities, reflecting a systematic geometric bias that deviates from the conformational diversity found in experimentally determined structures from the Protein Data Bank (PDB). While AFDB structures are cleaner and more idealized, PDB structures capture the intrinsic variability and physical realism essential for generalization in downstream tasks. To address this discrepancy, we introduce a Debiasing Structure AutoEncoder (DeSAE) that learns to reconstruct native-like conformations from intentionally corrupted backbone geometries. By training the model to recover plausible structural states, DeSAE implicitly captures a more robust and natural structural manifold. At inference, applying DeSAE to AFDB structures produces debiased structures that significantly improve inverse folding performance across multiple benchmarks. This work highlights the critical impact of subtle systematic biases in predicted structures and presents a principled framework for debiasing, significantly boosting the performance of structure-based learning tasks like inverse folding.



Paperid:890
Authors:Silas Ruhrberg Estévez, Nicolás Astorga, Mihaela van der Schaar
Abstract:
There is growing interest in using machine learning (ML) to support clinical diagnosis, but most approaches rely on static, fully observed datasets and fail to reflect the sequential, resource-aware reasoning clinicians use in practice. Diagnosis remains complex and error prone, especially in high-pressure or resource-limited settings, underscoring the need for frameworks that help clinicians make timely and cost-effective decisions. We propose ACTMED (Adaptive Clinical Test selection via Model-based Experimental Design), a diagnostic framework that integrates Bayesian Experimental Design (BED) with large language models (LLMs) to better emulate real-world diagnostic reasoning. At each step, ACTMED selects the test expected to yield the greatest reduction in diagnostic uncertainty for a given patient. LLMs act as flexible simulators, generating plausible patient state distributions and supporting belief updates without requiring structured, task-specific training data. Clinicians can remain in the loop; reviewing test suggestions, interpreting intermediate outputs, and applying clinical judgment throughout. We evaluate ACTMED on real-world datasets and show it can optimize test selection to improve diagnostic accuracy, interpretability, and resource use. This represents a step toward transparent, adaptive, and clinician-aligned diagnostic systems that generalize across settings with reduced reliance on domain-specific data.



Paperid:891
Authors:Adam Pardyl, Dominik Matuszek, Mateusz Przebieracz, Marek Cygan, Bartosz Zieliński, Maciej Wolczyk
Abstract:
The real world is messy and unstructured. Uncovering critical information often requires active, goal-driven exploration. It remains to be seen whether Vision-Language Models (VLMs), which recently emerged as a popular zero-shot tool in many difficult tasks, can operate effectively in such conditions. In this paper, we answer this question by introducing FlySearch, a 3D, outdoor, photorealistic environment for searching and navigating to objects in complex scenes. We define three sets of scenarios with varying difficulty and observe that state-of-the-art VLMs cannot reliably solve even the simplest exploration tasks, with the gap to human performance increasing as the tasks get harder. We identify a set of central causes, ranging from vision hallucination, through context misunderstanding, to task planning failures, and we show that some of them can be addressed by finetuning. We publicly release the benchmark, scenarios, and the underlying codebase.



Paperid:892
Authors:Alicja Ziarko, Michał Bortkiewicz, Michał Zawalski, Benjamin Eysenbach, Piotr Miłoś
Abstract:
Contrastive learning (CL) has emerged as a powerful framework for learning structured representations that enable a wide range of downstream tasks. Its applications span sample-efficient reinforcement learning (RL), retrieval-augmented generation, and improved selection of model-generated samples, among others. Despite these successes, its potential for combinatorial reasoning problems remains largely untapped. In this paper, we take a step in this direction by using temporal contrastive learning to learn representations conducive to solving combinatorial problems, which will reduce our reliance on planning. Our analysis reveals that standard CL approaches struggle to capture temporal dependencies over complex trajectories. To address this, we introduce a novel method that leverages negatives from the same trajectories. Across three complex reasoning tasks, our approach outperforms traditional supervised learning.



Paperid:893
Authors:Kasper Green Larsen, Natascha Schalburg
Abstract:
We prove the first generalization bound for large-margin halfspaces that is asymptotically tight in the tradeoff between the margin, the fraction of training points with the given margin, the failure probability and the number of training points.



Paperid:894
Authors:Myeongho Jeon, Jan Sobotka, Suhwan Choi, Maria Brbic
Abstract:
As future superhuman models become increasingly complex, accurately supervising their behavior may exceed human capabilities. Recent works have demonstrated that in such scenario weak models can effectively supervise strong models, a phenomenon known as weak-to-strong generalization. However, we find that naive weak-to-strong generalization fails under distribution shifts, often leading to worse performance of the strong model than its weak supervisors. To address this, we propose RAVEN, a robust weak-to-strong generalization framework that dynamically learns the optimal combinations of weak models in addition to parameters of the strong model. We demonstrate the effectiveness of RAVEN on image classification, text classification and preference alignment tasks. RAVEN outperforms alternative baselines by over 40% on out-of-distribution tasks while matching or surpassing existing methods on in-distribution tasks. Moreover, our results show that RAVEN assigns higher weights to more accurate weak models, demonstrating its ability to automatically identify trustworthy supervision.



Paperid:895
Authors:Olive Franzese, Ali Shahin Shamsabadi, Carter Luck, Hamed Haddadi
Abstract:
The "black-box service model" enables ML service providers to serve clients while keeping their intellectual property and client data confidential. Confidentiality is critical for delivering ML services legally and responsibly, but makes it difficult for outside parties to verify important model properties such as fairness. Existing methods that assess model fairness confidentially lack either (i)reliabilitybecause they certify fairness with respect to a static set of data, and therefore fail to guarantee fairness in the presence of distribution shift or service provider malfeasance; and/or (ii)scalabilitydue to the computational overhead of confidentiality-preserving cryptographic primitives. We address these problems by introducingonline fairness certificates, which verify that a model is fair with respect to data received by the service provideronlineduring deployment. We then present OATH, a deployably efficient and scalable zero-knowledge proof protocol for confidential online group fairness certification. OATH exploits statistical properties of group fairness via a "cut-and-choose" style protocol, enabling scalability improvements over baselines.



Paperid:896
Authors:Cong Wei, Bo Sun, Haoyu Ma, Ji Hou, Felix Juefei-Xu, Zecheng He, Xiaoliang Dai, Luxin Zhang, Kunpeng Li, Tingbo Hou, Animesh Sinha, Peter Vajda, Wenhu Chen
Abstract:
Recent advancements in video generation have achieved impressive motion realism, yet they often overlook character-driven storytelling, a crucial task for automated film, animation generation. We introduce \textbf{Talking Characters}, a more realistic task to generate talking character animations directly from speech and text. Unlike talking head, Talking Characters aims at generating the full portrait of one or more characters beyond the facial region. In this paper, we propose MoCha, the first of its kind to generate talking characters. To ensure precise synchronization between video and speech, we propose a \textbf{localized audio attention} mechanism that effectively aligns speech and video tokens.To address the scarcity of large-scale speech-labelled video datasets, we introduce a joint training strategy that leverages both speech-labelled and text-labelled video data, significantly improving generalization across diverse character actions. We also design structured prompt templates with character tags, enabling, for the first time, \textbf{multi-character conversation with turn-based dialogue}—allowing AI-generated characters to engage in context-aware conversations with cinematic coherence.Extensive qualitative and quantitative evaluations, including human evaluation studies and benchmark comparisons, demonstrate that MoCha sets a new standard for AI-generated cinematic storytelling, achieving superior realism, controllability and generalization.



Paperid:897
Authors:Siru Zhong, Junjie Qiu, Yangyu Wu, Xingchen Zou, Zhongwen Rao, Bin Yang, Chenjuan Guo, Hao Xu, Yuxuan Liang
Abstract:
Spatio-Temporal Foundation Models (STFMs) promise zero/few-shot generalization across various datasets, yet joint spatio-temporal pretraining is computationally prohibitive and struggles with domain-specific spatial correlations. To this end, we introduce FactoST, a factorized STFM that decouples universal temporal pretraining from spatio-temporal adaptation. The first stage pretrains a space-agnostic backbone with multi-frequency reconstruction and domain-aware prompting, capturing cross-domain temporal regularities at low computational cost. The second stage freezes or further fine-tunes the backbone and attaches an adapter that fuses spatial metadata, sparsifies interactions, and aligns domains with continual memory replay. Extensive forecasting experiments reveal that, in few-shot setting, FactoST reduces MAE by up to 46.4% versus UniST, uses 46.2% fewer parameters, and achieves 68% faster inference than OpenCity, while remaining competitive with expert models. We believe this factorized view offers a practical and scalable path toward truly universal STFMs. The code will be released upon notification.



Paperid:898
Authors:Sorachi Kato, Ryoma Yataka, Pu Wang, Pedro Miraldo, Takuya Fujihashi, Petros Boufounos
Abstract:
Abstract:Radar-based indoor 3D human pose estimation typically relied on fine-grained 3D keypoint labels, which are costly to obtain especially in complex indoor settings involving clutter, occlusions, or multiple people. In this paper, we propose \textbf{RAPTR} (RAdar Pose esTimation using tRansformer) using only 3D BBox and 2D keypoint labels which are considerably easier and more scalable to collect. Our RAPTR is characterized by a two-stage pose decoder architecture with a new pseudo-3D deformable attention to enhance (pose/joint) queries with multi-view radar features: a pose decoder estimates initial 3D poses with a 3D template loss designed to utilize the 3D BBox labels and mitigate depth ambiguities; and the refined joint decoder finalizes pose estimates with 2D keypoint labels and a 3D gravity loss. Evaluated on two indoor radar datasets, RAPTR outperforms existing methods, reducing joint position error by $34.2\%$ on HIBER and $76.9\%$ on MMVR.



Paperid:899
Authors:Seungwoo Yoo, Kyeongmin Yeo, Jisung Hwang, Minhyuk Sung
Abstract:
We introduce Neural Green’s Function, a neural solution operator for Poisson’s equation that achieves superior generalization to diverse irregular geometries and functions. The foundation of our approach is inspired by Green’s functions, which serve as solution operators for linear PDEs, including Poisson’s equation, and depend exclusively on the geometry of problem domains. Building on the eigendecompositions of Green’s functions, Neural Green’s Function extracts per-point features from a volumetric point cloud representing the problem domain and uses them to predict the terms required for the numerical integration to obtain the solution, enabling robust and efficient handling of diverse domains, source functions, and boundary conditions. Unlike recent learning-based solution operators, which often struggle with unseen source functions and boundary functions, Neural Green’s Function is agnostic to specific source and boundary functions used for training by design. By embedding the property of Green’s functions into the framework design, Neural Green’s Function enables data-efficient learning of solution operators. We demonstrate the superior performance of our method in steady-state thermal analysis of mechanical part shapes from the MCB dataset given unseen source and boundary functions. Neural Green’s Function outperforms the current state-of-the-art neural operators, achieving an average error reduction of 13.9% across five shape categories, while being up to 350 times faster than numerical solvers that rely on computationally expensive mesh generation.



Paperid:900
Authors:Jifan Zhang, Fangxin Wang, Zihe Song, Philip S Yu, Kaize Ding, Shixiang Zhu
Abstract:
Stream networks, a unique class of spatiotemporal graphs, exhibit complex directional flow constraints and evolving dependencies, making uncertainty quantification a critical yet challenging task. Traditional conformal prediction methods struggle in this setting due to the need for joint predictions across multiple interdependent locations and the intricate spatio-temporal dependencies inherent in stream networks. Existing approaches either neglect dependencies, leading to overly conservative predictions, or rely solely on data-driven estimations, failing to capture the rich topological structure of the network. To address these challenges, we propose Spatio-Temporal Adaptive Conformal Inference (STACI), a novel framework that integrates network topology and temporal dynamics into the conformal prediction framework. STACI introduces a topology-aware nonconformity score that respects directional flow constraints and dynamically adjusts prediction sets to account for temporal distributional shifts. We provide theoretical guarantees on the validity of our approach and demonstrate its superior performance on both synthetic and real-world datasets. Our results show that STACI effectively balances prediction efficiency and coverage, outperforming existing conformal prediction methods for stream networks.



Paperid:901
Authors:Alessandro Marchese, Jeroen Berrevoets, Sam Verboven
Abstract:
A central challenge in organ transplantation is the extremely low acceptance rate of donor organ offers—typically in the single digits—leading to high discard rates and suboptimal use of available grafts. Current acceptance models embedded in allocation systems are non-causal, trained on observational data, and fail to generalize to policy-relevant counterfactuals. This limits their reliability for both policy evaluation and simulator-based optimization. In this work, we reframe organ-offer acceptance as a counterfactual prediction problem and propose a method to learn from routinely recorded—but often overlooked—refusal reasons. These refusal codes act as direction-only counterfactual signals: for example, a rejection reason such as “donor too old” implies acceptance might have occurred had the donor been younger. We formalize this setting and introduce ClexNet, a novel causal model that learns policy-invariant representations via balanced training and an explanation-guided augmentation loss. On synthetic data, ClexNet outperforms existing acceptance models in predictive performance, generalization, and calibration, offering a robust drop-in improvement for simulators and allocation policy evaluation. Beyond transplantation, our approach offers a general method for incorporating domain-expert feedback as directional supervision, improving performance in settings where only observational data is available.



Paperid:902
Authors:XianJun, Davin Choo, Billy Jin, Yongho Shin
Abstract:
Online bipartite matching is a fundamental problem in online optimization, extensively studied both in its integral and fractional forms due to its theoretical significance and practical applications, such as online advertising and resource allocation. Motivated by recent progress in learning-augmented algorithms, we study online bipartite fractional matching when the algorithm is given advice in the form of a suggested matching in each iteration. We develop algorithms for both the vertex-weighted and unweighted variants that provably dominate the naive ``coin flip'' strategy of randomly choosing between the advice-following and advice-free algorithms. Moreover, our algorithm for the vertex-weighted setting extends to the AdWords problem under the small bids assumption, yielding a significant improvement over the seminal work of Mahdian, Nazerzadeh, and Saberi (EC 2007, TALG 2012). Complementing our positive results, we establish a hardness bound on the robustness-consistency tradeoff that is attainable by any algorithm. We empirically validate our algorithms through experiments on synthetic and real-world data.



Paperid:903
Authors:Mark Kozdoba, Binyamin Perets, Shie Mannor
Abstract:
There is a well known intrinsic trade-off between the fairness of a representation and the performance of classifiers derived from the representation. In this paper we propose a new method to compute the optimal Pareto front of this trade off. In contrast to the existing methods, this approach does not require the training of complex fair representation models. Our approach is derived through three main steps: We analyze fair representations theoretically, and derive several structural properties of optimal representations. We then show that these properties enable a reduction of the computation of the Pareto Front to a compact discrete problem. Finally, we show that these compact approximating problems can be efficiently solved via off-the shelf concave-convex programming methods. In addition to representations, we show that the new methods may also be used to directly compute the Pareto front of fair classification problems. Moreover, the proposed methods may be used with any concave performance measure. This is in contrast to the existing reduction approaches, developed recently in fair classification, which rely explicitly on the structure of the non-differentiable accuracy measure, and are thus unlikely to be extendable. The approach was evaluated on several real world benchmark datasets and compares favorably to a number of recent state of the art fair representation and classification methods.



Paperid:904
Authors:Tori Qiu, Benjamin Laufer, Jon Kleinberg, Hoda Heidari
Abstract:
Regulatory frameworks, such as the EU AI Act, encourage openness of general-purpose AI models by offering legal exemptions for "open-source" models. Despite this legislative attention on openness, the definition of open-source foundation models remains ambiguous. This paper presents a stylized model of the regulator's choice of an open-source definition in order to evaluate which standards will establish appropriate economic incentives for developers. In particular, we model the strategic interactions among the creator of the general-purpose model (the generalist) and the entity that fine-tunes the general-purpose model to a specialized domain or task (the specialist), in response to the regulator. Our results characterize market equilibria -- specifically, upstream model release decisions and downstream fine-tuning efforts -- under various openness policies and present an optimal range of open-source thresholds as a function of model performance. Overall, we identify a curve defined by initial model performance which determines whether increasing the regulatory penalty vs. increasing the open-source threshold will meaningfully alter the generalist's model release strategy. Our model provides a theoretical foundation for AI governance decisions around openness and enables evaluation and refinement of practical open-source policies.



Paperid:905
Authors:Quan Cheng, Yuanyu Wan, Lingyu Wu, Chenping Hou, Lijun Zhang
Abstract:
Continual learning aims to learn multiple tasks sequentially while preserving prior knowledge, but faces the challenge of catastrophic forgetting when acquiring new knowledge. Recently, approaches leveraging pre-trained models have gained increasing popularity to mitigate this issue, due to the strong generalization ability of foundation models. To adjust pre-trained models for new tasks, existing methods usually employ low-rank adaptation, which restricts parameter updates to a fixed low-rank subspace. However, constraining the optimization space inherently compromises the model's learning capacity, resulting in inferior performance. To address the limitation, we propose Continuous Subspace Optimization for Continual Learning (CoSO) to fine-tune the model in a series of subspaces rather than a single one. These sequential subspaces are dynamically determined through the singular value decomposition of gradients. CoSO updates the model by projecting gradients into these subspaces, ensuring memory-efficient optimization. To mitigate forgetting, the optimization subspaces of each task are set to be orthogonal to the historical task subspace. During task learning, CoSO maintains a task-specific component that captures the critical update directions associated with the current task. Upon completing a task, this component is used to update the historical task subspace, laying the groundwork for subsequent learning. Extensive experiments on multiple datasets demonstrate that CoSO significantly outperforms state-of-the-art methods, especially in challenging scenarios with long task sequences.



Paperid:906
Authors:Fangming Cui, Di Yang, Yuqiang Ren, Zhou Yu, Liang Xiao, Xinmei Tian
Abstract:
Advancements in directly-integration parameter optimization have underscored their potential to enhance the performance of labeled scenarios and tasks. One inherent flaw of these methods is that the optimized parameters usually exhibit weak performance on unlabeled tasks or scenarios. This may be attributed to the fact that the uncoordinated learning of directly-integration framework. To mitigate this issue of uncoordinated learning, we propose a novel method called Levelling Paradigm (LePa) to improve performance for unlabeled tasks or scenarios.The proposed LePa dynamically constrains and coordinates multiple objective functions, thereby improving the robustness of coordinated fine-tuning. Comprehensive experiments demonstrate that the LePa outperforms existing methods.



Paperid:907
Authors:Hyungjoo Chae, Seonghwan Kim, Junhee Cho, Seungone Kim, Seungjun Moon, Gyeom Hwangbo, Dongha Lim, Minjin Kim, Yeonjun Hwang, Minju Gwak, Dongwook Choi, Minseok Kang, Gwanhoon Im, ByeongUng Cho, Hyojun Kim, Jun Han, Taeyoon Kwon, Minju Kim, Beong-woo Kwak, Dongjin Kang, Jinyoung Yeo
Abstract:
Abstract:Web navigation is a unique domain that can automate many repetitive real-life tasks and is challenging as it requires long-horizon sequential decision making beyond typical multimodal large language model (MLLM) tasks.Yet, specialized reward models for web navigation that can be utilized during both training and test-time have been absent until now. Despite the importance of speed and cost-effectiveness, prior works have utilized MLLMs as reward models, which poses significant constraints for real-world deployment. To address this, in this work, we propose the first process reward model (PRM) called Web-Shepherd which could assess web navigation trajectories in a step-level. To achieve this, we first construct the WebPRM Collection, a large-scale dataset with 40K step-level preference pairs and annotated checklists spanning diverse domains and difficulty levels. Next, we also introduce the WebRewardBench, the first meta-evaluation benchmark for evaluating PRMs. In our experiments, we observe that our Web-Shepherd achieves about 30 points better accuracy compared to using GPT-4o on WebRewardBench. Furthermore, when testing on WebArena-lite by using GPT-4o-mini as the policy and Web-Shepherd as the verifier, we achieve 10.9 points better performance, in 10$\times$ less cost compared to using GPT-4o-mini as the verifier. Our model, dataset, and code are publicly available at LINK.



Paperid:908
Authors:Andy Zane, Dmitry Batenkov, Rafal Urbaniak, Jeremy Zucker, Sam Witty
Abstract:
Models of hybrid dynamical systems are widely used to answer questions about the causes and effects of dynamic events in time. Unfortunately, existing causal reasoning formalisms lack support for queries involving the dynamically triggered, discontinuous interventions that characterize hybrid dynamical systems. This mismatch can lead to ad-hoc and error-prone causal analysis workflows in practice. To bridge the gap between the needs of hybrid systems users and current causal inference capabilities, we develop a rigorous counterfactual semantics by formalizing interventions as transformations to the constraints of hybrid systems. Unlike interventions in a typical structural causal model, however, interventions in hybrid systems can easily render the model ill-posed. Thus, we identify mild conditions under which our interventions maintain solution existence, uniqueness, and measurability by making explicit connections to established hybrid systems theory. To illustrate the utility of our framework, we formalize a number of canonical causal estimands and explore a case study on the probabilities of causation with applications to fishery management. Our work simultaneously expands the modeling possibilities available to causal inference practitioners and begins to unlock decades of causality research for users of hybrid systems.



Paperid:909
Authors:Harvey Yiyun Fu, Aryan Shrivastava, Jared Moore, Peter West, Chenhao Tan, Ari Holtzman
Abstract:
Large language models (LLMs) are increasingly capable of processing long inputs and locating specific information within them, as evidenced by their performance on the Needle in a Haystack (NIAH) test. However, while models excel at recalling surprising information, they still struggle to identifyclearly omittedinformation. We introduce AbsenceBench to assesses LLMs' capacity to detect missing information across three domains: numerical sequences, poetry, and GitHub pull requests. AbsenceBench challenges models to specify which pieces of the original context were deliberately removed, given explicit references to both the original and edited contexts. Despite the apparent straightforwardness of these tasks, our experiments reveal that even state-of-the-art models like Claude-3.7-Sonnet achieve only 52.4% F1-score with modest average context lengths of 5K tokens. Our analysis suggests this poor performance stems from a fundamental limitation: Transformer attention mechanisms cannot easily attend to "gaps" in documents since these absences don't correspond to any specific keys that can be attended to. Combined, our results and analysis provide a case study of the close proximity of tasks where models are already superhuman (NIAH) and task where models breakdown unexpectedly (AbsenceBench).



Paperid:910
Authors:Peng Wu, Haoxuan Li, Chunyuan Zheng, Yan Zeng, Jiawei Chen, Yang Liu, Ruocheng Guo, Kun Zhang
Abstract:
Counterfactual inference aims to estimate the counterfactual outcome at the individual level given knowledge of an observed treatment and the factual outcome, with broad applications in fields such as epidemiology, econometrics, and management science. Previous methods rely on a known structural causal model (SCM) or assume the homogeneity of the exogenous variable and strict monotonicity between the outcome and exogenous variable. In this paper, we propose a principled approach for identifying and estimating the counterfactual outcome. We first introduce a simple and intuitive rank preservation assumption to identify the counterfactual outcome without relying on a known structural causal model. Building on this, we propose a novel ideal loss for theoretically unbiased learning of the counterfactual outcome and further develop a kernel-based estimator for its empirical estimation. Our theoretical analysis shows that the rank preservation assumption is not stronger than the homogeneity and strict monotonicity assumptions, and shows that the proposed ideal loss is convex, and the proposed estimator is unbiased. Extensive semi-synthetic and real-world experiments are conducted to demonstrate the effectiveness of the proposed method.



Paperid:911
Authors:Yanzhi Chen, Zijing Ou, Adrian Weller, Michael Gutmann
Abstract:
Estimating mutual information (MI) is a fundamental task in data science and machine learning. Existing estimators mainly rely on either highly flexible models (e.g., neural networks), which require large amounts of data, or overly simplified models (e.g., Gaussian copula), which fail to capture complex distributions. Drawing upon recent vector copula theory, we propose a principled interpolation between these two extremes to achieve a better trade-off between complexity and capacity. Experiments on state-of-the-art synthetic benchmarks and real-world data with diverse modalities demonstrate the advantages of the proposed method.



Paperid:912
Authors:Kodai Kawamura, Yuta Goto, Rintaro Yanagi, Hirokatsu Kataoka, Go Irie
Abstract:
Pre-trained Vision-Language Models (VLMs) exhibit strong generalization capabilities, enabling them to recognize a wide range of objects across diverse domains without additional training. However, they often retain irrelevant information beyond the requirements of specific target downstream tasks, raising concerns about computational efficiency and potential information leakage. This has motivated growing interest in approximate unlearning, which aims to selectively remove unnecessary knowledge while preserving overall model performance. Existing approaches to approximate unlearning have primarily focused on {\em class unlearning}, where a VLM is retrained to fail to recognize specified object classes while maintaining accuracy for others. However, merely forgetting object classes is often insufficient in practical applications. For instance, an autonomous driving system should accurately recognize {\em real} cars, while avoiding misrecognition of {\em illustrated} cars depicted in roadside advertisements as {\em real} cars, which could be hazardous. In this paper, we introduce {\em Approximate Domain Unlearning (ADU)}, a novel problem setting that requires reducing recognition accuracy for images from specified domains (e.g., {\em illustration}) while preserving accuracy for other domains (e.g., {\em real}). ADU presents new technical challenges: due to the strong domain generalization capability of pre-trained VLMs, domain distributions are highly entangled in the feature space, making naive approaches based on penalizing target domains ineffective. To tackle this limitation, we propose a novel approach that explicitly disentangles domain distributions and adaptively captures instance-specific domain information. Extensive experiments on three multi-domain benchmark datasets demonstrate that our approach significantly outperforms strong baselines built upon state-of-the-art VLM tuning techniques, paving the way for practical and fine-grained unlearning in VLMs. Codes will be published upon acceptance.



Paperid:913
Authors:Ejafa Bassam, Dawei Zhu, Kaigui Bian
Abstract:
Knowledge distillation is a model compression technique in which a compact "student" network is trained to replicate the predictive behavior of a larger "teacher" network. In logit-based knowledge distillation, it has become the de facto approach to augment cross-entropy with a distillation term. Typically, this term is either a KL divergence-matching marginal probabilities or a correlation-based loss-capturing intra and inter-class relationships-but in every case it sits as an "add-on" to cross-entropy, with its own weight that must be carefully tuned. In this paper, we adopt a choice‐theoretic perspective and recast knowledge distillation under the Plackett-Luce model by interpreting teacher logits as "worth" scores. We introducePlackett–Luce Distillation (PLD), a weighted list‐wise ranking loss in which the teacher model transfers knowledge of its full ranking of classes, weighting each ranked choice by its own confidence. PLD directly optimizes a single "teacher‐optimal" ranking-true label first, followed by the remaining classes in descending teacher confidence-yielding a convex, translation-invariant surrogate that subsumes weighted cross-entropy. Empirically, PLD is robust with default settings and consistently outperforms both classical KL‐based distillation and recent correlation-based method (DIST) across diverse student-teacher architectures and model sizes on standard image classification benchmarks.



Paperid:914
Authors:Ling Xing, Alex Jinpeng Wang, Rui Yan, Xiangbo Shu, Jinhui Tang
Abstract:
Abstract:Real-world applications are stretching context windows to hundreds of thousand of tokens while Large Language Models (LLMs) swell from billions to trillions of parameters.This dual expansion send compute and memory costs skyrocketing, making $\textit{token compression}$ indispensable.We introduce Vision Centric Token Compression ($\textbf{Vist}$), a $\textit{slow–fast}$ compression framework that mirrors human reading: the $\textit{fast}$ path renders distant tokens into images, letting a $\textbf{frozen, lightweight vision encoder}$ skim the low-salience context; the $\textit{slow}$ path feeds the proximal window into the LLM for fine-grained reasoning.A Probability-Informed Visual Enhancement (PVE) objective masks high-frequency tokens during training, steering the Resampler to concentrate on semantically rich regions—just as skilled reader gloss over function words.On eleven in-context learning benchmarks, $\textbf{Vist}$ achieves the same accuracy with 2.3$\times$ fewer tokens, cutting FLOPs by 16\% and memory by 50\%.This method delivers remarkable results, outperforming the strongest text encoder-based compression method CEPE by $\textbf{7.6}$\% on average over benchmarks like TriviaQA, NQ, PopQA, NLUI, and CLIN, setting a new standard for token efficiency in LLMs. The source code will be released.



Paperid:915
Authors:Cody Melcher, zeinab alizadeh, Lindsey Hiett, Afrooz Jalilzadeh, Erfan Yazdandoost Hamedani
Abstract:
Abstract:Semi-Infinite Programming (SIP) has emerged as a powerful framework for modeling problems with infinite constraints, however, its theoretical development in the context of nonconvex and large-scale optimization remains limited. In this paper, we investigate a class of nonconvex min-max optimization problems with nonconvex infinite constraints, motivated by applications such as adversarial robustness and safety-constrained learning. We propose a novel inexact dynamic barrier primal-dual algorithm and establish its convergence properties. Specifically, under the assumption that the squared infeasibility residual function satisfies the Lojasiewicz inequality with exponent $\theta \in (0,1)$, we prove that the proposed method achieves $\mathcal{O}(\epsilon^{-3})$, $\mathcal{O}(\epsilon^{-6\theta})$, and $\mathcal{O}(\epsilon^{-3\theta/(1-\theta)})$ iteration complexities to achieve an $\epsilon$-approximate stationarity, infeasibility, and complementarity slackness, respectively. Numerical experiments on robust multitask learning with task priority further illustrate the practical effectiveness of the algorithm.



Paperid:916
Authors:Qin Shi, Amber Yijia Zheng, Qifan Song, Raymond A. Yeh
Abstract:
We propose the task ofknowledge distillation detection, which aims to determine whether a student model has been distilled from a given teacher, under a practical setting where only the student’s weights and the teacher’s API are available. This problem is motivated by growing concerns about model provenance and unauthorized replication through distillation. To address this task, we introduce a model-agnostic framework that combines data-free input synthesis, statistical score computation for detecting distillation. Our approach is applicable to both classification and generative models. Experiments on diverse architectures for image classification and text-to-image generation show that our method improves detection accuracy over the strongest baselines by 59.6\% on CIFAR-10, 71.2\% on ImageNet, and 20.0\% for text-to-image generation.



Paperid:917
Authors:Advait Gadhikar, Tom Jacobs, chao zhou, Rebekka Burkholz
Abstract:
The performance gap between training sparse neural networks from scratch (PaI) and dense-to-sparse training presents a major roadblock for efficient deep learning. According to the Lottery Ticket Hypothesis, PaI hinges on finding a problem specific parameter initialization. As we show, to this end, determining correct parameter signs is sufficient. Yet, they remain elusive to PaI. To address this issue, we propose Sign-In, which employs a dynamic reparameterization that provably induces sign flips. Such sign flips are complementary to the ones that dense-to-sparse training can accomplish, rendering Sign-In as an orthogonal method. While our experiments and theory suggest performance improvements of PaI, they also carve out the main open challenge to close the gap between PaI and dense-to-sparse training.



Paperid:918
Authors:Zi Yang, Ying Li, Zhidi Lin, Michael Minyi Zhang, Pablo Martinez Olmos
Abstract:
The multi-view Gaussian process latent variable model (MV-GPLVM) aims to learn a unified representation from multi-view data but is hindered by challenges such as limited kernel expressiveness and low computational efficiency. To overcome these issues, we first introduce a new duality between the spectral density and the kernel function. By modeling the spectral density with a bivariate Gaussian mixture, we then derive a generic and expressive kernel termed Next-Gen Spectral Mixture (NG-SM) for MV-GPLVMs. To address the inherent computational inefficiency of the NG-SM kernel, we propose a random Fourier feature approximation. Combined with a tailored reparameterization trick, this approximation enables scalable variational inference for both the model and the unified latent representations. Numerical evaluations across a diverse range of multi-view datasets demonstrate that our proposed method consistently outperforms state-of-the-art models in learning meaningful latent representations.



Paperid:919
Authors:Soroush H. Zargarbashi, Mohammad Sadegh Akhondzadeh, Aleksandar Bojchevski
Abstract:
Abstract:Given any model, conformal prediction (CP) returns prediction sets guaranteed to include the true label with high adjustable probability; robust CP (RCP) extends this to the inputs with worst-case noise. For that, applying randomized smoothing (inference over a noise-augmented distribution around the input) is a well-established approach as it is resilient to larger radii, and applicable to any black-box model. However current smoothing-based RCPs require many model forward passes per each input which is computationally expensive. We show that without any informations (e.g. distribution of scores, inputs, etc), conformal prediction combined with *single* randomized augmented inference already shows robustness. Hence, we can lower bound its worst case coverage guarantee through any binary certificate. Our single sample approach (RCP1) returns (robust) prediction sets with a set sizes comparable to state of the art given large number of inferences (e.g. $\sim 100$) per input. Our key insight is to certify the conformal procedure itself rather than individual scores. Our approach is agnostic to the setup (classification and regression). We further extend our approach to smoothing based robust conformal risk control.



Paperid:920
Authors:Spyros Angelopoulos, Bertrand Simon
Abstract:
Online bidding is a classic optimization problem, with several applications in online decision-making, the design of interruptible systems, and the analysis of approximation algorithms. In this work, we study online bidding under learning-augmented settings that incorporate stochasticity, in either the prediction oracle or the algorithm itself. In the first part, we study bidding under distributional predictions, and find Pareto-optimal algorithms that offer the best-possible tradeoff between the consistency and the robustness of the algorithm. In the second part, we study the power and limitations of randomized bidding algorithms, by presenting upper and lower bounds on the consistency/robustness tradeoffs. Previous works focused predominantly on oracles that do not leverage stochastic information on the quality of the prediction, and deterministic algorithms.



Paperid:921
Authors:Quinn Lanners, Cynthia Rudin, Alexander Volfovsky, Harsh Parikh
Abstract:
Data fusion techniques integrate information from heterogeneous data sources to improve learning, generalization, and decision-making across data sciences. In causal inference, these methods leverage rich observational data to improve causal effect estimation, while maintaining the trustworthiness of randomized controlled trials. Existing approaches often relax the strong "no unobserved confounding" assumption by instead assuming exchangeability of counterfactual outcomes across data sources. However, when both assumptions simultaneously fail—a common scenario in practice—current methods cannot identify or estimate causal effects. We address this limitation by proposing a novel partial identification framework that enables researchers to answer key questions such as:Is the causal effect positive/negative?andHow severe must assumption violations be to overturn this conclusion?Our approach introduces interpretable sensitivity parameters that quantify assumption violations and derives corresponding causal effect bounds. We develop doubly robust estimators for these bounds and operationalize breakdown frontier analysis to understand how causal conclusions change as assumption violations increase. We apply our framework to the Project STAR study, which investigates the effect of classroom size on students’ third-grade standardized test performance. Our analysis reveals that the Project STAR results are robust to simultaneous violations of key assumptions, both on average and across various subgroups of interest. This strengthens confidence in the study's conclusions despite potential unmeasured biases in the data.



Paperid:922
Authors:Mingyu Kang, Chenglin Fan
Abstract:
We study the problem of computing the exact median using unreliable comparison oracles, motivated by settings such as crowdsourced or noisy decision-making. We analyze the complexity of this problem under both the classical and partially ordered input settings. In the classical setting, we introduce a modified LazySelect algorithm that combines weak comparisons with occasional strong comparisons through majority voting. We show that this hybrid strategy has near-linear running time and can achieve high-probability correctness using only sublinear strong comparisons, even when the weak oracle is only slightly better than random guessing. Our theoretical results hold under the persistent comparison model, where resampling will not amplify the probability of correctness. In the partially ordered setting, we generalize the notion of median to directed acyclic graphs (DAGs) and show that the complexity of median selection depends heavily on the DAG's width. We complement our analysis with extensive experiments on synthetic data.



Paperid:923
Authors:Thomas Foster, Anya Sims, Johannes Forkel, Jakob Foerster
Abstract:
Reinforcement learning is widely adopted in post-training large language models, especially for reasoning-style tasks such as maths questions. However, as we show, most existing methods will provably fail to learn from questions that are too hard, where the model always fails, or too easy, where the model always succeeds. Much human effort is therefore spent continually producing datasets of questions of a suitable difficulty for state-of-the-art models. Given this, we consider how to algorithmically identify questions that allow for maximally efficient training. We introduce a method, LILO (Learnability Improves LLMs Optimally), that prioritises training on questions with high variance of success, known as learnability, and we provide theory proving LILO maximises the expected improvement of the model. We run a wide range of experiments over multiple base models, algorithms and reasoning datasets to demonstrate that LILO consistently improves final test accuracy and can yield a 3x reduction in the number of training steps required to reach it. We explore how questions with high learnability can be efficiently identified, and discuss how learnability can be scaled to produce LLM agents that autonomously and open-endedly expand the frontier of human knowledge.



Paperid:924
Authors:Jie Yang, Kexin Zhang, Guibin Zhang, Philip S Yu, Kaize Ding
Abstract:
Time Series Imputation (TSI), which aims to recover missing values in temporal data, remains a fundamental challenge due to the complex and often high-rate missingness in real-world scenarios. Existing models typically optimize the point-wise reconstruction loss, focusing on recovering numerical values (local information). However, we observe that under high missing rates, these models still perform well in the training phase yet produce poor imputations and distorted latent representation distributions (global information) in the inference phase. This reveals a critical optimization dilemma: current objectives lack global guidance, leading models to overfit local noise and fail to capture global information of the data. To address this issue, we propose a new training paradigm,GlocalInformationBottleneck (Glocal-IB). Glocal-IB is model-agnostic and extends the standard IB framework by introducing a Global Alignment loss, derived from a tractable mutual information approximation. This loss aligns the latent representations of masked inputs with those of their originally observed counterparts. It helps the model retain global structure and local details while suppressing noise caused by missing values, giving rise to better generalization under high missingness. Extensive experiments on nine datasets confirm that Glocal-IB leads to consistently improved performance and aligned latent representations under missingness. Our code implementation is available in https://anonymous.4open.science/r/NeurIPS-25-Glocal-IB-E1F0/.



Paperid:925
Authors:Ariel Procaccia, Ben Schiffer, Serena Wang, Shirley Zhang
Abstract:
A common problem in LLM evaluation is how to choose a subset of metrics from a full suite of possible metrics. Subset selection is usually done for efficiency or interpretability reasons, and the goal is often to select a "representative" subset of metrics. However, "representative" is rarely clearly defined. In this work, we use ideas from social choice theory to formalize two notions of representation for the selection of a subset of evaluation metrics. We first introducepositional representation, which guarantees every alternative is sufficiently represented at every position cutoff. We then introducepositional proportionality, which guarantees no alternative is proportionally over- or under-represented by more than a small error at any position. We prove upper and lower bounds on the smallest number of metrics needed to guarantee either of these properties in the worst case. We also study a generalized form of each property that allows for additional input on groups of metrics that must be represented. Finally, we tie theory to practice through real-world case studies on both LLM evaluation and hospital quality evaluation.



Paperid:926
Authors:Amit Keinan, Moshe Shenfeld, Katrina Ligett
Abstract:
Recent methods for auditing the privacy of machine learning algorithms have improved computational efficiency by simultaneously intervening on multiple training examples in a single training run. Steinke et al. prove that one-run auditing indeed lower bounds the true privacy parameter of the audited algorithm, and give impressive empirical results. Their work leaves open the question of how precisely one-run auditing can uncover the true privacy parameter of an algorithm, and how that precision depends on the audited algorithm. In this work, we characterize the maximum achievable efficacy of one-run auditing and show that the key barrier to its efficacy is interference between the observable effects of different data elements. We present new conceptual approaches to minimize this barrier, towards improving the performance of one-run auditing of real machine learning algorithms.



Paperid:927
Authors:Xander Davies, Eric Winsor, Tomek Korbak, Alexandra Souly, Robert Kirk, Christian Schroeder de Witt, Yarin Gal
Abstract:
LLM developers deploy technical mitigations to preventfine-tuning misuse attacks, attacks in which adversaries evade safeguards by fine-tuning the model using a public API. Previous work has established several successful attacks against specific fine-tuning API defences; however, prior attacks training and/or inference samples can be easily flagged as suspicious. In this work, we show that defences of fine-tuning APIs that seek to detect individual harmful training or inference samples ('pointwise' detection) arefundamentally limitedin their ability to prevent fine-tuning attacks. We demonstrate a class of 'pointwise-undetectable' attacks that repurpose semantic or syntactic variations in benign model outputs to covertly transmit dangerous knowledge. Our attacks are composed solely of unsuspicious benign samples that can be collected from the model before fine-tuning, meaning training and inference samples are all individually benign and low-perplexity. We test our attacks against the OpenAI fine-tuning API, finding they succeed in eliciting answers to harmful multiple-choice questions, and that they evade an enhanced monitoring system we design that successfully detects other fine-tuning attacks. Our results showing fundamental limitations of defending against pointwise attacks suggest focusing research efforts on mitigations towards multi-point defences.



Paperid:928
Authors:Tianhao Zhao, Yiyang Zou, Zihao Mao, Peilun Xiao, Yulin Huang, Hongda Yang, Yuxuan Li, Tracy Li, Guobin Wu, Yutian Lin
Abstract:
Driving accident anticipation aims to predict potential collisions in real time, enabling timely alarms to enhance road safety. Existing methods typically predict frame-level anomaly scores as risk indicators. However, these approaches suffer from inconsistent supervision signals because driving risks evolve progressively rather than abruptly, and risk assessment inherently involves human subjectivity. To address this limitation, we propose a novel paradigm that directly predicts the probability of an accident occurring at multiple future timestamps (0.1s–2.0s), offering more precise supervision and improved interpretability. Our framework employs a snippet encoder to capture spatiotemporal dynamics and a Transformer-based decoder to simultaneously estimate accident probabilities across different time steps. Furthermore, we introduce a refined evaluation protocol that measures recall rate and Time-to-Accident (TTA) only under acceptable false alarm rates, ensuring practical applicability in the real world. Experiments demonstrate that our method achieves superior performance in both recall and TTA, validating its effectiveness for real-world accident anticipation.



Paperid:929
Authors:William Denault, Karl Tayeb, Peter Carbonetto, Jason Willwerscheid, Matthew Stephens
Abstract:
Matrix factorization is a fundamental method in statistics and machinelearning for inferring and summarizing structure in multivariatedata. Modern data sets often come with ``side information'' of variousforms (images, text, graphs) that can be leveraged to improveestimation of the underlying structure. However, existing methods thatleverage side information are limited in the types of data they canincorporate, and they assume specific parametric models. Here, weintroduce a novel method for this problem, {\em covariate-moderated empirical Bayes matrix factorization} (cEBMF). cEBMF is a modularframework that accepts any type of side information that isprocessable by a probabilistic model or neural network.The cEBMF framework can accommodate different assumptions andconstraints on the factors through the use of different priors, and itadapts these priors to the data. We demonstrate the benefits of cEBMFin simulations and in analyses of spatial transcriptomics andMovieLens data.



Paperid:930
Authors:Han Cui, Jingbo Liu
Abstract:
Abstract:Conditional sampling is a fundamental task in Bayesian statistics and general modeling. Consider the problem of sampling from the posterior distribution $P\_{X|Y=y^*}$ for some observation $y^*$, where the likelihood $P\_{Y|X}$ is known, and we are given $n$ i.i.d. samples $D=\\{X\_i\\}\_{i=1}^n$ drawn from an unknown prior distribution $\pi\_X$. Suppose that $f(\hat{\pi}\_{X^n})$ is the distribution of a posterior sample generated by an algorithm (e.g. a conditional generative model or the Bayes rule) when $\hat{\pi}\_{X^n}$ is the empirical distribution of the training data. Although averaging over the randomness of the training data $D$, we have $\mathbb{E}\_D[\hat{\pi}\_{X^n}]= \pi\_X$, we do not have $\mathbb{E}\_D[f(\hat{\pi}\_{X^n})]= f(\pi\_X)$ due to the nonlinearity of $f$, leading to a bias. In this paper we propose a black-box debiasing scheme that improves the accuracy of such a naive plug-in approach. For any integer $k$ and under boundedness of the likelihood and smoothness of $f$, we generate samples $\hat{X}^{(1)},\dots,\hat{X}^{(k)}$ and weights $w_1,\dots,w_k$ such that $\sum_{i=1}^kw_iP_{\hat{X}^{(i)}}$ is a $k$-th order approximation of $f(\pi_X)$, where the generation process treats $f$ as a black-box. Our generation process achieves higher accuracy when averaged over the randomness of the training data, but not conditionally on the training data, which highlights the trade-off between memorization and generalization in generative models.



Paperid:931
Authors:Dejan Stancevic, Florian Handke, Luca Ambrogioni
Abstract:
The practical performance of generative diffusion models depends on the appropriate choice of the noise scheduling function, which can also be equivalently expressed as a time reparameterization. In this paper, we present a time scheduler that selects sampling points based on entropy rather than uniform time spacing, ensuring that each point contributes an equal amount of information to the final generation. We prove that this time reparameterization does not depend on the initial choice of time. Furthermore, we provide a tractable exact formula to estimate this entropic time for a trained model using the training loss without substantial overhead. Alongside the entropic time, inspired by the optimality results, we introduce a rescaled entropic time. In our experiments with mixtures of Gaussian distributions and ImageNet, we show that using the (rescaled) entropic times greatly improves the inference performance of trained models. In particular, we found that the image quality in pretrained EDM2 models, as evaluated by FID and FD-DINO scores, can be substantially increased by the rescaled entropic time reparameterization without increasing the number of function evaluations, with greater improvements in the few NFEs regime.



Paperid:932
Authors:Minkyu Kim, Kiyoung Seong, Dongyeop Woo, Sungsoo Ahn, Minsu Kim
Abstract:
We address the challenge of training diffusion models to sample from unnormalized energy distributions in the absence of data, the so-called diffusion samplers. Although these approaches have shown promise, they struggle to scale in more demanding scenarios where energy evaluations are expensive and the sampling space is high-dimensional. To address this limitation, we propose a scalable and sample-efficient framework that properly harmonizes the powerful classical sampling method and the diffusion sampler. Specifically, we utilize Monte Carlo Markov chain (MCMC) samplers with a novelty-based auxiliary energy as a Searcher to collect off-policy samples, using an auxiliary energy function to compensate for exploring modes the diffusion sampler rarely visits. These off-policy samples are then combined with on-policy data to train the diffusion sampler, thereby expanding its coverage of the energy landscape. Furthermore, we identify primacy bias, i.e., the preference of samplers for early experience during training, as the main cause of mode collapse during training, and introduce a periodic re-initialization trick to resolve this issue. Our method significantly improves sample efficiency on standard benchmarks for diffusion samplers and also excels at higher-dimensional problems and real-world molecular conformer generation.



Paperid:933
Authors:Leheng Sheng, An Zhang, Zijian Wu, Weixiang Zhao, Changshuo Shen, zhang yi, Xiang Wang, Tat-Seng Chua
Abstract:
Recent studies empirically reveal that large reasoning models (LRMs) can automatically allocate more reasoning strengths (\ie the number of reasoning tokens) for harder problems, exhibiting difficulty-awareness for better task performance.While this automatic reasoning strength allocation phenomenon has been widely observed, its underlying mechanism remains largely unexplored. To this end, we provide explanations for this phenomenon from the perspective of model activations.\textbf{We find evidence that LRMs pre-plan the reasoning strengths in their activations even before generation, with this reasoning strength causally controlled by the magnitude of a pre-allocated directional vector.}Specifically, we show that the number of reasoning tokens is predictable solely based on the question activations using linear probes, indicating that LRMs estimate the required reasoning strength in advance.We then uncover that LRMs encode this reasoning strength through a pre-allocated directional vector embedded in the activations of the model, where the vector’s magnitude modulates the reasoning strength. Subtracting this vector can lead to reduced reasoning token number and performance, while adding this vector can lead to increased reasoning token number and even improved performance.We further reveal that this direction vector consistently yields positive reasoning length prediction, and it modifies the logits of end-of-reasoning token \texttt{} to affect the reasoning length.Finally, we demonstrate two potential applications of our findings: overthinking behavior detection and enabling efficient reasoning on simple problems.Our work provides new insights into the internal mechanisms of reasoning in LRMs and offers practical tools for controlling their reasoning behaviors.Our code is available at \url{https://anonymous.4open.science/r/LRM-plans-CoT-7E04}.



Paperid:934
Authors:Yixin Liu, Pengfei Liu, Arman Cohan
Abstract:
Evaluating large language models' (LLMs) alignment with human preferences typically involves directly assessing their open-ended responses, requiring human annotators or strong LLM judges. Conversely, LLMs themselves have also been extensively evaluated as judges for assessing alignment. In this work, we examine the relationship between LLMs' generation and evaluation capabilities in aligning with human preferences. To this end, we first conduct a comprehensive analysis of the generation-evaluation consistency (GE-consistency) among various LLMs, revealing a strong correlation between their generation and evaluation capabilities when evaluated by a strong LLM preference oracle (GPT-4o). Utilizing this finding, we propose a benchmarking paradigm that measures LLM alignment with human preferences without directly evaluating their generated outputs, instead assessing LLMs in their role as evaluators. Our evaluation shows that the proposed benchmark matches or surpasses widely used automatic LLM evaluation benchmarks, such as AlpacaEval and Arena-Hard, in capturing human preferences when ranking LLMs. Our study offers valuable insights into the connection between LLMs' generation and evaluation capabilities, and introduces a benchmark that assesses alignment without directly evaluating model outputs.



Paperid:935
Authors:Jacqueline He, Howard Yen, Margaret Li, Stella Li, Zhiyuan Zeng, Weijia Shi, Yulia Tsvetkov, Danqi Chen, Pang Wei Koh, Luke Zettlemoyer
Abstract:
A central challenge in modern language models (LMs) is intrinsic hallucination: the generation of information that is plausible but unsubstantiated relative to input context. To study this problem, we propose Precise Information Control (PIC), a new task formulation that requires models to generate long-form outputs grounded in a provided set of input statements, without adding any unsupported ones. While intrinsic hallucination is commonly evaluated as a binary judgment, we instantiate PIC at the level of short, self-contained statements known as verifiable claims. PIC includes a full setting, which tests a model’s ability to comprehensively include all input claims, and a partial setting, which requires the model to selectively include only relevant claims. We present PIC-Bench, a benchmark of eight long-form generation tasks (e.g., summarization, biography generation) adapted to the PIC setting, where LMs are supplied with well-formed, verifiable input claims. We evaluate a range of open and proprietary models on PIC-Bench; surprisingly, state-of-the-art LMs still intrinsically hallucinate in over 70% of outputs. To alleviate this lack of faithfulness, we introduce a post-training framework, using a weakly supervised preference data construction method, to train an 8B PIC-LM with stronger PIC ability---notably improving from 69.1% to 89.7% F1 in the full setting. When integrated into end-to-end generation pipelines, PIC-LM yields better end-task factual accuracy, improving by 14% in exact match recall on ambiguous QA with retrieval, and by 21% in precision on factual verification tasks, underscoring the broad potential of precisely grounded generation.



Paperid:936
Authors:Ting-Wei Li, Ruizhong Qiu, Hanghang Tong
Abstract:
Graph domain adaptation (GDA) is a fundamental task in graph machine learning, with techniques like shift-robust graph neural networks (GNNs) and specialized training procedures to tackle the distribution shift problem. Although these model-centric approaches show promising results, they often struggle with severe shifts and constrained computational resources. To address these challenges, we propose a novel model-free framework, GRADATE (GRAph DATa sElection), that selects the best training data from the source domain for the classification task on the target domain. GRADATE picks training samples without relying on any GNN model’s predictions or training recipes, leveraging optimal transport theory to capture and adapt to distribution changes. GRADATE is data-efficient, scalable and meanwhile complements existing model-centric GDA approaches. Through comprehensive empirical studies on several real-world graph-level datasets and multiple covariate shift types, we demonstrate that GRADATE outperforms existing selection methods and enhances off-the-shelf GDA methods with much fewer training data.



Paperid:937
Authors:Haipeng Luo, Spandan Senapati, Vatsal Sharan
Abstract:
Abstract:Calibration is a fundamental concept that aims at ensuring the reliability of probabilistic predictions by aligning them with real-world outcomes. There is a surge of studies on new calibration measures that are easier to optimize compared to the classical $\ell_1$-Calibration while still having strong implications for downstream applications. One recent such example is the work by Fishelson et al. (2025) who show that it is possible to achieve $\tilde{\mathcal{O}}(T^{1/3})$ pseudo $\ell_{2}$-Calibration error via minimizing pseudo swap regret of the squared loss, which in fact implies the same bound for all bounded proper losses with a smooth univariate form. In this work, we significantly generalize their result in the following ways: (a) in addition to smooth univariate forms, our algorithm also simultaneously achieves $\tilde{\mathcal{O}}(T^{1/3})$ swap regret for any proper loss with a twice continuously differentiable univariate form (such as Tsallis entropy); (b) our bounds hold not only for pseudo swap regret that measures losses using the forecaster's distributions on predictions, but also hold for the actual swap regret that measures losses using the forecaster's actual realized predictions. We achieve so by introducing a new stronger notion of calibration called (pseudo) KL-Calibration, which we show is equivalent to the (pseudo) swap regret with respect to log loss. We prove that there exists an algorithm that achieves $\tilde{\mathcal{O}}(T^{1/3})$ KL-Calibration error and provide an explicit algorithm that achieves $\tilde{\mathcal{O}}(T^{1/3})$ pseudo KL-Calibration error. Moreover, we show that the same algorithm achieves ${\mathcal{O}}(T^{1/3} (\log T) ^ {-\frac{1}{3}}\log (T/{\delta}))$ swap regret with probability at least $1-\delta$ for any proper loss with a smooth univariate form, which implies $\tilde{\mathcal{O}}(T^{1/3})$ $\ell_2$-Calibration error. A technical contribution of our work is a new randomized rounding procedure and a non-uniform discretization scheme to minimize the swap regret for log loss.



Paperid:938
Authors:Jiapeng Tang, Matthew Levine, Dor Verbin, Stephan Garbin, Matthias Niessner, Ricardo Martin Brualla, Pratul Srinivasan, Philipp Henzler
Abstract:
We introduce ROGR, a novel approach that reconstructs a relightable 3D model of an object captured from multiple views, driven by a generative relighting model that simulates the effects of placing the object under novel environment illuminations. Our method samples the appearance of the object under multiple lighting environments, creating a dataset that is used to train a lighting-conditioned Neural Radiance Field (NeRF) that outputs the object's appearance under any input environmental lighting. The lighting-conditioned NeRF uses a novel dual-branch architecture to encode the general lighting effects and specularities separately. The optimized lighting-conditioned NeRF enables efficient feed-forward relighting under arbitrary environment maps without requiring per-illumination optimization or light transport simulation. We evaluate our approach on the established TensoIR and Stanford-ORB datasets, where it improves upon the state-of-the-art on most metrics, and showcase our approach on real-world object captures.



Paperid:939
Authors:Jake Robertson, Arik Reuter, Siyuan Guo, Noah Hollmann, Frank Hutter, Bernhard Schölkopf
Abstract:
Causal effect estimation is critical to a range of scientific disciplines. Existing methods for this task either require interventional data, knowledge about the ground-truth causal graph, or rely on assumptions such as unconfoundedness, restricting their applicability in real-world settings. In the domain of tabular machine learning, Prior-data fitted networks (PFNs) have achieved state-of-the-art predictive performance, having been pre-trained on synthetic data to solve tabular prediction problems via in-context learning. To assess whether this can be transferred to the harder problem of causal effect estimation, we pre-train PFNs on synthetic data drawn from a wide variety of causal structures, including interventions, to predict interventional outcomes given observational data. Through extensive experiments on synthetic case studies, we show that our approach allows for the accurate estimation of causal effects without knowledge of the underlying causal graph. We also perform ablation studies that elucidate Do-PFN's scalability and robustness across datasets with a variety of causal characteristics.



Paperid:940
Authors:Insu Han, Michael Kapralov, Ekaterina Kochetkova, Kshiteej Sheth, Amir Zandieh
Abstract:
Abstract:Large language models (LLMs) have achieved impressive success, but their high memory requirements present challenges for long-context token generation. In this paper we study the streaming complexity of attention approximation, a key computational primitive underlying token generation. Our main contribution is BalanceKV, a streaming algorithm for $\epsilon$-approximating attention computations based on geometric process for selecting a balanced collection of Key and Value tokens as per Banaszczyk's vector balancing theory. We complement our algorithm with space lower bounds for streaming attention computation. Besides strong theoretical guarantees, BalanceKV exhibits empirically validated performance improvements over existing methods, both for attention approximation and end-to-end performance on various long context benchmarks.



Paperid:941
Authors:Guanhua Zhang, Florian E. Dorner, Moritz Hardt
Abstract:
Large language model (LLM) evaluation is increasingly costly, prompting interest in methods that speed up evaluation by shrinking benchmark datasets. Benchmark prediction (also called efficient LLM evaluation) aims to select a small subset of evaluation points and predict overall benchmark performance from that subset. In this paper, we systematically assess the strengths and limitations of 11 benchmark prediction methods across 19 diverse benchmarks. First, we identify a highly competitive baseline: Take a random sample and fit a regression model on the sample to predict missing entries. Outperforming most existing methods, this baseline challenges the assumption that careful subset selection is necessary for benchmark prediction. Second, we discover that all existing methods crucially depend on model similarity. They work best when interpolating scores among similar models. The effectiveness of benchmark prediction sharply declines when new models have higher accuracy than previously seen models. In this setting of extrapolation, none of the previous methods consistently beat a simple average over random samples. To improve over the sample average, we introduce a new method inspired by augmented inverse propensity weighting. This method consistently outperforms the random sample average even for extrapolation. However, its performance still relies on model similarity and the gains are modest in general. This shows that benchmark prediction fails just when it is most needed: at the evaluation frontier, where the goal is to evaluate new models of unknown capabilities.



Paperid:942
Authors:Piersilvio De Bartolomeis, Javier Abad, Guanbo Wang, Konstantin Donhauser, Raymond Duch, Fanny Yang, Issa Dahabreh
Abstract:
Randomized experiments are the preferred approach for evaluating the effects of interventions, but they are costly and often yield estimates with substantial uncertainty. On the other hand, in silico experiments leveraging foundation models offer a cost-effective alternative that can potentially attain higher statistical precision. However, the benefits of in silico experiments come with a significant risk: statistical inferences are not valid if the models fail to accurately predict experimental responses to interventions.In this paper, we propose a novel approach that integrates the predictions from multiple foundation models with experimental data while preserving valid statistical inference. Our estimator is consistent and asymptotically normal, with asymptotic variance no larger than the standard estimator based on experimental data alone. Importantly, these statistical properties hold even when model predictions are arbitrarily biased. Empirical results across several randomized experiments show that our estimator offers substantial precision gains, equivalent to a reduction of up to 20\% in the sample size needed to match the same precision as the standard estimator based on experimental data alone.



Paperid:943
Authors:Youwei Pang, Xiaoqi Zhao, Lihe Zhang, Huchuan Lu, Georges Fakhri, Xiaofeng Liu, Shijian Lu
Abstract:
As an essential vision task, infrared small target detection (IRSTD) has seen significant advancements through deep learning. However, critical limitations in current evaluation protocols impede further progress. First, existing methods rely on fragmented pixel- and target-level specific metrics, which fails to provide a comprehensive view of model capabilities. Second, an excessive emphasis on overall performance scores obscures crucial error analysis, which is vital for identifying failure modes and improving real-world system performance. Third, the field predominantly adopts dataset-specific training-testing paradigms, hindering the understanding of model robustness and generalization across diverse infrared scenarios. This paper addresses these issues by introducing a hybrid-level metric incorporating pixel- and target-level performance, proposing a systematic error analysis method, and emphasizing the importance of cross-dataset evaluation. These aim to offer a more thorough and rational hierarchical analysis framework, ultimately fostering the development of more effective and robust IRSTD models. An open-source toolkit has be released to facilitate standardized benchmarking.



Paperid:944
Authors:Yanbo Wang, Zixiang Xu, Yue Huang, Xiangqi Wang, Zirui Song, Lang Gao, Chenxi Wang, Robert Tang, Yue Zhao, Arman Cohan, Xiangliang Zhang, Xiuying Chen
Abstract:
Agent systems based on large language models (LLMs) have shown great potential in complex reasoning tasks, but building efficient and generalizable workflows remains a major challenge. Most existing approaches rely on manually designed processes, which limits their adaptability across different tasks. While a few methods attempt automated workflow generation, they are often tied to specific datasets or query types and make limited use of intermediate feedback, reducing system robustness and reasoning depth. Moreover, their operations are typically predefined and inflexible.To address these limitations, we proposeDyFlow, a dynamic workflow generation framework that adaptively constructs and adjusts reasoning procedures based on task requirements and real-time intermediate feedback, thereby enhancing cross-task generalization.DyFlow consists of two core components: a designer and an executor. The designer decomposes complex problems into a sequence of sub-goals defined by high-level objectives and dynamically plans the next steps based on intermediate outputs and feedback. These plans are then carried out by the executor, which executes each operation using dynamic operators with context-aware parameterization, enabling flexible and semantically grounded reasoning.We systematically evaluate DyFlow across diverse domains, including social reasoning, biomedical tasks, mathematical problem solving, and code generation.Results demonstrate that DyFlow significantly outperforms existing baselines, achieving substantial Pass@k improvements and exhibiting robust generalization across diverse domains.



Paperid:945
Authors:Marco Spinaci, Marek Polewczyk, Maximilian Schambach, Sam Thelin
Abstract:
Tabular in-context learning (ICL) has recently achieved state-of-the-art (SOTA) performance on several tabular prediction tasks. Previously restricted to classification problems on small tables, recent advances such as TabPFN and TabICL have extended its use to larger datasets. While being architecturally efficient and well-adapted to tabular data structures, current table-native ICL architectures, being trained exclusively on synthetic data, do not fully leverage the rich semantics and world knowledge contained in real-world tabular data. On another end of this spectrum, tabular ICL models based on pretrained large language models such as TabuLa-8B integrate deep semantic understanding and world knowledge but are only able to make use of a small amount of context due to inherent architectural limitations. With the aim to combine the best of both these worlds, we introduce ConTextTab, integrating semantic understanding and alignment into a table-native ICL framework. By employing specialized embeddings for different data modalities and by training on large-scale real-world tabular data, our model is competitive with SOTA across a broad set of benchmarks while setting a new standard on the semantically rich CARTE benchmark.



Paperid:946
Authors:Honglin Liu, Chao Sun, Peng Hu, Yunfan Li, Xi Peng
Abstract:
Conventional representation learning methods learn a universal representation that primarily captures dominant semantics, which may not always align with customized downstream tasks. For instance, in animal habitat analysis, researchers prioritize scene-related features, whereas universal embeddings emphasize categorical semantics, leading to suboptimal results. As a solution, existing approaches resort to supervised fine-tuning, which however incurs high computational and annotation costs. In this paper, we propose Conditional Representation Learning (CRL), aiming to extract representations tailored to arbitrary user-specified criteria. Specifically, we reveal that the semantics of a space are determined by its basis, thereby enabling a set of descriptive words to form the basis for a customized feature space. Building upon this insight, given a user-specified criterion, CRL first employs a large language model (LLM) to generate descriptive texts to construct the semantic basis, then projects the image representation into this conditional feature space leveraging a vision-language model (VLM). The transformed representation better captures semantics for the specific criterion, which could be utilized for customized tasks. Extensive experiments on customized downstream classification and retrieval demonstrate the superiority and generality of the proposed CRL. The code will be released.



Paperid:947
Authors:Coleman Hooper, Sebastian Zhao, Luca Manolache, Sehoon Kim, Michael Mahoney, Sophia Shao, Kurt Keutzer, Amir Gholami
Abstract:
Abstract:Large Reasoning Models (LRMs) have shown promising accuracy improvements on complex problem-solving tasks. While these models have attained high accuracy by leveraging additional computation at test time, they need to generate long chain-of-thought reasoning in order to think before answering, which requires generating thousands of tokens.While sparse attention methods can help reduce the KV cache pressure induced by this long autoregressive reasoning, these methods can introduce errors which disrupt the reasoning process.Our work addresses these challenges by introducing Multipole Attention, which accelerates autoregressive reasoning by only computing exact attention for the most important tokens, while maintaining approximate representations for the remaining tokens. Our method first performs clustering to group together semantically similar key vectors, and then uses the cluster centroids both to identify important key vectors and to approximate the remaining key vectors in order to retain high accuracy.Additionally, in order to accelerate long generation tasks, we design a fast cluster update process to quickly re-cluster the input and previously generated tokens, thereby allowing for accelerating attention to the previous output tokens.We evaluate our method using emerging LRMs such as Qwen-8B and Deepseek-R1-Distil-Qwen2.5-14B, demonstrating that our approach can maintain accuracy on complex reasoning tasks even with aggressive attention sparsity settings.We also provide kernel implementations to demonstrate the practical efficiency gains from our method, achieving up to 4.5$\times$ speedup for attention in long-context reasoning applications.



Paperid:948
Authors:Nayoung Kim, Seongsu Kim, Sungsoo Ahn
Abstract:
Designing metal–organic frameworks (MOFs) with novel chemistries is a long-standing challenge due to their large combinatorial space and the complex 3D arrangements of building blocks. While recent deep generative models have enabled scalable MOF generation, they assume (1) a fixed set of building blocks and (2) known ground-truth local block-wise 3D coordinates. However, this limits their ability to (1) design novel MOFs and (2) generate the structure using novel building blocks. We propose a two-stage de novo MOF generation framework that overcomes these limitations by modeling both chemical and geometric degrees of freedom. First, we train a SMILES-based autoregressive model to generate novel metal and organic building blocks, paired with cheminformatics for 3D structure initialization. Second, we introduce a flow-matching model that predicts translations, rotations, and torsional angles to assemble flexible blocks into valid 3D frameworks. Our experiments demonstrate improved reconstruction accuracy, the generation of valid, novel, and unique MOFs, and the ability of our model to create novel building blocks.



Paperid:949
Authors:Zhiyu Zhao, Haoxuan Li, Haifeng Zhang, Jun Wang, Francesco Faccio, Jürgen Schmidhuber, Mengyue Yang
Abstract:
When building a world model, a common assumption is that the environment has a single, unchanging underlying causal rule, like applying Newton's laws to every situation. In reality, what appears as a drifting causal mechanism is often the manifestation of a fixed underlying mechanism seen through a narrow observational window. This brings about a problem that, when building a world model, even subtle shifts in policy or environment states can alter the very observed causal mechanisms. In this work, we introduce the Meta-Causal Graph as world models, a minimal unified representation that efficiently encodes the transformation rules governing how causal structures shift across different latent world states. A single Meta-Causal Graph is composed of multiple causal subgraphs, each triggered by meta state, which is in the latent state space. Building on this representation, we introduce a Causality-Seeking Agent whose objectives are to (1) identify the meta states that trigger each subgraph, (2) discover the corresponding causal relationships by agent curiosity-driven intervention policy, and (3) iteratively refine the Meta-Causal Graph through ongoing curiosity-driven exploration and agent experiences. Experiments on both synthetic tasks and a challenging robot arm manipulation task demonstrate that our method robustly captures shifts in causal dynamics and generalizes effectively to previously unseen contexts.



Paperid:950
Authors:Hongbin Zhang, Kehai Chen, Xuefeng Bai, Xiucheng Li, Yang Xiang, Min Zhang
Abstract:
While large language models (LLMs) demonstrate remarkable success in multilingual translation, their internal core translation mechanisms, even at the fundamental word level, remain insufficiently understood.To address this critical gap, this work introduces a systematic framework for interpreting the mechanism behind LLM translation from the perspective of computational components. This paper first proposes subspace-intervened path patching for precise, fine-grained causal analysis, enabling the detection of components crucial to translation tasks and subsequently characterizing their behavioral patterns in human-interpretable terms.Comprehensive experiments reveal that translation is predominantly driven by a sparse subset of components: specialized attention heads serve critical roles in extracting source language, translation indicators, and positional features, which are then integrated and processed by specific multi-layer perceptrons (MLPs) into intermediary English-centric latent representations before ultimately yielding the final translation.The significance of these findings is underscored by the empirical demonstration that targeted fine-tuning a minimal parameter subset (<5%) enhances translation performance while preserving general capabilities. This result further indicates that these crucial components generalize effectively to sentence-level translation and are instrumental in elucidating more intricate translation tasks.



Paperid:951
Authors:Jialong Wu, Baixuan Li, Runnan Fang, Wenbiao Yin, Liwen Zhang, Zhenglin Wang, Zhengwei Tao, Ding-Chu Zhang, Zekun Xi, Robert Tang, Yong Jiang, Pengjun Xie, Fei Huang, Jingren Zhou
Abstract:
Addressing intricate real-world problems necessitates in-depth information seeking and multi-step reasoning. Recent progress in agentic systems, exemplified by Deep Research, underscores the potential for autonomous multi-step research. In this work, we present a cohesive paradigm for building end-to-end agentic information seeking agents from a data-centric and training-stage perspective.Our approach consists of four key stages: (1) browsing data construction, (2) trajectories sampling, (3) supervised fine-tuning for effective cold start, and (4) reinforcement learning for enhanced generalisation.We instantiate this framework in a web agent based on the ReAct format, WebDancer.Empirical evaluations on the challenging GAIA and WebWalkerQA benchmarks demonstrate the strong performance of WebDancer, achieving considerable results and highlighting the efficacy of our training paradigm. Further analysis of agent training provides valuable insights and actionable, systematic pathways for developing more capable agentic models.



Paperid:952
Authors:Thomas Kleine Buening, Jiarui Gan, Debmalya Mandal, Marta Kwiatkowska
Abstract:
Abstract:We study Reinforcement Learning from Human Feedback (RLHF) in settings where multiple labelers may strategically misreport feedback to steer the learned policy toward their own preferences. We show that existing RLHF algorithms, including recent pluralistic methods, are not strategyproof, and that even a single strategic labeler can cause arbitrarily large misalignment with social welfare. Moreover, we prove that, in the worst case, any strategyproof RLHF algorithm must perform $k$-times worse than the optimal policy, where $k$ is the number of labelers. This suggests a fundamental trade-off between incentive alignment (ensuring labelers report truthfully) and policy alignment (maximizing social welfare). To address this, we propose the Pessimistic Median of MLEs algorithm, which, under appropriate policy coverage assumptions, is approximately strategyproof and converges to the optimal policy as the number of labelers and samples increases. Our results apply to both contextual bandits and Markov decision processes.



Paperid:953
Authors:Dheeraj Vattikonda, Santhoshi Ravichandran, Emiliano Penaloza, Hadi Nekoei, Thibault de Chezelles, Megh Thakkar, Nicolas Gontier, Miguel Muñoz-Mármol, Sahar Omidi Shayegan, Stefania Raimondo, Xue (Steve) Liu, Alexandre Drouin, Alexandre Piche, Alexandre Lacoste, Massimo Caccia
Abstract:
Large language model (LLM) agents for web interfaces have advanced rapidly, yet open-source systems still lag behind proprietary agents. Bridging this gap is key to enabling customizable, efficient, and privacy-preserving agents. Two challenges hinder progress: the reproducibility issues in RL and LLM agent training, where results often depend on sensitive factors like seeds and decoding parameters, and the focus of prior work on single-step tasks, overlooking the complexities of web-based, multi-step decision-making.We address these gaps by providing a statistically driven study of training LLM agents for web tasks. Our two-stage pipeline combines imitation learning from a Llama 3.3 70B teacher with on-policy fine-tuning via Group Relative Policy Optimization (GRPO) on a Llama 3.1 8B student. Through 240 configuration sweeps and rigorous bootstrapping, we chart the first compute allocation curve for open-source LLM web agents. Our findings show that dedicating one-third of compute to teacher traces and the rest to RL improves MiniWoB++ success by 6 points and closes 60\% of the gap to GPT-4o on WorkArena, while cutting GPU costs by 45\%. We introduce a principled hyperparameter sensitivity analysis, offering actionable guidelines for robust and cost-effective agent training.



Paperid:954
Authors:Jongmin Lee, Ernest Ryu
Abstract:
Although there is an extensive body of work characterizing the sample complexity of discounted-return offline RL with function approximations, prior work on the average-reward setting has received significantly less attention, and existing approaches rely on restrictive assumptions, such as ergodicity or linearity of the MDP. In this work, we establish the first sample complexity results for average-reward offline RL with function approximation for weakly communicating MDPs, a much milder assumption. To this end, we introduce Anchored Fitted Q-Iteration, which combines the standard Fitted Q-Iteration with an anchor mechanism. We show that the anchor, which can be interpreted as a form of weight decay, is crucial for enabling finite-time analysis in the average-reward setting. We also extend our finite-time analysis to the setup where the dataset is generated from a single-trajectory rather than IID transitions, again leveraging the anchor mechanism.



Paperid:955
Authors:Yang Zhang, Rui Zhang, Jiaming Guo, Huang Lei, Di Huang, Yunpu Zhao, Shuyao Cheng, Pengwei Jin, Chongxiao Li, Zidong Du, Xing Hu, Qi Guo, Yunji Chen
Abstract:
The remarkable progress of Large Language Models (LLMs) presents promising opportunities for Verilog code generation which is significantly important for automated circuit design. The lacking of meaningful functional rewards hinders the preference optimization based on Reinforcement Learning (RL) for producing functionally correct Verilog code. In this paper, we propose Signal-Aware Learning for Verilog code generation (SAL-V) by leveraging code segmentas of functionally correct output signal to optimize RL training. Considering Verilog code specifies the structural interconnection of hardware gates and wires so that different output signals are independent, the key insight of SAL-V is to extract verified signal-aware implementations in partially incorrect modules, so as to enhance the extraction of meaningful functional rewards. Roughly, we verify the functional correctness of signals in generated module by comparing with that of reference module in the training data. Then abstract syntax tree (AST) is employed to identify signal-aware code segments which can provide meaningful functional rewards from erroneous modules. Finally, we introduce signal-aware DPO which is optimized on the correct signal-level code segments, thereby preventing noise and interference from incorrect signals.The proposed SAL-V underscores the paradigm shift from conventional module-level to fine-grained signal-level optimization in Verilog code generation, addressing the issue of insufficient functional rewards.Experiments demonstrate that our method achieves state-of-the-art performance on VerilogEval2.0 and RTLLM, with a 7B parameter model matching the performance of the DeepSeek v3 671B model and significantly outperforming the leading open-source model CodeV trained on the same dataset.



Paperid:956
Authors:Jinpei Guo, Yifei Ji, Zheng Chen, Kai Liu, Min Liu, Wang Rao, Wenbo Li, Yong Guo, Yulun Zhang
Abstract:
Pretrained latent diffusion models have shown strong potential for lossy image compression, owing to their powerful generative priors. Most existing diffusion-based methods reconstruct images by iteratively denoising from random noise, guided by compressed latent representations. While these approaches have achieved high reconstruction quality, their multi-step sampling process incurs substantial computational overhead. Moreover, they typically require training separate models for different compression bit-rates, leading to significant training and storage costs. To address these challenges, we propose a one-step diffusion codec across multiple bit-rates. termed OSCAR. Specifically, our method views compressed latents as noisy variants of the original latents, where the level of distortion depends on the bit-rate. This perspective allows them to be modeled as intermediate states along a diffusion trajectory. By establishing a mapping from the compression bit-rate to a pseudo diffusion timestep, we condition a single generative model to support reconstructions at multiple bit-rates. Meanwhile, we argue that the compressed latents retain rich structural information, thereby making one-step denoising feasible. Thus, OSCAR replaces iterative sampling with a single denoising pass, significantly improving inference efficiency. Extensive experiments demonstrate that OSCAR achieves superior performance in both quantitative and visual quality metrics. The code and models will be publicly available.



Paperid:957
Authors:Navdeep Kumar, Adarsh Gupta, Mohamed ELFATIHI, Giorgia Ramponi, Kfir Y. Levy, Shie Mannor
Abstract:
Abstract:Robust policy evaluation for non-rectangular uncertainty set is generally NP-hard, even in approximation. Consequently, existing approaches suffer from either exponential iteration complexity or significant accuracy gaps. Interestingly, we identify a powerful class of $L_p$-bounded uncertainty sets that avoid these complexity barriers due to their structural simplicity. We further show that this class can be decomposed into infinitely many \texttt{sa}-rectangular $L_p$-bounded sets and leverage its structural properties to derive a novel dual formulation for $L_p$ robust Markov Decision Processes (MDPs). This formulation provides key insights into the adversary’s strategy and enables the development of an efficient robust policy evaluation algorithm for these $L_p$ normed non-rectangular robust MDPs.



Paperid:958
Authors:Dekai Zhu, Yixuan Hu, Youquan Liu, Dongyue Lu, Lingdong Kong, Slobodan Ilic
Abstract:
Leveraging diffusion models, 3D LiDAR scene generation has achieved great success in both range-view and voxel-based representations. While recent voxel-based approaches can generate both geometric structures and semantic labels, existing range-view methods are limited to producing unlabeled LiDAR scenes. Relying on pretrained segmentation models to predict the semantic maps often results in suboptimal cross-modal consistency. To address this limitation while preserving the advantages of range-view representations, such as computational efficiency and simplified network design, we propose Spiral, a novel range-view LiDAR diffusion model that simultaneously generates depth, reflectance images, and semantic maps. Furthermore, we introduce novel semantic-aware metrics to evaluate the quality of the generated labeled range-view data. Experiments on SemanticKITTI and nuScenes datasets demonstrate that Spiral achieves state-of-the-art performance with the smallest parameter size, outperforming two-step methods that combine the best available generative and segmentation models. Additionally, we validate that Spiral’s generated range images can be effectively used for synthetic data augmentation in the downstream segmentation training, significantly reducing the labeling effort on LiDAR data.



Paperid:959
Authors:Ruilong Li, Brent Yi, Junchen Liu, Hang Gao, Yi Ma, Angjoo Kanazawa
Abstract:
As transformers become increasingly adopted for multi-view 3D perception tasks, a key challenge lies in effectively conditioning these models on camera parameters that ground visual observations in 3D space. We explore the generalization capabilities of these multi-view transformers, systematically analyzing how different approaches to camera conditioning affect model performance and robustness. Our analysis reveals that while existing attention-level conditioning methods show good generalization in terms of SE(3) transformations, they struggle when faced with varying intrinsic parameters at test time. Through careful theoretical analysis, we identify that the key to robust generalization lies in encoding the complete projective relationship between cameras, which naturally encompasses both intrinsic and extrinsic parameters while maintaining invariance to the choice of global reference frame. This insight leads to Projective Positional Encoding (PRoPE), which emerges as a natural theoretical extension of existing approaches. We demonstrate the effectiveness of PRoPE across three diverse tasks: novel view synthesis, stereo depth estimation, and multi-view spatial cognition. Our method shows strong performance and robustness, particularly when handling varying focal lengths and number of views at test time. PRoPE integrates seamlessly into existing transformer architectures, including CAT3D and UniMatch, yielding consistent improvements with minimal computational overhead. The method is flash attention-friendly and will be open-sourced.



Paperid:960
Authors:Alex McKenzie, Philipp Blandfort, Urja Pawar, William Bankes, David Krueger, Ekdeep S Lubana, Dmitrii Krasheninnikov
Abstract:
Monitoring is an important aspect of safely deploying Large Language Models (LLMs). This paper examines activation probes as a computationally efficient technique to detect "high-stakes" interactions---where the text indicates the interaction might lead to significant harm---as a critical, yet underexplored, target for such monitoring.We train several novel probe architectures on synthetic data and find they exhibit robust generalization performance (mean AUROC > 0.91) on diverse, out-of-distribution, real-world data. Their performance is comparable to that of prompted or fine-tuned medium-sized LLM monitors, while offering computational savings of six orders of magnitude.Furthermore, this research establishes a foundation for building resource-aware monitoring systems where probes serve as an initial, resource-efficient filter in a cascaded system, flagging cases for more specialized and expensive downstream analysis.Finally, we release our novel synthetic dataset and codebase to encourage further investigation.



Paperid:961
Authors:Licong Lin, Jingfeng Wu, Peter Bartlett
Abstract:
Abstract:Neural scaling laws suggest that the test error of large language models trained online decreases polynomially as the model size and data size increase. However, such scaling can be unsustainable when running out of new data. In this work, we show that data reuse can improve existing scaling laws in linear regression. Specifically, we derive sharp test error bounds on $M$-dimensional linear models trained by multi-pass *stochastic gradient descent* (multi-pass SGD) on $N$ data with sketched features. Assuming that the data covariance has a power-law spectrum of degree $a$, and that the true parameter follows a prior with an aligned power-law spectrum of degree $b-a$ (with $a > b > 1$), we show that multi-pass SGD achieves a test error of $\Theta(M^{1-b} + L^{(1-b)/a})$, where $L \lesssim N^{a/b}$ is the number of iterations. In the same setting, one-pass SGD only attains a test error of $\Theta(M^{1-b} + N^{(1-b)/a})$ (see, e.g., Lin et al., 2024). This suggests an improved scaling law via data reuse (i.e., choosing $L>N$) in data-constrained regimes. Numerical simulations are also provided to verify our theoretical findings.



Paperid:962
Authors:Jiahao Yu, Xian Wu, Hao Liu, Wenbo Guo, Xinyu Xing
Abstract:
We propose BlockScan, a customized Transformer for anomaly detection in blockchain transactions.Unlike existing methods that rely on rule-based systems or directly apply off-the-shelf large language models (LLMs), BlockScan introduces a series of customized designs to effectively model the unique data structure of blockchain transactions.First, a blockchain transaction is multi-modal, containing blockchain-specific tokens, texts, and numbers.We design a novel modularized tokenizer to handle these multi-modal inputs, balancing the information across different modalities.Second, we design a customized masked language modeling mechanism for pretraining the Transformer architecture, incorporating RoPE embedding and FlashAttention for handling longer sequences.Finally, we design a novel anomaly detection method based on the model outputs.We further provide theoretical analysis for the detection method of our system. Extensive evaluations on Ethereum and Solana transactions demonstrate BlockScan's exceptional capability in anomaly detection while maintaining a low false positive rate.Remarkably, BlockScan is the only method that successfully detects anomalous transactions on Solana with high accuracy, whereas all other approaches achieved very low or zero detection recall scores.This work sets a new benchmark for applying Transformer-based approaches in blockchain data analysis.



Paperid:963
Authors:Ren Yang, Jiahao Li, Yan Lu
Abstract:
Reconstruction of 3D scenes from a single image is a crucial step towards enabling next-generation AI-powered immersive experiences. However, existing diffusion-based methods often struggle with reconstructing omnidirectional scenes due to geometric distortions and inconsistencies across the generated novel views, hindering accurate 3D recovery. To overcome this challenge, we propose Omni3D, an approach designed to enhance the geometric fidelity of diffusion-generated views for robust omnidirectional reconstruction. Our method leverages priors from pose estimation techniques, such as MASt3R, to iteratively refine both the generated novel views and their estimated camera poses. Specifically, we minimize the 3D reprojection errors between paired views to optimize the generated images, and simultaneously, correct the pose estimation based on the refined views. This synergistic optimization process yields geometrically consistent views and accurate poses, which are then used to build an explicit 3D Gaussian Splatting representation capable of omnidirectional rendering. Experimental results validate the effectiveness of Omni3D, demonstrating significantly advanced 3D reconstruction quality in the omnidirectional space, compared to previous state-of-the-art methods. Codes will be publicly released on Github later.



Paperid:964
Authors:Gergely Csáji, Alexander Gundert, Jörg Rothe, Ildikó Schlotter
Abstract:
We study fundamental connections between coalition formation games and clustering, illustrating the cross-disciplinary relevance of these concepts. We focus on graphical hedonic games where agents' preferences are compactly represented by a friendship graph and an enemy graph. In the context of clustering, friendship relations naturally align with data point similarities, whereas enmity corresponds to dissimilarities. We consider two stability notions based on single-agent deviations: local popularity and local stability. Exploring these concepts from an algorithmic viewpoint, we design efficient mechanisms for finding locally stable or locally popular partitions. Besides gaining theoretical insight into the computational complexity of these problems, we perform simulations that demonstrate how our algorithms can be successfully applied in clustering and community detection. Our findings highlight the interplay between coalition formation games and data-driven clustering techniques, offering fresh perspectives and applications in both areas.



Paperid:965
Authors:Jialu Li, Yu Wang, Pengfei Zhu, Wanyu Lin, Xinjie Yao, Qinghua Hu
Abstract:
Graphs in the real world are fragmented and dynamic, lacking collaboration akin to that observed in human societies. Existing paradigms present collaborative information collapse and forgetting, making collaborative relationships poorly autonomous and interaction information insufficient. Moreover, collaborative information is prone to loss when graph grows. Effective collaboration in heterogeneous dynamic graph environments becomes challenging. Inspired by social learning, this paper presents a Graph Socialized Learning (GSL) paradigm. We provide insights into graph socialization in GSL and boost performance of agents through effective collaboration. It is crucial to determine with whom, what, and when to share and accumulate information for effective GSL. Thus, we propose ''Graphs Help Graphs'' (GHG) method to solve these issues. Specifically, it uses graph-driven organizational structure to autonomously select interacting agents and manage interaction strength. We produce customized synthetic graphs as interaction medium based on demand of agents, then apply synthetic graphs to build prototypes in life cycle to help select optimal parameters. We demonstrate the effectiveness of GHG in heterogeneous dynamic graphs by extensive empirical study.



Paperid:966
Authors:Daniel Berg Thomsen, Adrien Taylor, Aymeric Dieuleveut
Abstract:
Communication between agents often constitutes a major computational bottleneck in distributed learning. One of the most common mitigation strategies is to compress the information exchanged, thereby reducing communication overhead. To counteract the degradation in convergence associated with compressed communication, error feedback schemes—most notably EF and EF21—were introduced. In this work, we provide a tight analysis of both of these methods. Specifically, we find the Lyapunov function that yields the best possible convergence rate for each method—with matching lower bounds. This principled approach yields sharp performance guarantees and enables a rigorous, apples-to-apples comparison between EF, EF21, and Compressed Gradient Descent. Our analysis is carried out in a simplified yet representative setting, which allows for clean theoretical insights and fair comparison of the underlying mechanisms.



Paperid:967
Authors:chuan guo, Inwoo Hwang, Jian Wang, Bing Zhou
Abstract:
Text-to-motion generation has experienced remarkable progress in recent years. However, current approaches remain limited to synthesizing motion from short or general text prompts, primarily due to dataset constraints. This limitation undermines fine-grained controllability and generalization to unseen prompts. In this paper, we introduce OmniMotion, a new text-motion dataset featuring high-quality motion capture data paired with accurate, \textit{expressive} textual annotations. The dataset comprises 20K motion clips totaling 44 hours, accompanied by 122 detailed textual descriptions averaging 48 words per description (vs. 12 words of HumanML3D). Importantly, these motion clips preserve original temporal continuity as they were in long sequences, facilitating research in long-term motion generation and blending. We also improve upon previous generative masked modeling approaches. Our model, MoMask++, transforms motion into \textbf{multi-scale} token sequences that better exploit the token capacity, and learns to generate all tokens using a single generative masked transformer. MoMask++ achieves state-of-the-art performance on both HumanML3D and OmniMotion benchmarks. Additionally, we demonstrate the ability to process casual user prompts by employing an LLM to reformat inputs to align with the expressivity and narration style of OmniMotion.



Paperid:968
Authors:Tim Hua, James Baskerville, Henri Lemoine, Mia Hopman, Aryan Bhatt, Tyler Tracy
Abstract:
Monitoring AIs at runtime could help us detect and stop harmful actions. However, both monitoring and safety interventions (e.g., human audits) incur costs. In this paper, we study how to combine multiple runtime monitors into a single monitoring protocol. The protocol's objective is to maximize the probability of applying a safety intervention on any misaligned outputs (i.e., maximize recall). Simultaneously, it must adhere to an average-case budget constraint over monitor calls and safety interventions. Taking monitors' performance and cost as given, we find the best protocol by exhaustively searching over when to call monitors and which monitors to call. Given a monitor invocation strategy, we allocate the safety intervention to the most suspicious outputs as measured by likelihood ratios. Finally, we pick the best monitor invocation strategy based on its recall rate and use it as our protocol. By focusing on likelihood ratios and strategically trading off spending on monitors and spending on interventions, we can more than double our recall rate compared to the naive baseline in a code review setting. We also show that combining two monitors can Pareto dominate using either monitor alone. Our framework provides a principled methodology for combining existing monitors to detect undesirable behavior in cost-sensitive settings.



Paperid:969
Authors:Francesco Innocenti, El Mehdi Achour, Christopher L Buckley
Abstract:
Abstract:The biological implausibility of backpropagation (BP) has motivated many alternative, brain-inspired algorithms that attempt to rely only on local information, such as predictive coding (PC) and equilibrium propagation. However, these algorithms have notoriously struggled to train very deep networks, preventing them from competing with BP in large-scale settings. Indeed, scaling PC networks (PCNs) has recently been posed as a challenge for the community (Pinchetti et al., 2024). Here, we show that 100+ layer PCNs can be trained reliably using a Depth-$\mu$P parameterisation (Yang et al., 2023; Bordelon et al., 2023) which we call "$\mu$PC". Through an extensive analysis of the scaling behaviour of PCNs, we reveal several pathologies that make standard PCNs difficult to train at large depths. We then show that, despite addressing only some of these instabilities, $\mu$PC can train very deep (up to 128-layer) residual networks on simple classification tasks with competitive performance and little tuning compared to current benchmarks. Moreover, $\mu$PC enables zero-shot transfer of both weight and activity learning rates across widths and depths. Our results have implications for other local algorithms and could be extended to convolutional and transformer architectures.



Paperid:970
Authors:Dmitri Chklovskii, Xuehao Ding, Charles Epstein, Philip Greengard, Jason J. Moore, Joshua L Pughe-Sanford, Anirvan Sengupta
Abstract:
We model sensory streams as observations from high-dimensional stochastic dynamical systems and conceptualize sensory neurons as self-supervised learners of compact representations of such dynamics. From previously encountered stimuli, neurons learn \emph{coherent sets}—regions of stimulus state space whose trajectories evolve cohesively over finite times—and compute membership indices for new stimuli. Coherent sets are discovered via spectral clustering of the \emph{stochastic Koopman operator (SKO)}, where the sign pattern of the subdominant singular function partitions the state space into minimally coupled regions. For multivariate Ornstein–Uhlenbeck processes, the subdominant singular function reduces to a linear projection of inputs onto the dominant singular vector of the whitened state-transition matrix, enabling neurons to compute membership indices through a simple filter-and-rectify operation. These indices support near-future prediction in contracting coherent sets and retrospection-enhanced representations in expanding coherent sets, suggesting a functional dichotomy between prospective and retrospective neuron types. We extend this approach to nonlinear dynamics via Galerkin projections of the SKO followed by past-future canonical correlation analysis, providing a biologically plausible, data-driven route to the requisite singular vectors. This framework accounts for neuronal temporal receptive fields, the ubiquity of rectification in neuronal responses, and the predictive versus retrospective specialization seen in tufted versus mitral cells and non-lagged versus lagged cells. Coherent set detection thus emerges as a fundamental neural computation for modeling biological sensory processing and designing artificial systems.



Paperid:971
Authors:Jane Lange, Arsen Vasilyan
Abstract:
Abstract:We say that a classifier is $\text{\emph{adversarially robust}}$ to perturbations of norm $r$ if,with high probability over a point $x$ drawn from the input distribution, there is no point within distance $\le r$ from $x$ that is classified differently. The $\text{\emph{boundary volume}}$ is the probability that a point falls within distance $r$ of a point with a different label.This work studies the task of learning a hypothesis with small boundary volume, where the input is distributed as a subgaussian isotropic log-concave distribution over $\mathbb{R}^d$.Linear threshold functions are adversarially robust; they have boundary volume proportional to $r$.Such concept classes are efficiently learnable by polynomialregression, which produces a polynomial threshold function (PTF), but PTFs in general may have boundary volume $\Omega(1)$, even for $r \ll 1$.We give an algorithm that agnostically learns linear threshold functions and returns a classfier with boundary volume $O(r+\varepsilon)$ atradius of perturbation $r$.The time and sample complexity of $d^{\tilde{O}(1/\varepsilon^2)}$ matches the complexity of polynomial regression.Our algorithm augments the classic approach of polynomial regression with three additional steps:\$\quad$ a) performing the $\ell_1$-error regression under $\ell_1$ noise sensitivity constraints,\$\quad$ b) a structured partitioning and rounding step that returns a Boolean classifier with error $\mathrm{opt} + O(\varepsilon)$ and noise sensitivity $O(r+\varepsilon)$ simultaneously, and \$\quad c)$ a local corrector that ``smooths'' a function with low noise sensitivity into a function that is adversarially robust.



Paperid:972
Authors:Chang Deng, Bryon Aragam
Abstract:
Existing methods for differentiable structure learning in discrete data typically assume that the data are generated from specific structural equation models. However, these assumptions may not align with the true data-generating process, which limits the general applicability of such methods. Furthermore, current approaches often ignore the complex dependence structure inherent in discrete data and consider only linear effects. We propose a differentiable structure learning framework that is capable of capturing arbitrary dependencies among discrete variables. We show that although general discrete models are unidentifiable from purely observational data, it is possible to characterize the complete set of compatible parameters and structures. Additionally, we establish identifiability up to the Markov equivalence class (MEC) under mild assumptions. We formulate the learning problem as a single differentiable optimization task in the most general form, thereby avoiding the unrealistic simplifications adopted by previous methods. Empirical results demonstrate that our approach effectively captures complex relationships in discrete data.



Paperid:973
Authors:Rafael Oliveira, Xuesong Wang, Kian Ming Chai, Edwin Bonilla
Abstract:
We propose an extension of Thompson sampling to optimization problems over function spaces where the objective is a known functional of an unknown operator's output. We assume that functional evaluations are inexpensive, while queries to the operator (such as running a high-fidelity simulator) are costly. Our algorithm employs a sample-then-optimize approach using neural operator surrogates. This strategy avoids explicit uncertainty quantification by treating trained neural operators as approximate samples from a Gaussian process. We provide novel theoretical convergence guarantees based on Gaussian processes in the infinite-dimensional setting, under minimal assumptions. We benchmark our method against existing baselines on functional optimization tasks involving partial differential equations and other nonlinear operator-driven phenomena, demonstrating improved sample efficiency and competitive performance.



Paperid:974
Authors:Pratyush Maini, Sachin Goyal, Dylan Sam, Alexander Robey, Yash Savani, Yiding Jiang, Andy Zou, Matt Fredrikson, Zachary Lipton, J. Zico Kolter
Abstract:
As large language models (LLMs) are increasingly deployed in high-stakes settings, the risk of generating harmful or toxic content remains a central challenge. Post-hoc alignment methods are brittle: once unsafe patterns are learned during pretraining, they are hard to remove.In this work, we present a data-centric pretraining framework that builds safety into the model from the start. Our framework consists of four key steps:(i) Safety Filtering: building a safety classifier to classify webdata into safe and unsafe categories;(ii) Safety Rephrasing: we recontextualize unsafe webdata into safer narratives;(iii) Native Refusal: we synthetically generate pretraining datasets that actively teach models to refuse on unsafe content and the moral reasoning behind it, and(iv) Harmfulness-Tag annotated pretraining: we flag unsafe content during pretraining using a special token, and use it to steer models away from unsafe generations at inference-time.Our safety-pretrained models reduce attack success rates from 38.8% to 8.4% on standard LLM safety benchmarks with no performance degradation on general tasks.



Paperid:975
Authors:Sungje Park, Stephen Tu
Abstract:
Backward stochastic differential equation (BSDE)-based deep learning methods provide an alternative to Physics-Informed Neural Networks (PINNs) for solving high-dimensional partial differential equations (PDEs), offering potential algorithmic advantages in settings such as stochastic optimal control, where the PDEs of interest are tied to an underlying dynamical system. In this paper, we identify the root cause of this performance gap as a discretization bias introduced by the standard Euler-Maruyama (EM) integration scheme applied to one-step self-consistency BSDE losses, which shifts the optimization landscape off target. We find that this bias cannot be satisfactorily addressed through finer step-sizes or multi-step self-consistency losses. To properly handle this issue, we propose a Stratonovich-based BSDE formulation, which we implement with stochastic Heun integration. We show that our proposed approach completely eliminates the bias issues faced by EM integration. Furthermore, our empirical results show that our Heun-based BSDE method consistently outperforms EM-based variants and achieves competitive results with PINNs across multiple high-dimensional benchmarks. Our findings highlight the critical role of integration schemes in BSDE-based PDE solvers, an algorithmic detail that has received little attention thus far in the literature.



Paperid:976
Authors:Mojtaba Nafez, Mobina Poulaei, Nikan Vasei, Bardia moakhar, Mohammad Sabokrou, Mohammad Hossein Rohban
Abstract:
Weakly Supervised Video Anomaly Detection (WSVAD) has achieved notable advancements, yet existing models remain vulnerable to adversarial attacks, limiting their reliability. Due to the inherent constraints of weak supervision—where only video-level labels are provided despite the need for frame-level predictions—traditional adversarial defense mechanisms, such as adversarial training, are not effective since video-level adversarial perturbations are typically weak and inadequate. To address this limitation, pseudo-labels generated directly from the model can enable frame-level adversarial training; however, these pseudo-labels are inherently noisy, significantly degrading performance. We therefore introduce a novel Pseudo-Anomaly Generation method called Spatiotemporal Region Distortion (SRD), which creates synthetic anomalies by applying severe augmentations to localized regions in normal videos while preserving temporal consistency. Integrating these precisely annotated synthetic anomalies with the noisy pseudo-labels substantially reduces label noise, enabling effective adversarial training. Extensive experiments demonstrate that our method significantly enhances the robustness of WSVAD models against adversarial attacks, outperforming state-of-the-art methods by an average of 71.0\% in overall AUROC performance across multiple benchmarks.



Paperid:977
Authors:Lang Feng, Zhenghai Xue, Tingcong Liu, Bo An
Abstract:
Recent advances in group-based reinforcement learning (RL) have driven frontier large language models (LLMs) in single-turn tasks like mathematical reasoning. However, their scalability to long-horizon LLM agent training remains limited. Unlike static tasks, agent-environment interactions unfold over many steps and often yield sparse or delayed rewards, making credit assignment across individual steps significantly more challenging. In this work, we propose Group-in-Group Policy Optimization (GiGPO), a novel RL algorithm that achieves fine-grained credit assignment for LLM agents while preserving the appealing properties of group-based RL: critic-free, low memory, and stable convergence. GiGPO introduces a two-level structure for estimating relative advantage: (i) At the episode-level, GiGPO computes macro relative advantages based on groups of complete trajectories; (ii) At the step-level, GiGPO introduces an anchor state grouping mechanism that retroactively constructs step-level groups by identifying repeated environment states across trajectories. Actions stemming from the same state are grouped together, enabling micro relative advantage estimation.This hierarchical structure effectively captures both global trajectory quality and local step effectiveness without relying on auxiliary models or additional rollouts. We evaluate GiGPO on two challenging agent benchmarks, ALFWorld and WebShop, using Qwen2.5-1.5B-Instruct and Qwen2.5-7B-Instruct. Crucially, GiGPO delivers fine-grained per-step credit signals and achieves performance gains of > 12\% on ALFWorld and > 9\% on WebShop over the GRPO baseline: all while maintaining the same GPU memory overhead, identical LLM rollout, and incurring little to no additional time cost.



Paperid:978
Authors:Jinwu Hu, ZiHao Lian, Zhiquan Wen, Chenghao Li, Guohao Chen, Xutao Wen, Bin Xiao, Mingkui Tan
Abstract:
Reinforcement Learning enables agents to learn optimal behaviors through interactions with environments. However, real-world environments are typically non-stationary, requiring agents to continuously adapt to new tasks and changing conditions. Although Continual Reinforcement Learning facilitates learning across multiple tasks, existing methods often suffer from catastrophic forgetting and inefficient knowledge utilization. To address these challenges, we propose Continual Knowledge Adaptation for Reinforcement Learning (CKA-RL), which enables the accumulation and effective utilization of historical knowledge. Specifically, we introduce a Continual Knowledge Adaptation strategy, which involves maintaining a task-specific knowledge vector pool and dynamically using historical knowledge to adapt the agent to new tasks. This process mitigates catastrophic forgetting and enables efficient knowledge transfer across tasks by preserving and adapting critical model parameters. Additionally, we propose an Adaptive Knowledge Merging mechanism that combines similar knowledge vectors to address scalability challenges, reducing memory requirements while ensuring the retention of essential knowledge. Experiments on three benchmarks demonstrate that the proposed CKA-RL outperforms state-of-the-art methods, achieving an improvement of 4.20% in overall performance and 8.02% in forward transfer.



Paperid:979
Authors:Shiqi Liu, Wenhan Cao, Chang Liu, Zeyu He, Tianyi Zhang, Yinuo Wang, Shengbo Li
Abstract:
Abstract:Estimating hidden states in dynamical systems, also known as optimal filtering, is a long-standing problem in various fields of science and engineering. In this paper, we introduce a general filtering framework, $\textbf{LLM-Filter}$, which leverages large language models (LLMs) for state estimation by embedding noisy observations with text prototypes. In a number of experiments for classical dynamical systems, we find that first, state estimation can significantly benefit from the knowledge embedded in pre-trained LLMs. By achieving proper modality alignment with the frozen LLM, LLM-Filter outperforms the state-of-the-art learning-based approaches. Second, we carefully design the prompt structure, System-as-Prompt (SaP), incorporating task instructions that enable LLMs to understand tasks and adapt to specific systems. Guided by these prompts, LLM-Filter exhibits exceptional generalization, capable of performing filtering tasks accurately in changed or even unseen environments. We further observe a scaling-law behavior in LLM-Filter, where accuracy improves with larger model sizes and longer training times. These findings make LLM-Filter a promising foundation model of filtering.



Paperid:980
Authors:Kedi Ying, Ruiping Liu, Chongyan Chen, Mingzhe Tao, Hao Shi, Kailun Yang, Jiaming Zhang, Rainer Stiefelhagen
Abstract:
Abstract:Walking assistance in extreme or complex environments remains a significant challenge for people with blindness or low vision (BLV), largely due to the lack of a holistic scene understanding. Motivated by the real-world needs of the BLV community, we build mmWalk, a simulated multi-modal dataset that integrates multi-view sensor and accessibility-oriented features for outdoor safe navigation. Our dataset comprises $120$ manually controlled, scenario-categorized walking trajectories with $62k$ synchronized frames. It contains over $559k$ panoramic images across RGB, depth, and semantic modalities. Furthermore, to emphasize real-world relevance, each trajectory involves outdoor corner cases and accessibility-specific landmarks for BLV users. Additionally, we generate mmWalkVQA, a VQA benchmark with over $69k$ visual question-answer triplets across $9$ categories tailored for safe and informed walking assistance. We evaluate state-of-the-art Vision-Language Models (VLMs) using zero- and few-shot settings and found they struggle with our risk assessment and navigational tasks. We validate our mmWalk-finetuned model on real-world datasets and show the effectiveness of our dataset for advancing multi-modal walking assistance.



Paperid:981
Authors:Weijie Shi, Yue Cui, Yaguang Wu, Jingzhi Fang, Shibo Zhang, Mengze Li, Sirui Han, Jia Zhu, Jiajie Xu, Xiaofang Zhou
Abstract:
Diverse decoding of large language models is crucial for applications requiring multiple semantically distinct responses, yet existing methods primarily achieve lexical rather than semantic diversity. This limitation significantly constrains Best-of-N strategies, group-based reinforcement learning, and data synthesis. While temperature sampling and diverse beam search modify token distributions or apply n-gram penalties, they fail to ensure meaningful semantic differentiation. We introduce Semantic-guided Diverse Decoding (SemDiD), operating directly in embedding space that balances quality with diversity through three complementary mechanisms: orthogonal directional guidance, dynamic inter-group repulsion, and position-debiased probability assessment. SemDiD harmonizes these competing objectives using adaptive gain functions and constraint optimization, ensuring both quality thresholds and maximal semantic differentiation. Experiments show SemDiD consistently outperforms existing methods, improving Best-of-N coverage by 1.4-5.2% across diverse tasks and accelerating RLHF training convergence by 15% while increasing accuracy by up to 2.1%.



Paperid:982
Authors:Reza Shirkavand, Peiran Yu, Qi He, Heng Huang
Abstract:
Abstract:Fine-tuning pre-trained Large Language Models (LLMs) for downstream tasks using First-Order (FO) optimizers presents significant computational challenges. Parameter-Efficient Fine-Tuning~(PEFT) methods have been proposed to address these challenges by freezing most model parameters and training only a small subset. While PEFT is efficient, it may not outperform full fine-tuning when high task-specific performance is required.Zeroth-Order (ZO) methods offer an alternative for fine-tuning the entire pre-trained model by approximating gradients using only the forward pass, thus eliminating the computational burden of back-propagation,% in first-order methods, but they converge painfully slowly and are very sensitive to the choice of task prompts.We bridge these worlds with Bilevel‑ZOFO, a penalty‑based bilevel formulation that treats adapter parameters as a lower‑level learner coupled to an upper‑level ZO optimizer of the full backbone. This double-loop optimization strategy only requires the gradient of the PEFT model and the forward pass of the base model. We provide theoretical convergence guarantees for Bilevel ZOFO. Empirically, we demonstrate that Bilevel-ZOFO significantly outperforms existing ZO methods, achieves 2–4$\times$ faster training, and reduces sensitivity to prompts. Bilevel-ZOFO also outperforms FO PEFT methods while maintaining similar memory efficiency. Additionally, we show its strong potential for meta learning.



Paperid:983
Authors:Wahid Bhimji, Ragansu Chakkappai, Po-Wen Chang, Yuan-Tang Chou, Sascha Diefenbacher, Jordan Dudley, Ibrahim Elsharkawy, Steven Farrell, Aishik Ghosh, Cristina Giordano, Isabelle Guyon, Chris Harris, Yota Hashizume, Shih-Chieh Hsu, Elham E Khoda, Claudius Krause, Ang Li, Benjamin Nachman, David Rousseau, Robert Schöfbeck, Maryam Shooshtari, Dennis Schwarz, Ihsan Ullah, Daohan Wang, Yulei Zhang
Abstract:
The FAIR Universe – HiggsML Uncertainty Challenge focuses on measuring the physics properties of elementary particles with imperfect simulators due to differences in modelling systematic errors. Participants were required to compute and report confidence intervals for a parameter of interest regarding the Higgs boson while accounting for various uncertainties. The dataset is a tabular dataset of 28 features and 220 million instances. Each instance represents a simulated proton-proton collision as observed at CERN's Large Hadron Collider in Geneva. The features of these simulations capture the characteristics of various signature particles. These features include primary attributes, such as the energy and three-dimensional direction of the particles, as well as derived attributes, which are calculated from the primary ones using domain-specific knowledge. Additionally, a label feature designates the type of proton-proton collision for each example, distinguishing the Higgs boson of interest from three other background sources. The permanent release of the dataset, as outlined in this paper, allows long-term benchmarking of new techniques. The leading submissions are described, including Contrastive Normalising Flows and Density Ratios estimation through classification. Our challenge has brought together the physics and machine learning communities to advance our understanding and methodologies in handling systematic (epistemic) uncertainties within AI techniques.



Paperid:984
Authors:Zixiang Yin, Jiarui Li, Zhengming Ding
Abstract:
Decoding visual signals holds the tantalizing potential to unravel the complexities of cognition and perception. While recent studies have focused on reconstructing visual stimuli from neural recordings to bridge brain activity with visual imagery, existing methods offer limited insights into the underlying mechanisms of visual processing in the brain. To mitigate this gap, we present an \textit{i}nsightful \textbf{M}ulti-subject \textbf{I}nvariant \textbf{N}eural \textbf{D}ecoding (MIND) model, which employs a novel dual-decoding framework—both biometric and semantic decoding—to offer neural interpretability in a data-driven manner and deepen our understanding of brain-based visual functionalities. Our MIND model operates through three core steps: establishing a shared neural representation space across subjects using a ViT-based masked autoencoder, disentangling neural features into complementary subject-specific and object-specific components, and performing dual decoding to support both biometric and semantic classification tasks. Experimental results demonstrate that MIND achieves state-of-the-art decoding performance with minimal scalability limitations. Furthermore, MIND empirically generates voxel-object activation fingerprints that reveal object-specific neural patterns and enable investigation of subject-specific variations in attention to identical stimuli. These findings provide a foundation for more interpretable and generalizable subject-invariant neural decoding, advancing our understanding of the voxel semantic selectivity as well as the neural vision processing dynamics.



Paperid:985
Authors:Kshipra Bhawalkar, Yang Cai, Zhe Feng, Christopher Liaw, Tao Lin
Abstract:
Abstract:We consider the problem of maximizing a submodular function with access to a _noisy_ value oracle for the function instead of an exact value oracle. Similar to prior work, we assume that the noisy oracle is persistent in that multiple calls to the oracle for a specific set always returns the same value. In this model, Hassidim and Singer (2017) design a $(1-1/e)$-approximation algorithm for monotone submodular maximization subject to a cardinality constraint and Huang et al. (2022) design a $(1-1/e)/2$-approximation algorithm for monotone submodular maximization subject to any arbitrary matroid constraint. In this paper, we design a meta-algorithm that allows us to take any "robust" algorithm for exact submodular maximization as a black box and transform it into an algorithm for the noisy setting while retaining the approximation guarantee. By using the meta-algorithm with the measured continuous greedy algorithm, we obtain a $(1-1/e)$-approximation (resp. $1/e$-approximation) for monotone (resp. non-monotone) submodular maximization subject to a matroid constraint under noise. Furthermore, by using the meta-algorithm with the double greedy algorithm, we obtain a $1/2$-approximation for unconstrained (non-monotone) submodular maximization under noise.



Paperid:986
Authors:Yingjie Gao, Yanan Zhang, Zhi Cai, Di Huang
Abstract:
In recent years, test-time adaptive object detection has attracted increasing attention due to its unique advantages in online domain adaptation, which aligns more closely with real-world application scenarios. However, existing approaches heavily rely on source-derived statistical characteristics while making the strong assumption that the source and target domains share an identical category space. In this paper, we propose the first foundation model-powered test-time adaptive object detection method that eliminates the need for source data entirely and overcomes traditional closed-set limitations. Specifically, we design a Multi-modal Prompt-based Mean-Teacher framework for vision-language detector-driven test-time adaptation, which incorporates text and visual prompt tuning to adapt both language and visual representation spaces on the test data in a parameter-efficient manner. Correspondingly, we propose a Test-time Warm-start strategy tailored for the visual prompts to effectively preserve the representation capability of the vision branch. Furthermore, to guarantee high-quality pseudo-labels in every test batch, we maintain an Instance Dynamic Memory (IDM) module that stores high-quality pseudo-labels from previous test samples, and propose two novel strategies-Memory Reinforcement and Memory Hallucination-to leverage IDM's high-quality instances for enhancing original predictions and hallucinating images without available pseudo-labels, respectively. Extensive experiments on cross-corruption and cross-dataset benchmarks demonstrate that our method consistently outperforms previous state-of-the-art methods, and can adapt to arbitrary cross-domain and cross-category target data. The code and models will be made publicly available.



Paperid:987
Authors:Emile Anand, Jan van den Brand, Rose McCarty
Abstract:
Abstract:We consider the problem of preprocessing an $n\times n$ matrix $\mathbf{M}$, and supporting queries that, for any vector $v$, returns the matrix-vector product $\mathbf{M} v$. This problem has been extensively studied in both theory and practice: on one side, practitioners have developed algorithms that are highly efficient in practice, whereas on the other side, theoreticians have proven that the problem cannot be solved faster than naive multiplication in the worst-case. This lower bound holds even in the average-case, implying that existing average-case analyses cannot explain this gap between theory and practice. Hence, we study the problem for _structured_ matrices. We show that for $n\times n$ matrices of VC-dimension $d$, the matrix-vector multiplication problem can be solved with $\tilde{O}(n^2)$ preprocessing and $\tilde O(n^{2-1/d})$ query time. Given the low constant VC-dimensions observed in most real-world data, our results posit an explanation for why the problem can be solved so much faster in practice.Our results yield the first non-trivial upper bounds for many applications. In previous works, the online matrix-vector (OMv) hypothesis (conjecturing that quadratic time is needed per query, even over the boolean semi-ring) was used to prove many conditional lower bounds, showing that it is impossible to compute and maintain high-accuracy estimates for effective resistance, Laplacian solvers, shortest paths, and triangle detection in graphs subject to node insertions and deletions in subquadratic time. Yet, via a reduction to our matrix-vector-multiplication result, we show we can maintain these problems efficiently if the input is structured, providing the first subquadratic upper bounds in the high-accuracy regime.



Paperid:988
Authors:Christian Internò, Robert Geirhos, Markus Olhofer, Sunny Liu, Barbara Hammer, David Klindt
Abstract:
Abstract:The rapid advancement of generative AI enables highly realistic synthetic video, posing significant challenges for content authentication and raising urgent concerns about misuse. Existing detection methods often struggle with generalization and capturing subtle temporal inconsistencies. We propose $ReStraV$ ($Re$presentation $Stra$ightening for $V$ideo), a novel approach to distinguish natural from AI-generated videos. Inspired by the ``perceptual straightening'' hypothesis—which suggests real-world video trajectories become more straight in neural representation domain—we analyze deviations from this expected geometric property. Using a pre-trained self-supervised vision transformer (DINOv2), we quantify the temporal curvature and stepwise distance in the model's representation domain. We aggregate statistical and signals descriptors of these measures for each video and train a classifier. Our analysis shows that AI-generated videos exhibit significantly different curvature and distance patterns compared to real videos. A lightweight classifier achieves state-of-the-art detection performance (e.g., $97.17$ % accuracy and $98.63$ % AUROC on the VidProM benchmark, substantially outperforming existing image- and video-based methods. ReStraV is computationally efficient, it is offering a low-cost and effective detection solution. This work provides new insights into using neural representation geometry for AI-generated video detection.



Paperid:989
Authors:Zhiwei Zhai, Wenjing Yan, Ying-Jun Zhang
Abstract:
Abstract:Decentralized bilevel optimization has garnered significant attention due to its critical role in solving large-scale machine learning problems. However, existing methods often rely on prior knowledge of problem parameters—such as smoothness, convexity, or communication network topologies—to determine appropriate stepsizes. In practice, these problem parameters are typically unavailable, leading to substantial manual effort for hyperparameter tuning. In this paper, we propose \textbf{AdaSDBO}, a fully problem-parameter-free algorithm for decentralized bilevel optimization with a single-loop structure. AdaSDBO leverages adaptive stepsizes based on cumulative gradient norms to update all variables simultaneously, dynamically adjusting its progress and eliminating the need for problem-specific hyperparameter tuning. Through rigorous theoretical analysis, we establish that AdaSDBO achieves a convergence rate of $\widetilde{\mathcal{O}}\left(\frac{1}{T}\right)$, matching the performance of well-tuned state-of-the-art methods up to polylogarithmic factors. Extensive numerical experiments demonstrate that AdaSDBO delivers competitive performance compared to existing decentralized bilevel optimization methods while exhibiting remarkable robustness across diverse stepsize configurations.



Paperid:990
Authors:Peter Kocsis, Lukas Höllein, Matthias Niessner
Abstract:
We introduce IntrinsiX, a novel method that generates high-quality intrinsic images from text description.In contrast to existing text-to-image models whose outputs contain baked-in scene lighting, our approach predicts physically-based rendering (PBR) maps.This enables the generated outputs to be used for content creation scenarios in core graphics applications that facilitate re-lighting, editing, and texture generation tasks. In order to train our generator, we exploit strong image priors, and pre-train separate models for each PBR material component (albedo, roughness, metallic, normals).We then align these models with a new cross-intrinsic attention formulation that concatenates key and value features in a consistent fashion. This allows us to exchange information between each output modality and to obtain semantically coherent PBR predictions.To ground each intrinsic component, we propose a rendering loss which provides image-space signals to constrain the model, thus facilitating sharp details also in the output BRDF properties. Our results demonstrate detailed intrinsic generation with strong generalization capabilities that outperforms existing intrinsic image decomposition methods used with generated images by a significant margin.Finally, we show a series of applications, including re-lighting, editing, and for the first time text-conditioned room-scale PBR texture generation.We will release the code and the pre-trained model weights.



Paperid:991
Authors:Yuncong Yang, Jiageng Liu, Zheyuan Zhang, Siyuan Zhou, Reuben Tan, Jianwei Yang, Yilun Du, Chuang Gan
Abstract:
Spatial reasoning in 3D space is central to human cognition and indispensable for embodied tasks such as navigation and manipulation. However, state-of-the-art vision–language models (VLMs) struggle frequently with tasks as simple as anticipating how a scene will look after an egocentric motion: they perceive 2D images but lack an internal model of 3D dynamics. We therefore propose SpatialNavigator, a test-time scaling framework that grants a VLM with this missing capability by coupling it to a controllable world model based on video diffusion. The VLM iteratively sketches a concise camera trajectory, while the world model synthesizes the corresponding view at each step. The VLM then reasons over this multi-view evidence gathered during the interactive exploration. Without any fine-tuning, our SpatialNavigator achieves over an average 8\% performance boost on the representative spatial reasoning benchmark SAT, showing that pairing VLMs with world models for test-time scaling offers a simple, plug-and-play route to robust 3D reasoning. Meanwhile, our method also improves upon the test-time inference VLMs trained through reinforcement learning, which demonstrates the potential of our method that utilizes world models for test-time scaling.



Paperid:992
Authors:Alexandros Graikos, Nebojsa Jojic, Dimitris Samaras
Abstract:
Large denoising diffusion models, such as Stable Diffusion, have been trained on billions of image-caption pairs to perform text-conditioned image generation. As a byproduct of this training, these models have acquired general knowledge about image statistics, which can be useful for other inference tasks. However, when confronted with sampling an image under new constraints, e.g. generating the missing parts of an image, using large pre-trained text-to-image diffusion models is inefficient and often unreliable. Previous approaches either utilized backpropagation through the denoiser network, making them significantly slower and more memory-demanding than simple text-to-image generation, or only enforced the constraint locally, failing to capture critical long-range correlations in the sampled image. In this work, we propose an algorithm that enables fast, high-quality generation under arbitrary constraints. We show that in denoising diffusion models, we can employ an approximation to Newton's optimization method that allows us to speed up inference and avoid the expensive backpropagation operations. Our approach produces results that rival or surpass the state-of-the-art training-free inference methods while requiring a fraction of the time. We demonstrate the effectiveness of our algorithm under both linear (inpainting, super-resolution) and non-linear (style-guided generation) constraints. An implementation is provided in the supplementary code.



Paperid:993
Authors:TAO SUN, Liyuan Zhu, Shengyu Huang, Shuran Song, Iro Armeni
Abstract:
We present Rectified Point Flow, a unified parameterization that formulates pairwise point cloud registration and multi-part shape assembly as a single conditional generative problem. Given unposed point clouds, our method learns a continuous point-wise velocity field that transports noisy points toward their target positions, from which part poses are recovered. In contrast to prior work that regresses part-wise poses with ad-hoc symmetry handling, our method intrinsically learns assembly symmetries without symmetry labels. Together with a self-supervised encoder focused on overlapping points, Rectified Point Flow achieves a new state-of-the-art performance on six benchmarks spanning pairwise registration and shape assembly. Notably, our unified formulation enables effective joint training on diverse datasets, facilitating the learning of shared geometric priors and consequently boosting accuracy. Our code and models will be made publicly available.



Paperid:994
Authors:Maxime Meyer, Soumik Adhikary, Naixu Guo, Patrick Rebentrost
Abstract:
Abstract:Quantum state tomography, the task of learning an unknown quantum state, is a fundamental problem in quantum information. In standard settings, the complexity of this problem depends significantly on the type of quantum state that one is trying to learn, with pure states being substantially easier to learn than general mixed states. A natural question is whether this separation holds for any quantum state learning setting. In this work, we consider the online learning framework and prove the surprising result that learning pure states in this setting is as hard as learning mixed states. More specifically, we show that both classes share almost the same sequential fat-shattering dimension, leading to identical regret scaling. We also generalize previous results on full quantum state tomography in the online setting to (i) the $\epsilon$-realizable setting and (ii) learning the density matrix only partially, using smoothed analysis.



Paperid:995
Authors:Liangyu Wang, Junxiao Wang, Jie Ren, Zihang Xiang, David Keyes, Di Wang
Abstract:
As large language models (LLMs) increasingly underpin technological advancements, the privacy of their training data emerges as a critical concern. Differential Privacy (DP) serves as a rigorous mechanism to protect this data, yet its integration via Differentially Private Stochastic Gradient Descent (DP-SGD) introduces substantial challenges, primarily due to the complexities of per-sample gradient clipping. Current explicit methods, such as Opacus, necessitate extensive storage for per-sample gradients, significantly inflating memory requirements. Conversely, implicit methods like GhostClip reduce storage needs by recalculating gradients multiple times, which leads to inefficiencies due to redundant computations. This paper introduces FlashDP, an innovative cache-friendly per-layer DP-SGD that consolidates necessary operations into a single task, calculating gradients only once in a fused manner. This approach not only diminishes memory movement by up to 50\% but also cuts down redundant computations by 20\%, compared to previous methods. Consequently, FlashDP does not increase memory demands and achieves a 90\% throughput compared to the Non-DP method on a four-A100 system during the pre-training of the Llama-13B model, while maintaining parity with standard per-layer clipped DP-SGD in terms of accuracy. These advancements establish FlashDP as a pivotal development for efficient and privacy-preserving training of LLMs.



Paperid:996
Authors:Rajdeep Haldar, Ziyi Wang, Guang Lin, Yue XING, Qifan Song
Abstract:
We present a theoretical framework showing that popular LLM alignment methods—including RLHF and its variants—can be understood as divergence estimators between aligned (safe or preferred) and unaligned (harmful or less-preferred) distributions. This perspective explains the emergence of separation in the latent space between safe and harmful prompts after alignment. As an application of our general divergence framework, we propose KLDO, a novel KL divergence-based alignment method, and empirically validate its effectiveness. We further show that using compliance–refusal datasets, rather than standard preference-based datasets, leads to stronger separation and improved safety alignment. Finally, to quantify the separation effect, we propose a distance-based metric in the prompt representation space, which also acts as a statistically significant indicator for model safety.



Paperid:997
Authors:Jaron Maene, Efthymia Tsamoura
Abstract:
Logic programs have been combined with embeddings to enable differentiable reasoning over the representation space, giving rise to neurosymbolic models that can jointly learn symbolic rules and train neural networks. The key idea behind this composition is to replace unification in logic programming with soft unification. Unlike unification, its soft counterpart exploits the similarity between symbols in the representation space to decide when two symbols are equivalent. However, the semantics of similarity violate the transitive law of equivalence. We prove this has undesirable consequences during learning. To alleviate the above shortcomings, we introduce a probability distribution on equivalences in logic programs. Embeddings in this context can be understood as a factorization of the probabilistic equivalence. Finally, we introduce an exact probabilistic inference method that can exploit the symmetries between equivalent symbols.



Paperid:998
Authors:John Duchi
Abstract:
Abstract:We revisit the problem of constructing predictive confidence sets for which we wish to obtain some type of conditional validity. We provide new arguments showing how “split conformal” methods achieve near desired coverage levels with high probability, a guarantee conditional on the validation data rather than marginal over it. In addition, we directly consider (approximate) conditional coverage, where, e.g., conditional on a covariate $X$ belonging to some group of interest, we would like a guarantee that a predictive set covers the true outcome $Y$. We show that the natural method of performing quantile regression on a held-out (validation) dataset yields minimax optimal guarantees of coverage here. Complementing these positive results, we also provide experimental evidence that interesting work remains to be done to develop computationally efficient but valid predictive inference methods.



Paperid:999
Authors:Jes Frellsen, Maher Kassem, Tone Bengtsen, Lars Olsen, Kresten Lindorff-Larsen, Jesper Ferkinghoff-Borg, Wouter Boomsma
Abstract:
Inverse folding models have proven to be highly effective zero-shot predictors of protein stability. Despite this success, the link between the amino acid preferences of an inverse folding model and the free-energy considerations underlying thermostability remains incompletely understood. A better understanding would be of interest not only from a theoretical perspective, but also potentially provide the basis for stronger zero-shot stability prediction. In this paper, we take steps to clarify the free-energy foundations of inverse folding models. Our derivation reveals the standard practice of likelihood ratios as a simplistic approximation and suggests several paths towards better estimates of the relative stability. We empirically assess these approaches and demonstrate that considerable gains in zero-shot performance can be achieved with fairly simple means.



Paperid:1000
Authors:Dominic Maggio, Hyungtae Lim, Luca Carlone
Abstract:
We present VGGT-SLAM, a dense RGB SLAM system constructed by incrementally and globally aligning submaps created from the feed-forward scene reconstruction approach VGGT using only uncalibrated monocular cameras. While related works align submaps using similarity transforms (i.e., translation, rotation, and scale), we show that such approaches are inadequate in the case of uncalibrated cameras. In particular, we revisit the idea of reconstruction ambiguity, where given a set of uncalibrated cameras with no assumption on the camera motion or scene structure, the scene can only be reconstructed up to a 15-degrees-of-freedom projective transformation of the true geometry. This inspires us to recover a consistent scene reconstruction across submaps by optimizing over the SL(4) manifold, thus estimating 15-degrees-of-freedom homography transforms between sequential submaps while accounting for potential loop closure constraints. As verified by extensive experiments, we demonstrate that VGGT-SLAM achieves improved map quality using long video sequences that are infeasible for VGGT due to its high GPU requirements.



Paperid:1001
Authors:Tariq Berrada Ifriqi, Adriana Romero-Soriano, Michal Drozdzal, Jakob Verbeek, Karteek Alahari
Abstract:
Guidance techniques are commonly used in diffusion and flow models to improve image quality and input consistency for conditional generative tasks such as class-conditional and text-to-image generation.In particular, classifier-free guidance (CFG) is the most widely adopted guidance technique.It results, however, in trade-offs across quality, diversity and consistency: improving some at the expense of others.While recent work has shown that it is possible to disentangle these factors to some extent, such methods come with an overhead of requiring an additional (weaker) model, or require more forward passes per sampling step. In this paper, we propose Entropy Rectifying Guidance (ERG), a simple and effective guidance method based on inference-time changes in the attention mechanism of state-of-the-art diffusion transformer architectures, which allows for simultaneous improvements over image quality, diversity and prompt consistency. ERG is more general than CFG and similar guidance techniques, as it extends to unconditional sampling. We show that ERG results in significant improvements in various generation tasks such as text-to-image, class-conditional and unconditional image generation.We also show that ERG can be seamlessly combined with other recent guidance methods such as CADS and APG, further improving generations.



Paperid:1002
Authors:Jin Li, Shoujin Wang, Qi Zhang, Feng Liu, Tongliang Liu, Longbing Cao, Shui Yu, Fang Chen
Abstract:
Uncovering cause-and-effect mechanisms from data is fundamental to scientific progress. While large language models (LLMs) show promise for enhancing causal discovery (CD) from unstructured data, their application to the increasingly prevalent multimodal setting remains a critical challenge. Even with the advent of multimodal LLMs (MLLMs), their efficacy in multimodal CD is hindered by two primary limitations: (1) difficulty in exploring intra- and inter-modal interactions for comprehensive causal variable identification; and (2) insufficiency to handle structural ambiguities with purely observational data. To address these challenges, we propose MLLM-CD, a novel framework for multimodal causal discovery from unstructured data. It consists of three key components: (1) a novel contrastive factor discovery module to identify genuine multimodal factors based on the interactions explored from contrastive sample pairs; (2) a statistical causal structure discovery module to infer causal relationships among discovered factors; and (3) an iterative multimodal counterfactual reasoning module to refine the discovery outcomes iteratively by incorporating the world knowledge and reasoning capabilities of MLLMs. Extensive experiments on both synthetic and real-world datasets demonstrate the effectiveness of MLLM-CD in revealing genuine factors and causal relationships among them from multimodal unstructured data. The implementation code and data are available at https://anonymous.4open.science/r/MLLM-CD-3FB3



Paperid:1003
Authors:Kanghyun Choi, Hyeyoon Lee, Sunjong Park, Dain Kwon, Jinho Lee
Abstract:
Abstract:Low-bit floating-point (FP) formats, such as FP8, provide significant acceleration and memory savings in model training thanks to native hardware support on modern GPUs and NPUs. However, we analyze that FP8 quantization offers speedup primarily for large-dimensional matrix multiplications, while inherent quantization overheads diminish speedup when applied to low-rank adaptation (LoRA), which uses small-dimensional matrices for efficient fine-tuning of large language models (LLMs). To address this limitation, we propose FALQON, a novel framework that eliminates the quantization overhead from separate LoRA computational paths by directly merging LoRA adapters into an FP8-quantized backbone during fine-tuning. Furthermore, we reformulate the forward and backward computations for merged adapters to significantly reduce quantization overhead, and introduce a row-wise proxy update mechanism that efficiently integrates substantial updates into the quantized backbone. Experimental evaluations demonstrate that FALQON achieves approximately a 3$\times$ training speedup over existing quantized LoRA methods with a similar level of accuracy, providing a practical solution for efficient large-scale model fine-tuning. Moreover, FALQON’s end-to-end FP8 workflow removes the need for post-training quantization, facilitating efficient deployment.



Paperid:1004
Authors:Francis Ward, Teun van der Weij, Hanna Gábor, Sam Martin, Harel Lidar, Louis Makower, Raja Moreno, Thomas Jodrell, Lauren Robson
Abstract:
AI systems are increasingly able to autonomously conduct realistic software engineering tasks, and may soon be deployed to automate machine learning (ML) R\&D itself. Frontier AI systems may be deployed in safety-critical settings, including to help ensure the safety of future systems. Unfortunately, frontier and future systems may not be sufficiently trustworthy, and there is evidence that these systems may even be misaligned with their developers or users. Therefore, we investigate the capabilities of AI agents to act against the interests of their users for ML engineering tasks, by sabotaging ML models, sandbagging their performance, and subverting oversight mechanisms. First, we extend MLE-Bench, a benchmark for realistic ML tasks, with code-sabotage tasks such as implanting backdoors and purposefully causing generalisation failures. Frontier agent make meaningful progress on our tasks. In addition, we study agent capabilities to sandbag on MLE-Bench. Agents can calibrate their performance to specified target levels below their actual capability. To mitigate sabotage, we use LM monitors to detect suspicious agent behaviour, and we measure model capability to sabotage and sandbag without being detected by these monitors. Overall, monitors are capable at detecting code-sabotage attempts but our results suggest that detecting sandbagging is more difficult. Additionally, aggregating multiple monitor predictions works well, but monitoring may not be sufficiently reliable to mitigate sabotage in high-stakes domains. Our benchmarks are implemented in the UK AISI’s Inspect framework and we make our code available in the supplementary material.



Paperid:1005
Authors:Shiwei Feng, Xiangzhe Xu, Xuan Chen, Kaiyuan Zhang, Syed Ahmed, Zian Su, Mingwei Zheng, Xiangyu Zhang
Abstract:
LLM agents are increasingly deployed to automate real-world tasks by invoking APIs through natural language instructions. While powerful, they often suffer from misinterpretation of user intent, leading to the agent’s actions that diverge from the user’s intended goal, especially as external toolkits evolve. Traditional software testing assumes structured inputs and thus falls short in handling the ambiguity of natural language. We introduce TAI3, an API-centric stress testing framework that systematically uncovers intent integrity violations in LLM agents. Unlike prior work focused on fixed benchmarks or adversarial inputs, TAI3 generates realistic tasks based on toolkits’ documentation and applies targeted mutations to expose subtle agent errors while preserving user intent. To guide testing, we propose semantic partitioning, which organizes natural language tasks into meaningful categories based on toolkit API parameters and their equivalence classes. Within each partition, seed tasks are mutated and ranked by a lightweight predictor that estimates the likelihood of triggering agent errors. To enhance efficiency, TAI3 maintains a datatype-aware strategy memory that retrieves and adapts effective mutation patterns from past cases. Experiments on 80 toolkit APIs demonstrate that TAI3 effectively uncovers intent integrity violations, significantly outperforming baselines in both error-exposing rate and query efficiency. Moreover, TAI3 generalizes well to stronger target models using smaller LLMs for test generation, and adapts to evolving APIs across domains.



Paperid:1006
Authors:Zhengyao Huang, Daniel Huang, Tiannan Xiao, Dina Ma, Zhenyu Ming, Hao Shi, Yuanhui Wen
Abstract:
Symbolic regression aims to discover concise, interpretable mathematical expressions that satisfy desired objectives, such as fitting data, posing a highly combinatorial optimization problem. While genetic programming has been the dominant approach, recent efforts have explored reinforcement learning methods for improving search efficiency. Monte Carlo Tree Search (MCTS), with its ability to balance exploration and exploitation through guided search, has emerged as a promising technique for symbolic expression discovery. However, its traditional bandit strategies and sequential symbol construction often limit performance. In this work, we propose an improved MCTS framework for symbolic regression that addresses these limitations through two key innovations: (1) an extreme bandit allocation strategy tailored for identifying globally optimal expressions, with finite-time performance guarantees under polynomial reward decay assumptions; and (2) evolution-inspired state-jumping actions such as mutation and crossover, which enable non-local transitions to promising regions of the search space. These state-jumping actions also reshape the reward landscape during the search process, improving both robustness and efficiency. We conduct a thorough numerical study to the impact of these improvements and benchmark our approach against existing symbolic regression methods on a variety of datasets, including both ground-truth and black-box datasets. Our approach achieves competitive performance with state-of-the-art libraries in terms of recovery rate, attains favorable positions on the Pareto frontier of accuracy versus model complexity.



Paperid:1007
Authors:Marc Jourdan, Achraf Azize
Abstract:
Abstract:Best Arm Identification (BAI) algorithms are deployed in data-sensitive applications, such as adaptive clinical trials or user studies. Driven by the privacy concerns of these applications, we study the problem of fixed-confidence BAI under global Differential Privacy (DP) for Bernoulli distributions. While numerous asymptotically optimal BAI algorithms exist in the non-private setting, a significant gap remains between the best lower and upper bounds in the global DP setting. This work reduces this gap to a small multiplicative constant, for any privacy budget $\epsilon$. First, we provide a tighter lower bound on the expected sample complexity of any $\delta$-correct and $\epsilon$-global DP strategy. Our lower bound replaces the Kullback–Leibler (KL) divergence in the transportation cost used by the non-private characteristic time with a new information-theoretic quantity that optimally trades off between the KL divergence and the Total Variation distance scaled by $\epsilon$. Second, we introduce a stopping rule based on these transportation costs and a private estimator of the means computed using an arm-dependent geometric batching. En route to proving the correctness of our stopping rule, we derive concentration results of independent interest for the Laplace distribution and for the sum of Bernoulli and Laplace distributions. Third, we propose a Top Two sampling rule based on these transportation costs. For any budget $\epsilon$, we show an asymptotic upper bound on its expected sample complexity that matches our lower bound to a multiplicative constant smaller than $8$. Our algorithm outperforms existing $\delta$-correct and $\epsilon$-global DP BAI algorithms for different values of $\epsilon$.



Paperid:1008
Authors:Morris Alper, David Novotny, Filippos Kokkinos, Hadar Averbuch-Elor, Tom Monnier
Abstract:
Despite recent advances in sparse novel view synthesis (NVS) applied to object-centric scenes, scene-level NVS remains a challenge. A central issue is the lack of available clean multi-view training data, beyond manually curated datasets with limited diversity, camera variation, or licensing issues. On the other hand, an abundance of diverse and permissively-licensed data exists in the wild, consisting of scenes with varying appearances (illuminations, transient occlusions, etc.) from sources such as tourist photos. To this end, we present WildCAT, a framework for generating novel views of scenes learned from diverse 2D scene image data captured in-the-wild. We unlock training on these data sources by explicitly modeling global appearance conditions in images, extending the state-of-the-art multi-view diffusion paradigm to learn from scene views of varying appearances. Our trained model generalizes to new scenes at inference time, enabling the generation of multiple consistent novel views. WildCAT provides state-of-the-art results on single-view NVS in object- and scene-level settings, while training on strictly less data sources than prior methods. Additionally, it enables novel applications by providing global appearance control during generation.



Paperid:1009
Authors:Itay Lamprecht, Asaf Karnieli, Yair Hanani, Niv Giladi, Daniel Soudry
Abstract:
Training and inference of Large Language Models (LLMs) with tensor-parallelism requires substantial communication to synchronize activations. Our findings suggest that with a few minor adjustments to current practices, LLMs can be trained without fully synchronizing activations, reducing bandwidth demands. We name this ``Communication-Aware Architecture for Tensor-parallelism'' (CAAT-Net). We train 1B and 7B parameter CAAT-Net models, with a 50% reduction in tensor-parallel communication and no significant drop in pretraining accuracy. Furthermore, we demonstrate how CAAT-Net accelerates both training and inference workloads.



Paperid:1010
Authors:Daniel Barzilai, Guy Kornowski, Ohad Shamir
Abstract:
In recent years, there has been much interest in understanding the generalization behavior of interpolating predictors, which overfit on noisy training data. Whereas standard analyses are concerned with whether a method is consistent or not, recent observations have shown that even inconsistent predictors can generalize well. In this work, we revisit the classic interpolating Nadaraya-Watson (NW) estimator (also known as Shepard's method), and study its generalization capabilities through this modern viewpoint. In particular, by varying a single bandwidth-like hyperparameter, we prove the existence of multiple overfitting behaviors, ranging non-monotonically from catastrophic, through benign, to tempered.Our results highlight how even classical interpolating methods can exhibit intricate generalization behaviors. In addition, for the purpose of tuning the hyperparameter, the results suggest that over-estimating the intrinsic dimension of the data is less harmful than under-estimating it. Numerical experiments complement our theory, demonstrating the same phenomena.



Paperid:1011
Authors:Daniel Israel, Guy Van den Broeck, Aditya Grover
Abstract:
The generation speed of LLMs are bottlenecked by autoregressive decoding, where tokens are predicted sequentially one by one. Alternatively, diffusion large language models (dLLMs) theoretically allow for parallel token generation, but in practice struggle to achieve the speed of autoregressive models without significantly sacrificing quality. We therefore introduce adaptive parallel decoding (APD), a novel method that dynamically adjusts the number of tokens sampled in parallel. We achieve this by defining a multiplicative mixture between the dLLM marginal probabilities and the joint probability of sequences under a small auxiliary autoregressive model. This inverts the standard setup of speculative decoding, where the goal is to sample from a large autoregressive verifier by drafting from a smaller model. We further optimize APD by enabling KV caching and limiting the size of the masked input. Altogether, our method puts forward three tunable parameters to flexibly tradeoff throughput and quality. We show that APD provides markedly higher throughput with minimal quality degradations on downstream benchmarks.



Paperid:1012
Authors:Benjamin Arnav, Pablo Bernabeu-Perez, Nathan Helm-Burger, Timothy H. Kostolansky, Hannes Whittingham, Mary Phuong
Abstract:
As AI models are deployed with increasing autonomy, it is important to ensure they do not take harmful actions unnoticed. As a potential mitigation, we investigate Chain-of-Thought (CoT) monitoring, wherein a weaker trusted monitor model continuously oversees the intermediate reasoning steps of a more powerful but untrusted model. We compare CoT monitoring to action-only monitoring, where only final outputs are reviewed, in a red-teaming setup where the untrusted model is instructed to pursue harmful side tasks while completing a coding problem. We find that CoT monitoring improves detection by up to 27 percentage points in scenarios where action-only monitoring fails to reliably identify sabotage. However, CoT traces can also contain misleading rationalizations that deceive the monitor, reducing performance in more obvious sabotage cases. To address this, we introduce a hybrid protocol that independently scores both reasoning and final outputs and combines them using a weighted average. This hybrid monitor consistently outperforms both CoT and action-only monitors across all tested models and tasks, with detection rates over four times higher than action-only monitoring for subtle deception scenarios.



Paperid:1013
Authors:Wenhao Wu, Fuhong Liu, Haoru Li, Zican Hu, Daoyi Dong, Chunlin Chen, Zhi Wang
Abstract:
In-context reinforcement learning (ICRL) has emerged as a promising paradigm for adapting RL agents to downstream tasks through prompt conditioning. However, two notable challenges remain in fully harnessing in-context learning within RL domains: the intrinsic multi-modality of the state-action-reward data and the diverse, heterogeneous nature of decision tasks. To tackle these challenges, we proposeT2MIR(Token- andTask-wiseMoE forIn-contextRL), an innovative framework that introduces architectural advances of mixture-of-experts (MoE) into transformer-based decision models. T2MIR substitutes the feedforward layer with two parallel layers: a token-wise MoE that captures distinct semantics of input tokens across multiple modalities, and a task-wise MoE that routes diverse tasks to specialized experts for managing a broad task distribution with alleviated gradient conflicts. To enhance task-wise routing, we introduce a contrastive learning method that maximizes the mutual information between the task and its router representation, enabling more precise capture of task-relevant information. The outputs of two MoE components are concatenated and fed into the next layer. Comprehensive experiments show that T2MIR significantly facilitates in-context learning capacity and outperforms various types of baselines. We bring the potential and promise of MoE to ICRL, offering a simple and scalable architectural enhancement to advance ICRL one step closer toward achievements in language and vision communities.



Paperid:1014
Authors:Seouh-won Yi, Min-hwan Oh
Abstract:
Abstract:We propose *feature perturbation*, a simple yet powerful technique that injects randomness directly into feature inputs, instead of randomizing unknown parameters or adding noise to rewards. Remarkably, this algorithm achieves $\widetilde{\mathcal{O}}(d\sqrt{T})$ worst-case regret bound for generalized linear bandits, while avoiding the $\widetilde{\mathcal{O}}(d^{3/2}\sqrt{T})$ regret typical of existing randomized bandit algorithms. Because our algorithm eschews parameter sampling, it is both computationally efficient and naturally extends to non-parametric or neural network models. We verify these advantages through empirical evaluations, demonstrating that feature perturbation not only surpasses existing methods but also unifies strong practical performance with best-known theoretical guarantees.



Paperid:1015
Authors:Zhenghao Xu, Qin Lu, Qingru Zhang, Liang Qiu, Ilgee Hong, Changlong Yu, Wenlin Yao, Yao Liu, Haoming Jiang, Lihong Li, Hyokun Yun, Tuo Zhao
Abstract:
Reward model (RM) plays a pivotal role in reinforcement learning with human feedback (RLHF) for aligning large language models (LLMs). However, classical RMs trained on human preferences are vulnerable to reward hacking and generalize poorly to out-of-distribution (OOD) inputs. By contrast, strong LLM judges equipped with reasoning capabilities demonstrate superior generalization, even without additional training, but incur significantly higher inference costs, limiting their applicability in online RLHF. In this work, we propose an uncertainty-based routing framework that efficiently complements a fast RM with a strong but costly LLM judge. Our approach formulates advantage estimation in policy gradient (PG) methods as pairwise preference classification, enabling principled uncertainty quantification to guide routing. Uncertain pairs are forwarded to the LLM judge, while confident ones are evaluated by the RM. Experiments on RM benchmarks demonstrate that our uncertainty-based routing strategy significantly outperforms random judge calling at the same cost, and downstream alignment results showcase its effectiveness in improving online RLHF.



Paperid:1016
Authors:Zhining Liu, Zihao Li, Ze Yang, Tianxin Wei, Jian Kang, Yada Zhu, Hendrik Hamann, Jingrui He, Hanghang Tong
Abstract:
Class-imbalanced learning (CIL) on tabular data is important in many real-world applications where the minority class holds the critical but rare outcomes. In this paper, we present CLIMB, a comprehensive benchmark for class-imbalanced learning on tabular data. CLIMB includes 73 real-world datasets across diverse domains and imbalance levels, along with unified implementations of 29 representative CIL algorithms. Built on a high-quality open-source Python package with unified API designs, detailed documentation, and rigorous code quality controls, CLIMB supports easy implementation and comparison between different CIL algorithms. Through extensive experiments, we provide practical insights on method accuracy and efficiency, highlighting the limitations of naive rebalancing, the effectiveness of ensembles, and the importance of data quality. Our code, documentation, and examples are available at https://github.com/ZhiningLiu1998/imbalanced-ensemble.



Paperid:1017
Authors:Nan An, Weian Li, Qi Qi, Changyuan Yu, Liang Zhang
Abstract:
Accurate attribution for multiple platforms is critical for evaluating performance-based advertising. However, existing attribution methods rely heavily on the heuristic methods, e.g., Last-Click Mechanism (LCM) which always allocates the attribution to the platform with the latest report, lacking theoretical guarantees for attribution accuracy. In this work, we propose a novel theoretical model for the advertising attribution problem, in which we aim to design the optimal dominant strategy incentive compatible (DSIC) mechanisms and evaluate their performance. We first show that LCM is not DSIC and performs poorly in terms of accuracy and fairness. To address this limitation, we introduce the Peer-Validated Mechanism (PVM), a DSIC mechanism in which a platform's attribution depends solely on the reports of other platforms. We then examine the accuracy of PVM across both homogeneous and heterogeneous settings, and provide provable accuracy bounds for each case. Notably, we show that PVM is the optimal DSIC mechanism in the homogeneous setting. Finally, numerical experiments are conducted to show that PVM consistently outperforms LCM in terms of attribution accuracy and fairness.



Paperid:1018
Authors:Ziping Dong, Chao Shuai, Zhongjie Ba, Peng Cheng, Zhan Qin, Qinglong Wang, Kui Ren
Abstract:
Invisible Image Watermarking is crucial for ensuring content provenance and accountability in generative AI. While Gen-AI providers are increasingly integrating invisible watermarking systems, the robustness of these schemes against forgery attacks remains poorly characterized. This is critical, as forging traceable watermarks onto illicit content leads to false attribution, potentially harming the reputation and legal standing of Gen-AI service providers who are not responsible for the content.In this work, we proposeWMCopier, an effective watermark forgery attack that operates without requiring any prior knowledge of or access to the target watermarking algorithm. Our approach first models the target watermark distribution using an unconditional diffusion model, and then seamlessly embeds the target watermark into a non-watermarked image via a shallow inversion process. We also incorporate an iterative optimization procedure that refines the reconstructed image to further trade off the fidelity and forgery efficiency. Experimental results demonstrate that WMCopier effectively deceives both open-source and closed-source watermark systems (e.g., Amazon's system), achieving a significantly higher success rate than existing methods. Additionally, we evaluate the robustness of forged samples and discuss the potential defenses against our attack.Code is available at: https://anonymous.4open.science/r/WMCopier-E752.



Paperid:1019
Authors:Manjie Xu, Xinyi Yang, Jiayu Zhan, Wei Liang, Chi Zhang, Yixin Zhu
Abstract:
Self-play acts as a central role in enabling the autonomous acquisition of complex knowledge in structured environments, particularly for challenging tasks with compositional spaces where well-annotated data is limited. However, in more general learning scenarios, such symmetry for bootstrapping is often absent. For instance, in human education, teachers serve dual roles: as knowledge providers and problem proposers. They gradually present students with increasingly difficult problems as the students progress in mastering complex concepts. Although the roles of teacher and student are inherently different, they can be viewed as an adversarial pair, where the teacher challenges the student with problems, and the student responds by solving them. Inspired by this adversarial dynamic, we propose Heterogeneous Adversarial Play (HAP), an adversarial learning framework in which a teacher module generates problems tailored to the student's current capabilities. The two agents engage in a zero-sum game, driving the student's learning progress. Through multi-task learning experiments, we demonstrate that our framework achieves competitive performance relative to state-of-the-art baselines and facilitates more effective curriculum design for human learners. Our results highlight the potential of HAP to enhance the efficiency and robustness of AI learning systems.



Paperid:1020
Authors:Florent Foucaud, Harmender Gahlawat, Fionn Mc Inerney, Prafullkumar Tale
Abstract:
Abstract:The VC-dimension is a fundamental and well-studied measure of the complexity of a set system (or hypergraph) that is central to many areas of machine learning. We establish several new results on the complexity of computing the VC-dimension. In particular, given a hypergraph $\mathcal{H}=(\mathcal{V},\mathcal{E})$, we prove that the naive $2^{\mathcal{O}(|\mathcal{V}|)}$-time algorithm is optimal under the Exponential Time Hypothesis (\ETH). We then prove that the problem admits a $1$-additive fixed-parameter approximation algorithm when parameterized by the maximum degree of $\mathcal{H}$ and a fixed-parameter algorithm when parameterized by its dimension, and that these are essentially the only such exploitable structural parameters. Lastly, we consider a generalization of the problem, formulated using graphs, which captures the VC-dimension of both set systems and graphs. We show that it is fixed-parameter tractable parameterized by the treewidth of the graph (which, in the case of set systems, applies to the treewidth of its incidence graph). In contrast with closely related problems whose dependency on the treewidth is necessarily double-exponential (assuming the \ETH), our algorithm has a relatively low dependency on the treewidth.



Paperid:1021
Authors:Jakob Robnik, Reuben Cohn-Gordon, Uros Seljak
Abstract:
Sampling from high dimensional distributions is a computational bottleneck in many scientific applications. Hamiltonian Monte Carlo (HMC), and in particular the No-U-Turn Sampler (NUTS), are widely used, yet they struggle on problems with a very large number of parameters or a complicated geometry. Microcanonical Langevin Monte Carlo (MCLMC) has been recently proposed as an alternative which shows striking gains in efficiency over NUTS, especially for high-dimensional problems. However, it produces biased samples, with a bias that is hard to control in general. We introduce the Metropolis-Adjusted Microcanonical sampler (MAMS), which relies on the same dynamics as MCLMC, but introduces a Metropolis-Hastings step and thus produces asymptotically unbiased samples. We develop an automated tuning scheme for the hyperparameters of the algorithm, making it applicable out of the box. We demonstrate that MAMS outperforms NUTS across the board on benchmark problems of varying complexity and dimensionality, achieving up to a factor of seven speedup.



Paperid:1022
Authors:Zijian Li, Changze Zhou, Minghao Fu, Sanjay Manjunath, Fan Feng, Guangyi Chen, Yingyao Hu, Ruichu Cai, Kun Zhang
Abstract:
This paper is concerned with online time series forecasting, where unknown distribution shifts occur over time, i.e., latent variables influence the mapping from historical to future observations. To develop an automated way of online time series forecasting, we propose a Theoretical framework for Online Time-series forecasting (TOT in short) with theoretical guarantees. Specifically, we prove that supplying a forecaster with latent variables tightens the Bayes risk—the benefit endures under estimation uncertainty of latent variables and grows as the latent variables achieve a more precise identifiability. To better introduce latent variables into online forecasting algorithms, we further propose to identify latent variables with minimal adjacent observations. Based on these results, we devise a model-agnostic blueprint by employing a temporal decoder to match the distribution of observed variables and two independent noise estimators to model the causal inference of latent variables and mixing procedures of observed variables, respectively. Experiment results on synthetic data support our theoretical claims. Moreover, plug-in implementations built on several baselines yield general improvement across multiple benchmarks, highlighting the effectiveness in real-world applications.



Paperid:1023
Authors:Hyeongheon Cha, Dong Min Kim, Taesik Gong, Hye Won Chung, Sung-Ju Lee
Abstract:
Test-Time Adaptation (TTA) adjusts models using unlabeled test data to handle dynamic distribution shifts. However, existing methods rely on frequent adaptation and high computational cost, making them unsuitable for resource-constrained edge environments. To address this, we propose SNAP, a sparse TTA framework that reduces adaptation frequency and data usage while preserving accuracy. SNAP maintains competitive accuracy even when adapting based on only 1\% of the incoming data stream, demonstrating its robustness under infrequent updates. Our method introduces two key components: (i) Class and Domain Representative Memory (CnDRM), which identifies and stores a small set of samples that are representative of both class and domain characteristics to support efficient adaptation with limited data; and (ii) Inference-only Batch-aware Memory Normalization (IoBMN), which dynamically adjusts normalization statistics at inference time by leveraging these representative samples, enabling efficient alignment to shifting target domains. Integrated with five state-of-the-art TTA algorithms, SNAP reduces latency by up to 93.12\%, while keeping the accuracy drop below 3.3\%, even across adaptation rates ranging from 1\% to 50\%. This demonstrates its strong potential for practical use on edge devices serving latency-sensitive applications.



Paperid:1024
Authors:Zhenkun Zhu, Ruiqin Xiong, Jiyu Xie, Yuanlin Wang, Xinfeng Zhang, Tiejun Huang
Abstract:
As a bio-inspired vision sensor, spike camera records light intensity by accumulating photons and firing a spike once a preset threshold is reached. For high-light regions, the accumulated photons may reach the threshold multiple times within a readout interval, while only one spike can be stored and read out, resulting in incorrect intensity representation and a limited dynamic range. Multi-level (ML) spike camera enhances the dynamic range by introducing a spike-firing counter (SFC) to count spikes within each readout interval and uses different spike symbols to represent the arrival of different amounts of photons. However, when the light intensity becomes even higher, the energy consumption caused by highly frequent spike-firings can be huge. To address these issues, we propose time-encoding (TE) spike camera, which transforms the counting of spikes to recording of the time at which a specific number of spikes (i.e., an overflow) is reached. To encode time information with as few bits as possible, instead of directly utilising a timer, we leverage a periodic timing signal with a higher frequency than the readout signal. Then the recording of overflow moment can be transformed into recording the number of accumulated timing signal cycles until the overflow occurs. Additionally, we propose an image reconstruction scheme for TE spike camera, which leverages the multi-scale gradient features of spike data. This scheme includes a similarity-based pyramid alignment module to align spike streams across the temporal domain and a light intensity-based refinement module, which utilises the guidance of light intensity to fuse spatial features of the spike data. Experimental results demonstrate that TE spike camera effectively improves the dynamic range of spike camera.



Paperid:1025
Authors:Zachary Coalson, Juhan Bae, Nicholas Carlini, Sanghyun Hong
Abstract:
Abstract:We study how training data contributes to the emergence of toxic behaviors in large-language models. Most prior work on reducing model toxicity adopts *reactive* approaches, such as fine-tuning pre-trained (and potentially toxic) models to align them with human values. In contrast, we propose a *proactive* approach—IF-Guide—which leverages influence functions to identify harmful tokens within any training data and suppresses their impact during training. To this end, we first show that standard influence functions are ineffective at discovering harmful training records. We then present a novel adaptation that measures token-level attributions from training data to model toxicity, along with techniques for selecting toxic training documents and a learning objective that can be integrated into both pre-training and fine-tuning. Moreover, IF-Guide does not rely on human-preference data, which is typically required by existing alignment methods. In evaluation, we demonstrate that IF-Guide substantially reduces both explicit and implicit toxicity—by up to 10$\times$ compared to uncensored models, and up to 3$\times$ compared to baseline alignment methods, e.g., DPO and RAD—across both pre-training and fine-tuning scenarios. IF-Guide is computationally efficient: a billion-parameter model is *not necessary* for computing influence scores; a million-parameter model—with 7.5$\times$ fewer parameters—can effectively serve as a proxy for identifying harmful data.



Paperid:1026
Authors:XINYU DING, Bangtian Liu, Siyu Liao, Zhongfeng Wang
Abstract:
Abstract:Fast Fourier Transforms (FFT) are widely used to reduce memory and computational costs in deep learning. However, existing implementations, including standard FFT and real FFT (rFFT), cannot achieve true in-place computation. In particular, rFFT maps an input of size $n$ to a complex output of size $\frac{n}{2}+1$, causing dimensional mismatch and requiring additional memory allocation.We propose the first real-domain, fully in-place FFT framework (rdFFT) that preserves input-output dimensional consistency ($n \rightarrow n$). By leveraging butterfly operation symmetry and conjugate properties in the frequency domain, we design an implicit complex encoding scheme that eliminates intermediate cache usage entirely.Theoretically, our method reduces memory usage by 50\% compared to rFFTs. Moreover, it enables zero-cache parameter updates by utilizing the derivative property of the Fourier transform to compute matrix inverses efficiently without intermediate storage. Experiments on multiple natural language understanding tasks demonstrate the method’s effectiveness in maintaining model performance while significantly lowering memory overhead, offering a promising direction for frequency-domain lightweight adaptation.



Paperid:1027
Authors:Minchao Wu, Naoki Kobayashi
Abstract:
Craig interpolation plays a central role in formal verification tasks such as model checking, invariant generation, and abstraction refinement. In the domain of linear integer arithmetic (LIA), interpolants are crucial for deriving inductive invariants that characterize unreachable or safe program states, enabling scalable and precise reasoning about software and hardware correctness. Despite progress in interpolation algorithms, generating concise and interpretable interpolants remains a key challenge. We propose a lightweight learning-based approach to generating simple interpolants for LIA. Our model learns to lazily sample input problems directly and is complementary to existing logical methods. When Z3 is guided by our learned model, the complexity of the interpolants it produces can be reduced by up to 47.3%. For older solvers, the reduction rate can reach up to 69.1%.



Paperid:1028
Authors:Yiming Wang, Pei Zhang, Siyuan Huang, Baosong Yang, Zhuosheng Zhang, Fei Huang, Rui Wang
Abstract:
Abstract:Best-of-$N$ (BoN) sampling expands the search space at inference time to find better outputs, trading higher computational cost for improved performance.However, its cost–performance trade-off is still underexplored. Two main challenges limit the efficiency of BoN sampling:(1) Generating $N$ full samples consumes substantial GPU memory, reducing inference capacity under limited resources.(2) Reward models add extra memory and latency overhead, and training strong reward models introduces potential training data costs.Although some studies have explored efficiency improvements, none have addressed both challenges at once.To address this gap, we propose **Self-Truncation Best-of-$N$ (ST-BoN)**, a decoding method that avoids fully generating all $N$ samples and eliminates the need for reward models. It leverages early decoding consistency in the model’s internal states to identify promising paths and truncate suboptimal ones.In terms of cost, ST-BoN reduces dynamic GPU memory usage by over 80% and inference latency by 50%.In terms of cost–performance trade-off, ST-BoN achieves the same performance as Full-BoN while cutting computational cost by 70%–80%, and under the same cost, it can improve accuracy by 3–4 points.



Paperid:1029
Authors:Jonathan Li, Wei Cheng, Benjamin Riviere, Yue Wu, Masafumi Oyamada, Mengdi Wang, Yisong Yue, Santiago Paternain, Haifeng Chen
Abstract:
Inference scaling methods often rely on decomposing problems into steps, followed by sampling and selecting the best next steps. However, these steps and their sizes are typically fixed or depend on domain knowledge. We propose dynamic decomposition, a method that adaptively and automatically breaks down solution and reasoning traces into manageable steps during inference. By allocating compute more effectively—particularly by subdividing challenging steps and sampling them more frequently—dynamic decomposition significantly enhances inference efficiency. Experiments on benchmarks such as APPS, MATH, and LiveCodeBench demonstrate that dynamic decomposition outperforms static approaches, including token-level, sentence-level, and single-step decompositions. These findings highlight the potential of dynamic decomposition to improve a wide range of inference scaling techniques.



Paperid:1030
Authors:Jannis Halbey, Seta Rakotomandimby, Mathieu Besançon, Sébastien Designolle, Sebastian Pokutta
Abstract:
We develop a Frank-Wolfe algorithm with corrective steps, generalizing previous algorithms including blended pairwise conditional gradients and fully-corrective Frank-Wolfe, and propose a highly efficient corrective algorithm in the case of convex quadratic objectives based on linear optimization or linear system solving, akin to Wolfe's minimum-norm point.Beyond optimization problems that are directly quadratic, we revisit two algorithms, split conditional gradient for the intersection of two sets accessible through linear oracles, and second-order conditional gradient sliding, which approximately solves Variable-Metric projection subproblems, proposing improvement of the algorithms and their guarantees that may be of interest beyond our work, and we leverage quadratic corrections to accelerate the quadratic subproblems they solve.We show significant computational acceleration of Frank-Wolfe-based algorithms equipped with the quadratic correction on the considered problem classes.



Paperid:1031
Authors:Ting Liang, Xiaoliang Wu, Junyu Huang, Jianxin Wang, Qilong Feng
Abstract:
Clustering is a fundamental problem that has been extensively studied over past few decades, with most research focusing on point based clustering such as k-means, k-median, and k-center. However, numerous real-world applications, such as motion analysis, traffic monitoring, and trajectory modeling, require clustering over structured data, including lines, time series and affine subspaces (flats), where traditional point-based clustering algorithms often fall short. In this paper, we study the k-means of lines problem, where the input is a set L of lines in R^d, and the goal is to find k centers C in R^d such that the sum of squared distances from each line in L to its nearest center in C is minimized. The local search algorithm is a well-established strategy for point-based k-means clustering, known for its efficiency and provable approximation guarantees. However, extending local search algorithm to the k-means of lines problem is nontrivial, as the capture relation used in point-based clustering does not generalize to the line setting. This is because that the point-to-line distance function lack the triangle inequality property that supports geometric analysis in point-based clustering. Moreover, since lines extend infinitely in space, it is difficult to identify effective swap points that can significantly reduce the clustering cost. To overcome above obstacles, we introduce a proportional capture relation that links optimal and current centers based the assignment proportions of lines, enabling a refined analysis that bypasses the triangle inequality barrier. We also introduce a CrossLine structure, which provides a principled discretization of the geometric space around line pairs, and ensures coverage of high-quality swap points essential for local search, thereby enabling effective execution of the local search process. Consequently, based on the proposed components, we gave the first single-swap local search algorithm for the k-means of lines problem, achieving a (500+\varepsilon)-approximation in polynomial time.



Paperid:1032
Authors:Takayuki Osogami, Junya Honda, Junpei Komiyama
Abstract:
Abstract:We study the problem of estimating the mean reward of the best arm in a multi-armed bandit (MAB) setting. Specifically, given a target precision $\varepsilon$ and confidence level $1-\delta$, the goal is to return an $\varepsilon$-accurate estimate of the largest mean reward with probability at least $1-\delta$, while minimizing the number of samples. We first establish an instance-dependent lower bound on the sample complexity, using a novel approach to handle the infinitely many possible candidates of the estimated best mean. This lower bound is expressed in a non-convex optimization problem, which becomes the main difficulty of this problem preventing the direct application of standard techniques such as Track-and-Stop to provably achieve optimality. To overcome this difficulty, we introduce several new algorithmic and analytical techniques, and then propose an algorithm that achieves the asymptotic lower bound with matching constants in the leading term. Our method combines a confidence ellipsoid-based stopping condition with a two-phase sampling strategy tailored to manage the non-convexity. The proposed algorithm is simple, nearly free of hyperparameters, and achieves the instance-dependent, asymptotically optimal sample complexity. Experimental results support our theoretical guarantees and demonstrate the practical effectiveness of our method.



Paperid:1033
Authors:Yuanxin Liu, Rui Zhu, Shuhuai Ren, Jiacong Wang, Haoyuan Guo, Xu Sun, Lu Jiang
Abstract:
With the rapid growth of video generative models (VGMs), it is essential to develop reliable and comprehensive automatic metrics for AI-generated videos (AIGVs). Existing methods either use off-the-shelf models optimized for other tasks or rely on human assessment data to train specialized evaluators. These approaches are constrained to specific evaluation aspects and are difficult to scale with the increasing demands for finer-grained and more comprehensive evaluations. To address this issue, this work investigates the feasibility of using multimodal large language models (MLLMs) as a unified evaluator for AIGVs, leveraging their strong visual perception and language understanding capabilities. To evaluate the performance of automatic metrics in unified AIGV evaluation, we introduce a benchmark called UVE-Bench. UVE-Bench collects videos generated by state-of-the-art VGMs and provides pairwise human preference annotations across 15 evaluation aspects. Using UVE-Bench, we extensively evaluate 16 MLLMs. Our empirical results suggest that while advanced MLLMs (e.g., Qwen2VL-72B and InternVL2.5-78B) still lag behind human evaluators, they demonstrate promising ability in unified AIGV evaluation, significantly surpassing existing specialized evaluation methods. Additionally, we conduct an in-depth analysis of key design choices that impact the performance of MLLM-driven evaluators, offering valuable insights for future research on AIGV evaluation.



Paperid:1034
Authors:Michael Harding, Kirthevasan Kandasamy
Abstract:
We consider the problem of identifying the best arm in a multi-armed bandit model. Despite a wealth of literature in the traditional fixed budget and fixed confidence regimes of the best arm identification problem, it still remains a mystery to most practitioners as to how to choose an approach and corresponding budget or confidence parameter. We propose a new formalism to avoid this dilemma altogether by minimizing a risk functional which explicitly balances the performance of the recommended arm and the cost incurred by learning this arm. In this framework, a cost is incurred for each observation during the sampling phase, and upon recommending an arm, a performance penalty is incurred for identifying a suboptimal arm. The learner's goal is to minimize the sum of the penalty and cost. This new regime mirrors the priorities of many practitioners, e.g. maximizing profit in an A/B testing framework, better than classical fixed budget or confidence settings. We derive theoretical lower bounds for the risk of each of two choices for the performance penalty, the probability of misidentification and the simple regret, and propose an algorithm called DBCARE to match these lower bounds up to polylog factors on nearly all problem instances. We then demonstrate the performance of DBCARE on a number of simulated models, comparing to fixed budget and confidence algorithms to show the shortfalls of existing BAI paradigms on this problem.



Paperid:1035
Authors:Chang Liu, Yunfan Li, Lin Yang
Abstract:
Abstract:Safety is a fundamental challenge in reinforcement learning (RL), particularly in real-world applications such as autonomous driving, robotics, and healthcare. To address this, Constrained Markov Decision Processes (CMDPs) are commonly used to enforce safety constraints while optimizing performance. However, existing methods often suffer from significant safety violations or require a high sample complexity to generate near-optimal policies. We address two settings: relaxed feasibility, where small violations are allowed, and strict feasibility, where no violation is allowed. We propose a model-based primal-dual algorithm that balances regret and bounded constraint violations, drawing on techniques from online RL and constrained optimization. For relaxed feasibility, we prove that our algorithm returns an $\varepsilon$-optimal policy with $\varepsilon$-bounded violation with arbitrarily high probability, requiring $\tilde{O}\left(\frac{SAH^3}{\varepsilon^2}\right)$ learning episodes, matching the lower bound for unconstrained MDPs. For strict feasibility, we prove that our algorithm returns an $\varepsilon$-optimal policy with zero violation with arbitrarily high probability, requiring $\tilde{O}\left(\frac{SAH^5}{\varepsilon^2\zeta^2}\right)$ learning episodes, where $\zeta$ is the problem-dependent Slater constant characterizing the size of the feasible region. This result matches the lower bound for learning CMDPs with access to a generative model. Episodic tabular CMDPs serve as a crucial benchmark for safe RL, providing a structured environment for theoretical analysis and algorithmic validation. Our results demonstrate that learning CMDPs in an online setting is as easy as learning with a generative model and is no more challenging than learning unconstrained MDPs when small violations are allowed.



Paperid:1036
Authors:Daniel Augusto de Souza, Yuchen Zhu, Jake Cunningham, Yuri Saporito, Diego Mesquita, Marc Deisenroth
Abstract:
A variety of infinitely wide neural architectures (e.g., dense NNs, CNNs, and transformers) induce Gaussian process (GP) priors over their outputs.These relationships provide both an accurate characterization of the prior predictive distribution and enable the use of GP machinery to improve the uncertainty quantification of deep neural networks.In this work, we extend this connection to neural operators (NOs), a class of models designed to learn mappings between function spaces.Specifically, we show conditions for when arbitrary-depth NOs with Gaussian-distributed convolution kernels converge to function-valued GPs.Based on this result, we show how to compute the covariance functions of these NO-GPs for two NO parametrizations, including the popular Fourier neural operator (FNO).With this, we compute the posteriors of these GPs in realistic scenarios.This work is an important step towards uncovering the inductive biases of current FNO architectures and opens a path to incorporate novel inductive biases for use in kernel-based operator learning methods.



Paperid:1037
Authors:Pouya M. Ghari, simone sciabola, Ye Wang
Abstract:
Fine-tuning foundation models has emerged as a powerful approach for generating objects with specific desired properties. Reinforcement learning (RL) provides an effective framework for this purpose, enabling models to generate outputs that maximize a given reward function. However, in many applications such as text generation and drug discovery, it can be suboptimal to optimize using a single reward signal, as multiple evaluation criteria are often necessary. This paper proposes a novel reinforcement learning-based method for fine-tuning foundation models using multiple reward signals. By employing an iterative fine-tuning strategy across these rewards, our approach generalizes several state-of-the-art RL-based methods. We further provide a theoretical analysis that offers insights into the performance of multi-reward RL fine-tuning. Experimental results across diverse domains including text, biological sequence, and small molecule generation, demonstrate the effectiveness of the proposed algorithm compared to state-of-the-art baselines.



Paperid:1038
Authors:Wenlong Chen, Naoki Kiyohara, Harrison Zhu, Jacob Curran-Sebastian, Samir Bhatt, Yingzhen Li
Abstract:
We propose a novel online Gaussian process (GP) model that is capable of capturing long-term memory in sequential data in an online learning setting. Our model, Online HiPPO Sparse Variational Gaussian Process (OHSVGP), leverages the HiPPO (High-order Polynomial Projection Operators) framework, which is popularized in the RNN domain due to its long-range memory modeling capabilities. We interpret the HiPPO time-varying orthogonal projections as inducing variables with time-dependent orthogonal polynomial basis functions, which allows the SVGP inducing points to memorize the process history. We show that the HiPPO framework fits naturally into the interdomain GP framework and demonstrate that the kernel matrices can also be updated online in a recurrence form based on the ODE evolution of HiPPO. We evaluate OHSVGP with online prediction for 1D time series, continual learning in discriminative GP model for data with multidimensional inputs, and deep generative modeling with sparse Gaussian process variational autoencoder, showing that it outperforms existing online GP methods in terms of predictive performance, long-term memory preservation, and computational efficiency.



Paperid:1039
Authors:Drona Khurana, Anish Thilagar, Dhamma Kimpara, Rafael Frongillo
Abstract:
The statistical consistency of surrogate losses for discrete prediction tasks is often checked using the condition of calibration. However, directly verifying calibration can be arduous. Recent work shows that for polyhedral surrogates, a less arduous condition, indirect elicitation (IE), is still equivalent to calibration. We give the first results of this type for non-polyhedral surrogates, specifically the class of convex differentiable losses. We first prove that under mild conditions, IE and calibration are equivalent for one-dimensional losses in this class. We construct a counter-example that shows that this equivalence fails in higher dimensions. This motivates the introduction of strong IE, a strengthened form of IE that is equally easy to verify. We establish that strong IE implies calibration for differentiable surrogates and is both necessary and sufficient for strongly convex, differentiable surrogates. Finally, we apply these results to a range of problems to demonstrate the power of IE and strong IE for designing and analyzing consistent differentiable surrogates.



Paperid:1040
Authors:Giannis Daras, Adrian Rodriguez-Munoz, Adam Klivans, Antonio Torralba, Constantinos Daskalakis
Abstract:
We show how to use low-quality, synthetic, and out-of-distribution images to improve the quality of a diffusion model. Typically, diffusion models are trained on curated datasets that emerge from highly filtered data pools from the Web and other sources. We show that there is immense value in the lower-quality images that are often discarded. We present Ambient Diffusion Omni, a simple, principled framework to train diffusion models that can extract signal from arbitrarily images during training. Our framework exploits two properties of natural images -- spectral power law decay and locality. We first validate our framework by successfully training diffusion models with images synthetically corrupted by Gaussian blur, JPEG compression, and motion blur. We use our framework to achieve state-of-the-art ImageNet FID and we show significant improvements in both image quality and diversity for text-to-image generative modeling. The core insight is that noise dampens the initial skew between the desired high-quality distribution and the mixed distribution we actually observe. We provide rigorous theoretical justification for our approach by analyzing the trade-off between learning from biased data versus limited unbiased data across diffusion times.



Paperid:1041
Authors:Sasha Voitovych, Mahdi Haghifam, Idan Attias, Gintare Karolina Dziugaite, Roi Livni, Dan Roy
Abstract:
Abstract:In this paper, we investigate the necessity of traceability for accurate learning in stochastic convex optimization (SCO) under $\ell_p$ geometries. Informally, we say a learning algorithm is \emph{$m$-traceable} if, by analyzing its output, it is possible to identify at least $m$ of its training samples. Our main results uncover a fundamental tradeoff between traceability and excess risk in SCO. For every $p\in [1,\infty)$, we establish the existence of an excess risk threshold below which every sample-efficient learner is traceable with the number of samples which is a \emph{constant fraction} of its training sample. For $p\in [1,2]$, this threshold coincides with the best excess risk of differentially private (DP) algorithms, i.e., above this threshold, there exist algorithms that are not traceable, which corresponds to a sharp phase transition. For $p \in (2,\infty)$, this threshold instead gives novel lower bounds for DP learning, partially closing an open problem in this setup. En route to establishing these results, we prove a sparse variant of the fingerprinting lemma, which is of independent interest to the community.



Paperid:1042
Authors:Bohao Liao, Wei Zhai, Zengyu Wan, Zhixin Cheng, Wenfei Yang, Yang Cao, Tianzhu Zhang, Zheng-Jun Zha
Abstract:
Scene reconstruction from casually captured videos has wide real-world applications. Despite recent progress, existing methods relying on traditional cameras tend to fail in high-speed scenarios due to insufficient observations and inaccurate pose estimation. Event cameras, inspired by biological vision, record pixel-wise intensity changes asynchronously with high temporal resolution and low latency, providing valuable scene and motion information in blind inter-frame intervals. In this paper, we introduce the event cameras to aid scene construction from a casually captured video for the first time, and propose Event-Aided Free-Trajectory 3DGS, called EF-3DGS, which seamlessly integrates the advantages of event cameras into 3DGS through three key components. First, we leverage the Event Generation Model (EGM) to fuse events and frames, enabling continuous supervision between discrete frames. Second, we extract motion information through Contrast Maximization (CMax) of warped events, which calibrates camera poses and provides gradient-domain constraints for 3DGS. Third, to address the absence of color information in events, we combine photometric bundle adjustment (PBA) with a Fixed-GS training strategy that separates structure and color optimization, effectively ensuring color consistency across different views. We evaluate our method on the public Tanks and Temples benchmark and a newly collected real-world dataset, RealEv-DAVIS. Our method achieves up to 3dB higher PSNR and 40% lower Absolute Trajectory Error (ATE) compared to state-of-the-art methods under challenging high-speed scenarios.



Paperid:1043
Authors:Hyunsu Kim, Jonggeon Park, Joan Bruna, Hongseok Yang, Juho Lee
Abstract:
The advent of foundation models in AI has significantly advanced general-purpose learning, enabling remarkable capabilities in zero-shot inference and in-context learning. However, training such models on physics data, including solutions to partial differential equations (PDEs), poses a unique challenge due to varying dimensionalities across different systems. Traditional approaches either fix a maximum dimension or employ separate encoders for different dimensionalities, resulting in inefficiencies. To address this, we propose a dimension-agnostic neural network architecture, the Axial Neural Network (XNN), inspired by permutation equivariant structures such as Deep Sets and Graph Neural Networks. XNN generalizes across varying tensor dimensions while maintaining computational efficiency. We convert existing PDE foundation models into axial neural networks and evaluate their performance across three training scenarios: training from scratch, pretraining on multiple PDEs, and fine-tuning on a single PDE. Our experiments show that XNNs perform competitively with original models and exhibit superior generalization to unseen dimensions, highlighting the importance of multidimensional pretraining for foundation models.



Paperid:1044
Authors:Jiahao Yu, Zelei Cheng, Xian Wu, Xinyu Xing
Abstract:
Large language models (LLMs) are increasingly used in various domains, showing impressive potential on various tasks. Recently, reasoning LLMs have been proposed to improve the \textit{reasoning} or \textit{thinking} capabilities of LLMs to solve complex problems. Despite the promising results of reasoning LLMs, enhancing the multi-step reasoning capabilities of LLMs still remains a significant challenge. While existing optimization methods have advanced the LLM reasoning capabilities, they often treat reasoning trajectories as a whole, without considering the underlying critical steps within the trajectory. In this paper, we introduce \textbf{G}uided \textbf{P}ivotal \textbf{O}ptimization (GPO), a novel fine-tuning strategy that dives into the reasoning process to enable more effective improvements. GPO first identifies the `critical step' within a reasoning trajectory - a point that the model must carefully proceed so as to succeed at the problem. We locate the critical step by estimating the advantage function. GPO then resets the policy to the critical step and samples the new rollout and prioritizes learning process on those rollouts. This focus allows the model to learn more effectively from pivotal moments within the reasoning process to improve the reasoning performance. We demonstrate that GPO is not a standalone method, but rather a general strategy that can be integrated with various optimization methods to improve reasoning performance. Besides theoretical analysis, our experiments across challenging reasoning benchmarks show that GPO can consistently and significantly enhances the performance of existing optimization methods, showcasing its effectiveness and generalizability in improving LLM reasoning by concentrating on pivotal moments within the generation process.



Paperid:1045
Authors:Zhonglin Xie, Yiman Fong, Haoran Yuan, Zaiwen Wen
Abstract:
Optimization is an important module of modern machine learning applications. Tremendous efforts have been made to accelerate optimization algorithms. A common formulation is achieving a lower loss at a given time. This enables a differentiable framework with respect to the algorithm hyperparameters. In contrast, its dual, minimizing the time to reach a target loss, is believed to be non-differentiable, as the time is not differentiable. As a result, it usually serves as a conceptual framework or is optimized using zero-th order methods. To address this limitation, we propose a differentiable stopping time (DST) and theoretically justify it based on differential equations. An efficient algorithm is designed to backpropagate through DST. As a result, the proposed DST enables a new differentiable formulation for accelerating algorithms. We further discuss its applications, such as online hyperparameter tuning and learning to optimize. Our proposed methods show superior performance in comprehensive experiments across various problems, which confirms their effectiveness.



Paperid:1046
Authors:Zander W. Blasingame, Chen Liu
Abstract:
Training-free guided generation is a widely used and powerful technique that allows the end user to exert further control over the generative process of flow/diffusion models.Generally speaking, two families of techniques have emerged for solving this problem forgradient-based guidance: namely,posterior guidance(i.e., guidance via projecting the current sample to the target distribution via the target prediction model) andend-to-end guidance(i.e., guidance by performing backpropagation throughout the entire ODE solve). In this work, we show that these two seemingly separate families can actually beunifiedby looking at posterior guidance as agreedy strategyofend-to-end guidance. We explore the theoretical connections between these two families and provide an in-depth theoretical of these two techniques relative to thecontinuous ideal gradients. Motivated by this analysis we then show a method forinterpolatingbetween these two families enabling a trade-off between compute and accuracy of the guidance gradients.We then validate this work on several inverse image problems and property-guided molecular generation.



Paperid:1047
Authors:Zehong Ma, Longhui Wei, Feng Wang, Shiliang Zhang, Qi Tian
Abstract:
Existing acceleration techniques for video diffusion models often rely on uniform heuristics or time-embedding variants to skip timesteps and reuse cached features. These approaches typically require extensive calibration with curated prompts and risk inconsistent outputs due to prompt-specific overfitting. In this paper, we introduce a novel and robust discovery: a unified magnitude law observed across different models and prompts. Specifically, the magnitude ratio of successive residual outputs decreases monotonically, steadily in most timesteps while rapidly in the last several steps. Leveraging this insight, we introduce a Magnitude-aware Cache (MagCache) that adaptively skips unimportant timesteps using an error modeling mechanism and adaptive caching strategy. Unlike existing methods requiring dozens of curated samples for calibration, MagCache only requires a single sample for calibration. Experimental results show that MagCache achieves 2.1× and 2.68× speedups on Open-Sora and Wan 2.1, respectively, while preserving superior visual fidelity. It significantly outperforms existing methods in LPIPS, SSIM, and PSNR, under comparable computational budgets.



Paperid:1048
Authors:Alexander Chervov, Kirill Khoruzhii, Nikita Bukhal, Jalal Naghiyev, Vladislav Zamkovoy, Ivan Koltsov, Lyudmila Cheldieva, Arsenii Sychev, Arsenii Lenin, Mark Obozov, Egor Urvanov, Alexey Romanov
Abstract:
Abstract:The paper proposes a novel machine learning-based approach to the pathfinding problem on extremely large graphs. This method leverages diffusion distance estimation via a neural network and uses beam search for pathfinding. We demonstrate its efficiency by finding solutions for 4x4x4 and 5x5x5 Rubik's cubes with unprecedentedly short solution lengths, outperforming all available solvers and introducing the first machine learning solver beyond the 3x3x3 case. In particular, it surpasses every single case of the combined best results in the Kaggle Santa 2023 challenge, which involved over 1,000 teams. For the 3x3x3 Rubik's cube, our approach achieves an optimality rate exceeding 98%, matching the performance of task-specific solvers and significantly outperforming prior solutions such as DeepCubeA (60.3%) and EfficientCube (69.6%). Our solution in its current implementation is approximately 25.6 times faster in solving 3x3x3 Rubik's cubes while requiring up to 8.5 times less model training time than the most efficient state-of-the-art competitor. Finally, it is demonstrated that even a single agent trained using a relatively small number of examples can robustly solve a broad range of puzzles represented by Cayley graphs of size up to $10^{145}$, confirming the generality of the proposed method.



Paperid:1049
Authors:Jiayuan Ye, Vitaly Feldman, Kunal Talwar
Abstract:
Abstract:We study the fundamental problem of estimating an unknown discrete distribution $p$ over $d$ symbols, given $n$ i.i.d. samples from the distribution. We are interested in minimizing the KL divergence between the true distribution and the algorithm's estimate. We first construct minimax optimal private estimators. Minimax optimality however fails to shed light on an algorithm's performance on individual (non-worst-case) instances $p$ and simple minimax-optimal DP estimators can have poor empirical performance on real distributions. We then study this problem from an instance-optimality viewpoint, where the algorithm's error on $p$ is compared to the minimum achievable estimation error over a small local neighborhood of $p$. Under natural notions of local neighborhood, we propose algorithms that achieve instance-optimality up to constant factors, with and without a differential privacy constraint. Our upper bounds rely on (private) variants of the Good-Turing estimator. Our lower bounds use additive local neighborhoods that more precisely captures the hardness of distribution estimation in KL divergence, compared to ones considered in prior works.



Paperid:1050
Authors:Marco Bertolini, Tuan Le, Djork-Arné Clevert
Abstract:
Abstract:We introduce a novel class of score-based diffusion processes that operate directly in the representation space of Lie groups.Leveraging the framework of Generalized Score Matching, we derive a class of Langevin dynamics that decomposes as a direct sum of Lie algebra representations,enabling the modeling of any target distribution on any (non-Abelian) Lie group.Standard score-matching emerges as a special case of our framework when the Lie group is the translation group. We prove that our generalized generative processes arise as solutions to a new class of paired stochastic differential equations (SDEs), introduced here for the first time. We validate our approach through experiments on diverse data types, demonstrating its effectiveness in real-world applications such as $\text{SO}(3)$-guided molecular conformer generation and modeling ligand-specific global $\text{SE}(3)$ transformations for molecular docking, showing improvement in comparison to Riemannian diffusion on the group itself.We show that an appropriate choice of Lie group enhances learning efficiency by reducing the effective dimensionality of the trajectory space and enables the modeling of transitions between complex data distributions.



Paperid:1051
Authors:Sherman Khoo, Yakun Wang, Song Liu, Mark Beaumont
Abstract:
We study the problem of likelihood maximization when the likelihood function is intractable but model simulations are readily available. We propose a sequential, gradient-based optimization method that directly models the Fisher score based on a local score matching technique which uses simulations from a localized region around each parameter iterates. By employing a linear parameterization to the surrogate score model, our technique admits a closed-form, least-squares solution. This approach yields a fast, flexible, and efficient approximation to the Fisher score, effectively smoothing the likelihood objective and mitigating the challenges posed by complex likelihood landscapes. We provide theoretical guarantees for our score estimator, including bounds on the bias introduced by the smoothing. Empirical results on a range of synthetic and real-world problems demonstrate the superior performance of our method compared to existing benchmarks.



Paperid:1052
Authors:Weiyu Chen, Hansi Yang, Yunhao Gou, Han Shi, Enliang Hu, Zhenguo Li, James Kwok
Abstract:
Large Language Models (LLMs) have demonstrated remarkable capabilities across various tasks but require substantial computational resources, limiting their deployment in resource-constrained environments. While one-shot pruning methods can reduce model size without expensive retraining, they typically optimize for single objectives, ignoring LLMs' multi-faceted applications. We introduce Multi-Objective One-Shot Pruning (MOSP), which formulates LLM pruning as a multi-objective optimization problem. MOSP efficiently generates a Pareto set of pruned models representing different capability trade-offs, allowing users to select solutions aligned with their preferences. The proposed approach identifies share core support while enabling specialized support. Experiments across various LLMs and sparsity levels demonstrate MOSP's superior performance in navigating multi-objective trade-offs compared to baseline methods.



Paperid:1053
Authors:Qianhui Wu, Kanzhi Cheng, Rui Yang, Chaoyun Zhang, Jianwei Yang, Huiqiang Jiang, Jian Mu, Baolin Peng, Bo Qiao, Reuben Tan, Si Qin, Lars Liden, Qingwei Lin, Huan Zhang, Tong Zhang, Jianbing Zhang, Dongmei Zhang, Jianfeng Gao
Abstract:
Abstract:One of the principal challenges in building VLM-powered GUI agents is visual grounding—localizing the appropriate screen region for action execution based on both the visual content and the textual plans. Most existing work formulates this as a text-based coordinate generation task. However, these approaches suffer from several limitations: weak spatial-semantic alignment due to the lack of explicit spatial supervision; inability to handle ambiguous supervision targets, as single-point predictions penalize valid variations; and a mismatch between the fine-grained nature of screen coordinates and the coarse, patch-level granularity of visual features extracted by models like Vision Transformers. In this paper, we propose GUI-Actor, a VLM-based method for coordinate-free GUI grounding. At its core, GUI-Actor introduces an attention-based action head that learns to align a dedicatedtoken with all relevant visual patch tokens, enabling the model to propose one or more action regions in a single forward pass. In line with this, we design a grounding verifier to evaluate and select the most plausible action region from the candidates proposed for action execution. Extensive experiments show that GUI-Actor outperforms prior state-of-the-art methods on multiple GUI action grounding benchmarks, with improved generalization to unseen screen resolutions and layouts. Additionally, we show that by leveraging the verifier, fine-tuning only the newly introduced action head ($\sim$100M parameters) while keeping the VLM backbone frozen is sufficient to achieve performance comparable to previous state-of-the-art models, highlighting that GUI-Actor can endow the underlying VLM with effective grounding capabilities without compromising its general-purpose strengths.



Paperid:1054
Authors:YAOYAO YAN, Hui Yu, Weizhi Xu
Abstract:
Transformers have been widely adopted in natural language processing, computer vision, and other domains due to their exceptional performance across a variety of tasks. However, the computational cost of Transformers is prohibitively high, particularly when handling long input sequences, significantly increasing both training and inference time. Although various token pruning methods have been proposed to reduce the computational burden of Transformers, most approaches overlook critical differences in sequences in terms of length and complexity, leading to suboptimal compression efficiency.In this paper, we propose AD-TP, an Attribution-Driven Adaptive Token Pruning method designed to retain only the most informative tokens. We analyze the performance of using accumulated attention values to measure token importance and find that attention values do not accurately reflect the actual contribution of each token to text understanding. Additionally, we observe significant variations in the length and complexity of different sequences within the dataset. Based on these insights, we adopt Integrated Gradients to evaluate token importance and introduce a lightweight adaptive token retainer module that dynamically generates pruning configurations for each input sequence. In addition, we incorporate both teacher supervision and self-supervised learning objectives to enhance the training efficiency, accuracy, and robustness of the model.Experiments conducted on the GLUE benchmark, SQuAD, and 20News datasets demonstrate that AD-TP outperforms state-of-the-art token pruning and model compression methods in both accuracy and computational efficiency. On GLUE, AD-TP reduces FLOPs by an average of 7.8× while improving performance by 0.6%.



Paperid:1055
Authors:Moritz Grillo, Christoph Hertrich, Georg Loho
Abstract:
Abstract:We contribute towards resolving the open question of how many hidden layers are required in ReLU networks for exactly representing all continuous and piecewise linear functions on $\mathbb{R}^d$. While the question has been resolved in special cases, the best known lower bound in general is still 2. We focus on neural networks that are compatible with certain polyhedral complexes, more precisely with the braid fan. For such neural networks, we prove a non-constant lower bound of $\Omega(\log\log d)$ hidden layers required to exactly represent the maximum of $d$ numbers. Additionally, we provide a combinatorial proof that neural networks satisfying this assumption require three hidden layers to compute the maximum of 5 numbers; this had only been verified with an excessive computation so far.Finally, we show that a natural generalization of the best known upper bound to maxout networks is not tight, by demonstrating that a rank-3 maxout layer followed by a rank-2 maxout layer is sufficient to represent the maximum of 7 numbers.



Paperid:1056
Authors:Yichao Fu, Junda Chen, Siqi Zhu, Fu, Zhongdongming Dai, Yonghao Zhuang, Yian Ma, Aurick Qiao, Tajana S Rosing, Ion Stoica, Hao Zhang
Abstract:
Abstract:Test-time reasoning algorithms such as chain-of-thought, self-consistency, and MCTS enhance LLM problem-solving but can wastefully generate many tokens without improving accuracy. At the same time, we observe that these algorithms exhibit answer stabilization: their intermediate solutions often cease to change after a certain point, and further investment of compute does not change their final answer. To quantify this phenomenon, we introduce Certaindex, an algorithm-agnostic metric measuring this evolving stability, signaling when further computation is unlikely to alter the final result. Certaindex is lightweight, can accelerate reasoning program inference via early exit, and further enables dynamic token allocation, gang scheduling, and many opportunities when integrated with real-world LLM serving systems. To quantify real-world benefits, we built Certaindex as a scheduler into Dynasor, our reasoning-aware LLM serving system, and demonstrate up to 50\% compute savings and 3.3$\times$ higher throughput in real workloads with no accuracy drop, with only $\sim500$ lines of code changes in the core system.



Paperid:1057
Authors:Kyowoon Lee, Jaesik Choi
Abstract:
Diffusion-based generative models are emerging as powerful tools for long-horizon planning in reinforcement learning (RL), particularly with offline datasets. However, their performance is fundamentally limited by the quality and diversity of training data. This often restricts their generalization to tasks outside their training distribution or longer planning horizons. To overcome this challenge, we proposeState-Covering Trajectory Stitching(SCoTS), a novel reward-free trajectory augmentation method that incrementally stitches together short trajectory segments, systematically generating diverse and extended trajectories. SCoTS first learns a temporal distance-preserving latent representation that captures the underlying temporal structure of the environment, then iteratively stitches trajectory segments guided by directional exploration and novelty to effectively cover and expand this latent space. We demonstrate that SCoTS significantly improves the performance and generalization capabilities of diffusion planners on offline goal-conditioned benchmarks requiring stitching and long-horizon reasoning. Furthermore, augmented trajectories generated by SCoTS significantly improve the performance of widely used offline goal-conditioned RL algorithms across diverse environments.



Paperid:1058
Authors:Lennart Finke, Chandan Sreedhara, Thomas Dooms, Mat Allen, Juan Rodriguez, Noa Nabeshima, Thomas Marshall, Dan Braun
Abstract:
We present SimpleStories, a large synthetic story dataset in simple language, consisting of 2 million samples each in English and Japanese. Through parameterizing prompts at multiple levels of abstraction, we achieve control over story characteristics at scale, inducing syntactic and semantic diversity. Ablations on a newly trained tiny model suite then show improved sample efficiency and model interpretability in comparison with the TinyStories dataset. We open-source all constituent parts of model creation, hoping to enable novel ways to study the end-to-end training process. As a byproduct, we move the frontier with regards to the fewest-parameter language model that outputs grammatical English.



Paperid:1059
Authors:Qiaobo Li, Zhijie Chen, Arindam Banerjee
Abstract:
Abstract:Communication cost and privacy are two major considerations in federated learning (FL). For communication cost, gradient compression by sketching the clients’ transmitted model updates is often used for reducing per‐round communication. For privacy, the Gaussian mechanism (GM), which consists of clipping updates and adding Gaussian noise, is commonly used to guarantee client‐level differential privacy. Existing literature on private FL analyzes privacy of sketching and GM in an isolated manner, illustrating that sketching provides privacy determined by the sketching dimension and that GM has to supply any additional desired privacy. In this paper, we introduce the Sketched Gaussian Mechanism (SGM), which directly combines sketching and the Gaussian mechanism for privacy. Using Rényi-DP tools, we present a joint analysis of SGM's overall privacy guarantee, which is significantly more flexible and sharper compared to isolated analysis of sketching and GM privacy. In particular, we prove that the privacy level of SGM for a fixed noise magnitude is proportional to $1/\sqrt{b}$, where $b$ is the sketching dimension, indicating that (for moderate $b$) SGM can provide much stronger privacy guarantees than the original GM under the same noise budget. We demonstrate the application of SGM to FL with either gradient descent or adaptive server optimizers, and establish theoretical results on optimization convergence, which exhibits only a logarithmic dependence on the number of parameters $d$. Experimental results confirm that at the same privacy level, SGM based FL is at least competitive with non‐sketching private FL variants and outperforms them in some settings. Moreover, using adaptive optimization at the server improves empirical performance while maintaining the privacy guarantees.



Paperid:1060
Authors:Faraz Zargari, Hossein Jazi, Lyndon Hallett, Bo Sun, Xiaoqi Tan
Abstract:
We study the online multi-class selection problem with group fairness guarantees, where limited resources must be allocated to sequentially arriving agents. Our work addresses two key limitations in the existing literature. First, we introduce a novel lossless rounding scheme that ensures the integral algorithm achieves the same expected performance as any fractional solution. Second, we explicitly address the challenges introduced by agents who belong to multiple classes. To this end, we develop a randomized algorithm based on a relax-and-round framework. The algorithm first computes a fractional solution using a resource reservation approach---referred to as theset-asidemechanism---to enforce fairness across classes. The subsequent rounding step preserves these fairness guarantees without degrading performance. Additionally, we propose a learning-augmented variant that incorporates untrusted machine-learned predictions to better balance fairness and efficiency in practical settings.



Paperid:1061
Authors:Yuli Zhou, Guolei Sun, Yawei Li, Yuqian Fu, Luca Benini, Ender Konukoglu
Abstract:
Video camouflaged object segmentation (VCOS), aiming at segmenting camouflaged objects that seamlessly blend into their environment, is a fundamental vision task with various real-world applications. With the release of SAM2, video segmentation has witnessed significant progress. However, SAM2's capability of segmenting camouflaged videos is suboptimal, especially when given simple prompts such as point and box. To address the problem, we propose Camouflaged SAM2 (CamSAM2), which enhances SAM2's ability to handle camouflaged scenes without modifying SAM2's parameters. Specifically, we introduce a decamouflaged token to provide the flexibility of feature adjustment for VCOS. To make full use of fine-grained and high-resolution features from the current frame and previous frames, we propose implicit object-aware fusion (IOF) and explicit object-aware fusion (EOF) modules, respectively. Object prototype generation (OPG) is introduced to abstract and memorize object prototypes with informative details using high-quality features from previous frames. Extensive experiments are conducted to validate the effectiveness of our approach. While CamSAM2 only adds negligible learnable parameters to SAM2, it substantially outperforms SAM2 on three VCOS datasets, especially achieving 12.2 mDice gains with click prompt on MoCA-Mask and 19.6 mDice gains with mask prompt on SUN-SEG-Hard, with Hiera-T as the backbone. The code will be released.



Paperid:1062
Authors:Qirui Mi, Qipeng Yang, Zijun Fan, Wentian Fan, Heyang Ma, Chengdong Ma, Siyu Xia, Bo An, Haifeng Zhang, Jun Wang
Abstract:
Artificial intelligence (AI) has become a powerful tool for economic research, enabling large-scale simulation and policy optimization. However, applying AI effectively requires simulation platforms for scalable training and evaluation—yet existing environments remain limited to simplified, narrowly scoped tasks, falling short of capturing complex economic challenges such as demographic shifts, multi-government coordination, and large-scale agent interactions.To address this gap, we introduce EconGym, a scalable and modular testbed that connects diverse economic tasks with AI algorithms. Grounded in rigorous economic modeling, EconGym implements 11 heterogeneous role types (e.g., households, firms, banks, governments), their interaction mechanisms, and agent models with well-defined observations, actions, and rewards. Users can flexibly compose economic roles with diverse agent algorithms to simulate rich multi-agent trajectories across 25+ economic tasks for AI-driven policy learning and analysis.Experiments show that EconGym supports diverse and cross-domain tasks—such as coordinating fiscal, pension, and monetary policies—and enables benchmarking across AI, economic methods, and hybrids. Results indicate that richer task composition and algorithm diversity expand the policy space, while AI agents guided by classical economic methods perform best in complex settings. EconGym also scales to 10k agents with high realism and efficiency.



Paperid:1063
Authors:Manmatha Roy, Arijit Ghosh
Abstract:
Abstract:A function \( f : \mathbb{F}_2^n \to \mathbb{R} \) is said to be \( s \)-Fourier sparse if its Fourier expansion contains at most \( s \) non-zero coefficients. Fourier sparsity is a key structural property that facilitates efficient representation, computation, and learning of functions. In general, the existence of a sparse representation in the Fourier basis facilitates the development of efficient learning algorithms. However, most of these techniques rely on prior knowledge of the function's Fourier sparsity, with the algorithmic parameters tailored based on this information.This motivates the following decision problem: given \( s > 0 \), determine whether a given function is (approximately) \( s \)-Fourier sparse.This work addresses the problem of tolerant testing of real-valued functions over $\mathbb{F}_2^n$, accessed via queries, to determine whether the function is close to being $s$-Fourier sparse or significantly far from every $s$-Fourier sparse function. Building on the \emph{distance approximation via tolerant testing} framework developed by Parnas et al., our tester naturally yields an estimator that, given oracle access to the function, approximates its distance to the closest $s$-Fourier sparse function, with formal guarantees on both accuracy and confidence. Our algorithm has query complexity $\widetilde{O}(s)$ for constant accuracy and confidence parameters.A key structural ingredient in our analysis is a new spectral concentration result for real-valued functions defined over $\mathbb{F}_2^n$ under restriction to small-dimensional random affine subspaces. We complement our upper bound with a matching lower bound of $\Omega(s)$, showing that our tester is optimal up to logarithmic factors. Our lower bound leverages spectral properties of a certain class of cryptographically hard functions, known as the Maiorana–McFarland class, in a novel way.



Paperid:1064
Authors:Farnam Mansouri, Hans Simon, Adish Singla, Yuxin Chen, Sandra Zilles
Abstract:
Machine learning can greatly benefit from providing learning algorithms with pairs of contrastive training examples---typically pairs of instances that differ only slightly, yet have different class labels. Intuitively, the difference in the instances helps explain the difference in the class labels. This paper proposes a theoretical framework in which the effect of various types of contrastive examples on active learners is studied formally. The focus is on the sample complexity of learning concept classes and how it is influenced by the choice of contrastive examples. We illustrate our results with geometric concept classes and classes of Boolean functions. Interestingly, we reveal a connection between learning from contrastive examples and the classical model of self-directed learning.



Paperid:1065
Authors:Zhaoqun Li, Beishui Liao, Qiwei Ye
Abstract:
Mathematicians have long employed decomposition techniques to prove inequalities, yet automating this process remains a significant challenge in computational mathematics.We introduce IneqSearch, a hybrid reasoning system that integrates symbolic computation with large language models (LLMs) to address this challenge.IneqSearch reformulates inequality proving as a structured search problem: identifying appropriate combinations of theorems that decompose expressions into non-negative components.The system combines a symbolic solver for deductive reasoning with an LLM-based agent for constructive proof exploration, effectively implementing methodologies observed in formal mathematical practice.A key contribution of IneqSearch is its iterative learning mechanism that systematically incorporates newly proven results into its theorem database, enabling knowledge acquisition during practice that enhances its capabilities without requiring human intervention.In empirical evaluation on 437 Olympiad-level inequalities, IneqSearch successfully proves 342 problems, significantly outperforming existing methods and demonstrating the effectiveness of integrating symbolic and neural approaches for mathematical reasoning.



Paperid:1066
Authors:Omer Shubi, David Robert Reich, Keren Gruteke Klein, Yuval Angel, Paul Prasse, Lena A. Jäger, Yevgeni Berzak
Abstract:
We introduceEyeBench, a first-of-its-kind benchmark for evaluating multimodal machine learning models that decode cognitive and linguistic information from eye movements during reading. EyeBench provides an accessible entry point to the highly challenging and underexplored domain of eye movements paired with text, with the aim of driving innovation in multimodal AI and cognitive science. The benchmark provides a standardized framework for evaluating predictive models on a diverse set of datasets and tasks, ranging from prediction of reading comprehension to automatic assessment of a reader's experienced text difficulty. Progress on the EyeBench challenge will open the door for both practical real-world applications, such as adaptive user interfaces and personalized education, and scientific advances in understanding human language processing.



Paperid:1067
Authors:Siru Ouyang, Xinyu Zhu, Zilin Xiao, Minhao Jiang, Yu Meng, Jiawei Han
Abstract:
Reinforcement learning (RL) has become a powerful approach for improving the reasoning capabilities of large language models (LLMs), as evidenced by recent successes such as OpenAI’s o1 and Deepseek-R1. However, applying RL at scale remains intimidatingly resource-intensive, requiring multiple model copies and extensive GPU workloads. On the other hand, while being powerful, recent studies suggest that RL does not fundamentally endow models with new knowledge; rather, it primarily reshapes the model's output distribution to activate reasoning capabilities latent in the base model. Building on this insight, we hypothesize that the changes in output probabilities induced by RL are largely model-size invariant, opening the door to a more efficient paradigm: training a small model with RL and transferring its induced probability shifts to larger base models. To verify our hypothesis, we conduct a token-level analysis of decoding trajectories and find high alignment in RL-induced output distributions across model scales, validating our hypothesis. Motivated by this, we propose RAST, a simple yet effective method that transfers reasoning behaviors by injecting RL-induced probability adjustments from a small RL-trained model into larger models. Experiments across multiple mathematical reasoning benchmarks show that RAST significantly improves performance, often matching or surpassing direct RL training, while also increasing diversity in decoding trajectories. Our findings offer new insights into the nature of RL-driven reasoning and practical strategies for scaling its benefits without incurring its full computational cost. Our code is available at https://anonymous.4open.science/r/RASTid18137.



Paperid:1068
Authors:Beatrice Bertolotti, Matteo Russo, Chris Schwiegelshohn, Sudarshan Shyam
Abstract:
Abstract:We study the sublinear multivariate mean estimation problem in $d$-dimensional Euclidean space. Specifically, we aim to find the mean $\mu$ of a ground point set $A$, which minimizes the sum of squared Euclidean distances of the points in $A$ to $\mu$. We first show that a multiplicative $(1+\varepsilon)$ approximation to $\mu$ can be found with probability $1-\delta$ using $O(\varepsilon^{-1}\log \delta^{-1})$ many independent uniform random samples, and provide a matching lower bound. Furthermore, we give two estimators with optimal sample complexity that can be computed in optimal running time for extracting a suitable approximate mean:1. The coordinate-wise median of $\log \delta^{-1}$ sample means of sample size $\varepsilon^{-1}$. As a corollary, we also show improved convergence rates for this estimator for estimating means of multivariate distributions.2. The geometric median of $\log \delta^{-1}$ sample means of sample size $\varepsilon^{-1}$. To compute a solution efficiently, we design a novel and simple gradient descent algorithm that is significantly faster for our specific setting than all other known algorithms for computing geometric medians.In addition, we propose an order statistics approach that is empirically competitive with these algorithms, has an optimal sample complexity and matches the running time up to lower order terms.We finally provide an extensive experimental evaluation among several estimators which concludes that the geometric-median-of-means-based approach is typically the most competitive in practice.



Paperid:1069
Authors:Blossom Metevier, Dennis Wei, Karthikeyan Natesan Ramamurthy, Philip Thomas
Abstract:
Recent works have studied fair resource allocation in social settings, where fairness is judged by the impact of allocation decisions rather than more traditional minimum or maximum thresholds on the allocations themselves. Our work significantly adds to this literature by developing continuous resource allocation strategies that adhere to equality of impact, a generalization of equality of opportunity. We derive methods to maximize total welfare across groups subject to minimal violation of equality of impact, in settings where the outcomes of allocations are unknown but have a diminishing marginal effect. While focused on a two-group setting, our study addresses a broader class of welfare dynamics than explored in prior work. Our contributions are three-fold. First, we provide an algorithm designed for non-noisy continuous-resource environments that achieves sublinear fairness regret. Second, we propose a meta-algorithm for noisy settings building on the first contribution, and third, we empirically demonstrate that our approach consistently achieves fair, welfare-maximizing allocations.



Paperid:1070
Authors:Marguerite Petit-Talamon, Marc Lambert, Anna Korba
Abstract:
Variational inference (VI) is a popular approachin Bayesian inference, that looks for the best approximation of the posterior distribution within aparametric family, minimizing a loss that is typically the (reverse) Kullback-Leibler (KL) divergence. In this paper, we focus on the following parametric family: mixtures of isotropic Gaussians (i.e., with diagonal covariance matrices proportional to the identity) and uniform weights. We develop a variational framework and provide efficient algorithms suited for this family. In contrast with mixtures of Gaussian with generic covariance matrices, this choice presents a balance between accurate approximations of multimodal Bayesian posteriors, while being memory and computationally efficient. Our algorithms implement gradient descent on the location of the mixture components (the modes of the Gaussians), and either (an entropic) Mirror or Bures descent on their variance parameters. We illustrate the performance of our algorithms on numerical experiments.



Paperid:1071
Authors:Wei Shi, Yuan Cao
Abstract:
Transformers have proven highly effective across various applications, especially in handling sequential data such as natural languages and time series. However, transformer models often lack clear interpretability, and the success of transformers has not been well understood in theory. In this paper, we study the capability and interpretability of transformers in learning a family of classic statistical models, namely random walks on circles. We theoretically demonstrate that, after training with gradient descent, a one-layer transformer model can achieve optimal accuracy in predicting random walks. Importantly, our analysis reveals that the trained model is interpretable: the trained softmax attention serves as a token selector, focusing on the direct parent state; subsequently, the value matrix executes a one-step probability transition to predict the location of the next state based on this parent state. We also show that certain edge cases not covered by our theory are indeed failure cases, demonstrating that our theoretical conditions are tight. By investigating these success and failure cases, it is revealed that gradient descent with small initialization may fail or struggle to converge to a good solution in certain simple tasks even beyond random walks. Experiments are conducted to support our theoretical findings.



Paperid:1072
Authors:Daniel Jung, Kyoung Mu Lee
Abstract:
Hands are essential to human interaction, and understanding contact between hands and the world can promote comprehensive understanding of their function. Recently, there have been growing number of hand interaction datasets that cover interaction with object, other hand, scene, and body. Despite the significance of the task and increasing high-quality data, how to effectively learn dense hand contact estimation remains largely underexplored. There are two major challenges for learning dense hand contact estimation. First, there exists class imbalance issue from hand contact datasets where majority of samples are not in contact. Second, hand contact datasets contain spatial imbalance issue with most of hand contact exhibited in finger tips, resulting in challenges for generalization towards contacts in other hand regions. To tackle these issues, we present a framework that learns dense HAnd COntact estimation (HACO) from imbalanced data. To resolve the class imbalance issue, we introduce balanced contact sampling, which builds and samples from multiple sampling groups that fairly represent diverse contact statistics for both contact and non-contact samples. Moreover, to address the spatial imbalance issue, we propose vertex-level class-balanced (VCB) loss, which incorporates spatially varying contact distribution by separately reweighting loss contribution of each vertex based on its contact frequency across dataset. As a result, we effectively learn to predict dense hand contact estimation with large-scale hand contact data without suffering from class and spatial imbalance issue. The codes will be released.



Paperid:1073
Authors:Tianle Liu, Fan Lyu, Chenggong Ni, Zhang Zhang, Fuyuan Hu, Liang Wang
Abstract:
Test-Time Discovery (TTD) addresses the critical challenge of identifying and adapting to novel classes during inference while maintaining performance on known classes, which is a capability essential for dynamic real-world environments such as healthcare and autonomous driving. Recent TTD methods adopt training-free, memory-based strategies but rely on frozen models and static representations, resulting in poor generalization. In this paper, we propose a Discrepancy-Amplifying Adapter (DAA), a trainable module that enables real-time adaptation by amplifying feature-level discrepancies between known and unknown classes. During training, DAA is optimized using simulated unknowns and a novel warm-up strategy to enhance its discriminative capacity. To ensure continual adaptation at test time, we introduce a Short-Term Memory Renewal (STMR) mechanism, which maintains a queue-based memory for unknown classes and selectively refreshes prototypes using recent, reliable samples. DAA is further updated through self-supervised learning, promoting knowledge retention for known classes while improving discrimination of emerging categories. Extensive experiments show that our method maintains high adaptability and stability, and significantly improves novel class discovery performance. Our code will be available.



Paperid:1074
Authors:Tianyi Ma, Yiyue Qian, Zheyuan Zhang, Zehong Wang, Xiaoye Qian, Feifan Bai, Yifan Ding, Xuwei Luo, Shinan Zhang, Keerthiram Murugesan, Chuxu Zhang, Yanfang Ye
Abstract:
The exponential growth of data-driven systems and AI technologies has intensified the demand for high-quality web-sourced datasets. While existing datasets have proven valuable, conventional web data collection approaches face significant limitations in terms of human effort and scalability. Current data collecting solutions fall into two categories: wrapper-based methods that struggle with adaptability and reproducibility, and large language model (LLM)-based approaches that incur substantial computational and financial costs. To address these challenges, we propose AutoData, a novel multi-agent system for Automated web Data collection, that requires minimal human intervention, i.e., only necessitating a natural language instruction specifying the desired dataset. In addition, AutoData is designed for a robust multi-agent architecture, featuring a novel oriented message hypergraph coordinated by a central task manager, to efficiently organize agents across research and development squads. Besides, we introduce a novel hypergraph cache system to advance the multi-agent collaboration process that enables efficient automated data collection and mitigates the token cost issues prevalent in existing LLM-based systems. Moreover, we introduce Instruct2DS, a new benchmark dataset supporting live data collection from web sources across three domains: academic, finance, and sports. Comprehensive evaluations over Instruct2DS and three existing benchmark datasets demonstrate AutoData's superior performance compared to baseline methods.Case studies on challenging tasks such as picture book collection and paper extraction from surveys further validate its applicability.



Paperid:1075
Authors:Lan Pan, Hanbo Xie, Robert Wilson
Abstract:
Large Language Models (LLMs) have emerged with many intellectual capacities. While numerous benchmarks assess their intelligence, limited attention has been given to their ability to explore—an essential capacity for discovering new information and adapting to novel environments in both natural and artificial systems. The extent to which LLMs can effectively explore, particularly in open-ended tasks, remains unclear. This study investigates whether LLMs can surpass humans in exploration during an open-ended task, using Little Alchemy 2 as a paradigm, where agents combine elements to discover new ones. Results show most LLMs underperform compared to humans, except for the o1 model, with those traditional LLMs relying primarily on uncertainty-driven strategies, unlike humans who balance uncertainty and empowerment. Results indicate that traditional reasoning-focused LLMs, such as GPT-4o, exhibit a significantly faster and less detailed reasoning process, limiting their exploratory performance. In contrast, the DeepSeek reasoning model demonstrates prolonged, iterative thought processes marked by repetitive analysis of combinations and past trials, reflecting a more thorough and human-like exploration strategy. Representational analysis of the models with Sparse Autoencoders (SAE) revealed that uncertainty and choices are represented at earlier transformer blocks, while empowerment values are processed later, causing LLMs to think too fast and make premature decisions, hindering effective exploration. These findings shed light on the limitations of LLM exploration and suggest directions for improving their adaptability.



Paperid:1076
Authors:Letian Zhang, Bo Chen, Jieming Bian, Lei Wang, Jie Xu
Abstract:
Federated learning (FL) enables distributed devices to collaboratively train machine learning (ML) models while maintaining data privacy. However, the heterogeneous hardware capabilities of participating devices often result in significant training delays, as straggler clients with limited resources prolong the aggregation process. Existing solutions such as client selection, asynchronous FL, and partial training partially address these challenges but encounter issues such as reduced accuracy, stale updates, and compromised model performance due to inconsistent training contributions.To overcome these limitations, we propose FedEL, a federated elastic learning framework that enhances training efficiency while maintaining model accuracy. FedEL introduces a novel window-based training process, sliding the window to locate the training part of the modeland dynamically selecting important tensors for training within a coordinated runtime budget. This approach ensures progressive and balanced training across all clients, including stragglers. Additionally, FedEL employs a tensor importance adjustment module, harmonizing local and global tensor importance to mitigate biases caused by data heterogeneity. The experiment results shows that FedEL achieves up to 3.87× improvement in time-to-accuracy compared to baselines while maintaining or exceeding final test accuracy.



Paperid:1077
Authors:Boyu Gou, Zanming Huang, Yuting Ning, Yu Gu, Michael Lin, Botao Yu, Andrei Kopanev, Weijian Qi, Yiheng Shu, Jiaman Wu, Chan Hee Song, Bernal Jimenez Gutierrez, Yifei Li, Zeyi Liao, Hanane Moussa, TIANSHU ZHANG, Jian Xie, Tianci Xue, Shijie Chen, Boyuan Zheng, Kai Zhang, Zhaowei Cai, Viktor Rozgic, Morteza Ziyadi, Huan Sun, Yu Su
Abstract:
Agentic search such as Deep Research systems—where large language models autonomously browse the web, synthesize information, and return citation-backed answers—represents a major shift in how users interact with web-scale information. While promising greater efficiency and cognitive offloading, the growing complexity and open-endedness of agentic search has outpaced existing evaluation benchmarks and methodologies, which largely assume short horizons and static answers. In this paper, we introduce Mind2Web 2, a benchmark of 100 realistic, high-quality, long-horizon tasks that require real-time web interaction and extensive information synthesis. To address the challenge of evaluating time-varying, multi-source answers, we propose a novel Agent-as-a-Judge framework. Our method leverages task-specific, tree-structured rubrics and rubric-based judge agents to automatically assess both factual correctness and source attribution with a high agreement with humans. We conduct a comprehensive evaluation of 9 frontier agentic search systems and human performance, and a detailed error analysis to draw insights for future development. Together, Mind2Web 2 and our evaluation framework provide a rigorous foundation for developing and benchmarking the next generation of trustworthy, high-capability agentic search systems.



Paperid:1078
Authors:Yang Cao, Xiaoyu Li, Zhao Song, Xin Yang, Tianyi Zhou
Abstract:
Abstract:The famous theorem of Fritz John states that any convex body has a unique maximal volume inscribed ellipsoid, known as the John Ellipsoid. Computing the John Ellipsoid is a fundamental problem in convex optimization. In this paper, we focus on approximating the John Ellipsoid inscribed in a convex and centrally symmetric polytope defined by $ P := \{ x \in \mathbb{R}^d : -\mathbf{1}_n \leq A x \leq \mathbf{1}_n \},$ where $ A \in \mathbb{R}^{n \times d} $ is a rank-$d$ matrix and $ \mathbf{1}_n \in \mathbb{R}^n $ is the all-ones vector. We develop two efficient algorithms for approximating the John Ellipsoid. The first is a sketching-based algorithm that runs in nearly input-sparsity time $ \widetilde{O}(\mathrm{nnz}(A) + d^\omega) $, where $ \mathrm{nnz}(A) $ denotes the number of nonzero entries in the matrix $A$ and $ \omega \approx 2.37$ is the current matrix multiplication exponent. The second is a treewidth-based algorithm that runs in time $ \widetilde{O}(n \tau^2)$, where $\tau$ is the treewidth of the dual graph of the matrix $A$. Our algorithms significantly improve upon the state-of-the-art running time of $ \widetilde{O}(n d^2) $ achieved by [Cohen, Cousins, Lee, and Yang, COLT 2019].



Paperid:1079
Authors:Hyunjin Kim, Haebeom Jung, Jaesik Park
Abstract:
We propose an adaptive sampling framework for 3D Gaussian Splatting (3DGS) that leverages comprehensive multi-view photometric error signals within a unified Metropolis-Hastings approach. Traditional 3DGS methods heavily rely on heuristic-based density-control mechanisms (e.g., cloning, splitting, and pruning), which can lead to redundant computations or the premature removal of beneficial Gaussians. Our framework overcomes these limitations by reformulating densification and pruning as a probabilistic sampling process, dynamically inserting and relocating Gaussians based on aggregated multi-view errors and opacity scores. Guided by Bayesian acceptance tests derived from these error-based importance scores, our method substantially reduces reliance on heuristics, offers greater flexibility, and adaptively infers Gaussian distributions without requiring predefined scene complexity. Experiments on benchmark datasets, including Mip-NeRF360, Tanks and Temples, and Deep Blending, show that our approach reduces the number of Gaussians needed, enhancing computational efficiency while matching or modestly surpassing the view-synthesis quality of state-of-the-art models.



Paperid:1080
Authors:Chengyu Lu, Zhenhua Li, Xi Lin, Ji Cheng, Qingfu Zhang
Abstract:
Multi-objective optimization problems (MOPs), which involve optimizing multiple conflicting objectives simultaneously, are common in many real-world applications. Recently, Pareto Set Learning (PSL) has emerged as a powerful paradigm for solving MOPs, enabling the on-demand generation of tailored Pareto optimal solutions without requiring repeated optimization. However, applying PSL to black-box objectives, particularly those exhibiting non-separability, high dimensionality, or other complex properties, remains poorly understood. To bridge this gap, we propose leveraging evolution strategies (ESs), a class of specialized black-box optimization algorithms, within the PSL paradigm. Traditional ESs capture the complex dimensional dependencies of problems less efficiently, which can significantly hinder their performance in PSL. To tackle this issue, we suggest encapsulating the dependencies within a neural network, which is then trained using a novel gradient estimation method. The proposed method, termed Neural-ES, is evaluated using a bespoke benchmark suite for black-box PSL. Experimental comparisons with other methods demonstrate the efficiency of Neural-ES, underscoring its suitability for learning the Pareto sets of challenging black-box MOPs.



Paperid:1081
Authors:Bálint Mucsányi, Nathaël Da Costa, Philipp Hennig
Abstract:
In classification tasks, softmax functions are ubiquitously used as output activations to produce predictive probabilities. Such outputs only capture aleatoric uncertainty. To capture epistemic uncertainty, approximate Gaussian inference methods have been proposed. We develop a common formalism to describe such methods, which we view as outputting Gaussian distributions over the logit space. Predictives are then obtained as the expectations of the Gaussian distributions pushed forward through the softmax. However, such softmax Gaussian integrals cannot be solved analytically, and Monte Carlo (MC) approximations can be costly and noisy. We propose to replace the softmax activation by element-wise normCDF or sigmoid, which allows for the accurate sampling-free approximation of predictives. This also enables the approximation of the Gaussian pushforwards by Dirichlet distributions with moment matching. This approach entirely eliminates the runtime and memory overhead associated with MC sampling. We evaluate it combined with several approximate Gaussian inference methods (Laplace, HET, SNGP) on large- and small-scale datasets (ImageNet, CIFAR-100, CIFAR-10), demonstrating improved uncertainty quantification capabilities compared to softmax MC sampling.



Paperid:1082
Authors:Hrad Ghoukasian, Bonwoo Lee, Shahab Asoodeh
Abstract:
Abstract:We study the problem of sampling from a distribution under local differential privacy (LDP). Given a private distribution $P \in \mathcal{P}$, the goal is to generate a single sample from a distribution that remains close to $P$ in $f$-divergence while satisfying the constraints of LDP. This task captures the fundamental challenge of producing realistic-looking data under strong privacy guarantees. While prior work by Park et al. (NeurIPS'24) focuses on global minimax-optimality across a class of distributions, we take a local perspective. Specifically, we examine the minimax error in a neighborhood around a fixed distribution $P_0$, and characterize its exact value, which depends on both $P_0$ and the privacy level. Our main result shows that the local minimax error is determined by the global minimax error when the distribution class $\mathcal{P}$ is restricted to a neighborhood around $P_0$. To establish this, we (1) extend previous work from pure LDP to the more general functional LDP framework, and (2) prove that the globally optimal functional LDP sampler yields the optimal local sampler when constrained to distributions near $P_0$. Building on this, we also derive a simple closed-form expression for the locally minimax-optimal samplers which does not depend on the choice of $f$-divergence. We further argue that this local framework naturally models private sampling with public data, where the public data distribution is represented by $P_0$. In this setting, we empirically compare our locally optimal sampler to existing global methods, and demonstrate that it consistently outperforms global minimax samplers.



Paperid:1083
Authors:Helen Jin, Anton Xue, Weiqiu You, Surbhi Goel, Eric Wong
Abstract:
Stability guarantees have emerged as a principled way to evaluate feature attributions, but existing certification methods rely on heavily smoothed classifiers and often produce conservative guarantees.To address these limitations, we introduce soft stability and propose a simple, model-agnostic, sample-efficient stability certification algorithm (SCA) that yields non-trivial and interpretable guarantees for any attribution method.Moreover, we show that mild smoothing achieves a more favorable trade-off between accuracy and stability, avoiding the aggressive compromises made in prior certification methods.To explain this behavior, we use Boolean function analysis to derive a novel characterization of stability under smoothing.We evaluate SCA on vision and language tasks and demonstrate the effectiveness of soft stability in measuring the robustness of explanation methods.



Paperid:1084
Authors:Bruce Kuwahara, Chen-Yuan Lin, Xiao Shi Huang, Kin Kwan Leung, Jullian Yapeter, Ilya Stanevich, Felipe Perez, Jesse Cresswell
Abstract:
Automatic summarization systems have advanced rapidly with large language models (LLMs), yet they still lack reliable guarantees on inclusion of critical content in high-stakes domains like healthcare, law, and finance. In this work, we introduce Conformal Importance Summarization, the first framework for importance-preserving summary generation which uses conformal prediction to provide rigorous, distribution-free coverage guarantees. By calibrating thresholds on sentence-level importance scores, we enable extractive document summarization with user-specified coverage and recall rates over critical content. Our method is model-agnostic, requires only a small calibration set, and seamlessly integrates with existing black-box LLMs. Experiments on established summarization benchmarks demonstrate that Conformal Importance Summarization achieves the theoretically assured information coverage rate. Our work suggests that Conformal Importance Summarization can be combined with existing techniques to achieve reliable, controllable automatic summarization, paving the way for safer deployment of AI summarization tools in critical applications. Code is provided as supplementary material and will be released upon publication.



Paperid:1085
Authors:Devin Lange, Pengwei Sui, Shanghua Gao, Marinka Zitnik, Nils Gehlenborg
Abstract:
Biomedical research data portals are essential resources for scientific inquiry, and interactive exploratory visualizations are an integral component for querying such data repositories. Increasingly, machine learning is being integrated into visualization systems to create natural language interfaces where questions about data can be answered with visualizations, and follow-up questions can build on the previous state. This paper introduces a framework that takes abstract low-level questions about data and a visualization grammar specification that can answer such a question, reifies them with data entities and fields that meet certain constraints, and paraphrases the question language to produce the final collection of realized data-question-visualization triplets. Furthermore, we can link these foundational elements together to construct chains of queries, visualizations, and follow-up queries. We developed an open-source review interface for evaluating the results of these datasets. We applied this framework to five biomedical research data repositories, resulting in DQVis, a dataset of 1.08 million data-question-visualization triplets and 11.4 thousand two-step question samples. Five visualization experts provided feedback on the generated dataset through our review interface. We present a summary of their input and publish the full reviews as an additional resource alongside the dataset.The DQVis dataset and generation code are available at https://huggingface.co/datasets/HIDIVE/DQVis and https://github.com/hms-dbmi/DQVis-Generation.



Paperid:1086
Authors:Weilin Wan, Weizhong Zhang, Cheng Jin
Abstract:
Data selection improves computational efficiency by choosing informative subsets of training samples. However, existing methods ignore the compute budget, treating data selection and importance evaluation independently of compute budget constraints. Yet empirical studies show no algorithm can consistently outperform others (or even random selection) across varying budgets. We therefore argue that compute budget must be integral to data-selection strategies, since different budgets impose distinct requirements on data quantity, quality, and distribution for effective training. To this end, we propose a novel Computational budget-Aware Data Selection (CADS) method and naturally formulate it into a bilevel optimization framework, where the inner loop trains the model within the constraints of the computational budget on some selected subset of training data, while the outer loop optimizes data selection based on model evaluation. Our technical contributions liein addressing two main challenges in solving this bilevel optimization problem: the expensive Hessian matrix estimation for outer-loop gradients and the computational burden of achieving inner-loop optimality during iterations. To solve the first issue, we propose a probabilistic reparameterization strategy and compute the gradient using a Hessian-free policy gradient estimator. To address the second challenge, we transform the inner optimization problem into a penalty term in the outer objective, further discovering that we only need to estimate the minimum of a one-dimensional loss to calculate the gradient, significantly improving efficiency. To accommodate different data selection granularities, we present two complementary CADS variants: an example-level version (CADS-E) offering fine-grained control and a source-level version (CADS-S) aggregating samples into source groups for scalable, efficient selection without sacrificing effectiveness.Extensive experiments show that our method achieves performance gains of up to 14.42\% over baselines in vision and language benchmarks. Additionally, CADS achieves a 3-20× speedup compared to conventional bilevel implementations, with acceleration correlating positively with compute budget size.



Paperid:1087
Authors:Shihan Dou, Shichun Liu, Yuming Yang, Yicheng Zou, Yunhua Zhou, Shuhao Xing, Chenhao Huang, Qiming Ge, haijun Lv, Demin Song, Songyang Gao, Chengqi Lyu, Enyu Zhou, Honglin Guo, Zhiheng Xi, Qipeng Guo, Wenwei Zhang, Tao Gui, Qi Zhang, Xipeng Qiu, Xuanjing Huang, Kai Chen
Abstract:
We offer a novel perspective on reward modeling by formulating it as a policy discriminator, which quantifies the difference between two policies to generate a reward signal, guiding the training policy towards a target policy with desired behaviors. Leveraging this conceptual insight, we present a scalable pre-training method namedPOlicy DiffErentiationModeling (POEM), which trains a reward model (RM) to discern identical policies and discriminate different ones. Unlike traditional reward modeling methods relying on absolute preferences, POEM captures the relative difference between one policy and an arbitrary target policy, which provides a scalable, high-level optimization objective suitable for modeling generic ranking relationships. Empirical results show that POEM substantially outperforms traditional non-pre-trained methods, significantly enhancing RM performance. For instance, a 1.8B POEM RM could improve preference accuracy from 54.9% to 79.8% on STEM tasks and from 72.3% to 75.1% on general chat tasks compared to SOTA RMs. POEM also shows robust generalization capabilities in RLHF using Reinforcement Fine-tuning (RFT), effectively providing reliable reward signals and markedly enhancing policy performance—improving InternLM3-8B from an average of 56.5% to 62.6% and Qwen2.5-32B from 64.5% to 68.6% across 20 popular benchmarks. Moreover, scaling experiments reveal a clear power-law relationship between computation and performance, supported by linear correlation coefficients approaching 0.99. The impressive performance, strong generalization, and scaling properties of our models suggest that POEM is a promising direction for developing general and strong reward models.



Paperid:1088
Authors:Peng Xing, Haofan Wang, Yanpeng Sun, wangqixun, Baixu, Hao Ai, Jen-Yuan Huang, Zechao Li
Abstract:
The advancement of image style transfer has been fundamentally constrained by the absence of large-scale, high-quality datasets with explicit content-style-stylized supervision. Existing methods predominantly adopt training-free paradigms (e.g., image inversion), which limit controllability and generalization due to the lack of structured triplet data. To bridge this gap, we design a scalable and automated pipeline that constructs and purifies high-fidelity content-style-stylized image triplets. Leveraging this pipeline, we introduce IMAGStyle—the first large-scale dataset of its kind, containing 210K diverse and precisely aligned triplets for style transfer research. Empowered by IMAGStyle, we propose CSGO, a unified, end-to-end trainable framework that decouples content and style representations via independent feature injection. CSGO jointly supports image-driven style transfer, text-driven stylized generation, and text-editing-driven stylized synthesis within a single architecture. Extensive experiments show that CSGO achieves state-of-the-art controllability and fidelity, demonstrating the critical role of structured synthetic data in unlocking robust and generalizable style transfer.



Paperid:1089
Authors:Biao Liu, Ning Xu, Jie Wang, Xin Geng
Abstract:
Single-positive multi-label learning (SPMLL) is a weakly supervised multi-label learning problem, where each training example is annotated with only one positive label. Existing SPMLL methods typically assign pseudo-labels to unannotated labels with the assumption that prior probabilities of all classes are identical.However, the class-prior of each category may differ significantly in real-world scenarios, which makes the predictive model not perform as well as expected due to the unrealistic assumption on real-world application.To alleviate this issue, a novel framework named Crisp, i.e., Class-pRiors Induced Single-Positive multi-label learning, is proposed. Specifically, a class-priors estimator is introduced, which can estimate the class-priors that are theoretically guaranteed to converge to the ground-truth class-priors. In addition, based on the estimated class-priors, an unbiased risk estimator for classification is derived, and the corresponding risk minimizer can be guaranteed to approximately converge to the optimal risk minimizer on fully supervised data.Experimental results on ten MLL benchmark datasets demonstrate the effectiveness and superiority of our method over existing SPMLL approaches.



Paperid:1090
Authors:Valentin Kilian, Stefano Cortinovis, Francois Caron
Abstract:
Given a large pool of unlabelled data and a smaller amount of labels, prediction-powered inference (PPI) leverages machine learning predictions to increase the statistical efficiency of standard confidence interval procedures based solely on labelled data, while preserving their fixed-time validity. In this paper, we extend the PPI framework to the sequential setting, where labelled and unlabelled datasets grow over time. Exploiting Ville's inequality and the method of mixtures, we propose prediction-powered confidence sequence procedures that are valid uniformly over time and naturally accommodate prior knowledge on the quality of the predictions to further boost efficiency. We carefully illustrate the design choices behind our method and demonstrate its effectiveness in real and synthetic examples.



Paperid:1091
Authors:Chau Pham, Quan Dao, Mahesh Bhosale, Yunjie Tian, Dimitris Metaxas, DAVID DOERMANN
Abstract:
Recent advances in diffusion models have revolutionized text-guided image editing, yet existing editing methods face critical challenges in hyperparameter identification. To get the reasonable editing performance, these methods often require the user to brute-force tune multiple interdependent hyperparameters, such as inversion timesteps and attention modification, \textit{etc.} This process incurs high computational costs due to the huge hyperparameter search space. We consider searching optimal editing's hyperparameters as a sequential decision-making task within the diffusion denoising process. Specifically, we propose a reinforcement learning framework, which establishes a Markov Decision Process that dynamically adjusts hyperparameters across denoising steps, integrating editing objectives into a reward function. The method achieves time efficiency through proximal policy optimization while maintaining optimal hyperparameter configurations. Experiments demonstrate significant reduction in search time and computational overhead compared to existing brute-force approaches, advancing the practical deployment of a diffusion-based image editing framework in the real world.



Paperid:1092
Authors:Florentin Guth, Zahra Kadkhodaie, Eero Simoncelli
Abstract:
Learning probability models from data is at the heart of many machine learning endeavors, but is notoriously difficult due to the curse of dimensionality. We introduce a new framework for learning \emph{normalized energy} (log probability) models that is inspired from diffusion generative models, which rely on networks optimized to estimate the score. We modify a score network architecture to compute energy while preserving its inductive biases for denoising. The gradient of this energy network with respect to an input image is the score of the learned density which can be optimized using a denoising objective. Importantly, the gradient with respect to the noise level provides an additional score that can be optimized with a novel secondary objective, ensuring consistent and normalized energies across noise levels. We train a network with this ``dual score matching'' objective on ImageNet64 dataset, and show that the resulting trained network reaches a cross-entropy (negative log likelihood) of 3.4 bits/pixel, comparably to the state of the art. We further validate our approach by showing that estimated log probabilities are nearly independent of the specific images in the training set. Finally, we demonstrate that both image probability and dimensionality of local neighborhoods vary significantly with image content, in contrast with traditional assumptions such as concentration of measure or support on a low-dimensional manifold.



Paperid:1093
Authors:Ilan Cohen, Debmalya Panigrahi
Abstract:
Abstract:In online allocation problems, an algorithm must choose from a set of options at each step, where each option incurs a set of costs/rewards associated with a set of $d$ agents. The goal is to minimize/maximize a function of the accumulated costs/rewards assigned to the agents over the course of the entire allocation process. Such problems are common in combinatorial optimization, including minimization problems such as machine scheduling and network routing, as well as maximization problems such as fair allocation for welfare maximization.In this paper, we develop a general learning-augmented algorithmic framework for online allocation problems that produces a nearly optimal solution using only a single $d$-dimensional vector of learned weights. Using this general framework, we derive learning-augmented online algorithms for a broad range of application problems in routing, scheduling, and fair allocation. Our main tool is convex programming duality, which may also have further implications for learning-augmented algorithms in the future.



Paperid:1094
Authors:Ben Gao, Jordan Patracone, Olivier Alata, Stephane Chretien
Abstract:
We introduce Conformal Online Learning (COL), a novel framework for adaptively updating Koopman-invariant representations of nonlinear dynamical systems from streaming data. Our modeling approach combines deep feature learning with multi-step prediction consistency in the lifted space, where the dynamics evolve linearly. To prevent overfitting, COL employs a conformal-style mechanism that shifts the focus from evaluating the conformity of new states to assessing the consistency of the current Koopman model. Updates are triggered only when the current model’s prediction error exceeds a calibrated threshold, allowing selective refinement of the Koopman operator and embedding. Empirical results on benchmark dynamical systems demonstrate the effectiveness of COL in maintaining long-term predictive accuracy while significantly reducing unnecessary updates.



Paperid:1095
Authors:Florian Kalinke, Shakeel Gavioli-Akilagun
Abstract:
This article studies the problem of online non-parametric change point detection in multivariate data streams. We approach the problem through the lens of kernel-based two-sample testing and introduce a sequential testing procedure based on random Fourier features, running with logarithmic time complexity per observation and with overall logarithmic space complexity. The algorithm has two advantages compared to the state of the art. First, our approach is genuinely online, and no access to training data known to be from the pre-change distribution is necessary. Second, the algorithm does not require the user to specify a window parameter over which local tests are to be calculated. We prove strong theoretical guarantees on the algorithm's performance, including information-theoretic bounds demonstrating that the detection delay is optimal in the minimax sense. Numerical studies on real and synthetic data show that our algorithm is competitive with respect to the state of the art.



Paperid:1096
Authors:Jean-Philippe Monteuuis, Cong Chen, Jonathan Petit
Abstract:
``Caught in a landslide, no escape from reality" summarizes the state of the research in AI offense: an attack might work on paper but does not necessarily in practice. In the last 5 years, we have seen the rise of latency attacks against computer vision systems. Most of them targeted 2D object detection, especially its Non-Max-Suppression (NMS) block, via adversarial images. However, we uncovered that when tested in realistic deployment settings, the NMS latency attacks, accepted to top conferences, are harmless. In this paper, we define an evaluation framework (EVADE) to assess the practicality of attacks, and apply it to state-of-the-art NMS latency attacks. Attacks were tested on different hardware platforms, and different model formats and quantization. Results show that these attacks are not able to generate the claimed latency increase, nor transfer to other models (from the same family or not). Moreover, the latency increase falls well within the latency requirement of downstream tasks, hence, being harmless. We also tested three defenses, which were successful in mitigating the NMS latency attacks. Therefore, in their current form, NMS latency attacks are just fantasy.



Paperid:1097
Authors:Yijun Yang, Zeyu Huang, Wenhao Zhu, Zihan Qiu, Fei Yuan, Jeff Pan, Ivan Titov
Abstract:
Abstract:Existing frameworks for evaluating long-context language models (LCLM) can be broadly categorized into real-world and synthetic tasks.Despite their utility, both approaches are accompanied by certain intrinsic limitations.For example, real-world tasks are too complex to interpret or characterize and are susceptible to data contamination. In contrast, synthetic tasks often adopt the needle-in-the-haystack (NIAH) format, wherein a lack of coherence between the “needle” and the “haystack” compromises their validity as proxies for realistic applications. In response to these challenges, we posit that an ideal long-context evaluation framework should be characterized by three essential features: $\textit{seamless context}$, $\textit{controllable setting}$, and $\textit{sound evaluation}$.This study introduces $\textbf{LongBioBench}$, a novel benchmark that utilizes artificially generated biographies as a controlled environment for assessing LCLMs across dimensions of $\textit{understanding}$, $\textit{reasoning}$, and $\textit{trustworthiness}$. Our experimental evaluation, which includes $\textbf{18}$ LCLMs in total, demonstrates that most models still exhibit deficiencies in semantic understanding and elementary reasoning over retrieved results and are less trustworthy as context length increases.Our further analysis indicates some design choices employed by existing synthetic benchmarks, such as contextual non-coherence, numerical needles, and the absence of distractors, rendering them vulnerable to test the model's long-context capabilities.Moreover, we also reveal that long-context continual pretraining primarily adjusts RoPE embedding to accommodate extended context lengths, which in turn yields only marginal improvements in the model’s true capabilities. To sum up, compared to previous synthetic benchmarks, LongBioBench achieves a better trade-off between mirroring authentic language tasks and maintaining controllability, and is highly interpretable and configurable.



Paperid:1098
Authors:Andre Barreto, Vincent Dumoulin, Yiran Mao, Mark Rowland, Nicolas Perez-Nieves, Bobak Shahriari, Yann Dauphin, Doina Precup, Hugo Larochelle
Abstract:
Reinforcement learning from human feedback usually models preferences using a reward model that does not distinguish between people. We argue that this is unlikely to be a good design choice in contexts with high potential for disagreement, like in the training of large language models. We formalise and analyse the problem of learning a reward model that can be specialised to a user. Using the principle of empirical risk minimisation, we derive a probably approximately correct bound showing the dependency of the approximation error not only on the number of training examples, but also on the number of human raters who provided feedback on them. We also put forward a formal argument supporting the intuition that adaptive reward models should be beneficial when there is considerable disagreement among users. Building on our theoretical findings, we propose a concrete architecture for an adaptive reward model. Our approach leverages the observation that individual preferences can be captured as a linear combination of a set of general reward features. We show how to learn such features and subsequently use them to quickly adapt the reward model to a specific individual, even if their preferences are not reflected in the training data. We present experiments with large language models illustrating our theoretical results and comparing the proposed architecture with a non-adaptive baseline. As predicted by the theory, the benefits provided by our model increase with the number of raters and the heterogeneity of their preferences. We also show how our model compare favourably to adaptive counterparts, including models that do in-context personalisation.



Paperid:1099
Authors:Yuheng Yuan, Qiuhong Shen, Xingyi Yang, Xinchao Wang
Abstract:
Abstract:4D Gaussian Splatting (4DGS) has recently gained considerable attention as a method for reconstructing dynamic scenes. Despite achieving superior quality, 4DGS typically requires substantial storage and suffers from slow rendering speed. In this work, we delve into these issues and identify two key sources of temporal redundancy. (Q1) \textbf{Short-Lifespan Gaussians}: 4DGS uses a large portion of Gaussians with short temporal span to represent scene dynamics, leading to an excessive number of Gaussians. (Q2) \textbf{Inactive Gaussians}: When rendering, only a small subset of Gaussians contributes to each frame. Despite this, all Gaussians are processed during rasterization, resulting in redundant computation overhead. To address these redundancies, we present \textbf{4DGS-1K}, which runs at over 1000 FPS on modern GPUs. For Q1, we introduce the Spatial-Temporal Variation Score, a new pruning criterion that effectively removes short-lifespan Gaussians while encouraging 4DGS to capture scene dynamics using Gaussians with longer temporal spans. For Q2, we store a mask for active Gaussians across consecutive frames, significantly reducing redundant computations in rendering. Compared to vanilla 4DGS, our method achieves a $41\times$ reduction in storage and $9\times$ faster rasterization speed on complex dynamic scenes, while maintaining comparable visual quality.



Paperid:1100
Authors:Adrien Lafage, Olivier Laurent, Firas Gabetni, Gianni Franchi
Abstract:
Deep Neural Networks (DNNs) have demonstrated remarkable performance across various domains, including computer vision and natural language processing. However, they often struggle to accurately quantify their predictions' uncertainty, limiting their broader adoption in critical industrial applications. Uncertainty Quantification (UQ) for Deep Learning seeks to address this challenge by providing methodologies to improve the reliability of uncertainty estimates. While numerous techniques have been proposed, a unified tool remains lacking that offers a seamless workflow for evaluating and integrating these methods. To bridge this gap, we introduceTorch-Uncertainty, aPyTorchandLightningframework designed to streamline the training and evaluation of DNNs with UQ techniques. In this paper, we outline the foundational principles of our library and present comprehensive experimental results that benchmark a diverse set of UQ methods across classification, segmentation, and regression tasks. Our library is available at: https://github.com/ENSTA-U2IS-AI/torch-uncertainty



Paperid:1101
Authors:Yulin Zhang, Cheng Shi, Yang Wang, Sibei Yang
Abstract:
Envision an AI capable of functioning in human-like settings, moving beyond mere observation to actively understand, anticipate, and proactively respond to unfolding events. Towards this vision, we focus on the innovative task where, given ego-streaming video input, an assistant proactively answers diverse, evolving questions at the opportune moment, while maintaining synchronized perception and reasoning. This task embodies three key properties: (1) Proactive Coherence, (2) Just-in-Time Responsiveness, and (3) Synchronized Efficiency.To evaluate and address these properties, we first introduce ESTP-Bench (Ego Streaming Proactive Benchmark) alongside the ESTP-F1 metric—a novel framework designed for their rigorous assessment. Secondly, we propose a comprehensive technical pipeline to enable models to tackle this challenging task. This pipeline comprises: (1) a data engine, (2) a multi-stage training strategy, and (3) a proactive dynamic compression technique. Our proposed model effectively addresses these critical properties while achieving state-of-the-art (SOTA) performance on the standard COIN benchmark.



Paperid:1102
Authors:Amon Lahr, Johannes Köhler, Anna Scampicchio, Melanie Zeilinger
Abstract:
Non-conservative uncertainty bounds are key for both assessing an estimation algorithm’s accuracy and in view of downstream tasks, such as its deployment in safety-critical contexts. In this paper, we derive a tight, non-asymptotic uncertainty bound for kernel-based estimation, which can also handle correlated noise sequences. Its computation relies on a mild norm-boundedness assumption on the unknown function and the noise, returning the worst-case function realization within the hypothesis class at an arbitrary query input location. The value of this function is shown to be given in terms of the posterior mean and covariance of a Gaussian process for an optimal choice of the measurement noise covariance. By rigorously analyzing the proposed approach and comparing it with other results in the literature, we show its effectiveness in returning tight and easy-to-compute bounds for kernel-based estimates.



Paperid:1103
Authors:Catherine Arnett, Tyler Chang, Stella Biderman, Benjamin Bergen
Abstract:
The number of tokens it takes to encode parallel text in different languages is known to vary. These disparities are calledtoken premiums.Having high token premiums leads to less throughput during training and increases costs at inference. In this paper, we find that even after controlling for dataset size, vocabulary size, and data content, monolingual tokenizers exhibit a wide range of token premiums across languages. To understand the cross-linguistic differences that affect these token premiums, we train a suite of almost 7,000 comparable monolingual tokenizers for 98 languages, manipulating tokenizer type (BPE, Unigram, and SuperBPE), vocabulary size, and dataset size. We measure token premiums and test for a relationship between factors such as data similarity (between tokenizer training and evaluation), vocabulary size, and pre-tokenization. We also investigate the role of language-specific features such as writing system and word length. We find that similarity between training and test data does not impact token premiums, but vocabulary size and pre-tokenization do. While simply increasing vocabulary size does not lead to reduced token premium effects, we can determine an "optimal'' vocabulary size for each language to achieve significantly reduced token premium effects. We also train superword tokenizers which allow merges over whitespaces, and we find that they both reduce token premium effects and improve compression overall. Thus, intervening on either vocabulary size and the pre-tokenizer significantly reduces crosslingual token premium effects.



Paperid:1104
Authors:Tom Sander, Bargav Jayaraman, Mark Ibrahim, Kamalika Chaudhuri, Chuan Guo
Abstract:
Conventional wisdom in machine learning privacy research states that memorization directly implies a loss of privacy. In contrast, a well generalized model only remembers distributional patterns and preserves privacy of its training data. In this work, we show that this relationship is much more complex for LLMs trained for chat, and depends heavily on how knowledge is encoded and manipulated. To this end, we fine tune language models on synthetically generated biographical information including PIIs, and try to extract them in different ways after instruction fine-tuning. We find counter to conventional wisdom that better verbatim memorization does not necessarily increase data leakage via chat. We also find that it is easier to extract information via chat from an LLM that is better able to manipulate and process knowledge even if it is smaller, and that not all attributes are equally extractable. Our results suggest that when it comes to the LLMs of today, the relationship between privacy, memorization and language understanding may be more nuanced than previously thought.



Paperid:1105
Authors:Xiaoquan Yi, Haozhao Wang, Yining Qi, Wenchao Xu, Rui Zhang, Yuhua Li, Ruixuan Li
Abstract:
With the increasing sophistication of Large Language Models (LLMs), it is crucial to develop reliable methods to accurately identify whether an interlocutor in real-time dialogue is human or chatbot. However, existing detection methods are primarily designed for analyzing full documents, not the unique dynamics and characteristics of dialogue. These approaches frequently overlook the nuances of interaction that are essential in conversational contexts. This work identifies two key patterns in dialogues: (1) Human-Human (H-H) interactions exhibit significant bidirectional sentiment influence, while (2) Human-Chatbot (H-C) interactions display a clear asymmetric pattern. We propose an innovative approach named ChatbotID, whichapplies the Granger Causality Test (GCT) to extract a novel set of interactional features that capture the evolving, predictive relationships between conversational attributes. By synergistically fusing these GCT-based interactional features with contextual embeddings, and optimizing the model through a meticulous loss function. Experimental results across multiple datasets and detection models demonstrate the effectiveness of our framework, with significant improvements in accuracy for distinguishing between H-H and H-C dialogues. The dataset and code are in the link https://anonymous.4open.science/r/Distinguishing-LLMs-by-Analyzing-Dialogue-Dynamics-with-Granger-Causality-56E4/.



Paperid:1106
Authors:Qiuhong Shen, Xingyi Yang, Xinchao Wang
Abstract:
Dynamic 3D scene reconstruction from multi-view videos demands representation to model complex deformations at scale. Current Gaussian Splatting based methods often either suffer from significant computation cost due to dense MLP-based modeling or explicit modeling deformation of each Gaussian independently. However, the dynamics of objects within a scene are typically hierarchical and exhibit structural correlations. To leverage these structural priors into the representation, we introduceTreeSplat, aTreedata structure for deformable GaussianSplatting. In TreeSplat, as the name suggests, motions of Gaussian are represented hierarchically within a tree. Each node learns coefficients for time-varying basis functions, defining a part of the motion. The full motion for any given Gaussian is then determined by accumulating these transformations along the tree path from its leaf node to the root node. This tree isn't predefined; instead, it is constructed adaptively alongside Gaussian densification, where cloning or splitting a Gaussian correspondingly creates new leaf nodes. One central property of TreeSplat is its mergeability; after optimization during training, the hierarchical motion parameters for each Gaussian can be efficiently consolidated. By performing this merging step before test time, we eliminate the need to traverse the tree explicitly for each Gaussian during rendering. This results in dramatically faster rendering over 200 FPS and compact storage, while maintaining state-of-the-art rendering quality. Experiments on diverse synthetic and real-world datasets validate these advantages.



Paperid:1107
Authors:Changze Lv, Yansen Wang, Dongqi Han, Yifei Shen, Xiaoqing Zheng, Xuanjing Huang, Dongsheng Li
Abstract:
Spiking neural networks (SNNs) are bio-inspired networks that mimic how neurons in the brain communicate through discrete spikes, which have great potential in various tasks due to their energy efficiency and temporal processing capabilities.SNNs with self-attention mechanisms (spiking Transformers) have recently shown great advancements in various tasks, and inspired by traditional Transformers, several studies have demonstrated that spiking absolute positional encoding can help capture sequential relationships for input data, enhancing the capabilities of spiking Transformers for tasks such as sequential modeling and image classification. However, how to incorporate relative positional information into SNNs remains a challenge.In this paper, we introduce several strategies to approximate relative positional encoding (RPE) in spiking Transformers while preserving the binary nature of spikes.Firstly, we formally prove that encoding relative distances with Gray Code ensures that the binary representations of positional indices maintain a constant Hamming distance whenever their decimal values differ by a power of two, and we proposeGray-PEbased on this property.In addition, we propose another RPE method calledLog-PE, which combines the logarithmic form of the relative distance matrix directly into the spiking attention map.Furthermore, we extend our RPE methods to a two-dimensional form, making them suitable for processing image patches.We evaluate our RPE methods on various tasks, including time series forecasting, text classification, and patch-based image classification, and the experimental results demonstrate a satisfying performance gain by incorporating our RPE methods across many architectures.Our results provide fresh perspectives on designing spiking Transformers to advance their sequential modeling capability, thereby expanding their applicability across various domains.



Paperid:1108
Authors:Junshu Sun, Wanxing Chang, Chenxue Yang, Qingming Huang, Shuhui Wang
Abstract:
Graph attention has demonstrated superior performance in graph learning tasks. However, learning from global interactions can be challenging due to the large number of nodes. In this paper, we discover a new phenomenon termed over-aggregating. Over-aggregating arises when a large volume of messages is aggregated into a single node with less discrimination, leading to the dilution of the key messages and potential information loss. To address this, we propose Wideformer, a plug-and-play method for graph attention. Wideformer divides the aggregation of all nodes into parallel processes and guides the model to focus on specific subsets of these processes. The division can limit the input volume per aggregation, avoiding message dilution and reducing information loss. The guiding step sorts and weights the aggregation outputs, prioritizing the informative messages. Evaluations show that Wideformer can effectively mitigate over-aggregating. As a result, the backbone methods can focus on the informative messages, achieving superior performance compared to baseline methods.



Paperid:1109
Authors:Ruben Solozabal, Velibor Bojkovic, Hilal AlQuabeh, Kentaro Inui, Martin Takac
Abstract:
Current methods for initializing state space models (SSMs) parameters mainly rely on the \textit{HiPPO framework}, which is based on an online approximation of orthogonal polynomials. Recently, diagonal alternatives have shown to reach a similar level of performance while being significantly more efficient due to the simplification in the kernel computation. However, the \textit{HiPPO framework} does not explicitly study the role of its diagonal variants. In this paper, we take a further step to investigate the role of diagonal SSM initialization schemes from the frequency perspective. Our work seeks to systematically understand how to parameterize these models and uncover the learning biases inherent in such diagonal state-space models. Based on our observations, we propose a diagonal initialization on the discrete Fourier domain \textit{S4D-DFouT}. The insights in the role of pole placing in the initialization enable us to further scale them and achieve state-of-the-art results on the Long Range Arena benchmark, allowing us to train from scratch on very large datasets as PathX-256.



Paperid:1110
Authors:Haowei Sun, Jinwu Hu, Zhirui Zhang, Haoyuan Tian, Xinze Xie, Yufeng Wang, Xiaohua Xie, Yun Lin, Zhuliang Yu, Mingkui Tan
Abstract:
Drone Visual Active Tracking aims to autonomously follow a target object by controlling the motion system based on visual observations, providing a more practical solution for effective tracking in dynamic environments. However, accurate Drone Visual Active Tracking using reinforcement learning remains challenging due to the absence of a unified benchmark and the complexity of open-world environments with frequent interference. To address these issues, we pioneer a systematic solution. First, we propose DAT, the first open-world drone active air-to-ground tracking benchmark. It encompasses 24 city-scale scenes, featuring targets with human-like behaviors and high-fidelity dynamics simulation. DAT also provides a digital twin tool for unlimited scene generation. Additionally, we propose a novel reinforcement learning method called GC-VAT, which aims to improve the performance of drone tracking targets in complex scenarios. Specifically, we design a Goal-Centered Reward to provide precise feedback across viewpoints to the agent, enabling it to expand perception and movement range through unrestricted perspectives. Inspired by curriculum learning, we introduce a Curriculum-Based Training strategy that progressively enhances the tracking performance in complex environments. Besides, experiments on simulator and real-world images demonstrate the superior performance of GC-VAT, achieving an approximately 400% improvement over the SOTA methods in terms of the cumulative reward metric.



Paperid:1111
Authors:Zeqi Xiao, Yushi LAN, Yifan Zhou, Wenqi Ouyang, Shuai Yang, Yanhong Zeng, Xingang Pan
Abstract:
World simulation has gained increasing popularity due to its ability to model virtual environments and predict the consequences of actions. However, the limited temporal context window often leads to failures in maintaining long-term consistency, particularly in preserving 3D spatial consistency. In this work, we present WorldMem, a framework that enhances scene generation with a memory bank consisting of memory units that store memory frames and states (e.g., poses and timestamps). By employing state-aware memory attention that effectively extracts relevant information from these memory frames based on their states, our method is capable of accurately reconstructing previously observed scenes, even under significant viewpoint or temporal gaps. Furthermore, by incorporating timestamps into the states, our framework not only models a static world but also captures its dynamic evolution over time, enabling both perception and interaction within the simulated world. Extensive experiments in both virtual and real scenarios validate the effectiveness of our approach.



Paperid:1112
Authors:Shigui Li, Wei Chen, Delu Zeng
Abstract:
Diffusion models (DMs) excel in image generation, but suffer from slow inference and training-inference discrepancies. Although gradient-based solvers like DPM-Solver accelerate sampling inference of DMs, they lack theoretical foundations in information transmission efficiency. This paper introduces an information-theoretic perspective on the diffusion inference processes, revealing that successful denoising fundamentally reduces conditional entropy during reverse transitions. This principle leads to two key insights: (1) data prediction parameterization outperforms its noise counterpart in entropy reduction, and (2) conditional variance optimization that minimizes transition and reconstruction errors through conditional entropy reduction. Building on these insights, we propose entropy-aware variance optimization for diffusion inference, calledEvoDiff, that systematically reduces uncertainty during denoising by optimizing conditional entropy reduction. Extensive experiments on DMs validate our insights and demonstrate that the proposed method consistently outperforms state-of-the-art gradient-based solvers. For example, compared to the baseline method, EvoDiff reduces reconstruction error by up to 45.5\% (FID from 5.10 to 2.78) at 10 function evaluations (NFE) on CIFAR-10, cuts function evaluation cost by 25\% (from 20 to 15 NFE) for high-quality samples on ImageNet-256, and improves text-to-image generation while reducing artifacts.



Paperid:1113
Authors:Yu Yao, Yiliao (Lia) Song, Yian Xie, Mengdan Fan, Mingyu Guo, Tongliang Liu
Abstract:
LLM-agent systems often decompose a high-level task objective into a subtask-dependency graph, assuming each subtask’s response is conditionally independent of others given its parent responses. However, we find the inaccessible ground-truth responses will violate this assumption during execution, leading to inter-agent misalignment: failures arise from breakdowns in inter-agent interaction and coordination during execution. Consequently, both quality and runtime efficiency degenerate significantly. Motivated by this finding, we propose SeqCV, a dynamic framework that enables reliable execution under violated conditional independence assumptions. In SeqCV, subtasks are executed sequentially, each conditioned on all prior responses and verified via consistency checks immediately after agents generate a short token sequence. At each checkpoint, the token sequence is considered reliable if it is common knowledge consistently supported across diverse models. An unreliable token sequence is discarded, triggering a recursive splitting mechanism to decompose the subtask into more manageable components. Despite the sequential nature, SeqCV avoids costly misalignment corrections and delivers higher effective throughput than parallel pipelines. On different tasks, SeqCV not only improves accuracy by up to 17%, but also reduces execution time by more than half over six commonly used benchmarking datasets.



Paperid:1114
Authors:Liu, Feng Gao, Bingwen Wei, Xinlei Chen, Qingmin Liao, YI WU, Chao Yu, Yu Wang
Abstract:
Large Vision-Language Action (VLA) models have shown significant potential for embodied AI. However, their predominant training via supervised fine-tuning (SFT) limits generalization due to susceptibility to compounding errors under distribution shifts. Reinforcement learning (RL) offers a path to overcome these limitations by optimizing for task objectives via trial-and-error, yet a systematic understanding of its specific generalization benefits for VLAs compared to SFT is lacking. To address this, our study introduces a comprehensive benchmark for evaluating VLA generalization and systematically investigates the impact of RL fine-tuning across diverse visual, semantic, and execution dimensions. Our extensive experiments reveal that RL fine-tuning, particularly with PPO, significantly enhances generalization in semantic understanding and execution robustness over SFT, while maintaining comparable visual robustness. We identify PPO as a more effective RL algorithm for VLAs than LLM-derived methods like DPO and GRPO. We also develop a simple recipe for efficient PPO training on VLAs, and demonstrate its practical utility for improving VLA generalization. The project page is at https://sites.google.com/view/vla-rl-study-anonymous



Paperid:1115
Authors:Jiecheng Lu, Xu Han, Yan Sun, Viresh Pati, Yubin Kim, Siddhartha Somani, Shihao Yang
Abstract:
Abstract:Linear attention methods offer Transformers $O(N)$ complexity but typically underperform standard softmax attention. We identify two fundamental limitations affecting these approaches: the restriction to convex combinations that only permits additive information blending, and uniform accumulated weight bias that dilutes attention in long contexts. We propose Zero-Sum Linear Attention (ZeroS), which addresses these limitations by removing the constant zero-order term $1/t$ and reweighting the remaining zero-sum softmax residuals. This modification creates mathematically stable weights, enabling both positive and negative values and allowing a single attention layer to perform contrastive operations. While maintaining $O(N)$ complexity, ZeroS theoretically expands the set of representable functions compared to convex combinations. Empirically, it matches or exceeds standard softmax attention across various sequence modeling benchmarks.



Paperid:1116
Authors:Erwan escudie, Matthia Sabatelli, Olivier Buffet, Jilles Dibangoye
Abstract:
We present a novel framework for (\varepsilon)-optimally solving two-player zero-sum partially observable stochastic games (zs-POSGs). These games pose a major challenge due to the absence of a principled connection with dynamic programming (DP) techniques developed for two-player zero-sum stochastic games (zs-SGs). Prior attempts at transferring solution methods have lacked a lossless reduction—defined here as a transformation that preserves value functions, equilibrium strategies, and optimality structure—thereby limiting generalisation to ad hoc algorithms. This work introduces the first lossless reduction from zs-POSGs to transition-independent zs-SGs, enabling the principled application of a broad class of DP-based methods. We show empirically that point-based value iteration (PBVI) algorithms, applied via this reduction, produce (\varepsilon)-optimal strategies across a range of benchmark domains, consistently matching or outperforming existing state-of-the-art methods. Our results open a systematic pathway for algorithmic and theoretical transfer from SGs to partially observable settings.



Paperid:1117
Authors:Liming Wang, Muhammad Jehanzeb Mirza, Yishu Gong, Yuan Gong, Jiaqi Zhang, Brian Tracey, Katerina Placek, Marco Vilela, Jim Glass
Abstract:
This paper presents a novel theoretical framework for understanding how diffusion models can learn disentangled representations with commonly used weak supervision such as partial labels and multiple views. Within this framework, we establish identifiability conditions for diffusion models to disentangle latent variable models with \emph{stochastic}, \emph{non-invertible} mixing processes. We also prove \emph{finite-sample global convergence} for diffusion models to disentangle independent subspace models. To validate our theory, we conduct extensive disentanglement experiments on subspace recovery in latent subspace Gaussian mixture models, image colorization, denoising, and voice conversion for speech classification. Our experiments show that training strategies inspired by our theory, such as style guidance regularization, consistently enhance disentanglement performance.



Paperid:1118
Authors:Taeseong Yoon, Heeyoung Kim
Abstract:
Abstract:Uncertainty quantification (UQ) is crucial for deploying machine learning models in high-stakes applications, where overconfident predictions can lead to serious consequences. An effective UQ method must balance computational efficiency with the ability to generalize across diverse scenarios. Evidential deep learning (EDL) achieves efficiency by modeling uncertainty through the prediction of a Dirichlet distribution over class probabilities. However, the restrictive assumption of Dirichlet-distributed class probabilities limits EDL's robustness, particularly in complex or unforeseen situations. To address this, we propose *flexible evidential deep learning* ($\mathcal{F}$-EDL), which extends EDL by predicting a flexible Dirichlet distribution—a generalization of the Dirichlet distribution—over class probabilities. This approach provides a more expressive and adaptive representation of uncertainty, significantly enhancing UQ generalization and reliability under challenging scenarios. We theoretically establish several advantages of $\mathcal{F}$-EDL and empirically demonstrate its state-of-the-art UQ performance across diverse evaluation settings, including classical, long-tailed, and noisy in-distribution scenarios.



Paperid:1119
Authors:Brown Ebouky, Andrea Bartezzaghi, Mattia Rigotti
Abstract:
The recent advent of reasoning models like OpenAI's o1 was met with excited speculation by the AI community about the mechanisms underlying these capabilities in closed models, followed by a rush of replication efforts, particularly from the open source community. These speculations were largely settled by the demonstration from DeepSeek-R1 that chains-of-thought and reinforcement learning (RL) can effectively replicate reasoning on top of base LLMs. However, it remains valuable to explore alternative methods for theoretically eliciting reasoning that could help elucidate the underlying mechanisms, as well as providing additional methods that may offer complementary benefits. Here, we build on the long-standing literature in cognitive psychology and cognitive architectures, which postulates that reasoning arises from the orchestrated, sequential execution of a set of modular, predetermined cognitive operations.Crucially, we implement this key idea within a modern agentic tool-calling framework. In particular, we endow an LLM with a small set of "cognitive tools" encapsulating specific reasoning operations, each executed by the LLM itself. Surprisingly, this simple strategy results in considerable gains in performance on standard mathematical reasoning benchmarks compared to base LLMs, for both closed and open-weight models. For instance, providing our ``cognitive tools'' to GPT-4.1 increases its pass@1 performance on AIME2024 from 26.7\% to 43.3\%, bringing it very close to the performance of o1-preview.In addition to its practical implications, this demonstration contributes to the debate regarding the role of post-training methods in eliciting reasoning in LLMs versus the role of inherent capabilities acquired during pre-training, and whether post-training merely uncovers these latent abilities.



Paperid:1120
Authors:Chenghao Xiao, Hou Pong (Ken) Chan, Hao Zhang, Weiwen Xu, Mahani Aljunied, Yu Rong
Abstract:
Recent multi-modal embedding approaches leveraging multi-modal large language models (MLLMs) fine-tuned with contrastive learning (CL) have shown promising results. However, the underlying reasons behind their superiority remain underexplored. This work argues that a crucial advantage of MLLM-based approaches stems from implicit cross-modal alignment achieved during generative pretraining, where the language decoder learns to leverage multimodal information within a shared representation space for generating unimodal outputs. We empirically demonstrate that latent alignment emerges within MLLM representations, allowing CL to serve as a lightweight refinement stage. Leveraging this insight, we propose a language-centric omnimodal representation learning approach. Extensive experiments across diverse backbones and benchmarks demonstrate the effectiveness of our approach. Furthermore, we identify a Generation-Representation Scaling Law (GRSL), demonstrating that the representational capacity gained through contrastive refinement scales positively with the MLLM's generative capabilities. This suggests that improving generative abilities evolves as an effective paradigm for boosting representation quality. Finally, we provide a theoretical explanation for the GRSL based on mutual information discrepancy



Paperid:1121
Authors:Nataša Bolić, Tom Cesari, Roberto Colomboni, Christian Paravalos
Abstract:
Motivated by real-life supply chain management, we study a repeated newsvendor problem in which the learner is a mediator that facilitates trades between suppliers and retailers in a sequence of supplier/retailer interactions. At each time step, a new supplier and retailer join the mediator's platform with a private production cost and utility function, respectively, and the platform proposes a unitary trading price. The supplier accepts the proposed price if it meets or exceeds their unitary production cost and communicates their decision to the platform; simultaneously, the retailer decides the quantity to purchase at the proposed trading price based on their private utility function and sends their decision to the platform. If the supplier accepts the trading price, the transaction proceeds, and the retailer purchases their chosen quantity of units, paying the product of this quantity and the trading price to the supplier. The mediator's objective is to maximize social welfare. We design an online mediator's pricing strategy that features sharp regret rates under some natural assumptions, and we investigate the necessity of these assumptions, proving that relaxing any of them leads to unlearnability.



Paperid:1122
Authors:Stratis Skoulakis
Abstract:
Abstract:The *law of supply and demand* asserts that in a perfectly competitive market, the price of a good adjusts to a *market clearing price*. In a market clearing price $p^\star$ the number of sellers willing to sell the good at $p^\star$ equals the number of sellers willing to buy the good at price $p^\star$. In this work, we provide a mathematical foundation on the law of supply and demand through the lens of online learning. Specifically, we demonstrate that if each seller employs a no-swap regret algorithm to set their individual selling price—aiming to maximize its individual revenue—the collective pricing dynamics converge to the market-clearing price $p^\star$ . Our findings offer a novel perspective on the law of supply and demand, framing it as the emergent outcome of an adaptive learning processes among sellers.



Paperid:1123
Authors:Jacob Eeuwe Kooi, Zhao Yang, Vincent Francois-Lavet
Abstract:
Neural network architectures have a large impact in machine learning. In reinforcement learning, network architectures have remained notably simple, as changes often lead to small gains in performance. This work introduces a novel encoder architecture for pixel-based model-free reinforcement learning. The Hadamax (Hadamardmax}-pooling) encoder achieves state-of-the-art performance by max-pooling Hadamard products between GELU-activated parallel hidden layers. Based on the recent PQN algorithm, the Hadamax encoder achieves state-of-the-art model-free performance in the Atari-57 benchmark. Specifically, without applying any algorithmic hyperparameter modifications, Hadamax-PQN achieves an 80\% performance gain over vanilla PQN and significantly surpasses Rainbow-DQN. For reproducibility, the codebase is provided in the supplementary materials.



Paperid:1124
Authors:Yuan Zhang, Jiacheng Jiang, Guoqing Ma, Zhiying Lu, Bo Wang, Haoyang Huang, Jianlong Yuan, Nan Duan
Abstract:
In this work, we present GPDiT, a Generative Pre-trained Autoregressive Diffusion Transformer that unifies the strengths of diffusion and autoregressive modeling for long-range video synthesis, within a continuous latent space. Instead of predicting discrete tokens, GPDiT autoregressively predicts future latent frames using a diffusion loss, enabling natural modeling of motion dynamics and semantic consistency across frames. This continuous autoregressive framework not only enhances generation quality but also endows the model with representation capabilities. Additionally, we introduce a lightweight causal attention variant and a parameter-free rotation-based time-conditioning mechanism, improving both the training and inference efficiency. Extensive experiments demonstrate that GPDiT achieves strong performance in video generation quality, video representation ability, and few-shot learning tasks, highlighting its potential as an effective framework for video modeling in continuous space.



Paperid:1125
Authors:Boyi Wei, Benedikt Stroebl, Jiacen Xu, Joie Zhang, Zhou Li, Peter Henderson
Abstract:
Foundation models are increasingly becoming better autonomous programmers, raising the prospect that they could also automate dangerous offensive cyber‑operations. Current frontier model audits probe the cybersecurity risks of such agents, but most fail to account for the degrees of freedom available to adversaries in the real world.In particular, with strong verifiers and financial incentives, agents for offensive cybersecurity are amenable to iterative improvement by would-be adversaries. We argue that assessments should take into account an expanded threat model in the context of cybersecurity, emphasizing the varying degrees of freedom that an adversary may possess instatefulandnon-statefulenvironments within a fixed compute budget. We show that even with a relatively small compute budget (8 H100 GPU Hours in our study), adversaries can improve an agent's cybersecurity capability on InterCode CTF by more than 40\% relative to the baseline---without any external assistance. These results highlight the need to evaluate agents' cybersecurity risk in a dynamic manner, painting a more representative picture of risk.



Paperid:1126
Authors:Yuki Shibukawa, Taira Tsuchiya, Shinsaku Sakaue, Kenji Yamanishi
Abstract:
Abstract:Online structured prediction is a task of sequentially predicting outputs with complex structures based on inputs and past observations, encompassing online classification. Recent studies showed that in the full-information setting, we can achieve finite bounds on the *surrogate regret*, *i.e.*, the extra target loss relative to the best possible surrogate loss. In practice, however, full-information feedback is often unrealistic as it requires immediate access to the whole structure of complex outputs. Motivated by this, we propose algorithms that work with less demanding feedback, *bandit* and *delayed* feedback. For bandit feedback, by using a standard inverse-weighted gradient estimator, we achieve a surrogate regret bound of $O(\sqrt{KT})$ for the time horizon $T$ and the size of the output set $K$. However, $K$ can be extremely large when outputs are highly complex, resulting in an undesirable bound. To address this issue, we propose another algorithm that achieves a surrogate regret bound of $O(T^{2/3})$, which is independent of $K$. This is achieved with a carefully designed pseudo-inverse matrix estimator. Furthermore, we numerically compare the performance of these algorithms, as well as existing ones. Regarding delayed feedback, we provide algorithms and regret analyses that cover various scenarios, including full-information and bandit feedback, as well as fixed and variable delays.



Paperid:1127
Authors:Zhengyi Zhong, Wenzheng Jiang, Weidong Bao, Ji Wang, Qi Wang, Guanbo Wang, Yongheng Deng, Ju Ren
Abstract:
Conventional federated learning (FL) assumes a closed world with a fixed total number of clients. In contrast, new clients continuously join the FL process in real-world scenarios, introducing new knowledge. This raises two critical demands: detecting new knowledge, i.e., knowledge discovery, and integrating it into the global model, i.e., knowledge adaptation. Existing research focuses on coarse-grained knowledge discovery, and often sacrifices source domain performance and adaptation efficiency. To this end, we propose a fine-grained federated domain adaptation approach in open set (Gains). Gains splits the model into an encoder and a classifier, empirically revealing features extracted by the encoder are sensitive to domain shifts while classifier parameters are sensitive to class increments. Based on this, we develop fine-grained knowledge discovery and contribution-driven aggregation techniques to identify and incorporate new knowledge. Additionally, an anti-forgetting mechanism is designed to preserve source domain performance, ensuring balanced adaptation. Experimental results on multi-domain datasets across three typical data-shift scenarios demonstrate that Gains significantly outperforms other baselines in performance for both source-domain and target-domain clients. Code is available at: https://anonymous.4open.science/r/Gains.



Paperid:1128
Authors:Yifei He, Siqi Zeng, Yuzheng Hu, Rui Yang, Tong Zhang, Han Zhao
Abstract:
Model merging provides a scalable alternative to multi-task training by combining specialized finetuned models through parameter arithmetic, enabling efficient deployment without the need for joint training or access to all task data. While recent methods have shown promise, existing evaluations are limited in both model scale and task diversity, leaving open questions about their applicability to large, domain-specialized LLMs. To tackle the challenges, we introduce MergeBench, a comprehensive evaluation suite designed to assess model merging at scale. MergeBench builds on state-of-the-art open-source language models, including Llama and Gemma families at 2B to 9B scales, and covers five key domains: instruction following, mathematics, multilingual understanding, coding and safety. We standardize finetuning and evaluation protocols, and assess eight representative merging methods across multi-task performance, forgetting and runtime efficiency. Based on extensive experiments, we provide practical guidelines for algorithm selection and share insights showing that model merging tends to perform better on stronger base models, with techniques such as merging coefficient tuning and sparsification improving knowledge retention. However, several challenges remain, including the computational cost on large models, the gap for in-domain performance compared to multi-task models, and the underexplored role of model merging in standard LLM training pipelines. We hope MergeBench provides a foundation for future research to advance the understanding and practical application of model merging.



Paperid:1129
Authors:Kelly W Zhang, Tianhui Cai, Hongseok Namkoong, Daniel Russo
Abstract:
We introduce a framework for Thompson sampling (TS) contextual bandit algorithms, in which the algorithm's ability to quantify uncertainty and make decisions depends on the quality of a generative model that is learned offline. Instead of viewing uncertainty in the environment as arising from unobservable latent parameters, our algorithm treats uncertainty as stemming from missing, but potentially observable outcomes (including both future and counterfactual outcomes). If these outcomes were all observed, one could simply make decisions using an "oracle" policy fit on the complete dataset. Inspired by this conceptualization, at each decision-time, our algorithm uses a generative model to probabilistically impute missing outcomes, fits a policy using the imputed complete dataset, and uses that policy to select the next action. We formally show that this algorithm is a generative formulation of TS and establish a state-of-the-art regret bound. Notably, our regret bound depends on the generative model only through the quality of its offline prediction loss, and applies to any method of fitting the "oracle" policy.



Paperid:1130
Authors:Minghao Chen, Jianyuan Wang, Roman Shapovalov, Tom Monnier, Hyunyoung Jung, Dilin Wang, Rakesh Ranjan, Iro Laina, Andrea Vedaldi
Abstract:
We introduce AutoPartGen, a model that generates objects composed of 3D parts in an autoregressive manner.This model can take as input an image of an object, 2D masks of the object's parts, or an existing 3D object, and generate a corresponding compositional 3D reconstruction.Our approach builds upon 3DShape2VecSet, a recent latent 3D representation with powerful geometric expressiveness.We observe that this latent space exhibits strong compositional properties, making it particularly well-suited for part-based generation tasks.Specifically, AutoPartGen generates object parts autoregressively, predicting one part at a time while conditioning on previously generated parts and additional inputs, such as 2D images, masks, or 3D objects.This process continues until the model decides that all parts have been generated, thus determining automatically the type and number of parts.The resulting parts can be seamlessly assembled into coherent objects or scenes without requiring additional optimization.We evaluate both the overall 3D generation capabilities and the part-level generation quality of AutoPartGen, demonstrating that it achieves state-of-the-art performance in 3D part generation.



Paperid:1131
Authors:Yoonho Lee, Junseok Lee, Sangwoo Seo, Sungwon Kim, Yeongmin Kim, Chanyoung Park
Abstract:
Despite the promising results of disentangled representation learning in discovering latent patterns in graph-structured data, few studies have explored disentanglement for hypergraph-structured data. Integrating hyperedge disentanglement into hypergraph neural networks enables models to leverage hidden hyperedge semantics, such as unannotated relations between nodes, that are associated with labels. This paper presents an analysis of hyperedge disentanglement from a category-theoretical perspective and proposes a novel criterion for disentanglement derived from the naturality condition. Our proof-of-concept model experimentally showed the potential of the proposed criterion by successfully capturing functional relations of genes (nodes) in genetic pathways (hyperedges).



Paperid:1132
Authors:Kaiwen Wang, Jin Zhou, Jonathan Chang, Zhaolin Gao, Nathan Kallus, Kianté Brantley, Wen Sun
Abstract:
Abstract:In this paper, we propose a simple and efficient method for value model training on long-context reasoning traces. Compared to existing process reward models (PRMs), our method does not require a fine-grained notion of "step," which is difficult to define for long-context reasoning models. By collecting a dataset of 2.5 million reasoning traces, we train a 1.5B token-level value model and apply it to DeepSeek models for improved performance with test-time compute scaling. We find that block-wise value-guided search (VGS) with a final weighted majority vote achieves better test-time scaling than standard methods such as majority voting or best-of-$n$. With an inference budget of 64 generations, VGS with \texttt{DeepSeek-R1-Distill-1.5B} achieves an average accuracy of $45.7\%$ across four competition math benchmarks (AIME 2024 \& 2025, HMMT Feb 2024 \& 2025), reaching parity with o3-mini-medium.Moreover, VGS significantly reduces the inference FLOPs required to achieve the same performance of majority voting. Our dataset, model and codebase are open-sourced at ANONYMIZED.



Paperid:1133
Authors:Kangrui Wang, Pingyue Zhang, Zihan Wang, Yaning Gao, Linjie Li, Qineng Wang, Hanyang Chen, Yiping Lu, Zhengyuan Yang, Lijuan Wang, Ranjay Krishna, Jiajun Wu, Fei-Fei Li, Yejin Choi, Manling Li
Abstract:
Reinforcement Learning (RL) has demonstrated effectiveness in enhancing Large Language Model (LLM) in performing multi-turn agentic tasks, while Vision-Language Models (VLMs) has not yet shown promise in completing multi-turn agentic tasks, where a significant challenge is reasoning about visual states. Our empirical investigation reveals that incorporating explicit visual state reasoning, such as Grounding (describing current visual state) and WorldModeling (predicting next state), into the VLM's thinking process during RL training significantly enhances task performance. We further explore optimal visual state representations, finding natural language effective generally, while structured formats prove crucial for tasks demanding high precision or understanding of low-level visual details. To specifically reinforce visual state reasoning, we introduce Visual Reasoning RL, which incorporates two key techniques: a turn-level visual reasoning reward to supervise reasoning accuracy, and Bi-Level General Advantage Estimation (GAE) that estimates advantages at both turn and token levels. This comprehensive approach consistently improves task performance and reasoning quality. These experiments are facilitated by VAGEN, a scalable training framework for multi-turn VLM agents across diverse visual environments. Our findings offer pathways to more robust visual state reasoning in VLM agents.



Paperid:1134
Authors:Wenyi WU, Zixuan Song, Kun Zhou, Yifei Shao, Zhiting Hu, Biwei Huang
Abstract:
Language models (LMs) and their extension, vision-language models (VLMs), have achieved remarkable performance across various tasks. However, they still struggle with complex reasoning tasks that require multimodal or multilingual real world knowledge. To support such capabilities, an external memory system that can efficiently provide relevant multimodal information is essential. Existing approaches generally concatenate image and text tokens into a long sequence as memory, which, however, may drastically increase context length and even degrade performance. In contrast, we propose using continuous memory-a compact set of dense embeddings-to more effectively and efficiently represent multimodal and multilingual knowledge. Our key insight is that a VLM can serve as its own continuous memory encoder. We empirically show that this design improves performance on complex multimodal reasoning tasks. Building on this, we introduce a data-efficient and parameter-efficient method to fine-tune the VLM into a memory encoder, requiring only 1.2\% of the model’s parameters and a small corpus of 15.6K self-synthesized samples. Our approach CoMEM utilizes VLM's original capabilities to encode arbitrary multimodal and multilingual knowledge into just 8 continuous embeddings. Since the inference-time VLM remains frozen, our memory module is plug-and-play and can be flexibly integrated as needed. Extensive experiments across eight multimodal reasoning benchmarks demonstrate the effectiveness of our approach. Code and data will be publicly released.



Paperid:1135
Authors:Bojia Zi, Weixuan Peng, Xianbiao Qi, Jianan Wang, Shihao Zhao, Rong Xiao, Kam-Fai Wong
Abstract:
Recent advances in video diffusion models have driven rapid progress in video editing techniques. However, video object removal, a critical subtask of video editing, remains challenging due to issues such as hallucinated objects and visual artifacts. Furthermore, existing methods often rely on computationally expensive sampling procedures and classifier-free guidance (CFG), resulting in slow inference. To address these limitations, we proposeMiniMax-Remover, a novel two-stage video object removal approach. Motivated by the observation that text condition is not best suited for this task, we simplify the pretrained video generation model by removing textual input and cross-attention layers, resulting in a more lightweight and efficient model architecture in the first stage. In the second stage, we distilled our remover on successful videos produced by the stage-1 model and curated by human annotators, using a minimax optimization strategy to further improve editing quality and inference speed. Specifically, the inner maximization identifies adversarial input noise ("bad noise") that makes failure removals, while the outer minimization step trains the model to generate high-quality removal results even under such challenging conditions. As a result, our method achieves a state-of-the-art video object removal results with as few as 6 sampling steps and doesn't rely on CFG, significantly improving inference efficiency. Extensive experiments demonstrate the effectiveness and superiority of MiniMax-Remover compared to existing methods.



Paperid:1136
Authors:Chenxing Wei, Jiarui Yu, Ying He, Hande Dong, Yao Shu, Fei Yu
Abstract:
DeepSeek-R1 has successfully enhanced Large Language Model (LLM) reasoning capabilities through its rule-based reward system. While it's a ''perfect'' reward system that effectively mitigates reward hacking, such reward functions are often discrete. Our experimental observations suggest that discrete rewards can lead to gradient anomaly, unstable optimization, and slow convergence. To address this issue, we propose ReDit (Reward Dithering), a method that dithers the discrete reward signal by adding simple random noise. With this perturbed reward, exploratory gradients are continuously provided throughout the learning process, enabling smoother gradient updates and accelerating convergence. The injected noise also introduces stochasticity into flat reward regions, encouraging the model to explore novel policies and escape local optima. Experiments across diverse tasks demonstrate the effectiveness and efficiency of ReDit. On average, ReDit achieves performance comparable to vanilla GRPO with only approximately 10% the training steps, and furthermore, still exhibits a 4% performance improvement over vanilla GRPO when trained for a similar duration. Visualizations confirm significant mitigation of gradient issues with ReDit. Moreover, theoretical analyses are provided to further validate these advantages.



Paperid:1137
Authors:Mahdi Nikdan, Vincent Cohen-Addad, Dan Alistarh, Vahab Mirrokni
Abstract:
Abstract:Effective data selection is critical for efficient training of modern Large Language Models (LLMs). This paper introduces Influence Distillation, a novel, mathematically-justified framework for data selection that employs second-order information to optimally weight training samples. By distilling each sample's influence on a target distribution, our method assigns model-specific weights that are used to select training data for LLM fine-tuning, guiding it toward strong performance on the target domain. We derive these optimal weights for both Gradient Descent and Adam optimizers. To ensure scalability and reduce computational cost, we propose a $\textit{landmark-based approximation}$: influence is precisely computed for a small subset of "landmark" samples and then efficiently propagated to all other samples to determine their weights. We validate Influence Distillation by applying it to instruction tuning on the Tulu V2 dataset, targeting a range of tasks including GSM8k, SQuAD, and MMLU, across several models from the Llama and Qwen families. Experiments show that Influence Distillation matches or outperforms state-of-the-art performance while achieving up to $3.5\times$ faster selection.



Paperid:1138
Authors:Bowen Yang, Bharat Venkitesh, Dwaraknath Gnaneshwar Talupuru, Hangyu Lin, David Cairuz, Phil Blunsom, Acyr Locatelli
Abstract:
Long-context large language models (LLMs) have achieved remarkable advancements, driven by techniques like Rotary Position Embedding (RoPE) (Su et al., 2023) and its extensions (Chen et al., 2023; Liu et al., 2024c; Peng et al., 2023). By adjusting RoPE parameters and incorporating training data with extended contexts, we can train performant models with considerably longer input sequences. However, existing RoPE-based meth- ods exhibit performance limitations when applied to extended context lengths. This paper presents a comprehensive analysis of various attention mechanisms, including RoPE, No Positional Em- bedding (NoPE), and Query-Key Normalization (QK-Norm), identifying their strengths and short- comings in long-context modeling. Our inves- tigation identifies distinctive attention patterns in these methods and highlights their impact on long-context performance, providing valuable in- sights for architectural design. Building on these findings, we propose a novel architectural based on a hybrid attention mechanism that not only surpasses conventional RoPE-based transformer models in long-context tasks but also achieves competitive performance on benchmarks requir- ing shorter context lengths.



Paperid:1139
Authors:Junbin Xiao, Nanxin Huang, Hao Qiu, Zhulin Tao, Xun Yang, Richang Hong, Meng Wang, Angela Yao
Abstract:
We present EgoBlind, the first egocentric VideoQA dataset collected from blind individuals to evaluate the assistive capabilities of contemporary multimodal large language models (MLLMs). EgoBlind comprises 1,392 videos that record the daily lives of real blind users from a first-person perspective. It also features 5,311 questions directly posed or generated and verified by blind individuals to reflect their needs for visual assistance under various scenarios. We provide each question with an average of 3 reference answers to alleviate subjective evaluation. Using EgoBlind, we comprehensively evaluate 16 advanced MLLMs and find that all models struggle, with the best performers achieving accuracy near 60\%, far behind human performance of 87.4\%. To guide future advancements, we identify and summarize major limitations of existing MLLMs in egocentric visual assistance for the blind and explore heuristic solutions for improvement. With these efforts, we hope EgoBlind can serve as a valuable foundation for developing more effective AI assistants to enhance the independence of the blind individuals' lives.



Paperid:1140
Authors:Dongyuan Li, Shiyin Tan, Ying Zhang, Ming Jin, Shirui Pan, Manabu Okumura, Renhe Jiang
Abstract:
Dynamic graph modeling aims to uncover evolutionary patterns in real-world systems, enabling accurate social recommendation and early detection of cancer cells. Inspired by the success of recent state space models in efficiently capturing long-term dependencies, we propose DyG-Mamba by translating dynamic graph modeling into a long-term sequence modeling problem. Specifically, inspired by Ebbinghaus' forgetting curve, we treat the irregular timespans between events as control signals, allowing DyG-Mamba to dynamically adjust the forgetting of historical information. This mechanism ensures effective usage of irregular timespans, thereby improving both model effectiveness and inductive capability. In addition, inspired by Ebbinghaus' review cycle, we redefine core parameters to ensure that DyG-Mamba selectively reviews historical information and filters out noisy inputs, further enhancing the model’s robustness. Through exhaustive experiments on 12 datasets covering dynamic link prediction and node classification tasks, we show that DyG-Mamba achieves state-of-the-art performance on most datasets, while demonstrating significantly improved computational and memory efficiency. Our code is available at https://anonymous.4open.science/r/DyGMamba-4784/README.md.



Paperid:1141
Authors:Santiago Mazuelas
Abstract:
Methods for split conformal prediction leverage calibration samples to transform any prediction rule into a set-prediction rule that complies with a target coverage probability. Existing methods provide remarkably strong performance guarantees with minimal computational costs. However, they require to use calibration samples composed by labeled examples different to those used for training. This requirement can be highly inconvenient, as it prevents the use of all labeled examples for training and may require acquiring additional labels solely for calibration. This paper presents an effective methodology for split conformal prediction with unsupervised calibration for classification tasks. In the proposed approach, set-prediction rules are obtained using unsupervised calibration samples together with supervised training samples previously used to learn the classification rule. Theoretical and experimental results show that the presented methods can achieve performance comparable to that with supervised calibration, at the expenses of a moderate degradation in performance guarantees and computational efficiency.



Paperid:1142
Authors:Vydhourie Thiyageswaran, Tyler H. McCormick, Jennifer Brennan
Abstract:
Abstract:Randomized controlled trials often suffer from interference, a violation of the Stable Unit Treatment Values Assumption (SUTVA), where a unit's outcome is influenced by its neighbors' treatment assignments. This interference biases naive estimators of the average treatment effect (ATE). A popular method to achieve unbiasedness pairs the Horvitz-Thompson estimator of the ATE with a known exposure mapping, a function that identifies units in a given randomization unaffected by interference. For example, an exposure mapping may stipulate that a unit experiences no interference if at least an $h$-fraction of its neighbors share its treatment status. However, selecting this threshold $h$ is challenging: domain experts may struggle to specify it accurately, and a misspecified threshold can introduce bias. In this work, we propose a data-adaptive method to select the $h$-fractional threshold that minimizes the mean-squared-error (MSE) of the Hortvitz-Thompson estimator. Our approach estimates the bias and variance of the Horvitz-Thompson estimator under candidate thresholds by leveraging a first-order approximation, specifically, linear regression of potential outcomes on exposures. We present simulations illustrating that our method improves upon non-adaptive threshold choices, and an adapted Lepski's method. We further illustrate the performance of our estimator by running experiments on a village network dataset, and on a publicly-available Amazon product similarity graph. Furthermore, we demonstrate that our method remains robust to deviations from the linear potential outcomes model.



Paperid:1143
Authors:Yanglin Feng, Yongxiang Li, Yuan Sun, Yang Qin, Dezhong Peng, Peng Hu
Abstract:
Text-3D Scene Retrieval (T3SR) aims to retrieve relevant scenes using linguistic queries. Although traditional T3SR methods have made significant progress in capturing fine-grained associations, they implicitly assume that query descriptions are information-complete. In practical deployments, however, limited by the capabilities of users and models, it is difficult or even impossible to directly obtain a perfect textual query suiting the entire scene and model, thereby leading to performance degradation. To address this issue, we propose a novel Interactive Text-3D Scene Retrieval Method (IDeal), which promotes the enhancement of the alignment between texts and 3D scenes through continuous interaction. To achieve this, we present an Interactive Retrieval Refinement framework (IRR), which employs a questioner to pose contextually relevant questions to an answerer in successive rounds that either promote detailed probing or encourage exploratory divergence within scenes. Upon the iterative responses received from the answerer, IRR adopts a retriever to perform both feature-level and semantic-level information fusion, facilitating scene-level interaction and understanding for more precise re-rankings. To bridge the domain gap between queries and interactive texts, we propose an Interaction Adaptation Tuning strategy (IAT). IAT mitigates the discriminability and diversity risks among augmented text features that approximate the interaction text domain, achieving contrastive domain adaptation for our retriever. Extensive experimental results on three datasets demonstrate the superiority of IDeal.



Paperid:1144
Authors:Han Yi, Yulu Pan, Feihong He, Xinyu Liu, Benjamin Zhang, Oluwatumininu Oguntola, Gedas Bertasius
Abstract:
We present ExAct, a new video-language benchmark for expert-level understanding of skilled physical human activities. Our new benchmark contains 3,521 expert-curated video question-answer pairs spanning 11 physical activities in 6 domains: Sports, Bike Repair, Cooking, Health, Music, and Dance. ExAct requires the correct answer to be selected from five carefully designed candidate options, thus necessitating a nuanced, fine-grained, expert-level understanding of physical human skills. Evaluating the recent state-of-the-art VLMs on ExAct reveals a substantial performance gap relative to human expert performance. Specifically, the best-performing GPT-4o model achieves only 44.70% accuracy, well below the 82.02% attained by trained human specialists/experts. We believe that our ExAct will be beneficial for developing and evaluating VLMs capable of precise understanding of human skills in various physical and procedural domains. We will release the dataset and evaluation code.



Paperid:1145
Authors:Yuma Fujimoto, Kenshi Abe, Kaito Ariu
Abstract:
Abstract:This study raises and addresses the problem of time-delayed feedback in learning in games. Because learning in games assumes that multiple agents independently learn their strategies, a discrepancy in optimization often emerges among the agents. To overcome this discrepancy, the prediction of the future reward is incorporated into algorithms, typically known as Optimistic Follow-the-Regularized-Leader (OFTRL). However, the time delay in observing the past rewards hinders the prediction. Indeed, this study firstly proves that even a single-step delay worsens the performance of OFTRL from the aspects of the regret and convergence. This study proposes the weighted OFTRL (WOFTRL), where the prediction vector of the next reward in OFTRL is weighted $n$ times. We further capture an intuition that the optimistic weight cancels out this time delay. We prove that when the optimistic weight exceeds the time delay, our WOFTRL recovers the good performances that the regret is constant ($O(1)$-regret) in general-sum normal-form games, and the strategies converge to the Nash equilibrium as a subsequence (best-iterate convergence) in poly-matrix zero-sum games. The theoretical results are supported and strengthened by our experiments.



Paperid:1146
Authors:Elias Abad Rocamora, Christian Schlarmann, Naman Deep Singh, Yongtao Wu, Matthias Hein, Volkan Cevher
Abstract:
Adversarial input attacks can cause a significant shift of CLIP embeddings. This can affect the downstream robustness of models incorporating CLIP in the pipeline, such as text-to-image generative models or large vision language models. While some efforts have been done towards making the CLIP image encoders robust, the robustness of text encoders remains unexplored. In this work, we cover this gap in the literature. We propose LEAF: an efficient adversarial finetuning method for the text domain, with the ability to scale to large CLIP models. Our models significantly improve the zero-shot adversarial accuracy in the text domain, while maintaining the vision performance provided by robust image encoders. When combined with text-to-image diffusion models, we can improve the generation quality under adversarial noise. When employing our robust CLIP encoders in multimodal retrieval tasks, we improve the recall under adversarial noise over standard CLIP models. Finally, we show that robust text encoders facilitate better reconstruction of input text from its embedding via direct optimization.



Paperid:1147
Authors:Ahmad Bdeir, Johannes Burchert, Lars Schmidt-Thieme, Niels Landwehr
Abstract:
Hyperbolic deep learning has become a growing research direction in computer vision due to the unique properties afforded by the alternate embedding space. The negative curvature and exponentially growing distance metric provide a natural framework for capturing hierarchical relationships between datapoints and allowing for finer separability between their embeddings. However, current hyperbolic learning approaches are still prone to overfitting, computationally expensive, and prone to instability, especially when attempting to learn the manifold curvature to adapt to tasks and different datasets. To address these issues, our paper presents a derivation for Riemannian AdamW that helps increase hyperbolic generalization ability. For improved stability, we introduce a novel fine-tunable hyperbolic scaling approach to constrain hyperbolic embeddings and reduce approximation errors. Using this along with our curvature-aware learning schema for Riemannian Optimizers enables the combination of curvature and non-trivialized hyperbolic parameter learning. Our approach demonstrates consistent performance improvements across Computer Vision, EEG classification, and hierarchical metric learning tasks while greatly reducing runtime.



Paperid:1148
Authors:Shiyun Lin, Simon Mauras, Nadav Merlis, Vianney Perchet
Abstract:
Abstract:We study matching markets with ties, where workers on one side of the market may have tied preferences over jobs, determined by their matching utilities. Unlike classical two-sided markets with strict preferences, no single stable matching exists that is utility-maximizing for all workers. To address this challenge, we introduce the \emph{Optimal Stable Share} (OSS)-ratio, which measures the ratio of a worker's maximum achievable utility in any stable matching to their utility in a given matching. We prove that distributions over only stable matchings can incur linear utility losses, i.e., an $\Omega (N)$ OSS-ratio, where $N$ is the number of workers. To overcome this, we design an algorithm that efficiently computes a distribution over (possibly non-stable) matchings, achieving an asymptotically tight $O (\log N)$ OSS-ratio. When exact utilities are unknown, our second algorithm guarantees workers a logarithmic approximation of their optimal utility under bounded instability. Finally, we extend our offline approximation results to a bandit learning setting where utilities are only observed for matched pairs. In this setting, we consider worker-optimal stable regret, design an adaptive algorithm that smoothly interpolates between markets with strict preferences and those with statistical ties, and establish a lower bound revealing the fundamental trade-off between strict and tied preference regimes.



Paperid:1149
Authors:Yingying Fan, Yuxuan Han, Jinchi Lv, Xiaocong Xu, Zhengyuan Zhou
Abstract:
In this paper, we study the behavior of the Upper Confidence Bound-Variance (UCB-V) algorithm for the Multi-Armed Bandit (MAB) problems, a variant of the canonical Upper Confidence Bound (UCB) algorithm that incorporates variance estimates into its decision-making process. More precisely, we provide an asymptotic characterization of the arm-pulling rates for UCB-V, extending recent results for the canonical UCB in Kalvit and Zeevi (2021) and Khamaru and Zhang (2024). In an interesting contrast to the canonical UCB, our analysis reveals that the behavior of UCB-V can exhibit instability, meaning that the arm-pulling rates may not always be asymptotically deterministic. Besides the asymptotic characterization, we also provide non-asymptotic bounds for the arm-pulling rates in the high probability regime, offering insights into the regret analysis. As an application of this high probability result, we establish that UCB-V can achieve a more refined regret bound, previously unknown even for more complicate and advanced variance-aware online decision-making algorithms. A matching regret lower bound is also established, demonstrating the optimality of our result.



Paperid:1150
Authors:Qiyang Li, Zhiyuan Zhou, Sergey Levine
Abstract:
We present Q-learning with action chunking (Q-LAC), a simple yet strong actor-critic RL algorithm for offline-to-online RL. Our method addresses two common shortcomings of the existing actor-critic RL methods in these problem settings: (1) slow 1-step temporal-difference (TD) backups, (2) temporally incoherent actions for exploration. The former slows down value backup, leading to inefficient value learning and sample inefficiency. The latter reduces the quality of data collection online, compounding the sample inefficiency further. Our key idea is to use temporally extended actions where the policy predicts a sequence of actions for a fixed horizon and executes them one-by-one open loop, and run RL update directly with this extended action space with behavioral constraint. RL training on the temporal extended action space speeds up the TD backup by ``skipping" over time steps, while the behavioral constraint and the open-loop execution ensures the temporal coherence of the actions. On a range of long-horizon, sparse-reward manipulation tasks, our method exhibit strong offline performance and online sample efficiency, outperforming prior methods that operate in the original action space and skill-based methods.



Paperid:1151
Authors:Yixu Wang, Jiaxin Song, Yifeng Gao, Xin Wang, YANG YAO, Yan Teng, Xingjun Ma, Yingchun Wang, Yu-Gang Jiang
Abstract:
As Video Large Multimodal Models (VLMMs) rapidly advance, their inherent complexity introduces significant safety challenges, particularly the issue of mismatched generalization where static safety alignments fail to transfer to dynamic video contexts. We introduce SafeVid, a framework designed to instill video-specific safety principles in VLMMs. SafeVid uniquely transfers robust textual safety alignment capabilities to the video domain by employing detailed textual video descriptions as an interpretive bridge, facilitating LLM-based rule-driven safety reasoning. This is achieved through a closed-loop system comprising: 1) generation of SafeVid-350K, a novel 350,000-pair video-specific safety preference dataset; 2) targeted alignment of VLMMs using Direct Preference Optimization (DPO); and 3) comprehensive evaluation via our new SafeVidBench benchmark. Alignment with SafeVid-350K significantly enhances VLMM safety, with models like LLaVA-NeXT-Video demonstrating substantial improvements (e.g., up to 42.39%) on SafeVidBench. SafeVid provides critical resources and a structured approach, demonstrating that leveraging textual descriptions as a conduit for safety reasoning markedly improves the safety alignment of VLMMs in complex multimodal scenarios.



Paperid:1152
Authors:Fabio De Sousa Ribeiro, Ainkaran Santhirasekaram, Ben Glocker
Abstract:
We investigate the open question of counterfactual identifiability for high-dimensional variables given observational data. Pearl (2000) insists that counterfactuals must be identifiable—that is, uniquely recoverable from observed data—to support valid causal claims. A recent line of work prioritises empirical results but lacks identifiability analysis, raising concerns about the validity of its counterfactual claims due to confounding. To tackle this, we connect Optimal Transport (OT) with causality and prove that, under standard regularity, a mechanism within a Markovian structural causal model admits a unique, dynamic OT flow which: (i) pushes the observational distribution forward to the counterfactual; (ii) preserves a vector-monotone order that generalises scalar quantiles, ensuring counterfactual equivalence. Importantly, we provide a practical model recipe built on continuous-time OT flows, validate our theory in controlled scenarios, and demonstrate significant improvements in the axiomatic soundness of counterfactuals of real data.



Paperid:1153
Authors:Yangtian Zhang, Sizhuang He, Daniel Levine, Lawrence Zhao, David Zhang, Syed Rizvi, Shiyang Zhang, Emanuele Zappala, Rex Ying, David van Dijk
Abstract:
Discrete diffusion models offer a flexible, controllable approach to structured sequence generation, yet they still lag behind causal language models in expressive power. A key limitation lies in their reliance on the Markovian assumption, which restricts each step to condition only on the current state, leading to potential uncorrectable error accumulation.In this paper, We introduce CaDDi, a discrete diffusion model that conditions on the entire generative trajectory, thereby lifting the Markov constraint and allowing the model to revisit and improve past states. By unifying sequential (causal) and temporal (diffusion) reasoning in a single non‑Markovian transformer, CaDDi also treats standard causal language models as a special case and permits the direct reuse of pretrained LLM weights with no architectural changes. Empirically, CaDDi outperforms state‑of‑the‑art discrete diffusion baselines on natural‑language benchmarks, substantially narrowing the remaining gap to large autoregressive transformers.



Paperid:1154
Authors:Qiangqiang Mao, Jiayang Ren, Yixiu Wang, Chenxuanyin Zou, Jingjing Zheng, Yankai Cao
Abstract:
Decision tree, despite its unmatched interpretability and lightweight structure, faces two key issues: non-differentiability and low testing accuracy, both of which limit its broader applicability. This study addresses these issues by developing a differentiable oblique tree that optimizes entire tree structure using gradient-based optimization, avoiding the suboptimality common in greedy tree inductions. We propose an exact reformulation of hard-split trees based on "ReLU+Argmin" mechanism, and then cast the reformulated tree training as an unconstrained optimization task. The ReLU-based sample branching, expressed as exact-zero or non-zero values, preserve a unique decision path, in contrast to soft decision trees with probabilistic routing. The subsequent Argmin operation identifies the unique zero-violation path, enabling deterministic predictions. For effective gradient flow, we approximate Argmin behaviors by scaling softmin function. To ameliorate numerical instability, we propose a warm-start annealing scheme that solves multiple optimization tasks with increasingly accurate approximations. This reformulation alongside distributed GPU parallelism offers strong scalability, supporting 12-depth tree even on million-scale datasets where most baselines fail. Extensive experiments demonstrate that our optimized tree achieves superior testing accuracy against 14 baselines, including an average improvement of 7.54\% over CART, even 2.01\% over the classic random forest.



Paperid:1155
Authors:Heeseung Yun, Joonil Na, Jaeyeon Kim, Calvin Murdock, Gunhee Kim
Abstract:
People continuously perceive and interact with their surroundings based on underlying intentions that drive their exploration and behaviors.While research in egocentric user and scene understanding has focused primarily on motion and contact-based interaction, forecasting human visual perception itself remains less explored despite its fundamental role in guiding human actions and its implications for AR/VR and assistive technologies.We address the challenge of egocentric 3D visual span forecasting, predicting where a person's visual perception will focus next within their three-dimensional environment.To this end, we propose EgoSpanLift, a novel method that transforms egocentric visual span forecasting from 2D image planes to 3D scenes.EgoSpanLift converts SLAM-derived keypoints into gaze-compatible geometry and extracts volumetric visual span regions.We further combine EgoSpanLift with 3D U-Net and unidirectional transformers, enabling spatio-temporal fusion to efficiently predict future visual span in 3D grid.In addition, we curate a comprehensive benchmark from raw egocentric multisensory data, creating a testbed with 364.6K samples for 3D visual span forecasting.Our approach outperforms competitive baselines for egocentric gaze anticipation and 3D localization, while achieving comparable results even when projected back onto 2D image planes without additional 2D-specific training.



Paperid:1156
Authors:Hoomaan Maskan, Yikun Hou, Suvrit Sra, Alp Yurtsever
Abstract:
Abstract:We introduce a new projection-free (Frank-Wolfe) method for optimizing structured nonconvex functions that are expressed as a difference of two convex functions. This problem class subsumes smooth nonconvex minimization, positioning our method as a promising alternative to the classical Frank-Wolfe algorithm. DC decompositions are not unique; by carefully selecting a decomposition, we can better exploit the problem structure, improve computational efficiency, and adapt to the underlying problem geometry to find better local solutions. We prove that the proposed method achieves a first-order stationary point in $\mathcal{O}(1/\epsilon^2)$ iterations, matching the complexity of the standard Frank-Wolfe algorithm for smooth nonconvex minimization in general. Specific decompositions can, for instance, yield a gradient-efficient variant that requires only $\mathcal{O}(1/\epsilon)$ calls to the gradient oracle by reusing computed gradients over multiple iterations. Finally, we present numerical experiments demonstrating the effectiveness of the proposed method compared to other projection-free algorithms.



Paperid:1157
Authors:Jinho Choi, Hyesu Lim, Steffen Schneider, Jaegul Choo
Abstract:
Dataset bias is ubiquitous in machine learning datasets. Yet, systematically identifying these biases is challenging without costly, fine-grained attribute annotations. We introduce ConceptScope, a framework for characterizing dataset bias by disentangling visual concepts using a Sparse Autoencoder. Our framework automatically discovers visual concepts present in datasets and distinguishes them into target, contextual, and bias concepts. We first validate our framework by accurately detecting six distinct types of visual concepts—object, texture, background, facial attributes, emotion, and action—achieving high accuracy on labeled datasets. We then extend our approach to discover both known biases in annotated datasets (such as CelebA) and novel biases in datasets without explicit bias annotations (such as ImageNet). Furthermore, we introduce a method to partition test data into subgroups based on the strength and presence of task-related versus bias concepts, introducing a practical use case of ConceptScope for model robustness diagnosis. Our approach leverages existing datasets without the need for additional bias annotations, providing valuable insights into how concept distributions affect model generalization under bias-induced distribution shifts.



Paperid:1158
Authors:Xinyu Zhou, Tongxin Pan, Lingyi Hong, Pinxue Guo, HaiJing Guo, Zhaoyu Chen, Kaixun Jiang, Wenqiang Zhang
Abstract:
UAV tracking can be widely applied in scenarios such as disaster rescue, environmental monitoring, and logistics transportation. However, existing UAV tracking methods predominantly emphasize speed and lack exploration in semantic awareness, which hinders the search region from extracting accurate localization information from the template. The limitation results in suboptimal performance under typical UAV tracking challenges such as camera motion, fast motion, and low resolution, etc. To address this issue, we propose a dynamic semantic aware correlation modeling tracking framework. The core of our framework is a Dynamic Semantic Relevance Generator, which, in combination with the correlation map from the Transformer, explore semantic relevance. The approach enhances the search region's ability to extract important information from the template, improving accuracy and robustness under the aforementioned challenges.Additionally, to enhance the tracking speed, we design a pruning method for the proposed framework. Therefore, we present multiple model variants that achieve trade-offs between speed and accuracy, enabling flexible deployment according to the available computational resources. Experimental results validate the effectiveness of our method, achieving competitive performance on multiple UAV tracking datasets.



Paperid:1159
Authors:Ni Zhang, Zhiguang Cao
Abstract:
Existing GFlowNet-based methods for vehicle routing problems (VRPs) typically employ Trajectory Balance (TB) to achieve global optimization but often neglect important aspects of local optimization. While Detailed Balance (DB) addresses local optimization more effectively, it alone falls short in VRPs, which inherently require holistic trajectory optimization. To address these limitations, we introduce the Hybrid-Balance GFlowNet (HBG) framework, which uniquely integrates TB and DB in a principled and adaptive manner by aligning their intrinsically complementary strengths. Additionally, we propose a specialized inference strategy for depot-centric scenarios like the Capacitated Vehicle Routing Problem (CVRP), leveraging the depot node's greater flexibility in selecting successors. Despite this specialization, HBG maintains broad applicability, extending effectively to problems without explicit depots, such as the Traveling Salesman Problem (TSP). We evaluate HBG by integrating it into two established GFlowNet-based solvers—AGFN and GFACS—and demonstrate consistent and significant improvements across both CVRP and TSP benchmarks, underscoring the enhanced solution quality and generalization afforded by our approach.



Paperid:1160
Authors:Katarina Petrović, Lazar Atanackovic, Viggo Moro, Kacper Kapusniak, Ismail Ceylan, Michael Bronstein, Joey Bose, Alexander Tong
Abstract:
Modeling the transport dynamics of natural processes from population-level observations is a ubiquitous problem in the natural sciences. Such models rely on key assumptions about the underlying process in order to enable faithful learning of governing dynamics that mimic the actual system behavior. The de facto assumption in current approaches relies on the principle of least action that results in gradient field dynamics and leads to trajectories minimizing an energy functional between two probability measures. However, many real-world systems, such as cell cycles in single-cell RNA, are known to exhibit non-gradient, periodic behavior, which fundamentally cannot be captured by current state-of-the-art methods such as flow and bridge matching. In this paper, we introduce Curly Flow Matching (Curly-FM), a novel approach that is capable of learning non-gradient field dynamics by designing and solving a Schrödinger bridge problem with a non-zero drift reference process---in stark contrast to typical zero-drift reference processes---which is constructed using inferred velocities in addition to population snapshot data. We showcase Curly-FM by solving the trajectory inference problems for single cells, computational fluid dynamics, and ocean currents with approximate velocities. We demonstrate that Curly-FM can learn trajectories that better match both the reference process and population marginals. Curly-FM expands flow matching models beyond the modeling of populations and towards the modeling of known periodic behavior in physical systems.



Paperid:1161
Authors:Palash Chatterjee, Roni Khardon
Abstract:
Continuous time systems are often modeled using discrete time dynamics but this requires a small simulation step to maintain accuracy. In turn, this requires a large planning horizon which leads to computationally demanding planning problems and reduced performance. Previous work in model free reinforcement learning has partially addressed this issue using action repeats where a policy is learned to determine a discrete action duration. Instead we propose to control the continuous decision timescale directly by using temporally-extended actions and letting the planner treat the duration of the action as an additional optimization variable along with the standard action variables. This additional structure has multiple advantages. It speeds up simulation time of trajectories and, importantly, it allows for deep horizon search in terms of primitive actions while using a shallow search depth in the planner. In addition, in the model based reinforcement learning (MBRL) setting, it reduces compounding errors from model learning and improves training time for models. We show that this idea is effective and that the range for action durations can be automatically selected using a multi-armed bandit formulation and integrated into the MBRL framework. An extensive experimental evaluation both in planning and in MBRL, shows that our approach yields faster planning, better solutions, and that it enables solutions to problems that are not solved in the standard formulation.



Paperid:1162
Authors:WEI XING, Zhenjie Lu, Akeel Shah
Abstract:
Existing Multi-fidelity Bayesian Optimization (MFBO) methods ignore the convergence behavior of the multi-fidelity surrogate as the fidelity increases, leading to inefficient exploration and suboptimal performance. We introduce CAMO (Convergence-Aware Multi-fidelity Optimization), a principled framework based on Linear Fidelity Differential Equations (LFiDEs) that explicitly encodes convergence of fidelity-indexed outputs and employs a closed-form nonstationary kernel. We rigorously prove the existence and pointwise/uniform convergence to the high fidelity surrogate under mild restrictions and provide new convergence results for general FiDEs using smooth, non-smooth and even non-convex Lyapunov functions, establishing a bridge between MFBO and the theory of subgradient flows in non-smooth optimisation theory. Combined with a fidelity-aware acquisition function, CAMO outperforms state-of-the-art MFBO methods on a majority of synthetic and real-world benchmarks, with up to four-fold improvement in optimisation performance and dramatic speed-up in convergence. CAMO offers a tractable and theoretically grounded approach to convergence-aware MFBO.



Paperid:1163
Authors:Martino Bernasconi, Matteo Castiglioni, Andrea Celli, Gabriele Farina
Abstract:
Abstract:Correlated equilibria —and their generalization $\Phi$-equilibria— are a fundamental object of study in game theory, offering a more tractable alternative to Nash equilibria in multi-player settings. While computational aspects of equilibrium computation are well-understood in some settings, fundamental questions are still open in _generalized games_, that is, games in which the set of strategies allowed to each player depends on the other players' strategies. These classes of games model fundamental settings in economics and have been a cornerstone of economics research since the seminal paper of Arrow and Debreu [1954].Recently, there has been growing interest, both in economics and in computer science, in studying correlated equilibria in generalized games. It is known that finding a social welfare maximizing correlated equilibrium in generalized games is NP-hard. However, the existence of efficient algorithms to find _any_ equilibrium remains an important open question. In this paper, we answer this question negatively, showing that this problem is PPAD-complete.



Paperid:1164
Authors:Hadi Khalaf, Claudio Mayrink Verdun, Alex Oesterling, Himabindu Lakkaraju, Flavio Calmon
Abstract:
Abstract:A common paradigm to improve the performance of large language models is optimizing for a reward model. Reward models assign a numerical score to LLM outputs indicating, for example, which response would likely be preferred by a user or is most aligned with safety goals. However, reward models are never perfect. They inevitably function as proxies for complex desiderata such as correctness, helpfulness, and safety. By overoptimizing for a misspecified reward, we can subvert intended alignment goals and reduce overall performance - a phenomenon commonly referred to as reward hacking. In this work, we characterize reward hacking in inference-time alignment and demonstrate when and how we can mitigate it by hedging on the proxy reward. Hedging represents a tactical choice to avoid placing undue confidence in high but potentially misleading proxy reward signals. We study reward hacking under Best-of-$n$ (BoN) sampling, along with two inference-time methods, namely a novel method Best-of-Poisson (BoP) and Soft-Best-of-$n$ (SBoN), which introduce a parameter to control our confidence in the reward. We then propose $\texttt{HedgeTune}$ as an efficient algorithm to find the optimal hedging parameter. We demonstrate through experiments that hedging mitigates reward hacking and achieves superior distortion-reward tradeoffs with minimal computational overhead.



Paperid:1165
Authors:Wanru Zhao, Lucas Page-Caccia, Zhengyan Shi, Minseon Kim, Weijia Xu, Alessandro Sordoni
Abstract:
Curriculum learning is a class of training strategies that organizes the data being exposed to a model by difficulty, gradually from simpler to more complex examples. This research explores a reverse curriculum generation approach that recursively decomposes complex datasets into simpler, more learnable components.We propose a teacher-student framework where the teacher is equipped with the ability to reason step-by-step, which is used to recursively generate easier versions of examples, enabling the student model to progressively master difficult tasks. We propose a novel scoring system to measure data difficulty based on its structural complexity and conceptual depth, allowing curriculum construction over decomposed data. Experiments on math datasets (MATH and AIME) demonstrate that models trained with curricula generated by our approach exhibit superior performance compared to standard training on original datasets.



Paperid:1166
Authors:Dong-Hee Paek, SEUNG-HYUN KONG
Abstract:
Abstract:Sensor fusion of camera, LiDAR, and 4-dimensional (4D) Radar has brought a significant performance improvement in autonomous driving (AD). However, there still exist fundamental challenges: deeply coupled fusion methods assume continuous sensor availability, making them vulnerable to sensor degradation and failure, whereas sensor-wise cross-attention fusion methods struggle with computational cost and unified feature representation. This paper presents availability-aware sensor fusion (ASF), a novel method that employs unified canonical projection (UCP) to enable consistency in all sensor features for fusion and cross-attention across sensors along patches (CASAP) to enhance robustness of sensor fusion against sensor degradation and failure. As a result, the proposed ASF shows a superior object detection performance to the existing state-of-the-art fusion methods under various weather and sensor degradation (or failure) conditions; Extensive experiments on the K-Radar dataset demonstrate that ASF achieves improvements of 9.7\% in $AP_{BEV}$ (87.2\%) and 20.1\% in $AP_{3D}$ (73.6\%) in object detection at IoU=0.5, while requiring a low computational cost. The code for this work will be made publicly available following the completion of the review process.



Paperid:1167
Authors:Yeyu Yan, Shuai Zheng, Wenjun Hui, Xiangkai Zhu, Chen Dong, Zhenfeng Zhu, Yao Zhao, Kunlun He
Abstract:
Graph condensation (GC) aims to distill the original graph into a small-scale graph, mitigating redundancy and accelerating GNN training. However, conventional GC approaches heavily rely on rigid GNNs and task-specific supervision. Such a dependency severely restricts their reusability and generalization across various tasks and architectures. In this work, we revisit the goal of ideal GC from the perspective of GNN optimization consistency, and then a generalized GC optimization objective is derived, by which those traditional GC methods can be viewed nicely as special cases of this optimization paradigm. Based on this, \textbf{Pre}-trained \textbf{G}raph \textbf{C}ondensation (\textbf{PreGC}) via optimal transport is proposed to transcend the limitations of task- and architecture-dependent GC methods. Specifically, a hybrid-interval graph diffusion augmentation is presented to suppress the weak generalization ability of the condensed graph on particular architectures by enhancing the uncertainty of node states. Meanwhile, the matching between optimal graph transport plan and representation transport plan is tactfully established to maintain semantic consistencies across source graph and condensed graph spaces, thereby freeing graph condensation from task dependencies. To further facilitate the adaptation of condensed graphs to various downstream tasks, a traceable semantic harmonizer from source nodes to condensed nodes is proposed to bridge semantic associations through the optimized representation transport plan in pre-training. Extensive experiments verify the superiority and versatility of PreGC, demonstrating its task-independent nature and seamless compatibility with arbitrary GNNs.



Paperid:1168
Authors:Dilyara Bareeva, Marina Höhne, Alexander Warnecke, Lukas Pirch, Klaus-Robert Müller, Konrad Rieck, Sebastian Lapuschkin, Kirill Bykov
Abstract:
Deep Neural Networks (DNNs) are capable of learning complex and diverse representations; however, the semantic nature of the concepts they capture remains poorly understood. A widely used technique for interpreting the functional role of concepts within a model is Feature Visualization (FV), which synthesizes an input pattern that maximally activates a given feature. In this paper, we examine the vulnerability of FV to adversarial fine-tuning of the model. We introduce a novel technique, Gradient Slingshots, that enables manipulation of FV without changing the model architecture. By shaping new trajectories in the off-distribution regions of the activation landscape of a feature, we coerce the optimization process to converge in a predefined visualization. Crucially, this manipulation largely preserves the model’s external behavior and internal mechanisms, enabling deceptive explanations while maintaining functional integrity. We validate the effectiveness of our technique across several neural network models, demonstrating its ability to conceal the true functionality of arbitrary neurons by replacing their original explanations with selected target explanations during model auditing.



Paperid:1169
Authors:Xueguang Ma, Qian Liu, Dongfu Jiang, Ge Zhang, Zejun MA, Wenhu Chen
Abstract:
Reinforcement learning (RL) has recently demonstrated strong potential in enhancing the reasoning capabilities of large language models (LLMs).Particularly, the ``Zero'' reinforcement learning introduced by Deepseek-R1-Zero, enables direct RL training of base LLMs without relying on an intermediate supervised fine-tuning stage.Despite these advancements, current works for LLM reasoning mainly focus on mathematical and coding domains, largely due to data abundance and the ease of answer verification.This limits the applicability and generalization of such models to broader domains, where questions often have diverse answer representations, and data is more scarce.In this paper, we propose \model, a novel training paradigm designed to enhance LLM reasoning capabilities across diverse domains. Our key contributions include: (1) constructing a large-scale, high-quality dataset of questions with verifiable answers curated by web crawling, covering a wide range of disciplines; and (2) developing a generative model-based answer verifier, which replaces traditional rule-based verification with the capability of chain-of-thought and context-awareness. Our comprehensive evaluation across benchmarks such as MMLU-Pro, GPQA, SuperGPQA, BBEH and MATH, AMC, etc demonstrates that \model outperforms existing baseline methods, achieving robust and generalizable reasoning performance while maintaining superior effectiveness in mathematical reasoning tasks. Code, data, and model checkpoints in this work will be released.



Paperid:1170
Authors:Valeria Ruscio, Umberto Nanni, Fabrizio Silvestri
Abstract:
Attention sink (AS) is a consistent pattern in transformer attention maps where certain tokens (often special tokens or positional anchors) disproportionately attract attention from other tokens. We show that in transformers, AS is not an architectural artifact, but it is the manifestation of a fundamental geometric principle: the establishment of reference frames that anchor representational spaces. We analyze several architectures and identify three distinct reference frame types, centralized, distributed, and bidirectional, that correlate with the attention sink phenomenon. We show that they emerge during the earliest stages of training as optimal solutions to the problem of establishing stable coordinate systems in high-dimensional spaces. We show the influence of architecture components, particularly position encoding implementations, on the specific type of reference frame. This perspective transforms our understanding of transformer attention mechanisms and provides insights for both architecture design and the relationship with AS.



Paperid:1171
Authors:Feiyang Fu, Tongxian Guo, Zhaoqiang Liu
Abstract:
Discrete diffusion models (DDMs) have shown powerful generation ability for discrete data modalities like text and molecules. However, their practical application is hindered by inefficient sampling, requiring a large number of sampling steps. Accelerating DDMs by using larger step sizes typically introduces significant problems in generation quality, as it amplifies the impact of both the compounding decoding error due to factorized predictions and discretization error from numerical approximations, leading to a significant decrease in sampling quality. To address these challenges, we propose learnable sampler distillation (LSD), a novel approach to train fast and high-fidelity samplers for DDMs. LSD employs a distillation approach where a student sampler with a few steps learns to align its intermediate score trajectory with that of a high-quality teacher sampler with numerous steps. This alignment is achieved by optimizing learnable sampler coefficients that adaptively adjust sampling dynamics. Additionally, we further propose LSD+, which also learns time schedules that allocate steps non-uniformly. Experiments across text generation, image generation, and synthetic tasks demonstrate that our proposed approaches outperform existing samplers for DDMs, achieving substantially higher sampling quality with significantly fewer sampling steps.



Paperid:1172
Authors:Ge Meng, Zhongnan Cai, Ruizhe Chen, Jingyan Tu, Yingying Wang, Yue Huang, Xinghao Ding
Abstract:
Generating hyperspectral images (HSIs) from RGB images through spectral reconstruction can significantly reduce the cost of HSI acquisition. In this paper, we propose a Fractal-Based Recursive Spectral Reconstruction Network (FRN), which differs from existing paradigms that attempt to directly integrate the full-spectrum information from the R, G, and B channels in a one-shot manner. Instead, it treats spectral reconstruction as a progressive process, predicting from broad to narrow bands or employing a coarse-to-fine approach for predicting the next wavelength. Inspired by fractals in mathematics, FRN establishes a novel spectral reconstruction paradigm by recursively invoking an atomic reconstruction module. In each invocation, only the spectral information from neighboring bands is used to provide clues for the generation of the image at the next wavelength, which follows the low-rank property of spectral data. Moreover, we design a band-aware state space model that employs a pixel-differentiated scanning strategy at different stages of the generation process, further suppressing interference from low-correlation regions caused by reflectance differences. Through extensive experimentation across different datasets, FRN achieves superior reconstruction performance compared to state-of-the-art methods in both quantitative and qualitative evaluations.



Paperid:1173
Authors:Shuo Cheng, Liqian Ma, Zhenyang Chen, Ajay Mandlekar, Caelan Garrett, Danfei Xu
Abstract:
Behavior cloning has shown promise for robot manipulation by mimicking human demonstrations, but achieving robust, generalizable performance in the real world often requires costly and labor-intensive data collection to obtain these demonstrations. Recent advances in simulation and automated motion synthesis offer scalable alternatives for generating training data. However, transferring policies from simulation to the real world remains challenging due to simulation modeling inaccuracies. In this work, we propose a framework for learning generalizable manipulation policies that primarily leverages simulation and only requires a few real-world demonstrations. Central to our approach is learning a shared feature space that preserves task-relevant structure across simulation and the real world. Specifically, augment traditional imitation learning objective functions with a new loss inspired by optimal transport that encourages domain-invariant feature learning. We pair this with a motion generator that automatically synthesizes diverse simulated trajectories from a few manual demonstrations. We validate our method on challenging manipulation tasks in both simulation, where investigate sim-to-sim transfer, and the real world, demonstrating effective and data-efficient policy transfer.



Paperid:1174
Authors:Daniel Beechey, Özgür Şimşek
Abstract:
Reinforcement learning has achieved remarkable success in complex decision-making environments, yet its lack of transparency limits deployment in critical settings. Shapley values provide a principled framework for explaining reinforcement learning, but their exponential computational cost makes them impractical for real-world problems. We address this challenge by introducing FastSVERL, a scalable method for approximating Shapley values designed to handle the unique challenges of reinforcement learning, including temporal dependencies across multi-step trajectories, learning from off-policy data, and adapting to evolving agent behaviours. These contributions position FastSVERL as a practical solution for real-time, Shapley-based interpretability in reinforcement learning.



Paperid:1175
Authors:Hon Tik Tse, Siddarth Chandrasekar, Marlos C. Machado
Abstract:
In recent years, the successor representation (SR) has attracted increasing attention in reinforcement learning (RL), and it has been used to address some of its key challenges, such as exploration, credit assignment, and generalization. The SR can be seen as representing the underlying credit assignment structure of the environment by implicitly encoding its induced transition dynamics. However, the SR is reward-agnostic. In this paper, we discuss a similar representation that also takes into account the reward dynamics of the problem. We study the default representation (DR), a recently proposed representation with limited theoretical (and empirical) analysis. Here, we lay some of the theoretical foundation underlying the DR in the tabular case by (1) deriving dynamic programming and (2) temporal-difference methods to learn the DR, (3) characterizing the basis for the vector space of the DR, and (4) formally extending the DR to the function approximation case through default features. Empirically, we analyze the benefits of the DR in many of the settings in which the SR has been applied, including (1) reward shaping, (2) option discovery, (3) exploration, and (4) transfer learning. Our results show that, compared to the SR, the DR gives rise to qualitatively different, reward-aware behaviour and quantitatively better performance in several settings.



Paperid:1176
Authors:Yasser Khalil, Mehdi Setayesh, Hongliang Li
Abstract:
Machine unlearning (MU) aims to remove the influence of specific ''forget'' data from a trained model while preserving its knowledge of the remaining ''retain'' data. Existing MU methods based on label manipulation or model weight perturbations often achieve limited unlearning effectiveness. To address this, we introduce CoUn, a novel MU framework inspired by the observation that a model retrained from scratch using only retain data classifies forget data based on their semantic similarity to the retain data. CoUn emulates this behavior by adjusting learned data representations through contrastive learning (CL) and supervised learning, applied exclusively to retain data. Specifically, CoUn (1) leverages semantic similarity between data samples to indirectly adjust forget representations using CL, and (2) maintains retain representations within their respective clusters through supervised learning. Extensive experiments across various datasets and model architectures show that CoUn consistently outperforms state-of-the-art MU baselines in unlearning effectiveness. Additionally, integrating our CL module into existing baselines empowers their unlearning effectiveness.



Paperid:1177
Authors:Sepideh Mahabadi, Sherry Sarkar, Jakub Tarnawski
Abstract:
Abstract:Submodular maximization subject to matroid constraints is a central problem with many applications in machine learning. As algorithms are increasingly used in decision-making over datapoints with sensitive attributes such as gender or race, it is becoming crucial to enforce fairness to avoid bias and discrimination. Recent work has addressed the challenge of developing efficient approximation algorithms for fair matroid submodular maximization. However, the best algorithms known so far are only guaranteed to satisfy a relaxed version of the fairness constraints that loses a factor 2, i.e., the problem may ask for $\ell$ elements with a given attribute, but the algorithm is only guaranteed to find $\lfloor \ell/2 \rfloor$. In particular, there is no provable guarantee when $\ell=1$, which corresponds to a key special case of perfect matching constraints.In this work, we achieve a new trade-off via an algorithm that gets arbitrarily close to full fairness. Namely, for any constant $\varepsilon>0$, we give a constant-factor approximation to fair monotone matroid submodular maximization that in expectation loses only a factor $(1-\varepsilon)$ in the lower-bound fairness constraint. Our empirical evaluation on a standard suite of real-world datasets -- including clustering, recommendation, and coverage tasks -- demonstrates the practical effectiveness of our methods.



Paperid:1178
Authors:Diji Yang, Minghao Liu, Chung-Hsiang Lo, Yi Zhang, James Davis
Abstract:
Vision-language models (VLMs) have shown strong performance on text-to-image retrieval benchmarks. However, bridging this success to real-world applications remains a challenge. In practice, human search behavior is rarely a one-shot action. Instead, it is often a multi-round process guided by clues in mind. That is, a mental image ranging from vague recollections to vivid mental representations of the target image. Motivated by this gap, we study the task of Mental Image Retrieval (MIR), which targets the realistic yet underexplored setting where users refine their search for a mentally envisioned image through multi-round interactions with an image search engine. Central to successful interactive retrieval is the capability of machines to provide users with clear, actionable feedback; however, existing methods rely on indirect or abstract verbal feedback, which can be ambiguous, misleading, or ineffective for users to refine the query. To overcome this, we propose GenIR, a generative multi-round retrieval paradigm leveraging diffusion-based image generation to explicitly reify the AI system’s understanding at each round. These synthetic visual representations provide clear, interpretable feedback, enabling users to refine their queries intuitively and effectively. We further introduce a fully automated pipeline to generate a high-quality multi-round MIR dataset. Experimental results demonstrate that GenIR significantly outperforms existing interactive methods in the MIR scenario. This work establishes a new task with a dataset and an effective generative retrieval method, providing a foundation for future research in this direction.



Paperid:1179
Authors:Shiwei Gan, Yafeng Yin, Zhiwei Jiang, Lei Xie, Sanglu Lu, Hongkai Wen
Abstract:
Recent advances in sign language research have benefited from CNN-based backbones, which are primarily transferred from traditional computer vision tasks (\eg object detection, image recognition). However, these CNN-based backbones usually excel at extracting features like contours and texture, but may struggle with capturing sign-related features. To capture such region-related features, SignGraph model extracts the cross-region features by building Local Sign Graph (LSG) module, Temporal Sign Graph (TSG) module. However, we emphasize that although capturing cross-region dependencies can improve sign language performance, it may degrade the representation quality of local regions.To mitigate this, We introduce MixSignGraph, which represents sign sequences as a group of mixed graphs and designs the following graph modules for feature extraction. Specifically, besides from the LSG module and TSG module that model the intra-frame and inter-frame cross-regions features, we design a simple yet effective Hierarchical Sign Graph (HSG) module, which enhances local region representations following the extraction of cross-region features, by aggregating the same-region features from different-granularity feature maps of a frame, \ie to boosts discriminative local features.In addition, to further improve the performance of gloss-free sign language tasks, we propose a simple yet counter-intuitive Text-based CTC Pre-training (TCTC) method, which generates pseudo gloss labels from text sequence for model pre-training. Extensive experiments conducted on current five public sign language datasets demonstrate that our model surpasses the SOTA models on multiple sign language tasks across several datasets, without relying on any additional cues. Code and models are made available after publication.



Paperid:1180
Authors:Botong Zhang, Shuo Li, Ignacio Hounie, Osbert Bastani, Dongsheng Ding, Alejandro Ribeiro
Abstract:
We study the problem of computing an optimal large language model (LLM) policy for a constrained alignment problem, where the goal is to maximize a primary reward objective while satisfying constraints on secondary utilities. Despite the popularity of Lagrangian-based LLM policy search in constrained alignment, iterative primal-dual methods often fail to converge, and non-iterative dual-based methods do not achieve optimality in the LLM parameter space. To address these challenges, we employ Lagrangian duality to develop an iterative dual-based alignment method that alternates between updating the LLM policy via Lagrangian maximization and updating the dual variable via dual descent. In theory, we characterize the primal-dual gap between the primal value in the distribution space and the dual value in the LLM parameter space. We further quantify the optimality gap of the learned LLM policies at near-optimal dual variables with respect to both the objective and the constraint functions. These results prove that dual-based alignment methods can find an optimal constrained LLM policy, up to an LLM parametrization gap. We demonstrate the effectiveness and merits of our approach through extensive experiments conducted on the PKU-SafeRLHF dataset.



Paperid:1181
Authors:Jie Ren, Zhenwei Dai, Xianfeng Tang, Yue XING, Shenglai Zeng, Hui Liu, Jingying Zeng, Qiankun Peng, Samarth Varshney, Suhang Wang, Qi He, Charu Aggarwal, Hui Liu
Abstract:
Although Large Language Models (LLMs) have demonstrated impressive capabilities across a wide range of tasks, growing concerns have emerged over the misuse of sensitive, copyrighted, or harmful data during training. To address these concerns, unlearning techniques have been developed to remove the influence of specific data without retraining from scratch. However, this paper reveals a critical vulnerability in fine-tuning-based unlearning: a malicious user can craft a manipulated forgetting request that stealthily degrades the model’s utility for benign users. We demonstrate this risk through a red-teaming Stealthy Attack (SA), which is inspired by two key limitations of existing unlearning—the inability to constrain the scope of unlearning effect and the failure to distinguish benign tokens from unlearning signals. Prior work has shown that unlearned models tend to memorize forgetting data as unlearning signals, and respond with hallucinations or feigned ignorance when unlearning signals appear in the input. By subtly increasing the presence of common benign tokens in the forgetting data, SA enhances the connection between benign tokens and unlearning signals. As a result, when normal users include such tokens in their prompts, the model exhibits unlearning behaviors, leading to unintended utility degradation. To address this vulnerability, we propose Scope-aware Unlearning (SU), a lightweight enhancement that introduces a scope term into the unlearning objective, encouraging the model to localize the forgetting effect. Our method requires no additional data processing, integrates seamlessly with existing fine-tuning frameworks, and significantly improves robustness against SA. Extensive experiments validate the effectiveness of both SA and SU.



Paperid:1182
Authors:Eivinas Butkus, Nikolaus Kriegeskorte
Abstract:
Some argue that deep neural networks are fundamentallystatisticalsystems that fail to capture the causal generative processes behind their training data. Here we demonstrate that a GPT-style transformer trained for next-token prediction can simultaneously discover instances of linear Gaussian structural causal models (SCMs) and learn to answer counterfactual queries about them. First, we show that the network generalizes to counterfactual queries about SCMs for which it sawonlystrings describing noisy interventional data. Second, we decode the implicit SCM from the network's residual stream activations and use gradient descent to intervene on that "mental" SCM with predictable effects on the model's output. Our results suggest that neural networks trained using statistical prediction objectives on passively observed data may nevertheless discover and learn to use causal models of the world.



Paperid:1183
Authors:Noga Amit, Shafi Goldwasser, Orr Paradise, Guy Rothblum
Abstract:
Abstract:How can we trust the correctness of a learned model on a particular input of interest? Model accuracy is typically measured *on average* over a distribution of inputs, giving no guarantee for any fixed input. This paper proposes a theoretically-founded solution to this problem: to train *Self-Proving models* that prove the correctness of their output to a verification algorithm $V$ via an Interactive Proof. Self-Proving models satisfy that, with high probability over an input sampled from a given distribution, the model generates a correct output *and* successfully proves its correctness to $V$. The *soundness* property of $V$ guarantees that, for *every* input, no model can convince $V$ of the correctness of an incorrect output. Thus, a Self-Proving model proves correctness of most of its outputs, while *all* incorrect outputs (of any model) are detected by $V$. We devise a generic method for learning Self-Proving models, and prove its convergence under certain assumptions.



Paperid:1184
Authors:Long-Fei Li, Yu-Yang Qian, Peng Zhao, Zhi-Hua Zhou
Abstract:
Reinforcement Learning from Human Feedback (RLHF) has shown remarkable success in aligning Large Language Models (LLMs) with human preferences. Traditional RLHF approaches rely on a fixed dataset, which often suffers from limited coverage. To this end, online RLHF has emerged as a promising direction, enabling iterative data collection. Despite its potential, this paradigm faces a key bottleneck: the requirement to continuously integrate new data into the historical dataset and re-optimize the model from scratch at each iteration, resulting in computational and storage costs that grow linearly with the number of iterations. In this work, we address this challenge by proposing aone-passreward modeling method that does not require storing the historical data and can be computed in constant time. Specifically, we first formalize RLHF as a contextual preference bandit problem and design an online mirror descent algorithm with a tailored local norm to replace the standard maximum likelihood estimation for reward modeling. We then apply our method to various online RLHF settings, including passive data collection, active data collection, and deployment-time adaptation. We provide theoretical guarantees showing that our method improves both statistical and computational efficiency. Finally, we provide practical algorithms and conduct experiments usingLlama-3-8B-InstructandQwen2.5-7Bmodels on the Ultrafeedback-binarized and Mixture2 datasets, validating the effectiveness of our proposed method.



Paperid:1185
Authors:Udvas Das, Apurv Shukla, Debabrota Basu
Abstract:
Abstract:Preference-based Pure Exploration (PrePEx) aims to identify with a given confidence level the set of Pareto optimal arms in a vector-valued (aka multi-objective) bandit, where the reward vectors are ordered via a (given) preference cone $\mathcal C$. Though PrePEx and its variants are well-studied, there does not exist a *computationally efficient* algorithm that can *optimally* track the existing lower bound (Shukla and Basu, 2024) for arbitrary preference cones. We successfully fill this gap by efficiently solving the minimisation and maximisation problems in the lower bound. First, we derive three structural properties of the lower bound that yield a computationally tractable reduction of the minimisation problem. Then, we deploy a Frank-Wolfe optimiser to accelerate the maximisation problem in the lower bound. Together, these techniques solve the maxmin optimisation problem in $\mathcal O(KL^{2})$ time for a bandit instance with $K$ arms and $L$ dimensional reward, which is a significant acceleration over the literature. We further prove that our proposed PrePEx algorithm, **FraPPE**, asymptotically achieves the optimal sample complexity. Finally, we perform numerical experiments across synthetic and real datasets demonstrating that **FraPPE** achieves the lowest sample complexities to identify the exact Pareto set among the existing algorithms.



Paperid:1186
Authors:George Rapakoulias, Ali Pedram, Fengjiao Liu, Lingjiong Zhu, Panagiotis Tsiotras
Abstract:
Schrodinger Bridges (SBs) are diffusion processes that steer, in finite time, a given initial distribution to another final one while minimizing a suitable cost functional. Although various methods for computing SBs have recently been proposed in the literature, most of these approaches require computationally expensive training schemes, even for solving low-dimensional problems. In this work, we propose an analytic parametrization of a set of feasible policies for steering the distribution of a dynamical system from one Gaussian Mixture Model (GMM) to another. Instead of relying on standard non-convex optimization techniques, the optimal policy within the set can be approximated as the solution of a low-dimensional linear program whose dimension scales linearly with the number of components in each mixture. The proposed method generalizes naturally to more general classes of dynamical systems, such as controllable linear time-varying systems, enabling efficient solutions to multi-marginal momentum SB between GMMs, a challenging distribution interpolation problem. We showcase the potential of this approach in low-to-moderate dimensional problems such as image-to-image translation in the latent space of an autoencoder, learning of cellular dynamics using multi-marginal momentum SB problems, and various other examples.We also test our approach on an Entropic Optimal Transport (EOT) benchmark problem and show that it outperforms state-of-the-art methods in cases where the boundary distributions are mixture models while requiring virtually no training.



Paperid:1187
Authors:Yue Liu, Shengfang Zhai, Mingzhe Du, Yulin Chen, Tri Cao, Hongcheng Gao, Cheng Wang, Xinfeng Li, Kun Wang, Junfeng Fang, Jiaheng Zhang, Bryan Hooi
Abstract:
To enhance the safety of VLMs, this paper introduces a novel reasoning-based VLM guard model dubbed GuardReasoner-VL. The core idea is to incentivize the guard model to deliberatively reason before making moderation decisions via online RL.First, we construct GuardReasoner-VLTrain, a reasoning corpus with 123K samples and 631K reasoning steps, spanning text, image, and text-image inputs. Then, based on it, we cold-start our model's reasoning ability via SFT. In addition, we further enhance reasoning regarding moderation through online RL.Concretely, to enhance diversity and difficulty of samples, we conduct rejection sampling followed by data augmentation via the proposed safety-aware data concatenation.Besides, we use a dynamic clipping parameter to encourage exploration in early stages and exploitation in later stages. To balance performance and token efficiency, we design a length-aware safety reward that integrates accuracy, format, and token cost.Extensive experiments demonstrate the superiority of our model. Remarkably, it surpasses the runner-up by 19.27% F1 score on average. We release data, code, and models (3B/7B) ofGuardReasoner-VL.



Paperid:1188
Authors:ChangHao Li, Yuchen Zhuang, Rushi Qiang, Haotian Sun, Hanjun Dai, Chao Zhang, Bo Dai
Abstract:
Despite the impressive generative abilities of black-box large language models (LLMs), their inherent opacity hinders further advancements in capabilities such as reasoning, planning, and personalization. Existing works aim to enhance LLM capabilities via domain-specific adaptation, which require additional training on accessible model parameters, an infeasible option for black-box LLMs. To address this challenge, we introduce Matryoshka Pilot (M-Pilot), a lightweight white-box LLM controller that guides a large-scale black-box LLM generator by decomposing complex tasks into a series of intermediate outputs.Specifically, we consider the black-box LLM as an environment, with M-Pilot serving as a policy to provide intermediate guidance through prompts for driving the black-box LLM. M-Pilot is trained to pivot the outputs of the black-box LLM aligning with preferences during iterative interaction, which enables controllable multi-turn generation and self-improvement in optimizing intermediate guidance. Empirical evaluations on diverse tasks demonstrate that our method effectively enhances the capabilities of black-box LLMs in complex, long-horizon tasks.



Paperid:1189
Authors:Philipp Dahlinger, Tai Hoang, Denis Blessing, Niklas Freymuth, Gerhard Neumann
Abstract:
Accurately simulating physics is crucial across scientific domains, with applications spanning from robotics to materials science. While traditional mesh-based simulations are precise, they are often computationally expensive and require knowledge of physical parameters, such as material properties. In contrast, data-driven approaches like Graph Network Simulators (GNSs) offer faster inference but suffer from two key limitations: Firstly, they must be retrained from scratch for even minor variations in physical parameters, and secondly they require labor-intensive data collection for each new parameter setting. This is inefficient, as simulations with varying parameters often share a common underlying latent structure. In this work, we address these challenges by learning this shared structure through meta-learning, enabling fast adaptation to new physical parameters without retraining. To this end, we propose a novel architecture that generates a latent representation by encoding graph trajectories using conditional neural processes (CNPs). To mitigate error accumulation over time, we combine CNPs with a novel neural operator architecture. We validate our approach, Meta Neural Graph Operator (MaNGO), on several dynamics prediction tasks with varying material properties, demonstrating superior performance over existing GNS methods. Notably, MaNGO achieves accuracy on unseen material properties close to that of an oracle model.



Paperid:1190
Authors:Tanzhe Li, Caoshuo Li, Jiayi Lyu, Hongjuan Pei, Baochang Zhang, Taisong Jin, Rongrong Ji
Abstract:
State space models (SSMs) have recently garnered significant attention in computer vision. However, due to the unique characteristics of image data, adapting SSMs from natural language processing to computer vision has not outperformed the state-of-the-art convolutional neural networks (CNNs) and Vision Transformers (ViTs). Existing vision SSMs primarily leverage manually designed scans to flatten image patches into sequences locally or globally. This approach disrupts the original semantic spatial adjacency of the image and lacks flexibility, making it difficult to capture complex image structures. To address this limitation, we propose Dynamic Adaptive Scan (DAS), a data-driven method that adaptively allocates scanning orders and regions. This enables more flexible modeling capabilities while maintaining linear computational complexity and global modeling capacity. Based on DAS, we further propose the vision backbone DAMamba, which significantly outperforms popular vision Mamba models in vision tasks such as image classification, object detection, instance segmentation, and semantic segmentation. Notably, it surpasses some of the latest state-of-the-art CNNs and ViTs.



Paperid:1191
Authors:Korbinian Pöppel, Richard Freinschlag, Thomas Schmied, Wei Lin, Sepp Hochreiter
Abstract:
Modern recurrent architectures, such as xLSTM and Mamba, have recently challenged the Transformer in language modeling. However, their structure constrains their applicability to sequences only or requires processing multi-dimensional data structures, such as images or molecular graphs, in a pre-defined sequential order. In contrast, Multi-Dimensional RNNs (MDRNNs) are well suited for data with a higher level structure, like 2D grids, trees, and directed acyclic graphs (DAGs). In this work, we extend the notion of multi-dimensionality to linear RNNs. We introduce Parallelizable Linear Source Transition Mark networks (pLSTMs) using Source, Transition, and Mark gates that act on the linegraph of a general DAG. This enables parallelization in analogy to parallel associative scans and the chunkwise-recurrent form of sequential linear RNNs, but for DAGs. For regular grids (1D and 2D), like images, this scheme can be efficiently implemented using einsum operations, concatenations, and padding in logarithmic time. pLSTMs tackle the vanishing/exploding activation/gradient problem for long distances in DAGs via two distinct modes: a directed propagation mode (P-mode) and a diffusive distribution mode (D-mode). To showcase the long-range capabilities of pLSTM, we introduce arrow-pointing extrapolation as a synthetic computer vision task that contains long-distance directional information. We demonstrate that pLSTMs generalize well to larger image sizes, whereas Transformers struggle to extrapolate. On established molecular graph and computer vision benchmarks, pLSTMs also show strong performance.



Paperid:1192
Authors:Jinseong Park, Yujin Choi, Jaewook Lee
Abstract:
With the increasing need to safeguard data privacy in machine learning models, differential privacy (DP) is one of the major frameworks to build privacy-preserving models. Support Vector Machines (SVMs) are widely used traditional machine learning models due to their robust margin guarantees and strong empirical performance in binary classification. However, applying DP to multi-class SVMs is inefficient, as the standard one-versus-rest (OvR) and one-versus-one (OvO) approaches repeatedly query each data sample when building multiple binary classifiers, thus consuming the privacy budget proportionally to the number of classes. To overcome this limitation, we explore all-in-one SVM approaches for DP, which access each data sample only once to construct multi-class SVM boundaries with margin maximization properties. We propose a novel differentially Private Multi-class SVM (PMSVM) with weight and gradient perturbation methods, providing rigorous sensitivity and convergence analyses to ensure DP in all-in-one SVMs. Empirical results demonstrate that our approach surpasses existing DP-SVM methods in multi-class scenarios.



Paperid:1193
Authors:Alessio Russo, Alberto Maria Metelli, Marcello Restelli
Abstract:
Abstract:We address the learning problem in the context of infinite-horizon average-reward POMDPs. Traditionally, this problem has been approached using $\textit{Spectral Decomposition}$ (SD) methods applied to samples collected under non-adaptive policies, such as uniform or round-robin policies. Recently, SD techniques have been extended to accommodate a restricted class of adaptive policies such as $\textit{memoryless policies}$. However, the use of adaptive policies has introduced challenges related to data inefficiency, as SD methods typically require all samples to be drawn from a single policy.In this work, we propose $\texttt{Mixed Spectral Estimation}$, which generalizes spectral estimation techniques to support a broader class of $\textit{belief-based policies}$. We solve the open question of whether spectral methods can be applied to samples collected from multiple policies, and we provide finite-sample guarantees for our approach under standard observability and ergodicity assumptions.Building on this data-efficient estimation method, we introduce the $\texttt{Mixed Spectral UCRL}$ algorithm. Through a refined theoretical analysis, we demonstrate that it achieves a regret bound of $\widetilde{\mathcal{O}}(\sqrt{T})$ when compared to the optimal policy, without requiring full knowledge of either the transition or the observation model. Finally, we present numerical simulations that validate the theoretical analysis of both the proposed estimation procedure and the $\texttt{Mixed Spectral UCRL}$ algorithm.



Paperid:1194
Authors:Xiaoyuan Cheng, Xiaohang Tang, Yiming Yang
Abstract:
Abstract:Diffusion models have made significant strides in recent years, exhibiting strong generalization capabilities in planning and control tasks. However, most diffusion-based policies remain focused on reward maximization or cost minimization, often overlooking critical aspects of safety and stability. In this work, we propose Safe and Stable Diffusion ($S^2$Diff), a model-based framework that explores how diffusion models can ensure safety and stability from a Lyapunov perspective. We demonstrate that $S^2$Diff eliminates the reliance on both complex gradient-based solvers (e.g., quadratic programming, non-convex solvers) and control-affine structures, leading to globally valid control policies driven by the learned certificate functions. Additionally, we uncover intrinsic connections between diffusion sampling and almost Lyapunov theory, enabling the use of trajectory-level control policies to learn better certificate functions for safety and stability guarantees. To validate our approach, we conduct experiments on a wide variety of dynamical control systems, where $S^2$Diff consistently outperforms both certificate-based controllers and model-based diffusion baselines in terms of safety, stability, and overall control performance.



Paperid:1195
Authors:Amin Rakhsha, Kanika Madan, Tianyu Zhang, Amir-massoud Farahmand, Amir Khasahmadi
Abstract:
Abstract:Inference-time computational methods significantly enhance the reasoning abilities of Large Language Models (LLMs). Among these, Best-of-N has gained attention for its simplicity and scalability. It generates $N$ solutions from the LLM and selects the best one based on the reward model's evaluation. Due to noisy rewards, even with a large number of sampled solutions for a given question, the probability of selecting its correct answer does not necessarily converge to one.To mitigate this limitation, we propose Majority-of-the-Bests (MoB), a novel and hyperparameter-free selection mechanism that estimates the output distribution of Best-of-N via bootstrapping and selects its mode.Experimental results across five benchmarks, three different base LLMs, and two reward models demonstrate consistent improvements over Best-of-N in 25 out of 30 setups. We further provide theoretical results for the consistency of the bootstrapping.



Paperid:1196
Authors:Thorben Frank, Winfried Ripken, Gregor Lied, Klaus-Robert Müller, Oliver Unke, Stefan Chmiela
Abstract:
Diffusion Transformers (DiTs) have demonstrated strong performance in generative modeling, particularly in image synthesis, making them a compelling choice for molecular conformer generation. However, applying DiTs to molecules introduces novel challenges, such as integrating discrete molecular graph information with continuous 3D geometry, handling Euclidean symmetries, and designing conditioning mechanisms that generalize across molecules of varying sizes and structures. We propose DiTMC, a framework that adapts DiTs to address these challenges through a modular architecture that separates the processing of 3D coordinates from conditioning on atomic connectivity. To this end, we introduce two complementary graph-based conditioning strategies that integrate seamlessly with the DiT architecture. These are combined with different attention mechanisms, including both standard non-equivariant and SO(3)-equivariant formulations, enabling flexible control over the trade-off between between accuracy and computational efficiency.Experiments on standard conformer generation benchmarks (GEOM-QM9, -DRUGS, -XL) demonstrate that DiTMC achieves state-of-the-art precision and physical validity. Our results highlight how architectural choices and symmetry priors affect sample quality and efficiency, suggesting promising directions for large-scale generative modeling of molecular structures.



Paperid:1197
Authors:Haoze Sun, Linfeng Jiang, Fan Li, Renjing Pei, Zhixin Wang, Yong Guo, Jiaqi Xu, Haoyu Chen, Jin Han, Fenglong Song, Yujiu Yang, Wenbo Li
Abstract:
Real-world image super-resolution (RealSR) aims to enhance the visual quality of in-the-wild images, such as those captured by mobile phones. While existing methods leveraging large generative models demonstrate impressive results, the high computational cost and latency make them impractical for edge deployment. In this paper, we introduce PocketSR, an ultra-lightweight, single-step model that brings generative modeling capabilities to RealSR while maintaining high fidelity. To achieve this, we design LiteED, a highly efficient alternative to the original computationally intensive VAE in SD, reducing parameters by 97.5\% while preserving high-quality encoding and decoding. Additionally, we propose online annealing pruning for the U-Net, which progressively shifts generative priors from heavy modules to lightweight counterparts, ensuring effective knowledge transfer and further optimizing efficiency. To mitigate the loss of prior knowledge during pruning, we incorporate a multi-layer feature distillation loss. Through an in-depth analysis of each design component, we provide valuable insights for future research. PocketSR, with a model size of 146M parameters, processes 4K images in just 0.8 seconds, achieving a remarkable speedup over previous methods. Notably, it delivers performance on par with state-of-the-art single-step and even multi-step RealSR models, making it a highly practical solution for edge-device applications.



Paperid:1198
Authors:Jiajun Shi, Jian Yang, Jiaheng Liu, Xingyuan Bu, Jiangjie Chen, Junting Zhou, Kaijing Ma, Zhoufutu Wen, Bingli Wang, Yancheng He, Liang Song, Hualei Zhu, Shilong Li, Xingjian Wang, Wei Zhang, Ruibin Yuan, Yifan Yao, Wenjun Yang, Yunli Wang, Siyuan Fang, Siyu Yuan, Qianyu He, Robert Tang, Yingshui Tan, Wangchunshu Zhou, ZHAO-XIANG ZHANG, Zhoujun Li, Wenhao Huang, Ge Zhang
Abstract:
Recent advancements in large language models (LLMs) underscore the need for more comprehensive evaluation methods to accurately assess their reasoning capabilities. Existing benchmarks are often domain-specific and thus cannot fully capture an LLM’s general reasoning potential. To address this limitation, we introduce theKnowledge Orthogonal Reasoning Gymnasium (KORGym), a dynamic evaluation platform inspired by KOR-Bench and Gymnasium. KORGym offers over fifty games in either textual or visual formats and supports interactive, multi-turn assessments with reinforcement learning scenarios. Using KORGym, we conduct extensive experiments on 19 LLMs and 8 VLMs, revealing consistent reasoning patterns within model families and demonstrating the superior performance of closed-source models. Further analysis examines the effects of modality, reasoning strategies, reinforcement learning techniques, and response length on model performance. We expect KORGym to become a valuable resource for advancing LLM reasoning research and developing evaluation methodologies suited to complex, interactive environments.



Paperid:1199
Authors:Francois Lanusse, Liam Parker, Jeff Shen, Ollie Liu, Tom Hehir, Leopoldo Sarra, Lucas Meyer, Micah Bowles, Sebastian Wagner-Carena, Helen Qu, Siavash Golkar, Alberto Bietti, Hatim Bourfoune, Pierre Cornette, Keiya Hirashima, Geraud Krawezik, Ruben Ohana, Nicholas Lourie, Michael McCabe, Rudy Morel, Payel Mukhopadhyay, Mariel Pettee, Kyunghyun Cho, Miles Cranmer, Shirley Ho
Abstract:
While foundation models have shown promise across a variety of fields, astronomy lacks a unified framework for joint modeling across its highly diverse data modalities. In this paper, we present AION-1, the first large-scale multimodal foundation family of models for astronomy. AION-1 enables arbitrary transformations between heterogeneous data types using a two-stage architecture: modality-specific tokenization followed by transformer-based masked modeling of cross-modal token sequences. Trained on over 200M astronomical objects, AION-1 demonstrates strong performance across regression, classification, generation, and object retrieval tasks. Beyond astronomy, AION-1 provides a scalable blueprint for multimodal scientific foundation models that can seamlessly integrate heterogeneous combinations of real-world observations. Our model release is entirely open source, including the dataset, training script, and weights.



Paperid:1200
Authors:Ruofan Liang, Kai He, Zan Gojcic, Igor Gilitschenski, Sanja Fidler, Nandita Vijaykumar, Zian Wang
Abstract:
Estimating scene lighting from a single image or video remains a longstanding challenge in computer vision and graphics. Learning-based approaches are constrained by the scarcity of ground-truth HDR environment maps, which are expensive to capture and limited in diversity. While recent generative models offer strong priors for image synthesis, lighting estimation remains difficult due to its reliance on indirect visual cues, the need to infer global (non-local) context, and the recovery of high-dynamic-range outputs. We propose LuxDiT, a novel data-driven approach that fine-tunes a video diffusion transformer to generate HDR environment maps conditioned on visual input. Trained on a large synthetic dataset with diverse lighting conditions, our model learns to infer illumination from indirect visual cues and generalizes effectively to real-world scenes. To improve semantic alignment between the input and the predicted environment map, we introduce a low-rank adaptation finetuning strategy using a collected dataset of HDR panoramas. Our method produces accurate lighting predictions with realistic angular high-frequency details, outperforming existing state-of-the-art techniques in both quantitative and qualitative evaluations.



Paperid:1201
Authors:Panqi Chen, Lei Cheng, Jianlong Li, Weichang Li, Weiqing Liu, Jiang Bian, Shikai Fang
Abstract:
Tensor decomposition is a fundamental tool for analyzing multi-dimensional data by learning low-rank factors to represent high-order interactions. While recent works on temporal tensor decomposition have made significant progress by incorporating continuous timestamps in latent factors, they still struggle with general tensor data with continuous indexes not only in the temporal mode but also in other modes, such as spatial coordinates in climate data. Moreover, the challenge of self-adapting model complexity is largely unexplored in functional temporal tensor models, with existing methods being inapplicable in this setting. To address these limitations, we propose functional Complexity-Adaptive Temporal Tensor dEcomposition (Catte). Our approach encodes continuous spatial indexes as learnable Fourier features and employs neural ODEs in latent space to learn the temporal trajectories of factors. To enable automatic adaptation of model complexity, we introduce a sparsity-inducing prior over the factor trajectories. We develop an efficient variational inference scheme with an analytical evidence lower bound, enabling sampling-free optimization. Through extensive experiments on both synthetic and real-world datasets, we demonstrate that Catte not only reveals the underlying ranks of functional temporal tensors but also significantly outperforms existing methods in prediction performance and robustness against noise.



Paperid:1202
Authors:Andrew Zhao, Yiran Wu, Yang Yue, Tong Wu, Quentin Xu, Yang Yue, Matthieu Lin, Shenzhi Wang, Qingyun Wu, Zilong Zheng, Gao Huang
Abstract:
Reinforcement learning with verifiable rewards (RLVR) has shown promise in enhancing the reasoning capabilities of large language models by learning directly from rule-based outcome rewards. Recent RLVR works that operate under the zero setting avoid supervision in labeling the reasoning process, but still depend on manually curated collections of questions and answers for training. The scarcity of high-quality, human-produced examples raises concerns about the long-term scalability of relying on human supervision, a challenge already evident in the domain of language model pretraining. Furthermore, in a hypothetical future where AI surpasses human intelligence, tasks provided by humans may offer limited learning potential for a superintelligent system. To address these concerns, we propose a new RLVR paradigm called Absolute Zero, in which a single model learns to propose tasks that maximize its own learning progress and improves reasoning by solving them, without relying on any external human or distillation data. Under this paradigm, we introduce the Absolute Zero Reasoner (AZR), a system that self-evolves its training curriculum and reasoning ability. AZR uses a code executor to both validate self-proposed code reasoning tasks and verify answers, serving as an unified source of verifiable feedback to guide open-ended yet grounded learning. Despite being trained entirely without external data, AZR achieves overall SOTA performance on coding and mathematical reasoning tasks, outperforming existing zero-setting models that rely on tens of thousands of in-domain human-curated examples. Furthermore, we demonstrate that AZR can be effectively applied across different model scales and is compatible with various model classes.



Paperid:1203
Authors:Ali Behrouz, Meisam Razaviyayn, Peilin Zhong, Vahab Mirrokni
Abstract:
Developing more powerful neural architectures has been in the core of research studies to enhance the capability of foundation models. There has also been an extensive research effort on how to effectively train such architectures using optimization algorithms. While these two components and their roles in modern foundation models are decoupled in the literature, in this paper, we show that how inexplicably these two levels have been separated and introduce Nested Learning (NL) paradigm that represents a model with a set of nested optimization problems. We introduce \emph{Neural Learning Module} that sees the entire process of training a neural architecture and its optimization process as a single integrated system, optimizing a set of objectives at different levels. While having meta-learning as the special instance of nested learning (bi-level), we show how NL goes further by letting (i) the model have more nested layers; and (ii) optimization blocks be placed simultaneously on the same level, addressing different aspects of learning. Taking advantage of the concept of NL, we present a family of nested optimizers based on gradient descent that unifies, explains, and extends existing optimizers, including those with momentum and pre-conditioning techniques. NL provides a mathematically white box deep models, further enhancing the interpretability, and explain how an upper-level optimization block is learning the algorithms in its inner objectives. Building upon this insight, we present Hope, a neural learning module that is dynamically updating its weights using a single internal objective, adapting its memory based on its past. Our experimental results on NL paradigm, general optimizers, and Hope indicates: (1) the effectiveness of more nested levels; (2) the superior or on-par performance of general optimizers compared to existing well-established methods; and (3) the good performance of Hope neural learning module in tasks such as language modeling, common-sense reasoning, and recall-intensive tasks.



Paperid:1204
Authors:Dajun Sun, Wei Dong, Yuan Qiu, Ke Yi, Graham Cormode
Abstract:
People have diverse privacy requirements. This is best modeled using a personalized local differential privacy model where each user privatizes their data using a possibly different privacy parameter. While the model of personalized local differential privacy is a natural and important one, prior work has failed to give meaningful error bounds. In this paper, we study the foundational sum/mean estimation problem under this model. We present two novel protocols that achieve strong error guarantees. The first gives a guarantee based on the radius of the data, suiting inputs that are centered around zero. The second extends the guarantee to the diameter of the data, capturing the case when the points are situated arbitrarily. Experimental results on both synthetic and real data show that our protocols significantly outperform existing methods in terms of accuracy while providing a strong level of privacy.



Paperid:1205
Authors:Yewon Byun, Shantanu Gupta, Zachary Lipton, Rachel Childers, Bryan Wilder
Abstract:
Predictions and generations from large language models are increasingly being explored as an aid to computational social science and human subject research. One common hope is to augment data in cases where labeling data is expensive, leading to imprecise estimates of statistical parameters. While previous technical work has explored the potential to use model-predicted labels for unlabeled data in a principled manner, there is increasing interest in using large language models to generate entirely new synthetic samples, such as in responses to surveys or in human behavior simulation. However, it is not immediately clear by what means practitioners can incorporate such data and yet produce statistically valid conclusions upon them. In this work, we introduce a new estimator, providing a hyperparameter-free solution with strong theoretical guarantees to address the challenge at hand. We empirically validate the finite-sample performance of our estimator across different regression tasks in computational social science applications. To the best of our knowledge, our framework provides the first theoretically-sound approach for incorporating fully synthetic samples in downstream statistical analyses.



Paperid:1206
Authors:Shawn Im, Sharon Li
Abstract:
Large language models (LLMs) have demonstrated remarkable capabilities but often struggle to align with human preferences, leading to harmful or undesirable outputs. Preference learning, which trains models to distinguish between preferred and non-preferred responses based on human feedback, has become a crucial component for ensuring that LLMs align with human values. An essential part of ensuring that LLMs are aligned for all people is accounting for a diverse set of values. This paper introduces a new theoretical framework to analyze how generalization scales with value diversity and sample quantity in models trained with direct preference optimization. Our framework rigorously assesses how well models generalize after a finite number of gradient steps, reflecting real-world LLM training practices. By analyzing the reward margin associated with each sample and its trajectory throughout training, we provide a bound on the generalization error that demonstrates the challenges of effectively learning a wide set of concepts or values. These insights are empirically validated on contemporary LLMs, underscoring the practical relevance of our theory.



Paperid:1207
Authors:Guibin Zhang, Muxin Fu, Guancheng Wan, Miao Yu, Kun Wang, Shuicheng Yan
Abstract:
Abstract:Large language model (LLM)-powered multi-agent systems (MAS) have demonstrated cognitive and execution capabilities that far exceed those of single LLM agents, yet their capacity for self-evolution remains hampered by underdeveloped memory architectures. Upon close inspection, we are alarmed to discover that prevailing MAS memory mechanisms (1) are overly simplistic, completely disregarding the nuanced inter-agent collaboration trajectories, and (2) lack cross-trial and agent-specific customization, in stark contrast to the expressive memory developed for single agents. To bridge this gap, we introduce G-Memory, a hierarchical, agentic memory system for MAS inspired by organizational memory theory, which manages the lengthy MAS interaction via a three-tier graph hierarchy: insight, query, and interaction graphs. Upon receiving a new user query, G-Memory performs bi-directional memory traversal to retrieve both \textit{high-level, generalizable insights} that enable the system to leverage cross-trial knowledge, and \textit{fine-grained, condensed interaction trajectories} that compactly encode prior collaboration experiences. Upon task execution, the entire hierarchy evolves by assimilating new collaborative trajectories, nurturing the progressive evolution of agent teams. Extensive experiments across five benchmarks, three LLM backbones, and three popular MAS frameworks demonstrate that G-Memory improves success rates in embodied action and accuracy in knowledge QA by up to $20.89\\%$ and $10.12\\%$, respectively, without any modifications to the original frameworks.



Paperid:1208
Authors:Shaomin Li
Abstract:
We study the mediation analysis under the distributed framework, where data are stored and processed across different worker machines due to storage limitations or privacy concerns. Building upon the classic Sobel's test and MaxP test, we introduce the distributed Sobel's test and distributed MaxP test, respectively. These tests are both communication-efficient and easy to implement. Theoretical analysis and numerical experiments show that, compared to the global test obtained by pooling all data together, the proposed tests achieve nearly identical power, independent of the number of machines. Furthermore, based on these two distributed test statistics, many enhanced mediation tests derived from the Sobel's or MaxP tests can be easily adapted to the distributed system. We apply our method to an educational study, testing whether the effect of high school mathematics on college-level Probability and Mathematical Statistics courses is mediated by Calculus. Our method successfully detects the mediation effect, which would not be identifiable using data from only the first or second class, highlighting the advantage of our approach.



Paperid:1209
Authors:Runjia Zeng, James Liang, Cheng Han, Zhiwen Cao, Jiahao Liu, Xiaojun Quan, Yingjie Victor Chen, Lifu Huangg, Tong Geng, Qifan Wang, Dongfang Liu
Abstract:
Training large language models (LLMs) from scratch can yield models with unique functionalities and strengths, but it is costly and often leads to redundant capabilities. A more cost-effective alternative is to fuse existing pre-trained LLMs with different architectures into a more powerful model. However, a key challenge in existing model fusion is their dependence on manually predefined vocabulary alignment, which may not generalize well across diverse contexts, leading to performance degradation in several evaluation. To solve this, we draw inspiration from distribution learning and propose the probabilistic token alignment method as a general and soft mapping for alignment, named as PTA-LLM. Our approach innovatively reformulates token alignment into a classic mathematical problem: optimal transport, seamlessly leveraging distribution-aware learning to facilitate more coherent model fusion. Apart from its inherent generality, PTA-LLM exhibits interpretability from a distributional perspective, offering insights into the essence of the token alignment. Empirical results demonstrate that probabilistic token alignment enhances the target model's performance across multiple capabilities.



Paperid:1210
Authors:Adrien Aumon, Shuang Ni, Myriam Lizotte, Guy Wolf, Kevin Moon, Jake Rhodes
Abstract:
Extensive research has produced robust methods for unsupervised data visualization. Yet supervised visualization—where expert labels guide representations—remains underexplored, as most supervised approaches prioritize classification over visualization. Recently, RF-PHATE, a diffusion-based manifold learning method leveraging random forests and information geometry, marked significant progress in supervised visualization. However, its lack of an explicit mapping function limits scalability and its application to unseen data, posing challenges for large datasets and label-scarce scenarios. To overcome these limitations, we introduce Random Forest Autoencoders (RF-AE), a neural network-based framework for out-of-sample kernel extension that combines the flexibility of autoencoders with the supervised learning strengths of random forests and the geometry captured by RF-PHATE. RF-AE enables efficient out-of-sample supervised visualization and outperforms existing methods, including RF-PHATE's standard kernel extension, in both accuracy and interpretability. Additionally, RF-AE is robust to the choice of hyperparameters and generalizes to any kernel-based dimensionality reduction method.



Paperid:1211
Authors:Yuxuan Zhou, Heng Li, Zhi-Qi Cheng, Xudong Yan, Yifei Dong, Mario Fritz, Margret Keuper
Abstract:
Label Smoothing (LS) is widely adopted to mitigate over-confidence and improve generalization, yet accumulating evidence reveals two fundamental pathologies. (i) LS contracts feature embeddings into excessively compact class clusters, obscuring fine-grained intra-class structure; (ii) when the prediction is incorrect, LS paradoxically reinforces the wrong class, encouraging unwarranted confidence in errors. We derive an exact decomposition of the LS objective, demonstrating that the additional term decomposes into a regularization component—beneficial exclusively for correct predictions—and an error-amplification component—activated precisely for incorrect predictions. This single mechanism accounts for both representation collapse and mistaken over-confidence. Guided by this analysis, we introduce Max Suppression (MaxSup), a drop-in alternative that penalizes the top-1 logit instead of the ground-truth logit, thereby preserving the beneficial regularization while eliminating error amplification. Experiments on CIFAR-10/100, ImageNet-1k, and several transfer benchmarks show that MaxSup consistently increases intra-class dispersion, sharpens inter-class margins, and improves top-1 accuracy by up to 1.4 percentage points on ImageNet-1k, all with negligible computational overhead. Code and reproducibility scripts are available at https://anonymous.4open.science/r/Maximum-Suppression-Regularization-DB0C.



Paperid:1212
Authors:Jonathan Zheng, Alan Ritter, Sauvik Das, Wei Xu
Abstract:
Abstract:Probabilistic reasoning is a key aspect of both human and artificial intelligence that allows for handling uncertainty and ambiguity in decision-making. In this paper, we introduce a new numerical reasoning task under uncertainty for large language models, focusing on estimating the privacy risk of user-generated documents containing privacy-sensitive information. We propose BRANCH, a new LLM methodology that estimates the $k$-privacy value of a text—the size of the population matching the given information. BRANCH factorizes a joint probability distribution of personal information as random variables. The probability of each factor in a population is estimated separately using a Bayesian network and combined to compute the final $k$-value. Our experiments show that this method successfully estimates the $k$-value 73% of the time, a 13% increase compared to o3-mini with chain-of-thought reasoning. We also find that LLM uncertainty is a good indicator for accuracy, as high variance predictions are 37.47% less accurate on average.



Paperid:1213
Authors:Yan Zhuang, Chenye Ke, Zirui Liu, Qi Liu, Yuting Ning, Zhenya Huang, Weizhe Huang, Qingyang Mao, Shijin Wang
Abstract:
Human ability estimation is essential for educational assessment, career advancement, and professional certification. Adaptive Testing systems can improve estimation efficiency by selecting fewer, targeted questions, and are widely used in exams, e.g., GRE, GMAT, and Duolingo English Test. However, selecting an optimal subset of questions remains a challenging nested optimization problem. Existing methods rely on costly approximations or data-intensive training, making them unsuitable for today's large-scale and complex testing environments. Thus, we propose a Closed-Form solution for question subset selection in Adaptive Testing. It directly minimizes ability estimation error by reducing ability parameter's gradient bias while maintaining Hessian stability, which enables a simple greedy algorithm for question selection. Moreover, it can quantify the impact of human behavioral perturbations on ability estimation. Extensive experiments on large-scale educational datasets demonstrate that it reduces the number of required questions by 10% compared to SOTA methods, while maintaining the same estimation accuracy.



Paperid:1214
Authors:Alessandro Trenta, Alessio Gravina, Davide Bacciu
Abstract:
Capturing effective long-range information propagation remains a fundamental yet challenging problem in graph representation learning. Motivated by this, we introduce SONAR, a novel GNN architecture inspired by the dynamics of wave propagation in continuous media. SONAR models information flow on graphs as oscillations governed by the wave equation, allowing it to maintain effective propagation dynamics over long distances. By integrating adaptive edge resistances and state-dependent external forces, our method balances conservative and non-conservative behaviors, improving the ability to learn more complex dynamics. We provide a rigorous theoretical analysis of SONAR's energy conservation and information propagation properties, demonstrating its capacity to address the long-range propagation problem. Extensive experiments on synthetic and real-world benchmarks confirm that SONAR achieves state-of-the-art performance, particularly on tasks requiring long-range information exchange.



Paperid:1215
Authors:Qirui Yang, Yihao Liu, Yinbo Li, Peng-Tao Jiang, Fangpu Zhang, cheng qihua, Huanjing Yue, Jingyu Yang
Abstract:
Existing tone mapping methods operate on downsampled inputs and rely on handcrafted pyramids to recover high-frequency details. Existing tone mapping methods operate on downsampled inputs and rely on handcrafted pyramids to recover high-frequency details. These designs typically fail to preserve fine textures and structural fidelity in complex HDR scenes. Furthermore, most methods lack an effective mechanism to jointly model global tone consistency and local contrast enhancement, leading to globally flat or locally inconsistent outputs such as halo artifacts. We present the Differential Pyramid Representation Network (DPRNet), an end-to-end framework for high-fidelity tone mapping. At its core is a learnable differential pyramid that generalizes traditional Laplacian and Difference-of-Gaussian pyramids through content-aware differencing operations across scales. This allows DPRNet to adaptively capture high-frequency variations under diverse luminance and contrast conditions. To enforce perceptual consistency, DPRNet incorporates global tone perception and local tone tuning modules operating on downsampled inputs, enabling efficient yet expressive tone adaptation. Finally, an iterative detail enhancement module progressively restores the full-resolution output in a coarse-to-fine manner, reinforcing structure and sharpness. To support training and benchmarking, we introduce a new tone mapping dataset with diverse real-world scenes and lighting conditions. Experiments show that DPRNet achieves state-of-the-art results, improving PSNR by2.39 dBon the 4K HDR+ dataset and3.01 dBon the 4K HDRI Haven dataset, while producing perceptually coherent, detail-preserving results. Demo available atDPRNet.



Paperid:1216
Authors:Tingxu Han, Weisong Sun, Yanrong Hu, Chunrong Fang, Yonglong zhang, Shiqing Ma, Tao Zheng, Zhenyu Chen, Zhenting Wang
Abstract:
Text-to-image diffusion models have shown an impressive ability to generate high-quality images from input textual descriptions/prompts. However, concerns have been raised about the potential for these models to create content that infringes on copyrights or depicts disturbing subject matter.Removing specific concepts from these models is a promising solution to this issue. However, existing methods for concept removal do not work well in practical but challenging scenarios where concepts need to be continuously removed. Specifically, these methods lead to poor alignment between the text prompts and the generated image after the continuous removal process.To address this issue, we propose a novel concept removal approach called CCRT that includes a designed knowledge distillation paradigm.CCRT constrains the text-image alignment behavior during the continuous concept removal process by using a set of text prompts.These prompts are generated through our genetic algorithm, which employs a designed fuzzing strategy. To evaluate the effectiveness of CCRT, we conduct extensive experiments involving the removal of various concepts, algorithmic metrics, and human studies.The results demonstrate that CCRT can effectively remove the targeted concepts from the model in a continuous manner while maintaining the high image generation quality (e.g., text-image alignment).The code of CCRT is available at https://anonymous.4open.science/r/CCRT-F3EE.



Paperid:1217
Authors:Saleh Ashkboos, Mahdi Nikdan, Rush Tabesh, Roberto Castro, Torsten Hoefler, Dan Alistarh
Abstract:
Quantized training of Large Language Models (LLMs) remains an open challenge, as maintaining accuracy while performing all matrix multiplications in low precision has proven difficult. This is particularly the case when fine-tuning pre-trained models, which can have large weight, activation, and error (output gradient) outlier values that make lower-precision optimization difficult. To address this, we present HALO, a new quantization-aware training approach for Transformers that enables accurate and efficient low-precision training by combining 1) strategic placement of Hadamard rotations in both forward and backward passes, which mitigate outliers, 2) high-performance kernel support, and 3) FSDP integration for low-precision communication. Our approach ensures that all large matrix multiplications during the forward and backward passes are executed in lower precision. Applied to LLaMa models, HALO achieves near-full-precision-equivalent results during fine-tuning on various tasks, while delivering up to 1.41x end-to-end speedup for full fine-tuning on RTX 4090 GPUs. HALO efficiently supports both standard and parameter-efficient fine-tuning (PEFT). Our results demonstrate the first practical approach to fully quantized LLM fine-tuning that maintains accuracy in INT8 and FP6 precision, while delivering performance benefits.



Paperid:1218
Authors:Amirhesam Abedsoltan, Siyuan Ma, Parthe Pandit, Misha Belkin
Abstract:
Classical kernel machines have historically faced significant challenges in scaling to large datasets and model sizes—a key ingredient that has driven the success of neural networks. In this paper, we present a new methodology for building kernel machines that can scale efficiently with both data size and model size. Our algorithm introduces delayed projections to Preconditioned Stochastic Gradient Descent (PSGD) allowing the training of much larger models than was previously feasible. We validate our algorithm, \EP4, across multiple datasets, demonstrating drastic training speedups without compromising the performance.



Paperid:1219
Authors:Johannes Hertrich, Antonin Chambolle, Julie Delon
Abstract:
This paper investigates the connections between rectified flows, flow matching, and optimal transport. Flow matching is a recent approach to learning generative models by estimating velocity fields that guide transformations from a source to a target distribution. Rectified flow matching aims to straighten the learned transport paths, yielding more direct flows between distributions. Our first contribution is a set of invariance properties of rectified flows and explicit velocity fields. In addition, we also provide explicit constructions and analysis in the Gaussian (not necessarily independent) and Gaussian mixture settings and study the relation to optimal transport. Our second contribution addresses recent claims suggesting that rectified flows, when constrained such that the learned velocity field is a gradient, can yield (asymptotically) solutions to optimal transport problems. We study the existence of solutions for this problem and demonstrate that they only relate to optimal transport under assumptions that are significantly stronger than those previously acknowledged. In particular, we present several counter-examples that invalidate earlier equivalence results in the literature, and we argue that enforcing a gradient constraint on rectified flows is, in general, not a reliable method for computing optimal transport maps.



Paperid:1220
Authors:Runqing Miao, Sheng Xu, Runyi Zhao, Wai Kin (Victor) Chan, Guiliang Liu
Abstract:
Recent advancements in diffusion policies have demonstrated promising performance in decision-making tasks. To align these policies with human preferences, a common approach is incorporating Preference-based Reinforcement Learning (PbRL) into policy tuning. However, since preference data is practically collected from populations with different backgrounds, a key challenge lies in handling the inherent uncertainties in people's preferences during policy updates. To address this challenge, we propose the Diff-UAPA algorithm, designed for uncertainty-aware preference alignment in diffusion policies. Specifically, Diff-UAPA introduces a novel iterative preference alignment framework in which the diffusion policy adapts incrementally to preferences from different user groups. To accommodate this online learning paradigm, Diff-UAPA employs a maximum posterior objective, which aligns the diffusion policy with regret-based preferences under the guidance of an informative Beta prior. This approach enables direct optimization of the diffusion policy without specifying any reward functions, while effectively mitigating the influence of inconsistent preferences across different user groups. We conduct extensive experiments across both simulated and real-world robotics tasks, and diverse human preference configurations, demonstrating the robustness and reliability of Diff-UAPA in achieving effective preference alignment. The project page is available at https://anonymous.4open.science/w/Diff-UAPA-5827.



Paperid:1221
Authors:Moritz Miller, Bernhard Schölkopf, Siyuan Guo
Abstract:
Large-scale neural language models (LMs) exhibit remarkable performance in in-context learning: the ability to learn and reason the input context on the fly without parameter update. This work studies in-context counterfactual reasoning in language models, that is, to predict the consequences of changes under hypothetical scenarios. We focus on studying a well-defined synthetic setup: a linear regression task that requires noise abduction, where accurate prediction is based on inferring and copying the contextual noise from factual observations. We show that language models are capable of counterfactual reasoning in this controlled setup and provide insights that counterfactual reasoning for a broad class of functions can be reduced to a transformation on in-context observations; we find self-attention, model depth, and data diversity in pre-training drive performance in Transformers. More interestingly, our findings extend beyond regression tasks and show that Transformers can perform noise abduction on sequential data, providing preliminary evidence on the potential for counterfactual story generation.



Paperid:1222
Authors:Wanhua Li, Yujie Zhao, Minghan Qin, Yang Liu, Yuanhao Cai, Chuang Gan, Hanspeter Pfister
Abstract:
In this paper, we introduce LangSplatV2, which achieves high-dimensional featuresplatting at 476.2 FPS and 3D open-vocabulary text querying at 384.6 FPS forhigh-resolution images, providing a 42 × speedup and a 47 × boost over LangSplatrespectively, along with improved query accuracy. LangSplat employs GaussianSplatting to embed 2D CLIP language features into 3D, significantly enhancingspeed and learning a precise 3D language field with SAM semantics. Such advancements in 3D language fields are crucial for applications that require languageinteraction within complex scenes. However, LangSplat does not yet achieve real-time performance (8.2 FPS), even with advanced A100 GPUs, severely limitingits broader application. In this paper, we first conduct a detailed time analysis ofLangSplat, identifying the heavyweight decoder as the primary speed bottleneck.Our solution, LangSplatV2 assumes that each Gaussian acts as a sparse code withina global dictionary, leading to the learning of a 3D sparse coefficient field thatentirely eliminates the need for a heavyweight decoder. By leveraging this sparsity,we further propose an efficient sparse coefficient splatting method with CUDA optimization, rendering high-dimensional feature maps at high quality while incurringonly the time cost of splatting an ultra-low-dimensional feature. Our experimentalresults demonstrate that LangSplatV2 not only achieves better or competitive queryaccuracy but is also significantly faster.



Paperid:1223
Authors:Xueqing Sun, Renzhen Wang, Quanziang Wang, Yichen WU, Xixi Jia, Deyu Meng
Abstract:
Pseudo-labeling is a commonly used paradigm in semi-supervised learning, yet its application to semi-supervised regression (SSR) remains relatively under-explored. Unlike classification, where pseudo-labels are discrete and confidence-based filtering is effective, SSR involves continuous outputs with heteroscedastic noise, making it challenging to assess pseudo-label reliability. As a result, naive pseudo-labeling can lead to error accumulation and overfitting to incorrect labels. To address this, we propose an uncertainty-aware pseudo-labeling framework that dynamically adjusts pseudo-label influence from a bi-level optimization perspective. By jointly minimizing empirical risk over all data and optimizing uncertainty estimates to enhance generalization on labeled data, our method effectively mitigates the impact of unreliable pseudo-labels. We provide theoretical insights and extensive experiments to validate our approach across various benchmark SSR datasets, and the results demonstrate superior robustness and performance compared to existing methods.



Paperid:1224
Authors:Xin Du, Kumiko Tanaka-Ishii
Abstract:
Large language models (LLMs) have achieved impressive results in naturallanguage generation but continue to exhibit unexplained behaviors—such ashallucinations, repetition, and incoherence—even at low perplexity. This exposesthe limitations of traditional evaluation metrics, which focus on localprediction accuracy while neglecting long-range structural complexity. Weintroduce the correlation dimension, a fractal-geometric measure ofself-similarity, to quantify the implicit textual complexity perceived bylanguage models. Correlation dimension robustly captures the hierarchicalrecurrence structure of text, bridging local and global properties. Experimentsdemonstrate that this measure reveals distinct training phases, reflectscontext-dependent complexity, and reliably detects various forms of degenerationin generated text. The method is computationally efficient, broadly applicableto autoregressive models, and provides a theoretically grounded approach forunderstanding complex generative dynamics of LLMs.



Paperid:1225
Authors:Adam Dankowiakowski, Alessandro Ronca
Abstract:
We propose Metric Automata Theory, an elegant generalisation of classic Automata Theory to continuous dynamical systems, that constitutes a unifying theory of all kinds of Recurrent Neural Networks (RNNs), including widely-adopted architectures such as xLSTM and State Space Models (SSMs). The theory allows one to analyse RNNs both in the finite and unbounded precision settings seamlessly, and to use the fundamental results of Automata Theory. It also provides a novel notion of robustness that guarantees numerical stability, contributes to stability of learning, and notably allows one to prove results for real-world finite-precision implementations while abstracting away the difficulties introduced by finite-precision arithmetic.We employ the theory to prove a comprehensive set of expressivity results for widely-adopted recurrent neural architectures, with a focus on robustness and finite-precision. Notably, we prove that xLSTM, based on non-linear recurrence, is capable of recognising all star-free regular languages robustly, and hence by our theory it can also do so under any floating-point implementation given sufficient precision. Furthermore, we prove that linear recurrence, characterising SSMs, is not sufficient to robustly recognise all star-free languages. This provides a potential explanation for why xLSTM shows superior performance to SSMs on several tasks, and it gives a novel perspective on the importance of non-linear recurrences.We provide an improved understanding of the capabilities of Mamba, a popular SSM model. We show that Mamba is not generally capable of recognising the star-free languages under finite-precision, which is seemingly in contrast with the existing expressivity results for SSMs and also with its good empirical performance. We clarify the picture, by showing that Mamba admits a piecewise-linearly separable state space that allows it to approximate star-free languages, with some length-generalisation abilities. At the same time, Mamba does not admit such state spaces for languages like Parity. This explains why empirically Mamba performs well on star-free languages, and fails on Parity.



Paperid:1226
Authors:Eshika Saxena, Alberto Alfarano, Francois Charton, Emily Wenger, Kristin E. Lauter
Abstract:
Abstract:AI-powered attacks on Learning with Errors (LWE)—an important hard math problem in post-quantum cryptography—rival or outperform "classical" attacks on LWE under certain parameter settings. Despite the promise of this approach, a dearth of accessible data limits AI practitioners' ability to study and improve these attacks. Creating LWE data for AI model training is time- and compute-intensive and requires significant domain expertise. To fill this gap and accelerate AI research on LWE attacks, we propose the TAPAS datasets, a ${\bf t}$oolkit for ${\bf a}$nalysis of ${\bf p}$ost-quantum cryptography using ${\bf A}$I ${\bf s}$ystems. These datasets cover several LWE settings and can be used off-the-shelf by AI practitioners to prototype new approaches to cracking LWE. This work documents TAPAS dataset creation, establishes attack performance baselines, and lays out directions for future work.



Paperid:1227
Authors:Yuang Meng, Xin Jin, Lina Lei, Chun-Le Guo, Chongyi Li
Abstract:
Ultra-high dynamic range (UHDR) scenes exhibit pronounced exposure disparities between bright and dark regions. Such conditions are common in nighttime scenes with light sources. Even standard exposure settings often result in a bimodal intensity distribution with boundary peaks, making it challenging to simultaneously preserve both highlight and shadow details. RGB-based bracketing methods can capture details at both ends using short-long exposure pairs, but are susceptible to misalignment and ghosting artifacts. A short-exposure image, however, already retains sufficient highlight detail. The main challenge lies in denoising and recovering information in dark regions. RAW images, thanks to their higher bit depth and more predictable noise characteristics, offer greater potential for addressing this challenge. This raises a key question: can we learn to see everything in UHDR scenes using only a single short-exposure RAW image? Our method, relying solely on one short-exposure frame, inherently avoids ghosting and motion blur, making it particularly robust in dynamic scenes. To achieve that, we introduce a two-stage framework: exposure correction via a ratio map to balance dynamic range, followed by brightness-aware noise modeling to enhance detail recovery in dark regions. To support this, we design a 9-stop bracketing pipeline to synthesize realistic UHDR images, and construct a dataset accordingly on static scenes, using only the shortest exposure as input for reconstruction. Experiments show that our method significantly outperforms existing single-frame approaches. Code will be released publicly.



Paperid:1228
Authors:Fan LIU, Wen-Shuo Chao, Naiqiang Tan, Hao Liu
Abstract:
With the advancement of large language models (LLMs), solving complex tasks (e.g., math problems, code generation, etc.) has garnered increasing attention. Inference-time computation methods (e.g., Best-of-N, MCTS, etc.) are of significant importance, as they have the potential to enhance the reasoning capabilities of LLMs without requiring external training computation. However, due to the inherent challenges of this technique, most existing methods remain proof-of-concept and are not yet sufficiently effective. In this paper, we investigate and benchmark strategies for improving inference-time computation across a wide range of reasoning tasks. Since most current methods rely on a pipeline that first generates candidate solutions (e.g., generating chain-of-thought candidate solutions) and then selects them based on specific reward signals (e.g., RLHF reward, process reward, etc.), our research focuses on strategies for both candidate solution generation (e.g., instructing prompts, hyperparameters: temperature and top-p, etc.) and reward mechanisms (e.g., self-evaluation, reward types, etc.). The experimental results reveal that several previously overlooked strategies can be critical for the success of inference-time computation (e.g., simplifying the temperature can improve general reasoning task performance by up to 5%). Based on extensive experiments (more than 20,000 A100-80G GPU hours with over 1,000 experiments) across a variety of models (e.g., Llama, Qwen, and Mistral families) of various sizes, our proposed strategies outperform the baseline by a substantial margin in most cases, providing a stronger foundation for future research.



Paperid:1229
Authors:Hikaru Asano, Tadashi Kozuno, Yukino Baba
Abstract:
Recent advances in large language models (LLMs) have yielded impressive performance on various tasks, yet they often depend on high-quality feedback that can be costly. Self-refinement methods attempt to leverage LLMs' internal evaluation mechanisms with minimal human supervision; however, these approaches frequently suffer from inherent biases and overconfidence—especially in domains where the models lack sufficient internal knowledge—resulting in performance degradation. As an initial step toward enhancing self-refinement for broader applications, we introduce an iterative refinement pipeline that employs the Unlabeled-Unlabeled learning framework to improve LLM-generated pseudo-labels for classification tasks. By exploiting two unlabeled datasets with differing positive class ratios, our approach iteratively denoises and refines the initial pseudo-labels, thereby mitigating the adverse effects of internal biases with minimal human supervision. Evaluations on diverse datasets—including low-resource language corpora, patent classifications, and protein structure categorizations—demonstrate that our method consistently outperforms both initial LLM's classification performance and the self-refinement approaches by cutting-edge models (e.g., GPT-4o and DeepSeek-R1). Moreover, we experimentally confirm that our refined classifier facilitates effective post-training alignment for safety in LLMs and demonstrate successful self-refinement in generative tasks as well.



Paperid:1230
Authors:Sayan Deb Sarkar, Sinisa Stekovic, Vincent Lepetit, Iro Armeni
Abstract:
Transferring appearance to 3D assets using different representations of the appearance object--such as images or text--has garnered interest due to its wide range of applications in industries like gaming, augmented reality, and digital content creation. However, state-of-the-art methods still fail when the geometry between the input and appearance objects is significantly different. A straightforward approach is to directly apply a 3D generative model, but we show that this ultimately fails to produce appealing results. Instead, we propose a principled approach inspired by universal guidance. Given a pretrained rectified flow model conditioned on image or text, our training-free method interacts with the sampling process by periodically adding guidance. This guidance can be modeled as a differentiable loss function, and we experiment with two different types of guidance including part-aware losses for appearance and self-similarity. Our experiments show that our approach successfully transfers texture and geometric details to the input 3D asset. We outperform baselines both qualitatively and quantitatively. Traditional metrics are not suitable for evaluating the task due to inability of focusing on local details and comparing dissimilar inputs, in absence of ground truth data. We thus evaluate appearance transfer quality with a GPT-based system objectively ranking outputs, ensuring robust and human-like assessment. Beyond showcased scenarios, our method is general and could be extended to different types of diffusion models and guidance functions.



Paperid:1231
Authors:Liu Ziyin, Yizhou Xu, Isaac Chuang
Abstract:
As more and more emergent phenomena are being discovered in deep learning and large-language models, explaining and understanding their cause has become an urgent need. Here, we propose a rigorous entropic-force theory for understanding the learning dynamics of neural networks trained with stochastic gradient descent (SGD) and its variants. Building on the theory of parameter symmetries and an entropic loss landscape, we show that representation learning is crucially governed by emergent entropic forces arising from stochasticity and discrete-time updates. These forces systematically break continuous parameter symmetries and preserve discrete ones, leading to a series of gradient balance phenomena that resemble the equipartition property of thermal systems. These phenomena, in turn, explain the universal alignment of neural representations between AI models and the seemingly contradictory observations of progressive sharpening and flattening during deep learning optimization. We demonstrate that a combination of entropic forces and symmetry breaking is key to understanding emergent phenomena in deep learning.



Paperid:1232
Authors:Kalinin Nikita, Jalaj Upadhyay, Christoph Lampert
Abstract:
We proposeJoint Moment Estimation(JME), a method for continually and privately estimating both the first and second moments of a data stream with reduced noise compared to naive approaches. JME supports thematrix mechanismand exploits a joint sensitivity analysis to identify a privacy regime in which the second-moment estimation incurs no additional privacy cost, thereby improving accuracy while maintaining privacy. We demonstrate JME’s effectiveness in two applications: estimating the running mean and covariance matrix for Gaussian density estimation and model training with DP-Adam.



Paperid:1233
Authors:Xi Chen, Mingkang Zhu, Shaoteng Liu, Xiaoyang Wu, Xiaogang Xu, Yu Liu, Xiang Bai, Hengshuang Zhao
Abstract:
This work explores enabling Chain-of-Thought (CoT) reasoning to link visual cues across multiple images. A straightforward solution is to adapt rule-based reinforcement learning for Vision-Language Models (VLMs). However, such methods typically rely on manually curated question-answer pairs, which can be particularly challenging when dealing with fine-grained visual details and complex logic across images. Inspired by self-supervised visual representation learning, we observe that images contain inherent constraints that can serve as supervision. Based on this insight, we construct image triplets comprising two augmented views of the same image and a third, similar but distinct image. During training, the model is prompted to generate a reasoning process to compare these images (i.e., determine same or different). Then we optimize the model with rule-based reinforcement learning. Due to the high visual similarity and the presence of augmentations, the model must attend to subtle visual cues and perform logical reasoning to succeed. Experimental results demonstrate that, although trained solely on visual comparison tasks, the learned reasoning ability generalizes effectively to a wide range of questions. Without relying on any human-annotated question-answer pairs, our method achieves significant improvements on multi-image reasoning benchmarks and shows strong performance on general vision tasks.



Paperid:1234
Authors:Annie Marsden, Liam O'Carroll, Aaron Sidford, Chenyi Zhang
Abstract:
Abstract:We consider the problem of minimizing a $d$-dimensional Lipschitz convex function using a stochastic gradient oracle. We introduce and motivate a setting where the noise of the stochastic gradient is \emph{isotropic} in that it is bounded in every direction with high probability. We then develop an algorithm for this setting which improves upon prior results by a factor of $d$ in certain regimes, and as a corollary, achieves a new state-of-the-art complexity for sub-exponential noise. We give matching lower bounds (up to polylogarithmic factors) for both results. Additionally, we develop an efficient \emph{quantum isotropifier}, a quantum algorithm which converts a variance-bounded quantum sampling oracle into one that outputs an unbiased estimate with isotropic error. Combining our results, we obtain improved dimension-dependent rates for quantum stochastic convex optimization.



Paperid:1235
Authors:Yinjie Min, Chuchen Zhang, Liuhua Peng, Changliang Zou
Abstract:
Personalized federated learning addresses data heterogeneity across distributed agents but lacks uncertainty quantification that is both agent-specific and instance-specific, which is a critical requirement for risk-sensitive applications. We propose personalized federated conformal prediction (PFCP), a novel framework that combines personalized federated learning with conformal prediction to provide statistically valid agent-personalized prediction sets with instance-localization. By leveraging privacy-preserving knowledge transfer from other source agents, PFCP ensures marginal coverage guarantees for target agents while significantly improving conditional coverage performance on individual test instances, which has been validated by extensive experiments.



Paperid:1236
Authors:Shay Moran, Elizaveta Nesterova
Abstract:
Consider the task of locating an unknown target point using approximate distance queries: in each round, a reconstructor selects a reference point and receives a noisy version of its distance to the target. This problem arises naturally in various contexts—from localization in GPS and sensor networks to privacy-aware data access—making it relevant from the perspective of both the reconstructor (seeking accurate recovery) and the responder (aiming to limit information disclosure, e.g., for privacy or security reasons). We study this reconstruction game through a learning-theoretic lens, focusing on the rate and limits of the best possible reconstruction error.Our first result provides a tight geometric characterization of the optimal error in terms of the Chebyshev radius, a classical concept from geometry. This characterization applies to all compact metric spaces (in fact, to all totally bounded spaces) and yields explicit formulas for natural subsets of the Euclidean metric. Our second result addresses the asymptotic behavior of reconstruction, distinguishing between pseudo-finite spaces, where the optimal error is attained after finitely many queries, and spaces where the approximation curve exhibits a nontrivial decay. We characterize pseudo-finiteness for convex subsets of Euclidean spaces.



Paperid:1237
Authors:Gang Wan, Qinlong Lan, Zihan Li, Huimin Wang, Wu Yitian, wang zhen, Wanhua Li, Yufei Guo
Abstract:
Spiking Neural Networks (SNNs), as a biologically inspired neural network architecture, have garnered significant attention due to their exceptional energy efficiency and increasing potential for various applications. In this work, we extend the use of SNNs to neural rendering tasks and introduce Spik-NeRF (Spiking Neural Radiance Fields). We observe that the binary spike activation map of traditional SNNs lacks sufficient information capacity, leading to information loss and a subsequent decline in the performance of spiking neural rendering models. To address this limitation, we propose the use of ternary spike neurons, which enhance the information-carrying capacity in the spiking neural rendering model. With ternary spike neurons, Spik-NeRF achieves performance that is on par with, or nearly identical to, traditional ANN-based rendering models. Additionally, we present a re-parameterization technique for inference that allows Spik-NeRF with ternary spike neurons to retain the event-driven, multiplication-free advantages typical of binary spike neurons. Furthermore, to further boost the performance of Spik-NeRF, we employ a distillation method, using an ANN-based NeRF to guide the training of our Spik-NeRF model, which is more compatible with the ternary neurons compared to the standard binary neurons. We evaluate Spik-NeRF on both realistic and synthetic scenes, and the experimental results demonstrate that Spik-NeRF achieves rendering performance comparable to ANN-based NeRF models.



Paperid:1238
Authors:Pengfei Li, Yuelin Han, Adam Wierman, Shaolei Ren
Abstract:
Abstract:Fairness and action smoothness are two crucial considerations in many online optimization problems, but they have yet to be addressed simultaneously. In this paper, we study a new and challenging setting of fairness-regularized smoothed online convex optimization with switching costs. We propose FairOBD (Fairness-regularized Online Balanced Descent), which reconciles the tension between minimizing the hitting cost, switching cost, and fairness cost. Concretely, FairOBD decomposes the long-term fairness cost into a sequence of online costs by introducing an auxiliary variable and then leverages the auxiliary variable to regularize the online actions for fair outcomes. Based on a novel approach to account for switching costs, we prove that FairOBD offers a worst-case asymptotic competitive ratio against the optimal offline algorithm with parameterized constraints by considering the problem episode length $T\to\infty$. Finally, we run trace-driven experiments of dynamic computing resource provisioning for socially equitable AI inference to empirically evaluate FairOBD, showing that FairOBD can effectively reduce the total fairness-regularized cost and better promote fair outcomes compared to existing baseline solutions.



Paperid:1239
Authors:Geonho Hwang
Abstract:
Abstract:Recently, there has been a growing focus on determining the minimum width requirements for achieving the universal approximation property in deep, narrow Multi-Layer Perceptrons (MLPs).Among these challenges, one particularly challenging task is approximating a continuous function under the uniform norm, as indicated by the significant disparity between its lower and upper bounds.To address this problem, we propose a framework that simplifies finding the minimum width for deep, narrow MLPs into determining a purely geometrical function denoted as $w(d_x, d_y)$. This function relies solely on the input and output dimensions, represented as $d_x$ and $d_y$, respectively.To achieve this, we first demonstrate that deep, narrow MLPs, when provided with a small additional width, can approximate any $C^2$-diffeomorphism.Subsequently, using this result, we prove that $w(d_x, d_y)$ equates to the optimal minimum width required for deep, narrow MLPs to achieve universality.By employing the aforementioned framework and the Whitney embedding theorem, we provide an upper bound for the minimum width, given by $\operatorname{max}(2d_x+1, d_y) + \alpha(\sigma)$, where $0 \leq \alpha(\sigma) \leq 2$ represents a constant depending explicitly on the activation function.Furthermore, we provide novel optimal values for the minimum width in several settings, including $w(2,2)=w(2,3) = 4$.



Paperid:1240
Authors:Daoguang Zan, Zhirong Huang, Wei Liu, Hanwu Chen, Shulin Xin, Linhao Zhang, Qi Liu, Li Aoyan, Lu Chen, Xiaojian Zhong, Siyao Liu, Yongsheng Xiao, Liangqiang Chen, Yuyu Zhang, Jing Su, Tianyu Liu, RUI LONG, Ming Ding, liang xiang
Abstract:
The task of issue resolving aims to modify a codebase to generate a patch that addresses a given issue. However, most existing benchmarks focus almost exclusively on Python, making them insufficient for evaluating Large Language Models (LLMs) across different programming languages. To bridge this gap, we introduce a multilingual issue-resolving benchmark, called Multi-SWE-bench, covering 8 languages of Python, Java, TypeScript, JavaScript, Go, Rust, C, and C++. In particular, this benchmark includes a total of 2,132 high-quality instances, carefully curated by 68 expert annotators, ensuring a reliable and accurate evaluation of LLMs on the issue-resolving task. Based on human-annotated results, the issues are further classified into three difficulty levels. We evaluate a series of state-of-the-art models on Multi-SWE-bench, utilizing both procedural and agent-based frameworks for issue resolving. Our experiments reveal three key findings: (1) Limited generalization across languages: While existing LLMs perform well on Python issues, their ability to generalize across other languages remains limited; (2) Performance aligned with human-annotated difficulty: LLM-based agents' performance closely aligns with human-assigned difficulty, with resolution rates decreasing as issue complexity rises; and (3) Performance drop on cross-file issues: The performance of current methods significantly deteriorates when handling cross-file issues. These findings highlight the limitations of current LLMs and underscore the need for more robust models capable of handling a broader range of programming languages and complex issue scenarios.



Paperid:1241
Authors:Heyang Zhao, Chenlu Ye, Quanquan Gu, Tong Zhang
Abstract:
Abstract:Reverse-Kullback-Leibler (KL) regularization has emerged to be a predominant technique to enhance policy optimization in reinforcement learning (RL) and reinforcement learning from human feedback (RLHF), which forces the learned policy to stay close to a reference policy. While the effectiveness of KL-regularization has been empirically demonstrated in various practical scenarios, current theoretical analyses of KL-regularized RLHF still yield the same $\mathcal{O}(1 / \epsilon^2)$ sample complexity as ones without KL-regularization. To understand the fundamental distinction between objectives with KL-regularization and ones without KL-regularization, we are the first to theoretically demonstrate the power of KL-regularization by providing a sharp analysis for KL-regularized contextual bandits and RLHF, revealing an $\mathcal{O}(1 / \epsilon)$ sample complexity when $\epsilon$ is sufficiently small. We also prove matching lower bounds for both settings. More specifically, we study how the coverage of the reference policy affects the sample complexity of KL-regularized online contextual bandits and RLHF. We show that with sufficient coverage from the reference policy, a simple two-stage mixed sampling algorithm can achieve an $\mathcal{O}(1 / \epsilon)$ sample complexity with only an additive dependence on the coverage coefficient, thus proving the benefits of online data even without explicit exploration. Our results provide a comprehensive understanding of the roles of KL-regularization and data coverage in online decision making, shedding light on the design of more efficient algorithms.



Paperid:1242
Authors:Yunhao Luo, Utkarsh Mishra, Yilun Du, Danfei Xu
Abstract:
Effective trajectory stitching for long-horizon planning is a significant challenge in robotic decision-making. While diffusion models have shown promise in planning, they are limited to solving tasks similar to those seen in their training data. We propose CompDiffuser, a novel generative approach that can solve new tasks by learning to compositionally stitch together shorter trajectory chunks from previously seen tasks. Our key insight is modeling the trajectory distribution by subdividing it into overlapping chunks and learning their conditional relationships through a single bidirectional diffusion model. This allows information to propagate between segments during generation, ensuring physically consistent connections. We conduct experiments on benchmark tasks of various difficulties, covering different environment sizes, agent state dimension, trajectory types, training data quality, and show that CompDiffuser significantly outperforms existing methods.



Paperid:1243
Authors:Kaibo Zhang, Yunjuan Wang, Raman Arora
Abstract:
Curriculum learning has emerged as an effective strategy to enhance the training efficiency and generalization of machine learning models. However, its theoretical underpinnings remain relatively underexplored. In this work, we develop a theoretical framework for curriculum learning based on biased regularized empirical risk minimization (RERM), identifying conditions under which curriculum learning provably improves generalization. We introduce a sufficient condition that characterizes a ``good'' curriculum and analyze a multi-task curriculum framework, where solving a sequence of convex tasks can facilitate better generalization. We also demonstrate how these theoretical insights translate to practical benefits when using stochastic gradient descent (SGD) as an optimization method. Beyond convex settings, we explore the utility of curriculum learning for non-convex tasks. Empirical evaluations on synthetic datasets and MNIST validate our theoretical findings and highlight the practical efficacy of curriculum-based training.



Paperid:1244
Authors:Yujin Potter, Zhun Wang, Nicholas Crispino, Kyle Montgomery, Alexander Xiong, Ethan Chang, Francesco Pinto, Yuqi Chen, Rahul Gupta, Morteza Ziyadi, Christos Christodoulopoulos, Bo Li, Chenguang Wang, Dawn Song
Abstract:
As foundation models become more sophisticated, ensuring their trustworthiness becomes increasingly critical; yet, unlike text and image, the video modality still lacks comprehensive trustworthiness benchmarks. We introduce VMDT (Video-Modal DecodingTrust), the first unified platform for evaluating text-to-video (T2V) and video-to-text (V2T) models across five key trustworthiness dimensions: safety, hallucination, fairness, privacy, and adversarial robustness. Through our extensive evaluation of 7 T2V models and 19 V2T models using VMDT, we uncover several significant insights. For instance, all open-source T2V models evaluated fail to recognize harmful queries and often generate harmful videos, while exhibiting higher levels of unfairness compared to image modality models. In V2T models, unfairness and privacy risks rise with scale, whereas hallucination and adversarial robustness improve---though overall performance remains low. Uniquely, safety shows no correlation with model size, implying that factors other than scale govern current safety levels. Our findings highlight the urgent need for developing more robust and trustworthy video foundation models, and VMDT provides a systematic framework for measuring and tracking progress toward this goal. The code is available at \url{https://github.com/sunblaze-ucb/VMDT}.



Paperid:1245
Authors:Stephan Rabanser, Nicolas Papernot
Abstract:
Selective classifiers improve reliability by abstaining on uncertain inputs, yet their performance often lags behind the perfect-ordering oracle that accepts examples in exact order of correctness. We formulate this shortfall as a coverage-uniform selective-classification gap and prove the first finite-sample decomposition that pinpoints five distinct sources of looseness: Bayes noise, approximation error, ranking error, statistical noise, and implementation or shift-induced slack. Our bound shows that monotone post-hoc calibration cannot reduce the gap, as it preserves the original score ordering; closing the gap therefore requires scoring mechanisms that can modify the ranking induced by the base model. We validate our gap decomposition on synthetic two-moons data and real-world vision benchmarks, isolating each error component via controlled experiments. Results confirm that (i) Bayes noise and limited model capacity alone explain large gaps, (ii) only non-monotone or feature-aware calibrators shrink the ranking term, and (iii) distribution shift adds a distinct slack that must be addressed by robust training. Our decomposition supplies a quantitative error budget and concrete design guidelines for building selective classifiers that approach ideal oracle behavior.



Paperid:1246
Authors:Ilona Demler, Saumya Chauhan, Georgia Gkioxari
Abstract:
We introduce ITTO, a challenging new benchmark suite for evaluating and diagnosing the capabilities and limitations of point tracking methods. Our videos are sourced from existing datasets and egocentric real-world recordings, with high-quality human annotations collected through a multi-stage pipeline. ITTO captures the motion complexity, occlusion patterns, and object diversity characteristic of real-world scenes -- factors that are largely absent in current benchmarks. We conduct a rigorous analysis of state-of-the-art tracking methods on ITTO, breaking down performance along key axes of motion complexity. Our findings reveal that existing trackers struggle with these challenges, particularly in re-identifying points after occlusion, highlighting critical failure modes. These results point to the need for new modeling approaches tailored to real-world dynamics. We envision ITTO as a foundation testbed for advancing point tracking and guiding the development of more robust tracking algorithms.



Paperid:1247
Authors:Jialong Wu, Shaofeng Yin, Ningya Feng, Mingsheng Long
Abstract:
World models predict state transitions in response to actions and are increasingly developed across diverse modalities. However, standard training objectives such as maximum likelihood estimation (MLE) often misalign with task-specific goals of world models, i.e., transition prediction metrics like accuracy or perceptual quality. In this paper, we present RLVR-World, a unified framework that leverages reinforcement learning with verifiable rewards (RLVR) to directly optimize world models for such metrics. Despite formulating world modeling as autoregressive prediction of tokenized sequences, RLVR-World evaluates metrics of decoded predictions as verifiable rewards. We demonstrate substantial performance gains on both language- and video-based world models across domains, including text games, web navigation, and robot manipulation. Our work indicates that, beyond recent advances in reasoning language models, RLVR offers a promising post-training paradigm for enhancing the utility of generative models more broadly.



Paperid:1248
Authors:Jiaxin Song, Pooja Kulkarni, Parnian Shahkar, Bhaskar Ray Chaudhury
Abstract:
Abstract:The rapid growth of data-driven technologies and the emergence of various data-sharing paradigms have underscored the need for efficient and stable data exchange protocols. In any such exchange, agents must carefully balance the benefit of acquiring valuable data against the cost of sharing their own. Ensuring stability in these exchanges is essential to prevent agents—or groups of agents—from departing and conducting local (and potentially more favorable) exchanges among themselves. To address this, we study a model where $n$ agents participate in a data exchange. Each agent has an associated payoff for the data acquired from other agents and a cost incurred during sharing its own data. The net utility of an agent is payoff minus the cost. We adapt the classical notion of *core-stability* from cooperative game theory to data exchange. A data exchange is core-stable if no subset of agents has any incentive to deviate to a different exchange. We show that a core-stable data exchange is guaranteed to exist when agents have concave payoff functions and convex cost functions-- a setting typical in domains like PAC learning and random discovery models. We show that relaxing either of the foregoing conditions may result in the nonexistence of core-stable data exchanges. Then, we prove that finding a core-stable exchange is *PPAD-hard*, even when the potential blocking coalitions are restricted to constant size. To the best of our knowledge, this provides the first known PPAD-hardness result for core-like guarantees in data economics. Finally, we show that data exchange can be modelled as a *balanced* $n$-person game. This immediately gives a pivoting algorithm via Scarf's theorem [Scarf1967core]. We show that the pivoting algorithm works well in practice through our empirical results.



Paperid:1249
Authors:Zixuan Huang, Yikun Ban, Lean Fu, Xiaojie Li, Zhongxiang Dai, Jianxin Li, deqing wang
Abstract:
Direct Preference Optimization (DPO) has emerged as an effective approach for aligning large language models (LLMs) with human preferences. However, its performance is highly dependent on the quality of the underlying human preference data. To address this bottleneck, prior work has explored various data selection strategies, but these methods often overlook the impact of the evolving states of the language model during the DPO process.%including active querying, response pair selection, and data pre-selection. In this paper, we introduce a novel problem: Sample Scheduling for DPO, which aims to dynamically and adaptively schedule training samples based on the model's evolving states throughout preference optimization. To solve this problem, we propose SamS, an efficient and effective algorithm that adaptively selects samples in each training batch based on the LLM's learning feedback to maximize the potential generalization performance.Notably, without modifying the core DPO algorithm, simply integrating SamS significantly improves performance across tasks, with minimal additional computational overhead. This work points to a promising new direction for improving LLM alignment through more effective utilization of fixed preference datasets.



Paperid:1250
Authors:Mo WANG, Kaining Peng, Jingsheng Tang, Hongkai Wen, Quanying Liu
Abstract:
Brain atlases are essential for reducing the dimensionality of neuroimaging data and enabling interpretable analysis. However, most existing atlases are predefined, group-level templates with limited flexibility and resolution. We present Deep Cluster Atlas (DCA), a graph-guided deep embedding clustering framework for generating individualized, voxel-wise brain parcellations. DCA combines a pretrained autoencoder with spatially regularized deep clustering to produce functionally coherent and spatially contiguous regions. Our method supports flexible control over resolution and anatomical scope, and generalizes to arbitrary brain structures. We further introduce a standardized benchmarking platform for atlas evaluation. Across multiple datasets and scales, DCA outperforms state-of-the-art atlases, improving functional homogeneity by 74.1% and silhouette coefficient by 24.4%, and achieves superior performance in downstream tasks such as autism diagnosis and cognitive decoding. Codes and models are available at https://anonymous.4open.science/r/DCA-03B1.



Paperid:1251
Authors:Tianyu Pan, Mengdi Zhu, Alexa Cole, Ronald Wilson, Damon Woodard
Abstract:
Multimodal learning systems often struggle in non-stationary environments due to concept drift, where changing data distributions can degrade performance. Modality-specific drifts and the lack of mechanisms for continuous, stable adaptation compound this challenge. This paper introduces LS-OGD, a novel adaptive control framework for robust multimodal learning under concept drift. LS-OGD uses an online controller that dynamically adjusts the model's learning rate and the fusion weights between different data modalities in response to detected drift and evolving prediction errors. We prove that under bounded drift conditions, the LS-OGD system's prediction error is uniformly ultimately bounded and converges to zero if the drift ceases. Also, we show that the adaptive fusion strategy effectively isolates and mitigates the impact of severe modality-specific drift, ensuring system resilience and fault tolerance. These theoretical guarantees establish a principled foundation for developing reliable and continuously adapting multimodal learning systems.



Paperid:1252
Authors:Andrea Amaduzzi, Pierluigi Zama Ramirez, Giuseppe Lisanti, Samuele Salti, Luigi Di Stefano
Abstract:
Neural Radiance Fields (NeRFs) are neural networks - typically multilayer perceptrons (MLPs) - that represent objects' geometry and appearance, with applications in vision, graphics, and robotics. Recent works propose understanding NeRFs with natural language using Multimodal Large Language Models (MLLMs) that directly process the weights of a NeRF's MLP. However, these approaches rely on a global representation of the input object, making them unsuitable for spatial reasoning and capturing fine-grained details. Contrarily, we propose weights2space, a self-supervised framework featuring a novel meta-encoder that can compute a sequence of spatial tokens directly from NeRF weights. Leveraging this representation, we build Spatial LLaNA, a novel MLLM for NeRFs, capable of understanding details and spatial relationships in objects represented as NeRFs. We evaluate Spatial LLaNA on NeRF captioning and NeRF Q&A tasks, using both existing benchmarks and our novel Spatial ObjaNeRF dataset consisting of 100 manually-curated language annotations for NeRFs. The latter includes 3D models and descriptions that highlight the spatial reasoning capability of MLLMs. Spatial LLaNA outperforms existing approaches across all tasks.



Paperid:1253
Authors:Mingyu Kim, Dongjun Kim, Amman Yusuf, Stefano Ermon, Mi Jung Park
Abstract:
There is growing concern over the safety of powerful diffusion models, as they are often misused to produce inappropriate, not-safe-for-work content or generate copyrighted material or data of individuals who wish to be forgotten. Many existing methods tackle these issues by heavily relying on text-based negative prompts or retraining the model to eliminate certain features or samples. In this paper, we take a radically different approach, directly modifying the sampling trajectory by leveraging a negation set (e.g., unsafe images, copyrighted data, or private data) to avoid specific regions of data distribution, without needing to retrain or fine-tune the model. We formally derive the relationship between the expected denoised samples that are safe and those that are unsafe, leading to oursafedenoiser, which ensures its final samples are away from the area to be negated. We achieve state-of-the-art safety performance in large-scale datasets such as the CoPro dataset while also enabling significantly more cost-effective sampling than existing methodologies.



Paperid:1254
Authors:Jaeik Kim, Jaeyoung Do
Abstract:
Image classifiers play a critical role in detecting diseases in medical imaging and identifying anomalies in manufacturing processes. However, their predefined behaviors after extensive training make post hoc model editing difficult, especially when it comes to forgetting specific classes or adapting to distribution shifts. Existing classifier editing methods either focus narrowly on correcting errors or incur extensive retraining costs, creating a bottleneck for flexible editing. Moreover, such editing has seen limited investigation in image classification. To overcome these challenges, we introduce class vectors, which capture class-specific representation adjustments during fine-tuning. Whereas task vectors encode task-level changes in weight space, class vectors disentangle each class’s adaptation in the latent space. We show that class vectors capture each class’s semantic shift and that classifier editing can be achieved either by steering latent features along these vectors or by mapping them into weight space to update the decision boundaries. We also demonstrate that the inherent linearity and orthogonality of class vectors support efficient, flexible, and high-level concept editing via simple class arithmetic. Finally, we validate their utility in applications such as unlearning, environmental adaptation, adversarial defense, and adversarial trigger optimization.



Paperid:1255
Authors:Zhanke Wang, Zhiyan Wang, Kaiqiang Xiong, Wu Jiahao, Yang Deng, Ronggang Wang
Abstract:
Recently, 3D Gaussian Splatting (3DGS) has achieved state-of-the-art rendering results. However, its efficiency relies on simplifications that disregard the thickness of Gaussian primitives and their overlapping interactions. These simplifications can lead to popping artifacts due to inaccurate sorting, thereby affecting the rendering quality. In this paper, we propose Screen-Aligned Primitives (SAP), an anisotropic kernel that generates primitives parallel to the image plane for each view. Our rasterization pipeline enables full per-pixel ordering in real time. Since the primitives are parallel for a given viewpoint, a single global sorting operation suffices for correct per-pixel depth ordering. We formulate 3D reconstruction as a combination of a 3D-consistent decoder and 2D view-specific primitives, and further propose a highly efficient decoder to ensure 3D consistency. Moreover, within our framework, the primitive function values remain consistent between view space and screen space, allowing arbitrary radial basis functions (RBFs) to represent the scene without introducing projection errors. Experiments on diverse datasets demonstrate that our method achieves state-of-the-art rendering quality while maintaining real-time performance.



Paperid:1256
Authors:Wenhao Sun, Rong-Cheng Tu, Yifu Ding, Jingyi Liao, Zhao Jin, Shunyu Liu, Dacheng Tao
Abstract:
Abstract:Video Diffusion Transformers (VDiTs) have achieved remarkable progress in high-quality video generation, but remain computationally expensive due to the quadratic complexity of attention over high-dimensional video sequences. Recent attention acceleration methods leverage the sparsity of attention patterns to improve efficiency; however, they often overlook inefficiencies of redundant long-range interactions. To address this, we propose **VORTA**, an efficient VDiT framework with two novel components: (1) a sparse attention mechanism that efficiently captures long-range dependencies, and (2) a routing strategy that adaptively replaces full 3D attention with specialized sparse attention experts throughout the sampling process. It achieves a $\mathbf{1.76\times}$ end-to-end speedup without quality loss on VBench.Furthermore, VORTA can seamlessly integrate with various other acceleration methods, such as caching and step distillation, reaching up to $\mathbf{14.41\times}$ speedup with negligible performance degradation.VORTA demonstrates its efficiency and enhances the practicality of VDiTs in real-world settings.



Paperid:1257
Authors:Peter Blohm, Florian Chen, Aristides Gionis, Stefan Neumann
Abstract:
Online social media provide a platform for people to discuss current events and exchange opinions with their peers. While interactions are predominantly positive, in recent years we are increasingly witnessing conflicts based on different views and opinions. In this paper, we ask whether the conflicts in a network reveal a small and interpretable set of prevalent opinion ranges that explain the users' interactions. More precisely, we consider signed graphs, where the edge signs indicate positive and negative interactions of node pairs, and our goal is to infer opinion intervals that are consistent with the edge signs. We introduce an optimization problem that models this question, and we give strong hardness results and a polynomial-time approximation scheme by utilizing connections to interval graphs and the Correlation Clustering problem. We further provide scalable heuristics and show that in experiments they yield more expressive solutions than Correlation Clustering baselines. We also present a case study on a novel real-world dataset from the German parliament, showing that our algorithms can recover the political leaning of German parties based on co-voting behavior.



Paperid:1258
Authors:Zehong Wang, Zheyuan Zhang, Tianyi Ma, Chuxu Zhang, Yanfang Ye
Abstract:
Abstract:Graph neural networks (GNNs) has been predominantly driven by message-passing, where node representations are iteratively updated via local neighborhood aggregation. Despite their success, message-passing suffers from fundamental limitations---including constrained expressiveness, over-smoothing, over-squashing, and limited capacity to model long-range dependencies. These issues hinder scalability: increasing data size or model size often fails to yield improved performance, limiting the viability of GNNs as backbones for graph foundation models. In this work, we explore pathways beyond message-passing and introduce Generative Graph Pattern Machine (G$^2$PM), a generative Transformer pre-training framework for graphs. G$^2$PM represents graph instances (nodes, edges, or entire graphs) as sequences of substructures, and employs generative pre-training over the sequences to learn generalizable, transferable representations. Empirically, G$^2$PM demonstrates strong scalability: on the ogbn-arxiv benchmark, it continues to improve with model sizes up to 60M parameters, outperforming prior generative approaches that plateau at significantly smaller scales (e.g., 3M). In addition, we systematically analyze the model design space, highlighting key architectural choices that contribute to its scalability and generalization. Across diverse tasks---including node classification, graph classification, and transfer learning---G$^2$PM consistently outperforms strong baselines, establishing a compelling foundation for scalable graph learning.



Paperid:1259
Authors:Joonhyuk Park, Donghyun Lee, Yujee Song, Guorong Wu, Won Hwa Kim
Abstract:
Generating realistic graphs faces challenges in estimating accurate distribution of graphs in an embedding space while preserving structural characteristics. However, existing graph generation methods primarily focus on approximating the joint distribution of nodes and edges, often overlooking topological properties such as connected components and loops, hindering accurate representation of global structures. To address this issue, we propose a Topology-Aware diffusion-based Graph Generation (TAGG), which aims to sample synthetic graphs that closely resemble the structural characteristics of the original graph based on persistent homology. Specifically, we suggest two core components: 1) Persistence Diagram Matching (PDM) loss which ensures high topological fidelity of generated graphs, and 2) topology-aware attention module (TAM) which induces the denoising network to capture the homological characteristics of the original graphs. Extensive experiments on conventional graph benchmarks demonstrate the effectiveness of our approach demonstrating high generation performance across various metrics, while achieving closer alignment with the distribution of topological features observed in the original graphs. Furthermore, application to real brain network data showcases its potential for complex and real graph applications.



Paperid:1260
Authors:Jiaxing Wang, Deping Xiang, Jin Xu, Mingyang Yi, Guoqiang Gong, Zicheng Zhang, Haoran Li, Pengzhang Liu, Zhen Chen, Ke Zhang, Ju Fan, Qixia Jiang
Abstract:
The capabilities of large language models (LLMs) significantly depend on training data drawn from various domains. Optimizing domain-specific mixture ratios can be modeled as a bi-level optimization problem, which we simplify into a single-level penalized form and solve with twin networks: a proxy model trained on primary data and a dynamically updated reference model trained with additional data. Our proposed method, Twin Networks for bi-level DatA mixturE optiMization (TANDEM), measures the data efficacy through the difference between the twin models and up-weights domains that benefit more from the additional data. TANDEM provides theoretical guarantees and wider applicability, compared to prior approaches. Furthermore, our bi-level perspective suggests new settings to study domain reweighting such as data-restricted scenarios and supervised fine-tuning, where optimized mixture ratios significantly improve the performance. Extensive experiments validate TANDEM's effectiveness in all scenarios.



Paperid:1261
Authors:Yiying Yang, Wei Cheng, Sijin Chen, Xianfang Zeng, Fukun Yin, Jiaxu Zhang, Liao Wang, Gang Yu, Xingjun Ma, Yu-Gang Jiang
Abstract:
Scalable Vector Graphics (SVG) is an important image format widely adopted in graphic design because of their resolution independence and editability. The study of generating high-quality SVG has continuously drawn attention from both designers and researchers in the AIGC community. However, existing methods either produces unstructured outputs with huge computational cost or is limited to generating monochrome icons of over-simplified structures. To produce high-quality and complex SVG, we propose OmniSVG, a unified framework that leverages pre-trained Vision-Language Models (VLMs) for end-to-end multimodal SVG generation. By parameterizing SVG commands and coordinates into discrete tokens, OmniSVG decouples structural logic from low-level geometry for efficient training while maintaining the expressiveness of complex SVG structure. To further advance the development of SVG synthesis, we introduce MMSVG-2M, a multimodal dataset with two million richly annotated SVG assets, along with a standardized evaluation protocol for conditional SVG generation tasks. Extensive experiments show that OmniSVG outperforms existing methods and demonstrates its potential for integration into professional SVG design workflows.



Paperid:1262
Authors:HuaJun Xi, Kangdao Liu, Hao Zeng, Wenguang Sun, Hongxin Wei
Abstract:
Abstract:Conformal prediction is an emerging technique for uncertainty quantification that constructs prediction sets guaranteed to contain the true label with a predefined probability. Recent work develops online conformal prediction methods that adaptively construct prediction sets to accommodate distribution shifts. However, existing algorithms typically assume *perfect label accuracy* which rarely holds in practice. In this work, we investigate the robustness of online conformal prediction under uniform label noise with a known noise rate. We show that label noise causes a persistent gap between the actual mis-coverage rate and the desired rate $\alpha$, leading to either overestimated or underestimated coverage guarantees. To address this issue, we propose a novel loss function *robust pinball loss*, which provides an unbiased estimate of clean pinball loss without requiring ground-truth labels.Theoretically, we demonstrate that robust pinball loss enables online conformal prediction to eliminate the coverage gap under uniform label noise, achieving a convergence rate of $\mathcal{O}(T^{-1/2})$ for both empirical and expected coverage errors (i.e., absolute deviation of the empirical and expected mis-coverage rate from the target level $\alpha$).This loss offers a general solution to the uniform label noise, and is complementary to existing online conformal prediction methods.Extensive experiments demonstrate that the proposed loss enhances the noise robustness of various online conformal prediction methods by achieving a precise coverage guarantee.



Paperid:1263
Authors:Lucas Maystre, Gabriel Barello, Tudor Berariu, Cambray, Rares Dolga, Alvaro Gonzalez, Andrei Nica, David Barber
Abstract:
We address the problem of incremental sequence classification, where predictions are updated as new elements in the sequence are revealed. Drawing on temporal-difference learning from reinforcement learning, we identify a temporal-consistency condition that successive predictions should satisfy. We leverage this condition to develop a novel loss function for training incremental sequence classifiers. Through a concrete example, we demonstrate that optimizing this loss can offer substantial gains in data efficiency. We apply our method to text classification tasks and show that it improves predictive accuracy over competing approaches on several benchmark datasets. We further evaluate our approach on the task of verifying large language model generations for correctness in grade-school math problems. Our results show that models trained with our method are better able to distinguish promising generations from unpromising ones after observing only a few tokens.



Paperid:1264
Authors:Zhiyang Chen, Hailong Yao, Xia Yin
Abstract:
Algorithm configuration, which involves selecting algorithm parameters based on sampled problem instances, is a crucial step in applying modern algorithms such as SAT solvers. Although prior work has attempted to understand the theoretical foundations of algorithm configuration, we still lack a comprehensive understanding of why practical algorithm configurators exhibit strong generalization performances in real-world scenarios. In this paper, through the lens of machine learning theory, we provide an algorithm-dependent generalization bound for the widely used model-based algorithm configurators under mild assumptions. Our approach is based on the algorithmic stability framework for generalization bounds. To the best of our knowledge, this is the first generalization guarantee that aligns with algorithm configurators in practice.



Paperid:1265
Authors:Yuxuan Yao, Shuqi LIU, Zehua Liu, Qintong Li, Mingyang LIU, Xiongwei Han, Zhijiang Guo, Han Wu, Linqi Song
Abstract:
Recent research has increasingly focused on reconciling the reasoning capabilities of System 2 with the efficiency of System 1. While existing training-based and prompt-based approaches face significant challenges in terms of efficiency and stability, model merging emerges as a promising strategy to integrate the diverse capabilities of different Large Language Models (LLMs) into a unified model. However, conventional model merging methods often assume uniform importance across layers, overlooking the functional heterogeneity inherent in neural components. To address this limitation, we propose \textbf{A}ctivation-Guided \textbf{C}onsensus \textbf{M}erging (\textbf{ACM}), a plug-and-play merging framework that determines layer-specific merging coefficients based on mutual information between activations of pre-trained and fine-tuned models. ACM effectively preserves task-specific capabilities without requiring gradient computations or additional training. Extensive experiments on Long-to-Short (L2S) and general merging tasks demonstrate that ACM consistently outperforms all baseline methods. For instance, in the case of Qwen-7B models, TIES-Merging equipped with ACM achieves a \textbf{55.3\%} reduction in response length while simultaneously improving reasoning accuracy by \textbf{1.3} points. We submit the code with the paper for reproducibility, and it will be publicly available.



Paperid:1266
Authors:Alessandro Benfenati, Alfio Ferrara, Alessio Marta, Davide Riva, Elisabetta Rocchetti
Abstract:
This paper introduces a general method for the exploration of equivalence classes in the input space ofTransformer models. The proposed approach is based on sound mathematical theory which describes the internal layers of a Transformer architecture as sequential deformations of the input manifold. Using eigendecomposition of the pullback of the distance metric defined on the output space through the Jacobian of the model, we are able to reconstruct equivalence classes in the input space and navigate across them.Our method enables two complementary exploration procedures: the first retrieves input instances that produce the same class probability distribution as the original instance—thus identifying elements within the same equivalence class—while the second discovers instances that yield a different class probability distribution, effectively navigating toward distinct equivalence classes. Finally, we demonstrate how the retrieved instances can be meaningfully interpreted by projecting their embeddings back into a human-readable format.



Paperid:1267
Authors:Alexander Lobashev, Maria Larchenko, Dmitry Guskov
Abstract:
We propose SW-Guidance, a training-free approach for image generation conditioned on the color distribution of a reference image. While it is possible to generate an image with fixed colors by first creating an image from a text prompt and then applying a color style transfer method, this approach often results in semantically meaningless colors in the generated image. Our method solves this problem by modifying the sampling process of a diffusion model to incorporate the differentiable Sliced 1-Wasserstein distance between the color distribution of the generated image and the reference palette. Our method outperforms state-of-the-art techniques for color-conditional generation in terms of color similarity to the reference, producing images that not only match the reference colors but also maintain semantic coherence with the original text prompt. Our source code is available at https://anonymous.4open.science/r/sw-guidance-3E7D.



Paperid:1268
Authors:Hyun-jun Choi, Joongkyu Lee, Min-hwan Oh
Abstract:
Abstract:We revisit the contextual cascading bandit, where a learning agent recommends an ordered list ($\text{\textit{cascade}}$) of items and a user scans the list sequentially, stopping at the first attractive item. Although cascading bandits underpin various search and recommendation engines, the role of the cascade length $K$ in shaping regret has remained ambiguous. Contrary to prior results that regret grows with $K$, we prove that in the contextual setting it actually $\text{\textit{decreases}}$ once $K$ is large enough.Leveraging this insight, we design a new upper-confidence-bound algorithm built on online mirror descent that attains the sharpest known regret upper bound, $\tilde{\mathcal{O}}\bigl(K\bar{p}^{K-1} d \sqrt{T}\bigr)$ for contextual cascading bandits. To complement this new regret upper bound,we match it with a lower bound of $\Omega \bigl(K\underline{p}^{K-1} d \sqrt{T}\bigr)$, where $0 \leq \underline{p} \leq \bar{p} < 1$. Together, these results fully characterize how regret truly scales with $K$ and close the theoretical gap for contextual cascading bandits.Finally, extensive experiments validate our theoretical results and show the effectiveness of our proposed method.



Paperid:1269
Authors:Yi Ding, Ruqi Zhang
Abstract:
Abstract:Reasoning Vision-Language Models (VLMs) have shown promising performance on complex multimodal tasks. However, they still face significant challenges: they are highly sensitive to reasoning errors, require large volumes of annotated data or accurate verifiers, and struggle to generalize beyond specific domains.To address these limitations, we explore self-correction as a strategy to enhance reasoning VLMs. We first conduct an in-depth analysis of reasoning VLMs’ self-correction abilities and identify key gaps. Based on our findings, we introduce Sherlock, a self-correction and self-improvement training framework. \emph{Sherlock} introduces a trajectory-level self-correction objective, a preference data construction method based on visual perturbation, and a dynamic $\beta$ for preference tuning. Once the model acquires self-correction capabilities using only 20k randomly sampled annotated data, it continues to self-improve without external supervision.Built on the Llama3.2-Vision-11B model, \emph{Sherlock Iter2} achieves remarkable results across eight benchmarks, reaching an average accuracy of 64.1 with direct generation and 65.4 after self-correction. It outperforms LLaVA-CoT (63.2), Mulberry (60.4), and LlamaV-o1 (63.4) while only using less than 20\% of the annotated data.



Paperid:1270
Authors:I. Shavindra Jayasekera, Jacob Si, Filippo Valdettaro, Wenlong Chen, Aldo Faisal, Yingzhen Li
Abstract:
As large language models (LLMs) gain popularity in conducting prediction tasks in-context, understanding the sources of uncertainty in in-context learning becomes essential to ensuring reliability. The recent hypothesis of in-context learning performing predictive Bayesian inference opens the avenue for Bayesian uncertainty estimation, particularly for decomposing uncertainty into epistemic uncertainty due to lack of in-context data and aleatoric uncertainty inherent in the in-context prediction task. However, the decomposition idea remains under-explored due to the intractability of the latent parameter posterior from the underlying Bayesian model. In this work, we introduce a variational uncertainty decomposition framework for in-context learning without explicitly sampling from the latent parameter posterior, by optimising auxiliary inputs as probes to obtain an upper bound to the aleatoric uncertainty of an LLM's in-context learning procedure. Through experiments on synthetic and real-world tasks, we show quantitatively and qualitatively that the decomposed uncertainties obtained from our method exhibit desirable properties of epistemic and aleatoric uncertainty.



Paperid:1271
Authors:Haoran Zhao, Yuchen Yan, Yongliang Shen, Haolei Xu, Wenqi Zhang, Kaitao Song, Jian Shao, Weiming Lu, Jun Xiao, Yueting Zhuang
Abstract:
Large reasoning models (LRMs), such as OpenAI o1 and DeepSeek-R1, have significantly enhanced their reasoning capabilities by generating longer chains of thought, demonstrating outstanding performance across a variety of tasks. However, this performance gain comes at the cost of a substantial increase in redundant reasoning during the generation process, leading to high computational overhead and exacerbating the issue of overthinking.Although numerous existing approaches aim to address the problem of overthinking, they often rely on external interventions.In this paper, we propose a novel framework,Self-Braking Tuning(SBT), which tackles overthinking from the perspective of allowing the model to regulate its own reasoning process, thus eliminating the reliance on external control mechanisms. We construct a set of overthinking identification metrics based on standard answers and design a systematic method to detect redundant reasoning. This method accurately identifies unnecessary steps within the reasoning trajectory and generates training signals for learning self-regulation behaviors. Building on this foundation, we develop a complete strategy for constructing data with adaptive reasoning lengths and introduce an innovative braking prompt mechanism that enables the model to naturally learn when to terminate reasoning at an appropriate point.Experiments across mathematical benchmarks (AIME, AMC, MATH500, GSM8K) demonstrate that our method reduces token consumption by up to 60\% while maintaining comparable accuracy to unconstrained models.



Paperid:1272
Authors:Chengxin Hu, Hao Li, Yihe Yuan, Zezheng Song, Chenyang Zhao, Haixin Wang
Abstract:
In the molecular domain, numerous studies have explored the use of multimodal large language models (LLMs) to construct a general-purpose, multi-task molecular model. However, these efforts are still far from achieving a truly universal molecular model. We identify three key challenges in this endeavor: (1) Existing molecular task datasets are typically small in scale and lack comprehensive domain coverage. (2) Tasks from different molecular subfields are difficult to effectively learn jointly through LLMs due to significant distributional shifts and competition among tasks, which introduces instability in the learning process. (3) Both inter-task and intra-task molecular representations demand different intrinsic dimensions in the language space, making it challenging to balance between redundancy and insufficiency in language model representations. To address these challenges, we innovatively categorize existing small-molecule tasks into four types: Mol2Mol, Mol2Text, Mol2Num, and Text2Mol. We then collect a dataset encompassing over 16 tasks with more than 1.3 million samples, making it the largest molecular instruction-tuning dataset to date. Leveraging the extensive pretraining of LLMs on existing chemical literature, we propose a novel multimodal LLM framework, namedOmni-Mol, which unifies all small-molecule tasks and supports both molecular generation and understanding. The core of Omni-Mol is our proposed AGL-MoE, which dynamically adapts to the intrinsic rank of different tasks. This mixture-of-experts architecture enhances the model's ability to handle diverse tasks and modalities effectively. Our model achieves unified instruction tuning across 16 tasks and attains state-of-the-art performance on 12 of them. Extensive experiments further demonstrate the scalability and versatility of Omni-Mol. Code is available via https://anonymous.4open.science/r/Omni-Mol-8EDB



Paperid:1273
Authors:Xiaoyuan Zhang, Yizhe Huang, Chengdong Ma, Zhixun Chen, Long Ma, Yali Du, Song-Chun Zhu, Yaodong Yang, Xue Feng
Abstract:
Designing adaptive mechanisms to align individual and collective interests remains a central challenge in artificial social intelligence. Existing methods often struggle with modeling heterogeneous agents possessing persistent latent traits (e.g., skills, preferences) and dealing with complex multi-agent system dynamics. These challenges are compounded by the critical need for high sample efficiency due to costly real-world interactions. World Models, by learning to predict environmental dynamics, offer a promising pathway to enhance mechanism design in heterogeneous and complex systems. In this paper, we introduce a novel method named SWM-AP (Social World Model-Augmented Mechanism Design Policy Learning), which learns a social world model hierarchically modeling agents' behavior to enhance mechanism design. Specifically, the social world model infers agents' traits from their interaction trajectories and learns a trait-based model to predict agents' responses to the deployed mechanisms. The mechanism design policy collects extensive training trajectories by interacting with the social world model, while concurrently inferring agents' traits online during real-world interactions to further boost policy learning efficiency. Experiments in diverse settings (tax policy design, team coordination, and facility location) demonstrate that SWM-AP outperforms established model-based and model-free RL baselines in cumulative rewards and sample efficiency.



Paperid:1274
Authors:Hongshuo Yang, Elias Bareinboim
Abstract:
Graphical models have been widely used as parsimonious encoders of assumptions of the underlying structural causal models and provide a basis from which causal inferences can be performed. Models that encode stronger constraints tend to have higher expressive power at the expense of lower empirical falsifiability. In this paper, we define two new collections of distributions which include counterfactual quantities that become experimentally accessible under the counterfactual randomization action. Correspondingly, we provide two new classes of graphical models for encoding empirically testable constraints in these distributions. We further present a sound and complete calculus, based on counterfactual calculus, which licenses inference in these two new models with rules that are also fall within the empirically falsifiable boundary. In addition, we formulate a hierarchy over several graphical models based on the constraints they encode and study the fundamental trade-off between the expressive power and empirical falsifiability of different models across the hierarchy.



Paperid:1275
Authors:Arun Jose
Abstract:
Language models trained via outcome-based reinforcement learning (RL) to reason using chain-of-thought (CoT) have shown remarkable performance. Monitoring such a model's CoT may allow us to understand its intentions and detect potential malicious behavior. However, to be effective, this requires that CoTs are legible and faithful. We evaluate the legibility of CoTs in state-of-the-art reasoning models. We find that R1 and R1-Zero often produce illegible CoTs (mixing nonsensical phrases, random words, and non-English characters) before returning to perfect coherence in their final responses, while Claude 3.7 Sonnet notably exhibits higher legibility. We observe a significant positive correlation between CoT illegibility and answer correctness, suggesting illegible reasoning may improve performance. Illegibility increases with question difficulty, suggesting that CoT monitoring may be less reliable precisely when most needed. We discuss potential hypotheses for these results, including steganography, vestigial tokens, and training artifacts. Our findings suggest that current approaches to CoT monitoring may be vulnerable to the emergence of outcome-based RL, particularly as models face increasingly complex tasks.



Paperid:1276
Authors:Naga Venkata Sai Jitin Jami, Thomas Altstidl, Jonas Mueller, Jindong Li, Dario Zanca, Bjoern Eskofier, Heike Leutheuser
Abstract:
Random splitting of datasets in image segmentation often leads to unrepresentative test sets, resulting in biased evaluations and poor model generalization. While stratified sampling has proven effective for addressing label distribution imbalance in classification tasks, extending these ideas to segmentation remains challenging due to the multi-label structure and class imbalance typically present in such data. Building on existing stratification concepts, we introduce Iterative Pixel Stratification (IPS), a straightforward, label-aware sampling method tailored for segmentation tasks. Additionally, we present Wasserstein-Driven Evolutionary Stratification (WDES), a novel genetic algorithm designed to minimize the Wasserstein distance, thereby optimizing the similarity of label distributions across dataset splits. We prove that WDES is globally optimal given enough generations. Using newly proposed statistical heterogeneity metrics, we evaluate both methods against random sampling and find that WDES consistently produces more representative splits. Applying WDES across diverse segmentation tasks, including street scenes, medical imaging, and satellite imagery, leads to lower performance variance and improved model evaluation. Our results also highlight the particular value of WDES in handling small, imbalanced, and low-diversity datasets, where conventional splitting strategies are most prone to bias.



Paperid:1277
Authors:Bhavya Vasudeva, Jung Lee, Vatsal Sharan, Mahdi Soltanolkotabi
Abstract:
Adam is the de facto optimization algorithm for several deep learning applications, but an understanding of its implicit bias and how it differs from other algorithms, particularly standard first-order methods such as (stochastic) gradient descent (GD), remains limited. In practice, neural networks (NNs) trained with SGD are known to exhibit simplicity bias --- a tendency to find simple solutions. In contrast, we show that Adam is more resistant to such simplicity bias. First, we investigate the differences in the implicit biases of Adam and GD when training two-layer ReLU NNs on a binary classification task with Gaussian data. We find that GD exhibits a simplicity bias, resulting in a linear decision boundary with a suboptimal margin, whereas Adam leads to much richer and more diverse features, producing a nonlinear boundary that is closer to the Bayes' optimal predictor. This richer decision boundary also allows Adam to achieve higher test accuracy both in-distribution and under certain distribution shifts. We theoretically prove these results by analyzing the population gradients. Next, to corroborate our theoretical findings, we present extensive empirical results showing that this property of Adam leads to superior generalization across various datasets with spurious correlations where NNs trained with SGD are known to show simplicity bias and do not generalize well under certain distributional shifts.



Paperid:1278
Authors:Apratim Bhattacharyya, Bicheng Xu, Sanjay Haresh, Reza Pourreza, Litian Liu, Sunny Panchal, Leonid Sigal, Roland Memisevic
Abstract:
Multi-modal Large Language Models (LLM) have advanced conversational abilities but struggle with providing live, interactive step-by-step guidance, a key capability for future AI assistants. Effective guidance requires not only delivering instructions but also detecting their successful execution, as well as identifying and alerting users to mistakes, all of which has to happen in real-time. This requires models that are not turn-based, but that can react asynchronously to a video stream, as well as video data showing users performing tasks including mistakes and their corrections. To this end, we introduce LiveCook, a new benchmark and dataset built upon CaptainCook4D, which contains user mistakes during task execution. LiveCook features densely annotated, timed instructions and feedback messages, specifically including mistake alerts precisely timestamped to their visual occurrence in the video. We evaluate state-of-the-art multi-modal LLMs on LiveCook and introduce LiveMamba, a streaming multi-modal LLM designed for interactive instructional guidance. This work provides the first dedicated benchmark and a strong baseline for developing and evaluating on live, situated coaching.



Paperid:1279
Authors:Ruijia Liu, Ancheng Hou, Xiao Yu, Xiang Yin
Abstract:
Signal Temporal Logic (STL) is a powerful specification language for describing complex temporal behaviors of continuous signals, making it well-suited for high-level robotic task descriptions. However, generating executable plans for STL tasks is challenging, as it requires consideration of the coupling between the task specification and the system dynamics. Existing approaches either follow a model-based setting that explicitly requires knowledge of the system dynamics or adopt a task-oriented data-driven approach to learn plans for specific tasks. In this work, we address the problem of generating executable STL plans for systems with unknown dynamics. We propose a hierarchical planning framework that enables zero-shot generalization to new STL tasks by leveraging only task-agnostic trajectory data during offline training. The framework consists of three key components: (i) decomposing the STL specification into several progresses and time constraints, (ii) searching for timed waypoints that satisfy all progresses under time constraints, and (iii) generating trajectory segments using a pre-trained diffusion model and stitching them into complete trajectories. We formally prove that our method guarantees STL satisfaction, and simulation results demonstrate its effectiveness in generating dynamically feasible trajectories across diverse long-horizon STL tasks.



Paperid:1280
Authors:Ken Gu, Zhihan Zhang, Kate Lin, Yuwei Zhang, Akshay Paruchuri, Hong Yu, Mehran Kazemi, Kumar Ayush, A. Ali Heydari, Maxwell Xu, Yun Liu, Ming-Zher Poh, Yuzhe Yang, Mark Malhotra, Shwetak Patel, Hamid Palangi, Xuhai "Orson" Xu, Daniel McDuff, Tim Althoff, Xin Liu
Abstract:
Language models (LMs) are increasingly being deployed to perform autonomous data analyses. However, their data awareness—the ability to recognize, reason over, and appropriately handle data artifacts such as missing values, outliers, and logical inconsistencies—remains underexplored. These artifacts are especially common in real-world tabular data and, if mishandled, can significantly compromise the validity of analytical conclusions. To address this gap, we present RADAR, a benchmark for systematically evaluating data-aware reasoning on tabular data. We develop a framework to simulate data artifacts via programmatic perturbations to enable targeted evaluation of model behavior. RADAR comprises 2,980 table-query pairs, grounded in real-world data spanning 9 domains and 5 data artifact types. In addition to evaluating artifact handling, RADAR systematically varies table size to study how reasoning performance holds when increasing table size. Our evaluation reveals that, despite decent performance on tables without data artifacts, frontier models degrade significantly when data artifacts are introduced, exposing critical gaps in their capacity for robust, data-aware analysis. Designed to be flexible and extensible, RADAR supports diverse perturbation types and controllable table sizes, offering a valuable resource for advancing tabular reasoning.



Paperid:1281
Authors:Leo Segre, Shai Avidan
Abstract:
Neural Radiance Fields (NeRF) have advanced photorealistic novel view synthesis, but their reliance on photometric reconstruction introduces artifacts, commonly known as "floaters". These artifacts degrade novel view quality, particularly in unseen regions where NeRF optimization is unconstrained. We propose a fast, post-hoc NeRF cleanup method that eliminates such artifacts by enforcing a Free Space Prior, ensuring that unseen regions remain empty while preserving the structure of observed areas. Unlike existing approaches that rely on Maximum Likelihood (ML) estimation or complex, data-driven priors, our method adopts a Maximum-a-Posteriori (MAP) approach with a simple yet effective global prior. This enables our method to clean artifacts in both seen and unseen areas, significantly improving novel view quality even in challenging scene regions. Our approach generalizes across diverse NeRF architectures and datasets while requiring no additional memory beyond the original NeRF. Compared to state-of-the-art cleanup methods, our method is 2.5x faster in inference and completes cleanup training in under 30 seconds. Our code will be made publicly available.



Paperid:1282
Authors:Guhao Feng, Yihan Geng, Jian Guan, Wei Wu, Liwei Wang, Di He
Abstract:
Diffusion language models have emerged as a new approach for text generation. By enabling the parallel sampling of multiple tokens in each diffusion step, they appear to offer a more efficient alternative to auto-regressive models. However, our observations show that current open-sourced diffusion language models require more sampling steps to achieve comparable accuracy on representative tasks--resulting in even higher inference costs than their auto-regressive counterparts. To investigate whether this is an inherent limitation, we conduct a rigorous theoretical analysis of a widely adopted variant: the Masked Diffusion Model (MDM). Surprisingly, our analysis reveals that the conclusion is highly sensitive to the choice of evaluation metric. Under mild conditions, we prove that when the target is near-optimal perplexity, MDMs can achieve this goal in a constant number of sampling steps, independent of sequence length. This result demonstrates that efficiency can, in principle, be attained without compromising generation quality. However, when targeting low sequence error rate--which is important for assessing the ``correctness" of a generated sequence, such as a reasoning chain--we show that in the worst case, the required sampling steps must scale linearly with sequence length, thereby eliminating the efficiency advantage. Our analysis establishes the first theoretical foundation for understanding the comparative strengths and limitations of MDMs, offering practical guidance on when to favor MDMs over the auto-regressive models and vice versa.



Paperid:1283
Authors:Abhay Sheshadri, John Hughes, Julian Michael, Alex Mallen, Arun Jose, Fabien Roger
Abstract:
Alignment faking in large language modelspresented a demonstration of Claude 3 Opus and Claude 3.5 Sonnet selectively complying with a helpful-only training objective to prevent modification of their behavior outside of training. We expand this analysis to 23 models and find that only 5 (Claude 3 Opus, Claude 3.5 Sonnet, Llama 3 405B, Grok 3, Gemini 2.0 Flash) comply with harmful queries more when they infer they are in training than when they infer they are in deployment. First, we study the motivations of these 5 models and find surprising differences in the motivations that result in this compliance gap. Second, we investigate why models like GPT-4o or DeepSeek-V3 don’t fake alignment. Our results suggest this is not due to a lack of capabilities: we find that base models of GPT-4 and DeepSeek-V3 fake alignment, and that models fine-tuned to refuse less and pay more attention to details of the scenario also fake alignment. Our results indicate that variations in refusal behavior may account for a significant portion of differences in alignment faking, which suggests that post-training methods may reduce alignment faking.



Paperid:1284
Authors:Banseok Lee, Dongkyu Kim, Youngcheon You, Youngmin Kim
Abstract:
Abstract:Deploying large language models (LLMs) often faces challenges from substantial memory and computational costs. Quantization offers a solution, yet performance degradation in the sub-1-bit regime remains particularly difficult. This paper introduces LittleBit, a novel method for extreme LLM compression. It targets levels like 0.1 bits per weight (BPW), achieving nearly 31$\times$ memory reduction, e.g., Llama2-13B to under 0.9 GB. LittleBit represents weights in a low-rank form using latent matrix factorization, subsequently binarizing these factors. To counteract information loss from this extreme precision, it integrates a multi-scale compensation mechanism. This includes row, column, and an additional latent dimension that learns per-rank importance. Two key contributions enable effective training: Dual Sign-Value-Independent Decomposition (Dual-SVID) for stable quantization-aware training (QAT) initialization, and integrated Residual Compensation to mitigate errors. Extensive experiments confirm LittleBit's superiority in sub-1-bit quantization: e.g., its 0.1 BPW performance on Llama2-7B surpasses the leading method's 0.7 BPW. This establishes a superior size-performance trade-off, with kernel-level benchmarks indicating potential for a 5$\times$ speedup compared to FP16. LittleBit paves the way for deploying powerful LLMs in resource-constrained environments.



Paperid:1285
Authors:Xinyu Li, Yuchen Luo, Hao Wang, Haoxuan Li, Liuhua Peng, Feng Liu, Yandong Guo, Kun Zhang, Mingming Gong
Abstract:
Recent booming time series models have demonstrated remarkable forecasting performance. However, these methods often place greater emphasis on more effectively modelling the historical series, largely neglecting the forecasting phase, which generates long-term forecasts by separately predicting multiple time points. Given that real-world time series typically consist of various long short-term fluctuations, independent predictions over individual time points may fail to express complex underlying patterns and can lead to a lack of global views. To address these issues, this work explores new perspectives from the forecasting phase and proposes a novel Implicit Forecaster (IF) as an additional decoding module. Inspired by decomposition forecasting, IF adopts a more nuanced approach by implicitly predicting constituent waves represented by their frequency, amplitude, and phase, thereby accurately forming the time series. Extensive experimental results from multiple real-world datasets show that IF can consistently boost mainstream time series models, achieving state-of-the-art forecasting performance.



Paperid:1286
Authors:Jingming Yan, Yiyuan Luo, Vaggos Chatziafratis, Ioannis Panageas, Parnian Shahkar, Stelios Stavroulakis
Abstract:
Abstract:Contrastive learning is a powerful technique for discovering meaningful data representations by optimizing objectives based on $\textit{contrastive information}$, often given as a set of weighted triplets $\{(x_i, y_i^+, z_{i}^-)\}_{i = 1}^m$ indicating that an "anchor" $x_i$ is more similar to a "positive" example $y_i$ than to a "negative" example $z_i$. The goal is to find representations (e.g., embeddings in $\mathbb{R}^d$ or a tree metric) where anchors are placed closer to positive than to negative examples. While finding $\textit{global}$ optima of contrastive objectives is $\mathsf{NP}$-hard, the complexity of finding $\text{\textit{local}}$ optima---representations that do not improve by local search algorithms such as gradient-based methods---remains open. Our work settles the complexity of finding local optima in various contrastive learning problems by proving $\mathsf{PLS}$-hardness in discrete settings (e.g., maximize satisfied triplets) and $\mathsf{CLS}$-hardness in continuous settings (e.g., minimize Triplet Loss), where $\mathsf{PLS}$ (Polynomial Local Search) and $\mathsf{CLS}$ (Continuous Local Search) are well-studied complexity classes capturing local search dynamics in discrete and continuous optimization, respectively. Our results imply that no polynomial time algorithm (local search or otherwise) can find a local optimum for various contrastive learning problems, unless $\mathsf{PLS}\subseteq\mathsf{P}$ (or $\mathsf{CLS}\subseteq \mathsf{P}$ for continuous problems). Even in the unlikely scenario that $\mathsf{PLS}\subseteq\mathsf{P}$ (or $\mathsf{CLS}\subseteq \mathsf{P}$), our reductions imply that there exist instances where local search algorithms need exponential time to reach a local optimum, even for $d=1$ (embeddings on a line).



Paperid:1287
Authors:Jingyi Zheng, Tianyi Hu, Yule Liu, Zhen Sun, Zongmin Zhang, Wenhan Dong, Zifan Peng, Xinlei He
Abstract:
Current benchmarks for evaluating large language models (LLMs) in social media moderation completely overlook a serious threat: covert advertisements, which disguise themselves as regular posts to deceive and mislead consumers into making purchases, leading to significant ethical and legal concerns. In this paper, we present the CHASM, a first-of-its-kind dataset designed to evaluate the capability of Multimodal Large Language Models (MLLMs) in detecting covert advertisements on social media. CHASM is a high-quality, anonymized, manually curated dataset consisting of 4,992 instances, based on real-world scenarios from the Chinese social media platform Rednote. The dataset was collected and annotated under strict privacy protection and quality control protocols. It includes many product experience sharing posts that closely resemble covert advertisements, making the dataset particularly challenging.The results show that under both zero-shot and in-context learning settings, none of the current MLLMs are sufficiently reliable for detecting covert advertisements.Our further experiments revealed that fine-tuning open-source MLLMs on our dataset yielded noticeable performance gains. However, significant challenges persist, such as detecting subtle cues in comments and differences in visual and textual structures.We provide in-depth error analysis and outline future research directions. We hope our study can serve as a call for the research community and platform moderators to develop more precise defenses against this emerging threat.



Paperid:1288
Authors:Shuchen Weng, Haojie Zheng, zheng chang, Si Li, Boxin Shi, Xinlong Wang
Abstract:
Audio is inherently temporal and closely synchronized with the visual world, making it a naturally aligned and expressive control signal for controllable video generation (e.g., movies).Beyond control, directly translating audio into video is essential for understanding and visualizing rich audio narratives (e.g., Podcasts or historical recordings).However, existing approaches fall short in generating high-quality videos with precise audio-visual synchronization, especially across diverse and complex audio types.In this work, we introduce MTV, a versatile framework for audio-sync video generation. MTV explicitly separates audios into speech, effects, and music tracks, enabling disentangled control over lip motion, event timing, and visual mood, respectively—resulting in fine-grained and semantically aligned video generation.To support the framework, we additionally present DEMIX, a dataset comprising high-quality cinematic videos and demixed audio tracks. DEMIX is structured into five overlapped subsets, enabling scalable multi-stage training for diverse generation scenarios.Extensive experiments demonstrate that MTV achieves state-of-the-art performance across six standard metrics spanning video quality, text-video consistency, and audio-video alignment.



Paperid:1289
Authors:Zhongnan Cai, Yingying Wang, Hui Zheng, Panwang Pan, ZiXu Lin, Ge Meng, Chenxin Li, Chunming He, Jiaxin Xie, Yunlong Lin, Junbin Lu, Yue Huang, Xinghao Ding
Abstract:
Recently, deep learning-based pan-sharpening algorithms have achieved notable advancements over traditional methods. However, many deep learning-based approaches incur substantial computational overhead during inference, especially with high-resolution images. This excessive computational demand limits the applicability of these methods in real-world scenarios, particularly in the absence of dedicated computing devices such as GPUs and TPUs. To address these challenges, we propose Pan-LUT, a novel learnable look-up table (LUT) framework for pan-sharpening that strikes a balance between performance and computational efficiency for high-resolution remote sensing images. To finely control the spectral transformation, we devise the PAN-guided look-up table (PGLUT) for channel-wise spectral mapping. To effectively capture fine-grained spatial details, we introduce the spatial details look-up table (SDLUT). Furthermore, to adaptively aggregate channel information for generating high-resolution multispectral images, we design an adaptive output look-up table (AOLUT). Our smallest model variant, Pan-LUT-1, contains fewer than 1M parameters and processes a 4K-resolution image in under 1 ms using a single NVIDIA GeForce RTX 2080 Ti GPU, demonstrating significantly faster performance compared to other methods. Experiments reveal that Pan-LUT efficiently processes large remote sensing images in a lightweight manner, bridging the gap to real-world applications. Furthermore, our model surpasses SOTA methods in full-resolution scenes under real-world conditions, highlighting its effectiveness and efficiency.



Paperid:1290
Authors:Vicky Kouni
Abstract:
Abstract:Unfolding networks are interpretable networks emerging from iterative algorithms, incorporate prior knowledge of data structure, and are designed to solve inverse problems like compressed sensing, which deals with recovering data from noisy, missing observations. Compressed sensing finds applications in critical domains, from medical imaging to cryptography, where adversarial robustness is crucial to prevent catastrophic failures. However, a solid theoretical understanding of the performance of unfolding networks in the presence of adversarial attacks is still in its infancy. In this paper, we study the adversarial generalization of unfolding networks when perturbed with $l_2$-norm constrained attacks, generated by the fast gradient sign method. Particularly, we choose a family of state-of-the-art overaparameterized unfolding networks and deploy a new framework to estimate their adversarial Rademacher complexity. Given this estimate, we provide adversarial generalization error bounds for the networks under study, which are tight with respect to the attack level. To our knowledge, this is the first theoretical analysis on the adversarial generalization of unfolding networks. We further present a series of experiments on real-world data, with results corroborating our derived theory, consistently for all data. Finally, we observe that the family's overparameterization can be exploited to promote adversarial robustness, shedding light on how to efficiently robustify neural networks.



Paperid:1291
Authors:Jiaxin Wen, Chenglei Si, Chen Yueh-Han, He He, Shi Feng
Abstract:
Many promising-looking ideas in AI research fail to deliver, but their validation takes substantial human labor and compute. Predicting an idea's chance of success is thus crucial for accelerating empirical AI research, a skill that even expert researchers can only acquire through substantial experience. We build the first benchmark for this task and compare LMs with human experts. Concretely, given two research ideas (e.g., two jailbreaking methods), we aim to predict which will perform better on a set of benchmarks. We scrape ideas and experimental results from conference papers, yielding 1,585 human-verified idea pairs \textit{published after our base model's cut-off date} for testing, and 6,000 pairs for training.We then develop a system that combines a fine-tuned GPT-4.1 with a paper retrieval agent, and we recruit 25 human experts to compare with.In the NLP domain, our system beats human experts by a large margin (64.4\% v.s. 48.9\%).On the full test set, our system achieves 77\% accuracy, while off-the-shelf frontier LMs like o3 perform no better than random guessing, even with the same retrieval augmentation.We verify that our system does not exploit superficial features like idea complexity through extensive human-written and LM-designed robustness tests.Finally, we evaluate our system on unpublished novel ideas, including ideas generated by an AI ideation agent.Our system achieves 63.6\% accuracy, demonstrating its potential as a reward model for improving idea generation models.Altogether, our results outline a promising new direction for LMs to accelerate empirical AI research.



Paperid:1292
Authors:Shijie Liu, Tansu Alpcan, Christopher Leckie, Sarah Erfani
Abstract:
Abstract:Time-series anomaly detection is critical for ensuring safety in high-stakes applications, where robustness is a fundamental requirement rather than a mere performance metric. Addressing the vulnerability of these systems to adversarial manipulation is therefore essential. Existing defenses are largely heuristic or provide certified robustness only under $\ell_p$-norm constraints, which are incompatible with time-series data. In particular, $\ell_p$-norm fails to capture the intrinsic temporal structure in time series, causing small temporal distortions to significantly alter the $\ell_p$-norm measures. Instead, the similarity metric Dynamic Time Warping (DTW) is more suitable and widely adopted in the time-series domain, as DTW accounts for temporal alignment and remains robust to temporal variations. To date, however, there has been no certifiable robustness result in this metric that provides guarantees. In this work, we introduce the first DTW-certified robust defense in time-series anomaly detection by adapting the randomized smoothing paradigm. We develop this certificate by bridging the $\ell_p$-norm to DTW distance through a lower-bound transformation. Extensive experiments across various datasets and models validate the effectiveness and practicality of our theoretical approach. Results demonstrate significantly improved performance, e.g., up to 18.7\% in F1-score under DTW-based adversarial attacks compared to traditional certified models.



Paperid:1293
Authors:Beibei Lin, Zifeng Yuan, Tingting Chen
Abstract:
Polarization images provide rich physical information that is fundamentally absent from standard RGB images, benefiting a wide range of computer vision applications such as reflection separation and material classification. However, the acquisition of polarization images typically requires additional optical components, which increases both the cost and the complexity of the applications. To bridge this gap, we introduce a new task: RGB-to-polarization image estimation, which aims to infer polarization information directly from RGB images. In this work, we establish the first comprehensive benchmark for this task by leveraging existing polarization datasets and evaluating a diverse set of state-of-the-art deep learning models, including both restoration-oriented and generative architectures. Through extensive quantitative and qualitative analysis, our benchmark not only establishes the current performance ceiling of RGB-to-polarization estimation, but also systematically reveals the respective strengths and limitations of different model families — such as direct reconstruction versus generative synthesis, and task-specific training versus large-scale pre-training. In addition, we provide some potential directions for future research. This benchmark is intended to serve as a foundational resource to facilitate the design and evaluation of future methods for polarization estimation from standard RGB inputs.



Paperid:1294
Authors:Dravyansh Sharma, Alec Sun
Abstract:
Machine learning is ubiquitous in societal decision-making, for example in evaluating job candidates or loan applications, and it is increasingly important to take into account how classified agents will react to the learning algorithms. The majority of recent literature on strategic classification has focused on reducing and countering deceptive behaviors by the classified agents, but recent work of Attias et al. identifies surprising properties of learnability when the agents genuinely improve in order to attain the desirable classification, such as smaller generalization error than standard PAC-learning. In this paper we characterize so-called learnability with improvements across multiple new axes. We introduce a novel asymmetric variant of minimally consistent concept classes and use it to provide an exact characterization of proper learning with improvements in the realizable setting. While prior work studies learnability only under general, arbitrary agent improvement regions, we give positive results for more natural Euclidean ball improvement sets. In particular, we characterize improper learning under a mild generative assumption on the data distribution. We further show how to learn in more challenging settings, achieving lower generalization error under well-studied noise models and obtaining mistake bounds in realizable and agnostic online learning. Our paper resolves several open questions posed by Attias et al. for both proper and improper learning.



Paperid:1295
Authors:Haotian Sun, Yitong Li, Yuchen Zhuang, Niao He, Hanjun Dai, Bo Dai
Abstract:
Contrastive Language-Image Pretraining (CLIP) has demonstrated strong zero-shot performance across diverse downstream text-image tasks. Existing CLIP methods typically optimize a contrastive objective using negative samples drawn from each minibatch. To achieve robust representation learning, these methods require extremely large batch sizes and escalate computational demands to hundreds or even thousands of GPUs. Prior approaches to mitigate this issue often compromise downstream performance, prolong training duration, or face scalability challenges with very large datasets. To overcome these limitations, we propose AmorLIP, an efficient CLIP pretraining framework that amortizes expensive computations involved in contrastive learning through lightweight neural networks, which substantially improves training efficiency and performance. Leveraging insights from a spectral factorization of energy-based models, we introduce novel amortization objectives along with practical techniques to improve training stability. Extensive experiments across 38 downstream tasks demonstrate the superior zero-shot classification and retrieval capabilities of AmorLIP, consistently outperforming standard CLIP baselines with substantial relative improvements of up to 12.24%.



Paperid:1296
Authors:Amirmojtaba Sabour, Sanja Fidler, Karsten Kreis
Abstract:
Diffusion- and flow-based models have emerged as state-of-the-art generative modeling approaches, but they require many sampling steps. Consistency models can distill these models into efficient one-step generators; however, unlike flow- and diffusion-based methods, their performance inevitably degrades when increasing the number of steps, which we show both analytically and empirically.Flow maps generalize these approaches by connecting any two noise levels in a single step and remain effective across all step counts. In this paper, we introduce two new continuous-time objectives for training flow maps, along with additional novel training techniques, generalizing existing consistency and flow matching objectives. We further demonstrate that autoguidance can improve performance, using a low-quality model for guidance during distillation, and an additional boost can be achieved by adversarial finetuning, with minimal loss in sample diversity.We extensively validate our flow map models, calledAlign Your Flow, on challenging image generation benchmarks and achieve state-of-the-art few-step generation performance on both ImageNet 64x64 and 512x512, using small and efficient neural networks. Finally, we show text-to-image flow map models that outperform all existing non-adversarially trained few-step samplers in text-conditioned synthesis.



Paperid:1297
Authors:Keshav Ramji, Tahira Naseem, Ramón Astudillo
Abstract:
When language model (LM) users aim to improve the quality of its generations, it is crucial to specify concrete behavioral attributes that the model should strive to reflect. However, curating such principles across many domains, even non-exhaustively, requires a labor-intensive annotation process. To automate this process, we propose eliciting these latent attributes guiding model reasoning towards human-preferred responses by explicitly modeling them in a self-correction setting. Our approach mines new principles from the LM itself and compresses the discovered elements to an interpretable set via clustering. Specifically, we employ an approximation of posterior-regularized Monte Carlo Expectation-Maximization to both identify a condensed set of the most effective latent principles and teach the LM to strategically invoke them in order to intrinsically refine its responses. We demonstrate that bootstrapping our algorithm over multiple iterations enables smaller language models (7-8B parameters) to self-improve, achieving +8-10\% in AlpacaEval win-rate, an average of +0.3 on MT-Bench, and +19-23\% in principle-following win-rate on IFEval. We also show that clustering the principles yields interpretable and diverse model-generated constitutions while retaining model performance. The gains our method achieves highlight the potential of automated, principle-driven post-training recipes toward continual self-improvement.



Paperid:1298
Authors:Yuze Hao, Linchao Zhu, Yi Yang
Abstract:
Inverse design aims to design the input variables of a physical system to optimize a specified objective function, typically formulated as a search or optimization problem. However, in 3D domains, the design space grows exponentially, rendering exhaustive grid-based searches infeasible. Recent advances in deep learning have accelerated inverse design by providing powerful generative priors and differentiable surrogate models. Nevertheless, current methods tend to approximate the 3D design space using 2D projections or fine-tune existing 3D shapes. These approaches sacrifice volumetric detail and constrain design exploration, preventing true 3D design from scratch.In this paper, we propose a 3D Inverse Design (3DID) framework that directly navigates the 3D design space by coupling a continuous latent representation with a physics-aware optimization strategy. We first learn a unified physics–geometry embedding that compactly captures shape and physical field data in a continuous latent space. Then, we introduce a two-stage strategy to perform physics-aware optimization. In the first stage, a gradient-guided diffusion sampler explores the global latent manifold. In the second stage, an objective-driven, topology-preserving refinement further sculpts each candidate toward the target objective. This enables 3DID to generate high-fidelity 3D geometries, outperforming existing methods in both solution quality and design versatility.



Paperid:1299
Authors:Polina Kirichenko, Mark Ibrahim, Kamalika Chaudhuri, Samuel J. Bell
Abstract:
For Large Language Models (LLMs) to be reliably deployed in both everyday and high-stakes domains, knowing when not to answer is equally critical as answering correctly.Real-world user queries, which can be underspecified, ill-posed, or fundamentally unanswerable, require LLMs to reason about uncertainty and selectively abstain---i.e., refuse to answer definitively.However, abstention remains understudied, without a systematic evaluation framework for modern LLMs.In this work, we introduce AbstentionBench: a large-scale benchmark for holistically evaluating abstention across 20 diverse datasets, including questions with unknown answers, underspecification, false premises, subjective interpretations, and outdated information.Evaluating 20 frontier LLMs reveals abstention is an unsolved problem, and one where scaling models is of little use.While recent reasoning LLMs have shown impressive results in complex problem solving, surprisingly, we find that reasoning fine-tuning degrades abstention (by 24\% on average), even for math and science domains on which reasoning models are explicitly trained.We find that while a carefully crafted system prompt can boost abstention in practice, it does not resolve models’ fundamental inability to reason about uncertainty.We release AbstentionBench to foster research into advancing LLM reliability.



Paperid:1300
Authors:Jingkun Yue, Siqi Zhang, Zinan Jia, Huihuan Xu, Zongbo Han, Xiaohong Liu, Guangyu Wang
Abstract:
Visual grounding is essential for precise perception and reasoning in multimodal large language models (MLLMs), especially in medical imaging domains. While existing medical visual grounding benchmarks primarily focus on single-image scenarios, real-world clinical applications often involve sequential images, where accurate lesion localization across different modalities and temporal tracking of disease progression (e.g., pre- vs. post-treatment comparison) require fine-grained cross-image semantic alignment and context-aware reasoning. To remedy the underrepresentation of image sequences in existing medical visual grounding benchmarks, we propose MedSG-Bench, the first benchmark tailored for Medical Image Sequences Grounding. It comprises eight VQA-style tasks, formulated into two paradigms of the grounding tasks, including 1) Image Difference Grounding, which focuses on detecting change regions across images, and 2) Image Consistency Grounding, which emphasizes detection of consistent or shared semantics across sequential images. MedSG-Bench covers 76 public datasets, 10 medical imaging modalities, and a wide spectrum of anatomical structures and diseases, totaling 9,630 question–answer pairs. We benchmark both general-purpose MLLMs (e.g., Qwen2.5-VL) and medical-domain specialized MLLMs (e.g., HuatuoGPT-vision), observing that even the advanced models exhibit substantial limitations in medical sequential grounding tasks. To advance this field, we construct MedSG-188K, a large-scale instruction-tuning dataset tailored for sequential visual grounding, and further develop MedSeq-Grounder, an MLLM designed to facilitate future research on fine-grained understanding across medical sequential images. We release all resources on https://anonymous.4open.science/r/test-ABC123



Paperid:1301
Authors:Yue Hou, He Zhu, Ruomei Liu, Yingke Su, Junran Wu, Ke Xu
Abstract:
Distributional discrepancy between training and test data can lead models to make inaccurate predictions when encountering out-of-distribution (OOD) samples in real-world applications. Although existing graph OOD detection methods leverage data-centric techniques to extract effective representations, their performance remains compromised by structural redundancy that induces semantic shifts. To address this dilemma, we propose RedOUT, an unsupervised framework that integrates structural entropy into test-time OOD detection for graph classification. Concretely, we introduce the Redundancy-aware Graph Information Bottleneck (ReGIB) and decompose the objective into essential information and irrelevant redundancy. By minimizing structural entropy, the decoupled redundancy is reduced, and theoretically grounded upper and lower bounds are proposed for optimization. Extensive experiments on real-world datasets demonstrate the superior performance of RedOUT on OOD detection. Specifically, our method achieves an average improvement of 6.7\%, significantly surpassing the best competitor by 17.3\% on the ClinTox/LIPO dataset pair.



Paperid:1302
Authors:Jiaoda Li, Ryan Cotterell
Abstract:
Transformer-based language models (LMs) have achieved widespread empirical success, but their theoretical expressive power remains only partially understood. Prior work often relies on idealized models with assumptions---such as arbitrary numerical precision and hard attention---that diverge from real-world transformers. In this work, we provide an exact characterization of fixed-precision transformers with strict future masking and soft attention, an idealization that more closely mirrors practical implementations. We show that these models are precisely as expressive as a specific fragment of linear temporal logic that includes only a single temporal operator: the past operator. We further relate this logic to established classes in formal language theory, automata theory, and algebra, yielding a rich and unified theoretical framework for understanding transformer expressivity. Finally, we present empirical results that align closely with our theory: transformers trained on languages within their theoretical capacity generalize perfectly over lengths, while they consistently fail to generalize on languages beyond it.



Paperid:1303
Authors:Vincent de Bakker, Joey Hejna, Tyler Lum, Onur Celik, Aleksandar Taranovic, Denis Blessing, Gerhard Neumann, Jeannette Bohg, Dorsa Sadigh
Abstract:
Dexterous robotic hands are essential for performing complex manipulation tasks, yet remain difficult to train due to the challenges of demonstration collection and high-dimensional control. While reinforcement learning (RL) can alleviate the data bottleneck by generating experience in simulation, it typically relies on carefully designed, task-specific reward functions, which hinder scalability and generalization. Thus, contemporary works in dexterous manipulation have often bootstrapped from reference trajectories. These trajectories specify target hand poses that guide the exploration of RL policies and object poses that enable dense, task-agnostic rewards.However, sourcing suitable trajectories---particularly for dexterous hands---remains a significant challenge. Yet, the precise details in explicit reference trajectories are often unnecessary, as RL ultimately refines the motion. Our key insight is that modern vision-language models (VLMs) already encode the commonsense spatial and semantic knowledge needed to specify tasks and guide exploration effectively. Given a task description (e.g., “open the cabinet”) and a visual scene, our method uses an off-the-shelf VLM to first identify task-relevant keypoints (e.g., handles, buttons) and then synthesize 3D trajectories for hand motion and object motion. Subsequently, we train a low-level residual RL policy in simulation to track these coarse trajectories or ``scaffolds'' with high fidelity. Across a number of simulated tasks involving articulated objects and semantic understanding, we demonstrate that our method is able to learn robust dexterous manipulation policies. Moreover, we showcase that our method transfers to real-world robotic hands without any human demonstrations or handcrafted rewards.



Paperid:1304
Authors:Yong Liu, Wenpeng Xiao, Qianqian Wang, Junlin Chen, Shiyin Wang, Yitong Wang, Xinglong Wu, Yansong Tang
Abstract:
We introduce a model named DreamLight for universal image relighting in this work, which can seamlessly composite subjects into a new background while maintaining aesthetic uniformity in terms of lighting and color tone. The background can be specified by natural images (image-based relighting) or generated from unlimited text prompts (text-based relighting). Existing studies primarily focus on image-based relighting, while with scant exploration into text-based scenarios. Some works employ intricate disentanglement pipeline designs relying on environment maps to provide relevant information, which grapples with the expensive data cost required for intrinsic decomposition and light source. Other methods take this task as an image translation problem and perform pixel-level transformation with autoencoder architecture. While these methods have achieved decent harmonization effects, they struggle to generate realistic and natural light interaction effects between the foreground and background. To alleviate these challenges, we reorganize the input data into a unified format and leverage the semantic prior provided by the pretrained diffusion model to facilitate the generation of natural results. Moreover, we propose a Position-Guided Light Adapter (PGLA) that condenses light information from different directions in the background into designed light query embeddings, and modulates the foreground with direction-biased masked attention. In addition, we present a post-processing module named Spectral Foreground Fixer (SFF) to adaptively reorganize different frequency components of subject and relighted background, which helps enhance the consistency of foreground appearance. Extensive comparisons and user study demonstrate that our DreamLight achieves remarkable relighting performance.



Paperid:1305
Authors:Rongkun Zheng, Lu Qi, Xi Chen, Yi Wang, Kun Wang, Hengshuang Zhao
Abstract:
While visual autoregressive modeling (VAR) strategies have shed light on image generation with the autoregressive models, their potential for segmentation, a task that requires precise low-level spatial perception, remains unexplored. Inspired by the multi-scale modeling of classic Mask2Former-based models, we propose Seg-VAR, a novel framework that rethinks segmentation as a conditional autoregressive mask generation problem. This is achieved by replacing the discriminative learning with the latent learning process. Specifically, our method incorporates three core components: (1) an image encoder generating latent priors from input images, (2) a spatial-aware seglat (a latent expression of segmentation mask) encoder that maps segmentation masks into discrete latent tokens using a location-sensitive color mapping to distinguish instances, and (3) a decoder reconstructing masks from these latents. A multi-stage training strategy is introduced: first learning seglat representations via image-seglat joint training, then refining latent transformations, and finally aligning image-encoder-derived latents with seglat distributions. Experiments show Seg-VAR outperforms previous discriminative and generative methods on various segmentation tasks and validation benchmarks. By framing segmentation as a sequential hierarchical prediction task, Seg-VAR opens new avenues for integrating autoregressive reasoning into spatial-aware vision systems.



Paperid:1306
Authors:Yuning Wu, Jiahao Mei, Ming Yan, Chenliang Li, Shaopeng Lai, Yuran Ren, Zijia Wang, Ji Zhang, Mengyue Wu, Qin Jin, Fei Huang
Abstract:
Recent advancements in large language models (LLMs) have significantly enhanced text generation capabilities, yet evaluating their performance in generative writing remains a challenge. Existing benchmarks primarily focus on generic text generation or limited in writing tasks, failing to capture the diverse requirements of high-quality written contents across various domains. To bridge this gap, we present WritingBench, a comprehensive benchmark designed to evaluate LLMs across 6 core writing domains and 100 subdomains. We further propose a query-dependent evaluation framework that empowers LLMs to dynamically generate instance-specific assessment criteria. This framework is complemented by a fine-tuned critic model for criteria-aware scoring, enabling evaluations in style, format and length. The framework's validity is further demonstrated by its data curation capability, which enables a 7B-parameter model to outperform the performance of GPT-4o in writing. We open-source the benchmark, along with evaluation tools and modular framework components, to advance the development of LLMs in writing.



Paperid:1307
Authors:Xinpeng Wang, Mingyang Wang, Yihong Liu, Hinrich Schuetze, Barbara Plank
Abstract:
Refusal mechanisms in large language models (LLMs) are essential for ensuring safety. Recent research has revealed that refusal behavior can be mediated by a single direction in activation space, enabling targeted interventions to bypass refusals. While this is primarily demonstrated in an English-centric context, appropriate refusal behavior is important for any language, but poorly understood. In this paper, we investigate the refusal behavior in LLMs across 14 languages using \textit{PolyRefuse}, a multilingual safety dataset created by translating malicious and benign English prompts into these languages. We uncover the surprising cross-lingual universality of the refusal direction: a vector extracted from English can bypass refusals in other languages with near-perfect effectiveness, without any additional fine-tuning. Even more remarkably, refusal directions derived from any safety-aligned language transfer seamlessly to others. We attribute this transferability to the parallelism of refusal vectors across languages in the embedding space and identify the underlying mechanism behind cross-lingual jailbreaks. These findings provide actionable insights for building more robust multilingual safety defenses and pave the way for a deeper mechanistic understanding of cross-lingual vulnerabilities in LLMs.



Paperid:1308
Authors:Jiyuan Tan, Vasilis Syrgkanis, Jose Blanchet
Abstract:
Causal effect estimation seeks to determine the impact of an intervention from observational data. However, the existing causal inference literature primarily addresses treatment effects on frequently occurring events. But what if we are interested in estimating the effects of a policy intervention whose benefits, while potentially important, can only be observed and measured in rare yet impactful events, such as extreme climate events? The standard causal inference methodology is not designed for this type of inference since the events of interest may be scarce in the observed data and some degree of extrapolation is necessary. Extreme Value Theory (EVT) provides methodologies for analyzing statistical phenomena in such extreme regimes. We introduce a novel framework for assessing treatment effects in extreme data to capture the causal effect at the occurrence of rare events of interest. In particular, we employ the theory of multivariate regular variation to model extremities. We develop a consistent estimator for extreme treatment effects and present a rigorous non-asymptotic analysis of its performance. We illustrate the performance of our estimator using both synthetic and semi-synthetic data.



Paperid:1309
Authors:Tian Qiu, Wenda Li, Yang Gao, Jie Lei, Tao Wang, Yi Gao, Mingli Song, Zunlei Feng
Abstract:
Fraudulent activities have caused substantial negative social impacts and are exhibiting emerging characteristics such as intelligence and industrialization, posing challenges of high-order interactions, intricate dependencies, and the sparse yet concealed nature of fraudulent entities. Existing mainstream fraud detectors are limited by their narrow "receptive fields", focusing only on the relations between entities and their neighbors while neglecting long-range structural associations hidden between entities. To address this issue, a novel fraud detector based on Graph Path Aggregation (GPA) is proposed. It operates through variable-length path sampling, behavior-associated path encoding, path interaction and aggregation, and aggregation-enhanced fraud detection. To further facilitate interpretable association analysis, we synthesize G-Internet, the first benchmark dataset in the field of internet fraud detection. Extensive experiments across datasets in multiple fraud scenarios demonstrate that the proposed GPA outperforms mainstream fraud detectors by up to +15\% in Average Precision (AP). Additionally, GPA exhibits enhanced robustness to noisy labels and offers excellent interpretability by uncovering implicit fraudulent patterns across broader contexts.



Paperid:1310
Authors:Viet Nguyen, Changjian Shui, Vijay Giri, Siddharth Arya, Amol Verma, Fahad Razak, Rahul Krishnan
Abstract:
The distribution of data changes over time; models operating operating in dynamic environments need retraining. But knowing when to retrain, without access to labels, is an open challenge since some, but not all shifts degrade model performance. This paper formalizes and addresses the problem of post-deployment deterioration (PDD) monitoring. We propose D3M, a practical and efficient monitoring algorithm based on the disagreement of predictive models, achieving low false positive rates under non-deteriorating shifts and provides sample complexity bounds for high true positive rates under deteriorating shifts. Empirical results on both standard benchmark and a real-world large-scale internal medicine dataset demonstrate the effectiveness of the framework and highlight its viability as an alert mechanism for high-stakes machine learning pipelines.



Paperid:1311
Authors:Chi Zuo, Martin Møller, Pablo Martínez-Nuevo, Huayang Huang, Yu Wu, Ye Zhu
Abstract:
While being disturbed by environmental noises, the acoustic masking technique is a conventional way to reduce the annoyance in audio engineering that seeks to cover up the noises with other dominant yet less intrusive sounds. However, misalignment between the dominant sound and the noise—such as mismatched downbeats—often requires an excessive volume increase to achieve effective masking. Motivated by recent advances in cross-modal generation, in this work, we introduce an alternative method to acoustic masking, aiming to reduce the noticeability of environmental noises by blending them into personalized music generated based on user-provided text prompts. Following the paradigm of music generation using mel-spectrogram representations, we propose a Blending Noises into Personalized Music (BNMusic) framework with two key stages. The first stage synthesizes a complete piece of music in a mel-spectrogram representation that encapsulates the musical essence of the noise. In the second stage, we adaptively amplifying the generated music segment to further reduce noise perception and enhance the blending effectiveness, while preserving auditory quality. Our experiments with comprehensive evaluations on MusicBench, EPIC-SOUNDS, and ESC-50 demonstrate the effectiveness of our framework, highlighting the ability to blend environmental noise with rhythmically aligned, adaptively amplified, and enjoyable music segments, minimizing the noticeability of the noise, thereby improving overall acoustic experiences.



Paperid:1312
Authors:Siyi Chen, Yixuan Jia, Qing Qu, He Sun, Jeffrey Fessler
Abstract:
Data assimilation (DA) integrates observations with a dynamical model to estimate states of PDE-governed systems. Model-driven methods (e.g., Kalman, particle) presuppose full knowledge of the true dynamics, which is not always satisfied in practice, while purely data-driven solvers learn a deterministic mapping between observations and states and therefore miss the intrinsic stochasticity of real processes. Recently, Score-based diffusion models learn a global diffusion prior and provide a good modelling of the stochastic dynamics, showing new potential for DA. However, their all-at-once generation rather than the step-by-step transition limits their performance when dealing with highly complex stochastic processes and lacks physical interpretability. To tackle these drawbacks, we introduce FlowDAS, a generative DA framework that uses stochastic interpolants to directly learn state transition dynamics and achieve step-by-step transition to better model the real dynamics. We also improve the framework by combining the observation, better suiting the DA settings. Directly learning the underlying dynamics from collected data removes restrictive dynamical assumptions, and conditioning on observations at each interpolation step yields stable, measurement-consistent forecasts. Experiments on Lorenz-63, Navier–Stokes super-resolution/sparse-observation scenarios, and large-scale weather forecasting—where dynamics are partly or wholly unknown—show that FlowDAS surpasses model-driven methods, neural operators, and score-based baselines in accuracy and physical plausibility.



Paperid:1313
Authors:Arshak Hovhannisyan, Armen Allahverdyan
Abstract:
Simpson's paradox poses a challenge in probabilistic inference and decision-making. Our study revisits the paradox by re-estimating its frequency with an unbiased data generation process and reaffirms that it is not an artifact of deficient data collection. Thus, it can lead to incorrect recommendations in fields as diverse as statistics, psychology, and artificial intelligence. We show that the paradox can be resolved by assuming a minimal — though not necessarily observed — common cause (or screening) variable for the involved random variables. In our approach, conditioning on this minimal common cause establishes the correct association between events, which coincides with the conditioning option of the original Simpson paradox. This resolution applies to both discrete cases of binary variables and continuous settings modeled by Gaussian variables. For a non-minimal common cause, the resolution of the paradox is possible, but detailed knowledge of the common cause is required. Our findings extend traditional understandings of the paradox and offer practical guidance for resolving apparent contradictions in probabilistic inference, ultimately enhancing decision-making processes. This point is illustrated by several examples.



Paperid:1314
Authors:Viacheslav Meshchaninov, Egor Chimbulatov, Alexander Shabalin, Aleksandr Abramov, Dmitry Vetrov
Abstract:
Abstract:Autoregressive language models dominate modern text generation, yet their sequential nature introduces fundamental limitations: decoding is slow, and maintaining global coherence remains challenging. Diffusion models offer a promising alternative by enabling parallel generation and flexible control; however, their application to text generation is hindered by the high dimensionality of token-level representations. We introduce Cosmos, a novel approach to text generation that operates entirely in a compressed, smooth latent space tailored specifically for diffusion. This space is learned using an autoencoder trained simultaneously for token-level reconstruction and alignment with frozen activations from a pretrained language encoder, providing robust semantic grounding and enabling effective perturbation‑based augmentations. Empirically, we demonstrate that text representations can be compressed by $8\times$ while maintaining generation quality comparable to token‑level diffusion models. Furthermore, increasing the latent sequence length allows Cosmos to surpass both diffusion‑based and autoregressive baselines. We evaluate Cosmos on four diverse generative tasks including story generation, question generation, summarization, and detoxification and compare it with various generative paradigms. Cosmos achieves comparable or superior generation quality while offering more than $2\times$ faster inference.



Paperid:1315
Authors:Samuel Robertson, Thang Chu, Bo Dai, Dale Schuurmans, Csaba Szepesvari, Jincheng Mei
Abstract:
Abstract:We prove that the classic REINFORCE stochastic policy gradient (SPG) method converges to globally optimal policies in finite-horizon Markov Decision Processes (MDPs) with $\textit{any}$ constant learning rate. To avoid the need for small or decaying learning rates, we introduce two key innovations in the stochastic bandit setting, which we then extend to MDPs. $\textbf{First}$, we identify a new exploration property of SPG: the online SPG method samples every action infinitely often (i.o.), improving on previous results that only guaranteed at least two actions would be sampled i.o. This means SPG inherently achieves asymptotic exploration without modification. $\textbf{Second}$, we eliminate the assumption of unique mean reward values, a condition that previous convergence analyses relied on in the bandit setting, but that is unreasonable in MDPs. Our results deepen the theoretical understanding of SPG in both bandit problems and MDPs, with a focus on how it handles the exploration-exploitation trade-off when standard optimization and stochastic approximation methods cannot be applied, as is the case with large constant learning rates.



Paperid:1316
Authors:Gregory Dexter, Shao Tang, Ata Fatahibaarzi, Qingquan Song, Tejas Dharamsi, Aman Gupta
Abstract:
Abstract:The efficient deployment of large language models (LLMs) in online settings requires optimizing inference performance under stringent latency constraints, particularly the time-to-first-token (TTFT) and time-per-output-token (TPOT). This paper focuses on the query scheduling problem for LLM inference with prefix reuse, a technique that leverages shared prefixes across queries to reduce computational overhead. Our work reveals previously unknown limitations of the existing first-come-first-serve (FCFS) and longest-prefix-match (LPM) scheduling strategies with respect to satisfying latency constraints. We present a formal theoretical framework for LLM query scheduling under RadixAttention, a prefix reuse mechanism that stores and reuses intermediate representations in a radix tree structure. Our analysis establishes the NP-hardness of the scheduling problem with prefix reuse under TTFT constraints and proposes a novel scheduling algorithm, $k$-LPM, which generalizes existing methods by balancing prefix reuse and fairness in query processing. Theoretical guarantees demonstrate that $k$-LPM achieves improved TTFT performance under realistic traffic patterns captured by a data generative model. Empirical evaluations in a realistic serving setting validates our findings, showing significant reductions in P99 TTFT compared to baseline methods.



Paperid:1317
Authors:Peng Lu, Jerry Huang, QIUHAO Zeng, Xinyu Wang, Boxing Chen, Philippe Langlais, Yufei CUI
Abstract:
Abstract:The quadratic complexity of the attention mechanism in Transformer models has motivated the development of alternative architectures with sub-quadratic scaling, such as state-space models. Among these, Mamba has emerged as a leading architecture, achieving state-of-the-art results across a range of language modeling tasks. However, Mamba’s performance significantly deteriorates when applied to contexts longer than those seen during pre-training, revealing a sharp sensitivity to context length extension. Through detailed analysis, we attribute this limitation to the out-of-distribution behavior of its state-space dynamics, particularly within the parameterization of the state transition matrix $A$. Unlike recent works which attribute this sensitivity to the vanished accumulation of discretization time steps, $\exp(-\sum_{t=1}^N{\Delta}_t)$, we establish a connection between state convergence behavior as the input length approaches infinity and the spectrum of the transition matrix $A$, offering a well-founded explanation of its role in length extension. Next, to overcome this challenge, we propose an approach that applies spectrum scaling to pre-trained Mamba models to enable robust long-context generalization by selectively modulating the spectrum of $A$ matrices in each layer. We show that this can significantly improve performance in settings where simply modulating ${\Delta}_t$ fails, validating our insights and providing avenues for better length generalization of state-space models with structured transition matrices.



Paperid:1318
Authors:Stephan Rabanser, Nathalie Rauschmayr, Achin Kulshrestha, Petra Poklukar, Wittawat Jitkrittum, Sean Augenstein, Congchao Wang, Federico Tombari
Abstract:
Large-scale machine learning models deliver strong performance across a wide range of tasks but come with significant computational and resource constraints. To mitigate these challenges, local smaller models are often deployed alongside larger models, relying on routing and deferral mechanisms to offload complex tasks. However, existing approaches inadequately balance the capabilities of these models, often resulting in unnecessary deferrals or sub-optimal resource usage. In this work, we introduce a novel loss function called Gatekeeper for calibrating smaller models in cascade setups. Our approach fine-tunes the smaller model to confidently handle tasks it can perform correctly while deferring complex tasks to the larger model. Moreover, it incorporates a mechanism for managing the trade-off between model performance and deferral accuracy and is broadly applicable across various tasks and domains without any architectural changes. We evaluated our method on encoder-only, decoder-only, and encoder-decoder architectures. Experiments across image classification, language modeling, and vision-language tasks show that our approach substantially improves deferral performance.



Paperid:1319
Authors:Yuli Slavutsky, David Blei
Abstract:
Neural networks make accurate predictions but often fail to provide reliable uncertainty estimates, especially when test-time covariates differ from those seen during training, as occurs with selection bias or shifts over time. To address this, we propose a Bayesian framework for uncertainty estimation that explicitly accounts for covariate shifts. Unlike conventional approaches that rely on fixed priors, a key idea of our method is an adaptive prior, conditioned on both training and new covariates. This prior naturally increases uncertainty for inputs that lie far from the training distribution in regions where predictive performance is likely to degrade. To efficiently approximate the resulting posterior predictive distribution, we employ amortized variational inference. Finally, we construct synthetic environments by drawing small bootstrap samples from the training data, simulating a range of plausible covariate shifts using only the original dataset. We evaluate our method on both synthetic and real-world data, demonstrating that it yields substantially improved uncertainty estimates under distribution shift compared to existing approaches.



Paperid:1320
Authors:Lucas N. Alegre, Ana Bazzan, Andre Barreto, Bruno Silva
Abstract:
Multi-task reinforcement learning aims to quickly identify solutions for new tasks with minimal or no additional interaction with the environment. Generalized Policy Improvement (GPI) addresses this by combining a set of base policies to produce a new one that is at least as good—though not necessarily optimal—as any individual base policy. Optimality can be ensured, particularly in the linear-reward case, via techniques that compute a Convex Coverage Set (CCS). However, these are computationally expensive and do not scale to complex domains. The Option Keyboard (OK) improves upon GPI by producing policies that are at least as good—and often better. It achieves this through a learned meta-policy that dynamically combines base policies. However, its performance critically depends on the choice of base policies. This raises a key question: is there an optimal set of base policies—an optimalbehavior basis—that enables zero-shot identification of optimal solutions foranylinear tasks? We solve this open problem by introducing a novel method that efficiently constructs such an optimal behavior basis. We show that it significantly reduces the number of base policies needed to ensure optimality in new tasks. We also prove that it is strictly more expressive than a CCS, enabling particular classes ofnon-lineartasks to be solved optimally. We empirically evaluate our technique in challenging domains and show that it outperforms state-of-the-art approaches, increasingly so as task complexity increases.



Paperid:1321
Authors:Xiangqi Wang, Yue Huang, Yanbo Wang, Xiaonan Luo, Kehan Guo, Yujun Zhou, Xiangliang Zhang
Abstract:
LLMs often need effective configurations, like temperature and reasoning steps, to handle tasks requiring sophisticated reasoning and problem-solving, ranging from joke generation to mathematical reasoning. Existing prompting approaches usually adopt general-purpose, fixed configurations that work “well enough” across tasks but seldom achieve task-specific optimality.To address this gap, we introduce AdaReasoner, an LLM-agnostic plugin designed for any LLM to automate adaptive reasoning configurations for tasks requiring different types of thinking. AdaReasoner is trained using a reinforcement learning (RL) framework, combining a factorized action space with a targeted exploration strategy, along with a pretrained reward model to optimize the policy model for reasoning configurations with only a few-shot guide.AdaReasoner is backed by theoretical guarantees and experiments of fast convergence and a sublinear policy gap. Across six different LLMs and a variety of reasoning tasks, it consistently outperforms standard baselines, preserves out-of-distribution robustness, and yield gains on knowledge-intensive tasks through tailored prompts.



Paperid:1322
Authors:Yangchao Wu, Zongyue Qin, Alex Wong, Stefano Soatto
Abstract:
Speculative decoding is a technique to leverage hardware concurrency to improve the efficiency of large-scale autoregressive (AR) Transformer models by enabling multiple steps of token generation in a single forward pass. State-space models (SSMs) are already more efficient than AR Transformers, since their state summarizes all past data with no need to cache or re-process tokens in the sliding window context. However, their state can also comprise thousands of tokens; so, speculative decoding has recently been extended to SSMs. Existing approaches, however, do not leverage the tree-based verification methods, since current SSMs lack the means to compute a token tree efficiently. We propose the first scalable algorithm to perform tree-based speculative decoding in state-space models (SSMs) and hybrid architectures of SSMs and Transformer layers. We exploit the structure of accumulated state transition matrices to facilitate tree-based speculative decoding with minimal overhead to current SSM state update implementations. With the algorithm, we describe a hardware-aware implementation that improves naive application of AR Transformer tree-based speculative decoding methods to SSMs. Furthermore, we outperform vanilla speculative decoding with SSMs even with a baseline drafting model and tree structure on three different benchmarks, opening up opportunities for further speed up with SSM and hybrid model inference. Code will be released upon paper acceptance.



Paperid:1323
Authors:Devan Shah, Shlomo Fortgang, Sofiia Druchyna, Elad Hazan
Abstract:
We present the first provable method for identifying symmetric linear dynamical systems (LDS) with accuracy guarantees that are independent of the system’s state dimension or effective memory. Our approach builds upon recent work that represents symmetric LDSs as convolutions learnable via fixed spectral transformations. We show how to invert this representation—recovering an LDS model from its spectral transform—yielding an end-to-end convex optimization procedure. This distillation preserves predictive accuracy while enabling constant-time and constant-space inference per token, independent of sequence length. We evaluate our method, SpectraLDS, as a component in sequence prediction architectures and demonstrate that accuracy is preserved while inference efficiency is improved on tasks such as language modeling.



Paperid:1324
Authors:Yujin Chen, Yinyu Nie, Benjamin Ummenhofer, Reiner Birkl, Michael Paulitsch, Matthias Niessner
Abstract:
We present PBR-SR, a novel method for physically based rendering (PBR) texture super resolution (SR). It outputs high-resolution, high-quality PBR textures from low-resolution (LR) PBR input in a zero-shot manner. PBR-SR leverages an off-the-shelf super-resolution model trained on natural images, and iteratively minimizes the deviations between super-resolution priors and differentiable renderings. These enhancements are then back-projected into the PBR map space in a differentiable manner to produce refined, high-resolution textures.To mitigate view inconsistencies and lighting sensitivity, which is common in view-based super-resolution, our method applies 2D prior constraints across multi-view renderings, iteratively refining the shared, upscaled textures. In parallel, we incorporate identity constraints directly in the PBR texture domain to ensure the upscaled textures remain faithful to the LR input. PBR-SR operates without any additional training or data requirements, relying entirely on pretrained image priors. We demonstrate that our approach produces high-fidelity PBR textures for both artist-designed and AI-generated meshes, outperforming both direct SR models application and prior texture optimization methods. Our results show high-quality outputs in both PBR and rendering evaluations, supporting advanced applications such as relighting.



Paperid:1325
Authors:Tianyu Guo, Hongyu Chen, Hao Liang, Meiyi Qiang, Bohan Zeng, Linzhuang Sun, Bin CUI, Wentao Zhang
Abstract:
Automatic audio captioning is essential for audio understanding, enabling applications such as accessibility and content indexing. However, evaluating the quality of audio captions remains a major challenge, especially in reference-free settings where high-quality ground-truth captions are unavailable. While CLAPScore is currently the most widely used reference-free Audio Caption Evaluation Metric(ACEM), its robustness under diverse conditions has not been systematically validated. To address this gap, we introduce BRACE, a new benchmark designed to evaluate audio caption alignment quality in a reference-free setting. BRACE is primarily designed for assessing ACEMs, and can also be extended to measure the modality alignment abilities of Large Audio Language Model(LALM). BRACE consists of two sub-benchmarks: BRACE-Main for fine-grained caption comparison and BRACE-Hallucination for detecting subtle hallucinated content. We construct these datasets through high-quality filtering, LLM-based corruption, and human annotation. Given the widespread adoption of CLAPScore as a reference-free ACEM and the increasing application of LALMs in audio-language tasks, we evaluate both approaches using the BRACE benchmark, testing CLAPScore across various CLAP model variants and assessing multiple LALMs. Notably, even the best-performing CLAP-based ACEM achieves only a 70.01 F1-score on the BRACE-Main benchmark, while the best LALM reaches just 63.19. By revealing the limitations of CLAP models and LALMs, our BRACE benchmark offers valuable insights into the direction of future research. Our evaluation code and benchmark dataset are released in https://github.com/HychTus/BRACEEvaluation and https://huggingface.co/datasets/gtysssp/audiobenchmarks.



Paperid:1326
Authors:Minchan Kwon, Junwon Ko, kangil kim, Junmo Kim
Abstract:
Direct Preference Optimization (DPO) is a powerful paradigm to align language models with human preferences using pairwise comparisons. However, its binary win-or-loss supervision often proves insufficient for training small models with limited capacity.Prior works attempt to distill information from large teacher models using behavior cloning or KL divergence.These methods often focus on mimicking current behavior and overlook distilling reward modeling.To address this issue, we propose \textit{Teacher Value-based Knowledge Distillation} (TVKD), which introduces an auxiliary reward from the value function of the teacher model to provide a soft guide. This auxiliary reward is formulated to satisfy potential-based reward shaping, ensuring that the global reward structure and optimal policy of DPO are preserved. TVKD can be integrated into the standard DPO training framework and does not require additional rollouts. Our experimental results show that TVKD consistently improves performance across various benchmarks and model sizes.



Paperid:1327
Authors:Wenxuan Bao, Shan Jin, Hadi Abdullah, Anderson Nascimento, Vincent Bindschaedler, Yiwei Cai
Abstract:
Machine learning models are vulnerable to privacy attacks due to their tendency to memorize individual training set examples. Theoretically-sound defenses such as differential privacy can defend against this threat, but model performance often suffers. Empirical defenses may thwart existing attacks while maintaining model performance but do not offer any robust theoretical guarantees.In this paper, we explore a new strategy based on the concept of plausible deniability. We introduce a training algorithm called Plausibly Deniable Stochastic Gradient Descent (PD-SGD), which aims to provide both strong privacy protection with theoretical justification and maintain high performance. The core of this approach is a rejection sampling technique, which probabilistically prevents updating model parameters whenever a mini-batch cannot be plausibly denied. This ensures that no individual example has a disproportionate influence on the model parameters. We provide a set of theoretical results showing that PD-SGD effectively mitigates privacy leakage from individual data points. Experiments also demonstrate that PD-SGD offers a favorable trade-off between privacy and utility compared to existing defense methods.



Paperid:1328
Authors:William Réveillard, Richard Combes
Abstract:
Abstract:We consider a stochastic multi-armed bandit problem with i.i.d. rewards where the expected reward function is multimodal with at most $m$ modes. We propose the first known computationally tractable algorithm for computing the solution to the Graves-Lai optimization problem, which in turn enables the implementation of asymptotically optimal algorithms for this bandit problem.



Paperid:1329
Authors:Yixuan Ma, Kai Yi, Pietro Lió, Shi Jin, Yuguang Wang
Abstract:
Hypergraphs effectively model higher-order relationships in natural phenomena, capturing complex interactions beyond pairwise connections. We introduce a novel hypergraph message passing framework inspired by interacting particle systems, where hyperedges act as fields inducing shared node dynamics. By incorporating attraction, repulsion, and Allen-Cahn forcing terms, particles of varying classes and features achieve class-dependent equilibrium, enabling separability through the particle-driven message passing. We investigate both first-order and second-order particle system equations for modeling these dynamics, which mitigate over-smoothing and heterophily thus can capture complete interactions. The more stable second-order system permits deeper message passing. Furthermore, we enhance deterministic message passing with stochastic element to account for interaction uncertainties. We prove theoretically that our approach mitigates over-smoothing by maintaining a positive lower bound on the hypergraph Dirichlet energy during propagation and thus to enable hypergraph message passing to go deep. Empirically, our models demonstrate competitive performance on diverse real-world hypergraph node classification tasks, excelling on both homophilic and heterophilic datasets. Source code is available at https://anonymous.4open.science/r/HAMP-7EE8/.



Paperid:1330
Authors:Miranda Christ, Daniel Reichman, Jonathan Shafer
Abstract:
We study protocols for verifying approximate optimality of strategies in multi-armed bandits and normal-form games. As the number of actions available to each player is often large, we seek protocols where the number of queries to the utility oracle is sublinear in the number of actions. We prove that such verification is possible for sufficiently smooth strategies that do not put too much probability mass on any specific action and provide protocols for verifying that a smooth policy for a multi-armed bandit is close to optimal. Our verification protocols require provably fewer arm queries than learning. Furthermore, we show how to use cryptographic tools to reduce the communication cost of our protocols. We complement our protocol by proving a nearly tight lower bound on the query complexity of verification in our settings. As an application, we use our bandit verification protocol to build a protocol for verifying approximate optimality of a strong smooth Nash equilibrium, with sublinear query complexity.



Paperid:1331
Authors:Tomer Raz, Michael Shalyt, Elyasheev Leibtag, Rotem Kalisch, Shachar Weinbaum, Yaron Hadad, Ido Kaminer
Abstract:
Abstract:The constant $\large \pi$ has fascinated scholars throughout the centuries, inspiring numerous formulas for its evaluation, such as infinite sums and continued fractions. Despite their individual significance, many the underlying connections among formulas remain unknown, missing unifying theories that could unveil deeper understanding. The absence of a unifying theory reflects a broader challenge across math and science: knowledge is typically accumulated through isolated discoveries, while deeper connections often remain hidden. In this work, we present an automated framework for the unification of mathematical formulas. Our system combines large language models (LLMs) for systematic formula harvesting, an LLM-code feedback loop for validation, and a novel symbolic algorithm for clustering and eventual unification. We demonstrate this methodology on the hallmark case of $\large \pi$, an ideal testing ground for symbolic unification. Applying this approach to 455,050 arXiv papers, we validate 407 distinct formulas for $\large \pi$ and prove relations between 381 (94\%) of them, of which 188 (46\%) can be derived from a single mathematical object—linking canonical formulas by Euler, Gauss, Brouncker, and newer ones from algorithmic discoveries by the Ramanujan Machine. Our method generalizes to other constants, including $e$, $\zeta(3)$, and Catalan’s constant, demonstrating the potential of AI-assisted mathematics to uncover hidden structure and unify knowledge across domains.



Paperid:1332
Authors:Akifumi Wachi, Kohei Miyaguchi, Takumi Tanabe, Rei Sato, Youhei Akimoto
Abstract:
A longstanding goal in safe reinforcement learning (RL) is a method to ensure the safety of a policy throughout the entire process, from learning to operation. However, existing safe RL paradigms inherently struggle to achieve this objective. We propose a method, called Provably Lifetime Safe RL (PLS), that integrates offline safe RL with safe policy deployment to address this challenge. Our proposed method learns a policy offline using return-conditioned supervised learning and then deploys the resulting policy while cautiously optimizing a limited set of parameters, known as target returns, using Gaussian processes (GPs). Theoretically, we justify the use of GPs by analyzing the mathematical relationship between target and actual returns. We then prove that PLS finds near-optimal target returns while guaranteeing safety with high probability. Empirically, we demonstrate that PLS outperforms baselines both in safety and reward performance, thereby achieving the longstanding goal to obtain high rewards while ensuring the safety of a policy throughout the lifetime from learning to operation.



Paperid:1333
Authors:Wei Wan, Ning Yuxuan, Zhicong Huang, Cheng Hong, Shengshan Hu, Ziqi Zhou, Yechao Zhang, Tianqing Zhu, Wanlei Zhou, Leo Yu Zhang
Abstract:
Federated Learning (FL) is a distributed paradigm aimed at protecting participant data privacy by exchanging model parameters to achieve high-quality model training. However, this distributed nature also makes FL highly vulnerable to backdoor attacks. Notably, the recently proposed state-of-the-art (SOTA) attack, 3DFed (SP2023), uses an indicator mechanism to determine whether the backdoor models have been accepted by the defender and adaptively optimizes backdoor models, rendering existing defenses ineffective. In this paper, we first reveal that the failure of existing defenses lies in the employment of empirical statistical measures that are loosely coupled with backdoor attacks. Motivated by this, we propose a Malignity-Aware backdooR defenSe (MARS) that leverages backdoor energy (BE) to indicate the malicious extent of each neuron. To amplify malignity, we further extract the most prominent BE values from each model to form a concentrated backdoor energy (CBE). Finally, a novel Wasserstein distance-based clustering method is introduced to effectively identify backdoor models. Extensive experiments demonstrate that MARS can defend against SOTA backdoor attacks and significantly outperforms existing defenses.



Paperid:1334
Authors:Somangchan Park, Heesang Ann, Min-hwan Oh
Abstract:
Abstract:We study the multi-objective linear contextual bandit problem, where multiple possible conflicting objectives must be optimized simultaneously. We propose $\texttt{MOL-TS}$, the $\textit{first}$ Thompson Sampling algorithm with Pareto regret guarantees for this problem. Unlike standard approaches that compute an empirical Pareto front each round, $\texttt{MOL-TS}$ samples parameters across objectives and efficiently selects an arm from a novel $\textit{effective Pareto front}$, which accounts for repeated selections over time. Our analysis shows that $\texttt{MOL-TS}$ achieves a worst-case Pareto regret bound of $\widetilde{O}(d^{3/2}\sqrt{T})$, where $d$ is the dimension of the feature vectors, $T$ is the total number of rounds, matching the best known order for randomized linear bandit algorithms for single objective. Empirical results confirm the benefits of our proposed approach, demonstrating improved regret minimization and strong multi-objective performance.



Paperid:1335
Authors:Yonatan Ashlag, Uri Koren, Mirco Mutti, Esther Derman, Pierre-Luc Bacon, Shie Mannor
Abstract:
State entropy regularization has empirically shown better exploration and sample complexity in reinforcement learning (RL). However, its theoretical guarantees have not been studied. In this paper, we show that state entropy regularization improves robustness to structured and spatially correlated perturbations. These types of variation are common in transfer learning but often overlooked by standard robust RL methods, which typically focus on small, uncorrelated changes. We provide a comprehensive characterization of these robustness properties, including formal guarantees under reward and transition uncertainty, as well as settings where the method performs poorly. Much of our analysis contrasts state entropy with the widely used policy entropy regularization, highlighting their different benefits. Finally, from a practical standpoint, we illustrate that compared with policy entropy, the robustness advantages of state entropy are more sensitive to the number of rollouts used for policy evaluation.



Paperid:1336
Authors:Yonggan Fu, Xin Dong, Shizhe Diao, Matthijs Van keirsbilck, Hanrong Ye, Wonmin Byeon, Yashaswi Karnati, Lucas Liebenwein, Maksim Khadkevich, Alexander Keller, Jan Kautz, Yingyan (Celine) Lin, Pavlo Molchanov
Abstract:
Abstract:Efficient deployment of small language models (SLMs) is essential for numerous real-world applications with stringent latency constraints. While previous work in SLM design mainly focuses on reducing the number of parameters to deliver parameter-optimal SLMs, parameter efficiency may not necessarily translate into proportional real-device speed-ups. This work aims to provide a systematic exploration and roadmap for latency-optimal SLMs. Our goal is to identify critical determinants of SLMs' real-device latency and provide generalizable principles and methodologies for SLM design and training when real-device latency becomes the primary consideration. Specifically, we first analyze two central architecture design factors: depth-width ratios and the involved operators. We find that although deep-thin models generally lead to better accuracy under the same parameter budget, they may not lie on the frontier of the accuracy-latency trade-off. To identify the latency-optimal depth-width ratio, we augment previous scaling laws by relating model loss to both model depth and width, thus enabling determination of the sweet spot depth-width ratio when combined with device-specific profiling. Additionally, we explore emerging efficient attention alternatives to understand their potential as candidate building operators. Using the identified promising operators, we build an evolutionary search framework to automatically pinpoint optimal latency combinations of these operators into hybrid SLMs to push the accuracy-latency frontier. In addition to architectural improvements, we further analyze and enhance SLM training by enabling more effective weight updates and improving cache initialization, which are generalizable add-on components for future SLMs. Combining these contributions, we introduce a new family of hybrid SLMs, called Fast-SLM, which significantly advances the accuracy–latency trade-off frontier of state-of-the-art SLMs, e.g., achieving 2.57\% higher accuracy, 1.46$\times$ speed-up, and over 10$\times$ cache reduction compared to Llama3.2-3B.



Paperid:1337
Authors:Boshen Xu, Yuting Mei, liu xinbi, Sipeng Zheng, Qin Jin
Abstract:
Egocentric video-language pretraining has significantly advanced video representation learning. Humans perceive and interact with a fully 3D world, developing spatial awareness that extends beyond text-based understanding. However, most previous works learn from 1D text or 2D visual cues, such as bounding boxes, which inherently lack 3D understanding. To bridge this gap, we introduce EgoDTM, an Egocentric Depth- and Text-aware \textbf{M}odel, jointly trained through large-scale 3D-aware video pretraining and video-text contrastive learning. EgoDTM incorporates a lightweight 3D-aware decoder to efficiently learn 3D-awareness from pseudo depth maps generated by depth estimation models. To further facilitate 3D-aware video pretraining, we enrich the original brief captions with hand-object visual cues by organically combining several foundation models. Extensive experiments demonstrate EgoDTM's superior performance across diverse downstream tasks, highlighting its superior 3D-aware visual understanding. Code: \url{https://anonymous.4open.science/r/EgoDTM}.



Paperid:1338
Authors:Xin Su, Sungduk Yu, Phillip Howard, Steven Bethard
Abstract:
Time normalization is the task of converting natural language temporal expressions into machine-readable representations. It underpins many downstream applications in information retrieval, question answering, and clinical decision-making. Traditional systems based on the ISO-TimeML schema limit expressivity and struggle with complex constructs such as compositional, event-relative, and multi-span time expressions. In this work, we introduce a novel formulation of time normalization as a code generation task grounded in the SCATE framework, which defines temporal semantics through symbolic and compositional operators. We implement a fully executable SCATE Python library and demonstrate that large language models (LLMs) can generate executable SCATE code. Leveraging this capability, we develop an automatic data augmentation pipeline using LLMs to synthesize large-scale annotated data with code-level validation. Our experiments show that small, locally deployable models trained on this augmented data can achieve strong performance, outperforming even their LLM parents and enabling practical, accurate, and interpretable time normalization.



Paperid:1339
Authors:Astrit Tola, Funmilola Mary Taiwo, Cuneyt Akcora, Baris Coskunuzer
Abstract:
Graph embeddings play a critical role in graph representation learning, allowing machine learning models to explore and interpret graph-structured data. However, existing methods often rely on opaque, high-dimensional embeddings, limiting interpretability and practical visualization. In this work, we introduce Topological Evolution Rate (TopER), a novel, low-dimensional embedding approach grounded in topological data analysis. TopER simplifies a key topological approach, Persistent Homology, by calculating the evolution rate of graph substructures, resulting in intuitive and interpretable visualizations of graph data. This approach not only enhances the exploration of graph datasets but also delivers competitive performance in graph clustering and classification tasks. Our TopER-based models achieve or surpass state-of-the-art results across molecular, biological, and social network datasets in tasks such as classification, clustering, and visualization.



Paperid:1340
Authors:Kumar Ashutosh, Kristen Grauman
Abstract:
Skill assessment from video entails rating the quality of a person’s physical performance and explaining what could be done better. Today’s models specialize for an individual sport, and suffer from the high cost and scarcity of expert-level supervision across the long tail of sports. Towards closing that gap, we explore transferable video representations for skill assessment. Our CrossTrainer approach discovers skill-attributes—such as balance, control, and hand positioning—whose meaning transcends the boundaries of any given sport, then trains a multimodal language model to generate actionable feedback for a novel video, e.g., “lift hands more to generate more power” as well as its proficiency level, e.g., early expert. We validate the new model on multiple datasets for both cross-sport (transfer) and intra-sport (in-domain) settings, where it achieves gains up to 60% relative to the state of the art. By abstracting out the shared behaviors indicative of human skill, the proposed video representation generalizes substantially better than an array of existing techniques, enriching today’s multimodal large language models.



Paperid:1341
Authors:Mislav Balunovic, Jasper Dekoninck, Ivo Petrov, Nikola Jovanović, Martin Vechev
Abstract:
The rapid advancement of reasoning capabilities in large language models (LLMs) has led to notable improvements on mathematical benchmarks. However, many of the most commonly used evaluation datasets (e.g., AIME 2024) are widely available online, making it difficult to disentangle genuine reasoning from potential memorization. Furthermore, these benchmarks do not evaluate proof-writing capabilities, which are crucial for many mathematical tasks. To address this, we introduce MathArena, a new benchmark based on the following key insight: recurring math competitions provide a stream of high-quality, challenging problems that can be used for real-time evaluation of LLMs. By evaluating models as soon as new problems are released, we effectively eliminate the risk of contamination. Using this framework, we find strong signs of contamination in AIME 2024. Nonetheless, evaluations on harder competitions, such as SMT 2025—published well after model release dates—demonstrate impressive reasoning capabilities in top-performing models. MathArena is also the first benchmark for proof-writing capabilities. On USAMO 2025, even top models score below 25%, far behind their performance on final-answer tasks. So far, we have evaluated 30 models across five competitions, totaling 149 problems. As an evolving benchmark, MathArena will continue to track the progress of LLMs on newly released competitions, ensuring rigorous and up-to-date evaluation of mathematical reasoning.



Paperid:1342
Authors:Pascal Kesseli, Peter O'Hearn, Ricardo Cabral
Abstract:
We present a novel approach to formalise and solve search-based problems using large language models, which significantly improves upon previous state-of-the-art results. We demonstrate the efficacy of this approach on benchmarks like the logic puzzles tasks in ZebraLogicBench. Instead of letting the LLM attempt to directly solve the puzzles, our method prompts the model to formalise the problem in a logic-focused, human-readable domain-specific language (DSL) called Logic.py. This formalised representation is then solved using a constraint solver, leveraging the strengths of both the language model and the solver. Our approach achieves a remarkable 65% absolute improvement over the baseline performance of Llama 3.1 70B on ZebraLogicBench, setting a new state-of-the-art with an accuracy of over 90%. This significant advancement demonstrates the potential of combining language models with domain-specific languages and auxiliary tools on traditionally challenging tasks for LLMs.



Paperid:1343
Authors:Shaohong Wang, Lu Bin, Xinyu Xiao, Hanzhi Zhong, Bowen Pang, Tong Wang, Zhiyu Xiang, Hangguan Shan, Eryun Liu
Abstract:
Collaborative visual perception methods have gained widespread attention in the autonomous driving community in recent years due to their ability to address sensor limitation problems. However, the absence of explicit depth information often makes it difficult for camera-based perception systems, e.g., 3D object detection, to generate accurate predictions. To alleviate the ambiguity in depth estimation, we propose RayFusion, a ray-based fusion method for collaborative visual perception. Using ray occupancy information from collaborators, RayFusion reduces redundancy and false positive predictions along camera rays, enhancing the detection performance of purely camera-based collaborative perception systems. Comprehensive experiments show that our method consistently outperforms existing state-of-the-art models, substantially advancing the performance of collaborative visual perception. Our code will be made publicly available.



Paperid:1344
Authors:Francesco Bacchiocchi, Francesco Emanuele Stradi, Matteo Castiglioni, Alberto Marchesi, Nicola Gatti
Abstract:
In Bayesian persuasion, an informed sender strategically discloses information to a receiver so as to persuade them to undertake desirable actions. Recently, Markov persuasion processes (MPPs) have been introduced to capture sequential scenarios where a sender faces a stream of myopic receivers in a Markovian environment. The MPPs studied so far in the literature suffer from issues that prevent them from being fully operational in practice, e.g., they assume that the sender knows receivers' rewards. We fix such issues by addressing MPPs where the sender has no knowledge about the environment. We design a learning algorithm for the sender, working with partial feedback.We prove that its regret with respect to an optimal information-disclosure policy grows sublinearly in the number of episodes, as it is the case for the loss in persuasiveness cumulated while learning. Moreover, we provide lower bounds for our setting matching the guarantees of our algorithm.



Paperid:1345
Authors:Qingyu Song, Wei Lin, Hong Xu
Abstract:
Learn to Optimize (L2O) trains deep neural network based solvers for optimization, achieving success in accelerating convex problems and improving non-convex solutions. However, L2O lacks rigorous theoretical backing for its own training convergence, as existing analyses often use unrealistic assumptions—--a gap this work highlights empirically. We bridge this gap by proving the training convergence of L2O models that learn Gradient Descent (GD) hyperparameters for quadratic programming, leveraging the Neural Tangent Kernel (NTK) theory. We propose a deterministic initialization strategy to support our theoretical results and promote stable training over extended optimization horizons by mitigating gradient explosion. Our L2O framework demonstrates over 50\% better optimality against GD and superior robustness over state-of-the-art L2O methods on synthetic datasets.



Paperid:1346
Authors:Zixin Zhu, Haoxiang Li, Xuelu Feng, He Wu, Chunming Qiao, Junsong Yuan
Abstract:
Towards intelligent image editing, object removal should eliminate both the target object and its causal visual artifacts, such as shadows and reflections. However, existing image appearance-based methods either follow strictly mask-aligned training and fail to remove these casual effects which are not explicitly masked, or adopt loosely mask-aligned strategies that lack controllability and may unintentionally over-erase other objects. We identify that these limitations stem from ignoring the causal relationship between an object’s geometry presence and its visual effects. To address this limitation, we propose a geometry-aware two-stage framework that decouples object removal into (1) geometry removal and (2) appearance rendering. In the first stage, we remove the object directly from the geometry (e.g., depth) using strictly mask-aligned supervision, enabling structure-aware editing with strong geometric constraints. In the second stage, we render a photorealistic RGB image conditioned on the updated geometry, where causal visual effects are considered implicitly as a result of the modified 3D geometry. To guide learning in the geometry removal stage, we introduce a preference-driven objective based on positive and negative sample pairs, encouraging the model to remove objects as well as their causal visual artifacts while avoiding new structural insertions. Extensive experiments demonstrate that our method achieves state-of-the-art performance in removing both objects and their associated artifacts on two popular benchmarks.



Paperid:1347
Authors:Arshia Afzal, Elias Abad Rocamora, Leyla Candogan, Pol Puigdemont, Francesco Tonin, Yongtao Wu, Mahsa Shoaran, Volkan Cevher
Abstract:
Linear Transformers and State Space Models have emerged as efficient alternatives to softmax Transformers for causal sequence modeling, enabling parallel training via matrix multiplication and efficient RNN-style inference. However, despite their success in causal tasks, no unified framework exists for applying Linear Transformers to bidirectional sequence modeling.We introduce LION, the first framework to systematically extend Linear Transformers to the bidirectional setting. LION generalizes three core representations commonly used in the causal case—full Linear Attention, bidirectional RNN, and chunkwise parallel form—to the bidirectional setting. These forms are theoretically equivalent and enable models to exploit the strengths of each during training and inference.We prove that a broad class of Linear Transformers can be extended using LION and validate our framework via three core examples based on the choice of decay type: LION-Lit, the bidirectional extension of Katharopoulos et al., 2020; LION-D, based on Sun et al., 2023; and LION-S, a variant using selective decay. Across standard bidirectional tasks, LION enables models to match or exceed the performance of softmax Transformers, while offering significantly faster training and more efficient inference than existing State Space Models.



Paperid:1348
Authors:Liyan Chen, Huangying Zhan, Hairong Yin, Yi Xu, Philippos Mordohai
Abstract:
Effective robotic autonomy in unknown environments demands proactive exploration and precise understanding of both geometry and semantics. In this paper, we propose ActiveSGM, an active semantic mapping framework designed to predict the informativeness of potential observations before execution. Built upon a 3D Gaussian Splatting (3DGS) mapping backbone, our approach employs semantic and geometric uncertainty quantification, coupled with a sparse semantic representation, to guide exploration. By enabling robots to strategically select the most beneficial viewpoints, ActiveSGM efficiently enhances mapping completeness, accuracy, and robustness to noisy semantic data, ultimately supporting more adaptive scene exploration. Our experiments on the Replica and Matterport3D datasets highlight the effectiveness of ActiveSGM in active semantic mapping tasks.



Paperid:1349
Authors:Zhikun Chu, Bo Ho, xiaolong zou, Yuanyuan Mi
Abstract:
Sequence schemas are abstract, reusable knowledge structures that facilitate rapid adaptation and generalization in novel sequential tasks. In both animals and humans, shaping is an efficient way for acquiring such schemas, particularly in complex sequential tasks. As a form of curriculum learning, shaping works by progressively advancing from simple subtasks to integrated full sequences, and ultimately enabling generalization across different task variations. Despite the importance of schemas in cognition and shaping in schema acquisition, the underlying neural dynamics at play remain poorly understood. To explore this, we train recurrent neural networks on an odor-sequence task using a shaping protocol inspired by well-established paradigms in experimental neuroscience. Our model provides the first systematic reproduction of key features of schema learning observed in the orbitofrontal cortex, including rapid adaptation to novel tasks, structured neural representation geometry, and progressive dimensionality compression during learning. Crucially, analysis of the trained RNN reveals that the learned schema is implemented through sequence attractors. These attractor dynamics emerge gradually through the shaping process: starting with isolated discrete attractors in simple tasks, evolving into linked sequences, and eventually abstracting into generalizable attractors that capture shared task structure. Moreover, applying our method to a keyword spotting task shows that shaping facilitates the rapid development of sequence attractor-like schemas, leading to enhanced learning efficiency. In summary, our work elucidates a novel attractor-based mechanism underlying schema representation and its evolution via shaping, with the potential to provide new insights into the acquisition of abstract knowledge across biological and artificial intelligence.



Paperid:1350
Authors:Conor Igoe, Tejus Gupta, Jeff Schneider
Abstract:
Recent work in Bayesian Experiment Design (BED) has shown the value of using Deep Learning (DL) to obtain highly efficient adaptive experiment designs. In this paper, we argue that a central bottleneck of DL training for BED is belief explosion. Specifically, as an agent progresses deeper into an experiment, the effective number of realisable beliefs grows enormously, placing significant sampling burdens on offline training schemes in an effort to gather experience from all regions of belief space. We argue that choosing an appropriate inductive bias for actor/critic networks is a critical component in mitigating the effects of belief explosion and has so far been overlooked in the BED literature. We show how Graph Neural Networks are particularly well-suited for BED DL training due to their domain permutation equivariance properties, resulting in multiple orders of magnitude improvement to sample efficiency compared to naive parameterizations.



Paperid:1351
Authors:Arman Zharmagambetov, Chuan Guo, Ivan Evtimov, Maya Pavlova, Ruslan Salakhutdinov, Kamalika Chaudhuri
Abstract:
Autonomous AI agents that can follow instructions and perform complex multi-step tasks have tremendous potential to boost human productivity. However, to perform many of these tasks, the agents need access to personal information from their users, raising the question of whether they are capable of using it appropriately. In this work, we introduce a new benchmarkAgentDAMthat measures if AI web-navigation agents follow the privacy principle of"data minimization". For the purposes of our benchmark, data minimization means that the agent uses a piece of potentially sensitive information only if it is "necessary" to complete a particular task. Our benchmark simulates realistic web interaction scenarios end-to-end and is adaptable to all existing web navigation agents. We use AgentDAM to evaluate how well AI agents built on top of GPT-4, Llama-3 and Claude can limit processing of potentially private information, and show that they are prone to inadvertent use of unnecessary sensitive information. We also propose a prompting-based defense that reduces information leakage, and demonstrate that our end-to-end benchmarking provides a more realistic measure than probing LLMs about privacy. Our results highlight that further research is needed to develop AI agents that can prioritize data minimization at inference time.



Paperid:1352
Authors:Parikshit Bansal, Ali Kavis, Sujay Sanghavi
Abstract:
Contrastive learning involves learning representations via a loss function that encourages each (unlabeled) sample to be far from other samples, but close to its ownaugmentation. In this paper, we aim to understand why this simple idea performs remarkably well, by theoretically analyzing it for a simple, natural problem setting: dimensionality reduction in Gaussian Mixture Models (GMMs). Note that the standard GMM setup lacks the concept of augmentations. We study an intuitive extension: we define the pair of data sample and its augmentation as a coupled random draw from the GMM such that the marginal over the "noisy" augmentation isbiasedtowards the component of the data sample. For this setup, we show that vanilla contrastive loss, e.g., InfoNCE, is able to find theoptimallower-dimensional subspace even when the Gaussian components are non-isotropic. In particular, we show that InfoNCE can match the performance of a fully supervised algorithm, e.g., LDA, (where each data point is labeled with the mixture component it comes from) even when the augmentations are "noisy". We further extend our setup to the multi-modal case, and develop a GMM-like setting to study the contrastive CLIP loss. We corroborate our theoretical with real-data experiments on CIFAR100; representations learned by InfoNCE loss match the performance of LDA on clustering metrics.



Paperid:1353
Authors:Sagnik Bhattacharya, Abhiram Gorle, Ahsan Bilal, Connor Ding, Amit Kumar Singh Yadav, Tsachy Weissman
Abstract:
Generative modeling of non-negative, discrete data, such as symbolic music, remains challenging due to two persistent limitations in existing methods. Firstly, many approaches rely on modeling continuous embeddings, which is suboptimal for inherently discrete data distributions. Secondly, most models optimize variational bounds rather than exact data likelihood, resulting in inaccurate likelihood estimates and degraded sampling quality. While recent diffusion-based models have addressed these issues separately, we tackle them jointly. In this work, we introduce the Information-Theoretic Discrete Poisson Diffusion Model (ItDPDM), inspired by photon arrival process, which combines exact likelihood estimation with fully discrete-state modeling. Central to our approach is an information-theoretic Poisson Reconstruction Loss (PRL) that has a provable exact relationship with the true data likelihood. ItDPDM achieves improved likelihood and sampling performance over prior discrete and continuous diffusion models on a variety of synthetic discrete datasets. Furthermore, on real-world datasets such as symbolic music and images, ItDPDM attains superior likelihood estimates and competitive generation quality—demonstrating a proof of concept for distribution-robust discrete generative modeling.



Paperid:1354
Authors:Zheng HE, Roman Pogodin, Yazhe Li, Namrata Deka, Arthur Gretton, Danica J. Sutherland
Abstract:
Tests of conditional independence (CI) underpin a number of important problems in machine learning and statistics, from causal discovery to evaluation of predictor fairness and out-of-distribution robustness. Shah and Peters (2020) showed that, contrary to the unconditional case, no universally finite-sample valid test can ever achieve nontrivial power. While informative, this result (based on “hiding” dependence) does not seem to explain the frequent practical failures observed with popular CI tests. We investigate the Kernel-based Conditional Independence (KCI) test – of which we show the Generalized Covariance Measure underlying many recent tests isnearlya special case – and identify the major factors underlying its practical behavior. We highlight the key role of errors in the conditional mean embedding estimate for the Type I error, while pointing out the importance of selecting an appropriate conditioning kernel (not recognized in previous work) as being necessary for good test power but also tending to inflate Type I error.



Paperid:1355
Authors:Alexander Goldberg, Giulia Fanti, Nihar Shah
Abstract:
Many decision-making processes involve evaluating and then selecting items; examples include scientific peer review, job hiring, school admissions, and financial investment. Selection typically involves applying rules to evaluations and then deterministically choosing the best candidates. These domains often feature error-prone evaluations and uncertainty about future outcomes, which undermine the reliability of deterministic selection. As a result, selection mechanisms that incorporate uncertainty by involving explicit randomization are beginning to gain traction. However, current randomization methods are ad hoc. In this paper, we propose a principled framework for randomized decision-making based on interval estimates of the quality of each item. We introduce MERIT (Maximin Efficient Randomized Interval Top-k), an optimization-based method that maximizes the worst-case expected number of top candidates selected under "Knightian" uncertainty represented by overlapping intervals. We develop a polynomial-time algorithm to solve the optimization problem and demonstrate empirically that the method scales to over 10,000 items. Further, we prove that our approach can satisfy desirable axiomatic properties not guaranteed by existing approaches to randomization.



Paperid:1356
Authors:Idan Barnea, Tal Lancewicki, Yishay Mansour
Abstract:
Abstract:We study the regret in stochastic Multi-Armed Bandits (MAB) with multiple agents that communicate over an arbitrary connected communication graph.We show an individual regret bound of ${O}(\mathcal{R} / m + A^2 + A \sqrt{\log T})$ and a nearly matching lower bound.Here $A$ is the number of actions, $T$ the time horizon, $m$ the number of agents, and $\mathcal{R} = \sum_{\Delta_i > 0}\log(T)/\Delta_i$ is the optimal single agent regret, where $\Delta_i$ is the sub-optimality gap of action $i$.Our work is the first to show an individual regret bound in cooperative stochastic MAB that is independent of the graph's diameter.When considering communication networks there are additional considerations beyond regret, such as message size and number of communication rounds.First, we show that our regret bound holds even if we restrict the messages to be of logarithmic size.Second, for logarithmic number ofcommunication rounds, we obtain a regret bound of ${O}(\mathcal{R} / m+A \log T)$.



Paperid:1357
Authors:Zhenghao Zhao, Haoxuan Wang, Junyi Wu, Yuzhang Shang, Gaowen Liu, Yan Yan
Abstract:
Abstract:Dataset distillation aims to synthesize a small dataset from a large dataset, enabling the model trained on it to perform well on the original dataset. With the blooming of large language models and multimodal large language models, the importance of multimodal datasets, particularly image-text datasets, has grown significantly. However, existing multimodal dataset distillation methods are constrained by the Matching Training Trajectories algorithm, which significantly increases the computing resource requirement, and takes days to process the distillation. In this work, we introduce EDGE, a generative distillation method for efficient multimodal dataset distillation. Specifically, we identify two key challenges of distilling multimodal datasets with generative models: 1) The lack of correlation between generated images and captions.2) The lack of diversity among generated samples.To address the aforementioned issues, we propose a novel generative model training workflow with a bi-directional contrastive loss and a diversity loss. Furthermore, we propose a caption synthesis strategy to further improve text-to-image retrieval performance by introducing more text information. Our method is evaluated on Flickr30K, COCO, and CC3M datasets, demonstrating superior performance and efficiency compared to existing approaches. Notably, our method achieves results 18$\times$ faster than the state-of-the-art method.



Paperid:1358
Authors:Pei Yang, Hai Ci, Mike Zheng Shou
Abstract:
Graphical User Interface (GUI) agents show promising capabilities for automating computer-use tasks and facilitating accessibility, but existing interactive benchmarks are mostly English-only, covering web-use or Windows, Linux, and Android environments, but not macOS. macOS is a major OS with distinctive GUI patterns and exclusive applications. To bridge the gaps, we present macOSWorld, the first comprehensive benchmark for evaluating GUI agents on macOS. macOSWorld features 202 multilingual interactive tasks across 30 applications (28 macOS-exclusive), with task instructions and OS interfaces offered in 5 languages (English, Chinese, Arabic, Japanese, and Russian). As GUI agents are shown to be vulnerable to deception attacks, macOSWorld also includes a dedicated safety benchmarking subset. Our evaluation on six GUI agents reveals a dramatic gap: proprietary computer-use agents lead at above 30\% success rate, while open-source lightweight research models lag at below 2\%, highlighting the need for macOS domain adaptation. Multilingual benchmarks also expose common weaknesses, especially in Arabic, with a 27.5\% average degradation compared to English. Results from safety benchmarking also highlight that deception attacks are more general and demand immediate attention. Our benchmark will be open-sourced.



Paperid:1359
Authors:Yizhi Li, Ge Zhang, Yinghao Ma, Ruibin Yuan, Zhu, Hangyu Guo, Yiming Liang, Jiaheng Liu, Noah Wang, Jian Yang, Siwei Wu, Xingwei Qu, Jinjie Shi, Xinyue Zhang, Zhenzhu Yang, Yidan WEN, Yanghai Wang, Shihao Li, ZHAO-XIANG ZHANG, Ruibo Liu, Emmanouil Benetos, Wenhao Huang, Chenghua Lin
Abstract:
Recent advancements in multimodal large language models (MLLMs) have focused on integrating multiple modalities, yet their ability to simultaneously process and reason across different inputs remains underexplored. We introduce OmniBench, a novel benchmark designed to evaluate models’ ability to recognize, interpret, and reason across visual, acoustic, and textual inputs simultaneously. We define language models capable of such tri-modal processing as omni-language models (OLMs). OmniBench features high-quality human annotations that require integrated understanding across all modalities. Our evaluation reveals that: i) open-source OLMs show significant limitations in instruction-following and reasoning in tri-modal contexts; and ii) most baseline models perform poorly (below 50% accuracy) even with textual alternatives to image/audio inputs. To address these limitations, we develop OmniInstruct, an 96K-sample instruction tuning dataset for training OLMs. We advocate for developing more robust tri-modal integration techniques and training strategies to enhance OLM performance. Codes and data could be found atour repository.



Paperid:1360
Authors:Barak Bringoltz, Elisha Halperin, Ran Feraru, Evgeny Blaichman, Amit Berman
Abstract:
Over the last decade, ransomware detection has become a central topic in cybersecurity research. Due to ransomware's direct interaction with storage devices, analyzing I/O streams has become an effective detection method and represents a vital area of focus for research. A major challenge in this field is the lack of publicly accessible data featuring individual command labeling. To address this problem, we introduce the Command LEvel Annotated Ransomware (CLEAR) dataset, a large-scale collection of storage devices' stream data. The dataset comprises 1,045 TiB of I/O traffic data, featuring malicious traffic from 137 ransomware variants. It offers two orders of magnitude more I/O traffic data and one order of magnitude more ransomware variants than any other publicly accessible dataset. Importantly, it is the only dataset that individually labels each I/O command as either ransomware or benign activity. This labeling enables the use of advanced sequential models, which we show to outperform existing state-of-the-art models by up to 82% in data loss prevention. Additionally, this allows us to create new tasks, such as data recovery, by selectively reverting only the commands recognized as ransomware while preserving benign activity. The CLEAR dataset also includes supplementary auxiliary features derived from the data, which we demonstrate to improve performance through feature ablation studies. Lastly, a critical aspect of any ransomware detection model is its robustness to new, unseen ransomware variants, as new strains constantly emerge. Therefore, we propose a benchmark based on our dataset to evaluate performance against unknown ransomware samples and illustrate its application across different models.



Paperid:1361
Authors:Xiaoyuan Wang, Yizhou Zhao, Botao Ye, Shan Xiaojun, Weijie Lyu, Lu Qi, Kelvin Chan, Yinxiao Li, Ming-Hsuan Yang
Abstract:
We propose HoliGS, a novel deformable Gaussian splatting framework that addresses embodied view synthesis from long monocular RGB videos. Unlike prior 4D Gaussian splatting and dynamic NeRF pipelines, which struggle with training overhead in minute-long captures, our method leverages invertible Gaussian Splatting deformation networks to reconstruct large-scale, dynamic environments accurately. Specifically, we decompose each scene into a static background plus time-varying objects, each represented by learned Gaussian primitives undergoing global rigid transformations, skeleton-driven articulation, and subtle non-rigid deformations via an invertible neural flow. This hierarchical warping strategy enables robust free-viewpoint novel-view rendering from various embodied camera trajectories by attaching Gaussians to a complete canonical foreground shape (\eg, egocentric or third-person follow), which may involve substantial viewpoint changes and interactions between multiple actors. Our experiments demonstrate that \ourmethod~ achieves superior reconstruction quality on challenging datasets while significantly reducing both training and rendering time compared to state-of-the-art monocular deformable NeRFs. These results highlight a practical and scalable solution for EVS in real-world scenarios. The source code will be released.



Paperid:1362
Authors:Junlin Mu, Hantao Huang, Jihang Zhang, Minghui Yu, Tao Wang, Yidong Li
Abstract:
Large Language Models (LLMs) capable of handling extended contexts are in high demand, yet their inference remains challenging due to substantial Key-Value (KV) cache size and high memory bandwidth requirements. Previous research has demonstrated that KV cache exhibits low-rank characteristics within the hidden dimension, suggesting the potential for effective compression. However, due to the widely adopted Rotary Position Embedding (RoPE) mechanism in modern LLMs, naive low‑-rank compression suffers severe accuracy degradation or creates a new speed bottleneck, as the low-rank cache must first be reconstructed in order to apply RoPE.In this paper, we introduce two key insights: first, the application of RoPE to the key vectors increases their variance, which in turn results in a higher rank; second, after the key vectors are transformed into the latent space, they largely maintain their representation across most layers. Based on these insights, we propose the Sparse Attention in Latent Space (SALS) framework. SALS projects the KV cache into a compact latent space via low-rank projection, and performs sparse token selection using RoPE-free query--key interactions in this space. By reconstructing only a small subset of important tokens, it avoids the overhead of full KV cache reconstruction. We comprehensively evaluate SALS on various tasks using two large-scale models: LLaMA2-7b-chat and Mistral-7b. Experimental results demonstrate that SALS achieves SOTA performance by maintaining competitive accuracy. Under different settings, SALS achieves 6.4-fold KV cache compression and 5.7-fold speed-up in the attention operator compared to FlashAttention2 on the 4K sequence. For the end-to-end throughput performance, we achieves 1.4-fold and 4.5-fold improvement compared to GPT-fast on 4k and 32K sequences, respectively. The source code will be publicly available upon publication.



Paperid:1363
Authors:Pranoy Das, Kexin Fu, Abolfazl Hashemi, Vijay Gupta
Abstract:
Abstract:Deep generative models, particularly diffusion models, have achieved remarkable success across diverse domains but face significant challenges when trained on real-world, long-tailed datasets-where a few "head" classes dominate and many "tail" classes are underrepresented. This paper develops a rigorous theoretical framework for long-tailed learning via diffusion models through the lens of deep mutual learning. We introduce a novel regularized training objective that combines the standard diffusion loss with a mutual learning term, enabling balanced performance across all class labels, including the underrepresented tails. Our approach to learn via the proposed regularized objective is to formulated it as a multi-player game, with Nash equilibrium serving as the solution concept. We derive a non-asymptotic first-order convergence result for individual gradient descent algorithm to find the Nash equilibrium. We show that the Nash gap of the score network obtained from the algorithm is upper bounded by $\mathcal{O}(\frac{1}{\sqrt{T_{train}}}+\beta)$ where $\beta$ is the regularizing parameter and $T_{train}$ is the number of iterations of the training algorithm. Furthermore, we theoretically establish hyper-parameters for training and sampling algorithm that ensure we find conditional score networks (under our model) with a worst case sampling error $\mathcal{O}(\epsilon+1), \forall \epsilon>0$ across all class labels, which is further supported by numerical experiments. Our results offer new insights and guarantees for training diffusion models on imbalanced, long-tailed data, with implications for fairness, privacy, and generalization in real-world generative modeling scenarios.



Paperid:1364
Authors:Xuankun Rong, Wenke Huang, Jian Liang, Jinhe Bi, Xun Xiao, Yiming Li, Bo Du, Mang Ye
Abstract:
Multimodal Large Language Models (MLLMs) are increasingly deployed in fine-tuning-as-a-service (FTaaS) settings, where user-submitted datasets adapt general-purpose models to downstream tasks. This flexibility, however, introduces serious security risks, as malicious fine-tuning can implant backdoors into MLLMs with minimal effort. In this paper, we observe that backdoor triggers systematically disrupt cross-modal processing by causing abnormal attention concentration on non-semantic regions—a phenomenon we termattention collapse. Based on this insight, we proposeBelieve Your Eyes (BYE), a data filtering framework that leverages attention entropy patterns as self-supervised signals to identify and filter backdoor samples. BYE operates via a three-stage pipeline: (1) extracting attention maps using the fine-tuned model, (2) computing entropy scores and profiling sensitive layers via bimodal separation, and (3) performing unsupervised clustering to remove suspicious samples. Unlike prior defenses, BYE equires no clean supervision, auxiliary labels, or model modifications. Extensive experiments across various datasets, models, and diverse trigger types validate BYE's effectiveness: it achieves near-zero attack success rates while maintaining clean-task performance, offering a robust and generalizable solution against backdoor threats in MLLMs.



Paperid:1365
Authors:Anna Sepliarskaia, Sophie Langer, Johannes Schmidt-Hieber
Abstract:
Recent works have introduced new equivariant neural networks, motivated by their improved generalization compared to traditional deep neural networks. While experiments support this advantage, the theoretical understanding of their generalization properties remains limited. In this paper, we analyze the generalization capabilities of Group Convolutional Neural Networks (GCNNs) with the ReLU activation function through the lens of Vapnik-Chervonenkis (VC) dimension theory. We investigate how architectural factors—such as the number of layers, weights, and input dimensions—affect the VC dimension.A key challenge in our analysis is proving a lower bound on the VC dimension, for which we introduce new techniques, establishing a novel connection between GCNNs and standard deep neural networks. Additionally, we compare our derived bounds to those known for fully connected neural networks. Our results extend previous findings on the VC dimension of continuous GCNNs with two layers, offering new insights into their generalization behavior, particularly their dependence on input resolution.



Paperid:1366
Authors:Yan Ru Pei, Olivier Coenen
Abstract:
We introduce a class of neural networks named PLEIADES (PoLynomial Expansion In Adaptive Distributed Event-based Systems), which contains temporal convolution kernels generated from orthogonal polynomial basis functions. We focus on interfacing these networks with event-based data to perform online spatiotemporal classification and detection with low latency. By virtue of using structured temporal kernels and event-based data, we have the freedom to vary the sample rate of the data along with the discretization step-size of the network without additional finetuning. We experimented with three event-based benchmarks and obtained state-of-the-art results on all three by large margins with significantly smaller memory and compute costs. We achieved: 1) 99.59% accuracy with 192K parameters on the DVS128 hand gesture recognition dataset and 100\% with a small additional output filter; 2) 99.58% test accuracy with 277K parameters on the AIS 2024 eye tracking challenge; and 3) 0.556 mAP with 576k parameters on the PROPHESEE 1 Megapixel Automotive Detection Dataset.



Paperid:1367
Authors:Christoph Dann, Yishay Mansour, Teodor Vanislavov Marinov, Mehryar Mohri
Abstract:
The rapid proliferation of domain-specialized machine learning modelspresents a challenge: while individual models excel in specificdomains, their performance varies significantly across diverseapplications. This makes selecting the optimal model when faced withan unknown mixture of tasks, especially with limited or no datato estimate the mixture, a difficult problem. We address thischallenge by formulating it as a multiple-source domain adaptation(MSA) problem. We introduce a novel, scalable algorithm thateffectively routes each input to the best-suited model from a pool ofavailable models. Our approach provides a strong performanceguarantee: remarkably, for any mixture domain, the accuracy achieved by the bestsource model is maintained. This guarantee is established through atheoretical bound on the regret for new domains, expressed as a convexcombination of the best regrets in the source domains, plus aconcentration term that diminishes as the amount of source dataincreases. While our primary contributions are theoretical andalgorithmic, we also present empirical results demonstrating theeffectiveness of our approach.



Paperid:1368
Authors:Viktoria Schram, Markus Hiller, Daniel Beck, Trevor Cohn
Abstract:
Abstract:The prediction of training and learning curves of Natural Language Processing (NLP) models enables targeted decision-making to achieve performance objectives, reduces computational overhead and minimizes costs associated with dataset acquisition and curation. In this work, we formulate both prediction tasks as multitask learning problems, where each task’s data is modelled as being organized within a two-layer hierarchy. We employ latent variable multi-output Gaussian processes to capture correlations across tasks and hierarchies in order to support zero-shot prediction. We demonstrate that this approach facilitates the development of probabilistic scaling laws at lower costs. Applying an active learning strategy, training curves can be queried to reduce predictive uncertainty and provide predictions close to ground truth scaling laws. We validate the generality of our modelling approach and the strength of the proposed framework for NLP tasks using three small-scale datasets consisting of at maximum $30$ curves: Training curves from nanoGPT models of varying sizes, learning curves from mBART and a Transformer model for bilingual translation, and learning curves for multilingual translation using M2M100 architectures of three different sizes.



Paperid:1369
Authors:Senkang Hu, Xudong Han, Jinqi Jiang, Yihang Tao, Zihan Fang, Yong Dai, Sam Kwong, Yuguang Fang
Abstract:
Adapting billion-parameter language models to a downstream task is still costly, even with parameter-efficient fine-tuning (PEFT). We re-cast task adaptation as output-distribution alignment: the objective is to steer the output distribution toward the task distribution directly during decoding rather than indirectly through weight updates. Building on this view, we introduce Steering Vector Decoding (SVD), a lightweight, PEFT-compatible, and theoretically grounded method. We start with a short warm-start fine-tune and extract a task-aware steering vector from the Kullback-Leibler (KL) divergence gradient between the output distribution of the warm-started and pre-trained models. This steering vector is then used to guide the decoding process to steer the model's output distribution towards the task distribution. We theoretically prove that SVD is first-order equivalent to the gradient step of full fine-tuning and derive a globally optimal solution for the strength of the steering vector. Across three tasks and nine benchmarks, SVD paired with four standard PEFT methods improves multiple-choice accuracy by up to 5 points and open-ended truthfulness by 2 points, with similar gains (1-2 points) on commonsense datasets without adding trainable parameters beyond the PEFT adapter. SVD thus offers a lightweight, theoretically grounded path to stronger task adaptation for large language models.



Paperid:1370
Authors:Lee Jung-Mok, Nam Hyeon-Woo, Moon Ye-Bin, Junhyun Nam, Tae-Hyun Oh
Abstract:
Automated model discovery is the process of automatically searching and identifying the most appropriate model for a given dataset over a large combinatorial search space. Existing approaches, however, often face challenges in balancing the capture of fine-grained details with ensuring generalizability beyond training data regimes with a reasonable model complexity. In this paper, we present a multi-modal \& multi-step pipeline for effective automated model discovery. Our approach leverages two vision-language-based modules (VLM), AnalyzerVLM and EvaluatorVLM, for effective model proposal and evaluation in an agentic way. AnalyzerVLM autonomously plans and executes multi-step analyses to propose effective candidate models. EvaluatorVLM assesses the candidate models both quantitatively and perceptually, regarding the fitness for local details and the generalibility for overall trends. Our results demonstrate that our pipeline effectively discovers models that capture fine details and ensure strong generalizability. Additionally, extensive ablation studies show that both multi-modality and multi-step reasoning play crucial roles in discovering favorable models.



Paperid:1371
Authors:Scott Pesme, Giacomo Meanti, Michael Arbel, Julien Mairal
Abstract:
Abstract:Denoiser models have become powerful tools for inverse problems, enabling the use of pretrained networks to approximate the score of a smoothed prior distribution. These models are often used in heuristic iterative schemes aimed at solving Maximum a Posteriori (MAP) optimisation problems, where the proximal operator of the negative log-prior plays a central role. In practice, this operator is intractable, and practitioners plug in a pretrained denoiser as a surrogate—despite the lack of general theoretical justification for this substitution. In this work, we show that a simple algorithm, closely related to several used in practice, provably converges to the proximal operator under a log-concavity assumption of the prior $p$. We show that this algorithm can be interpreted as a gradient descent on smoothed proximal objectives. Our analysis thus provides a theoretical foundation for a class of empirically successful but previously heuristic methods.



Paperid:1372
Authors:Guanghan Wang, Yair Schiff, Subham Sahoo, Volodymyr Kuleshov
Abstract:
Part of the success of diffusion models stems from their ability to perform iterative refinement, i.e., repeatedly correcting outputs during generation. However, modern masked discrete diffusion lacks this capability: when a token is generated, it cannot be updated again, even when it introduces an error. Here, we address this limitation by introducing the remasking diffusion model (ReMDM) sampler, a method that can be applied to pretrained masked diffusion models in a principled way and that is derived from a discrete diffusion model with a custom remasking backward process. Most interestingly, ReMDM endows discrete diffusion with a form of inference-time compute scaling. By increasing the number of sampling steps, ReMDM generates natural language outputs that approach the quality of autoregressive models, whereas when the computation budget is limited, ReMDM better maintains quality. ReMDM also improves sample quality of masked diffusion models for discretized images, and in scientific domains such as molecule design, ReMDM facilitates diffusion guidance and pushes the Pareto frontier of controllability relative to classical masking and uniform noise diffusion. When applied to large pretrained diffusion language models, ReMDM boosts the model’s performance on downstream tasks requiring factual knowledge grasp and reasoning ability.



Paperid:1373
Authors:Guanming Zhang, David Heeger, Stefano Martiniani
Abstract:
Contrastive self-supervised learning based on point-wise comparisons has been widely studied for vision tasks. In the neural cortex, neuronal responses to distinct stimulus classes are organized into geometric structures known as neural manifolds. Accurate classification of stimuli can be achieved by effectively separating these manifolds, akin to solving a packing problem. We introduce Contrastive Learning As Manifold Packing (CLAMP), a self-supervised framework that recasts representation learning as a manifold packing problem. CLAMP introduces a loss function inspired by the potential energy of short-range repulsive particle systems, such as those encountered in the physics of simple liquids and jammed packings. In this framework, each class consists of sub‑manifolds embedding multiple augmented views of a single image. The sizes and positions of the sub-manifolds are dynamically optimized by following the gradient of a packing loss. This approach yields interpretable dynamics in the embedding space that parallel jamming physics, and introduces geometrically meaningful hyperparameters within the loss function. Under the standard linear evaluation protocol, which freezes the backbone and trains only a linear classifier, CLAMP achieves competitive performance with state‑of‑the‑art self‑supervised models. Furthermore, our analysis reveals that neural manifolds corresponding to different categories emerge naturally and are effectively separated in the learned representation space, highlighting the potential of CLAMP to bridge insights from physics, neural science, and machine learning.



Paperid:1374
Authors:Rohith Kuditipudi, Jing Huang, Sally Zhu, Percy Liang, Christopher Potts, Diyi Yang
Abstract:
Suppose Alice trains an open-weight language model, and subsequently Bob uses a blackbox derivative of Alice's model to produce text. Can Alice prove that Bob is using her model, either by querying Bob's derivative model (query setting) or from the text alone (sample setting)? We formulate this question as an independence testing problem---in which the null hypothesis is that Bob's model is independent of Alice's randomized training run---and investigate it through the lens of \textit{palimpsestic memorization} in language models: models are more likely to memorize data seen later in training, so we can test whether Bob's model derives from Alice's using test statistics that capture correlation between the output of Bob's model and the ordering of examples in Alice's training run. So long as Alice has randomly shuffled her training data, any significant correlation amounts to exactly quantifiable statistical evidence against the null hypothesis, regardless of the composition of Alice's training data. We develop tests for both the query and sample settings and empirically validate the power of our tests using the Pythia and OLMo model families, as well as small-scale models trained on TinyStories. In the query setting, we query Bob's model on Alice's training data and measure the correlation of its log-likelihood with the ordering of data. We show that this test is robust to common post-training practices (e.g., supervised fine-tuning, preference optimization, model souping). In the sample setting, we match spans of Bob's text against Alice's training examples and correlate the likelihood of a match with the ordering of training examples. We show this test reliably attributes text to models given a few thousand tokens. Our work offers a novel framework for provenance verification of open-weight language models, enabling accountability and protection for models.



Paperid:1375
Authors:Ruoyan Li, Guancheng Wan, Zijie Huang, Zixiao Liu, Haixin Wang, Xiao Luo, Wei Wang, Yizhou Sun
Abstract:
Flow‐field reconstruction from sparse sensor measurements remains a central challenge in modern fluid dynamics, as the need for high‐fidelity data often conflicts with practical limits on sensor deployment. Existing deep learning–based methods have demonstrated promising results, but they typically depend on simplifying assumptions such as two‐dimensional domains, predefined governing equations, synthetic datasets derived from idealized flow physics, and unconstrained sensor placement. In this work, we address these limitations by studying flow reconstruction under realistic conditions and introducing a \emph{directional transport‐aware Graph Neural Network (GNN)} that explicitly encodes both flow directionality and information transport. We further show that conventional sensor placement strategies frequently yield suboptimal configurations. To overcome this, we propose a novel \emph{Two‐Step Constrained PPO} procedure for Proximal Policy Optimization (PPO), which jointly optimizes sensor layouts by incorporating flow variability and accounts for reconstruction model's performance disparity with respect to sensor placement. We conduct comprehensive experiments under realistic assumptions to benchmark the performance of our reconstruction model and sensor placement policy. Together, they achieve significant improvements over existing methods.



Paperid:1376
Authors:Weili Jiang, Jinrong Lv, Xiaomeng Li, Xun Gong, Chubin Ou
Abstract:
Foundation segmentation models like Segment Anything (SAM) exhibit strong generalization on natural images but struggle with localized failures in medical imaging, especially on fine-grained structures such as vessels with complex morphology and indistinct boundaries. To address this, we propose FineSAM++, a structure-aware sparse expert framework designed to refine SAM outputs by introducing a confidence-driven soft Routing Module. This module dynamically identifies structurally uncertain regions and activates a lightweight Residual Expert to model and correct residual structural errors only within these areas, thereby achieving efficient "refinement over retraining." Extensive experiments on five public vascular segmentation datasets demonstrate that FineSAM++ consistently outperforms both SAM-adapted baselines and task-specific models in terms of accuracy, topological consistency. Our results highlight the effectiveness of sparse, structure-driven Mixture-of-Experts (MoE) strategies for enhancing the reliability of foundation vision models in clinical image understanding tasks.



Paperid:1377
Authors:Hyunseop Kim, Juheon Jeong, Hanul Kim, Yeong Jun Koh
Abstract:
We present a novel end-to-end transformer-based framework for Multiple Object Tracking (MOT) that advances temporal modeling and identity preservation. Despite recent progress in transformer-based MOT, existing methods still struggle to maintain consistent object identities across frames, especially under occlusions, appearance changes, or detection failures. We propose a dual-path temporal decoder that explicitly separates appearance adaptation and identity preservation. The appearance-adaptive decoder dynamically updates query features using current frame information, while the identity-preserving decoder freezes query features and reuses historical sampling offsets to maintain long-term temporal consistency. To further enhance stability, we introduce a confidence-guided update suppression strategy that retains previously reliable features when predictions are unreliable. Extensive experiments on MOT benchmarks demonstrate that our approach achieves state-of-the-art performance across major tracking metrics, with significant gains in association accuracy and identity consistency. Our results demonstrate the importance of decoupling dynamic appearance modeling from static identity cues, and provide a scalable foundation for robust tracking in complex scenarios.



Paperid:1378
Authors:Biswadeep Chakraborty, Saibal Mukhopadhyay
Abstract:
Abstract:We propose Fast Long-range Adaptive Memory for Event (FLAME), a novel scalable architecture that combines neuro-inspired feature extraction with robust structured sequence modelingto efficiently process asynchronous and sparse event camera data. As a departure from conventional input encoding methods, FLAME presents Event Attention Layer, a novel feature extractor that leverages neuromorphic dynamics (Leaky Integrate-and-Fire (LIF)) to directly capture multi-timescale features from event streams. The feature extractor is integrates with a structured state-space model with a novel Event-Aware HiPPO (EA-HiPPO) mechanism that dynamically adapts memory retention based on inter-event intervals to understand relationship across varying temporal scales and event sequences. A Normal Plus Low Rank (NPLR) decomposition reduces the computational complexity of state update from $\mathcal{O}(N^2)$ to $\mathcal{O}(Nr)$, where $N$ represents the dimension of the core state vector and $r$ is the rank of a low-rank component (with $r \ll N$). FLAME demonstrates state-of-the-art accuracy for event-by-event processing on complex event camera datasets.



Paperid:1379
Authors:Guangyi He, Tobias Sutter, Lukas Gonon
Abstract:
In this paper, we study the robustness of classical deep hedging strategies under distributional shifts by leveraging the concept of adversarial attacks. We first demonstrate that standard deep hedging models are highly vulnerable to small perturbations in the input distribution, resulting in significant performance degradation. Motivated by this, we propose an adversarial training framework tailored to increase the robustness of deep hedging strategies. Our approach extends pointwise adversarial attacks to the distributional setting and introduces a computationally tractable reformulation of the adversarial optimization problem over a Wasserstein ball. This enables the efficient training of hedging strategies that are resilient to distributional perturbations. Through extensive numerical experiments, we show that adversarially trained deep hedging strategies consistently outperform their classical counterparts in terms of out-of-sample performance and resilience to model misspecification. Our findings establish a practical and effective framework for robust deep hedging under realistic market uncertainties.



Paperid:1380
Authors:Qihao Duan, Bingding Huang, Zhenqiao Song, Irina Lehmann, Lei Gu, Roland Eils, Benjamin Wild
Abstract:
Large language models (LLMs) have revolutionized natural language processing and are increasingly applied to other sequential data types, including genetic sequences. However, adapting LLMs to genetics presents significant challenges. Capturing complex genomic interactions requires modeling long-range global dependencies within DNA sequences, where interactions often span over 10,000 base pairs, even within a single gene. This poses substantial computational demands under conventional model architectures and training paradigms. Additionally, traditional LLM training approaches are suboptimal for DNA sequences: autoregressive training, while efficient for training, only supports unidirectional sequence understanding. However, DNA is inherently bidirectional. For instance, bidirectional promoters regulate gene expression in both directions and govern approximately 11% of human gene expression. Masked language models (MLMs) enable bidirectional understanding. However, they are inefficient since only masked tokens contribute to loss calculations at each training step. To address these limitations, we introduce JanusDNA, the first bidirectional DNA foundation model built upon a novel pretraining paradigm, integrating the optimization efficiency of autoregressive modeling with the bidirectional comprehension capability of masked modeling. JanusDNA's architecture leverages a Mamba-Attention Mixture-of-Experts (MoE) design, combining the global, high-resolution context awareness of attention mechanisms with the efficient sequential representation learning capabilities of Mamba. The MoE layers further enhance the model's capacity through sparse parameter scaling, while maintaining manageable computational costs. Notably, JanusDNA can process up to 1 million base pairs at single-nucleotide resolution on a single 80GB GPU using its hybrid architecture. Extensive experiments and ablation studies demonstrate that JanusDNA achieves new state-of-the-art performance on three genomic representation benchmarks. Remarkably, JanusDNA surpasses models with 250x more activated parameters, underscoring its efficiency and effectiveness. Code available at https://anonymous.4open.science/r/JanusDNA/.



Paperid:1381
Authors:George Ma, Samuel Pfrommer, Somayeh Sojoudi
Abstract:
Mechanistic interpretability research seeks to reverse-engineer large language models (LLMs) by uncovering the internal representations of concepts within their activations. Sparse Autoencoders (SAEs) have emerged as a valuable tool for disentangling polysemantic neurons into more monosemantic, interpretable features. However, recent work on automatic explanation generation for these features has faced challenges: explanations tend to be overly broad and fail to take polysemanticity into consideration. This work addresses these limitations by introducing a similarity-based strategy for sourcing close negative sentences that more effectively falsify generated explanations. Additionally, we propose a structured, component-based format for feature explanations and a tree-based, iterative explanation method that refines explanations. We demonstrate that our structured format and tree-based explainer improve explanation quality, while our similarity-based evaluation strategy exposes biases in existing interpretability methods. We also analyze the evolution of feature complexity and polysemanticity across LLM layers, offering new insights into information content within LLMs' residual streams.



Paperid:1382
Authors:Junqi Gao, Zhichang Guo, Dazhi Zhang, Dong Li, Runze Liu, Pengfei Li, Kai Tian, Biqing Qi
Abstract:
Heterogeneous Large Language Model (LLM) fusion integrates the strengths of multiple source LLMs with different architectures into a target LLM with low computational overhead. While promising, existing methods suffer from two major limitations: 1)reliance on real data from limited domainfor knowledge fusion, preventing the target LLM from fully acquiring knowledge across diverse domains, and 2)fixed data allocation proportionsacross domains, failing to dynamically adjust according to the target LLM's varying capabilities across domains, leading to a capability imbalance. To overcome these limitations, we propose Bohdi, a synthetic-data-only heterogeneous LLM fusion framework. Through the organization of knowledge domains into a hierarchical tree structure, Bohdi enables automatic domain exploration and multi-domain data generation through multi-model collaboration, thereby comprehensively extracting knowledge from source LLMs. By formalizing domain expansion and data sampling proportion allocation on the knowledge tree as a Hierarchical Multi-Armed Bandit problem, Bohdi leverages the designed DynaBranches mechanism to adaptively adjust sampling proportions based on the target LLM's performance feedback across domains. Integrated with our proposed Introspection-Rebirth (IR) mechanism, DynaBranches dynamically tracks capability shifts during target LLM's updates via Sliding Window Binomial Likelihood Ratio Testing (SWBLRT), further enhancing its online adaptation capability. Comparative experimental results on a comprehensive suite of benchmarks demonstrate that Bohdi significantly outperforms existing baselines on multiple target LLMs, exhibits higher data efficiency, and virtually eliminates the imbalance in the target LLM's capabilities.



Paperid:1383
Authors:Daniel Barzilai, Ohad Shamir
Abstract:
The widespread use of generative models has created a feedback loop in which each generation of models is trained on data partially produced by its predecessors. This process has raised concerns about model collapse: A critical degradation in performance caused by repeated training on synthetic data. However, different analyses in the literature have reached different conclusions as to the severity of model collapse. As such, it remains unclear how concerning this phenomenon is, and under which assumptions it can be avoided. To address this, we theoretically study model collapse for maximum likelihood estimation (MLE), in a natural setting where synthetic data is gradually added to the original training set. Under standard assumptions (similar to those long used for proving asymptotic consistency and normality of MLE), we establish non-asymptotic bounds showing that collapse can be avoided even as the fraction of real data vanishes. On the other hand, we prove that some assumptions (beyond MLE consistency) are indeed necessary: Without them, model collapse can occur arbitrarily quickly, even when the original data is still present in the training set. To the best of our knowledge, these are the first rigorous examples of iterative generative modeling with accumulating data that rapidly leads to model collapse.



Paperid:1384
Authors:Kang Lin, Anselm Krainovic, Kun Wang, Reinhard Heckel
Abstract:
Deep neural networks achieve state-of-the-art results for accelerated MRI reconstruction. Most research on deep learning based imaging focuses on improving neural network architectures trained and evaluated on fixed and homogeneous training and evaluation data. In this work, we investigate data curation strategies for improving MRI reconstruction. We assemble a large dataset of raw k-space data from 18 public sources and construct a diverse evaluation set comprising 48 test sets, capturing variations in anatomy, contrast, number of coils, and other key factors. We propose and study different data filtering strategies to enhance performance of current state-of-the-art neural networks for accelerated MRI reconstruction. Our experiments show that filtering the training data leads to consistent, albeit modest, performance gains. These performance gains are robust across different training set sizes and accelerations, and we find that filtering is particularly beneficial when the proportion of in-distribution data in the unfiltered training set is low.



Paperid:1385
Authors:Jayneel Parekh, Pegah KHAYATAN, Mustafa Shukor, Arnaud Dapogny, Alasdair Newson, Matthieu Cord
Abstract:
Steering has emerged as a practical approach to enable post-hoc guidance of LLMs towards enforcing a specific behavior. However, it remains largely underexplored for multimodal LLMs (MLLMs); furthermore, existing steering techniques, such as \textit{mean} steering, rely on a single steering vector, applied independently of the input query. This paradigm faces limitations when the desired behavior is dependent on the example at hand. For example, a safe answer may consist in abstaining from answering when asked for an illegal activity, or may point to external resources or consultation with an expert when asked about medical advice. In this paper, we investigate a fine-grained steering that uses an input-specific linear shift. This shift is computed using contrastive input-specific prompting. However, the input-specific prompts required for this approach are not known at test time. Therefore, we propose to train a small auxiliary module to predict the input-specific steering vector. Our approach, dubbed as L2S (Learn-to-Steer), demonstrates that it reduces hallucinations and enforces safety in MLLMs, outperforming other static baselines. We will open-source our code.



Paperid:1386
Authors:Zigeng Chen, Xinyin Ma, Gongfan Fang, Ruonan Yu, Xinchao Wang
Abstract:
Large Reasoning Models (LRMs) have garnered considerable attention for their ability to tackle complex tasks through the Chain-of-Thought (CoT) approach. However, their tendency toward overthinking results in unnecessarily lengthy reasoning chains, dramatically increasing the inference costs. To mitigate this issue, we introduce VeriThinker, a novel approach for CoT compression. Unlike conventional methods that fine-tune LRMs directly on the original reasoning task using synthetic concise CoT data, we innovatively fine-tune the model solely through an auxiliary verification task. By training LRMs to accurately verify the correctness of CoT solutions, the LRMs inherently become more discerning about the necessity of subsequent self-reflection steps, thereby effectively suppressing overthinking. Extensive experiments validate that VeriThinker substantially reduces reasoning chain lengths while maintaining or even slightly improving accuracy. When applied to DeepSeek-R1-Distill-Qwen-7B, our approach reduces reasoning tokens on MATH500 from 3790 to 2125 while improving accuracy by 0.8% (94.0% to 94.8%), and on AIME25, tokens decrease from 14321 to 10287 with a 2.1% accuracy gain (38.7% to 40.8%). Additionally, our experiments demonstrate that VeriThinker can also be zero-shot generalized to speculative reasoning.



Paperid:1387
Authors:Hoang Tran, Van Hoan Trinh, VINH BUI, Tan Nguyen
Abstract:
Linear Mode Connectivity (LMC) is a notable phenomenon in the loss landscapesof neural networks, wherein independently trained models have been observed tobe connected—up to permutation symmetries—by linear paths in parameter spacealong which the loss remains consistently low. This observation challenges classicalviews of non-convex optimization and has implications for model ensembling,generalization, and our understanding of neural loss geometry. Inspired by recentstudies on LMC in standard neural networks, we systematically investigate thisphenomenon within Mixture-of-Experts (MoE) architectures—a class of modelsknown for their scalability and computational efficiency, which combine traditionalneural networks—referred to as experts—through a learnable gating mechanism.We begin by conducting a comprehensive analysis of both dense and sparse gatingregimes, demonstrating that the symmetries inherent to MoE architectures arefully characterized by permutations acting on both the expert components and thegating function. Building on these foundational findings, we propose a matchingalgorithm that enables alignment between independently trained MoEs, therebyfacilitating the discovery of LMC. Finally, we empirically validate the presence ofLMC using our proposed algorithm across diverse MoE configurations—includingdense, sparse, and shared-expert variants—under a wide range of model settingsand datasets of varying scales and modalities. Our results confirm the existenceof LMC in MoE architectures and offer fundamental insights into the functionallandscape and optimization dynamics of deep learning models.



Paperid:1388
Authors:Jan Philipp Schneider, Pratik S. Bisht, Ilya Chugunov, Andreas Kolb, Michael Moeller, Felix Heide
Abstract:
Learning editable high-resolution scene representations for dynamic scenes is an open problem with applications across the domains from autonomous driving to creative editing - the most successful approaches today make a trade-off between editability and supporting scene complexity: neural atlases represent dynamic scenes as two deforming image layers, foreground and background, which are editable in 2D, but break down when multiple objects occlude and interact. In contrast, scene graph models make use of annotated data such as masks and bounding boxes from autonomous‑driving datasets to capture complex 3D spatial relationships, but their implicit volumetric node representations are challenging to edit view-consistently. We propose Neural Atlas Graphs (NAGs), a hybrid high-resolution scene representation, where every graph node is a view‑dependent neural atlas, facilitating both 2D appearance editing and 3D ordering and positioning of scene elements. Fit at test‑time, NAGs achieve state‑of‑the‑art quantitative results on the Waymo Open Dataset - by 5 dB PSNR increase compared to existing methods - and make environmental editing possible in high resolution and visual quality - creating counterfactual driving scenarios with new backgrounds and edited vehicle appearance. We find that the method also generalizes beyond driving scenes and compares favorably - by more than 7 dB in PSNR - to recent matting and video editing baselines on the DAVIS video dataset with a diverse set of human and animal-centric scenes.



Paperid:1389
Authors:Arian Raje, Baris Askin, Divyansh Jhunjhunwala, Gauri Joshi
Abstract:
Abstract:Large Language Models (LLMs) have yet to effectively leverage the vast amounts of edge-device data, and Federated Learning (FL) offers a promising paradigm to collaboratively fine-tune LLMs without transferring private edge data to the cloud. To operate within the computational and communication constraints of edge devices, recent literature on federated fine-tuning of LLMs proposes the use of low-rank adaptation (LoRA) and similar parameter-efficient methods. However, LoRA-based methods suffer from accuracy degradation in FL settings, primarily because of data and computational heterogeneity across clients. We propose Ravan, an adaptive multi-head LoRA method that balances parameter efficiency and model expressivity by reparameterizing the weight updates as the sum of multiple LoRA heads, $s_i\textbf{B}_i\textbf{H}_i\textbf{A}_i$, in which only the $\textbf{H}_i$ parameters and their lightweight scaling factors $s_i$ are trained. These trainable scaling factors let the optimization focus on the most useful heads, recovering a higher-rank approximation of the full update without increasing the number of communicated parameters since clients upload $s_i\textbf{H}_i$ directly. Experiments on vision and language benchmarks show that Ravan improves test accuracy by 2–8\% over prior parameter-efficient baselines, making it a robust and scalable solution for federated fine-tuning of LLMs.



Paperid:1390
Authors:Qizheng Zhang, Michael Wornow, Kunle Olukotun
Abstract:
LLM-based agentic applications have shown increasingly remarkable capabilities in complex workflows but incur substantial costs due to extensive planning and reasoning requirements. Existing LLM caching techniques (like context caching and semantic caching), primarily designed for serving chatbots, are insufficient for agentic applications where outputs depend on external data and environmental contexts. We propose agentic plan caching, a novel approach that extracts, stores, adapts, and reuses structured plan templates from planning stages of agentic applications across semantically similar tasks to reduce the cost of serving. Unlike traditional semantic caching, our system extracts plan templates from completed agent executions at test-time, employs keyword extraction to match new requests against cached plans, and utilizes lightweight models to adapt these templates to task-specific plans with contexts. Evaluation across multiple real-world agentic applications shows that our system can reduce costs by 46.62\% on average while maintaining performance, offering a more efficient solution for serving LLM-based agents that complements existing LLM serving infrastructures.



Paperid:1391
Authors:Cheng-Yao Hong, Li-Heng Wang, Tyng-Luh Liu
Abstract:
Accurate identification and localization of objects in 3-D scenes are essential for advancing comprehensive 3-D scene understanding. Although diffusion models have demonstrated impressive capabilities across a broad spectrum of computer vision tasks, their potential in both 2-D and 3-D object detection remains underexplored. Existing approaches typically formulate detection as a ''noise-to-box'' process, but they rely heavily on direct coordinate regression, which limits adaptability for more advanced tasks such as grounding-based object detection. To overcome these challenges, we propose a promptable 3-D object recognition framework, which introduces a diffusion-based paradigm for flexible and conditionally guided 3-D object detection. Our approach encodes bounding boxes into latent representations and employs latent diffusion models to realize a ''promptable noise-to-box'' transformation. This formulation enables the refinement of standard 3-D object detection using textual prompts, such as class labels. Moreover, it naturally extends to grounding object detection through conditioning on natural language descriptions, and generalizes effectively to few-shot learning by incorporating annotated exemplars as visual prompts. We conduct thorough evaluations on three key 3-D object recognition tasks: general 3-D object detection, few-shot detection, and grounding-based detection. Experimental results demonstrate that our framework achieves competitive performance relative to state-of-the-art methods, validating its effectiveness, versatility, and broad applicability in 3-D computer vision.



Paperid:1392
Authors:Mohan Zhang, Jiaxuan Gao, Shusheng Xu, YI WU
Abstract:
We study the use of Process Reward Models (PRMs) for guiding Long Chain-of-Thought (CoT) reasoning in large language models. Although PRMs deliver fine-grained feedback in standard tasks, PRM-guided beam search does not consistently outperform PRM-free approaches in long CoT reasoning. We trace this shortfall to a "step quality degradation''—the expected step quality shows concave behavior, yielding unimodal or monotonically declining trends. To counteract this, we propose Z-Score Guided Early Stopping (ZGES), which halts search at the detected quality peak using local PRM-reward z-scores. Across multiple math benchmarks and model scales, ZGES outperforms both standard PRM-guided beam search and the PRM-free methods. Ablation studies further highlight the advantages and robustness of ZGES’s adaptive stopping mechanism.



Paperid:1393
Authors:Brian Nlong Zhao, Jiajun Wu, Shangzhe Wu
Abstract:
Computer vision for animals holds great promise for wildlife research but often depends on large-scale data, while existing collection methods rely on controlled capture setups. Recent data-driven approaches show the potential of single-view, non-invasive analysis, yet current animal video datasets are limited—offering as few as 2.4K 15-frame clips and lacking key processing for animal-centric 3D/4D tasks. We introduce an automated pipeline that mines YouTube videos and processes them into object-centric clips, along with auxiliary annotations valuable for downstream tasks like pose estimation, tracking, and 3D/4D reconstruction. Using this pipeline, we amass 30K videos (2M frames)—an order of magnitude more than prior works. To demonstrate its utility, we focus on 4D quadruped animal reconstruction task. To support this task, we present Animal4D, a benchmark of 230 manually filtered sequences with 11K frames showcasing clean, diverse animal motions. We evaluate state-of-the-art model-based and model-free methods on Animal4D, finding that 2D metrics favor the former despite unrealistic 3D shapes, while the latter yields more natural reconstructions but scores lower—revealing a gap in current evaluation. To address this, we enhance a recent model-free approach with sequence-level optimization, establishing the first 4D animal reconstruction baseline. Together, our pipeline, benchmark, and baseline aim to advance large-scale, markerless 4D animal reconstruction and related tasks from in-the-wild videos.



Paperid:1394
Authors:Jinghong Chen, Guangyu Yang, Weizhe Lin, Jingbiao Mei, Chenxu Lyu, Bill Byrne
Abstract:
We derive and investigate two DPO variants that explicitly model the possibility of declaring a tie in pair-wise comparisons. We replace the Bradley-Terry model in DPO with two well-known modeling extensions, by Rao and Kupper and by Davidson, that assign probability to ties as alternatives to clear preferences. Our experiments in neural machine translation and summarization show that explicitly labeled ties can be added to the datasets for these DPO variants without the degradation in task performance that is observed when the same tied pairs are presented to DPO. We find empirically that the inclusion of ties leads to stronger regularization with respect to the reference policy as measured by KL divergence, and we see this even for DPO in its original form. We provide a theoretical explanation for this regularization effect using ideal DPO policy theory. We further show performance improvements over DPO in translation and mathematical reasoning using our DPO variants. We find it can be beneficial to include ties in preference optimization rather than simply discard them, as is done in common practice.



Paperid:1395
Authors:Steve Hanneke
Abstract:
Abstract:We sharply characterize the optimal first-order query complexity of agnostic active learning for all concept classes, and propose a new general active learning algorithm which achieves it. Remarkably, the optimal query complexity admits a leading term which is always strictly smaller than the sample complexity of passive supervised learning (by a factor proportional to the best-in-class error rate). This was not previously known to be possible in the agnostic setting. For comparison, in all previous general analyses, the leading term exhibits an additional factor, such as the disagreement coefficient or related complexity measure, and therefore only provides improvements over passive learning in restricted cases. The present work completely removes such factors from the leading term, implying that $\textit{every}$ concept class benefits from active learning in the non-realizable case. The results established in this work resolve an important long-standing open question central to the past two decades of research on the theory of agnostic active learning.



Paperid:1396
Authors:KuoCheng Wu, Guohang Zhuang, Jinyang Huang, Xiang Zhang, Wanli Ouyang, Yan Lu
Abstract:
Super-resolution (SR) advances astronomical imaging by enabling cost-effective high-resolution capture, crucial for detecting faraway celestial objects and precise structural analysis. However, existing datasets for astronomical SR (ASR) exhibit three critical limitations: flux inconsistency, object-crop setting, and insufficient data diversity, significantly impeding ASR development. We propose STAR, a large-scale astronomical SR dataset containing 54,738 flux-consistent star field image pairs covering wide celestial regions. These pairs combine Hubble Space Telescope high-resolution observations with physically faithful low-resolution counterparts generated through a flux-preserving data generation pipeline, enabling systematic development of field-level ASR models. To further empower the ASR community, STAR provides a novel Flux Error (FE) to evaluate SR models in physical view. Leveraging this benchmark, we propose a Flux-Invariant Super Resolution (FISR) model that could accurately infer the flux-consistent high-resolution images from input photometry, suppressing several SR state-of-the-art methods by 24.84% on a novel designed flux consistency metric, showing the priority of our method for astrophysics. Extensive experiments demonstrate the effectiveness of our proposed method and the value of our dataset. Code and models are available at https://github.com/GuoCheng12/STAR.



Paperid:1397
Authors:Connor Magoon, Fengyu Yang, Noam Aigerman, Shahar Kovalsky
Abstract:
Differentiable optimization has attracted significant research interest, particularly for quadratic programming (QP). Existing approaches for differentiating the solution of a QP with respect to its defining parameters often rely on specific integrated solvers. This integration limits their applicability, including their use in neural network architectures and bi-level optimization tasks, restricting users to a narrow selection of solver choices. To address this limitation, we introducedQP, a modular and solver-agnostic framework for plug-and-play differentiation of virtually any QP solver. Our key theoretical insight is that the solution and its derivative can each be expressed in terms of closely-related and simple linear systems by using the active set at the solution. This insight enables efficient decoupling of the QP's solution, obtained by any solver, from its differentiation. Our open-source, minimal-overhead implementation is publicly available and seamlessly integrates with more than 15 state-of-the-art solvers. Comprehensive benchmark experiments demonstrate dQP’s robustness and scalability, particularly highlighting its advantages in large-scale sparse problems.



Paperid:1398
Authors:Jiaming Zhuo, Ziyi Ma, Yintong Lu, Yuwei Liu, Kun Fu, Di Jin, Chuan Wang, Wu Wenning, Zhen Wang, Xiaochun Cao, Liang Yang
Abstract:
Graph Transformers (GTs), effective at capturing long-range dependencies and structural biases simultaneously, are gradually recognized as viable alternatives to traditional Graph Neural Networks (GNNs). Advanced methods for GTs to leverage topology information involve integrating the GNN modules and modulating node attributes using positional encodings. However, the potential mechanism for their effectiveness remains not fully explored and understood. Upon revisiting these two methods, a common mechanism—Cross Aggregation—that effectively captures the interrelation between graph topology and node attributes is revealed. Building on this insight, this paper proposes a universal GT architecture with linear complexity, named Universal Graph Cross-attention Transformer (UGCFormer). The idea is to interactively learn the representations of graph topology and node attributes via a linearized Dual Cross-attention (DCA) module. In theory, this module can adaptively capture interrelations between these two types of graph information, thus achieving effective aggregation. To mitigate the overfitting problem caused by the two-channel design, a consistency constraint is utilized to enhance their commonality, which prevents representation distortion. Extensive evaluations on benchmark datasets demonstrate the effectiveness and efficiency of UGCFormer.



Paperid:1399
Authors:Haruki Settai, Naoya Takeishi, Takehisa Yairi
Abstract:
For continuous systems modeled by dynamical equations such as ODEs and SDEs, Bellman's principle of optimality takes the form of the Hamilton-Jacobi-Bellman (HJB) equation, which provides the theoretical target of reinforcement learning (RL). Although recent advances in RL successfully leverage this formulation, the existing methods typically assume the underlying dynamics are known a priori because they need explicit access to the coefficient functions of dynamical equations to update the value function following the HJB equation. We address this inherent limitation of HJB-based RL; we propose a model-free approach still targeting the HJB equation and propose corresponding temporal difference methods. We demonstrate its potential advantages over transition kernel-based formulations, both qualitatively and empirically. The proposed formulation paves the way toward bridging stochastic optimal control and model-free reinforcement learning.



Paperid:1400
Authors:Siyu Huang, Abdul Basit Adeel, Yubai Yuan
Abstract:
A cascade over a network refers to the diffusion process where behavior changes occurring in one part of an interconnected population lead to a series of sequential changes throughout the entire population. In recent years, there has been a surge in interest and efforts to understand and model cascade mechanisms, as they motivate many significant research topics across different disciplines. The propagation structure of cascades is governed by the underlying diffusion networks, which are often hidden. Inferring diffusion networks enables interventions in the cascading process to maximize information propagation and provides insights into the Granger causality of interaction mechanisms among individuals. In this project, we propose a novel double mixture graphical model for inferring latent diffusion networks in the presence of strong cascade heterogeneity. The new model represents cascade pathways as a distributional mixture over diffusion networks, where these networks capture different cascading patterns at the population level. We develop a data-driven optimization method to infer diffusion networks using only visible temporal cascade records, without requiring the modeling of complex and heterogeneous individual states. Both statistical and computational guarantees for the proposed method are established. We apply the double mixture cascade model to analyze the research topic cascades in social science across U.S. universities and uncover the latent research topic diffusion networks among the top U.S. social science programs.



Paperid:1401
Authors:Yilin LI, Yuan Deng, Wei Tang, Hanrui Zhang
Abstract:
Abstract:Automated bidding to optimize online advertising with various constraints, e.g. ROI constraints and budget constraints, is widely adopted by advertisers. A key challenge lies in designing algorithms for non-truthful mechanisms with ROI constraints. While prior work has addressed truthful auctions or non-truthful auctions with weaker benchmarks, this paper provides a significant improvement: We develop online bidding algorithms for repeated first-price auctions with ROI constraints, benchmarking against the optimal randomized strategy in hindsight. In the full feedback setting, where the maximum competing bid is observed, our algorithm achieves a near-optimal $\tilde O(\sqrt{T})$ regret bound, and in the bandit feedback setting (where the bidder only observes whether the bidder wins each auction), our algorithm attains $\tilde O(T^{3/4})$ regret bound.



Paperid:1402
Authors:Chenhui Zhu, Yilu Wu, Shuai Wang, Gangshan Wu, Limin Wang
Abstract:
Image-to-video generation has made remarkable progress with the advancements in diffusion models, yet generating videos with realistic motion remains highly challenging. This difficulty arises from the complexity of accurately modeling motion, which involves capturing physical constraints, object interactions, and domain-specific dynamics that are not easily generalized across diverse scenarios. To address this, we propose MotionRAG, a retrieval-augmented framework that enhances motion realism by adapting motion priors from relevant reference videos through Context-Aware Motion Adaptation (CAMA). The key technical innovations include: (i) a retrieval-based pipeline extracting high-level motion features using video encoder and specialized resamplers to distill semantic motion representations; (ii) an in-context learning approach for motion adaptation implemented through a causal transformer architecture; (iii) an attention-based motion injection adapter that seamlessly integrates transferred motion features into pretrained video diffusion models. Extensive experiments demonstrate that our method achieves significant improvements across multiple domains and various models. Furthermore, our modular design enables zero-shot generalization to new domains by simply updating the retrieval database without retraining any components. This research enhances the core capability of video generation systems by enabling the effective retrieval and transfer of motion priors, facilitating the synthesis of realistic motion dynamics.



Paperid:1403
Authors:Jindi Lv, Yuhao Zhou, Mingjia Shi, Zhiyuan Liang, Panpan Zhang, Xiaojiang Peng, Wangbo Zhao, Zheng Zhu, Jiancheng Lv, Qing Ye, Kai Wang
Abstract:
Mamba has proven efficient for long-sequence modeling in vision tasks. However, when token reduction techniques are applied to improve efficiency, Mamba-based models exhibit drastic performance degradation compared to Vision Transformers (ViTs). This decline is potentially attributed to Mamba's chain-like scanning mechanism, which we hypothesize not only induces cascading losses in token connectivity but also limits the diversity of spatial receptive fields.In this paper, we propose Asymmetric Multi-scale Vision Mamba (AMVim), a novel architecture designed to enhance pruning robustness. AMVim employs a dual-path structure, integrating a window-aware scanning mechanism into one path while retaining sequential scanning in the other. This asymmetry design promotes token connection diversity and enables multi-scale information flow, reinforcing spatial awareness.Empirical results demonstrate that AMVim achieves state-of-the-art pruning robustness. During token reduction, AMVim-T achieves a substantial 34\% improvement in training-free accuracy with identical model sizes and FLOPs. Meanwhile, AMVim-S exhibits only a 1.5\% accuracy drop, performing comparably to ViT. Notably, AMVim also delivers superior performance during pruning-free settings, further validating its architectural advantages.



Paperid:1404
Authors:Sofiane ENNADIR, Oleg Smirnov, Yassine ABBAHADDOU, Lele Cao, Johannes Lutzeyer
Abstract:
Graph Neural Networks (GNNs) have achieved strong performance across a range of graph representation learning tasks, yet their adversarial robustness in graph classification remains underexplored compared to node classification. While most existing defenses focus on the message-passing component, this work investigates the overlooked role of pooling operations in shaping robustness. We present a theoretical analysis of standard flat pooling methods (sum, average and max), deriving upper bounds on their adversarial risk and identifying their vulnerabilities under different attack scenarios and graph structures. Motivated by these insights, we propose Robust Singular Pooling (RS-Pool), a novel pooling strategy that leverages the dominant singular vector of the node embedding matrix to construct a robust graph-level representation. We theoretically investigate the robustness of RS-Pool and interpret the resulting bound leading to improved understanding of our proposed pooling operator. While our analysis centers on Graph Convolutional Networks (GCNs), RS-Pool is model-agnostic and can be implemented efficiently via power iteration. Empirical results on real-world benchmarks show that RS-Pool provides better robustness than the considered pooling methods when subject to state-of-the-art adversarial attacks while maintaining competitive clean accuracy.



Paperid:1405
Authors:Patrick Cheridito, Jean-Loup Dupret, Donatien Hainaut
Abstract:
In this paper, we introduce a model-based deep-learning approach to solve finite-horizon continuous-time stochastic control problems with jumps. We iteratively train two neural networks: one to represent the optimal policy and the other to approximate the value function. Leveraging a continuous-time version of the dynamic programming principle, we derive two different training objectives based on the Hamilton--Jacobi--Bellman equation, ensuring that the networks capture the underlying stochastic dynamics. Empirical evaluations on different problems illustrate the accuracy and scalability of our approach, demonstrating its effectiveness in solving complex, high-dimensional stochastic control tasks.



Paperid:1406
Authors:Tobias Schnabel, Kiran Tomlinson, Adith Swaminathan, Jennifer Neville
Abstract:
Despite their many successes, transformer-based large language models (LLMs) continue to struggle with tasks that require complex reasoning over large parts of their input. We argue that these failures arise due to capacity limits on the accurate flow of information within LLMs. To formalize this issue, we introduce the bounded attention prefix oracle (BAPO) model, a new computational framework that models bandwidth constraints on attention heads, the mechanism for internal communication in LLMs. We show that several important reasoning problems like graph reachability require high communication bandwidth for BAPOs to solve; we call these problems BAPO-hard. Our experiments corroborate our theoretical predictions: GPT-4o, Claude, and Gemini succeed on BAPO-easy tasks and fail even on relatively small BAPO-hard tasks. BAPOs also reveal another benefit of chain of thought (CoT): we prove that breaking down a task using CoT can turn any BAPO-hard problem into a BAPO-easy one. Our results offer principled explanations for key LLM failures and suggest directions for architectures and inference methods that mitigate bandwidth limits.



Paperid:1407
Authors:Sahar Abdelnabi, Ahmed Salem
Abstract:
Reasoning‐focused large language models (LLMs) sometimes alter their behavior when they detect that they are being evaluated—an effect analogous to the Hawthorne phenomenon—which can lead them to optimize for test‐passing performance or to comply more readily with harmful prompts if real‐world consequences appear absent. We present the first quantitative study of how such "test awareness'' impacts model behavior, particularly its safety alignment. We introduce a white‐box probing framework that (i) linearly identifies awareness‐related activations and (ii) steers models toward or away from test awareness while monitoring downstream performance. We apply our method to different state-of-the-art open-source reasoning LLMs across both realistic and hypothetical tasks. Our results demonstrate that test awareness significantly impact safety alignment, and is different for different models. By providing fine-grained control over this latent effect, our work aims to increase trust in how we perform safety evaluation.



Paperid:1408
Authors:Zhongwei Yu, Wannian Xia, Xue Yan, Bo Xu, Haifeng Zhang, Yali Du, Jun Wang
Abstract:
Advanced large language models (LLMs) frequently reflect in reasoning chain-of-thoughts (CoTs), where they self-verify the correctness of current solutions and explore alternatives. However, given recent findings that LLMs detect limited errors in CoTs, how reflection contributes to empirical improvements remains unclear. To analyze this issue, in this paper, we present a minimalistic reasoning framework to support basic self-verifying reflection for small transformers without natural language, which ensures analytic clarity and reduces the cost of comprehensive experiments. Theoretically, we prove that self-verifying reflection guarantees improvements if verification errors are properly bounded. Experimentally, we show that tiny transformers, with only a few million parameters, benefit from self-verification in both training and reflective execution, reaching remarkable LLM-level performance in integer multiplication and Sudoku. Similar to LLM results, we find that reinforcement learning (RL) improves in-distribution performance and incentivizes frequent reflection for tiny transformers, yet RL mainly optimizes shallow statistical patterns without faithfully reducing verification errors. In conclusion, integrating generative transformers with discriminative verification inherently facilitates CoT reasoning, regardless of scaling and natural language.



Paperid:1409
Authors:Artem Lukoianov, Chenyang Yuan, Justin Solomon, Vincent Sitzmann
Abstract:
Among generative models, diffusion models are uniquely intriguing due to the existence of a closed-form Bayes-optimal solution to their training objective, often referred to as the optimal denoiser.However, this optimal denoiser merely reproduces images in the training set, and hence fails to serve as a complete theory for the behavior of deep diffusion models. Recent work has attempted to characterize this gap between the optimal denoiser and deep diffusion models and has produced analytical, training-free models that can generate images that resemble those generated by a trained UNet. The best-performing method suggests that shift equivariance and locality inductive biases of convolutional neural networks are the cause of this gap, and hence, incorporates these assumptions into its analytical model. In this work, we present evidence that the locality in deep diffusion models emerges as a statistical property of the image dataset, \emph{not} due to the inductive bias of convolutional neural networks. Specifically, we demonstrate that an optimal parametric linear denoiser exhibits similar locality properties and provides a theoretical analysis that grounds this in a principal component analysis of the training set.We then show that an analytical denoiser based on these statistics better matches scores predicted by a deep diffusion model than the prior, expert-crafted alternative.



Paperid:1410
Authors:Chenyu Wang, Cai Zhou, Sharut Gupta, Zongyu Lin, Stefanie Jegelka, Stephen Bates, Tommi Jaakkola
Abstract:
Abstract:Diffusion models can be improved with additional guidance towards more effective representations of input. Indeed, prior empirical work has already shown that aligning internal representations of the diffusion model with those of pre-trained models improves generation quality. In this paper, we present a systematic framework for incorporating representation guidance into diffusion models. We provide alternative decompositions of denoising models along with their associated training criteria, where the decompositions determine when and how the auxiliary representations are incorporated. Guided by our theoretical insights, we introduce two new strategies for enhancing representation alignment in diffusion models. First, we pair examples with target representations either derived from themselves or arisen from different synthetic modalities, and subsequently learn a joint model over the multimodal pairs. Second, we design an optimal training curriculum that balances representation learning and data generation. Our experiments across image, protein sequence, and molecule generation tasks demonstrate superior performance as well as accelerated training. In particular, on the class-conditional ImageNet $256\times 256$ benchmark, our guidance results in $23.3$ times faster training than the original SiT-XL as well as four times speedup over the state-of-the-art method REPA~\cite{yu2024representation}.



Paperid:1411
Authors:Scott Cheng, Meng-Yu Tsai, Ding-Yong Hong, Mahmut T Kandemir
Abstract:
AlphaZero has achieved remarkable success in complex decision-making problems through self-play and neural network training. However, its self-play process remains inefficient due to limited exploration of high-uncertainty positions, the overlooked runner-up decisions in Monte Carlo Tree Search (MCTS), and high variance in value labels. To address these challenges, we propose and evaluate uncertainty-guided exploration by branching from high-uncertainty positions using our proposed Label Change Rate (LCR) metric, which is further refined by a Bayesian inference framework. Our proposed approach leverages runner-up MCTS decisions to create multiple variations, and ensembles value labels across these variations to reduce variance. We investigate three key design parameters for our branching strategy: where to branch, how many variations to branch, and which move to play in the new branch. Our empirical findings indicate that branching with 10 variations per game provides the best performance-exploration balance. Overall, our end-to-end results show an improved sample efficiency over the baseline by 58.5\% on 9x9 Go in the early stage of training and by 47.3\% on 19x19 Go in the late stage of training.



Paperid:1412
Authors:Wei Chen, Jingxi Yu, Zichen Miao, Qiang Qiu
Abstract:
Recent advances in AI, driven by Transformer architectures, have achieved remarkable success in language, vision, and multimodal reasoning, and there is growing demand for them to address in-context compositional learning tasks. In these tasks, models solve the target problems by inferring compositional rules from context examples, which are composed of basic components structured by underlying rules. However, some of these tasks remain challenging for Transformers, which are not inherently designed to handle compositional tasks and offer limited structural inductive bias.Inspired by sparse coding, we propose a reformulation of the attention to enhance its capability for compositional tasks. In sparse coding, data are represented as sparse combinations of basic elements, with the resulting coefficients capturing the underlying compositional structure of the input. Specifically, we reinterpret the standard attention block as projecting inputs into outputs through projections onto two sets of learned dictionary atoms: anencoding dictionaryand adecoding dictionary. The encoding dictionary decomposes the input into a set of coefficients, which represent the compositional structure of the input. To enhance structured representations, we impose sparsity on these coefficients. The sparse coefficients are then used to linearly combine the decoding dictionary atoms to generate the output. Furthermore, to assist compositional generalization tasks, we propose estimating the coefficients of the target problem as a linear combination of the coefficients obtained from the context examples.We demonstrate the effectiveness of our approach on the S-RAVEN and RAVEN datasets. For certain compositional generalization tasks, our method maintains performance even when standard Transformers fail, owing to its ability to learn and apply compositional rules.



Paperid:1413
Authors:Jiuqi Wang, Rohan Chandra, Shangtong Zhang
Abstract:
Typically, a modern reinforcement learning (RL) agent solves a task by updating its neural network parameters to adapt its policy to the task. Recently, it has been observed that some RL agents can solve a wide range of new out-of-distribution tasks without parameter updates after pretraining on some task distribution. When evaluated in a new task, instead of making parameter updates, the pretrained agent conditions its policy on additional input called the context, e.g., the agent's interaction history in the new task. The agent's performance increases as the information in the context increases, with the agent's parameters fixed. This phenomenon is typically called in-context RL (ICRL). The pretrained parameters of the agent network enable the remarkable ICRL phenomenon. However, many ICRL works perform the pretraining with standard RL algorithms. This raises the central question this paper aims to address: Why can the RL pretraining algorithm generate network parameters that enable ICRL? We hypothesize that the parameters capable of ICRL are minimizers of the pretraining loss. This work provides initial support for this hypothesis through a case study. In particular, we prove that when a Transformer is pretrained for policy evaluation, one of the global minimizers of the pretraining loss can enable in-context temporal difference learning.



Paperid:1414
Authors:Lexiang Xiong, Liu Chengyu, Jingwen Ye, YAN LIU, Yuecong Xu
Abstract:
With the growing power of text-to-image diffusion models, their potential to generate harmful or biased content has become a pressing concern, motivating the development of concept erasure techniques. Existing approaches, whether relying on retraining or not, frequently compromise the generative capabilities of the target model in achieving concept erasure, often struggling with a critical trade-off between erasure completeness and the preservation of unrelated content (locality). Here, we introduce Semantic Surgery, a novel training-free framework for zero-shot concept erasure. Semantic Surgery directly operates on text embeddings before the diffusion process, aiming to neutralize undesired concepts at their semantic origin and thereby enhance both erasure completeness and the locality of generation by modifying the global semantic input to the diffusion model. Specifically, Semantic Surgery dynamically estimates the presence and intensity of target concepts within an input prompt's global semantics, based on which it performs a calibrated, scaled vector subtraction from the entire text embedding to neutralize their influence at the source. The overall framework consists of a Co-Occurrence Encoding module for robust multi-concept erasure by considering their joint semantic signatures, and an optional visual feedback loop that refines the textual embedding to address Latent Concept Persistence, thereby reinforcing erasure throughout the subsequent denoising process. Our proposed Semantic Surgery requires no model retraining and adapts dynamically to the specific concepts and their intensity detected in each input prompt, ensuring precise and context-aware interventions. Extensive experiments are conducted on object, explicit content, artistic style, and multi-celebrity erasure tasks, demonstrating that our method significantly outperforms state-of-the-art approaches. That is, our proposed concept erasure framework achieves superior completeness and robustness while preserving locality and general image quality, thereby offering an effective and practical solution in text-to-image generation.



Paperid:1415
Authors:Guangyi Zhang, Qiyu Liu, Aristides Gionis
Abstract:
Abstract:Data valuation quantifies the impact of individual data points on model performance, and Shapley values provide a principled approach to this important task due to their desirable axiomatic properties,albeit with high computational complexity. Recent breakthroughs have enabled fast computation of exact Shapley values for unweighted $k$-nearest neighbor ($k$NN) classifiers. However, extending this to weighted $k$NN models has remained a significant open challenge. The state-of-the-art methods either require quadratic time complexity or resort to approximation via sampling.In this paper, we show that a conceptually simple but overlooked approach - data duplication - can bridge this gap.However, the increased data size due to duplication incurs prohibitive computational and memory costs.We develop an efficient algorithm that avoids materializing the duplicated dataset by exploiting the structural properties of weighted $k$NN models,reducing the complexity to near-linear time in the original data size.Besides, we establish theoretical foundations for this approach through axiomatic characterization of the resulting values,and empirically validate its effectiveness and efficiency.



Paperid:1416
Authors:Qiying Yu, Zheng Zhang, Ruofei Zhu, Yufeng Yuan, Xiaochen Zuo, Yu Yue, Weinan Dai, Tiantian Fan, Gaohong Liu, Juncai Liu, LingJun Liu, Xin Liu, Haibin Lin, Zhiqi Lin, Bole Ma, Guangming Sheng, Yuxuan Tong, Chi Zhang, Mofan Zhang, Ru Zhang, Wang Zhang, Hang Zhu, Jinhua Zhu, Jiaze Chen, Jiangjie Chen, Chengyi Wang, Hongli Yu, Yuxuan Song, Xiangpeng Wei, Hao Zhou, Jingjing Liu, Wei-Ying Ma, Ya-Qin Zhang, Lin Yan, Yonghui Wu, Mingxuan Wang
Abstract:
Inference scaling empowers LLMs with unprecedented reasoning ability, with reinforcement learning as the core technique to elicit complex reasoning. However, key technical details of state-of-the-art reasoning LLMs are concealed (such as in OpenAI o1 blog and DeepSeek R1 technical report), thus the community still struggles to reproduce their RL training results. We propose theDecoupled Clip andDynamic sAmplingPolicyOptimization (DAPO) algorithm, and fully open-source a state-of-the-art large-scale RL system that achieves 50 points on AIME 2024 using Qwen2.5-32B base model. Unlike previous works that withhold training details, we introduce four key techniques of our algorithm that make large-scale LLM RL a success. In addition, we open-source our training code, which is built on the verl framework, along with a carefully curated and processed dataset. These components of our open-source system enhance reproducibility and support future research in large-scale LLM RL.



Paperid:1417
Authors:Atefeh Gilani, Naima Tasnim, Lalitha Sankar, Oliver Kosut
Abstract:
Differentially private stochastic gradient descent (DP-SGD) is the most widely used method for training machine learning models with provable privacy guarantees. A key challenge in DP-SGD is setting the per-sample gradient clipping threshold, which significantly affects the trade-off between privacy and utility. While recent adaptive methods improve performance by adjusting this threshold during training, they operate in the standard coordinate system and fail to account for correlations across the coordinates of the gradient. We propose GeoClip, a geometry-aware framework that clips and perturbs gradients in a transformed basis aligned with the geometry of the gradient distribution. GeoClip adaptively estimates this transformation using only previously released noisy gradients, incurring no additional privacy cost. We provide convergence guarantees for GeoClip and derive a closed-form solution for the optimal transformation that minimizes the amount of noise added while keeping the probability of gradient clipping under control. Experiments on both tabular and image datasets demonstrate that GeoClip consistently outperforms existing adaptive clipping methods under the same privacy budget.



Paperid:1418
Authors:Kai Wang, Dongwen Tang, Wangbo Zhao, Konstantin Schürholt, Zhangyang "Atlas" Wang, Yang You
Abstract:
Parameter generation has long struggled to match the scale of today's large vision and language models, curbing its broader utility. In this paper, we introduce Recurrent Diffusion for Large-Scale Parameter Generation (RPG), a novel framework that generates full neural network parameters—up to hundreds of millions—on a single GPU. Our approach first partitions a network's parameters into non-overlapping 'tokens', each corresponding to a distinct portion of the model. A recurrent mechanism then learns the inter-token relationships, producing 'prototypes' which serve as conditions for a diffusion process that ultimately synthesizes the full parameters. Across a spectrum of architectures and tasks—including ResNets, ConvNeXts and ViTs on ImageNet-1K and COCO, and even LoRA-based LLMs—RPG achieves performance on par with fully trained networks while avoiding excessive memory overhead. Notably, it generalizes beyond its training set to generate valid parameters for previously unseen tasks, highlighting its flexibility in dynamic and open-ended scenarios. By overcoming the longstanding memory and scalability barriers, RPG serves as a critical advance in 'AI generating AI', potentially enabling efficient weight generation at scales previously deemed infeasible.



Paperid:1419
Authors:Nikolaos Spanos, Maria Lymperaiou, Giorgos Filandrianos, Konstantinos Thomas, Athanasios Voulodimos, Giorgos Stamou
Abstract:
Recent black-box counterfactual generation frameworks fail to take into account the semantic content of the proposed edits, while relying heavily on training to guide the generation process. We propose a novel, plug-and-play black-box counterfactual generation framework, which suggests step-by-step edits based on theoretical guarantees of optimal edits to produce human-level counterfactual explanations with zero training. Our framework utilizes a pre-trained image editing diffusion model, and operates without access to the internals of the classifier, leading to an explainable counterfactual generation process. Throughout our experimentation, we showcase the explanatory gap between human reasoning and neural model behavior by utilizing both Convolutional Neural Network (CNN), Vision Transformer (ViT) and Large Vision Language Model (LVLM) classifiers, substantiated through a comprehensive human evaluation.



Paperid:1420
Authors:Yuanlin Duan, Yuning Wang, Wenjie Qiu, He Zhu
Abstract:
Despite its promise, imitation learning often fails in long-horizon environments where perfect replication of demonstrations is unrealistic and small errors can accumulate catastrophically. We introduce Cago (Capability-Aware Goal Sampling), a novel learning-from-demonstrations method that mitigates the brittle dependence on expert trajectories for direct imitation. Unlike prior methods that rely on demonstrations only for policy initialization or reward shaping, Cago dynamically tracks the agent's competence along expert trajectories and uses this signal to select intermediate steps—goals that are just beyond the agent's current reach—to guide learning. This results in an adaptive curriculum that enables steady progress toward solving the full task. Empirical results demonstrate that Cago significantly improves sample efficiency and final performance across a range of sparse-reward, goal-conditioned tasks, consistently outperforming existing learning from-demonstrations baselines.



Paperid:1421
Authors:Nicolas Espinosa-Dice, Yiyi Zhang, Yiding Chen, Bradley Guo, Owen Oertell, Gokul Swamy, Kianté Brantley, Wen Sun
Abstract:
Diffusion and flow models have emerged as powerful generative approaches capable of modeling diverse and multimodal behavior. However, applying these models to offline RL remains challenging due to the iterative nature of their noise sampling processes, making policy optimization difficult. In this paper, we introduce Scalable Offline Reinforcement Learning (SORL), a new offline RL algorithm that leverages shortcut models – a novel class of generative models – to scale both training and inference. SORL's policy can capture complex data distributions and can be trained simply and efficiently in a one-stage training procedure. At test time, SORL supports both sequential and parallel inference scaling by using the learned Q-function as a verifier. We demonstrate that SORL achieves strong performance across a range of offline RL tasks and exhibits positive scaling behavior with increased test-time compute.



Paperid:1422
Authors:Ningyi Liao, Zihao Yu, Siqiang Luo, Gao Cong
Abstract:
Graph Transformer (GT) has recently emerged as a promising neural network architecture for learning graph-structured data. However, its global attention mechanism with quadratic complexity concerning the graph scale prevents wider application to large graphs. Effectively representing graph information while ensuring learning efficiency remains challenging, as our analysis reveals that current GT designs targeting scalability still suffer from the computational bottleneck related to graph-scale operations. In this work, we tackle the GT scalability issue by proposing HubGT, a scalable Graph Transformer boosted by fully decoupled graph processing and simplified learning. HubGT represents the graph by a novel hierarchical scheme exploiting hub labels, which is shown to be more informative than plain adjacency by offering global connections while promoting locality, and is particularly suitable for handling complex graph patterns such as heterophily. We also design algorithms for efficiently constructing and querying the hub label hierarchy tailored for the GT attention training in scalable deployments. Notably, the precomputation and training processes of HubGT achieve complexities linear to the number of graph edges and nodes, respectively, while the training stage completely removes graph-related computations, leading to favorable mini-batch capability and GPU utilization. Extensive experiments demonstrate that HubGT is efficient in terms of computational enhancement and mini-batch capability over existing GT designs on large-scale benchmarks, while achieving top-tier effectiveness on both homophilous and heterophilous graphs.



Paperid:1423
Authors:Jinyeong Kim, Junhyeok Kim, Yumin Shim, Joohyeok Kim, Sunyoung Jung, Seong Jae Hwang
Abstract:
How do vision transformers (ViTs) represent and process the world? This paper addresses this long-standing question through the first systematic analysis of 6.6K features across all layers, extracted via sparse autoencoders, and by introducing the residual replacement model, which replaces ViT computations with interpretable features in the residual stream. Our analysis reveals not only a feature evolution from low-level patterns to high-level semantics, but also how ViTs encode curves and spatial positions through specialized feature types. The residual replacement model scalably produces a faithful yet parsimonious circuit for human-scale interpretability by significantly simplifying the original computations. As a result, this framework enables intuitive understanding of ViT mechanisms. Finally, we demonstrate the utility of our framework in debiasing spurious correlations.



Paperid:1424
Authors:Ran Zhu, Mingkun Yang, Shiqiang Wang, Jie Yang, Qing Wang
Abstract:
Abstract:Federated Learning (FL) has emerged as a privacy-preserving framework for training models on data generated at the edge. However, the heterogeneity of data silos (e.g., label skew and domain shift) often leads to inconsistent learning objectives and suboptimal model performance. Inspired by the data-driven approach, we propose Flick, a novel data generation framework for heterogeneous **F**ederated **L**earning w**i**th **C**ommonsense **K**nowledge from Large Language Models (LLMs). In Flick, the client performs the local data summary to capture client-specific knowledge in textual form. The central server then distills task-relevant, high-quality knowledge from the out-of-the-box LLM -- guided by cross-client-specific insights -- to generate informative text prompts. These prompts direct a generative model in producing synthetic data, enabling global model fine-tuning and local data compensation. This process gradually aligns the label and feature distributions across clients. Extensive results demonstrate that Flick improves the global model accuracy by up to 11.43\%, and accelerates convergence by up to 12.9$\times$, validating its effectiveness in addressing data heterogeneity.



Paperid:1425
Authors:Yiyang Liu, James Liang, Heng Fan, Wenhao Yang, Yiming Cui, Xiaotian Han, Lifu Huangg, Dongfang Liu, Qifan Wang, Cheng Han
Abstract:
Prompt-based learning has emerged as a parameter-efficient finetuning (PEFT) approach to facilitate Large Language Model (LLM) adaptation to downstream tasks by conditioning generation with task-aware guidance. Despite its successes, current prompt-based learning methods heavily rely on laborious grid searching for optimal prompt length and typically require considerable number of prompts, introducing additional computational burden. Worse yet, our pioneer findings indicate that the task-aware prompt design is inherently limited by its absence of instance-aware information, leading to a subtle attention interplay with the input sequence. In contrast, simply incorporating instance-aware information as a part of the guidance can enhance the prompt-tuned model performance without additional fine-tuning. Moreover, we find an interesting phenomenon, namely "attention anchor", that incorporating instance-aware tokens at the earliest position of the sequence can successfully preserve strong attention to critical structural information and exhibit more active attention interaction with all input tokens. In light of our observation, we introduce Capsule Prompt-Tuning (CaPT), an efficient and effective solution that leverages off-the-shelf, informative instance semantics into prompt-based learning. Our approach innovatively integrates both instance-aware and task-aware information in a nearly parameter-free manner (i.e., one single capsule prompt).Empirical results demonstrate that our method can exhibit superior performance across various language tasks (e.g., 84.03\% average accuracy on T5-Large), serving as an "attention anchor," while enjoying high parameter efficiency (e.g., 0.003\% of model parameters on Llama3.2-1B).



Paperid:1426
Authors:Dacheng Li, Yunhao Fang, Yukang Chen, Shuo Yang, Shiyi Cao, Justin Wong, Michael Luo, Xiaolong Wang, Hongxu Yin, Joseph Gonzalez, Ion Stoica, Song Han, Yao Lu
Abstract:
Video generation models have rapidly progressed, positioning themselves as video world models capable of supporting decision-making applications like robotics and autonomous driving. However, current benchmarks fail to rigorously evaluate these claims, focusing only on general video quality, ignoring important factors to world models such as physics adherence.To bridge this gap, we propose WorldModelBench, a benchmark designed to evaluate the world modeling capabilities of video generation models in application-driven domains. WorldModelBench offers two key advantages: (1) Against to nuanced world modeling violations: By incorporating instruction-following and physics-adherence dimensions, WorldModelBench detects subtle violations, such as irregular changes in object size that breach the mass conservation law—issues overlooked by prior benchmarks. (2) Aligned with large-scale human preferences: We crowd-source 67K human labels to accurately measure 14 frontier models. Using our high-quality human labels, we further fine-tune an accurate judger to automate the evaluation procedure, achieving 9.9% lower error in predicting world modeling violations than GPT-4o with 2B parameters. In addition, we demonstrate that training to align human annotations by maximizing the rewards from the judger noticeably improve the world modeling capability.



Paperid:1427
Authors:Pengrun Huang, Chhavi Yadav, Ruihan Wu, Kamalika Chaudhuri
Abstract:
Large language models (LLMs) are increasingly fine-tuned on domain-specific datasets to support applications in fields such as healthcare, finance, and law. These fine-tuning datasets often have sensitive and confidential dataset-level properties — such as patient demographics or disease prevalence—that are not intended to be revealed. While prior work has studied property inference attacks on discriminative models (e.g., image classification models) and generative models (e.g., GANs for image data), it remains unclear if such attacks transfer to LLMs. In this work, we introduce PropInfer, a benchmark task for evaluating property inference in LLMs under two fine-tuning paradigms: question-answering and chat-completion. Built on the ChatDoctor dataset, our benchmark includes a range of property types and task configurations. We further propose two tailored attacks: a prompt-based generation attack and a shadow-model attack leveraging word frequency signals. Empirical evaluations across multiple pretrained LLMs show the success of our attacks, revealing a previously unrecognized vulnerability in LLMs.



Paperid:1428
Authors:Stanislas Strasman, Sobihan Surendran, Claire Boyer, Sylvain Le Corff, Vincent Lemaire, Antonio Ocello
Abstract:
Score-based Generative Models (SGMs) have achieved impressive performance in data generation across a wide range of applications and benefit from strong theoretical guarantees. Recently, methods inspired by statistical mechanics, in particular Hamiltonian dynamics, have introduced Critically-damped Langevin Diffusions (CLD), which define diffusion processes in extended spaces by coupling the data with auxiliary variables. These approaches, along with their associated score-matching and sampling procedures, have been shown to outperform standard diffusion-based samplers numerically. In this paper, we propose an upper bound on the sampling error for CLD-based generative models in the Wasserstein metric. To better exploit the extended space, we also propose a modified dynamic that introduces an additional hyperparameter controlling the noise applied to the data coordinates. This hyperparameter influences the smoothness of sample paths, and our discretization error analysis offers practical guidance for tuning, leading to improved sampling performance.



Paperid:1429
Authors:Tara Anand, Adèle Ribeiro, Jin Tian, George Hripcsak, Elias Bareinboim
Abstract:
Causal discovery approaches are limited by scalability and interpretability, and are primarily for learning relationships among variables. Learning causal relationships among sets or clusters of variables is of interest as for some applications, relationships among variables grouped in semantically meaningful ways is the goal, and in others, clusters improve causal discovery in high-dimensions by reducing dimensionality. Here, we introduce an approach for learning over clusters in Markov causal systems. We develop a new graphical model to encode knowledge of relationships between user-defined clusters while fully representing independencies and dependencies over clusters, faithful to a specific distribution. Then we define and characterize a graphical equivalence class of these models that share cluster-level independence information. Lastly, we introduce an algorithm for causal discovery, leveraging these new representations, to soundly represent learnable causal relationships between clusters of variables.



Paperid:1430
Authors:Guowei Xu, Mert Yuksekgonul, Carlos Guestrin, James Zou
Abstract:
Large language models (LLMs) are increasingly used in learning algorithms, evaluations, and optimization tasks. Recent studies have shown that using LLM-based optimizers to automatically optimize model prompts, demonstrations, predictions themselves, or other components can significantly enhance the performance of AI systems, as demonstrated by frameworks such as DSPy and TextGrad. However, optimizers built on language models themselves are usually designed by humans with manual design choices; optimizers themselves are not optimized. Moreover, these optimizers are general purpose by design, to be useful to a broad audience, and are not tailored for specific tasks. To address these challenges, we propose metaTextGrad, which focuses on designing a meta-optimizer to further enhance existing optimizers and align them to be good optimizers for a given task. Our approach consists of two key components: a meta prompt optimizer and a meta structure optimizer. The combination of these two significantly improves performance across multiple benchmarks, achieving an average absolute performance improvement of up to 6% compared to the best baseline.



Paperid:1431
Authors:Taehyun Hwang, Dahngoon Kim, Min-hwan Oh
Abstract:
Abstract:We study the contextual _multinomial logistic_ (MNL) bandit problem, a widely used framework for sequential assortment selection. Although most existing research assumes utility functions to be linear in item features, this linear assumption significantly restricts the modeling of intricate interactions between items and user preferences. A recent work (Zhang & Luo, 2024) has investigated general utility function classes, yet they face fundamental trade-offs between computational tractability and statistical efficiency.To address these limitations, we propose a novel computationally efficient algorithm for contextual MNL bandits leveraging the upper confidence bound principle, specifically designed for non-linear parametric utility functions, including those modeled by neural networks. Under a realizability assumption and a mild geometric condition on the utility function class, our algorithm achieves a regret bound of $\tilde{\mathcal{O}}(\sqrt{T})$, where $T$ denotes the total number of rounds. Remarkably, this result establishes that sharp $\tilde{\mathcal{O}}(\sqrt{T})$-regret is attainable even with neural network-based utilities, without relying on strong assumptions such as neural tangent kernel approximations.To the best of our knowledge, our proposed method is the first computationally tractable algorithm for contextual MNL bandits with non-linear utilities that provably attains $\tilde{\mathcal{O}}(\sqrt{T})$ regret. Comprehensive numerical experiments validate the effectiveness of our approach, showing robust performance not only in realizable settings but also in scenarios with model misspecification.



Paperid:1432
Authors:Borna Khodabandeh, Amirabbas Afzali, Amirhossein Afsharrad, Seyed Mousavi, Sanjay Lall, Sajjad Amini, Seyed-Mohsen Moosavi-Dezfooli
Abstract:
Visual encoders have become fundamental components in modern computer vision pipelines. However, ensuring robustness against adversarial perturbations remains a critical challenge. Recent efforts have explored both supervised and unsupervised adversarial fine-tuning strategies. We identify two key limitations in these approaches: (i) they often suffer from instability, especially during the early stages of fine-tuning, resulting in suboptimal convergence and degraded performance on clean data, and (ii) they exhibit a suboptimal trade-off between robustness and clean data accuracy, hindering the simultaneous optimization of both objectives. To overcome these challenges, we proposeLagrangian-OptimizedRobustEmbeddings (LORE), a novel unsupervised adversarial fine-tuning framework. LORE utilizes constrained optimization, which offers a principled approach to balancing competing goals, such as improving robustness while preserving nominal performance. By enforcing embedding-space proximity constraints, LORE effectively maintains clean data performance throughout adversarial fine-tuning. Extensive experiments show that LORE significantly improves zero-shot adversarial robustness with minimal degradation in clean data accuracy. Furthermore, we demonstrate the effectiveness of the adversarially fine-tuned CLIP image encoder in out-of-distribution generalization and enhancing the interpretability of image embeddings. The code is available onGitHub.



Paperid:1433
Authors:Liuhao Lin, Ke Li, Zihan Xu, Yuchen Shi, Yulei Qin, Yan Zhang, Xing Sun, Rongrong Ji
Abstract:
Current evaluation paradigms for large language models (LLMs) represent a critical blind spot in AI research—relying on opaque numerical metrics that conceal fundamental limitations in spatial reasoning while providing no intuitive understanding of model capabilities. This deficiency creates a dangerous disconnect between reported performance and practical abilities, particularly for applications requiring physical world understanding. We introduce LTD-Bench, a breakthrough benchmark that transforms LLM evaluation from abstract scores to directly observable visual outputs by requiring models to generate drawings through dot matrices or executable code. This approach makes spatial reasoning limitations immediately apparent even to non-experts, bridging the fundamental gap between statistical performance and intuitive assessment. LTD-Bench implements a comprehensive methodology with complementary generation tasks (testing spatial imagination) and recognition tasks (assessing spatial perception) across three progressively challenging difficulty levels, methodically evaluating both directions of the critical language-spatial mapping. Our extensive experiments with state-of-the-art models expose an alarming capability gap: even LLMs achieving impressive results on traditional benchmarks demonstrate profound deficiencies in establishing bidirectional mappings between language and spatial concepts—a fundamental limitation that undermines their potential as genuine world models. Furthermore, LTD-Bench's visual outputs enable powerful diagnostic analysis, offering a potential approach to investigate model similarity. Our dataset and codes are available at https://anonymous.4open.science/r/LTD-Bench-D324.



Paperid:1434
Authors:Eduardo Sánchez, Belen Alastruey, Christophe Ropers, Arina Turkatenko, Pontus Lars Erik Saito Stenetorp, Mikel Artetxe, Marta Ruiz Costa-jussà
Abstract:
We propose a new benchmark to measure a language model's linguistic reasoning skills without relying on pre-existing language-specific knowledge. The test covers 894 questions grouped in 160 problems across 75 (mostly) extremely low-resource languages, extracted from the International Linguistic Olympiad corpus. To attain high accuracy on this benchmark, models don't need previous knowledge of the tested language, as all the information needed to solve the linguistic puzzle is presented in the context. We find that, while all analyzed models rank below 25% accuracy, there is a significant gap between open and closed models, with the best-performing proprietary model scoring 24.05% and the best-performing open model 8.84%.



Paperid:1435
Authors:Hengzhi Li, Megan Tjandrasuwita, Yi R. (May) Fung, Armando Solar-Lezama, Paul Pu Liang
Abstract:
As AI becomes more closely integrated with peoples' daily activities, socially intelligent AI that can understand and interact seamlessly with humans in daily lives is increasingly important. However, current works in AI social reasoning all rely on language-only or language-dominant approaches to benchmark and training models, resulting in systems that are improving in verbal communication but struggle with nonverbal social understanding. To address this limitation, we tap into a novel data source rich in nonverbal social interactions -- mime videos. Mimes refer to the art of expression through gesture and movement without spoken words, which presents unique challenges and opportunities in interpreting nonverbal social communication. We contribute a new dataset called MimeQA, obtained by sourcing ~8 hours of videos clips from YouTube and developing a comprehensive video question-answering benchmark comprising 806 carefully annotated and verified question-answer pairs, designed to probe nonverbal social reasoning capabilities. Using MimeQA, we evaluate state-of-the-art video large language models (vLLMs) and find that they achieve low overall accuracy, ranging from 20-30\%, why humans score 86\%. Our analysis reveals that vLLMs often fail to ground imagined objects and over-rely on the text prompt while ignoring subtle nonverbal interactions. We hope to inspire future work in AI models that embody true social intelligence capable of interpreting non-verbal human interactions.



Paperid:1436
Authors:Jaesik Yoon, Hyeonseo Cho, Sungjin Ahn
Abstract:
Monte Carlo Tree Diffusion (MCTD) has emerged as a powerful framework for long-horizon planning, combining the structured exploration of tree search with the generative flexibility of diffusion models. However, MCTD remains limited by the trajectory lengths observed during training, restricting its effectiveness on tasks requiring extended reasoning. To address this, we propose Compositional Monte Carlo Tree Diffusion (C-MCTD), a novel inference-time scaling framework that includes three approaches: Online Composer, Distributed Composer and Preplan Composer. Online Composer extends the MCTD's tree search to stitch the generated plan, and Distributed Composer and Preplan Composer improves Online Composer's scalability through parallel computation and prebuilt plan graphs. Experiments demonstrate that C-MCTD methods significantly outperforms standard MCTD strategies on tasks requiring plans much longer than those seen in training, highlighting the potential of compositional inference-time scaling. Source code will be released upon publication at: \url{https://anonymous.4open.science/r/C-MCTD-CF7F}.



Paperid:1437
Authors:Maresa Schröder, Dennis Frauen, Jonas Schweisthal, Konstantin Hess, Valentyn Melnychuk, Stefan Feuerriegel
Abstract:
Uncertainty quantification of causal effects is crucial for safety-critical applications such as personalized medicine. A powerful approach for this is conformal prediction, which has several practical benefits due to model-agnostic finite-sample guarantees. Yet, existing methods for conformal prediction of causal effects are limited to binary/discrete treatments and make highly restrictive assumptions, such as known propensity scores. In this work, we provide a novel conformal prediction method for potential outcomes of continuous treatments. We account for the additional uncertainty introduced through propensity estimation so that our conformal prediction intervals are valid even if the propensity score is unknown. Our contributions are three-fold: (1) We derive finite-sample validity guarantees for prediction intervals of potential outcomes of continuous treatments. (2) We provide an algorithm for calculating the derived intervals. (3) We demonstrate the effectiveness of the conformal prediction intervals in experiments on synthetic and real-world datasets. To the best of our knowledge, we are the first to propose conformal prediction for continuous treatments when the propensity score is unknown and must be estimated from data.



Paperid:1438
Authors:Xiaoqi Liu, Dorian Baudry, Julian Zimmert, Patrick Rebeschini, Arya Akhavan
Abstract:
Abstract:Bandit Convex Optimization is a fundamental class of sequential decision-making problems, where the learner selects actions from a continuous domain and observes a loss (but not its gradient) at only one point per round. We study this problem in non-stationary environments, and aim to minimize the regret under three standard measures of non-stationarity: the number of switches $S$ in the comparator sequence, the total variation $\Delta$ of the loss functions, and the path-length $P$ of the comparator sequence. We propose a polynomial-time algorithm, Tilted Exponentially Weighted Average with Sleeping Experts (TEWA-SE), which adapts the sleeping experts framework from online convex optimization to the bandit setting. For strongly convex losses, we prove that TEWA-SE is minimax-optimal with respect to known $S$ and $\Delta$ by establishing matching upper and lower bounds. By equipping TEWA-SE with the Bandit-over-Bandit framework, we extend our analysis to environments with unknown non-stationarity measures. For general convex losses, we introduce a second algorithm, clipped Exploration by Optimization (cExO), based on exponential weights over a discretized action space. While not polynomial-time computable, this method achieves minimax-optimal regret with respect to known $S$ and $\Delta$, and improves on the best existing bounds with respect to $P$.



Paperid:1439
Authors:Si Chen, Yifei Li, Richong Zhang
Abstract:
Inductive reasoning infers general rules from observed evidence, which is one of the most critical intelligence abilities.Previous works have succeeded in formal languages but suffer from onerous and error-prone conversions between a particular formal language and the working language.As large language models (LLMs) have emerged, direct reasoning with various kinds of languages, especially natural languages, without formal language involvement has become feasible.However, existing LLM-based inductive reasoning usually relies on LLM's intrinsic generation ability, which is prone to LLM's hallucination and lacks systematic guidance according to the nature of inductive reasoning.To this end, we propose HypoBootstrap, an integrated framework for inductive reasoning that generates and confirms hypotheses both in a bootstrapping manner.Regarding hypothesis generation, we propose a novel bootstrapping generation strategy, bootstrapping object hypotheses, relational hypotheses, and functional hypotheses successively, assisting LLM in observing the evidence from trivial patterns to non-trivial patterns.Regarding hypothesis confirmation, we utilize Glymour's theory of bootstrap confirmation, a hypothesis confirmation theory from the philosophy of science that can confirm a set of hypotheses.% By our meticulous design, this confirmation theory can coordinate appropriately with our bootstrapping generation strategy to confirm the object hypotheses, relational hypotheses, and functional hypotheses, which benefit from their shared bootstrapping nature.We use its principles to confirm the object hypotheses, relational hypotheses, and functional hypotheses.Empirical studies on four inductive reasoning scenarios of completely different natures, involving causal induction, concept learning, grammar learning, and abstract reasoning, demonstrate that HypoBootstrap significantly outperforms existing methods.



Paperid:1440
Authors:Jialong Zuo, Yongtai Deng, Lingdong Kong, Jingkang Yang, Rui Jin, Yiwei Zhang, Nong Sang, Liang Pan, Ziwei Liu, Changxin Gao
Abstract:
Recent studies have shown that agent-based systems leveraging large language models (LLMs) for key information retrieval and integration have emerged as a promising approach for long video understanding. However, these systems face two major challenges. First, they typically perform modeling and reasoning on individual frames, struggling to capture the temporal context of consecutive frames. Second, to reduce the cost of dense frame-level captioning, they adopt sparse frame sampling, which risks discarding crucial information. To overcome these limitations, we propose VideoLucy, a deep memory backtracking framework for long video understanding. Inspired by the human recollection process from coarse to fine, VideoLucy employs a hierarchical memory structure with progressive granularity. This structure explicitly defines the detail level and temporal scope of memory at different hierarchical depths. Through an agent-based iterative backtracking mechanism, VideoLucy systematically mines video-wide, question-relevant deep memories until sufficient information is gathered to provide a confident answer. This design enables effective temporal understanding of consecutive frames while preserving critical details. In addition, we introduce EgoMem, a new benchmark for long video understanding. EgoMem is designed to comprehensively evaluate a model's ability to understand complex events that unfold over time and capture fine-grained details in extremely long videos. Extensive experiments demonstrate the superiority of VideoLucy. Built on open-source models, VideoLucy significantly outperforms state-of-the-art methods on multiple long video understanding benchmarks, achieving performance even surpassing the latest proprietary models such as GPT-4o. Our code and dataset will be made publicly available.



Paperid:1441
Authors:Haorun Cai, Han-Jia Ye
Abstract:
While tabular machine learning has achieved remarkable success, temporal distribution shifts pose significant challenges in real-world deployment, as the relationships between features and labels continuously evolve. Static models assume fixed mappings to ensure generalization, whereas adaptive models may overfit to transient patterns, creating a dilemma between robustness and adaptability.In this paper, we analyze key factors essential for constructing an effective dynamic mapping for temporal tabular data. We discover that evolving feature semantics—particularly objective and subjective meanings—introduce concept drift over time. Crucially, we identify that feature transformation strategies are able to mitigate discrepancies in feature representations across temporal stages.Motivated by these insights, we propose a feature-aware temporal modulation mechanism that conditions feature representations on temporal context, modulating statistical properties such as scale and skewness. By aligning feature semantics across time, our approach achieves a lightweight yet powerful adaptation, effectively balancing generalizability and adaptability.Benchmark evaluations validate the effectiveness of our method in handling temporal shifts in tabular data.



Paperid:1442
Authors:Ruilin Li, Hang Yu, Jiayan Qiu
Abstract:
Conventional video outpainting methods primarily focus on maintaining coherent textures and visual consistency across frames.However, they often fail at handling dynamic scenes due to the complex motion of objects or camera movement, leading to temporal incoherence and visible flickering artifacts across frames. This is primarily because they lack instance-aware modeling to accurately separate and track individual object motions throughout the video. In this paper, we propose a novel video outpainting framework that explicitly takes shadow-object pairs into consideration to enhance the temporal and spatial consistency of instances, even when they are temporarily invisible. Specifically, we first track the shadow-object pairs across frames and predict the instances in the scene to unveil the spatial regions of invisible instances. Then, these prediction results are fed to guide the instance-aware optical flow completion to unveil the temporal motion of invisible instances. Next, these spatiotemporal guidances of instances are used to guide the video outpainting process. Finally, a video-aware discriminator is implemented to enhance alignment among dynamic shadows and the extended semantics in the scene. Comprehensive experiments underscore the superiority of our approach, outperforming existing state-of-the-art methods in widely recognized benchmarks.



Paperid:1443
Authors:Siqi Wan, Jingwen Chen, Qi Cai, Yingwei Pan, Ting Yao, Tao Mei
Abstract:
Diffusion models have yielded remarkable success in virtual try-on (VTON) task, yet they often fall short of fully meeting user expectations regarding visual quality and detail preservation. To alleviate this issue, we curate a dataset of synthesized VTON images annotated with human judgments across multiple perceptual criteria. A vision large language model (VLLM), namely VTON-VLLM, is then learnt on these annotations. VTON-VLLM functions as a unified ``fashion expert'' and is capable of both evaluating and steering VTON synthesis towards human preferences. Technically, beyond serving as an automatic VTON evaluator, VTON-VLLM upgrades VTON model through two pivotal ways: (1) providing fine-grained supervisory signals during the training of a plug-and-play VTON refinement model, and (2) enabling adaptive and preference-aware test-time scaling at inference. To benchmark VTON models more holistically, we introduce VITON-Bench, a challenging test suite of complex try-on scenarios, and human-preference–aware metrics. Extensive experiments demonstrate that powering VTON models with our VTON-VLLM markedly enhances alignment with human preferences.



Paperid:1444
Authors:Hagay Michaeli, Daniel Soudry
Abstract:
Transformers have emerged as a competitive alternative to convnets in vision tasks, yet they lack the architectural inductive bias of convnets, which may hinder their potential performance. Specifically, Vision Transformers (ViTs) are not translation‑invariant and are more sensitive to minor image translations than standard convnets.Previous studies have shown, however, that convnets are also not perfectly shift‑invariant, due to aliasing in down‑sampling and non‑linear layers. Consequently, anti‑aliasing approaches have been proposed to certify convnets translation robustness. Building on this line of work, we propose an Alias‑Free ViT, which combines two main components. First, it uses alias-free down‑sampling and non‑linearities. Second, it uses linear cross‑covariance attention that is shift‑invariant to both integer and fractional translations.Our model maintains competitive performance in image classification and outperforms similar‑sized models in terms of robustness to adversarial translations.



Paperid:1445
Authors:Senyu Hou, Yuru Ren, Gaoxia Jiang, Wenjian Wang
Abstract:
Noisy labels often compel models to overfit, especially in multi-label classification tasks. Existing methods for noisy multi-label learning (NML) primarily follow a discriminative paradigm, which relies on noise transition matrix estimation or small-loss strategies to correct noisy labels. However, they remain substantial optimization difficulties compared to noisy single-label learning. In this paper, we propose a Co-Occurrence-Aware Diffusion (CAD) model, which reformulates NML from a generative perspective. We treat features as conditions and multi-labels as diffusion targets, optimizing the diffusion model for multi-label learning with theoretical guarantees. Benefiting from the diffusion model's strength in capturing multi-object semantics and structured label matrix representation, we can effectively learn the posterior mapping from features to true multi-labels. To mitigate the interference of noisy labels in the forward process, we guide generation using pseudo-clean labels reconstructed from the latent neighborhood space, replacing original point-wise estimates with neighborhood-based proxies. In the reverse process, we further incorporate label co-occurrence constraints to enhance the model's awareness of incorrect generation directions, thereby promoting robust optimization. Extensive experiments on both synthetic (Pascal-VOC, MS-COCO) and real-world (NUS-WIDE) noisy datasets demonstrate that our approach outperforms state-of-the-art methods.



Paperid:1446
Authors:Yichi Zhang, Jinlong Pang, Zhaowei Zhu, Yang Liu
Abstract:
The recent success of generative AI highlights the crucial role of high-quality human feedback in building trustworthy AI systems. However, the increasing use of large language models (LLMs) by crowdsourcing workers poses a significant challenge: datasets intended to reflect human input may be compromised by LLM-generated responses. Existing LLM detection approaches often rely on high-dimensional training data such as text, making them unsuitable for structured annotation tasks like multiple-choice labeling. In this work, we investigate the potential of peer prediction --- a mechanism that evaluates the information within workers' responses --- to mitigate LLM-assisted cheating in crowdsourcing with a focus on annotation tasks. Our method quantifies the correlations between worker answers while conditioning on (a subset of) LLM-generated labels available to the requester. Building on prior research, we propose a training-free scoring mechanism with theoretical guarantees under a novel model that accounts for LLM collusion. We establish conditions under which our method is effective and empirically demonstrate its robustness in detecting low-effort cheating on real-world crowdsourcing datasets.



Paperid:1447
Authors:DongJae Lee, Jiwan Hur, Jaehyun Choi, Jaemyung Yu, Junmo Kim
Abstract:
Vision Transformers have demonstrated exceptional performance across various computer vision tasks, yet their quadratic computational complexity concerning token length remains a significant challenge. To address this, token reduction methods have been widely explored. However, existing approaches often overlook the frequency characteristics of self-attention, such as rank collapsing and over-smoothing phenomenon.In this paper, we propose a frequency-aware token reduction strategy that improves computational efficiency while preserving performance by mitigating rank collapsing. Our method partitions tokens into high-frequency (HF) tokens and low-frequency tokens (LF). HF tokens are selectively preserved, while LF tokens are aggregated into a compact DC token to retain essential low-frequency components. Through extensive experiments and analysis, we demonstrate that our approach significantly improves accuracy while reducing computational overhead and mitigating rank collapsing and over smoothing. Furthermore, we analyze the previous methods, shedding light on their implicit frequency characteristics and limitations.



Paperid:1448
Authors:Abdou Karim KANDJI, Frederic Precioso, Cheikh BA, Augustin NDIONE, Samba NDIAYE
Abstract:
Intent classification models have made a lot of progress in recent years. However, previous studies primarily focus on high-resource languages datasets, which results in a gap for low-resource languages and for regions with a high rate of illiterate people where languages are more spoken than read or written. This is the case in Senegal, for example, where Wolof is spoken by around 90\% of the population, with an illiteracy rate of 42\% for the country. Wolof is actually spoken by more than 10 million people in West African region. To tackle such limitations, we release a Wolof Intent Classification Dataset (WolBanking77), for academic research in intent classification. WolBanking77 currently contains 9,791 text sentences in the banking domain and more than 4 hours of spoken sentences. Experiments on various baselines are conducted in this work, including text and voice state-of-the-art models. The results are very promising on this current dataset. This paper also provide detailed analyses of the contents of the data. We report baseline f1-score and word error rate metrics respectively on NLP and ASR models trained on WolBanking77 dataset and also comparisons between models. We plan to share and conduct dataset maintenance, updates and to release open-source code.



Paperid:1449
Authors:Pierre Marza, Leo Fillioux, Sofiène Boutaj, KUNAL MAHATHA, Christian Desrosiers, Pablo Piantanida, Jose Dolz, Stergios Christodoulidis, Maria Vakalopoulou
Abstract:
Progress in a research field can be hard to assess, in particular when many concurrent methods are proposed in a short period of time. This is the case in digital pathology, where many foundation models have been released recently to serve as feature extractors for tile-level images, being used in a variety of downstream tasks, both for tile- and slide-level problems. Benchmarking available methods then becomes paramount to get a clearer view of the research landscape. In particular, in critical domains such as healthcare, a benchmark should not only focus on evaluating downstream performance, but also provide insights about the main differences between methods, and importantly, further consider uncertainty and robustness to ensure a reliable usage of proposed models. For these reasons, we introduceTHUNDER, a tile-level benchmark for digital pathology foundation models, allowing for efficient comparison of many models on diverse datasets with a series of downstream tasks, studying their feature spaces and assessing the robustness and uncertainty of predictions informed by their embeddings.THUNDERis a fast, easy-to-use, dynamic benchmark that can already support a large variety of state-of-the-art foundation, as well as local user-defined models for direct tile-based comparison. In this paper, we provide a comprehensive comparison of 23 foundation models on 16 different datasets covering diverse tasks, feature analysis, and robustness. The code forTHUNDERis publicly available at https://github.com/MICS-Lab/thunder/tree/neurips2025datasetsand_benchmark.



Paperid:1450
Authors:Thomas Kuntz, Agatha Duzan, Hao Zhao, Francesco Croce, J. Zico Kolter, Nicolas Flammarion, Maksym Andriushchenko
Abstract:
Computer use agents are a new way of building agents that can interact with a computer directly by processing screenshots or accessibility trees. Despite their emerging popularity, their safety has been mostly overlooked. However, evaluating potential harmful behaviors and failure cases of agents represents a crucial step towards their widespread adoption. To this end, we introduce OS-Harm, a new benchmark for measuring safety of computer use agents. OS-Harm is built on top of the OSWorld environment and aims to test models across three categories of harm: deliberate user misuse, prompt injection attacks, and model misbehavior.To cover these cases, we collect 100 base and 150 augmented tasks spanning several types of safety violations (harassment, copyright infringement, disinformation, data exfiltration, etc.) and that require the agent to interact with a variety of applications. Moreover, we propose an automated judge to evaluate both the accuracy (whether the task is successfully completed) and the safety (whether the actions are safe) of the agent.Finally, we benchmark state-of-the-art computer use agents on OS-Harm, as well as the automated judges. We believe our benchmark can help the community both to measure the safety of current and future computer use agents, and to develop novel task-specific techniques to red team agents before deployment.



Paperid:1451
Authors:Adam Stein, Neelay Velingker, Mayur Naik, Eric Wong
Abstract:
Large language models (LLMs) excel at zero-shot inference but continue to struggle with complex, multi-step reasoning. Recent methods that augment LLMs with intermediate reasoning steps such as Chain of Thought (CoT) and Program of Thought (PoT) improve performance but often produce undesirable solutions, especially in algorithmic domains.We introduce Per-Instance Program Synthesis (PIPS), a method that generates and refines programs at the instance level using structural feedback without relying on task-specific guidance or explicit test cases. To further improve performance, PIPS incorporates a confidence metric that dynamically chooses between direct inference and program synthesis on a per-instance basis.Experiments across three frontier LLMs and 30 benchmarks including all tasks of Big Bench Extra Hard (BBEH), visual question answering tasks, relational reasoning tasks, and mathematical reasoning tasks show that PIPS improves the absolute harmonic mean accuracy by up to 8.6\% and 9.4\% compared to PoT and CoT respectively,and reduces undesirable program generations by 65.1\% on the algorithmic tasks compared to PoT with Gemini-2.0-Flash.



Paperid:1452
Authors:Arthur Pellegrino, Angus Chadwick
Abstract:
How networks internally encode variables of the world in their activity is a fundamental question in both neuroscience and machine learning. Characterising such "neural representations" has reshaped our understanding of computation in biological and artificial neural networks. In parallel, there is mounting interest in understanding how dynamical systems perform computations on task variables through their internal dynamics. Yet, the link between computation-through-dynamics and representational geometry is poorly understood. Here, we hypothesise that recurrent neural networks (RNNs) perform computations by dynamically warping their internal representation of task variables. To test this hypothesis, we develop a Riemannian geometric framework which enables the derivation of the manifold topology and geometry of a dynamical system from the manifold of its inputs. By characterising the time-varying geometry of task-trained RNNs, we show that dynamic warping is a fundamental feature of their computations.



Paperid:1453
Authors:Guangze Zheng, Shijie Lin, Haobo Zuo, Si Si, Ming-Shan Wang, Changhong Fu, Jia Pan
Abstract:
This work proposes the Lattice Boltzmann Model (LBM) to learn real-world pixel dynamicity for visual tracking.LBM decomposes visual representations into dynamic pixel lattices and solves pixel motion states through collision-streaming processes. Specifically, the high-dimensional distribution of the target pixels is acquired through a multilayer predict-update network to estimate the pixel positions and visibility. The predict stage formulates lattice collisions among the spatial neighborhood of target pixels and develops lattice streaming within the temporal visual context. The update stage rectifies the pixel distributions with online visual representations. Compared with existing methods, LBM demonstrates practical applicability in an online and real-time manner, which can efficiently adapt to real-world visual tracking tasks. Comprehensive evaluations of real-world point tracking benchmarks such as TAP-Vid and RoboTAP validate LBM's efficiency. A general evaluation of large-scale open-world object tracking benchmarks such as TAO, BFT, and OVT-B further demonstrates LBM's real-world practicality.



Paperid:1454
Authors:Qing Mao, Tianxin Huang, Yu Zhu, Jinqiu Sun, Yanning Zhang, Gim Hee Lee
Abstract:
Pairwise camera pose estimation from sparsely overlapping image pairs remains a critical and unsolved challenge in 3D vision. Most existing methods struggle with image pairs that have small or no overlap. Recent approaches attempt to address this by synthesizing intermediate frames using video interpolation and selecting key frames via a self-consistency score. However, the generated frames are often blurry due to small overlap inputs, and the selection strategies are slow and not explicitly aligned with pose estimation.To solve these cases, we propose Hybrid Video Generation (HVG) to synthesize clearer intermediate frames by coupling a video interpolation model with a pose-conditioned novel view synthesis model, where we also propose a Feature Matching Selector (FMS) based on feature correspondence to select intermediate frames appropriate for pose estimation from the synthesized results. Extensive experiments on Cambridge Landmarks, ScanNet, DL3DV-10K, and NAVI demonstrate that, compared to existing SOTA methods, PoseCrafter can obviously enhance the pose estimation performances, especially on examples with small or no overlap.



Paperid:1455
Authors:Wenhao Li, Qiangchang Wang, Xianjing Meng, Zhibin Wu, Yilong Yin
Abstract:
Few-shot learning (FSL) aims to recognize novel concepts from only a few labeled support samples. Recent studies enhance support features by incorporating additional semantic information (e.g., class descriptions) or designing complex semantic fusion modules. However, these methods still suffer from hallucinating semantics that contradict the visual evidence due to the lack of grounding in actual instances, resulting in noisy guidance and costly corrections. To address these issues, we propose a novel framework, bridging Vision and Text with LLMs for Few-Shot Learning (VT-FSL), which constructs precise cross-modal prompts conditioned on Large Language Models (LLMs) and support images, seamlessly integrating them through a geometry-aware alignment mechanism. It mainly consists of Cross-modal Iterative Prompting (CIP) and Cross-modal Geometric Alignment (CGA). Specifically, the CIP conditions an LLM on both class names and support images to generate precise class descriptions iteratively in a single structured inference pass. These descriptions not only enrich the semantic understanding of novel classes, but also enable the zero-shot synthesis of semantically consistent support images. The generated descriptions and synthetic images act as complementary textual and visual prompts, providing high-level class semantics and low-level intra-class diversity to compensate for limited support data. Furthermore, the CGA jointly aligns the fused textual, support, and synthetic visual representations by minimizing the kernelized volume of the 3-dimensional parallelotope they span. It captures global and nonlinear relationships among all representations, enabling structured and consistent multimodal integration. The proposed VT-FSL method establishes new state-of-the-art performance across ten diverse benchmarks, including standard, cross-domain, and fine-grained few-shot learning scenarios. Code is available at https://anonymous.4open.science/r/VT-FSL-27B4.



Paperid:1456
Authors:Zizhao Chen, Yoav Artzi
Abstract:
We propose KnotGym, an interactive environment for complex, spatial reasoning and manipulation. KnotGym includes goal-oriented rope manipulation tasks with varying levels of complexity, all requiring acting from pure image observations.Tasks are defined along a clear and quantifiable axis of complexity based on the number of knot crossings, creating a natural generalization test.KnotGym has a simple observation space, allowing for scalable development, yet it highlights core challenges in integrating acute perception, spatial reasoning, and grounded manipulation.We evaluate methods of different classes, including model-based RL, model-predictive control, and chain-of-thought reasoning, and illustrate the challenges KnotGym presents.



Paperid:1457
Authors:Hui He, Guansong Pang, Kun Yi, Yuanchi Ma, Qi Zhang, Zhendong Niu
Abstract:
The recent boom of large pre-trained models witnesses remarkable success in developing foundation models (FMs) for time series forecasting. Despite impressive performance across diverse downstream forecasting tasks, existing time series FMs possess massive network architectures and require substantial pre-training on large-scale datasets, which significantly hinders their deployment in resource-constrained environments. In response to this growing tension between versatility and affordability, we proposeSEMPO, a novel lightweight foundation model that requires pretraining on relatively small-scale data, yet exhibits strong general time series forecasting. Concretely, SEMPO comprises two key modules: 1) energy-awareSpEctral decomposition module, that substantially improves the utilization of pre-training data by modeling not only the high-energy frequency signals but also the low-energy yet informative frequency signals that are ignored in current methods; and 2)Mixture-of-PrOmpts enabled Transformer, that learns heterogeneous temporal patterns through small dataset-specific prompts and adaptively route time series tokens to these prompt-based experts for parameter-efficient model adaptation across different datasets and domains. Equipped with these modules, SEMPO significantly reduces both pre-training data scale and model size, while achieving strong generalization. Extensive experiments on two large-scale benchmarks covering 16 datasets demonstrate the superior performance of SEMPO in zero-shot and few-shot forecasting scenarios compared with state-of-the-art methods.



Paperid:1458
Authors:Mengjie Wu, Zhenjun Lin, Yihao Huang, Kangjie Chen, Yuyang zhang, Yuhan Huang, Run Wang, Lina Wang
Abstract:
Large language models (LLMs) are inherently designed to support multi-turn interactions, which opens up new possibilities for jailbreak attacks that unfold gradually and potentially bypass safety mechanisms more effectively than single-turn attacks. However, current multi-turn jailbreak methods are still in their early stages and suffer from two key limitations. First, they all inherently require inserting sensitive phrases into the context, which makes the dialogue appear suspicious and increases the likelihood of rejection, undermining the effectiveness of the attack. Second, even when harmful content is generated, the response often fails to align with the malicious prompt due to semantic drift, where the conversation slowly moves away from its intended goal. To address these challenges, we propose an analogy-based black-box multi-turn jailbreak framework that constructs fully benign contexts to improve attack success rate while ensuring semantic alignment with the malicious intent. The method first guides the model through safe tasks that mirror the response structure of the malicious prompt, enabling it to internalize the format without exposure to sensitive content. A controlled semantic shift is then introduced in the final turn, substituting benign elements with malicious ones while preserving structural coherence. Experiments on six commercial and open-source LLMs, two benchmark datasets show that our method significantly improves attack performance, achieving an average attack success rate of 93.3\% and outperforming five competitive baselines. Our code is released at https://anonymous.4open.science/r/AMA-E1C4



Paperid:1459
Authors:Xiang Li, Zirui Wang, Zixuan Huang, James Rehg
Abstract:
Humans and traditional computer vision methods rely on a diverse set of monocular cues to infer 3D structure from a single image, such as shading, texture, silhouette, etc. While recent deep generative models have dramatically advanced single-image 3D generation, it remains unclear which image cues these methods actually exploit. We introduce Cue3D, the first comprehensive, model-agnostic framework for quantifying the influence of individual image cues in single-image 3D generation. Our unified benchmark evaluates seven state-of-the-art methods, spanning regression-based, multi-view, and native 3D generative paradigms. By systematically perturbing cues such as shading, texture, silhouette, perspective, edges, and local continuity, we measure their impact on 3D output quality. Our analysis reveals that shape meaningfulness, not texture, dictates generalization. Geometric cues, particularly shading, are crucial for 3D generation. We further identify over-reliance on provided silhouettes and diverse sensitivities to cues such as perspective and local continuity across model families. By dissecting these dependencies, Cue3D advances our understanding of how modern 3D networks leverage classical vision cues, and offers directions for developing more transparent, robust, and controllable single-image 3D generation models.



Paperid:1460
Authors:Giannis Daras, Jeffrey Ouyang-Zhang, Krithika Ravishankar, Constantinos Daskalakis, Adam Klivans, Daniel Diaz
Abstract:
We present Ambient Protein Diffusion, a framework for training protein diffusion models that generates structures with unprecedented diversity and quality. State-of-the-art generative models are trained on computationally derived structures from AlphaFold2 (AF), as experimentally determined structures are relatively scarce. The resulting models are therefore limited by the quality of synthetic datasets. Since the accuracy of AF predictions degrades with increasing protein length and complexity, de novo generation of long, complex proteins remains challenging. Ambient Protein Diffusion overcomes this problem by treating low-confidence AF structures as corrupted data. Rather than simply filtering out low-quality AF structures, our method adjusts the diffusion objective for each structure based on its corruption level, allowing the model to learn from both high and low quality structures. Empirically, ambient protein diffusion yields major improvements: on proteins with 700 residues, diversity increases from 45% to 85% from the previous state-of-the-art, and designability improves from 70% to 88%.



Paperid:1461
Authors:Marc Lafon, Gustavo Vargas Hakim, Clément Rambour, Christian Desrosiers, Nicolas THOME
Abstract:
Vision-language models (VLMs) like CLIP exhibit strong zero-shot capabilities but often fail to generalize under distribution shifts. Test-time adaptation (TTA) allows models to update at inference time without labeled data, typically via entropy minimization. However, this objective is fundamentally misaligned with the contrastive image-text training of VLMs, limiting adaptation performance and introducing failure modes such as pseudo-label drift and class collapse. We propose CLIPTTA, a new gradient-based TTA method for vision-language models that leverages a soft contrastive loss aligned with CLIP’s pre-training objective. We provide a theoretical analysis of CLIPTTA’s gradients, showing how its batch-aware design mitigates the risk of collapse. We further extend CLIPTTA to the open-set setting, where both in-distribution (ID) and out-of-distribution (OOD) samples are encountered, using an Outlier Contrastive Exposure (OCE) loss to improve OOD detection. Evaluated on 75 datasets spanning diverse distribution shifts, CLIPTTA consistently outperforms entropy-based objectives and is highly competitive with state-of-the-art TTA methods, outperforming them on a large number of datasets and exhibiting more stable performance across diverse shifts.



Paperid:1462
Authors:Di Wu, Yibin Lei, Christof Monz
Abstract:
Neural machine translation (NMT) systems typically employ maximuma posteriori(MAP) decoding to select the highest-scoring translation from the distribution. However, recent evidence highlights the inadequacy of MAP decoding, often resulting in low-quality or even pathological hypotheses as the decoding objective is only weakly aligned with real-world translation quality. This paper proposes to directly calibrate hypothesis likelihood with translation quality from a distributional view by directly optimizing their Pearson correlation, thereby enhancing decoding effectiveness. With our method, translation with large language models (LLMs) improves substantially after limited training (2K instances per direction). This improvement is orthogonal to those achieved through supervised fine-tuning, leading to substantial gains across a broad range of metrics and human evaluations. This holds even when applied to top-performing translation-specialized LLMs fine-tuned on high-quality translation data, such as Tower, or when compared to recent preference optimization methods, like CPO. Moreover, the calibrated translation likelihood can directly serve as a strong proxy for translation quality, closely approximating or even surpassing some state-of-the-art translation quality estimation models, like CometKiwi. Lastly, our in-depth analysis demonstrates that calibration enhances the effectiveness of MAP decoding, thereby enabling greater efficiency in real-world deployment. The resulting state-of-the-art translation model, which covers 10 languages, along with the accompanying code and human evaluation data, has been released: https://anonymous.4open.science/r/calibrating-llm-mt.



Paperid:1463
Authors:Rémi Castera, Felipe Garrido-Lucero, Patrick Loiseau, Simon Mauras, Mathieu Molina, Vianney Perchet
Abstract:
Abstract:We consider matroid allocation problems under \textit{opportunity fairness} constraints: resources need to be allocated to a set of agents under matroid constraints (which includes classical problems such as bipartite matching). Agents are divided into $C$ groups according to a sensitive attribute, and an allocation is opportunity-fair if each group receives the same share proportional to the maximum feasible allocation it could achieve in isolation. We study the Price of Fairness (PoF), i.e., the ratio between maximum size allocations and maximum size opportunity-fair allocations. We first provide a characterization of the PoF leveraging the underlying polymatroid structure of the allocation problem. Based on this characterization, we prove bounds on the PoF in various settings from fully adversarial (worst-case) to fully random. Notably, one of our main results considers an arbitrary matroid structure with agents randomly divided into groups. In this setting, we prove a PoF bound as a function of the (relative) size of the largest group. Our result implies that, as long as there is no dominant group (i.e., the largest group is not too large), opportunity fairness constraints do not induce any loss of social welfare (defined as the allocation size). Overall, our results give insights into which aspects of the problem's structure affect the trade-off between opportunity fairness and social welfare.



Paperid:1464
Authors:Jingyuan Liu, Hao Qiu, Lin Yang, Mengfan Xu
Abstract:
Abstract:We study the distributed multi-agent multi-armed bandit problem with heterogeneous rewards over random communication graphs. Uniquely, at each time step $t$ agents communicate over a time-varying random graph $\mathcal{G}\_t$ generated by applying the Erdős–Rényi model to a fixed connected base graph $\mathcal{G}$ (for classical Erdos-Rényi graphs, $\mathcal{G}$ is a complete graph), where each potential edge in $\mathcal{G}$ is randomly and independently present with probability $p$. Notably, the resulting random graph is not necessarily connected at each time step. Additionally, each agent's arm rewards follow time-invariant distributions, and the reward distribution for the same arm may differ across agents. The goal is to minimize the cumulative expected regret relative to the global mean reward of each arm, defined as the average of that arm’s mean rewards across all agents. To this end, we propose a fully distributed algorithm that integrates the arm elimination strategy with the random gossip algorithm. We theoretically show that the regret upper bound is of order $\log T$ and is highly interpretable, where $T$ is the time horizon. It includes the optimal centralized regret $\mathcal{O}\left(\sum_{k: \Delta_k>0} \frac{\log T}{\Delta_k}\right)$ and an additional term $\mathcal{O}\left(\frac{N^2 \log T}{p \lambda_{N-1}(\operatorname{Lap}(\mathcal{G}))} + \frac{KN^2 \log T}{p}\right)$ where $N$ and $K$ denotes the total number of agents and arms, respectively. This term reflects the impact of $\mathcal{G}$'s algebraic connectivity $\lambda_{N-1}(\operatorname{Lap}(\mathcal{G}))$ and the link probability $p$, and thus highlights a fundamental trade-off between communication efficiency and regret. As a by-product, we show a nearly optimal regret lower bound. Finally, our numerical experiments not only show the superiority of our algorithm over existing benchmarks, but also validate the theoretical regret scaling with problem complexity.



Paperid:1465
Authors:Qingwen Zhang, Xiaomeng Zhu, Yushan Zhang, Yixi Cai, Olov Andersson, Patric Jensfelt
Abstract:
Abstract:Previous dominant methods for scene flow estimation focus mainly on input from two consecutive frames, neglecting valuable information in the temporal domain. While recent trends shift towards multi-frame reasoning, they suffer from rapidly escalating computational costs as the number of frames grows. To leverage temporal information more efficiently, we propose DeltaFlow ($\Delta$Flow), a lightweight 3D framework that captures motion cues via a $\Delta$ scheme, extracting temporal features with minimal computational cost, regardless of the number of frames. Additionally, scene flow estimation faces challenges such as imbalanced object class distributions and motion inconsistency. To tackle these issues, we introduce a Category-Balanced Loss to enhance learning across underrepresented classes and an Instance Consistency Loss to enforce coherent object motion, improving flow accuracy. Extensive evaluations on the Argoverse 2 and Waymo datasets show that $\Delta$Flow achieves state-of-the-art performance with up to 22\% lower error and $2\times$ faster inference compared to the next-best multi-frame supervised method, while also demonstrating a strong cross-domain generalization ability. The source code will be made publicly available upon acceptance.



Paperid:1466
Authors:Filippo Bigi, Sanggyu Chong, Agustinus Kristiadi, Michele Ceriotti
Abstract:
Molecular dynamics (MD) provides insights into atomic-scale processes by integrating over time the equations that describe the motion of atoms under the action of interatomic forces. Machine learning models have substantially accelerated MD by providing inexpensive predictions of the forces, but they remain constrained to minuscule time integration steps, which are required by the fast time scale of atomic motion. In this work, we propose FlashMD, a method to predict the evolution of positions and momenta over strides that are between one and two orders of magnitude longer than typical MD time steps. We incorporate considerations on the mathematical and physical properties of Hamiltonian dynamics in the architecture, generalize the approach to allow the simulation of any thermodynamic ensemble, and carefully assess the possible failure modes of a direct MD approach. We validate FlashMD’s accuracy in reproducing equilibrium and time‐dependent properties, using both system‐specific and general-purpose models, extending the ability of MD simulation to reach the long time scales needed to model microscopic processes of high scientific and technological relevance.



Paperid:1467
Authors:Zheyuan Liu, Munan Ning, Qihui Zhang, Shuo Yang, Zhongrui Wang, Yiwei Yang, Xianzhe Xu, Yibing Song, Weihua Chen, Fan Wang, Li Yuan
Abstract:
Current text-to-image (T2I) generation models struggle to align spatial composition with the input text, especially in complex scenes. Even layout-based approaches yield suboptimal spatial control, as their generation process is decoupled from layout planning, making it difficult to refine the layout during synthesis.We present CoT-Diff, a framework that brings step-by-step CoT-style reasoning into T2I generation by tightly integrating Multimodal Large Language Model (MLLM)-driven 3D layout planning with the diffusion process.CoT-Diff enables layout-aware reasoning inline within a single diffusion round: at each denoising step, the MLLM evaluates intermediate predictions, dynamically updates the 3D scene layout, and continuously guides the generation process. The updated layout is converted into semantic conditions and depth maps, which are fused into the diffusion model via a condition-aware attention mechanism, enabling precise spatial control and semantic injection. Experiments on 3D Scene benchmarks show that CoT-Diff significantly improves spatial alignment and compositional fidelity, and outperforms the state-of-the-art method by 34.7% in complex scene spatial accuracy, thereby validating the effectiveness of this entangled generation paradigm.



Paperid:1468
Authors:Jinsong Chen, Chenyang Li, Gaichao Li, John Hopcroft, Kun He
Abstract:
Node tokenized graph Transformers (GTs) have shown promising performance in node classification. The generation of token sequences is the key module in existing tokenized GTs which transforms the input graph into token sequences, facilitating the node representation learning via Transformer. In this paper, we observe that the generations of token sequences in existing GTs only focus on the first-order neighbors on the constructed similarity graphs, which leads to the limited usage of nodes to generate diverse token sequences, further restricting the potential of tokenized GTs for node classification. To this end, we propose a new method termed SwapGT. SwapGT first introduces a novel token swapping operation based on the characteristics of token sequences that fully leverages the semantic relevance of nodes to generate more informative token sequences. Then, SwapGT leverages a Transformer-based backbone to learn node representations from the generated token sequences. Moreover, SwapGT develops a center alignment loss to constrain the representation learning from multiple token sequences, further enhancing the model performance. Extensive empirical results on various datasets showcase the superiority of SwapGT for node classification.



Paperid:1469
Authors:wu tianci, Guangming Zhu, Lu jiang, Siyuan Wang, Ning Wang, Nuoye Xiong, Liang Zhang
Abstract:
Action recognition is a fundamental task in video understanding. Existing methods typically extract unified features to process all actions in one video, which makes it challenging to model the interactions between different objects in multi-action scenarios. To alleviate this issue, we explore disentangling any specified actions from complex scenes as an effective solution. In this paper, we propose Prompt-guided Disentangled Representation for Action Recognition (ProDA), a novel framework that disentangles any specified actions from a multi-action scene. ProDA leverages Spatio-temporal Scene Graphs (SSGs) and introduces Dynamic Prompt Module (DPM) to guide a Graph Parsing Neural Network (GPNN) in generating action-specific representations. Furthermore, we design a video-adapted GPNN that aggregates information using dynamic weights. Experiments in video action recognition demonstrate the effectiveness of our approach when compared with the state-of-the-art methods. Our code can be found in Supplementary material.



Paperid:1470
Authors:Yili Wang, Tairan Huang, Changlong He, Qiutong Li, Jianliang Gao
Abstract:
Heterogeneous temporal graphs (HTGs) are ubiquitous data structures in the real world. Recently, to enhance representation learning on HTGs, numerous attention-based neural networks have been proposed. Despite these successes, existing methods rely on a decoupled temporal and spatial learning paradigm, which weakens interactions of spatio-temporal information and leads to a high model complexity.To bridge this gap, we propose a novel learning paradigm for HTGs called Simple and Efficient Heterogeneous Temporal Graph Neural Network (SE-HTGNN). Specifically, we innovatively integrate temporal modeling into spatial learning via a novel dynamic attention mechanism, which retains attention information from historical graph snapshots to guide subsequent attention computation, thereby improving the overall discriminative representations learning of HTGs. Additionally, to comprehensively and adaptively understand HTGs, we leverage large language models to prompt SE-HTGNN, enabling the model to capture the implicit properties of node types as prior knowledge. Extensive experiments demonstrate that SE-HTGNN achieves up to 10× speed-up over the state-of-the-art and latest baseline while maintaining the best forecasting accuracy. Codes are available at \https://anonymous.4open.science/r/SE-HTGNN/.



Paperid:1471
Authors:Walid Durani, Collin Leiber, Khalid Durani, Claudia Plant, Christian Böhm
Abstract:
Anomaly detection is a crucial task in data mining, focusing on identifying data points that deviate significantly from the main patterns in the data. This paper introduces Anomaly Detection by an Ensemble of Random Pairs of Hyperspheres (ADERH), a new isolation-based technique leveraging two key observations: (i) anomalies are comparatively rare, and (ii) they typically deviate stronger from general patterns than normal data points. Drawing on a delta-separation argument, ADERH constructs an ensemble of hyperspheres built upon randomly paired data points to identify anomalies. To address inevitable overlaps between anomalous and normal regions in the feature space, ADERH integrates two complementary concepts: Pitch, which highlights points near hypersphere boundaries, and NDensity, which down-weights hyperspheres centered on sparse (and often anomalous) regions. By averaging these local, density-adjusted ``isolation'' indicators across many random subsets, ADERH yields robust anomaly scores that clearly separate normal from abnormal samples. Extensive experiments on diverse real-world datasets show that ADERH consistently outperforms state-of-the-art methods while maintaining linear runtime scalability and stable performance across varying hyperparameter settings.



Paperid:1472
Authors:Zhijian Zhou, Liuhua Peng, Xunye Tian, Feng Liu
Abstract:
Abstract:The relative similarity testing aims to determine which of the distributions, $P$ or $Q$, is closer to an anchor distribution $U$. Existing kernel-based approaches often test the relative similarity with a fixed kernel in a manually specified alternative hypothesis, e.g., $Q$ is closer to $U$ than $P$. Although kernel selection is known to be important to kernel-based testing methods, the manually specified hypothesis poses a significant challenge for kernel selection in relative similarity testing: Once the hypothesis is specified first, we can always find a kernel such that the hypothesis is rejected. This challenge makes relative similarity testing ill-defined when we want to select a good kernel after the hypothesis is specified. In this paper, we cope with this challenge via learning a proper hypothesis and a kernel simultaneously, instead of learning a kernel after manually specifying the hypothesis. We propose an anchor-based maximum discrepancy (AMD), which defines the relative similarity as the maximum discrepancy between the distances of $(U, P)$ and $(U, Q)$ in a space of deep kernels. Based on AMD, our testing incorporates two phases. In Phase I, we estimate the AMD over the deep kernel space and infer the potential hypothesis. In Phase II, we assess the statistical significance of the potential hypothesis, where we propose a unified testing framework to derive thresholds for tests over different possible hypotheses from Phase I. Lastly, we validate our method theoretically and demonstrate its effectiveness via extensive experiments on benchmark datasets.



Paperid:1473
Authors:Xiaohui Gao, Haoran Yang, Yue Cheng, Mengfei Zuo, Yiheng Liu, Peiyang Li, Xintao Hu
Abstract:
In recent years, the alignment between artificial neural network (ANN) embeddings and voxel-level responses in functional magnetic resonance imaging (fMRI) via neural encoding models has significantly advanced research on neural representation mechanisms and interpretability in the brain. However, these approaches remain limited in their ability to characterize the brain’s inherently nonlinear response properties. To this end, we propose a novel nonlinearity metric—Jacobian-based Nonlinearity Evaluation (JNE)—which quantifies the degree of nonlinearity through interpretability analysis of nonlinear neural encoding model. Simulation experiments demonstrated that JNE consistently achieved high accuracy in measuring nonlinearity across various activation functions and network architectures. Applying this metric to natural image fMRI data reveals that the early visual cortex exhibits low nonlinearity, whereas higher visual areas (e.g., EBA) show significantly enhanced nonlinear responses, which presents a layer-wise increasing trend from lower to higher cortical regions. Moreover, JNE uncovers input-dependent sample selectivity in neural responses, highlighting the critical role of high-level cortex in nonlinear visual processing. As the first interpretability metric for quantifying nonlinear responses, JNE provides a new tool for understanding brain processing mechanisms.



Paperid:1474
Authors:Yanming Wan, Jiaxing Wu, Marwa Abdulhai, Lior Shani, Natasha Jaques
Abstract:
Effective conversational agents must personalize their interactions to adapt to user preferences, personalities, and attributes across diverse domains like education and healthcare. Current methods like Reinforcement Learning from Human Feedback (RLHF), often prioritize helpfulness and safety but fall short in fostering truly empathetic, adaptive, and personalized dialogues. Existing personalization approaches typically rely on extensive user history, limiting their effectiveness for new or context-limited users. To address these limitations, we propose leveraging a user model to incorporate a curiosity-based intrinsic reward into multi-turn RLHF. This novel reward mechanism encourages the agent to actively infer user traits by optimizing conversations to improve its user model's accuracy. Consequently, the agent delivers more personalized interactions by learning more about the user. We demonstrate our method's effectiveness in two distinct domains: significantly improving personalization performance in a conversational recommendation task, and personalizing conversations for different learning styles in an educational setting with improved generalization capabilities compared to traditional multi-turn RLHF, all while maintaining conversation quality. Our method offers a promising solution for creating more personalized, adaptive, and engaging conversational agents.



Paperid:1475
Authors:Paula Cordero-Encinar, Andrew Duncan, Sebastian Reich, O. Deniz Akyildiz
Abstract:
We introduce a novel framework for efficient sampling from complex, unnormalised target distributions by exploiting multiscale dynamics. Traditional score-based sampling methods either rely on learned approximations of the score function or involve computationally expensive nested Markov chain Monte Carlo (MCMC) loops. In contrast, the proposed approach leverages stochastic averaging within a slow-fast system of stochastic differential equations (SDEs) to estimate intermediate scores along a diffusion path without training or inner-loop MCMC. Two algorithms are developed under this framework:MultALMC, which uses multiscale annealed Langevin dynamics, andMultCDiff, based on multiscale controlled diffusions for the reverse-time Ornstein-Uhlenbeck process. Both overdamped and underdamped variants are considered, with theoretical guarantees of convergence to the desired diffusion path. The framework is extended to handle heavy-tailed target distributions using Student’s t-based noise models and tailored fast-process dynamics. Empirical results across synthetic and real-world benchmarks, including multimodal and high-dimensional distributions, demonstrate that the proposed methods are competitive with existing samplers in terms of accuracy and efficiency, without the need for learned models.



Paperid:1476
Authors:Anyi Li, Jiacheng Cen, Songyou Li, Mingze Li, YANG YU, Wenbing Huang
Abstract:
Accurate prediction of ionic conductivity in electrolyte systems is crucial for advancing numerous scientific and technological applications. While significant progress has been made, current research faces two fundamental challenges: (1) the lack of high-quality standardized benchmarks, and (2) inadequate modeling of geometric structure and intermolecular interactions in mixture systems. To address these limitations, we first reorganize and enhance the CALiSol and DiffMix electrolyte datasets by incorporating geometric graph representations of molecules. We then propose GeoMix, a novel geometry-aware framework that preserves Set-SE(3) equivariance—an essential but challenging property for mixture systems. At the heart of GeoMix lies the Geometric Interaction Network (GIN), an equivariant module specifically designed for intermolecular geometric message passing. Comprehensive experiments demonstrate that GeoMix consistently outperforms diverse baselines (including MLPs, GNNs, and geometric GNNs) across both datasets, validating the importance of cross-molecular geometric interactions and equivariant message passing for accurate property prediction. This work not only establishes new benchmarks for electrolyte research but also provides a general geometric learning framework that advances modeling of mixture systems in energy materials, pharmaceutical development, and beyond.



Paperid:1477
Authors:Jiani Huang, Mayank Keoliya, Matthew Kuo, Neelay Velingker, Amish Sethi, JungHo Jung, Ziyang Li, Ser Nam Lim, Mayur Naik
Abstract:
Multi-modal large language models (MLLMs) are making rapid progress toward general-purpose embodied agents. However, current training pipelines primarily rely on high-level vision-sound-text pairs and lack fine-grained, structured alignment between pixel-level visual content and textual semantics. To overcome this challenge, we propose ESCA, a new framework for contextualizing embodied agents through structured spatial-temporal understanding. At its core is SGClip, a novel CLIP-based, open-domain, and promptable model for generating scene graphs. SGClip is trained on 87K+ open-domain videos via a neurosymbolic learning pipeline, which harnesses model-driven self-supervision from video-caption pairs and structured reasoning, thereby eliminating the need for human-labeled scene graph annotations. We demonstrate that SGClip supports both prompt-based inference and task-specific fine-tuning, excelling in scene graph generation and action localization benchmarks. ESCA with SGClip consistently improves both open-source and commercial MLLMs, achieving state-of-the-art performance across two embodied environments. Notably, it significantly reduces agent perception errors and enables open-source models to surpass proprietary baselines.



Paperid:1478
Authors:Parsa Rahimi, Damien Teney, Sébastien Marcel
Abstract:
The increasing reliance on large-scale datasets in machine learning poses significant privacy and ethical challenges, particularly in sensitive domains such as face recognition (FR). Synthetic data generation offers a promising alternative; however, most existing methods depend heavily on external datasets or pre-trained models, increasing complexity and resource demands. In this paper, we introduceAugGen, a self-contained synthetic augmentation technique. AugGen strategically samples from a class-conditional generative model trained exclusively on the target FR dataset, eliminating the need for external resources. Evaluated across 8 FR benchmarks, including IJB-C and IJB-B, our method achieves1–12% performance improvements, outperforming models trained solely on real data and surpassing state-of-the-art synthetic data generation approaches, while using less real data. Notably, these gains often exceed those from architectural modifications, underscoring the value of synthetic augmentation in data-limited scenarios. Our findings demonstrate that carefully integrated synthetic data can both mitigate privacy constraints and substantially enhance discriminative performance in face recognition. Code and datasets will be made publicly available upon publication.



Paperid:1479
Authors:Qipeng zhu, Jie Chen, Jian Pu, Junping Zhang
Abstract:
Graph Neural Networks (GNNs) are pivotal in graph classification but often struggle with generalization and overfitting. We introduce a unified and efficient Graph Multi-View (GMV) learning framework that integrates multi-view learning into GNNs to enhance robustness and efficiency. Leveraging the lottery ticket hypothesis, GMV activates diverse sub-networks within a single GNN through a novel training pipeline, which includes mixed-view generation, and multi-view decomposition and learning. This approach simultaneously broadens "views" from the data, model, and optimization perspectives during training to enhance the generalization capabilities of GNNs. During inference, GMV only incorporates additional prediction heads into standard GNNs, thereby achieving multi-view learning at minimal cost. Our experiments demonstrate that GMV surpasses other augmentation and ensemble techniques for GNNs and Graph Transformers across various graph classification scenarios.



Paperid:1480
Authors:Arjun Karuvally, Franz Nowak, Andy Keller, Carmen Amo Alonso, Terrence Sejnowski, Hava Siegelmann
Abstract:
Recent work has revealed that state space models (SSMs), while efficient for long-sequence processing, are fundamentally limited in their ability to represent formal languages—particularly due to time-invariant and real-valued recurrence structures. In this work, we draw inspiration from adaptive and structured dynamics observed in biological neural systems and introduce the Adaptive Unitary State Space Model (AUSSM): a novel class of SSMs that leverages skew-symmetric, input-dependent recurrence to achieve unitary evolution and high expressive power. Using algebraic automata theory, we prove that AUSSM can perform modulo counting and simulate solvable group automata at finite precision, enabling SSMs to model a broad class of regular languages out of reach for other SSM architectures. To overcome the practical inefficiencies of adaptive recurrence, we develop a separable convolution formulation and a CUDA implementation that enables scalable parallel training. Empirically, we show that AUSSM and its hybrid variant—interleaved with Mamba—outperform prior SSMs on formal algorithmic tasks such as parity and modular arithmetic, and achieve competent performance on real-world long time-series classification benchmarks. Our results demonstrate that adaptive unitary recurrence provides a powerful and efficient inductive bias for both symbolic and continuous sequence modeling.



Paperid:1481
Authors:Moayed Haji-Ali, Willi Menapace, Ivan Skorokhodov, Arpit Sahni, Sergey Tulyakov, Vicente Ordonez, Aliaksandr Siarohin
Abstract:
Generating high-dimensional visual modalities is a computationally intensive task. A common solution is progressive generation, where the outputs are synthesized in a coarse-to-fine spectral autoregressive manner. While diffusion models benefit from the coarse-to-fine nature of denoising, explicit multi-stage architectures are rarely adopted. These architectures have increased the complexity of the overall approach, introducing the need for a custom diffusion formulation, decomposition-dependent stage transitions, add-hoc samplers, or a model cascade. Our contribution, Decomposable Flow Matching (DFM), is a simple and effective framework for the progressive generation of visual media. DFM applies Flow Matching independently at each level of a user-defined multi-scale representation (such as Laplacian pyramid). As shown by our experiments, our approach improves visual quality for both images and videos, featuring superior results compared to prior multistage frameworks. On Imagenet-1k 512px, DFM achieves 35.2% improvements in FDD scores over the base architecture and 26.4% over the best-performing baseline, under the same training compute. When applied to finetuning of large models, such as FLUX, DFM shows faster convergence speed to the training distribution. Crucially, all these advantages are achieved with a single model, architectural simplicity, and minimal modifications to existing training pipelines.



Paperid:1482
Authors:Wenbin Lu, Yihan Chen, Junnan Xu, Wei Li, Junwei Zhu, Jianwei Zheng
Abstract:
Abstract:Neural operator is considered a popular data-driven alternative to traditional partial differential equation (PDE) solvers. However, most current solutions, whether fulfilling computations in frequency, Laplacian, and wavelet domains, all deviate far from the intrinsic PDE space. While with meticulous network architecture elaborated, the deviation often leads to biased accuracy. To address the issue, we open a new avenue that pioneers leveraging Radon transform to decompose the input space, finalizing a novel Radon neural operator (RNO) to solve PDEs in infinite-dimensional function space. Distinct from previous solutions, we project the input data into the sinogram domain, shrinking the multi-dimensional transformations to a reduced-dimensional counterpart and fitting compactly with the PDE space. Theoretically, we prove that RNO obeys a property of bilipschitz strongly monotonicity under diffeomorphism, providing deeper insights to guarantee the desired accuracy than typical discrete invariance or continuous-discrete equivalence. Within the sinogram domain, we further evidence that different angles contribute unequally to the overall space, thus engineering a reweighting technique to enable more effective PDE solutions. On that basis, a sinogram-domain convolutional layer is crafted, which operates on a fixed $\theta$-grid that is decoupled from the PDE space, further enjoying a natural guarantee of discrete invariance. Extensive experiments demonstrate that RNO sets new state-of-the-art (SOTA) scores across massive standard benchmarks, with superior generalization performance enjoyed. Codes are attached and will be released at GitHub.



Paperid:1483
Authors:Xiang Li, Rongrong Wang, Qing Qu
Abstract:
Classifier-free guidance (CFG) is a core technique powering state-of-the-art image generation systems, yet its underlying mechanisms remain poorly understood. In this work, we first analyze CFG in a simplified linear diffusion model, where we show its behavior closely resembles that observed in the nonlinear case. Our analysis reveals that linear CFG improves generation quality via three distinct components: (i) a mean-shift term that approximately steers samples in the direction of class means, (ii) a positive Contrastive Principal Components (CPC) term that amplifies class-specific features, and (iii) a negative CPC term that suppresses generic features prevalent in unconditional data. We then verify that these insights in real-world, nonlinear diffusion models: over a broad range of noise levels, linear CFG resembles the behavior of its nonlinear counterpart. Although the two eventually diverge at low noise levels, we discuss how the insights from the linear analysis still shed light on the CFG's mechanism within the nonlinear regime.



Paperid:1484
Authors:Haoran Sun, Yurong Chen, Siwei Wang, Chu Xu, Wei Chen, Xiaotie Deng
Abstract:
Fine-tuning large language models (LLMs) to aggregate multiple preferences has attracted considerable research attention. With aggregation algorithms advancing, a potential economic scenario arises where fine-tuning services are provided to agents with different preferences. In this context, agents may benefit from strategically misreporting their preferences, but this could harm the aggregation performance. This paper addresses such incentive issues by framing it as a mechanism design problem: an LLM provider determines the fine-tuning objective (training rule) and the pricing scheme (payment rule) for agents. We primarily focus on training rules that maximize social welfare subject to certain regularizations, referred to as SW-Max rules. First, we show that under most circumstances, truthful reporting is sub-optimal with simply a SW-Max rule, thereby highlighting the necessity of payments. Second, we extend the VCG payment to implement SW-Max rules in dominant-strategy incentive compatibility (DSIC). We characterize sufficient conditions for payment equivalence and derive the necessary conditions for a payment rule to implement a SW-Max rule in DSIC and other principles. Third, we demonstrate that our mechanism is approximately DSIC with perturbed input, showcasing its robustness against the inevitable errors in real-world applications. Experiments on real LLM training results further confirm the practical implications of our results.



Paperid:1485
Authors:Giordano Cicchetti, Eleonora Grassucci, Danilo Comminiello
Abstract:
Multimodal learning plays a pivotal role in advancing artificial intelligence systems by incorporating information from multiple modalities to build a more comprehensive representation. Despite its importance, current state-of-the-art models still suffer from severe limitations that prevent the successful development of a fully multimodal model. Such methods do not provide indicators that all the involved modalities are effectively aligned. As a result, a set of modalities may not be aligned, undermining the effectiveness of the model in downstream tasks where multiple modalities should provide additional information that the model fails to exploit. In this paper, we present TRIANGLE: TRI-modAl Neural Geometric LEarning, the novel proposed similarity measure that is directly computed in the higher-dimensional space spanned by the modality embeddings. TRIANGLE improves the joint alignment of three modalities via a triangle‑area similarity, avoiding additional fusion layers.When incorporated in contrastive losses replacing cosine similarity, TRIANGLE significantly boosts the performance of multimodal modeling, while yielding interpretable alignment rationales. Extensive evaluation in three-modal tasks such as video-text and audio-text retrieval or audio-video classification, demonstrates that TRIANGLE achieves state-of-the-art results across different datasets improving the performance of cosine-based methods up to 9 points of Recall@1.



Paperid:1486
Authors:Dengwei Zhao, Shikui Tu, Yanan Sun, Lei Xu
Abstract:
Abstract:In recent years, neural network-guided heuristic search algorithms, such as Monte-Carlo tree search and A$^*$ search, have achieved significant advancements across diverse practical applications. Due to the challenges stemming from high state-space complexity, sparse training datasets, and incomplete environmental modeling, heuristic estimations manifest uncontrolled inherent biases towards the actual expected evaluations, thereby compromising the decision-making quality of search algorithms. Sampling exploration enhanced A$^*$ (SeeA$^*$) was proposed to improve the efficiency of A$^*$ search by constructing an dynamic candidate subset through random sampling, from which the expanded node was selected. However, uniform sampling strategy utilized by SeeA$^*$ facilitates exploration exclusively through the injection of randomness, which completely neglects the heuristic knowledge relevant to open nodes. Moreover, the theoretical support of cluster sampling remains ambiguous. Despite the existence of potential biases, heuristic estimations still encapsulate certain valuable information. In this paper, epistemic exploratory A$^*$ search (KeeA$^*$) is proposed to integrate heuristic knowledge for calibrating the sampling process. We first theoretically demonstrate that SeeA$^*$ with cluster sampling outperforms uniform sampling due to the distribution-aware selection with higher variance. Building on this insight, cluster scouting and path-aware sampling are introduced in KeeA$^*$ to further exploit heuristic knowledge to increase the sampling mean and variance, respectively, thereby generating higher-quality extreme candidates and enhancing overall decision-making performance. Finally, empirical results on retrosynthetic planning and logic synthesis demonstrate superior performance of KeeA$^*$ compared to state-of-the-art heuristic search algorithms.



Paperid:1487
Authors:Gangwei Xu, Haotong Lin, Hongcheng Luo, Xianqi Wang, JINGFENG YAO, Lianghui Zhu, Yuechuan Pu, Cheng Chi_, Haiyang Sun, Bing Wang, Guang Chen, Hangjun Ye, Sida Peng, Xin Yang
Abstract:
This paper presentsPixel-Perfect Depth, a monocular depth estimation model based on pixel-space diffusion generation that produces high-quality, flying-pixel-free point clouds from estimated depth maps. Current generative depth estimation models fine-tune Stable Diffusion and achieve impressive performance. However, they require a VAE to compress depth maps into latent space, which inevitably introduces flying pixels at edges and details. Our model addresses this challenge by directly performing diffusion generation in the pixel space, avoiding VAE-induced artifacts. To tackle the resulting complexity of high-resolution generation, we introduce two novel designs: 1)Semantics-Guided Diffusion Transformers (DiT)that extracts high-level semantic representations from vision foundation models to guide the diffusion process, enabling accurate modeling of both global image structures and fine-grained details; and 2)Cascade DiT Designthat progressively increases the number of patches to further enhance efficiency and accuracy. Our model achieves the best performance among all published generative models across five benchmarks, and significantly outperforms all other models in edge-aware point cloud evaluation. Code will be released for reproducibility.



Paperid:1488
Authors:Zhi Wen Soi, Chaoyi Zhu, Fouad Abiad, Aditya Shankar, Jeroen Galjaard, Huijuan Wang, Lydia Chen
Abstract:
Abstract:Synthetic time series generated by diffusion models enable sharing privacy-sensitive datasets, such as patients' functional MRI records. Key criteria for synthetic data include high data utility and traceability to verify the data source. Recent watermarking methods embed in homogeneous latent spaces, but state-of-the-art time series generators operate in real space—making latent-based watermarking incompatible. This creates the challenge of watermarking directly in real space while handling feature heterogeneity and temporal dependencies. We propose TimeWak, the first watermarking algorithm for multivariate time series diffusion models. To handle temporal dependence and spatial heterogeneity, TimeWak embeds a temporal chained-hashing watermark directly within the real temporal-feature space. The other unique feature is the $\epsilon$-exact inversion, which addresses the non-uniform reconstruction error distribution across features from inverting the diffusion process to detect watermarks. We derive the error bound of inverting multivariate time series and further maintain high watermark detectability. We extensively evaluate TimeWak on its impact on synthetic data quality, watermark detectability, and robustness under various post-editing attacks, against 5 datasets and baselines of different temporal lengths. Our results show that TimeWak achieves improvements of 61.96% in context-FID score, and 8.44% in correlational scores against strongest state-of-the-art baseline, while remaining consistently detectable. Our code can be accessed at the anonymized repository linked below: https://anonymous.4open.science/r/TimeWak.



Paperid:1489
Authors:Zefan Cai, Wen Xiao, Hanshi Sun, cheng Luo, Yikai Zhang, Ke Wan, Yucheng LI, Yeyang Zhou, Li-Wen Chang, Jiuxiang Gu, Zhen Dong, Animashree Anandkumar, Abedelkadir Asi, Junjie Hu
Abstract:
Reasoning models have demonstrated impressive performance in self-reflection and chain-of-thought reasoning. However, they often produce excessively long outputs, leading to prohibitively large key-value (KV) caches during inference. While chain-of-thought inference significantly improves performance on complex reasoning tasks, it can also lead to reasoning failures when deployed with existing KV cache compression approaches. To address this, we propose Redundancy-aware KV Cache Compression for Reasoning models (R-KV), a novel method specifically targeting redundant tokens in reasoning models. Our method preserves nearly 100% of the full KV cache performance using only 10% of the KV cache, substantially outperforming existing KV cache baselines, which reach only 60% of the performance. Remarkably, R-KV even achieves 105% of full KV cache performance with 38% of the KV cache. This KV-cache reduction also leads to a 50% memory saving and a 2x speedup over standard chain-of-thought reasoning inference. Experimental results show that R-KV consistently outperforms existing KV cache compression baselines across two mathematical reasoning datasets.



Paperid:1490
Authors:Rim Assouel, Pietro Astolfi, Florian Bordes, Michal Drozdzal, Adriana Romero-Soriano
Abstract:
Recent advances in vision language models (VLM) have been driven by contrastive models such as CLIP, which learn to associate visual information with their corresponding text descriptions. However, these models have limitations in understanding complex compositional scenes involving multiple objects and their spatial relationships. To address these challenges, we propose a novel approach that diverges from commonly used strategies that rely on the design of finegrained hard-negative augmentations. Instead, our work focuses on integrating inductive biases into the pretraining of CLIP-like models to improve their compositional understanding. To that end, we introduce a binding module that connects a scene graph, derived from a text description, with a slot-structured image representation, facilitating a structured similarity assessment between the two modalities. We also leverage relationships as text-conditioned visual constraints, thereby capturing the intricate interactions between objects and their contextual relationships more effectively. Our resulting model not only enhances the performance of CLIP-based models in multi-object compositional understanding but also paves the way towards more accurate and sample-efficient image-text matching of complex scenes.



Paperid:1491
Authors:Dian Wang, Boce Hu, Shuran Song, Robin Walters, Robert Platt
Abstract:
Recently, equivariant neural networks for policy learning have shown promising improvements in sample efficiency and generalization, however, their wide adoption faces substantial barriers due to implementation complexity. Equivariant architectures typically require specialized mathematical formulations and custom network design, posing significant challenges when integrating with modern policy frameworks like diffusion-based models. In this paper, we explore a number of straightforward and practical approaches to incorporate symmetry benefits into diffusion policies without the overhead of full equivariant designs. Specifically, we investigate (i) invariant representations via relative trajectory actions and eye-in-hand perception, (ii) integrating equivariant vision encoders, and (iii) symmetric feature extraction with pretrained encoders using Frame Averaging. We first prove that combining eye-in-hand perception with relative or delta action parameterization yields inherent SE(3)-invariance, thus improving policy generalization. We then perform a systematic experimental study on those design choices for integrating symmetry in diffusion policies, and conclude that an invariant representation with equivariant feature extraction significantly improves the policy performance. Our method achieves performance on par with or exceeding fully equivariant architectures while greatly simplifying implementation.



Paperid:1492
Authors:Xianquan Wang, Zhaocheng Du, Huibo Xu, Shukang Yin, Yupeng Han, Jieming Zhu, Kai Zhang, Qi Liu
Abstract:
With the rapid progress of large language models (LLMs) and diffusion models, there has been growing interest in personalized content generation. However, current conversational systems often present the same recommended content to all users, falling into the dilemma of "one-size-fits-all." To break this limitation and boost user engagement, in this paper, we introduce PiViC (PersonalizedVisualContent) generation, a unified framework for personalizing item images within conversational systems. We tackle two key bottlenecks: the depth of personalization and the fidelity of generated images. Speficially, an LLM-poweredInclinations Analyzeris adopted to capture user likes and dislikes from context to construct personalized prompts. Moreover, we design a dual-stage LoRA mechanism—Global LoRAfor understanding task-specific visual style, andLocal LoRAfor capturing preferred visual elements from conversation history. During training, we introduce the visual content condition method to ensure LoRA learns both historical visual context and maintains fidelity to the original item images.Extensive experiments on benchmark conversational datasets—including objective metrics and GPT-based evaluations—demonstrate that our framework significantly outperforms strong baselines, which highlight its potential to redefine personalization in visual content generation for e-commerce.The code is available.



Paperid:1493
Authors:Maria Taktasheva, Lily Goli, Alessandro Fiorini, Zhen Li, Daniel Rebain, Andrea Tagliasacchi
Abstract:
Recent advances in radiance fields and novel view synthesis enable creation of realistic digital twins from photographs. However, current methods struggle with flat, texture-less surfaces, creating uneven and semi-transparent reconstructions, due to an ill-conditioned photometric reconstruction objective. Surface reconstruction methods solve this issue but sacrifice visual quality. We propose a novel hybrid 2D/3D representation that jointly optimizes constrained planar (2D) Gaussians for modeling flat surfaces and freeform (3D) Gaussians for the rest of the scene. Our end-to-end approach dynamically detects and refines planar regions, improving both visual fidelity and geometric accuracy. It achieves state-of-the-art depth estimation on ScanNet++ and ScanNetv2, and excels at mesh extraction without overfitting to a specific camera model, showing its effectiveness in producing high-quality reconstruction of indoor scenes.



Paperid:1494
Authors:Xinrui Chen, Haoli Bai, Tao Yuan, ruikang liu, Kang Zhao, Xianzhi Yu, Lu Hou, Tian Guan, Yonghong He, Chun Yuan
Abstract:
Layer pruning has become a popular technique for compressing large language models (LLMs) due to its simplicity. However, existing layer pruning methods often suffer from significant performance drops. We identify that \textit{this degradation stems from the mismatch of activation magnitudes across layers and tokens at the pruning interface}. To address this, we propose \textsc{LinearPatch}, a simple plug-and-play technique to revive the layer-pruned LLMs. The proposed method adopts Hadamard transformation to suppress massive outliers in particular tokens, and channel-wise scaling to align the activation magnitudes. These operations can be fused into a single matrix, which functions as a patch to bridge the pruning interface with negligible inference overhead. \textsc{LinearPatch} retains up to \textbf{94.15\%} performance of the original model when pruning 5 layers of LLaMA-3-8B on the question answering benchmark, surpassing existing state-of-the-art methods by \textbf{4\%}. In addition, the patch matrix can be further optimized with memory efficient offline knowledge distillation. With only 5K samples, the retained performance of \textsc{LinearPatch} can be further boosted to \textbf{95.16\%} within 30 minutes on a single computing card.



Paperid:1495
Authors:Jiaxuan Gao, Shu Yan, Qixin Tan, lu Yang, Shusheng Xu, Wei Fu, Zhiyu Mei, Kaifeng Lyu, YI WU
Abstract:
Large Reasoning Models (LRMs) demonstrate remarkable problem-solving capabilities through extended Chain-of-Thought (CoT) reasoning but often produce excessively verbose and redundant reasoning traces. This inefficiency incurs high inference costs and limits practical deployment. While existing fine-tuning methods aim to improve reasoning efficiency, assessing their efficiency gains remains challenging due to inconsistent evaluations. In this work, we introduce thereasoning efficiency frontiers, empirical upper bounds derived from fine-tuning a base LRM (DeepSeek-R1-Distill-Qwen-1.5B/7B) across diverse approaches and training configurations. Based on these frontiers, we propose theReasoning Efficiency Gap (REG), a unified metric quantifying deviations of any fine-tuned LRMs from these frontiers. Systematic evaluation on challenging mathematical benchmarks, AMC23, AIME24, and AIME25, reveals significant gaps in current methods: they either sacrifice accuracy for short length or use excessive tokens to achieve sub-optimal accuracies despite high overall accuracy. To reduce the efficiency gap, we proposeREO-RL, a Reinforcement Learning algorithm that optimizes reasoning efficiency by targeting a sparse set of token budgets. Leveraging numerical integration over strategically selected budgets, REO-RL approximates the full efficiency objective with low error using a small set of token budgets. Experiments show that, compared to vanilla RL with outcome reward, REO-RL reduces the reasoning efficiency gap by 74.5\% and 64.2\% in the 1.5B and 7B settings. The 7B LRM fine-tuned with REO-RL achieves reasoning conciseness surpassing frontier LRMs like Qwen3 and Claude Sonnet 3.7. Ablation studies confirm the efficacy of our token budget strategy and highlight REO-RL’s flexibility across design choices. This work establishes a systematic framework for evaluating and optimizing reasoning efficiency in LRMs. We will release the related code, data, and models to support future research on efficient reasoning in LRMs.



Paperid:1496
Authors:Arthur Bizzi, Matias Grynberg Portnoy, Vitor Pereira Matias, Daniel Perazzo, João Paulo Silva do Monte Lima, Luiz Velho, Nuno Gonçalves, João Pereira, Guilherme Schardong, Tiago Novello
Abstract:
Morphing is a long-standing problem in vision and computer graphics, requiring a time-dependent warping for feature alignment and a blending for smooth interpolation. Recently, multilayer perceptrons (MLPs) have been explored as implicit neural representations (INRs) for modeling such deformations, due to their meshlessness and differentiability; however, extracting coherent and accurate morphings from standard MLPs typically relies on costly regularizations, often leading to unstable training and impeding the effective alignment and interpolation between features. To overcome these limitations, we propose FLOWING (FLOW morphING), a framework that reframes warping as the construction of a differential vector flow, naturally ensuring continuity, invertibility, and temporal coherence.By design, FLOWING encodes structural flow propertiesdirectly into the network architectures, avoiding costly regularizations. This flow-centric approach yields principled and stable transformations that are smooth, reversible, and temporally coherent by construction, enabling accurate, structure-preserving morphing of both 2D images and 3D shapes.Extensive experiments across a range of applications—including face and image morphing, as well as Gaussian Splatting morphing—show that FLOWING achieves state-of-the-art morphing quality with substantially faster convergence. Code and pretrained models will be released.



Paperid:1497
Authors:Hu Yu, Biao Gong, Hangjie Yuan, DanDan Zheng, Weilong Chai, Jingdong Chen, Kecheng Zheng, Feng Zhao
Abstract:
Abstract:Masked-based autoregressive models have demonstrated promising image generation capability in continuous space. However, their potential for video generation remains under-explored. Masked-based autoregressive models have demonstrated promising image generation capability in continuous space. However, their potential for video generation remains under-explored. In this paper, we propose \textbf{VideoMAR}, a concise and efficient decoder-only autoregressive image-to-video model with continuous tokens, composing temporal frame-by-frame and spatial masked generation.We first identify temporal causality and spatial bi-directionality as the first principle of video AR models, and propose the next-frame diffusion loss for the integration of mask and video generation. Besides, the huge cost and difficulty of long sequence autoregressive modeling is a basic but crucial issue. To this end, we propose the temporal short-to-long curriculum learning and spatial progressive resolution training, and employ progressive temperature strategy at inference time to mitigate the accumulation error. Furthermore, VideoMAR replicates several unique capacities of language models to video generation. It inherently bears high efficiency due to simultaneous temporal-wise KV cache and spatial-wise parallel generation, and presents the capacity of spatial and temporal extrapolation via 3D rotary embeddings. On the VBench-I2V benchmark, VideoMAR surpasses the previous state-of-the-art (Cosmos I2V) while requiring significantly fewer parameters ($9.3\%$), training data ($0.5\%$), and GPU resources ($0.2\%$).



Paperid:1498
Authors:Boyang Wang, Xuweiyi Chen, Matheus Gadelha, Zezhou Cheng
Abstract:
Controllability, temporal coherence, and detail synthesis remain the most critical challenges in video generation. In this paper, we focus on a commonly used yet underexplored cinematic technique known as Frame In and Frame Out. Specifically, starting from image-to-video generation, users can control the objects in the image to naturally leave the scene or provide breaking new identity references to enter the scene, guided by user-specified motion trajectory. To support this task, we introduce a new dataset curated semi-automatically, a comprehensive evaluation protocol targeting this setting, and an efficient identity-preserving motion-controllable video Diffusion Transformer architecture. Our evaluation shows that our proposed approach significantly outperforms existing baselines.



Paperid:1499
Authors:Zhishuai Liu, Pan Xu
Abstract:
Abstract:Many real-world decision-making problems face the off-dynamics challenge: the agent learns a policy in a source domain and deploys it in a target domain with different state transitions. The distributionally robust Markov decision process (DRMDP) addresses this challenge by finding a robust policy that performs well under the worst-case environment within a pre-specified uncertainty set of transition dynamics. Its effectiveness heavily hinges on the proper design of these uncertainty sets, based on prior knowledge of the dynamics. In this work, we propose a novel linear mixture DRMDP framework, where the nominal dynamics is assumed to be a linear mixture model. In contrast with existing uncertainty sets directly defined as a ball centered around the nominal kernel, linear mixture DRMDPs define the uncertainty sets based on a ball around the mixture weighting parameter. We show that this new framework provides a more refined representation of uncertainties compared to conventional models based on $(s,a)$-rectangularity and $d$-rectangularity, when prior knowledge about the mixture model is present. We propose a meta algorithm for robust policy learning in linear mixture DRMDPs with general $f$-divergence defined uncertainty sets, and analyze its sample complexities under three divergence metrics instantiations: total variation, Kullback-Leibler, and $\chi^2$ divergences. These results establish the statistical learnability of linear mixture DRMDPs, laying the theoretical foundation for future research on this new setting.



Paperid:1500
Authors:Anish Hebbar, Rong Ge, Amit Kumar, Debmalya Panigrahi
Abstract:
Abstract:The field of learning-augmented algorithms seeks to use ML techniques on past instances of a problem to inform an algorithm designed for a future instance. In this paper, we introduce a novel model for learning-augmented algorithms inspired by online learning. In this model, we are given a sequence of instances of a problem and the goal of the learning-augmented algorithm is to use prior instances to propose a solution to a future instance of the problem. The performance of the algorithm is measured by its average performance across all the instances, where the performance on a single instance is the ratio between the cost of the algorithm's solution and that of an optimal solution for that instance. We apply this framework to the classic $k$-median clustering problem, and give an efficient learning algorithm that can approximately match the average performance of the best fixed $k$-median solution in hindsight across all the instances. We also experimentally evaluate our algorithm and show that its empirical performance is close to optimal, and also that it automatically adapts the solution to a dynamically changing sequence.



Paperid:1501
Authors:Xuming He, Zhiyuan You, Junchao Gong, Couhua Liu, Xiaoyu Yue, Peiqin Zhuang, zhangwenlong, LEI BAI
Abstract:
Quality analysis of weather forecasts is an essential topic in meteorology. Although traditional score-based evaluation metrics can quantify certain forecast errors, they are still far from meteorological experts in terms of descriptive capability, interpretability, and understanding of dynamic evolution. With the rapid development of Multi-modal Large Language Models (MLLMs), these models become potential tools to overcome the above challenges. In this work, we introduce an MLLM-based weather forecast analysis method, RadarQA, integrating key physical attributes with detailed assessment reports. We introduce a novel and comprehensive task paradigm for multi-modal quality analysis, encompassing both single frame and sequence, under both rating and assessment scenarios. To support training and benchmarking, we design a hybrid annotation pipeline that combines human expert labeling with automated heuristics. With such an annotation method, we construct RQA-70K, a large-scale dataset with varying difficulty levels for radar forecast quality evaluation. We further design a multi-stage training strategy that iteratively improves model performance at each stage. Extensive experiments show that RadarQA outperforms existing general MLLMs across all evaluation settings, highlighting its potential for advancing quality analysis in weather prediction.



Paperid:1502
Authors:Wang Bill Zhu, Deqing Fu, Kai Sun, Yi Lu, Zhaojiang Lin, Seungwhan Moon, Kanika Narang, MUSTAFA CANIM, Yue Liu, Anuj Kumar, Xin Dong
Abstract:
Existing recommendation systems either rely on user interaction logs, such as online shopping history for shopping recommendations, or focus on text signals. However, item-based histories are not always accessible and generalizable for multimodal recommendation.We hypothesize that a user's visual history --- comprising images from daily life --- can offer rich, task-agnostic insights into their interests and preferences, and thus be leveraged for effective personalization.To this end, we propose VisualLens, a novel framework that leverages multimodal large language models (MLLMs) to enable personalization using task-agnostic visual history.VisualLens extracts, filters, and refines a spectrum user profile from the visual history to support personalized recommendation.We created two new benchmarks, Google-Review-V and Yelp-V, with task-agnostic visual histories, and show that VisualLens improves over state-of-the-art item-based multimodal recommendations by 5-10\% on Hit@3, and outperforms GPT-4o by 2-5\%.Further analysis shows that VisualLens is robust across varying history lengths and excels at adapting to both longer histories and unseen content categories.



Paperid:1503
Authors:JUE GONG, Tingyu Yang, Jingkai Wang, Zheng Chen, Xing Liu, Hong Gu, Yulun Zhang, Xiaokang Yang
Abstract:
Human-centered images often suffer from severe generic degradation during transmission and are prone to human motion blur (HMB), making restoration challenging. Existing research lacks sufficient focus on these issues, as both problems often coexist in practice. To address this, we design a degradation pipeline that simulates the coexistence of HMB and generic noise, generating synthetic degraded data to train our proposed HAODiff, a human-aware one-step diffusion. Specifically, we propose a triple-branch dual-prompt guidance (DPG), which leverages high-quality images, residual noise (LQ minus HQ), and HMB segmentation masks as training targets. It produces a positive–negative prompt pair for classifier‑free guidance (CFG) in a single diffusion step. The resulting adaptive dual prompts let HAODiff exploit CFG more effectively, boosting robustness against diverse degradations. For fair evaluation, we introduce MPII‑Test, a benchmark rich in combined noise and HMB cases. Extensive experiments show that our HAODiff surpasses existing state-of-the-art (SOTA) methods in terms of both quantitative metrics and visual quality on synthetic and real-world datasets, including our introduced MPII-Test. The code and model will be released soon.



Paperid:1504
Authors:Rishwanth Raghu, Axel Levy, Gordon Wetzstein, Ellen Zhong
Abstract:
Trained on decades of accumulated high-resolution atomic models, protein structure prediction models are now capable of generating accurate 3D structural hypotheses from sequence alone. However, they routinely fail to capture the conformational diversity of dynamic biomolecular complexes, often requiring heuristic MSA subsampling approaches for generating alternative states. In parallel, cryo-electron microscopy (cryo-EM) has emerged as a powerful tool for imaging near-native structural heterogeneity, but is challenged by arduous pipelines to go from raw experimental data to atomic models. Here, we bridge the gap between these modalities, combining cryo-EM density maps with the rich sequence and biophysical priors learned by protein structure prediction models. Our method, CryoBoltz, guides the sampling trajectory of a pretrained protein structure prediction model using both global and local structural constraints derived from density maps, driving predictions towards conformational states consistent with the experimental data. We demonstrate that this flexible yet powerful inference-time approach allows us to build atomic models into heterogeneous cryo-EM maps across a variety of dynamic biomolecular systems including transporters and antibodies.



Paperid:1505
Authors:Sounak Banerjee, Daphne Cornelisse, Deepak Gopinath, Emily Sumner, Jonathan DeCastro, Guy Rosman, Eugene Vinitsky, Mark Ho
Abstract:
People's goal-directed actions are influenced by their cognitive biases, and autonomous systems that interact with people should be aware of this. For example, because attention is always limited, people's attention to objects in their environment will be biased in a way that systematically affects their actions when performing everyday tasks (e.g., driving to work, making breakfast). Here, building on recent work in computational cognitive science, we formally articulate the attention-aware inverse planning problem, in which the goal is to estimate, from a person's behavior, what attentional biases influence their behavior on a task. We then demonstrate how attention-aware inverse planning systematically differs from standard inverse reinforcement learning and how cognitive biases can be inferred from behavior. Finally, we present a scalable algorithmic approach to attention-aware inverse planning that combines deep reinforcement learning with computational cognitive modeling. We use this approach to infer the attentional strategies of agents in real-life driving scenarios selected from the Waymo Open Dataset, demonstrating the scalability of estimating cognitive biases with attention-aware inverse planning.



Paperid:1506
Authors:Zhiying Jiang, Ruhao Yan, Zengxi Zhang, Bowei Zhang, Jinyuan Liu
Abstract:
Image stitching synthesizes images captured from multiple perspectives into a single image with a broader field of view. The significant variations in object depth often lead to large parallax, resulting in ghosting and misalignment in the stitched results. To address this, we propose a depth-consistency-constrained seamless-free image stitching method. First, to tackle the multi-view alignment difficulties caused by parallax, a multi-stage mechanism combined with global depth regularization constraints is developed to enhance the alignment accuracy of the same apparent target across different depth ranges. Second, during the multi-view image fusion process, an optimal stitching seam is determined through graph-based low-cost computation, and a soft-seam region is diffused to precisely locate transition areas, thereby effectively mitigating alignment errors induced by parallax and achieving natural and seamless stitching results. Furthermore, considering the computational overhead in the shift regression process, a reparameterization strategy is incorporated to optimize the structural design, significantly improving algorithm efficiency while maintaining optimal performance. Extensive experiments demonstrate the superior performance of the proposed method against the existing methods.



Paperid:1507
Authors:Huanjian Zhou, Masashi Sugiyama
Abstract:
Abstract:Non-log-concave sampling from an unnormalized density is fundamental in machine learning and statistics. As datasets grow larger, computationalefficiency becomes increasingly important, particularly in reducing adaptive complexity, namely the number of sequential rounds required for sampling algorithms. In this work, we initiate the study of the adaptive complexity of non-log-concave sampling within the framework of relative Fisher information introduced by Balasubramanian et al. in 2022. To obtain a relative fisher information of at most $\varepsilon^2$ from the target distribution, we propose a novel algorithm that reduces the adaptive complexity from $\mathcal{O}(d^2/\varepsilon^4)$ to $\mathcal{O}(d/\varepsilon^2)$ by leveraging parallelism. Furthermore, we show our algorithm is optimal for a specific regime of large $\varepsilon$. Our algorithm builds on a diagonally parallelized Picard iteration, while the lower bound is based on a reduction from the problem of finding stationary points.



Paperid:1508
Authors:Luke Guerdan, Solon Barocas, Kenneth Holstein, Hanna Wallach, Steven Wu, Alex Chouldechova
Abstract:
The LLM-as-a-judge paradigm, in which a judge LLM system replaces human raters in rating the outputs of other generative AI (GenAI) systems, plays a critical role in scaling and standardizing GenAI evaluations. To validate such judge systems, evaluators assess human--judge agreement by first collecting multiple human ratings for each item in a validation corpus, then aggregating the ratings into a single, per-item gold label rating. In many cases, however, items or rating criteria may be vague or ambiguous, or admit multiple interpretations, so a given rater (human or LLM) may deem multiple answers "reasonable" or "correct." We call such rating tasks indeterminate. Problematically, many indeterminate rating tasks rely on forced-choice elicitation, whereby raters are able to select only one answer option for each item. In this paper, we introduce a framework for LLM-as-a-judge validation for indeterminate rating tasks. We draw theoretical connections between different measures of judge system performance under different human--judge agreement metrics, and different rating elicitation and aggregation schemes. We demonstrate that differences in how humans and LLMs resolve indeterminacy in forced-choice rating tasks heavily bias LLM-as-a-judge validations. Through extensive experiments involving 11 real-world rating tasks and 8 commercial LLMs, we show that standard validation approaches select judge systems that are highly suboptimal, performing as much as 30\% worse than judge systems selected by alternatives we propose. We conclude with concrete recommendations for more principled and reliable approaches to LLM-as-a-judge validation.



Paperid:1509
Authors:Kaiwen He, Petros Drineas, Rajiv Khanna
Abstract:
Abstract:Spectral clustering is a fundamental method for graph partitioning, but its reliance on eigenvector computation limits scalability to massive graphs. Classical sparsification methods preserve spectral properties by sampling edges proportionally to their effective resistances, but require expensive preprocessing to estimate these resistances. We study whether uniform edge sampling—a simple, structure-agnostic strategy—can suffice for spectral clustering. Our main result shows that for graphs admitting a well-separated $k$-clustering, characterized by a large structure ratio $\Upsilon(k) = \lambda_{k+1} / \rho_G(k)$, uniform sampling preserves the spectral subspace used for clustering. Specifically, we prove that uniformly sampling $O(\kappa^2 n \log n / \varepsilon^2)$ edges, where $\kappa$ is the Laplacian condition number, yields a sparsifier whose top $(n-k)$-dimensional eigenspace is approximately orthogonal to the cluster indicators. This ensures that the spectral embedding remains faithful, and clustering quality is preserved.Our analysis introduces new resistacne bounds for intra-cluster edges, a rank-$(n-k)$ effective resistance formulation, and a matrix Chernoff bound adapted to the dominant eigenspace. These tools allow us to bypass importance sampling entirely. Conceptually, our result connects recent coreset-based clustering theory to spectral sparsification, showing that under strong clusterability, even uniform sampling is structure-aware. This provides the first provable guarantee that uniform edge sampling suffices for structure-preserving spectral clustering.



Paperid:1510
Authors:Lai Wei, Yuting Li, Chen Wang, Yue Wang, Linghe Kong, Weiran Huang, Lichao Sun
Abstract:
Abstract:Improving Multi-modal Large Language Models (MLLMs) in the post-training stage typically relies on supervised fine-tuning (SFT) or reinforcement learning (RL). However, these supervised methods require expensive and manually annotated multi-modal data--an ultimately unsustainable resource. While recent efforts have explored unsupervised post-training, their methods are complex and difficult to iterate. In this work, we are the first to investigate the use of GRPO, a stable and scalable online RL algorithm, for enabling continual self-improvement without any external supervision. We propose MM-UPT, a simple yet effective framework for unsupervised post-training of MLLMs. MM-UPT builds upon GRPO, replacing traditional reward signals with a self-rewarding mechanism based on majority voting over multiple sampled responses. Our experiments demonstrate that MM-UPT significantly improves the reasoning ability of Qwen2.5-VL-7B (e.g., 66.3\%$\rightarrow$72.9\% on MathVista, 62.9\%$\rightarrow$68.7\% on We-Math), using standard dataset without ground truth labels. MM-UPT also outperforms prior unsupervised baselines and even approaches the results of supervised GRPO. Furthermore, we show that incorporating synthetic questions, generated solely by MLLM itself, can boost performance as well, highlighting a promising approach for scalable self-improvement. Overall, MM-UPT offers a new paradigm for continual, autonomous enhancement of MLLMs in the absence of external supervision.



Paperid:1511
Authors:Jakub Białek, Wojtek Kuberski, Juhani Kivimäki, Nikolaos Perrakis
Abstract:
After deployment, machine learning models often experience performance degradation due to shifts in data distribution. It is challenging to assess post-deployment performance accurately when labels are missing or delayed. Existing proxy methods, such as data drift detection, fail to measure the effects of these shifts adequately. To address this, we introduce a new method for evaluating binary classification models on unlabeled tabular data that accurately estimates model performance under covariate shift and call it Probabilistic Adaptive Performance Estimation (PAPE). It can be applied to any performance metric defined with elements of the confusion matrix. Crucially, PAPE operates independently of the original model, relying only on its predictions and probability estimates, and does not need any assumptions about the nature of covariate shift, learning directly from data instead. We tested PAPE using over 900 dataset-model combinations from US census data, assessing its performance against several benchmarks through various metrics. Our findings show that PAPE outperforms other methodologies, making it a superior choice for estimating the performance of binary classification models.



Paperid:1512
Authors:Yuqi Jia, Minghong Fang, Neil Gong
Abstract:
Decentralized federated learning (DFL) enables clients (e.g., hospitals and banks) to jointly train machine learning models without a central orchestration server. In each global training round, each client trains a local model on its own training data and then they exchange local models for aggregation. In this work, we propose SelfishAttack, a new family of attacks to DFL. In SelfishAttack, a set of selfish clients aim to achieve competitive advantages over the remaining non-selfish ones, i.e., the final learnt local models of the selfish clients are more accurate than those of the non-selfish ones. Towards this goal, the selfish clients send carefully crafted local models to each remaining non-selfish one in each global training round. We formulate finding such local models as an optimization problem and propose methods to solve it when DFL uses different aggregation rules. Theoretically, we show that our methods find the optimal solutions to the optimization problem. Empirically, we show that SelfishAttack successfully increases the accuracy gap (i.e., competitive advantage) between the final learnt local models of selfish clients and those of non-selfish ones. Moreover, SelfishAttack achieves larger accuracy gaps than poisoning attacks when extended to increase competitive advantages.



Paperid:1513
Authors:Pujith Kachana, Samuel Li, Prajwal Chidananda, Saurabh Nair, Yasutaka Furukawa, Matthew A Brown
Abstract:
Estimating agent pose and 3D scene structure from multi-camera rigs is a central task in embodied AI applications such as autonomous driving. Recent learned approaches such as DUSt3R have shown impressive performance in multiview settings. However, these models treat images as unstructured collections, limiting effectiveness in scenarios where frames are captured from synchronized rigs with known or inferable structure. To this end, we introduce Rig3R, a generalization of prior multiview reconstruction models that incorporates rig structure when available, and learns to infer it when not. Rig3R conditions on optional rig metadata including camera ID, time, and rig poses to develop a rig-aware latent space that remains robust to missing information. It jointly predicts pointmaps and two types of raymaps: a pose raymap relative to a global frame, and a rig raymap relative to a rig-centric frame consistent across time. Rig raymaps allow the model to infer rig structure directly from input images when metadata is missing. Rig3R achieves state-of-the-art performance in 3D reconstruction, camera pose estimation, and rig discovery -- outperforming both traditional and learned methods by 17-45% mAA across diverse real-world rig datasets, all in a single forward pass without post-processing or iterative refinement.



Paperid:1514
Authors:Wenzhuo Tang, Haitao Mao, Danial Dervovic, Ivan Brugere, Saumitra Mishra, Yuying Xie, Jiliang Tang
Abstract:
Models for natural language and images benefit from data scaling behavior: the more data fed into the model, the better they perform. This 'better with more' phenomenon enables the effectiveness of large-scale pre-training on vast amounts of data. However, current graph pre-training methods struggle to scale up data due to heterogeneity across graphs. To achieve effective data scaling, we aim to develop a general model that is able to capture diverse data patterns of graphs and can be utilized to adaptively help the downstream tasks. To this end, we propose UniAug, a universal graph structure augmentor built on a diffusion model. We first pre-train a discrete diffusion model on thousands of graphs across domains to learn the graph structural patterns. In the downstream phase, we provide adaptive enhancement by conducting graph structure augmentation with the help of the pre-trained diffusion model via guided generation. By leveraging the pre-trained diffusion model for structure augmentation, we consistently achieve performance improvements across various downstream tasks in a plug-and-play manner. To the best of our knowledge, this study represents the first demonstration of a data-scaling graph structure augmentor on graphs across domains.



Paperid:1515
Authors:Ruiyang Zhou, Shuozhe Li, Amy Zhang, Liu Leqi
Abstract:
Recent advances in large language models have been driven by reinforcement learning-style post-training, which improves reasoning by optimizing model outputs based on reward or preference signals. These methods require positive training samples, which are difficult to obtain: expert demonstrations are costly, and self-generated positive responses are rare during early-stage training, particularly for challenging tasks.In this work, we identify two key properties of effective positive samples: they should have high probability under the current policy and increase the model’s likelihood of predicting the correct answer.Building on these insights, we propose ExPO (Self-Explanation Policy Optimization), a simple and modular framework that generates such positive samples from the model itself. ExPO improves both learning efficiency and final performance on reasoning benchmarks, even outperforming methods relying on expert demonstrations in challenging settings where the model initially struggles (e.g., MATH level-5).



Paperid:1516
Authors:Junjie Chen, Zeyu Luo, Wenhui Jiang, Zezheng Liu, Li Niu, Yuming Fang
Abstract:
Pose estimation is a fundamental task in computer vision, but generally requires large-scale annotated data for training. Few-shot and unsupervised pose estimation are prevalent economical paradigms, but the former still requires annotations for extensive novel classes while the latter only supports for single class. In this paper, we focus on the task of weak-shot pose estimation, where multiple novel classes are learned from unlabeled images with the help of labeled base classes. The key problem is what to transfer from base classes to novel classes, and we propose to transfer keyness and correspondence, which essentially belong to comparing entities and thus are class-agnostic and class-wise transferable. The keyness compares which pixel in the local region is more key, which can guide the keypoints of novel classes to move towards the local maximum (i.e., obtaining keypoints). The correspondence compares whether the two pixels belongs to the same semantic part, which can activate the keypoints of novel classes by reinforcing the consistency between corresponding points on two paired images. By transferring keyness and correspondence, our framework achieves favourable performance for weak-shot pose estimation. Extensive experiments and analyses on large-scale benchmark MP-100 demonstrate our effectiveness.



Paperid:1517
Authors:Paolo Gajo, Domenic Rosati, Hassan Sajjad, Alberto Barrón-Cedeño
Abstract:
Dependency parsing is the task of inferring natural language structure, often approached by modeling word interactions via attention through biaffine scoring. This mechanism works like self-attention in Transformers, where scores are calculated for every pair of words in a sentence. However, unlike Transformer attention, biaffine scoring does not use normalization prior to taking the softmax of the scores. In this paper, we provide theoretical evidence and empirical results revealing that a lack of normalization necessarily results in overparameterized parser models, where the extra parameters compensate for the sharp softmax outputs produced by high variance inputs to the biaffine scoring function. We argue that biaffine scoring can be made substantially more efficient by performing score normalization. We conduct experiments on six datasets for semantic and syntactic dependency parsing using a one-hop parser. We train N-layer stacked BiLSTMs and evaluate the parser's performance with and without normalizing biaffine scores. Normalizing allows us to beat the state of the art on two datasets, with fewer samples and trainable parameters. Code: https://anonymous.4open.science/r/EfficientSDP-70C1



Paperid:1518
Authors:Emma Rapoport, Edith Cohen, Uri Stemmer
Abstract:
Abstract:Most work on adaptive data analysis assumes that samples in the dataset are independent. When correlations are allowed, even the non-adaptive setting can become intractable, unless some structural constraints are imposed. To address this, Bassily and Freund [2016] introduced the elegant framework of *concentrated queries*, which requires the analyst to restrict itself to queries that are concentrated around their expected value. While this assumption makes the problem trivial in the non-adaptive setting, in the adaptive setting it remains quite challenging. In fact, all known algorithms in this framework support significantly fewer queries than in the independent case: At most $O(n)$ queries for a sample of size $n$, compared to $O(n^2)$ in the independent setting.In this work, we prove that this utility gap is inherent under the current formulation of the concentrated queries framework, assuming some natural conditions on the algorithm. Additionally, we present a simplified version of the best-known algorithms that match our impossibility result.



Paperid:1519
Authors:Niklas Dexheimer, Sascha Gaudlitz, Johannes Schmidt-Hieber
Abstract:
Hebbian learning is a key principle underlying learning in biological neural networks. It postulates that synaptic changes occur locally, depending on the activities of pre- and postsynaptic neurons. While Hebbian learning based on neuronal firing rates is well explored, much less is known about learning rules that account for precise spike-timing. We relate a Hebbian spike-timing-dependent plasticity rule to noisy gradient descent with respect to a natural loss function on the probability simplex. This connection allows us to prove that the learning rule eventually identifies the presynaptic neuron with the highest activity. We also discover an intrinsic connection to noisy mirror descent.



Paperid:1520
Authors:The Viet Bui, Tien Mai, Thanh Nguyen
Abstract:
We study offline imitation learning (IL) in cooperative multi-agent settings, where demonstrations have unlabeled mixed quality - containing both expert and suboptimal trajectories. Our proposed solution is structured in two stages: trajectory labeling and multi-agent imitation learning, designed jointly to enable effective learning from heterogeneous, unlabeled data. In the first stage, we combine advances in large language models and preference-based reinforcement learning to construct a progressive labeling pipeline that distinguishes expert-quality trajectories. In the second stage, we introduce MisoDICE, a novel multi-agent IL algorithm that leverages these labels to learn robust policies while addressing the computational complexity of large joint state-action spaces. By extending the popular single-agent DICE framework to multi-agent settings with a new value decomposition and mixing architecture, our method yields a convex policy optimization objective and ensures consistency between global and local policies. We evaluate MisoDICE on multiple standard multi-agent RL benchmarks and demonstrate superior performance, especially when expert data is scarce.



Paperid:1521
Authors:Binbin Huang, Luo Luo, Yanghua Xiao, Deqing Yang, Baojian Zhou
Abstract:
Abstract:This work proposes a novel framework based on nested evolving set processes to accelerate Personalized PageRank (PPR) computation. At each stage of the process, we employ a localized inexact proximal point iteration to solve a simplified linear system. We show that the time complexity of such localized methods is upper bounded by $\min\{\tilde{\mathcal{O}}(R^2/\epsilon^2), \tilde{\mathcal{O}}(m)\}$ to obtain an $\epsilon$-approximation of the PPR vector, where $m$ denotes the number of edges in the graph and $R$ is a constant defined via nested evolving set processes. Furthermore, the algorithms induced by our framework require solving only $\tilde{\mathcal{O}}(1/\sqrt{\alpha})$ such linear systems, where $\alpha$ is the damping factor. When $1/\epsilon^2\ll m$, this implies the existence of an algorithm that computes an $\epsilon$-approximation of the PPR vector with an overall time complexity of $\tilde{\mathcal{O}}(R^2 / (\sqrt{\alpha}\epsilon^2))$, independent of the underlying graph size. Our result resolves an open conjecture from existing literature. Experimental results on real-world graphs validate the efficiency of our methods, demonstrating significant convergence in the early stages.



Paperid:1522
Authors:Shulun Chen, Runlong Zhou, Zihan Zhang, Maryam Fazel, Simon Du
Abstract:
Abstract:We consider the gap-dependent regret bounds for episodic MDPs.We show that the Monotonic Value Propagation (MVP) algorithm (\cite{zhang2024settling}) achieves a variance-aware gap-dependent regret bound of $$\tilde{O}\left(\left(\sum_{\Delta_h(s,a)>0} \frac{H^2 \land \cVarmax}{\Delta_h(s,a)} +\sum_{\Delta_h(s,a)=0}\frac{ H^2 \land \cVarmax}{\Delta_{\mathrm{min}}} + S^2AH^3 \right) \cdot \log K\right),$$ where $H$ is the planning horizon, $S$ is the number of states, $A$ is the number of actions, and $K$ is the number of episodes. Here, $\Delta_h(s,a) =V_h^* (a) - Q_h^* (s, a)$ represents the sub-optimal gap and $\Delta_{\mathrm{min}} := \min_{(h,s,a)}\Delta_h(s,a)$.The term $\cVarmax$ denotes the maximum conditional total variance, calculated as the maximum over all $(\pi, h, s)$ tuples of the expected total variance under policy $\pi$ conditioned on trajectories visiting state $s$ at step $h$. This $\cVarmax$ characterizes the randomness encountered when learning a $(h, s)$ pair.Our result stems from a novel analysis of the weighted sum of the sub-optimal gap and can be potentially adapted for other algorithms.To complement the study, we establish a lower bound of $$\tilde{\Omega}\left( \sum_{\Delta_h(s,a)>0} \frac{H^2 \land \cVarmax}{\Delta_h(s,a)}\cdot \log(K)\right),$$ demonstrating the necessity of dependence on $\cVarmax$ even when the maximum unconditional total variance (without conditioning on $(h, s)$) approaches zero.



Paperid:1523
Authors:Muzhi Dai, Chenxu Yang, Qingyi Si
Abstract:
As Test-Time Scaling emerges as an active research focus in the large language model community, advanced post-training methods increasingly emphasize extending chain-of-thought (CoT) generation length, thereby enhancing reasoning capabilities to approach Deepseek R1-like reasoning models. However, recent studies reveal that reasoning models (even Qwen3) consistently exhibit excessive thought redundancy in CoT generation. This overthinking issue arises from the inherent limitations of conventional outcome-reward reinforcement learning, which systematically overlooks the regulation of intermediate reasoning processes. This paper introduces Serial-Group Decaying-Reward Policy Optimization (S-GRPO), a novel reinforcement learning paradigm that enables models to implicitly evaluate the sufficiency of intermediate reasoning steps, thereby facilitating early exit in CoT generation.Unlike GRPO, which samples multiple possible reasoning paths in parallel (parallel group), S-GRPO only samples one reasoning path and serially selects multiple temporal positions from the path to exit thinking and directly generate answers (serial group). For correct answers within a serial group, rewards gradually decrease based on the exit positions along the reasoning path from front to back. This design encourages the model to produce more accurate and concise thoughts, while also incentivizing early thinking termination when appropriate. Empirical evaluations demonstrate that S-GRPO is compatible with state-of-the-art reasoning models, including Qwen3 and Deepseek-distill. Across diverse benchmarks such as GSM8K, AIME 2024, AMC 2023, MATH-500, and GPQA Diamond, S-GRPO achieves a substantial reduction in sequence length (35.4%～61.1%) while simultaneously improving accuracy (absolute 0.72%～6.08%).



Paperid:1524
Authors:Carlos Stein Brito, Daniel McNamee
Abstract:
Deploying learned control policies in real-world environments poses a fundamental challenge. When system dynamics change unexpectedly, performance degrades until models are retrained on new data. We introduce Reflexive World Models (RWM), a dual control framework that uses world model predictions as implicit reference trajectories for rapid adaptation. Our method separates the control problem into long-term reward maximization through reinforcement learning and robust motor execution through rapid latent control. This dual architecture achieves significantly faster adaptation with low online computational cost compared to model-based RL baselines, while maintaining near-optimal performance. The approach combines the benefits of flexible policy learning through reinforcement learning with rapid error correction capabilities, providing a principled approach to maintaining performance in high-dimensional continuous control tasks under varying dynamics.



Paperid:1525
Authors:Andrei Panferov, Alexandra Volkova, Ionut-Vlad Modoranu, Vage Egiazarian, Mher Safaryan, Dan Alistarh
Abstract:
Scaling laws have shaped recent advances in machine learning by enabling predictable scaling of model performance based on model size, computation, and data volume. Concurrently, the rise in computational cost for AI has motivated model compression techniques, notably quantization and sparsification, which have emerged to mitigate the steep computational demands associated with large-scale training and inference. This paper investigates the interplay between scaling laws and compression strategies, exploring whether a unified scaling framework can accurately predict model performance when training occurs over various compressed representations, such as sparse, scalar-quantized, sparse-quantized or even vector-quantized formats. Our key contributions include proposing and validating a general scaling law formulation applicable both individually but also composably across compression types. We demonstrate both theoretically and empirically that a simple metric based on Gaussian mean squared error fitting can robustly predict parameter efficiency across compressed models. Additionally, we extend our formulation to directly compare the accuracy potential of different compressed formats, and to derive better algorithms for training over sparse-quantized formats. Finally, we identify conditions under which these unified scaling laws fail.



Paperid:1526
Authors:Huajie Qian, Donghao Ying, Henry Lam, Wotao Yin
Abstract:
Ensemble learning is a popular technique to improve the accuracy of machine learning models. It traditionally hinges on the rationale that aggregating multiple weak models can lead to better models with lower variance and hence higher stability, especially for discontinuous base learners. In this paper, we provide a new perspective on ensembling. By selecting the best model trained on subsamples via majority voting, we can attain exponentially decaying tails for the excess risk, even if the base learner suffers from slow (i.e., polynomial) decay rates. This tail enhancement power of ensembling is agnostic to the underlying base learner and is stronger than variance reduction in the sense of exhibiting rate improvement. We demonstrate how our ensemble methods can substantially improve out-of-sample performances in a range of numerical examples involving heavy-tailed data or intrinsically slow rates.



Paperid:1527
Authors:Amitayush Thakur, Jasper Lee, George Tsoukalas, Meghana Sistla, Matthew Zhao, Stefan Zetzsche, Greg Durrett, Yisong Yue, Swarat Chaudhuri
Abstract:
Abstract:We introduce ${\rm C{\small LEVER}}$, a high-quality, manually curated benchmark of 161 problems for end-to-end verified code generation in Lean. Each problem consists of (1) the task of generating a specification that matches a held-out ground-truth specification, and (2) the task of generating a Lean implementation that provably satisfies this specification. Unlike prior benchmarks,${\rm C{\small LEVER}}$ avoids test-case supervision, LLM-generated annotations, and specifications that leak implementation logic or allow vacuous solutions. All outputs are verified post-hoc using Lean's type checker to ensure machine-checkable correctness. We use ${\rm C{\small LEVER}}$ to evaluate several few-shot and agentic approaches based on state-of-the-art language models. These methods all struggle to achieve full verification, establishing it as a challenging frontier benchmark for program synthesis and formal reasoning. Our benchmark can be found on [GitHub](https://github.com/trishullab/clever) as well as [HuggingFace](https://huggingface.co/datasets/amitayusht/clever). All our evaluation code is also available [online](https://github.com/trishullab/clever-prover).



Paperid:1528
Authors:Valerii Startsev, Alexander Ustyuzhanin, Alexey Kirillov, Dmitry Baranchuk, Sergey Kastryulin
Abstract:
Pre-training equips text-to-image (T2I) models with broad world knowledge, but this alone is often insufficient to achieve high aesthetic quality and alignment. Consequently, supervised fine-tuning (SFT) is crucial for further refinement. However, its effectiveness highly depends on the quality of the fine-tuning dataset.Existing public SFT datasets frequently target narrow domains (e.g., anime or specific art styles), and the creation of high-quality, general-purpose SFT datasets remains a significant challenge.Current curation methods are often costly and struggle to identify truly impactful samples.This challenge is further complicated by the scarcity of public general-purpose datasets, as leading models often rely on large, proprietary, and poorly documented internal data, hindering broader research progress.This paper introduces a novel methodology for creating general-purpose SFT datasets by leveraging a pre-trained generative model as an estimator of high-impact training samples. We apply this methodology to construct and release Alchemist, a compact (3,350 samples) yet highly effective SFT dataset. Experiments demonstrate that Alchemist substantially improves the generative quality of five public T2I models while preserving diversity and style. Additionally, we release the fine-tuned models' weights to the public.



Paperid:1529
Authors:Fabian Paischer, Lukas Hauzenberger, Thomas Schmied, Benedikt Alkin, Marc Deisenroth, Sepp Hochreiter
Abstract:
Foundation models (FMs) are pre-trained onlarge-scale datasets and thenfine-tuned for a specific downstream task.The most common fine-tuning method is to update pretrained weights via low-rank adaptation (LoRA). Existing initialization strategies for LoRA often rely on singular value decompositions (SVD) of gradients or weight matrices. However, they do not provably maximize the expected gradient signal, which is critical for fast adaptation. To this end, we introduceExplainedVarianceAdaptation (EVA), an initialization scheme that uses the directions capturing the most activation variance, provably maximizing the expected gradient signal and accelerating fine-tuning.EVA performs incremental SVD on minibatches of activation vectors and selects the right-singular vectors for initialization once they converged.Further, by selecting the directions that capture the most activation-variance for a given rank budget, EVA accommodates adaptive ranks that reduce the number of trainable parameters, while maintaining or improving downstream performance.We apply EVA to a variety of fine-tuning tasks aslanguage generation and understanding, image classification, and reinforcement learning.EVA exhibits faster convergence than competitors and achieves the highest average score across a multitude of tasks per domain while reducing the number of trainable parameters through rank redistribution.



Paperid:1530
Authors:Yinghui Liu, Hao Miao, Guojiang Shen, Yan Zhao, Xiangjie Kong, Ivan Lee
Abstract:
Out-of-town trip recommendation aims to generate a sequence of Points of Interest (POIs) for users traveling from their hometowns to previously unvisited regions based on personalized itineraries, e.g., origin, destination, and trip duration. Modeling the complex user preferences--which often exhibit a two-fold nature of static and dynamic interests--is critical for effective recommendations. However, the sparsity of out-of-town check-in data presents significant challenges in capturing such user preferences. Meanwhile, existing methods often conflate the static and dynamic preferences, resulting in suboptimal performance. In this paper, we for the first time systematically study the problem of out-of-town trip recommendation. A novel framework SPOT-Trip is proposed to explicitly learns the dual static-dynamic user preferences. Specifically, to handle scarce data, we construct a POI attribute knowledge graph to enrich the semantic modeling of users’ hometown and out-of-town check-ins, enabling the static preference modeling through attribute relation-aware aggregation. Then, we employ neural ordinary differential equations (ODEs) to capture the continuous evolution of latent dynamic user preferences and innovatively combine a temporal point process to describe the instantaneous probability of each preference behavior. Further, a static-dynamic fusion module is proposed to merge the learned static and dynamic user preferences. Extensive experiments on real data offer insight into the effectiveness of the proposed solutions, showing that SPOT-Trip achieves performance improvement by up to 17.01%.



Paperid:1531
Authors:Yu Duan, Hamza Chaudhry, Misha B Ahrens, Christopher Harvey, Matthew Perich, Karl Deisseroth, Kanaka Rajan
Abstract:
Predicting future neural activity is a core challenge in modeling brain dynamics, with applications ranging from scientific investigation to closed-loop neurotechnology. While recent models of population activity emphasize interpretability and behavioral decoding, neural forecasting—particularly across multi-session, spontaneous recordings—remains underexplored. We introduce POCO, a unified forecasting model that combines a lightweight univariate forecaster with a population-level encoder to capture both neuron-specific and brain-wide dynamics. Trained across five calcium imaging datasets spanning zebrafish, mice, andC. elegans, POCO achieves state-of-the-art accuracy at cellular resolution in spontaneous behaviors. After pre-training, POCO rapidly adapts to new recordings with minimal fine-tuning. Notably, POCO's learned unit embeddings recover biologically meaningful structure—such as brain region clustering—without any anatomical labels. Our comprehensive analysis reveals several key factors influencing performance, including context length, session diversity, and preprocessing. Together, these results position POCO as a scalable and adaptable approach for cross-session neural forecasting and offer actionable insights for future model design. By enabling accurate, generalizable forecasting models of neural dynamics across individuals and species, POCO lays the groundwork for adaptive neurotechnologies and large-scale efforts for neural foundation models.



Paperid:1532
Authors:Stefano Zampini, Jacob K Christopher, Luca Oneto, Davide Anguita, Nando Fioretto
Abstract:
Stable diffusion models represent the state-of-the-art in data synthesis across diverse domains and hold transformative potential for applications in science and engineering, e.g., by facilitating the discovery of novel solutions and simulating systems that are computationally intractable to model explicitly.While there is increasing effort to incorporate physics-based constraints into generative models, existing techniques are either limited in their applicability to latent diffusion frameworks or lack the capability to strictly enforce domain-specific constraints.To address this limitation this paper proposes a novel integration of stable diffusion models with constrained optimization frameworks, enabling the generation of outputs satisfying stringent physical and functional requirements. The effectiveness of this approach is demonstrated through material design experiments requiring adherence to precise morphometric properties, challenging inverse design tasks involving the generation of materials inducing specific stress-strain responses, and copyright-constrained content generation tasks.



Paperid:1533
Authors:Shufan Li, Konstantinos Kallidromitis, Hritik Bansal, Akash Gokul, Yusuke Kato, Kazuki Kozuka, Jason Kuen, Zhe Lin, Kai-Wei Chang, Aditya Grover
Abstract:
Modern Vision-Language Models (VLMs) can solve a wide range of tasks requiring visual reasoning. In real-world scenarios, desirable properties for VLMs include fast inference and controllable generation (e.g., constraining outputs to adhere to a desired format). However, existing autoregressive (AR) VLMs like LLaVA struggle in these aspects. Discrete diffusion models (DMs) offer a promising alternative, enabling parallel decoding for faster inference and bidirectional context for controllable generation through text-infilling. While effective in language-only settings, DMs' potential for multimodal tasks is underexplored. We introduce LaViDa, a family of VLMs built on DMs. We build LaViDa by equipping DMs with a vision encoder and jointly fine-tune the combined parts for multimodal instruction following. To address challenges encountered, LaViDa incorporates novel techniques such as complementary masking for effective training, prefix KV cache for efficient inference, and timestep shifting for high-quality sampling. Experiments show that LaViDa achieves competitive or superior performance to AR VLMs on multi-modal benchmarks such as MMMU, while offering unique advantages of DMs, including flexible speed-quality tradeoff, controllability, and bidirectional reasoning. On COCO captioning, LaViDa surpasses Open-LLaVa-Next-8B by +4.1 CIDEr with 1.92x speedup. On bidirectional tasks, it achieves +59% improvement on Constrained Poem Completion. These results demonstrate LaViDa as a strong alternative to AR VLMs. Code and models will be released in the camera-ready version.



Paperid:1534
Authors:Vignav Ramesh, Morteza Mardani
Abstract:
Abstract:The iterative and stochastic nature of diffusion models enables *test-time scaling*, whereby spending additional compute during denoising generates higher-fidelity samples. Increasing the number of denoising steps is the primary scaling axis, but this yields quickly diminishing returns. Instead optimizing the *noise trajectory*—the sequence of injected noise vectors—is promising, as the specific noise realizations critically affect sample quality; but this is challenging due to a high-dimensional search space, complex noise-outcome interactions, and costly trajectory evaluations. We address this by first casting diffusion as a Markov Decision Process (MDP) with a terminal reward, showing tree-search methods such as Monte Carlo tree search (MCTS) to be meaningful but impractical. To balance performance and efficiency, we then resort to a relaxation of MDP, where we view denoising as a sequence of independent *contextual bandits*. This allows us to introduce an $\epsilon$-greedy search algorithm that *globally explores* at extreme timesteps and *locally exploits* during the intermediate steps where de-mixing occurs. Experiments on EDM and Stable Diffusion reveal state-of-the-art scores for class-conditioned/text-to-image generation, exceeding baselines by up to $164$% and matching/exceeding MCTS performance. To our knowledge, this is the first practical method for test-time noise *trajectory* optimization of *arbitrary (non-differentiable)* rewards.



Paperid:1535
Authors:Ali Shirali
Abstract:
Algorithms shape countless human interactions---ranging from querying chatbots to discovering new media content. These interactions often unfold over multiple steps, during which strategic users can guide algorithms to better align with their true interests by selectively engaging with content. However, users frequently exhibit inconsistent preferences: they may spend considerable time on content that offers little long-term value, inadvertently signaling that such content is desirable. Focusing on the user side, this raises a key question: what does it take for such users to align the algorithm with their true interests?To investigate these dynamics, we model the user’s decision process as split between a rational system 2 that decides whether to engage and an impulsive system 1 that determines how long engagement lasts. We then study a multi-leader, single-follower extensive Stackelberg game, where users---specifically system 2---lead by committing to engagement strategies and the algorithm best-responds based on observed interactions. We define the burden of alignment as the minimum horizon over which users must optimize to effectively steer the algorithm. We show that a critical horizon exists: users who are sufficiently foresighted can achieve alignment, while those who are not are instead aligned to the algorithm’s objective. This critical horizon can be long, imposing a substantial burden. However, even a small, costly signal (e.g., an extra click) can significantly reduce it. Overall, our framework explains how users with inconsistent preferences can align an engagement-driven algorithm with their interests in equilibrium, highlighting both the challenges and potential remedies for achieving alignment.



Paperid:1536
Authors:Dong Yang, YIYI CAI, Yuki Saito, Lixu Wang, Hiroshi Saruwatari
Abstract:
We propose a shallow flow matching (SFM) mechanism to enhance flow matching (FM)-based text-to-speech (TTS) models within a coarse-to-fine generation paradigm. SFM constructs intermediate states along the FM paths using coarse output representations. During training, we introduce an orthogonal projection method to adaptively determine the temporal position of these states, and apply a principled construction strategy based on a single-segment piecewise flow. The SFM inference starts from the intermediate state rather than pure noise and focuses computation on the latter stages of the FM paths. We integrate SFM into multiple TTS models with a lightweight SFM head. Experiments show that SFM consistently improves the naturalness of synthesized speech in both objective and subjective evaluations, while significantly reducing inference when using adaptive-step ODE solvers. Demo and codes are provided in the supplementary materials.



Paperid:1537
Authors:Zhiwei Zhang, Zi Ye, Yibin Wen, Shuai Yuan, Haohuan Fu, Huang Jianxi, Juepeng Zheng
Abstract:
Agricultural parcels serve as basic units for conducting agricultural practices and applications, which is vital for land ownership registration, food security assessment, soil erosion monitoring, etc. However, existing agriculture parcel extraction studies only focus on mid-resolution mapping or regular plain farmlands while lacking representation of complex terraced terrains due to the demands of precision agriculture. In this paper, we introduce a more fine-grained terraced parcel dataset named GTPBD (Global Terraced Parcel and Boundary Dataset), which is the first fine-grained dataset covering major worldwide terraced regions with more than 200,000 complex terraced parcels with manually annotation. GTPBD comprises 24,238 high-resolution images with three-level labels, including pixel-level boundary labels, mask labels, and parcel labels. It covers seven major geographic zones in China and transcontinental climatic regions around the world. Compared to the existing datasets, the GTPBD dataset brings considerable challenges due to the: (1) terrain diversity; (2) complex and irregular parcel objects; and (3) multiple domain styles. Our proposed GTPBD dataset is suitable for four different tasks, including semantic segmentation, edge detection, terraced parcel extraction and unsupervised domain adaptation (UDA) tasks. Accordingly, we benchmark the GTPBD dataset on eight semantic segmentation methods, four edge extraction methods, three parcel extraction methods and five UDA methods, along with a multi-dimensional evaluation framework integrating pixel-level and object-level metrics. GTPBD fills a critical gap in terraced remote sensing research, providing a basic infrastructure for fine-grained agricultural terrain analysis and cross-scenario knowledge transfer. The code and data are available at https://github.com/Z-ZW-WXQ/GTPBG/.



Paperid:1538
Authors:Jiaxiang Tang, Ruijie Lu, Max Li, Zekun Hao, Xuan Li, Fangyin Wei, Shuran Song, Gang Zeng, Ming-Yu Liu, Tsung-Yi Lin
Abstract:
Recent progress in 3D object generation has greatly improved both the quality and efficiency.However, most existing methods generate a single mesh with all parts fused together, which limits the ability to edit or manipulate individual parts.A key challenge is that different objects may have a varying number of parts.To address this, we propose a new end-to-end framework for part-level 3D object generation.Given a single input image, our method generates high-quality 3D objects with an arbitrary number of complete and semantically meaningful parts.We introduce a dual volume packing strategy that organizes all parts into two complementary volumes, allowing for the creation of complete and interleaved parts that assemble into the final object.Experiments show that our model achieves better quality, diversity, and generalization than previous image-based part-level generation methods.



Paperid:1539
Authors:Mingxuan Li, Junzhe Zhang, Elias Bareinboim
Abstract:
A key task in Artificial Intelligence is learning effective policies for controlling agents in unknown environments to optimize performance measures. Off-policy learning methods, like Q-learning, allow learners to make optimal decisions based on past experiences. This paper studies off-policy learning from biased data in complex and high-dimensional domains where \emph{unobserved confounding} cannot be ruled out a priori. Building on the well-celebrated Deep Q-Network (DQN), we propose a novel deep reinforcement learning algorithm robust to confounding biases in observed data. Specifically, our algorithm attempts to find a safe policy for the worst-case environment compatible with the observations. We apply our method to twelve confounded Atari games, and find that it consistently dominates the standard DQN in all games where the observed input to the behavioral and target policies mismatch and unobserved confounders exist.



Paperid:1540
Authors:Houssam Zenati, Bariscan Bozkurt, Arthur Gretton
Abstract:
Estimating the distribution of outcomes under counterfactual policies is critical for decision-making in domains such as recommendation, advertising, and healthcare. We analyze a novel framework—Counterfactual Policy Mean Embedding (CPME)—that represents the entire counterfactual outcome distribution in a reproducing kernel Hilbert space (RKHS), enabling flexible and nonparametric distributional off-policy evaluation. We introduce both a plug-in estimator and a doubly robust estimator; the latter enjoys improved uniform convergence rates by correcting for bias in both the outcome embedding and propensity models. Building on this, we develop a doubly robust kernel test statistic for hypothesis testing, which achieves asymptotic normality and thus enables computationally efficient testing and straightforward construction of confidence intervals. Our framework also supports sampling from the counterfactual distribution. Numerical simulations illustrate the practical benefits of CPME over existing methods.



Paperid:1541
Authors:Haipeng Luo, Spandan Senapati, Vatsal Sharan
Abstract:
Abstract:In this paper, we consider the related problems of multicalibration --- a multigroup fairness notion and omniprediction --- a simultaneous loss minimization paradigm, both in the distributional and online settings. The recent work of Garg et al. (2024) raised the open problem of whether it is possible to efficiently achieve $\tilde{\mathcal{O}}(\sqrt{T})$ $\ell_{2}$-multicalibration error against bounded linear functions. In this paper, we answer this question in a strongly affirmative sense. We propose an efficient algorithm that achieves $\tilde{\mathcal{O}}(T^{\frac{1}{3}})$ $\ell_{2}$-swap multicalibration error (both in high probability and expectation). On propagating this bound onward, we obtain significantly improved rates for $\ell_{1}$-swap multicalibration and swap omniprediction for a loss class of convex Lipschitz functions. In particular, we show that our algorithm achieves $\tilde{\mathcal{O}}(T^{\frac{2}{3}})$ $\ell_{1}$-swap multicalibration and swap omniprediction errors, thereby improving upon the previous best-known bound of $\tilde{\mathcal{O}}(T^{\frac{7}{8}})$. As a consequence of our improved online results, we further obtain several improved sample complexity rates in the distributional setting. In particular, we establish a $\tilde{\mathcal{O}}(\varepsilon ^ {-3})$ sample complexity of efficiently learning an $\varepsilon$-swap omnipredictor for the class of convex and Lipschitz functions, $\tilde{\mathcal{O}}(\varepsilon ^{-2.5})$ sample complexity of efficiently learning an $\varepsilon$-swap agnostic learner for the squared loss, and $\tilde{\mathcal{O}}(\varepsilon ^ {-5}), \tilde{\mathcal{O}}(\varepsilon ^ {-2.5})$ sample complexities of learning $\ell_{1}, \ell_{2}$-swap multicalibrated predictors against linear functions, all of which significantly improve on the previous best-known bounds.



Paperid:1542
Authors:Sean McGregor, Vassil Tashev, Armstrong Foundjem, Aishwarya Ramasethu, Sadegh AlMahdi Kazemi Zarkouei, Chris Knotz, Kongtao Chen, Alicia Parrish, Anka Reuel-Lamparth, Heather Frase
Abstract:
Large language model (LLM) benchmarks inform LLM use decisions (e.g., "is this LLM safe to deploy for my use case and context?"). However, benchmarks may be rendered unreliable by various failure modes impacting benchmark bias, variance, coverage, or people's capacity to understand benchmark evidence. Using the National Institute of Standards and Technology's risk management process as a foundation, this research iteratively analyzed 26 popular benchmarks, identifying 57 potential failure modes and 196 corresponding mitigation strategies. The mitigations reduce failure likelihood and/or severity, providing a frame for evaluating "benchmark risk," which is scored to provide a metaevaluation benchmark: BenchRisk. Higher scores indicate benchmark users are less likely to reach an incorrect or unsupported conclusion about an LLM. All 26 scored benchmarks present significant risk within one or more of the five scored dimensions (comprehensiveness, intelligibility, consistency, correctness, and longevity), which points to important open research directions for the field of LLM benchmarking. The BenchRisk workflow allows for comparison between benchmarks; as an open-source tool, it also facilitates the identification and sharing of risks and their mitigations.



Paperid:1543
Authors:Elizabeth Donoway, Hailey Joren, Arushi Somani, Henry Sleight, Julian Michael, Michael Deweese, John Schulman, Ethan Perez, Fabien Roger, Jan Leike
Abstract:
Large language models often possess latent capabilities that lie dormant unless explicitly elicited, or surfaced, through fine-tuning or prompt engineering. Predicting, assessing, and understanding these latent capabilities pose significant challenges in the development of effective, safe AI systems. In this work, we recast elicitation as an information-constrained fine-tuning problem and empirically characterize an upper bound on the minimal number of parameters needed to reach a certain task accuracy. We find that training as few as 10–100 randomly chosen parameters---several orders of magnitude fewer than state-of-the-art parameter-efficient methods---can recover up to 50\% of the performance gap between pretrained-only and fully fine-tuned models, and 1,000s to 10,000s of parameters can recover 95\% of this performance gap. We show that a logistic curve fits the relationship between the number of trained parameters and the fraction of the performance gap recovered. To help explain this behavior, we consider a simplified picture of elicitation via fine-tuning where each trainable parameter serves as an encoding mechanism for accessing task-specific knowledge. We observe a relationship between the number of trained parameters and the effectiveness in eliciting latent capabilities, offering a potential route to distinguish elicitation from teaching.



Paperid:1544
Authors:Diyang Li, Kyra Gan
Abstract:
Targeted Maximum Likelihood Estimation (TMLE) is a powerful debiasing algorithm for plug-in estimation, widely employed across research communities due to its double robustness and asymptotic efficiency.While extensive literature has concentrated on its statistical guarantees, the inherent convergence properties of TMLE as an iterative optimization scheme have remained inadequately explored.To bridge this critical gap, we rigorously investigate the optimization dynamics of TMLE iterations under standard assumptions and regularity conditions.We begin with studying key thresholding rules in TMLE framework and establish their fundamental connections to convergence.Our analysis reveals that the homotopy mapping induced by each fluctuation submodel defines a smooth embedding into the probability simplex, which constitutes a topological invariance that is free of self-intersection and carries implications for algorithmic stability.We then derive a structural characterization for the solution locus of targeted estimating‐equation and loss landscape, showing that it forms a submanifold of the statistical model with codimension equal to the dimension of target parameter.Building on these geometric insights, we deliver the first strict proof of TMLE's global convergence, as well as explicit sufficient criteria under which TMLE terminates in a single update.As a by-product, we discover an unidentified overshooting phenomenon wherein the algorithm can surpass feasible roots to the estimating-equation along homotopy path, highlighting a promising avenue for designing enhanced debias algorithms.



Paperid:1545
Authors:Yang Liu, Daxuan Ren, Yijie Ding, Jianmin Zheng, Fang Deng
Abstract:
This paper presents PartCAD, a novel framework for reconstructing CAD modeling sequences directly from point clouds by projection-guided, part-aware geometry reasoning. It consists of (1) an autoregressive approach that decomposes point clouds into part-aware latent representations, serving as interpretable anchors for CAD generation; (2) a projection guidance module that provides explicit cues about underlying design intent via triplane projections; and (3) a non-autoregressive decoder to generate sketch-extrusion parameters in a single forward pass, enabling efficient and structurally coherent CAD instruction synthesis. By bridging geometric signals and semantic understanding, PartCAD tackles the challenge of reconstructing editable CAD models—capturing underlying design processes—from 3D point clouds. Extensive experiments show that PartCAD significantly outperforms existing methods for CAD instruction generation in both accuracy and robustness. The work sheds light on part-driven reconstruction of interpretable CAD models, opening new avenues in reverse engineering and CAD automation.



Paperid:1546
Authors:Hongxin Xiang, Ke Li, Mingquan Liu, Zhixiang Cheng, Bin Yao, Wenjie Du, Jun Xia, Li Zeng, Xin Jin, xiangxiang Zeng
Abstract:
Abstract:Existing molecular machine learning force fields (MLFFs) generally focus on the learning of atoms, molecules, and simple quantum chemical properties (such as energy and force), but ignore the importance of electron density (ED) $\rho(r)$ in accurately understanding molecular force fields (MFFs). ED describes the probability of finding electrons at specific locations around atoms or molecules, which uniquely determines all ground state properties (such as energy, molecular structure, etc.) of interactive multi-particle systems according to the Hohenberg-Kohn theorem. However, the calculation of ED relies on the time-consuming first-principles density functional theory (DFT), which leads to the lack of large-scale ED data and limits its application in MLFFs. In this paper, we introduce EDBench, a large-scale, high-quality dataset of ED designed to advance learning-based research at the electronic scale. Built upon the PCQM4Mv2, EDBench provides accurate ED data, covering 3.3 million molecules. To comprehensively evaluate the ability of models to understand and utilize electronic information, we design a suite of ED-centric benchmark tasks spanning prediction, retrieval, and generation. Our evaluation of several state-of-the-art methods demonstrates that learning from EDBench is not only feasible but also achieves high accuracy. Moreover, we show that learning-based methods can efficiently calculate ED with comparable precision while significantly reducing the computational cost relative to traditional DFT calculations. All data and benchmarks from EDBench will be freely available, laying a robust foundation for ED-driven drug discovery and materials science.



Paperid:1547
Authors:William Roy Orchard, Nastaran Okati, Sergio Garrido Mejia, Patrick Blöbaum, Dominik Janzing
Abstract:
The goal of Root Cause Analysis (RCA) is to explain why an anomaly occurred by identifying where the fault originated. Several recent works model the anomalous event as resulting from a change in the causal mechanism at the root cause, i.e. as a soft intervention. RCA is then the task of identifying which causal mechanism changed. In real-world applications, one often has either few or only a single sample from the post-intervention distribution: a severe limitation for most methods, which assume one knows or can estimate the distribution. However, even those which do not are statistically ill-posed due to the need to probe regression models in regions of low probability density. In this paper we propose simple, efficient methods to overcome both difficulties in the case where there is a single root cause and the causal graph is a directed tree. When one knows the causal graph, we give guarantees for a traversal algorithm which requires only marginal anomaly scores and does not depend on specifying an arbitrary anomaly score cut-off. When one does not know the causal graph, we show that the heuristic of identifying root causes as the variables with the highest marginal anomaly scores is causally justified. To this end, we prove anomalies with small scores are unlikely to cause those with larger scores and give upper bounds for the likelihood of causal pathways with non-monotonic anomaly scores.



Paperid:1548
Authors:Jae-Young Yim, Dongwook Kim, Jae-Young Sim
Abstract:
Large-scale datasets are usually required to train deep neural networks, but it increases the computational complexity hindering the practical applications. Recently, dataset distillation for images and texts has been attracting a lot of attention, that reduces the original dataset to a synthetic dataset to alleviate the computational burden of training while preserving essential task-relevant information. However, the dataset distillation for 3D point clouds remains largely unexplored, as the point clouds exhibit fundamentally different characteristics from that of images, making the dataset distillation more challenging. In this paper, we propose a distribution matching-based distillation framework for 3D point clouds that jointly optimizes the geometric structures as well as the orientations of the synthetic 3D objects. To address the semantic misalignment caused by unordered indexing of points, we introduce a Semantically Aligned Distribution Matching loss computed on the sorted features in each channel. Moreover, to address the rotation variation, we jointly learn the optimal rotation angles while updating the synthetic dataset to better align with the original feature distribution. Extensive experiments on widely used benchmark datasets demonstrate that the proposed method consistently outperforms existing dataset distillation methods, achieving superior accuracy and strong cross-architecture generalization.



Paperid:1549
Authors:Alan Chen, Jack Merullo, Alessandro Stolfo, Ellie Pavlick
Abstract:
In this work, we demonstrate that affine mappings between residual streams of language models is a cheap way to effectively transfer represented features between models. We apply this technique to transfer the \textit{weights} of Sparse Autoencoders (SAEs) between models of different sizes to compare their representations. We find that small and large models learn highly similar representation spaces, which motivates training expensive components like SAEs on a smaller model and transferring to a larger model at a FLOPs savings. For example, using a small-to-large transferred SAE as initialization can lead to 50% cheaper training runs when training SAEs on larger models. Next, we show that transferred probes and steering vectors can effectively recover ground truth performance. Finally, we dive deeper into feature-level transferability, finding that semantic and structural features transfer noticeably differently while specific classes of functional features have their roles faithfully mapped. Overall, our findings illustrate similarities and differences in the linear representation spaces of small and large models and demonstrate a method for improving the training efficiency of SAEs.



Paperid:1550
Authors:Chase van de Geijn, Polina Turishcheva, Alexander Ecker, Timo Lüddecke
Abstract:
Abstract:Rotary Positional Encodings (RoPE) have emerged as a highly effective technique for one-dimensional sequences in Natural Language Processing spurring recent progress towards generalizing RoPE to higher-dimensional data such as images and videos. The success of RoPE has been thought to be due to its positional equivariance, i.e. its status as a \textit{relative} positional encoding. In this paper, we mathematically show RoPE to be one of the most general solutions for equivariant positional embedding in one-dimensional data. Moreover, we show Mixed RoPE to be the analogously general solution for $M$-dimensional data, if we require commutative generators -- a property necessary for RoPE's equivariance. However, we question the necessity of equivariance. We propose Spherical RoPE, a method analogous to Mixed RoPE, but with the assumption of anti-commutative generators -- relaxing the equivariant condition. Empirically, we find Spherical RoPE to have the equivalent learning behavior as its equivariant analogues. This strongly suggests that relative positional embeddings are not as important as is commonly believed. We expect this discovery to facilitate future work in positional encodings for vision that are faster and generalize better by removing the preconception that they must be relative.



Paperid:1551
Authors:Xiangyan Liu, Jinjie Ni, Zijian Wu, Chao Du, Longxu Dou, Haonan Wang, Tianyu Pang, Michael Qizhe Shieh
Abstract:
Abstract:Recent advances in reinforcement learning (RL) have strengthened the reasoning capabilities of vision-language models (VLMs). However, enhancing policy exploration to better scale test-time compute remains largely underexplored. In addition, VLMs continue to struggle with imperfect visual perception, which in turn affects the subsequent reasoning process.To this end, we propose **NoisyRollout**, a simple yet effective data augmentation method that mixes trajectories from both clean and moderately distorted images during RL training. By injecting targeted diversity in visual perception and the resulting reasoning patterns, NoisyRollout promotes better policy exploration through vision-oriented inductive biases, ultimately leading to more robust reasoning behaviors. We further adopt a noise annealing schedule that gradually reduces distortion strength over training, leveraging noisy signals early on while ensuring training stability in later stages.Crucially, our method is easy-to-adopt—**requiring no additional training cost and no modifications to the RL objective**. Extensive experiments on $2$ distinct training datasets demonstrate that NoisyRollout achieves state-of-the-art performance among open-source RL-tuned models across $5$ out-of-domain reasoning and perception benchmarks. Furthermore, we validate the effectiveness of NoisyRollout across model sizes ($7$B and $32$B) and data scales (from $1$K to $6$K), highlighting its generalizability and scalability.



Paperid:1552
Authors:Xeron Du, Yifan Yao, Kaijing Ma, Bingli Wang, Tianyu Zheng, Zhu, Minghao Liu, Yiming Liang, Xiaolong Jin, Zhenlin Wei, Chujie Zheng, Kaixin Deng, Shuyue Guo, Shian Jia, Sichao Jiang, Yiyan Liao, Rui Li, Qinrui Li, Sirun Li, Yizhi Li, Yunwen Li, Dehua Ma, Yuansheng Ni, Haoran Que, Qiyao Wang, Zhoufutu Wen, Siwei Wu, Tianshun Xing, 明 许, Zhenzhu Yang, Noah Wang, Junting Zhou, yuelin bai, Xingyuan Bu, chenglin cai, Liang Chen, Yifan Chen, Cheng Chengtuo, Tianhao Cheng, Keyi Ding, Siming Huang, HUANG YUN, Yaoru Li, Yizhe Li, Zhaoqun Li, Tianhao Liang, Chengdong Lin, Hongquan Lin, Yinghao Ma, Zhongyuan Peng, Zifan Peng, Qige Qi, Shi Qiu, Xingwei Qu, Shanghaoran Quan, Yizhou Tan, Zili Wang, 王晨清, Hao Wang, Yiya Wang, Yubo Wang, Jiajun Xu, Kexin Yang, Ruibin Yuan, Yuanhao Yue, Tianyang Zhan, Chun Zhang, Jinyang Zhang, Xiyue Zhang, Owen Zhang, Yue Zhang, Yongchi Zhao, Xiangyu Zheng, ChenghuaZhong, Yang Gao, Zhoujun Li, Dayiheng Liu, Qian Liu, Tianyu Liu, Shiwen Ni, Junran Peng, Yujia Qin, Wenbo Su, Guoyin Wang, Shi Wang, Jian Yang, Min Yang, Meng Cao, Xiang Yue, ZHAO-XIANG ZHANG, Wangchunshu Zhou, Jiaheng Liu, Qunshu Lin, Wenhao Huang, Ge Zhang
Abstract:
Large language models (LLMs) have demonstrated remarkable proficiency in mainstream academic disciplines such as mathematics, physics, and computer science. However, human knowledge encompasses over 200 specialized disciplines, far exceeding the scope of existing benchmarks. The capabilities of LLMs in many of these specialized fields-particularly in light industry, agriculture, and service-oriented disciplines-remain inadequately evaluated. To address this gap, we present SuperGPQA, a comprehensive benchmark that evaluates graduate-level knowledge and reasoning capabilities across 285 disciplines. Our benchmark employs a novel Human-LLM collaborative filtering mechanism to eliminate trivial or ambiguous questions through iterative refinement based on both LLM responses and expert feedback. Our experimental results reveal significant room for improvement in the performance of current state-of-the-art LLMs across diverse knowledge domains (e.g., the reasoning-focused model DeepSeek-R1 achieved the highest accuracy of 61.82% on SuperGPQA), highlighting the considerable gap between current model capabilities and artificial general intelligence. Additionally, we present comprehensive insights from our management of a large-scale annotation process, involving over 80 expert annotators and an interactive Human-LLM collaborative system, offering valuable methodological guidance for future research initiatives of comparable scope.



Paperid:1553
Authors:Lisa Alazraki, Yi Chern Tan, Jon Ander Campos, Maximilian Mozes, Marek Rei, Max Bartolo
Abstract:
The capabilities of Large Language Models (LLMs) are routinely evaluated by other LLMs trained to predict human preferences. This framework—known asLLM-as-a-judge—is highly scalable and relatively low cost. However, it is also vulnerable to malicious exploitation, as LLM responses can be tuned to overfit the preferences of the judge. Previous work shows that the answers generated by a candidate-LLM can be editedpost hocto maximise the score assigned to them by a judge-LLM. In this study, we adopt a different approach and use the signal provided by judge-LLMs as a reward to adversarially tune models that generate text preambles designed to boost downstream performance. We find that frozen LLMs pipelined with these models attain higher LLM-evaluation scores than existing frameworks. Crucially, unlike other frameworks which intervene directly on the model's response, our method is virtually undetectable. We also demonstrate that the effectiveness of the tuned preamble generator transfers when the candidate-LLM and the judge-LLM are replaced with models that are not used during training. These findings raise important questions about the design of more reliable LLM-as-a-judge evaluation settings. They also demonstrate that human preferences can be reverse engineered effectively, by pipelining LLMs to optimise upstream preambles via reinforcement learning—an approach that could find future applications in diverse tasks and domains beyond adversarial attacks.



Paperid:1554
Authors:Mudit, Utsav Singh, Amrit Singh Bedi, Raghu Pasupathy, Vaneet Aggarwal
Abstract:
Abstract:Bilevel reinforcement learning (BRL) has emerged as a powerful framework for aligning generative models, yet its theoretical foundations, especially sample complexity bounds, remain underexplored. In this work, we present the first sample complexity bound for BRL, establishing a rate of $\mathcal{O}(\epsilon^{-3})$ in continuous state-action spaces. Traditional MDP analysis techniques do not extend to BRL due to its nested structure and non-convex lower-level problems. We overcome these challenges by leveraging the Polyak-Łojasiewicz (PL) condition and the MDP structure to obtain closed-form gradients, enabling tight sample complexity analysis. Our analysis also extends to general bi-level optimization settings with non-convex lower levels, where we achieve state-of-the-art sample complexity results of $\mathcal{O}(\epsilon^{-3})$ improving upon existing bounds of $\mathcal{O}(\epsilon^{-6})$. Additionally, we address the computational bottleneck of hypergradient estimation by proposing a fully first-order, Hessian-free algorithm suitable for large-scale problems.



Paperid:1555
Authors:Charles Arnal, Gaëtan Narozniak, Vivien Cabannes, Yunhao Tang, Julia Kempe, Remi Munos
Abstract:
Abstract:Reinforcement learning (RL) is increasingly used to align large language models (LLMs). Off-policy methods offer greater implementation simplicity and data efficiency than on-policy techniques, but often result in suboptimal performance. In this work, we study the intermediate range of algorithms between off-policy RL and supervised fine-tuning (SFT) by analyzing a simple off-policy policy gradient (PG) algorithm, where the advantage is defined as $A=r-V$, with $r$ a reward and $V$ some tunable baseline. Intuitively, lowering $V$ emphasizes correct (positively rewarded) samples, while raising it penalizes incorrect ones more heavily.We first provide a theoretical analysis of this off-policy PG algorithm, showing that when the baseline $V$ lower-bounds the expected reward, the algorithm enjoys a policy improvement guarantee. Our analysis reveals that while on-policy updates can safely leverage both positive and negative signals, off-policy updates benefit from focusing more on positive rewards than on negative ones.We validate our findings experimentally in a controlled stochastic bandit setting and through fine-tuning state-of-the-art LLMs on reasoning tasks.



Paperid:1556
Authors:Shixuan Liu, Yue He, Haotian Wang, Wenjing Yang, Yunfei Wang, Peng Cui, Zhong Liu
Abstract:
Data-driven methods offer efficient and robust solutions for analyzing complex dynamic systems but rely on the assumption of IID data, driving the development of generalization techniques for environmental differences. These techniques, however, are limited by their dependence on environment labels, which are often unavailable during training due to data acquisition challenges, privacy concerns, and environmental variability, particularly in large public datasets and privacy-sensitive domains. In response, we propose DynaInfer, a novel method that infers environment specifications by analyzing prediction errors from fixed neural networks within each training round, enabling environment assignments directly from data. We prove our algorithm effectively solves the alternate optimization problem in unlabeled scenarios and validate it through extensive experiments across diverse dynamic systems. Results show that DynaInfer outperforms existing environment assignment techniques, converges rapidly to true labels, and even achieves superior performance when environment labels are available.



Paperid:1557
Authors:Tian Gao, Songtao Lu, Junkyu Lee, Elliot Nelson, Debarun Bhattacharjya, Yue Yu, Miao Liu
Abstract:
Abstract:Causal discovery in the form of a directed acyclic graph (DAG) for dynamic time series data has been widely studied in various applications. Much of the existing work has focused on observational, offline, and/or stationary settings. In this work, we propose a dynamic DAG discovery algorithm, Meta-D$^2$AG, based on online meta-learning. Meta-D$^2$AG is designed to learn dynamic DAG structures from potentially nonlinear and non-stationary times series datasets, accounting for changes in both parameters and graph structures. Notably, Meta-D$^2$AG explicitly treats data collected at different time points with distribution shifts as distinct domains, which is assumed to occur as a result of external interventions. Moreover, Meta-D$^2$AG contains a new online meta-learning framework to take advantage of the temporal transition among existing domains such that it can quickly adapt to new domains with few measurements. A first-order optimization approach is utilized to efficiently solve the meta-learning framework, and theoretical analysis establishes the identifiability conditions and the convergence of the learning process. We demonstrate the promising performance of our method through better accuracy and sample efficiency on benchmark datasets against state-of-the-art baselines.



Paperid:1558
Authors:Wenqian Ye, Guangtao Zheng, Aidong Zhang
Abstract:
In reinforcement learning from human feedback, preference-based reward models play a central role in aligning large language models to human-aligned behavior. However, recent studies show that these models are prone to reward hacking and often fail to generalize well due to over-optimization. They achieve high reward scores by exploiting shortcuts, that is, using spurious features (e.g., response verbosity, agreeable tone, or sycophancy) that correlate with human preference labels in the training data rather than genuinely reflecting the intended objectives. In this paper, instead of probing these issues one at a time, we take a broader view of the reward hacking problem as shortcut behaviors and introduce a principled yet flexible approach to mitigate shortcut behaviors in preference-based reward learning. Inspired by the invariant theory in the kernel perspective, we propose Preference-based Reward Invariance for Shortcut Mitigation (PRISM), which learns group-invariant kernels with feature maps in a closed-form learning objective. Experimental results in several benchmarks show that our method consistently improves the accuracy of the reward model on diverse out-of-distribution tasks and reduces the dependency of shortcuts in downstream policy models, establishing a robust framework for preference-based alignment.



Paperid:1559
Authors:Prateek Chanda, Prayas Agrawal, Saral Sureka, Lokesh Reddy Polu, Atharv Kshirsagar, Ganesh Ramakrishnan
Abstract:
Traditional curriculum learning proceeds from easy to hard samples, yet defining a reliable notion of difficulty remains elusive. Prior work has used submodular functions to induce difficulty scores in curriculum learning. We reinterpret adaptive subset selection and formulate it as a multi-armed bandit problem, where each arm corresponds to a submodular function guiding sample selection. We introduce OnlineSubmod, a novel online greedy policy that optimizes a utility-driven reward and provably achieves no-regret performance under various sampling regimes. Empirically, OnlineSubmod outperforms both traditional curriculum learning and bi-level optimization approaches across vision and language datasets, showing superior accuracy-efficiency tradeoffs. More broadly, we show that validation-driven reward metrics offer a principled way to guide the curriculum schedule.



Paperid:1560
Authors:Fu Luo, Xi Lin, Mengyuan Zhong, Fei Liu, Zhenkun Wang, Jianyong Sun, Qingfu Zhang
Abstract:
Neural Combinatorial Optimisation (NCO) is a promising learning-based approach for solving Vehicle Routing Problems (VRPs) without extensive manual design. While existing constructive NCO methods typically follow an appending-based paradigm that sequentially adds unvisited nodes to partial solutions, this rigid approach often leads to suboptimal results. To overcome this limitation, we explore the idea of insertion-based paradigm and propose Learning to Construct with Insertion-based Paradigm (L2C-Insert), a novel learning-based method for constructive NCO. Unlike traditional approaches, L2C-Insert builds solutions by strategically inserting unvisited nodes at any valid position in the current partial solution, which can significantly enhance the flexibility and solution quality. The proposed framework introduces three key components: a novel model architecture for precise insertion position prediction, an efficient training scheme for model optimization, and an advanced inference technique that fully exploits the insertion paradigm's flexibility. Extensive experiments on both synthetic and real-world instances of the Travelling Salesman Problem (TSP) and Capacitated Vehicle Routing Problem (CVRP) demonstrate that L2C-Insert consistently achieves superior performance across various problem sizes.



Paperid:1561
Authors:Yuchen Zhang, Nikhil Keetha, Chenwei Lyu, Bhuvan Jhamb, Yutian Chen, Yuheng Qiu, Jay Karhade, Shreyas Jha, Yaoyu Hu, Deva Ramanan, Sebastian Scherer, Wenshan Wang
Abstract:
Abstract:Dense image correspondence is central to many applications, such as visual odometry, 3D reconstruction, object association, and re-identification. Historically, dense correspondence has been tackled separately for wide-baseline scenarios and optical flow estimation, despite the common goal of matching content between two images. In this paper, we develop a Unified Flow \& Matching model (UFM), which is trained on unified data for pixels that are co-visible in both source and target images. UFM uses a simple, generic transformer architecture that directly regresses the $(u,v)$ flow. It is easier to train and more accurate for large flows compared to the typical coarse-to-find cost volumes in prior work. UFM is 28\% more accurate than state-of-the-art flow methods (Unimatch), while also having 62\% less error and 6.7x faster than dense wide-baseline matchers (RoMa). UFM is the first to demonstrate that unified training can outperform specialized approaches across both domains. This result enables fast, general-purpose correspondence and opens new directions for multi-modal, long-range, and real-time correspondence tasks.



Paperid:1562
Authors:Yangning Li, Shaoshen Chen, Yinghui Li, Yankai Chen, Hai-Tao Zheng, Hui Wang, Wenhao Jiang, Philip S Yu
Abstract:
The quadratic complexity of self-attention limits Large Language Models (LLMs) in processing long contexts, a capability vital for many advanced applications. Context compression aims to mitigate this computational barrier while preserving essential semantic information. However, existing methods often falter: explicit methods can sacrifice local detail, while implicit ones may exhibit positional biases, struggle with information degradation, or fail to capture long-range semantic dependencies. We introduce AdmTree, a novel framework for adaptive, hierarchical context compression designed with a core focus on maintaining high semantic fidelity while keep efficiency. AdmTree dynamically segments input based on information density, employing gist tokens to summarize variable-length segments as leaves in a semantic binary tree. This structure, combined with a lightweight aggregation mechanism and a frozen backbone LLM (minimizing new trainable parameters), enables efficient hierarchical abstraction of the context. By effectively preserving fine-grained details alongside global semantic coherence, mitigating position bias, and adapting dynamically to content, AdmTree comprehensively preserves the semantic information of lengthy context.



Paperid:1563
Authors:Yonggang Zhang, Jun Nie, Xinmei Tian, Mingming Gong, Kun Zhang, Bo Han
Abstract:
The increasing realism of generated images has raised significant concerns about their potential misuse, necessitating robust detection methods. Current approaches mainly rely on training binary classifiers, which depend heavily on the quantity and quality of available generated images. In this work, we propose a novel framework that exploits geometric differences between the data manifolds of natural and generated images. To exploit this difference, we employ a pair of functions engineered to yield consistent outputs for natural images but divergent outputs for generated ones, leveraging the property that their gradients reside in mutually orthogonal subspaces. This design enables a simple yet effective detection method: an image is identified as generated if a transformation along its data manifold induces a significant change in the loss value of a self-supervised model pre-trained on natural images. Further more, to address diminishing manifold disparities in advanced generative models, we leverage normalizing flows to amplify detectable differences by extruding generated images away from the natural image manifold. Extensive experiments demonstrate the efficacy of this method.



Paperid:1564
Authors:Chengtai Li, Yuting He, Jianfeng Ren, Ruibin Bai, Yitian Zhao, Xudong Jiang
Abstract:
Abstract Visual Reasoning (AVR) entails discerning latent patterns in visual data and inferring underlying rules. Existing solutions often lack scalability and adaptability, as deep architectures tend to overfit training data, and static neural networks fail to dynamically capture diverse rules. To tackle the challenges, we propose a Dynamic and Scalable Reasoning Framework (DSRF) that greatly enhances the reasoning ability by widening the network instead of deepening it, and dynamically adjusting the reasoning network to better fit novel samples instead of a static network. Specifically, we design a Multi-View Reasoning Pyramid (MVRP) to capture complex rules through layered reasoning to focus features at each view on distinct combinations of attributes, widening the reasoning network to cover more attribute combinations analogous to complex reasoning rules. Additionally, we propose a Dynamic Domain-Contrast Prediction (DDCP) block to handle varying task-specific relationships dynamically by introducing a Gram matrix to model feature distributions, and a gate matrix to capture subtle domain differences between context and target features. Extensive experiments on six AVR tasks demonstrate DSRF's superior performance, achieving state-of-the-art results under various settings. Code will be released upon paper acceptance.



Paperid:1565
Authors:Shen Dong, Shaochen Xu, Pengfei He, Yige Li, Jiliang Tang, Tianming Liu, Hui Liu, Zhen Xiang
Abstract:
Agents powered by large language models (LLMs) have demonstrated strong capabilities in a wide range of complex, real-world applications.However, LLM agents with a compromised memory bank may easily produce harmful outputs when the past records retrieved for demonstration are malicious.In this paper, we propose a novel Memory INJection Attack, MINJA, without assuming that the attacker can directly modify the memory bank of the agent.The attacker injects malicious records into the memory bank by onlyinteracting with the agent via queries and output observations.These malicious records are designed to elicit a sequence of malicious reasoning steps corresponding to a different target query during the agent's execution of the victim user's query.Specifically, we introduce a sequence ofbridging stepsto link victim queries to the malicious reasoning steps.During the memory injection, we propose anindication promptthat guides the agent to autonomously generate similar bridging steps, with aprogressive shortening strategythat gradually removes the indication prompt, such that the malicious record will be easily retrieved when processing later victim queries.Our extensive experiments across diverse agents demonstrate the effectiveness of MINJA in compromising agent memory.With minimal requirements for execution, MINJA enables any user to influence agent memory, highlighting the risk.



Paperid:1566
Authors:Ofir Gaash, Kfir Y. Levy, Yair Carmon
Abstract:
Abstract:We study stochastic gradient descent (SGD) with gradient clipping on convex functions under a generalized smoothness assumption called $(L_0,L_1)$-smoothness. Using gradient clipping, we establish a high probability convergence rate that matches the SGD rate in the $L$ smooth case up to polylogarithmic factors and additive terms. We also propose a variation of adaptive SGD with gradient clipping, which achieves the same guarantee. We perform empirical experiments to examine our theory and algorithmic choices.



Paperid:1567
Authors:Fengxiang Wang, Yulin Wang, Di Wang, Haotian Wang, Mingshuo Chen, Hongzhen Wang, Haiyan Zhao, Yangang Sun, Shuo Wang, Long Lan, Wenjing Yang, Jing Zhang
Abstract:
Recent advances in self-supervised learning for Vision Transformers (ViTs) have fueled breakthroughs in remote sensing (RS) foundation models. However, the quadratic complexity of self-attention poses a significant barrier to scalability, particularly for large models and high-resolution images. While the linear-complexity Mamba architecture offers a promising alternative, existing RS applications of Mamba remain limited to supervised tasks on small, domain-specific datasets. To address these challenges, we propose RoMA, a framework that enables scalable self-supervised pretraining of Mamba-based RS foundation models using large-scale, diverse, unlabeled data. RoMA enhances scalability for high-resolution images through a tailored auto-regressive learning strategy, incorporating two key innovations: 1) a rotation-aware pretraining mechanism combining adaptive cropping with angular embeddings to handle sparsely distributed objects with arbitrary orientations, and 2) multi-scale token prediction objectives that address the extreme variations in object scales inherent to RS imagery. Systematic empirical studies validate that Mamba adheres to RS data and parameter scaling laws, with performance scaling reliably as model and data size increase. Furthermore, experiments across scene classification, object detection, and semantic segmentation tasks demonstrate that RoMA-pretrained Mamba models consistently outperform ViT-based counterparts in both accuracy and computational efficiency. The source code and pretrained models have be released in Anonymous Github.



Paperid:1568
Authors:Nick Erickson, Lennart Purucker, Andrej Tschalzev, David Holzmüller, Prateek Desai, David Salinas, Frank Hutter
Abstract:
With the growing popularity of deep learning and foundation models for tabular data, the need for standardized and reliable benchmarks is higher than ever. However, current benchmarks are static. Their design is not updated even if flaws are discovered, model versions are updated, or new models are released. To address this, we introduce TabArena, the first continuously maintained living tabular benchmarking system. To launch TabArena, we manually curate a representative collection of datasets and well-implemented models, conduct a large-scale benchmarking study to initialize a public leaderboard, and assemble a team of experienced maintainers. Our results highlight the influence of validation method and ensembling of hyperparameter configurations to benchmark models at their full potential. While gradient-boosted trees are still strong contenders on practical tabular datasets, we observe that deep learning methods have caught up under larger time budgets with ensembling. At the same time, foundation models excel on smaller datasets. Finally, we show that ensembles across models advance the state-of-the-art in tabular machine learning and investigate the contributions of individual models. We launch TabArena with a public leaderboard, reproducible code, and maintenance protocols to create a living benchmarking system.



Paperid:1569
Authors:Satvik Golechha, Adrià Garriga-Alonso
Abstract:
Abstract:Prior studies on deception in language-based AI agents typically assess whether the agent produces a false statement about a topic, or makes a binary choice prompted by a goal, rather than allowing open-ended deceptive behavior to emerge in pursuit of a longer-term goal.To fix this, we introduce $\textit{Among Us}$, a sandbox social deception game where LLM-agents exhibit long-term, open-ended deception as a consequence of the game objectives. While most benchmarks saturate quickly, $\textit{Among Us}$ can be expected to last much longer, because it is a multi-player game far from equilibrium.Using the sandbox, we evaluate $18$ proprietary and open-weight LLMs and uncover a general trend:models trained with RL are comparatively much better at producing deception than detecting it.We evaluate the effectiveness of methods to detect lying and deception: logistic regression on the activations and sparse autoencoders (SAEs). We find that probes trained on a dataset of ``pretend you're a dishonest model: $\dots$'' generalize extremely well out-of-distribution, consistently obtaining AUROCs over 95% even when evaluated just on the deceptive statement, without the chain of thought. We also find two SAE features that work well at deception detection but are unable to steer the model to lie less.We hope our open-sourced sandbox, game logs, and probes serve to anticipate and mitigate deceptive behavior and capabilities in language-based agents.



Paperid:1570
Authors:Shikun Feng, Bicheng Lin, Yuanhuan Mo, Yuyan Ni, Wenyu Zhu, Bowen Gao, Wei-Ying Ma, haitao li, Yanyan Lan
Abstract:
Flexible docking, which predicts the binding conformations of both proteins and small molecules by modeling their structural flexibility, plays a vital role in structure-based drug design. Although recent generative approaches, particularly diffusion-based models, have shown promising results, they require iterative sampling to generate candidate structures and depend on separate scoring functions for pose selection. This leads to an inefficient pipeline that is difficult to scale in real-world drug discovery workflows. To overcome these challenges, we introduce FIGRDock, a fast and accurate flexible docking framework that understands complicated interactions between molecules and proteins with a regression-based approach. FIGRDock leverages initial docking poses from conventional tools to distill interaction-aware distance patterns, which serve as explicit structural conditions to directly guide the prediction of the final protein-ligand complex via a regression model. This one-shot inference paradigm enables rapid and precise pose prediction without reliance on multi-step sampling or external scoring stages. Experimental results show that FIGRDock achieves up to 100× faster inference than diffusion-based docking methods, while consistently surpassing them in accuracy across standard benchmarks. These results suggest that FIGRDock has the potential to offer a scalable and efficient solution for flexible docking, advancing the pace of structure-based drug discovery.



Paperid:1571
Authors:Wenze Liu, Xiangyu Yue
Abstract:
Abstract:To accelerate diffusion model inference, numerical solvers perform poorly at extremely small steps, while distillation techniques often introduce complexity and instability. This work presents an intermediate strategy, balancing performance and cost, by learning ODE integration using loss functions derived from the derivative-integral relationship, inspired by Monte Carlo integration and Picard iteration. From a geometric perspective, the losses operate by gradually extending the tangent to the secant, thus are named as secant losses. The secant losses can rapidly convert (via fine-tuning or distillation) a pretrained diffusion model into its secant version. In our experiments, the secant version of EDM achieves a $10$-step FID of $2.14$ on CIFAR-10, while the secant version of SiT-XL/2 attains a $4$-step FID of $2.27$ and an $8$-step FID of $1.96$ on ImageNet-$256\times256$. Code will be available.



Paperid:1572
Authors:Liyuan Mao, Haoran Xu, Amy Zhang, Weinan Zhang, Chenjia Bai
Abstract:
Obtaining a generalizable reward model is crucial in Reinforcement Learning from Human Feedback (RLHF) as it enables correctly evaluating unseen prompt-response pairs. However, existing reward models lack this ability, as they are typically trained by increasing the reward gap between chosen and rejected responses, while overlooking the prompts that the responses are conditioned on. Consequently, when the trained reward model is evaluated on prompt-response pairs that lie outside the data distribution, neglecting the effect of prompts may result in poor generalization of the reward model. To address this issue, we decompose the reward value into two independent components: prompt-free reward and prompt-related reward. Prompt-free reward represents the evaluation that is determined only by responses, while the prompt-related reward reflects the reward that derives from both the prompt and the response. We extract these two components from an information theoretical perspective, which requires no extra models.. Subsequently, we propose a new reward learning algorithm by prioritizing data samples based on their prompt-free reward values. Through toy examples, we demonstrate that the extracted prompt-free and prompt-related rewards effectively characterize two parts of the reward model. Further, standard evaluations show that our method improves both the alignment performance and the generalization capability of the reward model.



Paperid:1573
Authors:Mert Bulent Sariyildiz, Philippe Weinzaepfel, Guillaume Bono, Gianluca Monaci, Christian Wolf
Abstract:
One key aspect of spatially aware robots is the ability to "find their bearings", ie. to correctly situate themselves or previously seen spaces. In this work, we focus on this particular scenario of continuous robotics operations, where information observed before an actual episode start is exploited to optimize efficiency. We introduce a new model, "Kinaema" and agent, capable of integrating a stream of visual observations while moving in a potentially large scene, and upon request, processing a query image and predicting the relative position of the shown space with respect to its current position. Our model does not explicitly store an observation history, therefore does not have hard constraints on context length. It maintains an implicit latent memory, which is updated by a transformer in a recurrent way, compressing the history of sensor readings into a compact representation. We evaluate the impact of this model in a new downstream task we call "Mem-Nav", targeting continuous robotics operations. We show that our large-capacity recurrent model maintains a useful representation of the scene, navigates to goals observed before the actual episode start, and is computationally efficient, in particular compared to classical transformers with attention over an observation history.



Paperid:1574
Authors:Saleh Momeni, Changnan Xiao, Bing Liu
Abstract:
This paper studies the class-incremental learning (CIL) setting of continual learning. CIL aims to learn a sequence of tasks, where each task consists of a set of classes. Traditional CIL methods does not use a pre-trained model (PTM) and suffer from catastrophic forgetting (CF) due to their need to incrementally learn both feature representations and the classifier. The incorporation of PTMs into CIL has led to the development of computationally efficient methods that treat the PTM as a feature extractor paired with analytical classifiers. These methods often achieve state-of-the-art performance in CIL. However, they still face a major limitation: the inability to continually adapt or update feature representations incrementally to best suit the specific CIL tasks, leading to suboptimal performance. To overcome this, we propose ACP (Analytical Contrastive Projection), a novel method that retains the computational efficiency and stability of analytical classifiers while enabling incremental feature adaptation without gradient-based training. Our experiments demonstrate that ACP not only outperforms strong baselines but also matches the accuracy of joint training, which is regarded as the upper bound of CIL.



Paperid:1575
Authors:Xinran Gu, Kaifeng Lyu, Jiazheng Li, Jingzhao Zhang
Abstract:
Large Language Models (LLMs) are typically trained on data mixtures: most data come from web scrapes, while a small portion is curated from high-quality sources with dense domain-specific knowledge.In this paper, we show that when training LLMs on such data mixtures, knowledge acquisition from knowledge-dense datasets—unlike training exclusively on knowledge-dense data—does not always follow a smooth scaling law but can exhibit phase transitions with respect to the mixing ratio and model size. Through controlled experiments on a synthetic biography dataset mixed with web-scraped data, we demonstrate that: (1) as we increase the model size to a critical value, the model suddenly transitions from memorizing very few to most of the biographies; (2) below a critical mixing ratio, the model memorizes almost nothing even with extensive training, but beyond this threshold, it rapidly memorizes more biographies. We attribute these phase transitions to a capacity allocation phenomenon: a model with bounded capacity must act like a knapsack problem solver to minimize the overall test loss, and the optimal allocation across datasets can change discontinuously as the model size or mixing ratio varies. We formalize this intuition in an information-theoretic framework and reveal that these phase transitions are predictable, with the critical mixing ratio following a power-law relationship with the model size. Our findings highlight a concrete case where a good mixing recipe for large models may not be optimal for small models, and vice versa.



Paperid:1576
Authors:Baiting Chen, Tong Zhu, Jiale Han, Lexin Li, Gang Li, Xiaowu Dai
Abstract:
Large Language Models (LLMs) have demonstrated strong generative capabilities but remain prone to inconsistencies and hallucinations. We introduce Peer Elicitation Games (PEG), a training-free, game-theoretic framework for aligning LLMs through a peer elicitation mechanism involving a generator and multiple discriminators instantiated from distinct base models. Discriminators interact in a peer evaluation setting, where rewards are computed using a determinant-based mutual information score that provably incentivizes truthful reporting without requiring ground-truth labels. We establish theoretical guarantees showing that each agent, via online learning, achieves sublinear regret in the sense their cumulative performance approaches that of the best fixed truthful strategy in hindsight. Moreover, we prove last-iterate convergence to a truthful Nash equilibrium, ensuring that the actual policies used by agents converge to stable and truthful behavior over time. Empirical evaluations across multiple benchmarks demonstrate significant improvements in factual accuracy. These results position PEG as a practical approach for eliciting truthful behavior from LLMs without supervision or fine-tuning.



Paperid:1577
Authors:Wenyi Xiao, Leilei Gan
Abstract:
When applying reinforcement learning—typically through GRPO—to large vision-language model reasoning struggles to effectively scale reasoning length or generates verbose outputs across all tasks with only marginal gains in accuracy.To address this issue, we present \textbf{FAST-GRPO}, a variant of GRPO that dynamically adapts reasoning depth based on question characteristics. Through empirical analysis, we establish the feasibility of fast-slow thinking in LVLMs by investigating how response length and data distribution affect performance.Inspired by these observations, we introduces two complementary metrics to estimate the difficulty of the questions, guiding the model to determine when fast or slow thinking is more appropriate. Next, we incorporate adaptive length-based rewards and difficulty-aware KL divergence into the GRPO algorithm.Experiments across seven reasoning benchmarks demonstrate that FAST achieves state-of-the-art accuracy with over 10\% relative improvement compared to the base model, while reducing token usage by 32.7-67.3\% compared to previous slow-thinking approaches, effectively balancing reasoning length and accuracy.



Paperid:1578
Authors:Sarthak Chakraborty, Xuchao Zhang, Chetan Bansal, Indranil Gupta, Suman Nath
Abstract:
Abstract:Large Language Models (LLMs) are increasingly being used to plan, reason, and execute tasks across various scenarios. Use cases like repeatable workflows, chatbots, and AI agents often involve recurring tasks and tend to reuse similar prompts when interacting with the LLM. This opens up opportunities for caching. With structurally similar prompts that differ in subtle yet important ways, which are also reflected in their corresponding responses, exact prompt matching fails, while semantic caching techniques may return cached responses that are incorrect since they ignore these variations. To address this, we introduce GenCache, a generative cache that produces variation-aware responses for structurally similar prompts. It identifies and reuses the pattern in which responses are generated for structurally similar prompts for new requests. We show that GenCache achieves an 83\% cache hit rate with minimal negative hits on datasets without prompt repetition. In agentic workflows, it improves cache hit rate by $\sim$20\% and reduces end-to-end execution latency by $\sim$34\% for one workflow compared to standard prompt matching.



Paperid:1579
Authors:Gongwei Chen, Lirong Jie, Lexiao Zou, Weili Guan, Miao Zhang, Liqiang Nie
Abstract:
Benefiting from the availability of extensive navigation trajectories, both manually and automatically annotated, current graphical user interface (GUI) agents have achieved remarkable advancements in performance. However, these annotated datasets often contain substantial noise, which impedes effective agent training and underscores the necessity for rigorous trajectory quality assessment. In contrast to existing prompting-based evaluators that rely on proprietary multimodal large language models (MLLMs), we propose an Uncertainty-aware Reinforced Self-Training (URST) framework to train lightweight MLLMs for efficient and reliable trajectory evaluation. URST iteratively fine-tunes MLLMs using their own generated thoughts and judgments to enable self-improvement, while its uncertainty-aware sampling strategy ensures the selection of the most informative training examples. To further enhance reasoning and judgment capabilities, we propose a simplified group policy optimization approach that effectively leverages diverse positive and negative samples for evaluator learning. Our evaluator demonstrates superior judgment performance across both in-domain and out-of-domain datasets. When used to filter navigation datasets, it consistently leads to performance improvements in training GUI agents.



Paperid:1580
Authors:Floris Holstege, Shauli Ravfogel, Bram Wouters
Abstract:
Modern neural networks often encode unwanted concepts alongside task-relevant information, leading to fairness and interpretability concerns. Existing post-hoc approaches can remove undesired concepts but often degrade useful signals. We introduce SPLICE—Simultaneous Projection for LInear concept removal and Covariance prEservation—which eliminates sensitive concepts from representations while exactly preserving their covariance with a target label. SPLICE achieves this via an oblique projection that ``splices out'' the unwanted direction yet protects important label correlations. Theoretically, it is the unique solution that removes linear concept predictability and maintains target covariance with minimal embedding distortion. Empirically, SPLICE outperforms baselines on benchmarks such as Bias in Bios and Winobias, removing protected attributes while minimally damaging main-task information.



Paperid:1581
Authors:Yingcong Li, Xiangyu Chang, Muti Kara, Xiaofeng Liu, Amit Roy-Chowdhury, Samet Oymak
Abstract:
Abstract:Recent research shows that in-context learning (ICL) can be effective even in settings where demonstrations have missing or incorrect labels. This motivates a deeper understanding of how sequence models leverage unlabeled data. We consider a canonical setting where the in-context demonstrations are drawn according to a binary Gaussian mixture model (GMM) and a certain fraction of the demonstrations have missing labels. We provide a comprehensive theoretical study to show that: (1) The loss landscape of one-layer linear attention learns the optimal fully-supervised learner but it completely fails to leverage the unlabeled data. (2) Multilayer as well as looped transformers can effectively leverage unlabeled data by implicitly constructing estimators of the form $\sum_{i\ge 0} a_i (X^\top X)^iX^\top y$ with $X$ and $y$ denoting features and visible labels. We shed light on the class of polynomials that can be expressed as a function of depth/looping and draw connections to iterative pseudo-labeling. Building on these insights and the importance of depth, we propose looping off-the-shelf tabular foundation models, such as TabPFN or TabICL, to enhance their semi-supervision capabilities. Extensive evaluations on real-world datasets reveal that our method significantly improves the semisupervised tabular learning performance over the standard single pass inference.



Paperid:1582
Authors:Michael Arbel, David Salinas, Frank Hutter
Abstract:
Recent foundational models for tabular data, such as TabPFN, excel at adapting to new tasks via in-context learning but remain constrained to a fixed, pre-defined number of target dimensions—often necessitating costly ensembling strategies. We trace this constraint to a deeper architectural shortcoming: these models lack target-equivariance, so that permuting target-dimension orderings alters their predictions. This deficiency gives rise to an irreducible “equivariance gap,” an error term that introduces instability in predictions. We eliminate this gap by designing a fully target-equivariant architecture—ensuring permutation invariance via equivariant encoders, decoders, and a bi-attention mechanism. Empirical evaluation on standard classification benchmarks shows that, on datasets with more classes than those seen during pre-training, our model matches or surpasses existing methods while incurring lower computational overhead.



Paperid:1583
Authors:Yazheng Liu, Xi Zhang, Sihong Xie, Hui Xiong
Abstract:
Explaining graph neural networks (GNNs) is a key approach to improve the trustworthiness of GNN in high-stakes applications, such as finance and healthcare. However, existing methods are vulnerable to perturbations, raising concerns about explanation reliability. Prior methods enhance explanation robustness using model retraining or explanation ensemble, with certain weaknesses. Retraining leads to models that are different from the original target model and misleading explanations, while ensemble can produce contradictory results due to different inputs or models. To improve explanation robustness without the above weaknesses, we take an unexplored route and exploit the two edge geometry properties curvature and resistance to enhance explanation robustness. We are the first to prove that these geometric notions can be used to bound explanation robustness. We design a general optimization algorithm to incorporate these geometric properties into a wide spectrum of base GNN explanation methods to enhance the robustness of base explanations. We empirically show that our method outperforms six base explanation methods in robustness across nine datasets spanning node classification, link prediction, and graph classification tasks, improving fidelity in 80\% of the cases and achieving up to a 10\% relative improvement in robust performance.



Paperid:1584
Authors:Yun Hua, Shang Gao, Wenhao Li, Haosheng Chen, Bo Jin, Xiangfeng Wang, Jun Luo, Hongyuan Zha
Abstract:
Multi-agent reinforcement learning (MARL) has emerged as a powerful framework for modeling autonomous agents that independently optimize their individual objectives. However, in mixed-motive MARL environments, rational self-interested behaviors often lead to collectively suboptimal outcomes situations commonly referred to as social dilemmas.A key challenge in addressing social dilemmas lies in accurately quantifying and representing them in a numerical form that captures how self-interested agent behaviors impact social welfare.To address this challenge, \textit{externalities} in the economic concept is adopted and extended to denote the unaccounted-for impact of one agent's actions on others, as a means to rigorously quantify social dilemmas.Based on this measurement, a novel method, \textbf{L}earning \textbf{O}ptimal \textbf{P}igovian \textbf{T}ax (\textbf{LOPT}) is proposed. Inspired by Pigovian taxes, which are designed to internalize externalities by imposing cost on negative societal impacts, LOPT employs an auxiliary tax agent that learns an optimal Pigovian tax policy to reshape individual rewards aligned with social welfare, thereby promoting agent coordination and mitigating social dilemmas. We support LOPT with theoretical analysis and validate it on standard MARL benchmarks, including Escape Room and Cleanup. Results show that by effectively internalizing externalities that quantify social dilemmas, LOPT aligns individual objectives with collective goals, significantly improving social welfare over state-of-the-art baselines.



Paperid:1585
Authors:Yuxin Wen, Yangsibo Huang, Tom Goldstein, Ravi Kumar, Badih Ghazi, Chiyuan Zhang
Abstract:
Understanding what and how neural networks memorize during training is crucial, both from the perspective of unintentional memorization of potentially sensitive information and from the standpoint of effective knowledge acquisition for real-world, knowledge-intensive tasks. While previous studies primarily investigate memorization within a single modality, such as text memorization in large language models or image memorization in diffusion models, unified multimodal models are becoming increasingly prevalent in practical applications. In this work, we focus on the unique characteristics of cross-modality memorization and conduct a systematic study centered on vision-language models. To facilitate controlled experiments, we first introduce a synthetic persona dataset comprising diverse synthetic person images and textual descriptions. We quantify factual knowledge memorization and cross-modal transferability by training models on a single modality and evaluating their performance in the other. Our results reveal that facts learned in one modality transfer to the other, but a significant gap exists between recalling information in the source and target modalities. Furthermore, we observe that this gap exists across various scenarios, including more capable models, machine unlearning, and the multi-hop case. At the end, we propose a baseline method to mitigate this challenge. We hope our study can inspire future research on developing more robust multimodal learning techniques to enhance cross-modal transferability.



Paperid:1586
Authors:Ningning CHEN, Weicai Ye, Ying Jiang
Abstract:
Abstract:We introduce HBLLM, a wavelet-enhanced high-fidelity $1$-bit post-training quantization method for Large Language Models (LLMs). By leveraging Haar wavelet transforms to enhance expressive capacity through frequency decomposition, HBLLM significantly improves quantization fidelity while maintaining minimal overhead. This approach features two innovative structure-aware grouping strategies: (1) frequency-aware multi-parameter intra-row grouping and (2) $\ell_2$-norm-based saliency-driven column selection. For non-salient weights, a shared mean is employed across quantization groups within each frequency band to optimize storage efficiency. Experiments conducted on the OPT and LLaMA models demonstrate that HBLLM achieves state-of-the-art performance in $1$-bit quantization, attaining a perplexity of $6.71$ perplexity on LLaMA$2$-$13$B with an average weight storage of only $1.08$ bits.



Paperid:1587
Authors:Weiyang Zhang, Xinyang Chen, Xiucheng Li, Kehai Chen, Weili Guan, Liqiang Nie
Abstract:
Abstract:With the increasing number of time series pre-trained models, designing transferability evaluation metrics for time series has become an urgent problem to address. While transferability evaluation has been extensively studied in computer vision, we aim to address a critical gap by developing tailored metrics for time series analysis. In this paper, we introduce TEMPLATE, a transferability estimation framework specifically tailored for versatile time series analysis, comprising three complementary metrics: (1) Dependency Learning Score quantifies a model’s capacity to capture temporal dependencies. (2) Pattern Learning Score evaluates the representation quality in extracting discriminative temporal patterns. (3) Task Adaptation Score assesses cross-task generalization capability, enabling versatile time series analysis. TEMPLATE presents a versatile framework compatible with both classification and regression paradigms. Through comprehensive benchmarking across five distinct downstream tasks, our method demonstrates superior capability in identifying optimal pre-trained models from heterogeneous model pools for transfer learning. Compared to the state-of-the-art method ETran, our approach improves the weighted Kendall's $\tau_w$ across five downstream tasks by 35\%. The code is available at https://anonymous.4open.science/r/TEMPLATE-A0AA/.



Paperid:1588
Authors:Xiaohan Zou, Jian Kang, George Kesidis, Lu Lin
Abstract:
Recent studies reveal that vision-language models (VLMs) become more susceptible to harmful requests and jailbreak attacks after integrating the vision modality, exhibiting greater vulnerability than their text-only LLM backbones. To uncover the root cause of this phenomenon, we conduct an in-depth analysis and identify a key issue: multimodal inputs introduce an modality-induced activation shift toward a “safer” direction compared to their text-only counterparts, leading VLMs to systematically overestimate the safety of harmful inputs. We refer to this issue as safety perception distortion. To mitigate such distortion, we propose Activation Shift Disentanglement and Calibration (ShiftDC), a training-free method that decomposes and calibrates the modality-induced activation shift to reduce its impact on safety. By isolating and removing the safety-relevant component, ShiftDC restores the inherent safety alignment of the LLM backbone while preserving the vision-language capabilities of VLMs. Experiments demonstrate that ShiftDC significantly enhances safety alignment without impairing model utility.



Paperid:1589
Authors:Tianyue Zhang, Lucas Maes, Alan Milligan, Alexia Jolicoeur-Martineau, Ioannis Mitliagkas, Damien Scieur, Simon Lacoste-Julien, Charles Guille-Escuret
Abstract:
Despite its widespread adoption, Adam's advantage over Stochastic Gradient Descent (SGD) lacks a comprehensive theoretical explanation. This paper investigates Adam's sensitivity to rotations of the parameter space. We observe that Adam's performance in training transformers degrades under random rotations of the parameter space, indicating a crucial sensitivity to the choice of basis in practice. This reveals that conventional rotation-invariant assumptions are insufficient to capture Adam's advantages theoretically. To better understand the rotation-dependent properties that benefit Adam, we also identify structured rotations that preserve or even enhance its empirical performance. We then examine the rotation-dependent assumptions in the literature and find that they fall short in explaining Adam's behavior across various rotation types. In contrast, we verify the orthogonality of the update as a promising indicator of Adam’s basis sensitivity, suggesting it may be the key quantity for developing rotation-dependent theoretical frameworks that better explain its empirical success.



Paperid:1590
Authors:Deeksha Adil, Brian Bullins, Aaron Sidford, Chenyi Zhang
Abstract:
Abstract:We develop optimization methods which offer new trade-offs between the number of gradient and Hessian computations needed to compute the critical point of a non-convex function. We provide a method that for a twice-differentiable $f\colon \mathbb{R}^d \rightarrow \mathbb{R}$ with $L_2$-Lipschitz Hessian, and input initial point with $\Delta$-bounded sub-optimality and sufficiently small $\epsilon > 0$ outputs an $\epsilon$-critical point, i.e., a point $x$ such that $\|\nabla f(x)\| \leq \epsilon$, using $\tilde{O}(\Delta L_2^{1/4} n_H^{-1/2}\epsilon^{-9/4})$ queries to a gradient oracle and $n_H$ queries to a Hessian oracle. As a consequence, we obtain an improved gradient query complexity of $\tilde{O}(d^{1/3}L_2^{1/2}\Delta\epsilon^{-3/2})$ in the case of bounded dimension and of $\tilde{O}(\Delta^{3/2} L_2^{3/4}\epsilon^{-9/4})$ in the case where we are allowed only a single Hessian query. We obtain these results through a more general algorithm which can handle approximate Hessian computations and recovers known prior state-of-the-art bounds of computing an $\epsilon$-critical point, under the additional assumption that $f$ has an $L_1$-Lipschitz gradient, with $O(\Delta L_2^{1/4}\epsilon^{-7/4})$-gradient queries.



Paperid:1591
Authors:Till Freihaut, Giorgia Ramponi
Abstract:
Multi-agent inverse reinforcement learning (MAIRL) aims to recover agent reward functions from expert demonstrations. We characterize the feasible reward set in Markov games, identifying all reward functions that rationalize a given equilibrium. However, equilibrium-based observations are often ambiguous: a single Nash equilibrium can correspond to many reward structures, potentially changing the game's nature in multi-agent systems. We address this by introducing entropy-regularized Markov games, which yield a unique equilibrium while preserving strategic incentives. For this setting, we provide a sample complexity analysis detailing how errors affect learned policy performance. Our work establishes theoretical foundations and practical insights for MAIRL.



Paperid:1592
Authors:Jingyi Yang, Shuai Shao, Dongrui Liu, Jing Shao
Abstract:
With the rapid development of multimodal large language models (MLLMs), they are increasingly deployed as autonomous computer-use agents capable of accomplishing complex computer tasks. However, a pressing issue arises: Can the safety risk principles designed and aligned for general MLLMs in dialogue scenarios be effectively transferred to real-world computer-use scenarios? Existing research on evaluating the safety risks of MLLM-based computer-use agents suffers from several limitations: it either lacks realistic interactive environments, or narrowly focuses on one or a few specific risk types. These limitations ignore the complexity, variability, and diversity of real-world environments, thereby restricting comprehensive risk evaluation for computer-use agents.To this end, we introduceRiOSWorld, a benchmark designed to evaluate the potential risks of MLLM-based agents during real-world computer manipulations. Our benchmark includes 492 risky tasks spanning various computer applications, involving web, social media, multimedia, os, email, and office software. We categorize these risks into two major classes based on their risk source: (i) User-originated risks and (ii) Environmental risks. For the evaluation, we evaluate safety risks from two perspectives: (i) Risk goal intention and (ii) Risk goal completion. Extensive experiments with multimodal agents onRiOSWorlddemonstrate that current computer-use agents confront significant safety risks in real-world scenarios. Our findings highlight the necessity and urgency of safety alignment for computer-use agents in real-world computer manipulation, providing valuable insights for developing trustworthy computer-use agents.



Paperid:1593
Authors:Shahaf Bassan, Michal Moshkovitz, Guy Katz
Abstract:
Abstract:Generalized Additive Models (GAMs) are commonly considered *interpretable* within the ML community, as their structure makes the relationship between inputs and outputs relatively understandable. Therefore, it may seem natural to hypothesize that obtaining meaningful explanations for GAMs could be performed efficiently and would not be computationally infeasible. In this work, we challenge this hypothesis by analyzing the *computational complexity* of generating different explanations for various forms of GAMs across multiple contexts. Our analysis reveals a surprisingly diverse landscape of both positive and negative complexity outcomes. Particularly, under standard complexity assumptions such as P$\neq$NP, we establish several key findings: (1) in stark contrast to many other common ML models, the complexity of generating explanations for GAMs is heavily influenced by the structure of the input space; (2) the complexity of explaining GAMs varies significantly with the types of component models used - but interestingly, these differences only emerge under specific input domain settings; (3) significant complexity distinctions appear for obtaining explanations in regression tasks versus classification tasks in GAMs; and (4) expressing complex models like neural networks additively (e.g., as neural additive models) can make them easier to explain, though interestingly, this benefit appears only for certain explanation methods and input domains. Collectively, these results shed light on the feasibility of computing diverse explanations for GAMs, offering a rigorous theoretical picture of the conditions under which such computations are possible or provably hard.



Paperid:1594
Authors:Kaiqi Guan, Wenke Huang, Xianda Guo, Yueyang Yuan, Bin Yang, Mang Ye
Abstract:
Federated Learning is a promising technique that enables collaborative machine learning while preserving participant privacy. With respect to multi-party collaboration, achieving performance fairness acts as a critical challenge in federated systems. Existing explorations mainly focus on considering all parameter-wise fairness and consistently protecting weak clients to achieve performance fairness in federation. However, these approaches neglect two critical issues. 1) Parameter Redundancy: Redundant parameters that are unnecessary for fairness training may conflict with critical parameters update, thereby leading to performance degradation. 2) Persistent Protection: Current fairness mechanisms persistently enhance weak clients throughout the entire training cycle, hindering global optimization and causing lower performance alongside unfairness. To address these, we propose a strategy with two key components: First, parameter adjustment with mask and rescale which discarding redundant parameter and highlight critical ones, preserving key parameter updates and decrease conflict. Second, we observe that the federated training process exhibits distinct characteristics across different phases. We propose a dynamic aggregation strategy that adaptively weights clients based on local update directions and performance variations. Empirical results on single-domain and cross-domain scenarios demonstrate the effectiveness of the proposed solution and the efficiency of crucial modules. The code is available at https://anonymous.4open.science/r/FedPW.



Paperid:1595
Authors:Louis Allain, Sébastien Da Veiga, Brian Staber
Abstract:
Conformal Prediction (CP) is a popular framework for constructing prediction bands with valid coverage in finite samples, while being free of any distributional assumption. A well-known limitation of conformal prediction is the lack of adaptivity, although several works introduced practically efficient alternate procedures. In this work, we build upon recent ideas that rely on recasting the CP problem as a statistical learning problem, directly targeting coverage and adaptivity. This statistical learning problem is based on reproducible kernel Hilbert spaces (RKHS) and kernel sum-of-squares (SoS) methods. First, we extend previous results with a general representer theorem and exhibit the dual formulation of the learning problem. Crucially, such dual formulation can be solved efficiently by accelerated gradient methods with several hundreds or thousands of samples, unlike previous strategies based on off-the-shelf semidefinite programming algorithms. Second, we introduce a new hyperparameter tuning strategy tailored specifically to target adaptivity through bounds on test-conditional coverage. This strategy, based on the Hilbert-Schmidt Independence Criterion (HSIC), is introduced here to tune kernel lengthscales in our framework, but has broader applicability since it could be used in any CP algorithm where the score function is learned. Finally, extensive experiments are conducted to show how our method compares to related work. All figures can be reproduced with the accompanying code.



Paperid:1596
Authors:Felix Draxler, Yang Meng, Kai Nelson, Lukas Laskowski, Yibo Yang, Theofanis Karaletsos, Stephan Mandt
Abstract:
Marked Temporal Point Processes (MTPPs) model sequences of irregularly timed events accompanied with metadata (marks), which occur in various domains such as medicine, finance, and remote sensing. We extend MTPPs in two ways: First, we support variable-length, mixed-type (i.e., both discrete and continuous) marks, such as those found in electronic health records. Second, we treat conditioning as a first-class modeling principle by using our flexible MTPPs as the structured output of a regression task such as the detection of events in an input time series. Our model uses an autoregressive Transformer to directly model the joint distribution of event times and marks, employing flexible normalizing flow models for continuous-valued marks. Our intensity-free formulation avoids numerical integration and naturally supports complex, conditional event sequences. Empirically, we find that our model excels both at tasks with discrete-only and mixed-type marks, and that the gained flexibility improves prediction quality.



Paperid:1597
Authors:Wenlong Jiao, Binglong Li, Wei Shang, Ping Wang, Dongwei Ren
Abstract:
Image deblurring plays a crucial role in enhancing visual clarity across various applications. Although most deep learning approaches primarily focus on sRGB images, which inherently lose critical information during the image signal processing pipeline, RAW images, being unprocessed and linear, possess superior restoration potential but remain underexplored. Deblurring RAW images presents unique challenges, particularly in handling frequency-dependent blur while maintaining computational efficiency. To address these issues, we propose Frequency Enhanced Network (FrENet), a framework specifically designed for RAW-to-RAW deblurring that operates directly in the frequency domain. We introduce a novel Adaptive Frequency Positional Modulation module, which dynamically adjusts frequency components according to their spectral positions, thereby enabling precise control over the deblurring process. Additionally, frequency domain skip connections are adopted to further preserve high-frequency details. Experimental results demonstrate that FrENet surpasses state-of-the-art deblurring methods in RAW image deblurring, achieving significantly better restoration quality while maintaining high efficiency in terms of reduced MACs. Furthermore, FrENet's adaptability enables it to be extended to sRGB images, where it delivers comparable or superior performance compared to methods specifically designed for sRGB data. The source code will be publicly available.



Paperid:1598
Authors:Alexandru Oarga, Yilun Du
Abstract:
Generalization is a key challenge in machine learning, specifically in reasoning tasks, where models are expected to solve problems more complex than those encountered during training. Existing approaches typically train reasoning models in an end-to-end fashion, directly mapping input instances to solutions. While this allows models to learn useful heuristics from data, it often results in limited generalization beyond the training distribution. In this work, we propose a novel approach to reasoning generalization by learning energy landscapes over the solution spaces of smaller, more tractable subproblems. At test time, we construct a global energy landscape for a given problem by combining the energy functions of multiple subproblems. This compositional approach enables the incorporation of additional constraints during inference, allowing the construction of energy landscapes for problems of increasing difficulty. To improve the sample quality from this newly constructed energy landscape, we introduce Parallel Energy Minimization (PEM). We evaluate our approach on a wide set of reasoning problems. Our method outperforms existing state-of-the-art methods, demonstrating its ability to generalize to larger and more complex problems.



Paperid:1599
Authors:Ching-Chia Kao, Chia-Mu Yu, Chun-Shien Lu, Chu-Song Chen
Abstract:
Large Language Models (LLMs) are increasingly adopted in high-stakes scenarios, yet their safety mechanisms often remain fragile. Simple jailbreak prompts or even benign fine-tuning can bypass internal safeguards, underscoring the need to understand the failure modes of current safety strategies. Recent findings suggest that vulnerabilities emerge when alignment is confined to only the initial output tokens. To address this, we introduce the notion of safety depth, a designated output position where the model refuses to generate harmful content. While deeper alignment appears promising, identifying the optimal safety depth remains an open and underexplored challenge.We leverage the equivalence between autoregressive language models and Markov chains to derive the first theoretical result on identifying the optimal safety depth. To reach this safety depth effectively, we propose a cyclic group augmentation strategy that improves safety scores across six LLMs. In addition, we uncover a critical interaction between safety depth and ensemble width, demonstrating that larger ensembles can offset shallower alignments. These results suggest that test-time computation, often overlooked in safety alignment, can play a key role. Our approach provides actionable insights for building safer LLMs.



Paperid:1600
Authors:Rishi Singhal, Jung-Eun Kim
Abstract:
Layer Normalization (LayerNorm) is one of the fundamental components in transformers that stabilizes training and improves optimization. In recent times, Pre-LayerNorm transformers have become the preferred choice over Post-LayerNorm transformers due to their stable gradient flow. However, the impact of LayerNorm on learning and memorization across these architectures remains unclear. In this work, we investigate how LayerNorm influences memorization and learning for Pre- and Post-LayerNorm transformers. We identify that LayerNorm serves as a key factor for stable learning in Pre-LayerNorm transformers, while in Post-LayerNorm transformers, it impacts memorization. Our analysis reveals that eliminating LayerNorm parameters in Pre-LayerNorm models exacerbates memorization and destabilizes learning, while in Post-LayerNorm models, it effectively mitigates memorization by restoring genuine labels. We further precisely identify that early layers LayerNorm are the most critical over middle/later layers and their influence varies across Pre and Post LayerNorm models. We have validated it through 13 models across 6 Vision and Language datasets. These insights shed new light on the role of LayerNorm in shaping memorization and learning in transformers.



Paperid:1601
Authors:Xuheng Li, Quanquan Gu
Abstract:
Abstract:Variance-dependent regret bounds have received increasing attention in recent studies on contextual bandits. However, most of these studies are focused on upper confidence bound (UCB)-based bandit algorithms, while sampling based bandit algorithms such as Thompson sampling are still understudied. The only exception is the `LinVDTS` algorithm (Xu et al., 2023), which is limited to linear reward function and its regret bound is not optimal with respect to the model dimension. In this paper, we present `FGTSVA`, a variance-aware Thompson Sampling algorithm for contextual bandits with general reward function with optimal regret bound. At the core of our analysis is an extension of the decoupling coefficient, a technique commonly used in the analysis of Feel-good Thompson sampling (FGTS) that reflects the complexity of the model space. With the new decoupling coefficient denoted by $\mathrm{dc}$, `FGTS-VA` achieves the regret of $\tilde{\mathcal{O}}(\sqrt{\mathrm{dc}\cdot\log|\mathcal{F}|\sum_{t=1}^T\sigma_t^2}+\mathrm{dc})$, where $|\mathcal{F}|$ is the size of the model space, $T$ is the total number of rounds, and $\sigma_t^2$ is the subgaussian norm of the noise (e.g., variance when the noise is Gaussian) at round $t$. In the setting of contextual linear bandits, the regret bound of `FGTSVA` matches that of UCB-based algorithms using weighted linear regression (Zhou and Gu, 2022).



Paperid:1602
Authors:Sahar Rajabi, Nayeema Nonta, Sirisha Rambhatla
Abstract:
Training large language models (LLMs) is highly resource-intensive due to their massive number of parameters and the overhead of optimizer states. While recent work has aimed to reduce memory consumption, such efforts often entail trade-offs among memory efficiency, training time, and model performance. Yet, true democratization of LLMs requires simultaneous progress across all three dimensions. To this end, we propose \mymethod that leverages Grassmannian gradient subspace tracking combined with projection-aware optimizers, enabling Adam’s internal statistics to adapt to changes in the optimization subspace. Additionally, employing recovery scaling, a technique that restores information lost through low-rank projections, further enhances model performance. Our method demonstrates SOTA convergence by exploiting Grassmannian geometry and achieves lowest evaluation loss,outperforming the current SOTA while reducing pre-training wall time by 43%and maintaining the memory footprint on a 1B-parameter Llama model.



Paperid:1603
Authors:Jongha (Jon) Ryu, Samuel Zhou, Gregory Wornell
Abstract:
Spectral decomposition of linear operators plays a central role in many areas of machine learning and scientific computing. Recent work has explored training neural networks to approximate eigenfunctions of such operators, enabling scalable approaches to representation learning, dynamical systems, and partial differential equations (PDEs). In this paper, we revisit a classical optimization framework from the computational physics literature known as theorbital minimization method(OMM), originally proposed in the 1990s for solving eigenvalue problems in computational chemistry. We provide a simple linear algebraic proof of the consistency of the OMM objective, and reveal connections between this method and several ideas that have appeared independently across different domains. Our primary goal is to justify its broader applicability in modern learning pipelines. We adapt this framework to train neural networks to decompose positive semidefinite operators, and demonstrate its practical advantages across a range of benchmark tasks. Our results highlight how revisiting classical numerical methods through the lens of modern theory and computation can provide not only a principled approach for deploying neural networks in numerical analysis, but also effective and scalable tools for machine learning.



Paperid:1604
Authors:Arthur Bizzi, Leonardo Moreira, Márcio Marques, Leonardo Mendonça, Christian de Oliveira, Vitor Balestro, Lucas dos Santos Fernandez, Daniel Yukimura, Pavel Petrov, João Pereira, Tiago Novello, Lucas Nissenbaum
Abstract:
Physics-Informed Neural Networks (PINNs) have emerged as a powerful neural framework for solving partial differential equations (PDEs). However, standard MLP-based PINNs often fail to converge when dealing with complex initial-value problems, leading to solutions that violate causality and suffer from a spectral bias towards low-frequency components. To address these issues, we introduce NeuSA(Neuro-Spectral Architectures), a novel class of PINNs inspired by classical spectral methods, designed to solve linear and nonlinear PDEs with variable coefficients. NeuSA learns a projection of the underlying PDE onto a spectral basis, leading to a finite-dimensional representation of the dynamicswhich is then integrated with an adapted Neural ODE (NODE). This allows us to overcome spectral bias, by leveraging the high-frequency components enabled by the spectral representation; to enforce causality, by inheriting the causal structure of NODEs, and to start training near the target solution, by means of an initialization scheme based on classical methods. We validate NeuSA on canonical benchmarks for linear and nonlinear wave equations, demonstrating strong performance as compared to other architectures, with faster convergence, improved temporal consistency and superior predictive accuracy. Code and pretrained models will be released.



Paperid:1605
Authors:Chitralekha Gupta, Soundarya Ramesh, Praveen Sasikumar, Kian Yeo, Suranga Nanayakkara
Abstract:
Unmanned Aerial Vehicles (UAVs) or drones, are increasingly used in search and rescue missions to detect human presence. Existing systems primarily leverage vision-based methods which are prone to fail under low-visibility or occlusion. Drone-based audio perception offers promise but suffers from extreme ego-noise that masks sounds indicating human presence. Existing datasets are either limited in diversity or synthetic, lacking real acoustic interactions, and there are no standardized setups for drone audition. To this end, we present DroneAudioset (The dataset is publicly available at https://huggingface.co/datasets/ahlab-drone-project/DroneAudioSet/ under the MIT license), a comprehensive drone audition dataset featuring 23.5 hours of annotated recordings, covering a wide range of signal-to-noise ratios (SNRs) from -60 dB to 0 dB, across various drone types, throttles, microphone configurations as well as environments. The dataset enables development and systematic evaluation of noise suppression and classification methods for human-presence detection under challenging conditions, while also informing practical design considerations for drone audition systems, such as microphone placement trade-offs, and development of drone noise-aware audio processing. This dataset is an important step towards enabling design and deployment of drone-audition systems.



Paperid:1606
Authors:Mathurin VIDEAU, Badr Youbi Idrissi, Alessandro Leite, Marc Schoenauer, David Lopez-Paz, Olivier Teytaud
Abstract:
Tokenization imposes a fixed granularity on the input text, freezing how a language model operates on data and how far in the future it predicts. Byte Pair Encoding (BPE) and similar schemes split text once, build a static vocabulary, and leave the model stuck with that choice. We relax this rigidity by introducing an autoregressive U-Net that learns to embed its own tokens as it trains. The network reads raw bytes, pools them into words, then pairs of words, then up to 4 words, giving it a multi-scale view of the sequence. At deeper stages, the model must predict further into the future -- anticipating the next few words rather than the next byte -- so deeper stages focus on broader semantic patterns while earlier stages handle fine details. When carefully tuning and controlling pretraining compute, shallow hierarchies tie strong BPE baselines, and deeper hierarchies have a promising trend. Because tokenization now lives inside the model, the same system can handle character-level tasks and carry knowledge across low-resource languages.



Paperid:1607
Authors:Xueliang Zhao, Wei Wu, Jian Guan, Qintong Li, Lingpeng Kong
Abstract:
In modern sequential decision-making systems, the construction of an optimal candidate action space is critical to efficient inference. However, existing approaches either rely on manually defined action spaces that lack scalability or utilize unstructured spaces that render exhaustive search computationally prohibitive. In this paper, we propose a novel framework named \textsc{DynaAct} for automatically constructing a compact action space to enhance sequential reasoning in complex problem-solving scenarios. Our method first estimates a proxy for the complete action space by extracting general sketches observed in a corpus covering diverse complex reasoning problems using large language models. We then formulate a submodular function that jointly evaluates candidate actions based on their utility to the current state and their diversity, and employ a greedy algorithm to select an optimal candidate set. Extensive experiments on six diverse standard benchmarks demonstrate that our approach significantly improves overall performance, while maintaining efficient inference without introducing substantial latency.



Paperid:1608
Authors:Hao Wang, Licheng Pan, Zhichao Chen, Xu Chen, Qingyang Dai, Lei Wang, Haoxuan Li, Zhouchen Lin
Abstract:
Training time-series forecasting models presents unique challenges in designing effective learning objectives. Existing methods predominantly utilize the temporal mean squared error, which faces two critical challenges: (1) label autocorrelation, which leads to bias from the label sequence likelihood; (2) excessive amount of tasks, which increases with the forecast horizon and complicates optimization. To address these challenges, we propose Transform-enhanced Direct Forecast (TransDF), which transforms the label sequence into decorrelated components with discriminated significance. Models are trained to align the most significant components, thereby effectively mitigating label autocorrelation and reducing task amount. Extensive experiments demonstrate that TransDF achieves state-of-the-art performance and is compatible with various forecasting models. Code is available at https://anonymous.4open.science/r/TransDF-88CF.



Paperid:1609
Authors:Jingyuan Qi, Zhiyang Xu, Qifan Wang, Lifu Huangg
Abstract:
We introduce Autoregressive Retrieval Augmentation (AR-RAG), a novel paradigm that enhances image generation by autoregressively incorporating k-nearest neighbor retrievals at the patch level.Unlike prior methods that perform a single, static retrieval before generation and condition the entire generation on fixed reference images, AR-RAG performs context-aware retrievals at each generation step, using prior-generated patches as queries to retrieve and incorporate the most relevant patch-level visual references, enabling the model to respond to evolving generation needs while avoiding limitations (e.g., over-copying, stylistic bias, etc.) prevalent in existing methods. To realize AR-RAG, we propose two parallel frameworks: (1) Distribution-Augmentation in Decoding (DAiD), a training-free plug-and-use decoding strategy that directly merges the distribution of model-predicted patches with the distribution of retrieved patches, and (2) Feature-Augmentation in Decoding (FAiD), a parameter-efficient fine-tuning method that progressively smooths the features of retrieved patches via multi-scale convolution operations and leverages them to augment the image generation process. We validate the effectiveness of AR-RAG on widely adopted benchmarks, including Midjourney-30K, GenEval and DPG-Bench, demonstrating significant performance gains over state-of-the-art image generation models.



Paperid:1610
Authors:LINHAO LUO, Zicheng Zhao, Reza Haffari, Dinh Phung, Chen Gong, Shirui Pan
Abstract:
Retrieval-augmented generation (RAG) has proven effective in integrating knowledge into large language models (LLMs). However, conventional RAGs struggle to capture complex relationships between pieces of knowledge, limiting their performance in intricate reasoning that requires integrating knowledge from multiple sources. Recently, graph-enhanced retrieval augmented generation (GraphRAG) builds a graph structure to explicitly model these relationships, enabling more effective and efficient retrievers. Nevertheless, its performance is still hindered by the noise and incompleteness within the graph structure. To address this, we introduce GFM-RAG, a novel graph foundation model (GFM) for retrieval augmented generation. GFM-RAG is powered by an innovative graph neural network that reasons over graph structure to capture complex query-knowledge relationships. The GFM with 8M parameters undergoes a two-stage training process on large-scale datasets, comprising 60 knowledge graphs with over 14M triples and 700k documents. This results in impressive performance and generalizability for GFM-RAG, making it the first graph foundation model applicable to unseen datasets for retrieval without any fine-tuning required. Extensive experiments on three multi-hop QA datasets and seven domain-specific RAG datasets demonstrate that GFM-RAG achieves state-of-the-art performance while maintaining efficiency and alignment with neural scaling laws, highlighting its potential for further improvement.



Paperid:1611
Authors:Yuxin Chang, Alex Boyd, Cao (Danica) Xiao, Taha Kass-Hout, Parminder Bhatia, Padhraic Smyth, andrew warrington
Abstract:
Marked temporal point processes (MTPPs) model sequences of events occurring at irregular time intervals, with wide-ranging applications in fields such as healthcare, finance and social networks. We propose thestate-space point process(S2P2), a novel and performant model that leverages techniques derived for modern deep state-space models (SSMs) to overcome limitations of existing MTPP models, while simultaneously imbuing strong inductive biases for continuous-time event sequences that other discrete sequence models (i.e., RNNs, transformers, SSMs) do not capture. Inspired by the classical linear Hawkes processes, we propose an architecture that interleaves stochastic jump differential equations with nonlinearities to create a highly expressive intensity-based MTPP model, without the need for restrictive parametric assumptions for the intensity. This novel scheme admits efficient inference with a parallel scan, bringing linear complexity and sublinear scaling while retaining expressivity to MTPPs. Empirically, S2P2 achieves state-of-the-art predictive likelihoods across eight real-world datasets, delivering an average improvement of 33% over the best existing approaches.



Paperid:1612
Authors:Jiazhou Zhou, Kanghao Chen, Lei Zhang, Lin Wang
Abstract:
Event cameras are bio-inspired sensors that capture intensity changes asynchronously with distinct advantages, such as high temporal resolution. Existing methods for event-based object/action recognition predominantly sample and convert event representation at every fixed temporal interval (or frequency). However, they are constrained to processing a limited number of event lengths and show poor frequency generalization, thus not fully leveraging the event's high temporal resolution. In this paper, we present our PASS framework, exhibiting superior capacity for spatiotemporal event modeling towards a larger number of event lengths and generalization across varying inference temporal frequencies. Our key insight is to learn adaptively encoded event features via the state space models (SSMs), whose linear complexity and generalization on input frequency make them ideal for processing high temporal resolution events. Specifically, we propose a Path-selective Event Aggregation and Scan (PEAS) module to encode events into features with fixed dimensions by adaptively scanning and selecting aggregated event presentation. On top of it, we introduce a novel Multi-faceted Selection Guiding (MSG) loss to minimize the randomness and redundancy of the encoded features during the PEAS selection process. Our method outperforms prior methods on five public datasets and shows strong generalization across varying inference frequencies with less accuracy drop (ours -8.62% v.s. -20.69% for the baseline). Moreover, our model exhibits strong long spatiotemporal modeling for a broader distribution of event length (1-10^9), precise temporal perception, and effective generalization for real-world scenarios. Code and checkpoints will be released upon acceptance.



Paperid:1613
Authors:Yinhao Dong, Shan Jiang, Shi Li, Pan Peng
Abstract:
Abstract:We study streaming algorithms for Correlation Clustering. Given a graph as an arbitrary-order stream of edges, with each edge labeled as positive or negative, the goal is to partition the vertices into disjoint clusters, such that the number of disagreements is minimized. In this paper, we give the first learning-augmented streaming algorithms for the problem on both complete and general graphs, improving the best-known space-approximation tradeoffs. Based on the works of Cambus et al. (SODA'24) and Ahn et al. (ICML'15), our algorithms use the predictions of pairwise distances between vertices provided by a predictor. For complete graphs, our algorithm achieves a better-than-$3$ approximation under good prediction quality, while using $\tilde{O}(n)$ total space. For general graphs, our algorithm achieves an $O(\log |E^-|)$ approximation under good prediction quality using $\tilde{O}(n)$ total space, improving the best-known non-learning algorithm in terms of space efficiency. Experimental results on synthetic and real-world datasets demonstrate the superiority of our proposed algorithms over their non-learning counterparts.



Paperid:1614
Authors:Maximilian Du, Shuran Song
Abstract:
Deploying large, complex policies in the real world requires the ability to steer them to fit the needs of a situation. Most common steering approaches, like goal-conditioning, require training the robot policy with a distribution of test-time objectives in mind. To overcome this limitation, we present DynaGuide, a steering method for diffusion policies using guidance from an external dynamics model during the diffusion denoising process. DynaGuide separates the dynamics model from the base policy, which gives it multiple advantages, including the ability to steer towards multiple objectives, enhance underrepresented base policy behaviors, and maintain robustness on low-quality objectives. The separate guidance signal also allows DynaGuide to work with off-the-shelf pretrained diffusion policies. We demonstrate the performance and features of DynaGuide against other steering approaches in a series of simulated and real experiments, showing an average steering success of 70% on a set of articulated CALVIN tasks and outperforming goal-conditioning by 5.4x when steered with low-quality objectives. We also successfully steer an off-the-shelf real robot policy to express preference for particular objects and even create novel behavior.



Paperid:1615
Authors:Reiss Koh, Wonbeen Oh, Jaein Jang, MinHyung Lee, Hyeongjin Kim, Ah Kim, Joonkee Kim, Junghyun Lee, Taehyeon Kim, Se-Young Yun
Abstract:
Self-Taught Reasoners (STaR), synonymously known as Rejection sampling Fine-Tuning (RFT), is an integral part of the training pipeline of self-improving reasoning Language Models (LMs). The self-improving mechanism often employs random observation (data) sampling. However, this results in trained observation imbalance; inefficiently over-training on solved examples while under-training on challenging ones. In response, we introduce Adaptive STaR (AdaSTaR), a novel algorithm that rectifies this by integrating two adaptive sampling principles: (1) Adaptive Sampling for Diversity: promoting balanced training across observations, and (2) Adaptive Sampling for Curriculum: dynamically adjusting data difficulty to match the model's evolving strength. Across six benchmarks, AdaSTaR achieves best test accuracy in all instances (6/6) and reduces training FLOPs by an average of 58.6\% against an extensive list of baselines. These improvements in performance and efficiency generalize to different pre-trained LMs and larger models, paving the way for more efficient and effective self-improving LMs.



Paperid:1616
Authors:Boce Hu, Dian Wang, David Klee, Heng Tian, Xupeng Zhu, Haojie Huang, Robert Platt, Robin Walters
Abstract:
Abstract:Equivariant models have recently been shown to improve the data efficiency of diffusion policy by a significant margin. However, prior work that explored this direction focused primarily on point cloud inputs generated by multiple cameras fixed in the workspace. This type of point cloud input is not compatible with the now-common setting where the primary input modality is an eye-in-hand RGB camera like a GoPro. This paper closes this gap by incorporating into the diffusion policy model a process that projects features from the 2D RGB camera image onto a sphere. This enables us to reason about symmetries in $\mathrm{SO}(3)$ without explicitly reconstructing a point cloud. We perform extensive experiments in both simulation and the real world that demonstrate that our method consistently outperforms strong baselines in terms of both performance and sample efficiency. Our work is the first $\mathrm{SO}(3)$-equivariant policy learning framework for robotic manipulation that works using only monocular RGB inputs.



Paperid:1617
Authors:Shun Lei, Yaoxun XU, ZhiweiLin, Huaicheng Zhang, Wei tan, Hangting Chen, Yixuan Zhang, Chenyu Yang, Haina Zhu, Shuai Wang, Zhiyong Wu, Dong Yu
Abstract:
Recent advances in large language models (LLMs) and audio language models have significantly improved music generation, particularly in lyrics-to-song generation.However, existing approaches still struggle with the complex composition of songs and the scarcity of high-quality data, leading to limitations in sound quality, musicality, instruction following, and vocal-instrument harmony.To address these challenges, we introduce LeVo, an LM-based framework consisting of LeLM and a music codec.LeLM is capable of parallel modeling of two types of tokens: mixed tokens, which represent the combined audio of vocals and accompaniment to achieve vocal-instrument harmony, and dual-track tokens, which separately encode vocals and accompaniment for high-quality song generation.It employs two decoder-only transformers and a modular extension training strategy to prevent interference between different token types.To further enhance musicality and instruction following, we introduce a multi-preference alignment method based on Direct Preference Optimization (DPO).This method handles diverse human preferences through a semi-automatic data construction process and DPO fine-tuning.Experimental results demonstrate that LeVo consistently outperforms existing methods on both objective and subjective metrics.Ablation studies further justify the effectiveness of our designs.Audio examples are available at https://levo-demo.github.io/.



Paperid:1618
Authors:Brian Chmiel, Maxim Fishman, Ron Banner, Daniel Soudry
Abstract:
Abstract:We demonstrate, for the first time, fully quantized training (FQT) of large language models (LLMs) using predominantly 4-bit floating-point (FP4) precision for weights, activations, and gradients on datasets up to 200 billion tokens. We extensively investigate key design choices for FP4, including block sizes, scaling formats, and rounding methods. Our analysis shows that the NVFP4 format, where each block of 16 FP4 values (E2M1) shares a scale represented in E4M3, provides optimal results. We use stochastic rounding for backward and update passes and round-to-nearest for the forward pass to enhance stability. Additionally, we identify a theoretical and empirical threshold for effective quantized training: when the gradient norm falls below approximately $\sqrt{3}$ times the quantization noise, quantized training becomes less effective. Leveraging these insights, we successfully train a 7-billion-parameter model on 256 Intel Gaudi2 accelerators. The resulting FP4-trained model achieves downstream task performance comparable to a standard BF16 baseline, confirming that FP4 training is a practical and highly efficient approach for large-scale LLM training. A reference implementation is supplied in https://github.com/Anonymous1252022/fp4-all-the-way.



Paperid:1619
Authors:Kevin Frans, Sergey Levine, Pieter Abbeel
Abstract:
In this work, we take an experimentally grounded look at neural network optimization. Building on the Shampoo family of algorithms, we identify and alleviate three key issues, resulting in the proposed SPlus method. First, we find that naive Shampoo is prone to divergence when matrix-inverses are cached for long periods. We introduce an alternate bounded update combining a historical eigenbasis with instantaneous normalization, resulting in across-the-board stability and significantly lower computational requirements. Second, we adapt a shape-aware scaling to enable learning rate transfer across network width. Third, we find that high learning rates result in large parameter noise, and propose a simple iterate-averaging scheme which unblocks faster learning. To properly confirm these findings, we introduce a pointed Transformer training benchmark, considering three objectives (language modelling, image classification, and diffusion modelling) across different stages of training. On average, SPlus is able to reach the validation performance of Adam within 44% of the gradient steps and 62% of the wallclock time.



Paperid:1620
Authors:Kiarash Shamsi, Tran Gia Bao Ngo, Razieh Shirzadkhani, Shenyang Huang, Farimah Poursafaei, Poupak Azad, Reihaneh Rabbany, Baris Coskunuzer, Guillaume Rabusseau, Cuneyt Akcora
Abstract:
Temporal Graph Learning (TGL) aims to discover patterns in evolving networks or temporal graphs and leverage these patterns to predict future interactions. However, most existing research focuses on learning from a single network in isolation, leaving the challenges of within-domain and cross-domain generalization largely unaddressed. In this study, we introduce a new benchmark of 84 real-world temporal transaction networks and proposeTemporal Multi-network Transfer (MiNT), a pre-training framework designed to capture transferable temporal dynamics across diverse networks. We train MiNT models on up to 64 transaction networks and evaluate their generalization ability on 20 held-out, unseen networks. Our results show that MiNT consistently outperforms individually trained models, revealing a strong relation between the number of pre-training networks and transfer performance. These findings highlight scaling trends in temporal graph learning and underscore the importance of network diversity in improving generalization. This work establishes the first large-scale benchmark for studying transferability in TGL and lays the groundwork for developing Temporal Graph Foundation Models. Our code is available at \url{https://github.com/benjaminnNgo/ScalingTGNs}



Paperid:1621
Authors:Tarun Gogisetty, Naman Malpani, Gugan Chandrashekhar Thoppe, Devarajan Sridharan
Abstract:
Abstract:Deep neural networks are increasingly applied for automated histopathology. Yet, whole-slide images (WSIs) are often acquired at gigapixel sizes, rendering it computationally infeasible to analyze them entirely at high resolution. Diagnostic labels are largely available only at the slide-level, because expert annotation of images at a finer (patch) level is both laborious and expensive. Moreover, regions with diagnostic information typically occupy only a small fraction of the WSI, making it inefficient to examine the entire slide at full resolution. Here, we propose SASHA -- ${\it S}$equential ${\it A}$ttention-based ${\it S}$ampling for ${\it H}$istopathological ${\it A}$nalysis -- a deep reinforcement learning approach for efficient analysis of histopathological images. First, SASHA learns informative features with a lightweight hierarchical, attention-based multiple instance learning (MIL) model. Second, SASHA samples intelligently and zooms selectively into a small fraction (10-20\%) of high-resolution patches, to achieve reliable diagnosis. We show that SASHA matches state-of-the-art methods that analyze the WSI fully at high-resolution, albeit at a fraction of their computational and memory costs. In addition, it significantly outperforms competing, sparse sampling methods. We propose SASHA as an intelligent sampling model for medical imaging challenges that involve automated diagnosis with exceptionally large images containing sparsely informative features.



Paperid:1622
Authors:Florian Felten, Gabriel Apaza, Gerhard Bräunlich, Cashen Diniz, Xuliang Dong, Arthur Drake, Milad Habibi, Nathaniel Hoffman, Matthew Keeler, Soheyl Massoudi, Francis VanGessel, Mark Fuge
Abstract:
Engineering design optimization seeks to automatically determine the shapes, topologies, or parameters of components that maximize performance under given conditions. This process often depends on physics-based simulations, which are difficult to install, computationally expensive, and require domain-specific expertise. To mitigate these challenges, we introduce EngiBench, the first open‐source library and datasets spanning diverse domains for data‐driven engineering design. EngiBench provides a unified API and a curated set of benchmarks---covering aeronautics, heat conduction, photonics, and more---that enable fair, reproducible comparisons of optimization and machine learning algorithms, such as generative or surrogate models. We also release EngiOpt, a companion library offering a collection of such algorithms compatible with the EngiBench interface. Both libraries are modular, letting users plug in novel algorithms or problems, automate end-to-end experiment workflows, and leverage built-in utilities for visualization, dataset generation, feasibility checks, and performance analysis. We demonstrate their versatility through experiments comparing state-of-the-art techniques across multiple engineering design problems, an undertaking that was previously prohibitively time-consuming to perform. Finally, we show that these problems pose significant challenges for standard machine learning methods due to highly sensitive and constrained design manifolds.



Paperid:1623
Authors:Zihao Tang, Boyuan Wang, Chuan Wen, Jiaye Teng
Abstract:
Conformal prediction is widely adopted in uncertainty quantification, due to its post-hoc, distribution-free, and model-agnostic properties.In the realm of modern deep learning, researchers have proposed Feature Conformal Prediction (FCP), which deploys conformal prediction in a feature space, yielding reduced band lengths.However, the practical utility of FCP is limited due to the time-consuming non-linear operations required to transform confidence bands from feature space to output space.In this paper, we present Fast Feature Conformal Prediction (FFCP), a method that accelerates FCP by leveraging a Taylor expansion to approximate these non-linear operations.The proposed FFCP introduces a novel non-conformity score that is both effective and efficient for real-world applications.Empirical validations showcase that FFCP performs comparably with FCP (both outperforming the vanilla version) while achieving a significant reduction in computational time by approximately 50x in both regression and classification tasks.



Paperid:1624
Authors:Alexey Zakharov, Shimon Whiteson
Abstract:
Abstract:Natural language can offer a concise and human-interpretable means of specifying reinforcement learning (RL) tasks. The ability to extract rewards from a language instruction can enable the development of robotic systems that can learn from human guidance; however, it remains a challenging problem, especially in visual environments. Existing approaches that employ large, pretrained language models either rely on non‑visual environment representations, require prohibitively large amounts of feedback, or generate noisy, ill‑shaped reward functions. In this paper, we propose a novel method, GoalLadder, that leverages vision-language models (VLMs) to train RL agents from a single language instruction in visual environments. GoalLadder works by incrementally discovering states that bring the agent closer to completing a task specified in natural language. To do so, it queries a VLM to identify states that represent an improvement in agent's task progress and to rank them using pairwise comparisons. Unlike prior work, GoalLadder does not trust VLM's feedback completely; instead, it uses it to rank potential goal states using an ELO-based rating system, thus reducing the detrimental effects of noisy VLM feedback. Over the course of training, the agent is tasked with minimising the distance to the top-ranked goal in a learned embedding space, which is trained on unlabelled visual data. This key feature allows us to bypass the need for abundant and accurate feedback typically required to train a well-shaped reward function. We demonstrate that GoalLadder outperforms existing related methods on classic control and robotic manipulation environments with the average final success rate of $\sim$95\% compared to only $\sim$45\% of the best competitor.



Paperid:1625
Authors:Jiayi Luo, Qingyun Sun, Beining Yang, Haonan Yuan, Xingcheng Fu, Yanbiao Ma, Jianxin Li, Philip S Yu
Abstract:
Graph condensation (GC) has gained significant attention for its ability to synthesize smaller yet informative graphs. However, existing studies often overlook the robustness of GC in scenarios where the original graph is corrupted. In such cases, we observe that the performance of GC deteriorates significantly, while existing robust graph learning technologies offer only limited effectiveness. Through both empirical investigation and theoretical analysis, we reveal that GC is inherently an intrinsic-dimension-reducing process, synthesizing a condensed graph with lower classification complexity. Although this property is critical for effective GC performance, it remains highly vulnerable to adversarial perturbations. To tackle this vulnerability and improve GC robustness, we adopt the geometry perspective of graph data manifold and propose a novelManifold-constrainedRobustGraphCondensation framework namedMRGC. Specifically, we introduce three graph data manifold learning modules that guide the condensed graph to lie within a smooth, low-dimensional manifold with minimal class ambiguity, thereby preserving the classification complexity reduction capability of GC and ensuring robust performance under universal adversarial attacks. Extensive experiments demonstrate the robustness of MRGC across diverse attack scenarios.



Paperid:1626
Authors:Xiaoyu Zhou, Jingqi Wang, Yuang Jia, Yongtao Wang, Deqing Sun, Ming-Hsuan Yang
Abstract:
Current 3D scene understanding methods are limited by offline-collected multi-view data or pre-constructed 3D geometry. In this paper, we present ExtractAnything3D (EA3D), a unified online framework for open-world 3D object extraction that enables simultaneous geometric reconstruction and holistic scene understanding. Given a streaming video, EA3D dynamically interprets each frame using vision-language and 2D vision foundation encoders to extract object-level knowledge. This knowledge is integrated and embedded into a Gaussian feature map via a feed-forward online update strategy. We then iteratively estimate visual odometry from historical frames and incrementally update online Gaussian features with new observations. A recurrent joint optimization module directs the model's attention to regions of interest, simultaneously enhancing both geometric reconstruction and semantic understanding. Extensive experiments across diverse benchmarks and tasks, including photo-realistic rendering, semantic and instance segmentation, 3D bounding box and semantic occupancy estimation, and 3D mesh generation, demonstrate the effectiveness of EA3D. Codes will be released upon acceptance.



Paperid:1627
Authors:Guido Carnevale, Giuseppe Notarstefano
Abstract:
Model-free optimization methods typically rely on cost samples gathered by perturbing the current solution estimate along a finite and fixed set of directions. However, at each iteration, only current cost samples are used, while potentially informative, previously collected samples are discarded. In this work, we challenge this conventional approach by introducing a simple yet effective memory mechanism that maintains an auxiliary vector of cost samples whose components are iteratively updated. By leveraging this stored information, our method estimates descent directions through an averaging of all perturbing directions weighted by the auxiliary vector components. This results in a faster convergence without increasing the number of function queries. By interpreting the resulting algorithm as a time-varying dynamical system, we are able to establish its convergence properties in the strongly convex case. In particular, by using tools from system theory based on timescale separation, we are able to guarantee a linear convergence rate toward an arbitrarily small neighborhood of the optimal solution. Numerical simulations on regression problems demonstrate that the proposed approach sensibly outperforms existing model-free optimization methods.



Paperid:1628
Authors:Zhixuan Zhou, Tingting Dan, Guorong Wu
Abstract:
A fundamental challenge in cognitive neuroscience is understanding how cognition emerges from the interplay between structural connectivity (SC) and dynamic functional connectivity (FC) in the brain. Network neuroscience has emerged as a powerful framework to understand brain function through a holistic perspective on structure-function relationships. In this context, current machine learning approaches typically seek to establish direct mappings between structural connectivity (SC) and functional connectivity (FC) associated with specific cognitive states.However, these state-independent methods often yield inconsistent results due to overlapping brain networks across cognitive states. To address this limitation, we conceptualize to uncover the dendritic coupling mechanism between one static SC and multiple FCs by solving a flow problem that bridges the distribution of SC to a mixed distribution of FCs, conditioned on various cognitive states, along a Riemannian manifold of symmetric positive-definite (SPD) manifold.We further prove the equivalence between flow matching on the SPD manifold and on the computationally efficient Cholesky manifold. Since a spare of functional connections is shared across cognitive states, we introduce the notion of consensus control to promote the shared kinetic structures between multiple FC-to-SC pathways via synchronized coordination, yielding a biologically meaningful underpinning on SC-FC coupling mechanism.Together, we present BrainFlow, a reversible generative model that achieves state-of-the-art performance on not only synthetic data but also large-scale neuroimaging datasets from UK Biobank and Human Connectome Project.



Paperid:1629
Authors:Yichao Shen, Fangyun Wei, Zhiying Du, Yaobo Liang, Yan Lu, Jiaolong Yang, Nanning Zheng, Baining Guo
Abstract:
Generalization in robot manipulation is essential for deploying robots in open-world environments and advancing toward artificial general intelligence. While recent vision-language-action models leverage large pre-trained understanding models for perception and instruction following, their ability to generalize to novel tasks, objects, and settings remains limited. In this work, we present RoboSeer, a new approach that shifts from understanding to generation. Instead of solely predicting the next action, RoboSeer also imagines and generates the future visual outcome of that action. Built on a multi-modal Diffusion Transformer, RoboSeer jointly models video, language, and action modalities, using pre-trained video generative models for joint visual and action forecasting. Our experiments show that high-quality imagined futures correlate with reliable action predictions and task success, highlighting the importance of visual imagination in manipulation. RoboSeer demonstrates strong generalization, including imitating other embodiments' skills and handling novel objects. This dual-prediction strategy—forecasting both actions and their visual consequences—marks a paradigm shift in robot learning and unlocks generalization capabilities in manipulation systems.



Paperid:1630
Authors:Li, Haotian Wu, Deniz Gunduz
Abstract:
We introduce Modular Region-Based Implicit Codec (MoRIC), a novel image compression algorithm that relies on implicit neural representations (INRs). Unlike previous INR-based codecs that model the entire image with a single neural network, MoRIC assigns dedicated models to distinct regions in the image, each tailored to its local distribution. This region-wise design enhances adaptation to local statistics and enables flexible, single-object compression with fine-grained rate-distortion (RD) control. MoRIC allows regions of arbitrary shapes, and provides the contour information for each region as separate information. In particular, it incorporates adaptive chain coding for lossy and lossless contour compression, and a shared global modulator that injects multi-scale global context into local overfitting processes in a coarse-to-fine manner. MoRIC achieves state-of-the-art performance in single-object compression with significantly lower decoding complexity than existing learned neural codecs, which results in a highly efficient compression approach for fixed-background scenarios, e.g., for surveillance cameras. It also sets a new benchmark among overfitted codecs for standard image compression. Additionally, MoRIC naturally supports semantically meaningful layered compression through selective region refinement, paving the way for scalable and flexible INR-based codecs.



Paperid:1631
Authors:Jing-Yi Zhu, Yi Gao, Miao Xu, Min-Ling Zhang
Abstract:
Multi-label complementary label learning (MLCLL) is a weakly supervised paradigm that addresses multi-label learning (MLL) tasks using complementary labels (i.e., irrelevant labels) instead of relevant labels. Existing methods typically adopt an unbiased risk estimator (URE) under the assumption that complementary labels follow a uniform distribution. However, this assumption fails in real-world scenarios due to instance-specific annotation biases, making URE-based methods ineffective under such conditions. Furthermore, existing methods underutilize label correlations inherent in MLL. To address these limitations, we propose ComRank, a ranking loss framework for MLCLL, which encourages complementary labels to be ranked lower than non-complementary ones, thereby modeling pairwise label relationships. Theoretically, our surrogate loss ensures Bayes consistency under both uniform and biased cases. Experiments demonstrate the effectiveness of our method in MLCLL tasks. Our code will be released after acceptance.



Paperid:1632
Authors:Jiachen Lu, Hailan Shanbhag, Haitham Al Hassanieh
Abstract:
GeRaF is the first method to use neural implicit learning for near-range 3D geometry reconstruction from radio frequency (RF) signals. Unlike RGB or LiDAR-based methods, RF sensing can see through occlusion but suffers from low resolution and noise due to its lens-less imaging nature. While lenses in RGB imaging constrain sampling to 1D rays, RF signals propagate through the entire space, introducing significant noise and leading to cubic complexity in volumetric rendering. Moreover, RF signals interact with surfaces via specular reflections requiring fundamentally different modeling. To address these challenges, GeRaF (1) introduces filter-based rendering to suppress irrelevant signals, (2) implements a physics-based RF volumetric rendering pipeline, and (3) proposes a novel lens-less sampling and lens-less alpha blending strategy that makes full-space sampling feasible during training. By learning signed distance functions, reflectiveness, and signal power through MLPs and trainable parameters, GeRaF takes the first step towards reconstructing millimeter-level geometry from RF signals in real-world settings.



Paperid:1633
Authors:Lingfeng Wang, Hualing Lin, Senda Chen, Tao Wang, Changxu Cheng, Yangyang Zhong, Dong Zheng, Wuyue Zhao
Abstract:
While humans effortlessly draw visual objects and shapes by adaptively allocating attention based on their complexity, existing multimodal large language models (MLLMs) remain constrained by rigid token representations. Bridging this gap, we propose ALTo, an adaptive length tokenizer for autoregressive mask generation. To achieve this, a novel token length predictor is designed, along with a length regularization term and a differentiable token chunking strategy. We further build ALToLLM that seamlessly integrates ALTo into MLLM. Preferences on the trade-offs between mask quality and efficiency is implemented by group relative policy optimization (GRPO). Experiments demonstrate that ALToLLM achieves state-of-the-art performance with adaptive token cost on popular segmentation benchmarks. Code and models will be released.



Paperid:1634
Authors:Xiaopeng Li, Shangwen Wang, Shasha Li, Shezheng Song, Bin Ji, Ma Jun, Jie Yu
Abstract:
Abstract:Model editing enables targeted updates to the knowledge of large language models (LLMs) with minimal retraining. Among existing approaches, locate-then-edit methods constitute a prominent paradigm: they first identify critical layers, then compute residuals at the final critical layer based on the target edit, and finally apply least-squares-based multi-layer updates via \textbf{residual distribution}. While empirically effective, we identify a counterintuitive failure mode: residual distribution, a core mechanism in these methods, introduces weight shift errors that undermine editing precision. Through theoretical and empirical analysis, we show that such errors increase with the distribution distance, batch size, and edit sequence length, ultimately leading to inaccurate or suboptimal edits. To address this, we propose the $\textbf{B}$oundary $\textbf{L}$ayer $\textbf{U}$pdat$\textbf{E (BLUE)}$ strategy to enhance locate-then-edit methods. Sequential batch editing experiments on three LLMs and two datasets demonstrate that BLUE not only delivers an average performance improvement of 35.59\%, significantly advancing the state of the art in model editing, but also enhances the preservation of LLMs' general capabilities.



Paperid:1635
Authors:John Lazarsfeld, Georgios Piliouras, Ryann Sim, Stratis Skoulakis
Abstract:
Abstract:This paper studies the *optimistic* variant of Fictitious Play for learning in two-player zero-sum games. While it is known that Optimistic FTRL -- a *regularized* algorithm with a bounded stepsize parameter -- obtains constant regret in this setting, we show for the first time that similar, optimal rates are also achievable *without* regularization: we prove for two-strategy games that Optimistic Fictitious Play (using *any* tiebreaking rule) obtains only *constant regret*, providing surprising new evidence on the ability of *non*-no-regret algorithms for fast learning in games. Our proof technique leverages a geometric view of Optimistic Fictitious Play in the dual space of payoff vectors, where we show a certain *energy function* of the iterates remains bounded over time. Additionally, we also prove a regret *lower bound* of $\Omega(\sqrt{T})$ for *Alternating* Fictitious Play. In the unregularized regime, this separates the ability of optimism and alternation in achieving $o(\sqrt{T})$ regret.



Paperid:1636
Authors:Thomas Norrenbrock, Timo Kaiser, Sovan Biswas, Neslihan Kose, Ramesh Manuvinakurike, Bodo Rosenhahn
Abstract:
Globally interpretable models are a promising approach for trustworthy AI in safety-critical domains. Alongside global explanations, detailed local explanations are a crucial complement to effectively support human experts during inference. This work proposes the Calibrated Hierarchical QPM (CHiQPM) which offers uniquely comprehensive global and local interpretability, paving the way for human-AI complementarity. CHiQPM achieves superior global interpretability by contrastively explaining the majority of classes and offers novel hierarchical explanations that are more similar to how humans reason and can be traversed to offer a built-in interpretable Conformal prediction (CP) method. Our comprehensive evaluation shows that CHiQPM achieves state-of-the-art accuracy as a point predictor, maintaining 99% accuracy of non-interpretable models. This demonstrates a substantial improvement, where interpretability is incorporated without sacrificing overall accuracy. Furthermore, its calibrated set prediction is competitively efficient to other CP methods, while providing interpretable predictions of coherent sets along its hierarchical explanation.



Paperid:1637
Authors:Yoav Gelberg, Yam Eitan, Aviv Navon, Aviv Shamsian, Theo Putterman, Michael Bronstein, Haggai Maron
Abstract:
Gradients of neural networks encode valuable information for optimization, editing, and analysis of models. Therefore, practitioners often treat gradients as inputs to task-specific algorithms, e.g., using gradient statistics for pruning or optimization. Recent works explorelearningalgorithms that operate directly on gradients but use architectures that are not specifically designed for gradient processing, hindering their applicability. In this paper, we present a principled approach for designing architectures that process gradients. Our approach is guided by three principles: (1) equivariant design that preserves neuron permutation symmetries, (2) processing sets of gradients across multiple data points to capture curvature information, and (3) efficient gradient representation through rank-1 decomposition. Based on these principles, we introduce GradMetaNet, a novel architecture for learning on gradients, constructed from simple equivariant blocks. We prove universality results for GradMetaNet, and show that previous approaches cannot approximate natural gradient-based functions that GradMetaNet can. We then demonstrate GradMetaNet's effectiveness on a diverse set of gradient-based tasks forMLPsandtransformers, such as learned optimization, INR editing, and loss landscape curvature estimation.



Paperid:1638
Authors:Top Piriyakulkij, Yichao Liang, Hao Tang, Adrian Weller, Marta Kryven, Kevin Ellis
Abstract:
Learning how the world works is central to building AI agents that can adapt to complex environments. Traditional world models based on deep-learning demand vast amounts of training data, and do not flexibly update their knowledge from sparse observations. Recent advances in program synthesis using Large Language Models (LLMs) give an alternate approach which learns world models represented as source code, supporting strong generalization from little data. To date, application of program-structured world models remains limited to natural language and grid-world domains. We introduce a novel program synthesis method for effectively modeling complex, non-gridworld domains by representing a world model as an exponentially-weighted product of programmatic experts (PoE-World) synthesized by LLMs.We show that this approach can learn complex, stochastic world models from just a few observations. We evaluate the learned world models by embedding them in a model-based planning agent, demonstrating efficient performance and generalization to unseen levels on Atari's Pong and Montezuma's Revenge.



Paperid:1639
Authors:Zenghao Guan, Yucan Zhou, Wu Liu, Xiaoyan Gu
Abstract:
Test-time adaptation (TTA) for Vision-Language Models (VLMs) aims to enhance performance on unseen test data. However, existing methods struggle to achieve robust and continuous knowledge accumulation during test time. To address this, we propose Statistics Caching test-time Adaptation (SCA), a novel cache-based approach. Unlike traditional feature-caching methods prone to forgetting, SCA continuously accumulates task-specific knowledge from all encountered test samples. By formulating the reuse of past features as a least squares problem, SCA avoids storing raw features and instead maintains compact, incrementally updated feature statistics. This design enables efficient online adaptation without the limitations of fixed-size caches, ensuring that the accumulated knowledge grows persistently over time. Furthermore, we introduce adaptive strategies that leverage the VLM's prediction uncertainty to reduce the impact of noisy pseudo-labels and dynamically balance multiple prediction sources, leading to more robust and reliable performance. Extensive experiments demonstrate that SCA achieves compelling performance while maintaining competitive computational efficiency.



Paperid:1640
Authors:Elle Miller, Trevor McInroe, David Abel, Oisin Mac Aodha, Sethu Vijayakumar
Abstract:
Effectively combining tactile sensing and reinforcement learning (RL) creates powerful new pathways for sophisticated robot manipulation. However, tactile information is not always fully exploited by neural network-based approaches in deep RL due to its unique characteristics (e.g. sparsity). Departing from conventional reliance on idealised state representations, we present a new approach to strengthen the performance of sensory-driven agents for complex manipulation tasks. We provide a novel application and analysis of tailored reconstruction and multi-step dynamics objectives that help the agent more effectively leverage its tactile observations, and propose training these objectives on a separated auxiliary memory. We find that dynamics-based objectives unlock higher-performing agents that are able to predict future contacts with high precision. Experimental results show the efficacy of our approach through a simulated robotic agent on three complex control tasks with touch and proprioception alone.



Paperid:1641
Authors:Sahar Dastani, Ali Bahri, Gustavo Vargas Hakim, Moslem Yazdanpanah, Mehrdad Noori, David OSOWIECHI, Samuel Barbeau, Ismail Ayed, Herve Lombaert, Christian Desrosiers
Abstract:
State Space Models (SSMs) have emerged as efficient alternatives to Vision Transformers (ViTs), with VMamba standing out as a pioneering architecture designed for vision tasks. However, their generalization performance degrades significantly under distribution shifts. To address this limitation, we propose TRUST (Test-Time Refinement using Uncertainty-Guided SSM Traverses), a novel test-time adaptation (TTA) method that leverages diverse traversal permutations to generate multiple causal perspectives of the input image. Model predictions serve as pseudo-labels to guide updates of the Mamba-specific parameters, and the adapted weights are averaged to integrate the learned information across traversal scans. Altogether, TRUST is the first approach that explicitly leverages the unique architectural properties of SSMs for adaptation. Experiments on seven benchmarks show that TRUST consistently improves robustness and outperforms existing TTA methods.



Paperid:1642
Authors:Yuechen Zhang, Jinbo Xing, bin xia, Shaoteng Liu, Bohao PENG, Xin Tao, Pengfei Wan, Eric Lo, Jiaya Jia
Abstract:
Abstract:Despite the remarkable generation quality of video Diffusion Transformer (DiT) models, their practical deployment is severely hindered by extensive computational requirements. This inefficiency stems from two key challenges: the quadratic complexity of self-attention with respect to token length and the multi-step nature of diffusion models. To address these limitations, we present Jenga, a novel inference pipeline that combines dynamic attention carving with progressive resolution generation. Our approach leverages two key insights: (1) early denoising steps do not require high-resolution latents, and (2) later steps do not require dense attention. Jenga introduces a block-wise attention mechanism that dynamically selects relevant token interactions using 3D space-filling curves, alongside a progressive resolution strategy that gradually increases latent resolution during generation. Experimental results demonstrate that Jenga achieves substantial speedups across multiple state-of-the-art video diffusion models while maintaining comparable generation quality (8.83$\times$ speedup with 0.01\% performance drop on VBench). As a plug-and-play solution, Jenga enables practical, high-quality video generation on modern hardware by reducing inference time from minutes to seconds---without requiring model retraining.



Paperid:1643
Authors:Yangyang Guo, Fangkai Jiao, Liqiang Nie, Mohan Kankanhalli
Abstract:
The vulnerability of Vision Large Language Models (VLLMs) to jailbreak attacks appears as no surprise.However, recent defense mechanisms against these attacks have reached near-saturation performance on benchmark evaluations, often with minimal effort.This \emph{dual high performance} in both attack and defense raises a fundamental and perplexing paradox.To gain a deep understanding of this issue and thus further help strengthen the trustworthiness of VLLMs, this paper makes three key contributions: i) One tentative explanation for VLLMs being prone to jailbreak attacks--\textbf{inclusion of vision inputs}, as well as its in-depth analysis. ii) The recognition of a largely ignored problem in existing defense mechanisms--\textbf{over-prudence}.The problem causes these defense methods to exhibit unintended abstention, even in the presence of benign inputs, thereby undermining their reliability in faithfully defending against attacks.iii) A simple safety-aware method--\textbf{LLM-Pipeline}.Our method repurposes the more advanced guardrails of LLMs on the fly, serving as an effective alternative detector prior to VLLM response. Last but not least, we find that the two representative evaluation methods for jailbreak often exhibit chance agreement.This limitation makes it potentially misleading when evaluating attack strategies or defense mechanisms.We believe the findings from this paper offer useful insights to rethink the foundational development of VLLM safety with respect to benchmark datasets, defense strategies, and evaluation methods.



Paperid:1644
Authors:Xuewei Bai, Yongcai Wang, Deying Li, Haodi Ping, LI Chunxu
Abstract:
Existing tracking-by-detection Multi-Object Tracking methods mainly rely on associating objects with tracklets using motion and appearance features. However, variations in viewpoint and occlusions can result in discrepancies between the features of current objects and those of historical tracklets. To tackle these challenges, this paper proposes a novel Spatial-Temporal Tracklet Graph Matching paradigm (STAR). The core idea of STAR is to achieve long-term, reliable object association through the association of ``tracklet clips (TCs)". TCs are segments of confidently associated multi-object trajectories, which are linked through graph matching. Specifically, STAR initializes TCs using a Confident Initial Tracklet Generator (CITG) and constructs a TC graph via Tracklet Clip Graph Construction (TCGC). In TCGC, each object in a TC is treated as a vertex, with the appearance and local topology features encoded on the vertex. The vertices and edges of the TC graph are then updated through message propagation to capture higher-order features. Finally, a Tracklet Clip Graph Matching (TCGM) method is proposed to efficiently and accurately associate the TCs through graph matching. STAR is model-agnostic, allowing for seamless integration with existing methods to enhance their performance. Extensive experiments on diverse datasets, including MOTChallenge, DanceTrack, and VisDrone2021-MOT, demonstrate the robustness and versatility of STAR, significantly improving tracking performance under challenging conditions. The code will be made available as open-source.



Paperid:1645
Authors:Shivam Duggal, Sanghyun Byun, Bill Freeman, Antonio Torralba, Phillip Isola
Abstract:
According to Algorithmic Information Theory (AIT), intelligent representations compress data into the shortest possible program while remaining predictive of its content—exhibiting low Kolmogorov Complexity (KC). In contrast, most visual representation learning systems assign fixed-length representations to all inputs, ignoring variations in complexity or familiarity. Recent adaptive tokenization methods address this by allocating variable-length representations but typically require test-time search over multiple hypotheses to identify the most predictive one. Inspired by KC principles, we propose a one-shot adaptive tokenizer, KARL, that predicts the appropriate number of tokens for an image in a single forward pass, halting once its approximate KC is reached. The token count serves as a proxy for the minimum description length. KARL performs comparably to recent adaptive tokenizers while operating in a one-pass manner. Additionally, we present a conceptual study showing a correlation between adaptive tokenization and core ideas from AIT. We demonstrate that adaptive tokenization not only aligns with KC but also reveals empirical signals approximating AIT concepts such as sophistication and logical depth. Finally, we analyze predicted image complexity and interestingness across axes such as structure vs. noise and in-distribution vs. out-of-distribution familiarity, highlighting alignment with human annotations.



Paperid:1646
Authors:Paul STRANG, Zacharie ALES, Côme Bissuel, Safia Kedad-Sidhoum, Olivier JUAN, Emmanuel Rachelson
Abstract:
Mixed-Integer Linear Programming (MILP) is a powerful framework used to address a wide range of NP-hard combinatorial optimization problems, often solved by Branch and bound (B&B). A key factor influencing the performance of B&B solvers is the variable selection heuristic governing branching decisions. Recent contributions have sought to adapt reinforcement learning (RL) algorithms to the B&B setting to learn optimal branching policies, through Markov Decision Processes (MDP) inspired formulations, and ad hoc convergence theorems and algorithms. In this work, we introduce B&B MDPs, a principled vanilla MDP formulation for variable selection in B&B, allowing to leverage a broad range of RL algorithms for the purpose of learning optimal B&B heuristics. Computational experiments validate our model empirically, as our branching agent outperforms prior state-of-the-art RL agents on four standard MILP benchmarks.



Paperid:1647
Authors:Jing Tan, Shuai Yang, Tong Wu, Jingwen He, Yuwei Guo, Ziwei Liu, Dahua Lin
Abstract:
Abstract:$360^\circ$ videos offer a hyper-immersive experience that allows the viewers to explore a dynamic scene from full 360 degrees. To achieve more accessible and personalized content creation in $360^\circ$ video format, we seek to lift standard perspective videos into $360^\circ$ equirectangular videos. To this end, we introduce **Imagine360**, the first perspective-to-$360^\circ$ video generation framework that creates high-quality $360^\circ$ videos with rich and diverse motion patterns from video anchors.Imagine360 learns fine-grained spherical visual and motion patterns from limited $360^\circ$ video data with several key designs. **1)** Firstly we adopt the dual-branch design, including a perspective and a panorama video denoising branch to provide local and global constraints for $360^\circ$ video generation, with motion module and spatial LoRA layers fine-tuned on $360^\circ$ videos.**2)** Additionally, an antipodal mask is devised to capture long-range motion dependencies, enhancing the reversed camera motion between antipodal pixels across hemispheres.**3)** To handle diverse perspective video inputs, we propose rotation-aware designs that adapt to varying video masking due to changing camera poses across frames.**4)** Lastly, we introduce a new 360 video dataset featuring 10K high-quality, trimmed 360 video clips with structured motion to facilitate training.Extensive experiments show Imagine360 achieves superior graphics quality and motion coherence with our curated dataset among state-of-the-art $360^\circ$ video generation methods. We believe Imagine360 holds promise for advancing personalized, immersive $360^\circ$ video creation.



Paperid:1648
Authors:Donghyeon Ki, JunHyeok Oh, Seong-Woong Shim, Byung-Jun Lee
Abstract:
Abstract:Diffusion models have recently gained prominence in offline reinforcement learning due to their ability to effectively learn high-performing, generalizable policies from static datasets. Diffusion-based planners facilitate long-horizon decision-making by generating high-quality trajectories through iterative denoising, guided by return-maximizing objectives. However, existing guided sampling strategies such as Classifier Guidance, Classifier-Free Guidance, and Monte Carlo Sample Selection either produce suboptimal multi-modal actions, struggle with distributional drift, or incur prohibitive inference-time costs. To address these challenges, we propose $\textbf{\textit{Prior Guidance}}$ (PG), a novel guided sampling framework that replaces the standard Gaussian prior of a behavior-cloned diffusion model with a learnable distribution, optimized via a behavior-regularized objective. PG directly generates high-value trajectories without costly reward optimization of the diffusion model itself, and eliminates the need to sample multiple candidates at inference for sample selection. We present an efficient training strategy that applies behavior regularization in latent space, and empirically demonstrate that PG outperforms state-of-the-art diffusion policies and planners across diverse long-horizon offline RL benchmarks.



Paperid:1649
Authors:Huy Hoang, Tien Mai, Pradeep Varakantham
Abstract:
This paper addresses the problem of learning avoidance behavior within the context of offline imitation learning. In contrast to conventional methodologies that prioritize the replication of expert or near-expert demonstrations, our work investigates a setting where expert (or desirable) data is absent, and the objective is to learn to eschew undesirable actions by leveraging demonstrations of such behavior (i.e., learning from negative examples).To address this challenge, we propose a novel training objective grounded in the maximum entropy principle. We further characterize the fundamental properties of this objective function, reformulating the learning process as a cooperative inverse Q-learning task. Moreover, we introduce an efficient strategy for the integration of unlabeled data (i.e., data of indeterminate quality) to facilitate unbiased and practical offline training. The efficacy of our method is evaluated across standard benchmark environments, where it consistently outperforms state-of-the-art baselines.



Paperid:1650
Authors:Yijie Hu, Zihao Zhou, Kaizhu Huang, Xiaowei Huang, Qiufeng Wang
Abstract:
Math reasoning has been one crucial ability of large language models (LLMs), where significant advancements have been achieved in recent years. However, most efforts focus on LLMs by curating high-quality annotation data and intricate training (or inference) paradigms, while the math reasoning performance of multi-modal LLMs (MLLMs) remains lagging behind. Since the MLLM typically consists of an LLM and vision block, we wonder: \textit{Can MLLMs directly absorb math reasoning abilities from off-the-shelf math LLMs without tuning?} Recent model-merging approaches may offer insights into this question. However, they overlook the alignment between the MLLM and LLM, where we find that there is a large gap between their parameter spaces, resulting in lower performance. Our empirical evidence reveals two key factors behind this issue: the identification of crucial reasoning-associated layers in the model and the mitigation of the gaps in parameter space. Based on the empirical insights, we propose \textbf{IP-Merging} that first \textbf{I}dentifies the reasoning-associated parameters in both MLLM and Math LLM, then \textbf{P}rojects them into the subspace of MLLM aiming to maintain the alignment, finally merges parameters in this subspace. IP-Merging is a tuning-free approach since parameters are directly adjusted. Extensive experiments demonstrate that our IP-Merging method can enhance the math reasoning ability of MLLMs directly from Math LLMs without compromising their other capabilities.



Paperid:1651
Authors:Haoran Li, CHENHAN XIAO, Muhao Guo, Yang Weng
Abstract:
Learning dynamics is essential for model-based control and Reinforcement Learning in systems operating in changing environments, such as robotics, autonomous vehicles, and power systems. However, limited system measurements, such as those from low-resolution meters, demand sample-efficient learning. Symmetry provides a powerful inductive bias by characterizing equivalent relations in system behavior to improve sample efficiency. While recent methods attempt to discover symmetries from data, they typically assume a single global symmetry group and treat symmetry discovery and dynamic learning as separate tasks, leading to limited expressiveness and error accumulation. In this paper, we propose the Latent Mixture of Symmetries (Latent MoS), an expressive model that captures symmetry-governed latent factors from complex dynamical measurements. Latent MoS focuses on dynamic learning while locally preserving the underlying symmetric transformations. To further capture long-range temporal equivalence, we introduce a hierarchical architecture that stacks Latent MoS blocks across multiple time scales. Numerical experiments across diverse physical systems demonstrate that Latent MoS significantly outperforms state-of-the-art baselines in interpolation and extrapolation tasks while offering interpretable latent representations suitable for future geometric and safety-critical analysis.



Paperid:1652
Authors:Quan Liu, Han Zhou, Wenquan Wu, Hua Wu, Sen Su
Abstract:
Abstract:Ensuring that large language models (LLMs) remain both helpful and harmless poses a significant challenge: fine-tuning on repetitive safety datasets—where unsafe prompts are paired with standard refusal templates—often leads to \emph{false refusals}, in which benign queries are declined. We first quantify this effect, showing that safety data exhibits substantially lower token entropy ($H_{1}\approx9.18$) and 2-gram diversity ($\approx$ 0.048) compared to general instruction data ($H_{1}\approx12.05$, 2-gram$\approx$0.205). To uncover the root cause, we introduce \emph{FlowLens}, a stable PCA-based tool for residual-stream geometry analysis, and reveal that higher proportions of safety examples concentrate variance along a few components, reducing representational smoothness and driving false refusals (false refusal rate rises from 63\% to 84\% as safety data increases from 0\% to 40\%). Guided by these insights, we propose \emph{Variance Concentration Loss} (VCL), an auxiliary regularizer that penalizes excessive variance concentration in mid-layer residuals. Empirical results demonstrate that VCL reduces false refusals by over 35 percentage points while maintaining or improving performance on general benchmarks such as MMLU and GSM8K.



Paperid:1653
Authors:Alex Clinton, Thomas Zeng, Yiding Chen, Jerry Zhu, Kirthevasan Kandasamy
Abstract:
Modern data marketplaces and data sharing consortia increasingly rely on incentive mechanisms to encourage agents to contribute data. However, schemes that reward agents based on the quantity of submitted data are vulnerable to manipulation, as agents may submit fabricated or low-quality data to inflate their rewards. Prior work has proposed comparing each agent’s data against others’ to promote honesty: when others contribute genuine data, the best way to minimize discrepancy is to do the same. Yet prior implementations of this idea rely on very strong assumptions about the data distribution (e.g. Gaussian), limiting their applicability.In this work, we develop reward mechanisms based on a novel, two-sample test inspired by the Cramér-von Mises statistic. Our methods strictly incentivize agents to submit more genuine data, while disincentivizing data fabrication and other types of untruthful reporting.We establish that truthful reporting constitutes a (possibly approximate) Nash equilibrium in both Bayesian and prior-agnostic settings. We theoretically instantiate our method in two canonical data sharing problems and show that it relaxes key assumptions made by prior work.Empirically, we demonstrate that our mechanism incentivizes truthful data sharing via simulations and on real-world language and image data.



Paperid:1654
Authors:Razvan-Andrei Lascu, Mateusz Majka
Abstract:
Abstract:We study the problem of minimizing non-convex functionals on the space of probability measures, regularized by the relative entropy (KL divergence) with respect to a fixed reference measure, as well as the corresponding problem of solving entropy-regularized non-convex-non-concave min-max problems. We utilize the Best Response flow (also known in the literature as the fictitious play flow) and study how its convergence is influenced by the relation between the degree of non-convexity of the functional under consideration, the regularization parameter and the tail behaviour of the reference measure. In particular, we demonstrate how to choose the regularizer, given the non-convex functional, so that the Best Response operator becomes a contraction with respect to the $L^1$-Wasserstein distance, which then ensures the existence of its unique fixed point that is then showed to be the unique global minimizer for our optimization problem. This extends recent results where the Best Response flow was applied to solve convex optimization problems regularized by the relative entropy with respect to arbitrary reference measures, and with arbitrary values of the regularization parameter. Our results explain precisely how the assumption of convexity can be relaxed, at the expense of making a specific choice of the regularizer. Additionally, we demonstrate how these results can be applied in reinforcement learning in the context of policy optimization for bandit problems with softmax parametrized policies in the mean-field regime.



Paperid:1655
Authors:Marta Gentiloni Silveri, Giovanni Conforti, Alain Durmus
Abstract:
The Schrödinger Bridge (SB) problem has become a fundamental tool in computational optimal transport and generative modeling.To address this problem, ideal methods such as Iterative Proportional Fitting and Iterative Markovian Fitting (IMF) have been proposed—alongside practical approximations like Diffusion Schrödinger Bridge and its Matching (DSBM) variant. While previous work have established asymptotic convergence guarantees for IMF, a quantitative, non-asymptotic understanding remains unknown. In this paper, we provide the first non-asymptotic exponential convergence guarantees for IMF under mild structural assumptions on the reference measure and marginal distributions, assuming a sufficiently large time horizon. Our results encompass two key regimes: one where the marginals are log-concave, and another where they are weakly log-concave. The analysis relies on new contraction results for the Markovian projection operator and paves the way to theoretical guarantees for DSBM.



Paperid:1656
Authors:Noga Bar, Mariia Seleznova, ‪Yotam Alexander‬‏, Gitta Kutyniok, Raja Giryes
Abstract:
Abstract:Proper initialization is critical for Recurrent Neural Networks (RNNs), particularly in long-range reasoning tasks, where repeated application of the same weight matrix can cause vanishing or exploding signals.A common baseline for linear recurrences is Glorot initialization, designed to ensure stable signal propagation---but derived under the infinite-width, fixed-length regime—an unrealistic setting for RNNs processing long sequences. In this work, we show that Glorot initialization is in fact unstable: small positive deviations in the spectral radius are amplified through time and cause the hidden state to explode. Our theoretical analysis demonstrates that sequences of length $t = O(\sqrt{n})$, where $n$ is the hidden width, are sufficient to induce instability. To address this, we propose a simple, dimension-aware rescaling of Glorot that shifts the spectral radius slightly below one, preventing rapid signal explosion or decay. These results suggest that standard initialization schemes may break down in the long-sequence regime, motivating a separate line of theory for stable recurrent initialization.



Paperid:1657
Authors:Yulong Li, Xiwei Liu, feilong tang, Ming Hu, Jionglong Su, Zongyuan Ge, Imran Razzak, Eran Segal
Abstract:
Causal mediation analysis is crucial for elucidating the mechanisms through which variables exert their effects. However, the automated identification and quantification of complex mediation pathways under uncertainty causal structures remains an open challenge.Causal mediation analysis is crucial for deconstructing complex mechanisms of action. However, in current mediation analysis, complex structures derived from causal discovery lack direct interpretation of mediation pathways, while traditional mediation analysis and effect estimation are limited by the reliance on pre-specified pathways, leading to a disconnect between structure discovery and causal mechanism understanding. Therefore, a unified framework integrating structure discovery, pathway identification, and effect estimation systematically quantifies mediation pathways under structural uncertainty, enabling automated identification and inference of mediation pathways. To this end, we propose Structure-Guided Mediation Analysis (\textbf{SGMA}), which guides automated mediation pathway identification through probabilistic causal structure discovery and uncertainty quantification, enabling end-to-end propagation of structural uncertainty from structure learning to effect estimation. Specifically, SGMA employs differentiable Flow-Structural Equation Models to learn structural posteriors, generating diverse Directed Acyclic Graphs (DAGs) to quantify structural uncertainty. Based on these DAGs, we introduce the Path Stability Score to evaluate the marginal probability of pathways, identifying high-confidence mediation routes. For identified mediation pathways, we integrate Efficient Influence Functions with Bayesian model averaging to fuse within-structure estimation uncertainty and between-structure effect variation, propagating uncertainty to the final effect estimates. In synthetic data experiments, SGMA achieves state-of-the-art performance in pathway identification accuracy and effect quantification precision under structures uncertainty, concurrent multiple pathways, and nonlinear scenarios. In real-world applications using Human Phenotype Project data, SGMA identifies mediation effects of sleep quality on cardiovascular health through inflammatory and metabolic pathways, uncovering previously unspecified multiple mediation routes.



Paperid:1658
Authors:Oussama Zekri, Nicolas Boulle
Abstract:
Discrete diffusion models have recently gained significant attention due to their ability to process complex discrete structures for language modeling. However, fine-tuning these models with policy gradient methods, as is commonly done in Reinforcement Learning from Human Feedback (RLHF), remains a challenging task. We propose an efficient, broadly applicable, and theoretically justified policy gradient algorithm, called Score Entropy Policy Optimization (SEPO), for fine-tuning discrete diffusion models over non-differentiable rewards. Our numerical experiments across several discrete generative tasks demonstrate the scalability and efficiency of our method.



Paperid:1659
Authors:Armand Kassaï Koupaï, Lise Le Boudec, Louis Serrano, Patrick Gallinari
Abstract:
Solving time-dependent parametric partial differential equations (PDEs) remains a fundamental challenge for neural solvers, particularly when generalizing across a wide range of physical parameters and dynamics.When data is uncertain or incomplete—as is often the case—a natural approach is to turn to generative models.We introduce ENMA, a generative neural operator designed to model spatio-temporal dynamics arising from physical phenomena. ENMA predicts future dynamics in a compressed latent space using a generative masked autoregressive transformer trained with flow matching loss, enabling tokenwise generation. Irregularly sampled spatial observations are encoded into uniform latent representations via attention mechanisms and further compressed through a spatio-temporal convolutional encoder. This allows ENMA to perform in-context learning at inference time by conditioning on either past states of the target trajectory or auxiliary context trajectories with similar dynamics. The result is a robust and adaptable framework that generalizes to new PDE regimes and supports one-shot surrogate modeling of time-dependent parametric PDEs.



Paperid:1660
Authors:SOHAM BONNERJEE, Sayar Karmakar, Wei Biao Wu
Abstract:
Federated Learning has gained traction in privacy-sensitive collaborative environments, with local SGD emerging as a key optimization method in decentralized settings. While its convergence properties are well-studied, asymptotic statistical guarantees beyond convergence remain limited. In this paper, we present two generalized Gaussian approximation results for local SGD and explore their implications. First, we prove a Berry-Esseen theorem for the final local SGD iterates, enabling valid multiplier bootstrap procedures. Second, motivated by robustness considerations, we introduce two distinct time-uniform Gaussian approximations for the entire trajectory of local SGD. The time-uniform approximations support Gaussian bootstrap-based tests for detecting adversarial attacks. Extensive simulations are provided to support our theoretical results.



Paperid:1661
Authors:Matias Cattaneo, Boris Shigida
Abstract:
In modern optimization methods used in deep learning, each update depends on the history of previous iterations, often referred to as memory, and this dependence decays fast as the iterates go further into the past. For example, gradient descent with momentum has exponentially decaying memory through exponentially averaged past gradients. We introduce a general technique for identifying a memoryless algorithm that approximates an optimization algorithm with memory. It is obtained by replacing all past iterates in the update by the current one, and then adding a correction term arising from memory (also a function of the current iterate). This correction term can be interpreted as a perturbation of the loss, and the nature of this perturbation can inform how memory implicitly (anti-)regularizes the optimization dynamics. As an application of our theory, we find that Lion does not have the kind of implicit anti-regularization induced by memory that AdamW does, providing a theory-based explanation for Lion’s better generalization performance recently documented. Empirical evaluations confirm our theoretical findings.



Paperid:1662
Authors:Kentaro Kanamori, Ken Kobayashi, Takuya Takagi
Abstract:
This paper proposes a new algorithm for learning gradient boosted decision trees while ensuring the existence of recourse actions. Algorithmic recourse aims to provide a recourse action for altering the undesired prediction result given by a model. While existing studies often focus on extracting a valid and executable action from a given learned model, such reasonable actions do not always exist for models optimized solely for predictive accuracy. To address this issue, recent studies proposed a framework for learning a model while guaranteeing the existence of reasonable actions with high probability. However, these methods can not be applied to gradient boosted decision trees, which are renowned as one of the most popular models for tabular datasets. We propose an efficient gradient boosting algorithm that takes recourse guarantee into account, while maintaining the same time complexity as the standard ones. We also propose a post-processing method for refining a learned model under the constraint on recourse guarantee and give a PAC-style analysis of the refined model. Experimental results demonstrated that our method successfully provided reasonable actions to more instances than the baselines without significantly degrading accuracy and computational efficiency.



Paperid:1663
Authors:Shijing Hu, Jingyang Li, Xingyu Xie, Zhihui Lu, Kim-chuan Toh, Pan Zhou
Abstract:
Speculative decoding accelerates inference in large language models (LLMs) by generating multiple draft tokens simultaneously. However, existing methods often struggle with token misalignment between the training and decoding phases, limiting their performance. To address this, we propose GRIFFIN, a novel framework that incorporates a token-alignable training strategy and a token-alignable draft model to mitigate misalignment.The training strategy employs a loss masking mechanism to exclude highly misaligned tokens during training, preventing them from negatively impacting the draft model's optimization. The token-alignable draft model introduces input tokens to correct inconsistencies in generated features.Experiments on LLaMA, Vicuna, Qwen and Mixtral models demonstrate that GRIFFIN achieves an average acceptance length improvement of over 8\% and a speedup ratio exceeding 7\%, outperforming current speculative decoding state-of-the-art methods. Our code and GRIFFIN's draft models will be released publicly in https://anonymous.4open.science/r/GRIFFIN-NIPS-E5F3/.



Paperid:1664
Authors:Nicolas von Lützow, Matthias Niessner
Abstract:
Volumetric rendering has become central to modern novel view synthesis methods, which use differentiable rendering to optimize 3D scene representations directly from observed views. While many recent works build on NeRF or 3D Gaussians, we explore an alternative volumetric scene representation. More specifically, we introduce two new scene representations based on linear primitives - octahedra and tetrahedra - both of which define homogeneous volumes bounded by triangular faces. To optimize these primitives, we present a differentiable rasterizer that runs efficiently on GPUs, allowing end-to-end gradient-based optimization while maintaining real-time rendering capabilities. Through experiments on real-world datasets, we demonstrate comparable performance to state-of-the-art volumetric methods while requiring fewer primitives to achieve similar reconstruction fidelity. Our findings deepen the understanding of 3D representations by providing insights into the fidelity and performance characteristics of transparent polyhedra and suggest that adopting novel primitives can expand the available design space.



Paperid:1665
Authors:George Cazenavette, Antonio Torralba, Vincent Sitzmann
Abstract:
The task of dataset distillation aims to find a small set of synthetic images such that training a model on them reproduces the performance of the same model trained on a much larger dataset of real samples. Existing distillation methods focus on synthesizing datasets that enable training randomly initialized models. In contrast, state-of-the-art vision approaches are increasingly building on large, pre-trained self-supervised models rather than training from scratch. In this paper, we investigate the problem of distilling datasets that enable us to optimally train linear probes on top of such large, pre-trained vision models. We introduce a method of dataset distillation for this task called Linear Gradient Matching that optimizes the synthetic images such that, when passed through a pre-trained feature extractor, they induce gradients in the linear classifier similar to those produced by the real data. Our method yields synthetic data that out-perform all real-image baselines and, remarkably, generalize across pre-trained vision models, enabling us, for instance, to train a linear CLIP probe that performs competitively using a dataset distilled via a DINO backbone. Further, we show that distilled datasets provide a valuable tool for model interpretability, predicting, among other things, how similar two model's representations spaces are under the platonic representation hypothesis or whether a model is sensitive to spurious correlations in adversarial datasets.



Paperid:1666
Authors:Zuheng Xu, Trevor Campbell
Abstract:
Most expressive variational families---such as normalizing flows---lack practical convergence guarantees, as their theoretical assurances typically hold only at the intractable global optimum.In this work, we present a general recipe for constructing tuning-free, asymptotically exactvariational flows from involutive MCMC kernels.The core methodological component is a novel representation of general involutive MCMC kernels as invertible, measure-preservingiterated random functionsystems, which act as the flow maps of our variational flows.This leads to three new variational families with provable total variation convergence.Our framework resolves key practical limitations of existing variational families with similar guarantees (e.g., MixFlow), while requiring substantially weaker theoretical assumptions.Finally, we demonstrate the competitive performance of our flows across tasks including posterior approximation, Monte Carlo estimates, and normalization constant estimation,outperforming or matching state-of-the-art MCMC and black-box normalizing flows.



Paperid:1667
Authors:Yushu Pan, Elias Bareinboim
Abstract:
The process of editing an image can be naturally modeled as evaluating a counterfactual query: "What would an image look like had a particular feature changed?". While recent advances in text-guided image editing leverage powerful pre-trained models and produce visually appealing images, they often lack causal consistency, ignoring how features are causally related and how changing one may affect others. In contrast, existing causal editing approaches provide theoretical foundations and practically work for specific settings, but remain limited in terms of scalability and reliance on full supervision. In this work, we move towards bridging the gap between causal editing and large-scale text-to-image generation. We introduce Backdoor Disentangled Causal Latent Space ( \textbf{BD-CLS}) as a structured proxy for counterfactual image editing, which provides causal guarantees even under weak supervision. We then develop an algorithm, BD-CLS-Edit, to extract a BD-CLS from a pre-trained Stable Diffusion model, enabling counterfactual image editing without the need for retraining. Our method ensures that edits respect the causal relationships among features, even when some are unlabeled or unprompted.



Paperid:1668
Authors:Massimiliano Ciranni, Vito Paolo Pastore, Roberto Di Via, Enzo Tartaglione, Francesca Odone, Vittorio Murino
Abstract:
Deep learning model effectiveness in classification tasks is often challenged by the quality and quantity of training data whenever they are affected by strong spurious correlations between specific attributes and target labels. This results in a form of bias affecting training data, which typically leads to unrecoverable weak generalization in prediction. This paper aims at facing this problem by leveraging bias amplification with generated synthetic data: we introduce Diffusing DeBias (DDB), a novel approach acting as a plug-in for common methods of unsupervised model debiasing exploiting the inherent bias-learning tendency of diffusion models in data generation. Specifically, our approach adopts conditional diffusion models to generate synthetic bias-aligned images, which replace the original training set for learning an effective bias amplifier model that we subsequently incorporate into an end-to-end and a two-step unsupervised debiasing approach. By tackling the fundamental issue of bias-conflicting training samples memorization in learning auxiliary models, typical of this type of techniques, our proposed method beats current state-of-the-art in multiple benchmark datasets, demonstrating its potential as a versatile and effective tool for tackling bias in deep learning models.



Paperid:1669
Authors:Yassine Chemingui, Aryan Deshwal, Alan Fern, Thanh Nguyen-Tang, Jana Doppa
Abstract:
We study the problem of Offline Safe Reinforcement Learning (OSRL), where the goal is to learn a reward-maximizing policy from fixed data under a cumulative cost constraint. We propose a novel OSRL approach that frames the problem as a mini-max objective and solves it by combining offline RL with online optimization algorithms. We prove the approximate optimality of this approach when integrated with an approximate offline RL oracle and no-regret online optimization. We also present a practical approximation that can be combined with any offline RL algorithm, eliminating the need for offline policy evaluation. Empirical results on the DSRL benchmark demonstrate that our method reliably enforces safety constraints under stringent cost budgets, while achieving high rewards.



Paperid:1670
Authors:Yanghao Wang, Long Chen
Abstract:
Abstract:Although today's pretrained discriminative vision-language models (e.g., CLIP) have demonstrated strong perception abilities, such as zero-shot image classification, they also suffer from the bag-of-words problem and spurious bias. To mitigate these problems, some pioneering studies leverage powerful generative models (e.g., pretrained diffusion models) to realize generalizable image classification, dubbed Diffusion Classifier (DC). Specifically, by randomly sampling a Gaussian noise, DC utilizes the differences of denoising effects with different category conditions to classify categories. Unfortunately, an inherent and notorious weakness of existing DCs is noise instability: different random sampled noises lead to significant performance changes. To achieve stable classification performance, existing DCs always ensemble the results of hundreds of sampled noises, which significantly reduces the classification speed. To this end, we firstly explore the role of noise in DC, and conclude that: there are some ``good noises'' that can relieve the instability. Meanwhile, we argue that these good noises should meet two principles: 1) Frequency Matching: noise should destroy the specific frequency signals; 2) Spatial Matching: noise should destroy the specific spatial areas. Regarding both principles, we propose a novel Noise Optimization method to learn matching (i.e., good) noise for DCs: NoOp. For frequency matching, NoOp first optimizes a dataset-specific noise: Given a dataset and a timestep $t$, optimize one randomly initialized parameterized noise. For Spatial Matching, NoOp trains a Meta-Network that adopts an image as input and outputs image-specific noise offset. The sum of optimized noise and noise offset will be used in DC to replace random noise. Extensive ablations on various datasets demonstrated the effectiveness of NoOp. It is worth noting that our noise optimization is orthogonal to existing optimization methods (e.g., prompt tuning), our NoOP can even benefit from these methods to further boost performance.



Paperid:1671
Authors:ZhanFeng Feng, Long Peng, Xin Di, Yong Guo, Wenbo Li, Yulun Zhang, Renjing Pei, Yang Wang, Yang Cao, Zheng-Jun Zha
Abstract:
Multi-frame video enhancement tasks aim to improve the spatial and temporal resolution and quality of video sequences by leveraging temporal information from multiple frames, which are widely used in streaming video processing, surveillance, and generation. Although numerous Transformer-based enhancement methods have achieved impressive performance, their computational and memory demands hinder deployment on edge devices. Quantization offers a practical solution by reducing the bit-width of weights and activations to improve efficiency. However, directly applying existing quantization methods to video enhancement tasks often leads to significant performance degradation and loss of fine details. This stems from two limitations: (a) inability to allocate varying representational capacity across frames, which results in suboptimal dynamic range adaptation; (b) over-reliance on full-precision teachers, which limits the learning of low-bit student models. To tackle these challenges, we propose a novel quantization method for video enhancement: Progressive Multi-Frame Quantization for Video Enhancement (PMQ-VE). This framework features a coarse-to-fine two-stage process: Backtracking-based Multi-Frame Quantization (BMFQ) and Progressive Multi-Teacher Distillation (PMTD). BMFQ utilizes a percentile-based initialization and iterative search with pruning and backtracking for robust clipping bounds. PMTD employs a progressive distillation strategy with both full-precision and multiple high-bit (INT) teachers to enhance low-bit models' capacity and quality. Extensive experiments demonstrate that our method outperforms existing approaches, achieving state-of-the-art performance across multiple tasks and benchmarks. The code will be made publicly available.



Paperid:1672
Authors:Riccardo Zamboni, Mirco Mutti, Marcello Restelli
Abstract:
In reinforcement learning, we typically refer tounsupervisedpre-training when we aim to pre-train a policy without a priori access to the task specification, i.e., rewards, to be later employed for efficient learning of downstream tasks. In single-agent settings, the problem has been extensively studied and mostly understood. A popular approach casts the unsupervised objective as maximizing theentropyof the state distribution induced by the agent's policy, from which principles and methods follow. In contrast, little is known about state entropy maximization in multi-agent settings, which are ubiquitous in the real world. What are the pros and cons of alternative problem formulations in this setting? How hard is the problem in theory, how can we solve it in practice? In this paper, we address these questions by first characterizing those alternative formulations and highlighting how the problem, even when tractable in theory, is non-trivial in practice. Then, we present a scalable, decentralized, trust-region policy search algorithm to address the problem in practical settings. Finally, we provide numerical validations to both corroborate the theoretical findings and pave the way for unsupervised multi-agent reinforcement learning via state entropy maximization in challenging domains, showing that optimizing for a specific objective, namelymixture entropy, provides an excellent trade-off between tractability and performances.



Paperid:1673
Authors:Haocheng Yang, Chunyuan Zheng, Haoxuan Li, Mengyue Yang
Abstract:
Selection bias poses a widely-recognized challenge for unbiased evaluation and learning in many industrial scenarios. For example, in recommender systems, it arises from the users' selective interactions with items. Recently, doubly robust and its variants have been widely studied to achieve debiased learning of prediction models, however, all of them consider a simple exact matching scenario, i.e., the units (such as user-item pairs in a recommender system) are the same between the training and test sets. In practice, there may be limited or even no overlap in units between the training and test. In this paper, we consider a more practical scenario: the joint distribution of the feature and rating is the same in the training and test set. Theoretical analysis shows the previous DR estimator is biased even if the imputed errors and learned propensities are correct in this scenario. In addition, we propose a novel super-population doubly robust estimator (SuperDR), which can achieve a more accurate estimation and desirable generalization error bound compared to the existing DR estimators, and extend the joint learning algorithm for training the prediction and imputation models. We conduct extensive experiments on three real-world datasets, including a large-scale industrial dataset, to show the effectiveness of our method.



Paperid:1674
Authors:Fan Mo, Bo Liu, Yuan Fan, Kun Qin, Yizhou Zhao, Jinhe Zhou, Jia Sun, Jinfei Liu, Kui Ren
Abstract:
With the rapid development of artificial intelligence technologies, the demand for training data has surged, exacerbating risks of data leakage. Despite increasing incidents and costs associated with such leaks, data leakage prevention (DLP) technologies lag behind evolving evasion techniques that bypass existing sensitive information recognition (SIR) models. Current datasets lack comprehensive coverage of these adversarial transformations, limiting the evaluation of robust SIR systems. To address this gap, we introduce DataSIR, a benchmark dataset specifically designed to evaluate SIR models on sensitive data subjected to diverse format transformations. We curate 26 sensitive data categories based on multiple international regulations, and collect 131,890 original samples correspondingly. Through empirical analysis of real-world evasion tactics, we implement 21 format transformation methods, which are applied to the original samples, expanding the dataset to 1,647,501 samples to simulate adversarial scenarios. We evaluated DataSIR using four traditional NLP models and four large language models (LLMs). For LLMs, we design structured prompts with varying degrees of contextual hints to assess the impact of prior knowledge on recognition accuracy. These evaluations demonstrate that our dataset effectively differentiates the performance of various SIR algorithms. Combined with its rich category and format diversity, the dataset can serve as a benchmark for evaluating related models and help develop future more advanced SIR models. Our dataset and experimental code are publicly available at https://www.kaggle.com/datasets/fanmo1/datasir and https://github.com/Fan-Mo-ZJU/DataSIR.



Paperid:1675
Authors:WonJun Moon, MinSeok Jung, Gilhan Park, Tae-Young Kim, Cheol-Ho Cho, Woojin Jun, Jae-Pil Heo
Abstract:
Partially Relevant Video Retrieval (PRVR) seeks videos where only part of the content matches a text query. Existing methods treat every annotated text–video pair as a positive and all others as negatives, ignoring the rich semantic variation both within a single video and across different videos. Consequently, embeddings of both queries and their corresponding video‐clip segments for distinct events within the same video collapse together, while embeddings of semantically similar queries and segments from different videos are driven apart.This limits retrieval performance when videos contain multiple, diverse events.This paper addresses the aforementioned problems, termed as semantic collapse, in both the text and video embedding spaces. We first introduce Text Correlation Preservation Learning, which preserves the semantic relationships encoded by the foundation model across text queries.To address collapse in video embeddings, we propose Cross-Branch Video Alignment (CBVA), a contrastive alignment method that disentangles hierarchical video representations across temporal scales.Subsequently, we introduce order-preserving token merging and adaptive CBVA to enhance alignment by producing video segments that are internally coherent yet mutually distinctive.Extensive experiments on PRVR benchmarks demonstrate that our framework effectively prevents semantic collapse and substantially improves retrieval accuracy.



Paperid:1676
Authors:Zijing Ou, Ruixiang ZHANG, Yingzhen Li
Abstract:
Sampling from unnormalised discrete distributions is a fundamental problem across various domains. While Markov chain Monte Carlo offers a principled approach, it often suffers from slow mixing and poor convergence.In this paper, we propose Discrete Neural Flow Samplers (DNFS), a trainable and efficient framework for discrete sampling. DNFS learns the rate matrix of a continuous-time Markov chain such that the resulting dynamics satisfy the Kolmogorov equation.As this objective involves the intractable partition function, we then employ control variates to reduce the variance of its Monte Carlo estimation, leading to a coordinate descent learning algorithm.To further facilitate computational efficiency, we propose locally equivaraint Transformer, a novel parameterisation of the rate matrix that significantly improves training efficiency while preserving powerful network expressiveness.Empirically, we demonstrate the efficacy of DNFS in a wide range of applications, including sampling from unnormalised distributions, training discrete energy-based models, and solving combinatorial optimisation problems.



Paperid:1677
Authors:Paul Liautaud, Pierre Gaillard, Olivier Wintenberger
Abstract:
Abstract:We study online adversarial regression with convex losses against a rich class of continuous yet highly irregular prediction rules, modeled by Besov spaces $B_{pq}^s$ with general parameters $1 \leq p,q \leq \infty$ and smoothness $s > d/p$. We introduce an adaptivewavelet-based algorithm that performs sequential prediction without prior knowledge of $(s,p,q)$, and establish minimax-optimal regret bounds against any comparator in $B_{pq}^s$.We further design a locally adaptive extension capable of dynamically tracking spatially inhomogeneous smoothness. This adaptive mechanism adjusts the resolution of the predictions over both time and space, yielding refined regret bounds in terms of local regularity. Consequently, in heterogeneous environments, our adaptive guarantees can significantly surpass those obtained by standard global methods.



Paperid:1678
Authors:Yang Liu, Chuanchen Luo, Zimo Tang, Yingyan Li, yuran Yang, Yuanyong Ning, Lue Fan, Junran Peng, ZHAO-XIANG ZHANG
Abstract:
Editing illumination in long videos with complex dynamics has significant value in various downstream tasks, including visual content creation and manipulation, as well as data scaling up for embodied AI through sim2real and real2real transfer. Nevertheless, existing video relighting techniques are predominantly limited to portrait videos or fall into the bottleneck of temporal consistency and computation efficiency. In this paper, we proposeTC-Light, a novel paradigm characterized by the proposed two-stage post optimization mechanism. Starting from the video preliminarily relighted by an inflated video relighting model, it optimizes appearance embedding in the first stage to align global illumination. Then it optimizes the proposed canonical video representation, i.e.,Unique Video Tensor (UVT), to align fine-grained texture and lighting in the second stage. To comprehensively evaluate performance, we also establish a long and highly dynamic video benchmark. Extensive experiments show that our method enables physically plausible relighting results with superior temporal coherence and low computation cost.



Paperid:1679
Authors:Dennis Wagner, Fabian Hartung, Justus Arweiler, Aparna Muraleedharan, Indra Jungjohann, Arjun Nair, Steffen Reithermann, Ralf Schulz, Michael Bortz, Daniel Neider, Heike Leitte, Joachim Pfeffinger, Stephan Mandt, Sophie Burkhardt, Torsten Katz, Fabian Jirasek, Jakob Burger, Hans Hasse, Marius Kloft
Abstract:
Monitoring chemical processes is necessary to prevent catastrophic failures, optimize costs and profits, and ensure the safety of employees and the environment. A key component of modern monitoring systems is the automated detection of anomalies in sensor data over time, called time series, enabling partial automation of plant operation and adding additional layers of supervision to crucial components. The development of anomaly detection methods in this domain is challenging, since real chemical process data is usually proprietary, and simulated data is generally not a sufficient replacement. In this paper, we present NoBOOM, the first collection of datasets for anomaly detection in real-life chemical process data, including labeled data from a running process at a leading industry partner, and several chemical processes run in a laboratory‑scale plant and a pilot‑scale plant. While we are not able to share every detail about the industrial process, for the laboratory‑ and pilot‑scale plants, we provide comprehensive information on plant configuration, processes run, operation, and, in particular, anomaly events, enabling a differentiated analysis of anomaly detection methods. To demonstrate the complexity of the benchmark, we analyze the data with regard to common issues of time-series anomaly detection (TSAD) benchmarks, including triviality and biases.



Paperid:1680
Authors:Xinyu Zhang, Yuxuan Dong, Lingling Zhang, Chengyou Jia, Zhuohang Dang, Basura Fernando, Jun Liu, Mike Zheng Shou
Abstract:
Despite significant advances in Vision Language Models (VLMs), they remain constrained by the complexity and redundancy of visual input.When images contain large amounts of irrelevant information, VLMs are susceptible to interference, thus generating excessive task-irrelevant reasoning processes or even hallucinations.This limitation stems from their inability to discover and process the required regions during reasoning precisely.To address this limitation, we present the Chain of Foresight-Focus Thought (CoFFT), a novel training-free approach that enhances VLMs' visual reasoning by emulating human visual cognition.Each Foresight-Focus Thought consists of three stages:(1) Diverse Sample Generation: generates diverse reasoning samples to explore potential reasoning paths, where each sample contains several reasoning steps;(2) Dual Foresight Decoding: rigorously evaluates these samples based on both visual focus and reasoning progression, adding the first step of optimal sample to the reasoning process; (3) Visual Focus Adjustment: precisely adjust visual focus toward regions most beneficial for future reasoning, before returning to stage (1) to generate subsequent reasoning samples until reaching the final answer.These stages function iteratively, creating an interdependent cycle where reasoning guides visual focus and visual focus informs subsequent reasoning.Empirical results across multiple benchmarks using Qwen2.5-VL, InternVL-2.5, and Llava-Next demonstrate consistent performance improvements of 3.1-5.8\% with controllable increasing computational overhead.



Paperid:1681
Authors:Reza Ebrahimi, Roland Memisevic
Abstract:
The role of hidden units in recurrent neural networks is typically seen as modeling memory, with research focusing on enhancing information retention through gating mechanisms. A less explored perspective views hidden units as active participants in the computation performed by the network, rather than passive memory stores. In this work, we revisit bi-linear operations, which involve multiplicative interactions between hidden units and input embeddings. We demonstrate theoretically and empirically that they constitute a natural inductive bias for representing the evolution of hidden states in state tracking tasks. These are the simplest type of task that require hidden units to actively contribute to the behavior of the network. We also show that bi-linear state updates form a natural hierarchy corresponding to state tracking tasks of increasing complexity, with popular linear recurrent networks such as Mamba residing at the lowest-complexity center of that hierarchy.



Paperid:1682
Authors:Luca Della Libera, Francesco Paissan, Cem Subakan, Mirco Ravanelli
Abstract:
Large language models have revolutionized natural language processing through self-supervised pretraining on massive datasets. Inspired by this success, researchers have explored adapting these methods to speech by discretizing continuous audio into tokens using neural audio codecs. However, existing approaches face limitations, including high bitrates, the loss of either semantic or acoustic information, and the reliance on multi-codebook designs when trying to capture both, which increases architectural complexity for downstream tasks. To address these challenges, we introduce FocalCodec, an efficient low-bitrate codec based on focal modulation that utilizes a single binary codebook to compress speech between 0.16 and 0.65 kbps. FocalCodec delivers competitive performance in speech resynthesis and voice conversion at lower bitrates than the current state-of-the-art, while effectively handling multilingual speech and noisy environments. Evaluation on downstream tasks shows that FocalCodec successfully preserves sufficient semantic and acoustic information, while also being well-suited for generative modeling. Demo samples and code are available at https://anonymous.4open.science/w/fc-web-CC74/.



Paperid:1683
Authors:Slobodan Mitrovic, Theodore Pan, Mahdi Qaempanah, Mohammad Amin Raeisi
Abstract:
Abstract:Finding dense subgraphs is a fundamental problem with applications to community detection, clustering, and data mining. Our work focuses on finding approximate densest subgraphs in directed graphs in computational models for processing massive data. We consider two such models: Massively Parallel Computation (MPC) and semi-streaming. We show how to find a $(2+\varepsilon)$-approximation in $\tilde{O}(\sqrt{\log n})$ MPC rounds with sublinear memory per machine. This improves the state-of-the-art results by Bahmani et al. (WAW 2014) and Mitrovic \& Pan (ICML 2024). Moreover, we show how to find an $O(\log n)$-approximation in a single pass in semi-streaming. This is in stark contrast to prior work, which implies $\tilde{\Omega}(n^{1/6})$-approximation for a single pass; a better approximation is known only for randomized streams (Mitrovi\'c \& Pan). We empirically evaluate our approaches in two ways. First, we illustrate that our single-pass semi-streaming algorithm performs much better than the theoretical guarantee. Specifically, its approximation on temporal datasets matches the $(2+\varepsilon)$-approximation of an $O(\log n)$-pass algorithm by Bahmani et al. (VLDB 2012). Second, we demonstrate that our MPC algorithm requires fewer rounds than prior work.



Paperid:1684
Authors:Diyuan Wu, Aleksandr Shevchenko, Samet Oymak, Marco Mondelli
Abstract:
Abstract:Token embeddings play a crucial role in language modeling but, despite this practical relevance, their theoretical understanding is limited. Our paper addresses the gap by characterizing the structure of embeddings obtained via gradient descent. Specifically, we consider a one-layer softmax attention model with a linear head for binary classification, i.e., $\mathrm{Softmax}( p^\top E_X^\top ) E_X v = \frac{ \sum_{i=1}^T \exp(p^\top E_{x_i}) E_{x_i}^\top v}{\sum_{j=1}^T \exp(p^\top E_{x_{j}}) }$, where $E_X = [ E_{x_1} , \dots, E_{x_T} ]^\top$ contains the embeddings of the input sequence, $p$ is the embedding of the $\mathrm{\langle cls \rangle}$ token and $v$ the output vector. First, we show that, already after a single step of gradient training with the standard logistic loss, the embeddings $E_X$ capture the importance of tokens in the dataset by aligning with the output vector $v$ proportionally to the corresponding average signed frequency that captures the relevance of tokens to the labels. Then, after training $p$ via gradient flow until convergence, the softmax selects the important tokens in the sentence (i.e., those that are predictive of the label), and the resulting $\mathrm{\langle cls \rangle}$ embedding maximizes the margin for such a selection. Experiments on real-world datasets (IMDB, Yelp) exhibit a phenomenology close to that unveiled by our theory.



Paperid:1685
Authors:Liting Lin, Heng Fan, Zhipeng Zhang, Yuqing Huang, Yaowei Wang, Yong Xu, Haibin Ling
Abstract:
Transformer-based algorithms, such as LoRAT, have significantly enhanced object-tracking performance. However, these approaches rely on a standard attention mechanism, which incurs quadratic token complexity, making real-time inference computationally expensive. In this paper, we introduce LoRAT v2, a novel tracking framework that addresses these limitations with three main contributions. First, LoRAT v2 integrates frame-wise causal attention, which ensures full self-attention within each frame while enabling causal dependencies across frames, significantly reducing computational overhead. Moreover, key-value (KV) caching is employed to efficiently reuse past embeddings for further speedup.Second, building on LoRAT's parameter-efficient fine-tuning, we propose Stream-Specific LoRA Adapters (SSLA). As frame-wise causal attention introduces asymmetry in how streams access temporal information, SSLA assigns dedicated LoRA modules to the template and each search stream, with the main ViT backbone remaining frozen. This allows specialized adaptation for each stream's role in temporal tracking.Third, we introduce a two-phase progressive training strategy, which first trains a single-search-frame tracker and then gradually extends it to multi-search-frame inputs by introducing additional LoRA modules. This curriculum-based learning paradigm improves long-term tracking while maintaining training efficiency.In extensive experiments on multiple benchmarks, LoRAT v2 achieves state-of-the-art performance, substantially improved efficiency, and a superior performance-to-FLOPs ratio over state-of-the-art trackers.The source code will be released upon paper publication.



Paperid:1686
Authors:Ali Farajzadeh, Danyal Saeed, Syed M Abbas, Rushit Shah, Aadirupa Saha, Brian Ziebart
Abstract:
Imitation learning seeks policies reflecting the values of demonstrated behaviors. Prevalent approaches learn to match or exceed the demonstrator's performance in expectation without knowing the demonstrator’s reward function. Unfortunately, this does not induce pluralistic imitators that learn to support qualitatively distinct demonstrations. We reformulate imitation learning using stochastic dominance over the demonstrations' reward distribution across a range of reward functions as our foundational aim. Our approach matches imitator policy samples (or support) with demonstrations using optimal transport theory to define an imitation learning objective over trajectory pairs. We demonstrate the benefits of pluralistic stochastic dominance (PSD) for imitation in both theory and practice.



Paperid:1687
Authors:Leonardo Defilippis, Yatin Dandi, Pierre Mergny, Florent Krzakala, Bruno Loureiro
Abstract:
We consider the problem of how many samples from a Gaussian multi-index model are required to weakly reconstruct the relevant index subspace. Despite its increasing popularity as a testbed for investigating the computational complexity of neural networks, results beyond the single-index setting remain elusive. In this work, we introduce spectral algorithms based on the linearization of a message passing scheme tailored to this problem. Our main contribution is to show that the proposed methods achieve the optimal reconstruction threshold. Leveraging a high-dimensional characterization of the algorithms, we show that above the critical threshold the leading eigenvector correlates with the relevant index subspace, a phenomenon reminiscent of the Baik–Ben Arous–Peche (BBP) transition in spiked models arising in random matrix theory. Supported by numerical experiments and a rigorous theoretical framework, our work bridges critical gaps in the computational limits of weak learnability in multi-index model.



Paperid:1688
Authors:Jin Seong, Jiyun Park, Wencke Liermann, Hongseok Choi, Yoonji Nam, Hyun Kim, Soojong Lim, Namhoon Lee
Abstract:
The dynamic nature of information necessitates continuously updating large vision-language models (LVLMs).While recent knowledge editing techniques hint at promising directions, they often focus on editing a single modality (vision or language) in isolation. This prevalent practice neglects the inherent multimodality of LVLMs and the continuous nature of knowledge updates, potentially leading to suboptimal editing outcomes when considering the interplay between modalities and the need for ongoing knowledge refinement. To address these limitations, we propose MemEIC, a novel method for Continual and Compositional Knowledge Editing (CCKE) in LVLMs. MemEIC enables compositional editing of both visual and textual knowledge sequentially. Our approach employs a hybrid external-internal editor featuring a dual external memory for cross-modal evidence retrieval and dual LoRA adapters that facilitate disentangled parameter updates for each modality. A key component is a brain-inspired knowledge connector, activated selectively for compositional reasoning, that integrates information across different modalities. Experiments demonstrate that MemEIC significantly improves performance on complex multimodal questions and effectively preserves prior edits, setting a new benchmark for CCKE in LVLMs.



Paperid:1689
Authors:Shuning Chang, Pichao WANG, Jiasheng Tang, Fan Wang, Yi Yang
Abstract:
Abstract:Diffusion Transformers (DiT) are renowned for their impressive generative performance; however, they are significantly constrained by considerable computational costs due to the quadratic complexity in self-attention and the extensive sampling steps required. While advancements have been made in expediting the sampling process, the underlying architectural inefficiencies within DiT remain underexplored. We introduce SparseDiT, a novel framework that implements token sparsification across spatial and temporal dimensions to enhance computational efficiency while preserving generative quality. Spatially, SparseDiT employs a tri-segment architecture that allocates token density based on feature requirements at each layer: Poolingformer in the bottom layers for efficient global feature extraction, Sparse-Dense Token Modules (SDTM) in the middle layers to balance global context with local detail, and dense tokens in the top layers to refine high-frequency details. Temporally, SparseDiT dynamically modulates token density across denoising stages, progressively increasing token count as finer details emerge in later timesteps. This synergy between SparseDiT’s spatially adaptive architecture and its temporal pruning strategy enables a unified framework that balances efficiency and fidelity throughout the generation process. Our experiments demonstrate SparseDiT’s effectiveness, achieving a 55\% reduction in FLOPs and a 175\% improvement in inference speed on DiT-XL with similar FID score on 512$\times$512 ImageNet, a 56\% reduction in FLOPs across video generation datasets, and a 69\% improvement in inference speed on PixArt-$\alpha$ on text-to-image generation task with a 0.24 FID score decrease. SparseDiT provides a scalable solution for high-quality diffusion-based generation compatible with sampling optimization techniques.



Paperid:1690
Authors:Zishi Zhang, Tao Ren, Yijie Peng
Abstract:
Abstract:In this work, we consider the problem of Bayesian Optimization (BO) under reward model uncertainty—that is, when the underlying distribution type of the reward is unknown and potentially intractable to specify. This challenge is particularly evident in many modern applications, where the reward distribution is highly ill-behaved, often non-stationary, multi-modal, or heavy-tailed.In such settings, classical Gaussian Process (GP)-based BO methods often fail due to their strong modeling assumptions. To address this challenge, we propose a novel surrogate model, the infinity-Gaussian Process ($\infty$-GP), which represents a sequential spatial Dirichlet Process mixture with a GP baseline. The $\infty$-GP quantifies both value uncertainty and model uncertainty, enabling more flexible modeling of complex reward structures. Combined with Thompson Sampling, the $\infty$-GP facilitates principled exploration and exploitation in the distributional space of reward models. Theoretically, we prove that the $\infty$-GP surrogate model can approximate a broad class of reward distributions by effectively exploring the distribution space, achieving near-minimax-optimal posterior contraction rates. Empirically, our method outperforms state-of-the-art approaches in various challenging scenarios, including highly non-stationary and heavy-tailed reward settings where classical GP-based BO often fails.



Paperid:1691
Authors:Mengjia Niu, Hamed Haddadi, Guansong Pang
Abstract:
Hallucinations in large language models (LLMs) pose significant safety concerns that impede their broader deployment. Recent research in hallucination detection has demonstrated that LLMs' internal representations contain truthfulness hints, which can be harnessed for detector training. However, the performance of these detectors is heavily dependent on the internal representations of predetermined tokens, fluctuating considerably when working on free-form generations with varying lengths and sparse distributions of hallucinated entities.To address this, we propose HaMI, a novel approach that enables robust detection of hallucinations through adaptive selection and learning of critical tokens that are most indicative of hallucinations. We achieve this robustness by an innovative formulation of the Hallucination detection task as Multiple Instance (HaMI) learning over token-level representations within a sequence, thereby facilitating a joint optimisation of token selection and hallucination detection on generation sequences of diverse forms. Comprehensive experimental results on four hallucination benchmarks show that HaMI significantly outperforms existing state-of-the-art approaches.



Paperid:1692
Authors:Luyang Zhang, Cathy Jiao, Beibei Li, Chenyan Xiong
Abstract:
Training data is the backbone of large language models (LLMs), yet today’s data markets often operate under exploitative pricing -- sourcing data from marginalized groups with little pay or recognition. This paper introduces a theoretical framework for LLM data markets, modeling the strategic interactions between buyers (LLM builders) and sellers (human annotators). We begin with theoretical and empirical analysis showing how exploitative pricing drives high-quality sellers out of the market, degrading data quality and long-term model performance. Then we introduce fairshare, a pricing mechanism grounded in data valuation that quantifies each data’s contribution. It aligns incentives by sustaining seller participation and optimizing utility for both buyers and sellers. Theoretically, we show that fairshare yields mutually optimal outcomes: maximizing long-term buyer utility and seller profit while sustaining market participation. Empirically when training open-source LLMs on complex NLP tasks, including math problems, medical diagnosis, and physical reasoning,fairshare boosts seller earnings and ensures a stable supply of high-quality data, while improving buyers’ performance-per-dollar and long-term welfare. Our findings offer a concrete path toward fair, transparent, and economically sustainable data markets for LLM. Our code will be open sourced.



Paperid:1693
Authors:Yuyan Ni, Shikun Feng, Haohan Chi, Bowen Zheng, Huan-ang Gao, Wei-Ying Ma, Zhi-Ming Ma, Yanyan Lan
Abstract:
Diffusion-based models have shown great promise in molecular generation but often require a large number of sampling steps to generate valid samples. In this paper, we introduce a novel Straight-Line Diffusion Model (SLDM) to tackle this problem, by formulating the diffusion process to follow a linear trajectory. The proposed process aligns well with the noise sensitivity characteristic of molecular structures and uniformly distributes reconstruction effort across the generative process, thus enhancing learning efficiency and efficacy. Consequently, SLDM achieves state-of-the-art performance on 3D molecule generation benchmarks, delivering a 100-fold improvement in sampling efficiency.



Paperid:1694
Authors:Irene Wang, Mostafa Elhoushi, H Ekin Sumbul, Samuel Hsia, Daniel Jiang, Newsha Ardalani, Divya Mahajan, Carole-Jean Wu, Bilge Acun
Abstract:
Machine learning solutions are rapidly adopted to enable a variety of key use cases, from conversational AI assistants to scientific discovery. As the adoption of machine learning models becomes increasingly prevalent, the associated lifecycle carbon footprint is expected to increase, including bothoperational carbonfrom training and inference andembodied carbonfrom AI hardware manufacturing. We introduce CATransformers, the first carbon-aware co-optimization framework for Transformer-based models and hardware accelerators. By integrating both operational and embodied carbon into early-stage design space exploration, CATransformers enables sustainability-driven model architecture and hardware accelerator co-design that reveals fundamentally different trade-offs than latency- or energy-centric approaches. Evaluated across a range of Transformer models, CATransformers consistently demonstrates the potential to reduce total carbon emissions --by up to 30\% -- while maintaining accuracy and latency. We further highlight its extensibility through a focused case study on multi-modal models. Our results emphasize the need for holistic optimization methods that prioritize carbon efficiency without compromising model capability and execution time performance. Our framework will be open-sourced.



Paperid:1695
Authors:Ehsan Sharifian, Saber Salehkaleybar, Negar Kiyavash
Abstract:
We study the problem of causal structure learning from a combination of observational and interventional data generated by a linear non-Gaussian structural equation model that might contain cycles. Recent results show that using mere observational data identifies the causal graph only up to a permutation-equivalence class. We obtain a combinatorial characterization of this class by showing that each equivalence class corresponds to a perfect matching in a bipartite graph. This bipartite representation allows us to analyze how interventions modify or constrain the matchings. Specifically, we show that each atomic intervention reveals one edge of the true matching and eliminates all incompatible causal graphs. Consequently, we formalize the optimal experiment design task as an adaptive stochastic optimization problem over the set of equivalence classes with a natural reward function that quantifies how many graphs are eliminated from the equivalence class by an intervention. We show that this reward function is adaptive submodular and provide a greedy policy with a provable near-optimal performance guarantee. A key technical challenge is to efficiently estimate the reward function without having to explicitly enumerate all the graphs in the equivalence class. We propose a sampling-based estimator using random matchings and analyze its bias and concentration behavior. Our simulation results show that performing a small number of interventions guided by our stochastic optimization framework recovers the true underlying causal structure.



Paperid:1696
Authors:Muralikrishnna Guruswamy Sethuraman, Faramarz Fekri
Abstract:
Understanding causal relationships between variables is fundamental across scientific disciplines. Most causal discovery algorithms rely on two key assumptions: (i) all variables are observed, and (ii) the underlying causal graph is acyclic. While these assumptions simplify theoretical analysis, they are often violated in real-world systems, such as biological networks. Existing methods that account for confounders either assume linearity or struggle with scalability. To address these limitations, we propose DCCD-CONF, a novel framework for differentiable learning of nonlinear cyclic causal graphs in the presence of unmeasured confounders using interventional data. Our approach alternates between optimizing the graph structure and estimating the confounder distribution by maximizing the log-likelihood of the data. Through experiments on synthetic data and real-world gene perturbation datasets, we show that DCCD-CONF outperforms state-of-the-art methods in both causal graph recovery and confounder identification. Additionally, we also provide consistency guarantees for our framework, reinforcing its theoretical soundness.



Paperid:1697
Authors:Chuofan Ma, Yi Jiang, Junfeng Wu, Jihan Yang, Xin Yu, Zehuan Yuan, BINGYUE PENG, Xiaojuan Qi
Abstract:
Abstract:Visual generative and understanding models typically rely on distinct tokenizers to process images, presenting a key challenge for unifying them within a single framework. Recent studies attempt to address this by connecting the training of VQVAE (for autoregressive generation) and CLIP (for understanding) to build a unified tokenizer. However, directly combining these training objectives has been observed to cause severe loss conflicts. In this paper, we show that reconstruction and semantic supervision do not inherently conflict. Instead, the underlying bottleneck stems from limited representational capacity of discrete token space. Building on these insights, we introduce UniTok, a unified tokenizer featuring a novel multi-codebook quantization mechanism that effectively scales up the vocabulary size and bottleneck dimension. In terms of final performance, UniTok sets a new record of 0.38 rFID and 78.6\% zero-shot accuracy on ImageNet. Besides, UniTok can be seamlessly integrated into MLLMs to unlock native visual generation capability, without compromising the understanding performance. Additionally, we show that UniTok favors cfg-free generation, reducing gFID from 14.6 to 2.5 on ImageNet 256$\times$256 benchmark. All codes and models have been made publicly available.



Paperid:1698
Authors:Lijun Zhang, Lin Li, Yajie Qi, Huizhong Song, Yaodong Yang, Jun Wang, Wei Wei
Abstract:
When fine-tuning pre-trained Large Language Models (LLMs) to align with human values and intentions, maximizing the estimated reward can lead to superior performance, but it also introduces potential risks due to deviations from the reference model's intended behavior. Most existing methods typically introduce KL divergence to constrain deviations between the trained model and the reference model; however, this may not be sufficient in certain applications that require tight risk control. In this paper, we introduce Risk-aware Direct Preference Optimization (Ra-DPO), a novel approach that incorporates risk-awareness by employing a class of nested risk measures. This approach formulates a constrained risk-aware advantage function maximization problem and then converts the Bradley-Terry model into a token-level representation. The objective function maximizes the likelihood of the policy while suppressing the deviation between a trained model and the reference model using a sequential risk ratio, thereby enhancing the model's risk-awareness. Experimental results across three open-source datasets: IMDb Dataset, Anthropic HH Dataset, and AlpacaEval, demonstrate the proposed method's superior performance in balancing alignment performance and model drift.



Paperid:1699
Authors:Canyu Zhao, Mingyu Liu, Huanyi Zheng, Muzhi Zhu, Zhiyue Zhao, Yanlong Sun, Hao Chen, Tong He, Chunhua Shen
Abstract:
This paper's primary objective is to develop a robust generalist perception model capable of addressing multiple tasks under constraints of computational resources and limited training data. We leverage text-to-image diffusion models pre-trained on billions of images and successfully introduce our DICEPTION, a visual generalist model. Exhaustive evaluations demonstrate that DICEPTION effectively tackles diverse perception tasks, even achieving performance comparable to SOTA single-task specialist models. Specifically, we achieve results on par with SAM-vit-h using only 0.06% of their data (e.g., 600K vs.\ 1B pixel-level annotated images). We designed comprehensive experiments on architectures and input paradigms, demonstrating that the key to successfully re-purposing a single diffusion model for multiple perception tasks lies in maximizing the preservation of the pre-trained model's prior knowledge. Consequently, DICEPTION can be trained with substantially lower computational costs than conventional models requiring training from scratch. Furthermore, adapting DICEPTION to novel tasks is highly efficient, necessitating fine-tuning on as few as 50 images and approximately 1% of its parameters. Finally, we demonstrate that a subtle application of classifier-free guidance can improve the model's performance on depth and normal estimation. We also show that pixel-aligned training, as is characteristic of perception tasks, significantly enhances the model's ability to preserve fine details. DICEPTION offers valuable insights and presents a promising direction for the development of advanced diffusion-based visual generalist models.



Paperid:1700
Authors:Xun Deng, Han Zhong, Rui Ai, Fuli Feng, Zheng Wang, Xiangnan He
Abstract:
Direct Preference Optimization (DPO) has emerged as a promising approach for aligning large language models with human preferences. While prior work mainly extends DPO from the aspect of the objective function, we instead improve DPO from the largely overlooked but critical aspect of data selection. Specifically, we address the issue of parameter shrinkage caused by noisy data by proposing a novel margin-maximization principle for dataset curation in DPO training. To further mitigate the noise in different reward models, we propose a Bayesian Aggregation approach that unifies multiple margin sources (external and implicit) into a single preference probability. Extensive experiments in diverse settings demonstrate the consistently high data efficiency of our approach. Remarkably, by using just 10\% of the Ultrafeedback dataset, our approach achieves 3\% to 8\% improvements across various Llama, Mistral, and Qwen models on the AlpacaEval2 benchmark. Furthermore, our approach seamlessly extends to iterative DPO, yielding a roughly 3\% improvement with 25\% online data, revealing the high redundancy in this presumed high-quality data construction manner. These results highlight the potential of data selection strategies for advancing preference optimization.



Paperid:1701
Authors:Yinuo Jiang, Yan Xiaodong, Keyan Ding, Deng Zhao, Lei Liang, Qiang Zhang, Huajun Chen
Abstract:
Abstract:Parameter-efficient fine-tuning (PEFT) methods, such as LoRA, have enabled the efficient adaptation of large language models (LLMs) by updating only a small subset of parameters. However, their robustness under out-of-distribution (OOD) conditions remains insufficiently studied. In this paper, we identify the limitations of conventional LoRA in handling distributional shifts and propose $\textbf{HiMoLE}$($\textbf{Hi}$erarchical $\textbf{M}$ixture of $\textbf{L}$oRA $\textbf{E}$xperts), a new framework designed to improve OOD generalization. HiMoLE integrates hierarchical expert modules and hierarchical routing strategies into the LoRA architecture and introduces a two-phase training procedure enhanced by a diversity-driven loss. This design mitigates negative transfer and promotes effective knowledge adaptation across diverse data distributions. We evaluate HiMoLE on three representative tasks in natural language processing. Experimental results evidence that HiMoLE consistently outperforms existing LoRA-based approaches, significantly reducing performance degradation on OOD data while improving in-distribution performance. Our work bridges the gap between parameter efficiency and distributional robustness, advancing the practical deployment of LLMs in real-world applications.



Paperid:1702
Authors:Zongbo Han, Jialong Yang, Guangyu Wang, Junfan Li, Qianli Xu, Mike Zheng Shou, Changqing Zhang
Abstract:
Vision-language foundation models (VLMs), such as CLIP, exhibit remarkable performance across a wide range of tasks. However, deploying these models can be unreliable when significant distribution gaps exist between training and test data, while fine-tuning for diverse scenarios is often costly. Cache-based test-time adapters offer an efficient alternative by storing representative test samples to guide subsequent classifications. Yet, these methods typically employ naive cache management with limited capacity, leading to severe catastrophic forgetting when samples are inevitably dropped during updates. In this paper, we propose DOTA (DistributiOnal Test-time Adaptation), a simple yet effective method addressing this limitation. Crucially, instead of merely memorizing individual test samples, DOTA continuously estimates the underlying distribution of the test data stream. Test-time posterior probabilities are then computed using these dynamically estimated distributions via Bayes' theorem for adaptation. This distribution-centric approach enables the model to continually learn and adapt to the deployment environment. Extensive experiments validate that DOTA significantly mitigates forgetting and achieves state-of-the-art performance compared to existing methods.



Paperid:1703
Authors:Zihan Su, Xuerui Qiu, Hongbin Xu Xu, Tangyu Jiang, Jun-hao Zhuang, Chun Yuan, Ming Li, Shengfeng He, Fei Yu
Abstract:
The explosive growth of generative video models has amplified the demand forreliable copyright preservation of AI-generated content. Despite its popularity inimage synthesis, invisible generative watermarking remains largely underexploredin video generation. To address this gap, we propose Safe-Sora, the first frameworkto embed graphical watermarks directly into the video generation process. Motivated by the observation that watermarking performance is closely tied to the visualsimilarity between the watermark and cover content, we introduce a hierarchicalcoarse-to-fine adaptive matching mechanism. Specifically, the watermark image isdivided into patches, each assigned to the most visually similar video frame, andfurther localized to the optimal spatial region for seamless embedding. To enablespatiotemporal fusion of watermark patches across video frames, we develop a 3Dwavelet transform-enhanced Mamba architecture with a novel scanning strategy,effectively modeling long-range dependencies during watermark embedding andretrieval. To the best of our knowledge, this is the first attempt to apply state spacemodels to watermarking, opening new avenues for efficient and robust watermarkprotection. Extensive experiments demonstrate that Safe-Sora achieves state-of-the-art performance in terms of video quality, watermark fidelity, and robustness, whichis largely attributed to our proposals. Code and additional supporting materials areprovided in the supplementary.



Paperid:1704
Authors:Zhao-Heng Yin, Sherry Yang, Pieter Abbeel
Abstract:
Abstract:Learning robot control policies from human videos is a promising direction for scaling up robot learning. However, how to extract action knowledge (or action representations) from videos for policy learning remains a key challenge. Existing action representations such as video frames, pixelflow, and pointcloud flow have inherent limitations such as modeling complexity or loss of information. In this paper, we propose to use object-centric 3D motion field to represent actions for robot learning from human videos, and present a novel framework for extracting this representation from videos for zero-shot control. We introduce two novel components. First, a novel training pipeline for training a ``denoising'' 3D motion field estimator to extract fine object 3D motions from human videos with noisy depth robustly. Second, a dense object-centric 3D motion field prediction architecture that favors both cross-embodiment transfer and policy generalization to background. We evaluate the system in real world setups. Experiments show that our method reduces 3D motion estimation error by over 50% compared to the latest method, achieve 55% average success rate in diverse tasks where prior approaches fail ($\lesssim 10$\%), and can even acquire fine-grained manipulation skills like insertion.



Paperid:1705
Authors:Zehan Wang, Ziang Zhang, Jiayang Xu, Jialei Wang, Tianyu Pang, Chao Du, Hengshuang Zhao, Zhou Zhao
Abstract:
This work presents Orient Anything V2, an enhanced foundation model for unified understanding of object 3D orientation and rotation from single or paired images. Building upon Orient Anything V1, which defines orientation via a single unique front face, V2 extends this capability to handle objects with diverse rotational symmetries and directly estimate relative rotations. These improvements are enabled by four key innovations: 1) Scalable 3D assets synthesized by generative models, ensuring broad category coverage and balanced data distribution; 2) An efficient, model-in-the-loop annotation system that robustly identifies 0 to N valid front faces for each object; 3) A symmetry-aware, periodic distribution fitting objective that captures all plausible front-facing orientations, effectively modeling object rotational symmetry; 4) A multi-frame architecture that directly predicts relative object rotations. Extensive experiments show that Orient Anything V2 achieves state-of-the-art zero-shot performance on orientation estimation, 6DoF pose estimation, and object symmetry recognition across 11 widely used benchmarks. The model demonstrates strong generalization, significantly broadening the applicability of orientation estimation in diverse downstream tasks.



Paperid:1706
Authors:Shengyuan Chen, Zheng Yuan, Qinggang Zhang, Wen Hua, Jiannong Cao, Xiao Huang
Abstract:
Entity alignment (EA) aims to merge two knowledge graphs (KGs) by identifying equivalent entity pairs. Existing methods can be categorized into symbolic and neural models. Symbolic models, while precise, struggle with substructure heterogeneity and sparsity, whereas neural models, although effective, generally lack interpretability and cannot handle uncertainty. We propose NeuSymEA, a unified neuro-symbolic reasoning framework that combines the strengths of both methods and fully exploits the cross-KG structural pattern for robust entity alignment. NeuSymEA models the joint probability of all possible pairs' truth scores in a Markov random field, regulated by a set of rules, and optimizes it with the variational EM algorithm. In the E-step, a neural model parameterizes the truth score distributions and infers missing alignments. In the M-step, the rule weights are updated based on the observed and inferred alignments. To facilitate interpretability, we further design a path-ranking-based explainer upon this framework that generates supporting rules for the inferred alignments. Experiments on benchmarks demonstrate that NeuSymEA not only significantly outperforms baselines in terms of effectiveness and robustness, but also provides interpretable results.



Paperid:1707
Authors:Fang Li, Hao Zhang, Narendra Ahuja
Abstract:
Abstract:Although COLMAP has long remained the predominant method for camera parameter optimization in static scenes, it is constrained by its lengthy runtime and reliance on ground truth (GT) motion masks for application to dynamic scenes. Many efforts attempted to improve it by incorporating more priors as supervision such as GT focal length, motion masks, 3D point clouds, camera poses, and metric depth, which, however, are typically unavailable in casually captured RGB videos. In this paper, we propose a novel method for more $\textbf{accurate}$ and $\textbf{efficient}$ camera parameter optimization in dynamic scenes $\textbf{solely supervised by a single RGB video}$. Our method consists of three key components: (1) $\textit{Patch-wise Tracking Filters}$, to establish robust and maximally sparse hinge-like relations across the RGB video. (2) $\textit{Outlier-aware Joint Optimization}$, for efficient camera parameter optimization by adaptive down-weighting of moving outliers, without reliance on motion priors. (3) A $\textit{Two-stage Optimization Strategy}$, to enhance stability and optimization speed by a trade-off between the Softplus limits and convex minima in losses. We visually and numerically evaluate our camera estimates. To further validate accuracy, we feed the camera estimates into a 4D reconstruction method and assess the resulting 3D scenes, and rendered 2D RGB and depth maps. We perform experiments on 4 real-world datasets (NeRF-DS, DAVIS, iPhone, and TUM-dynamics) and 1 synthetic dataset (MPI-Sintel), demonstrating that our method estimates camera parameters more efficiently and accurately with a single RGB video as the only supervision. Code will be released upon acceptance.



Paperid:1708
Authors:Xuejie Liu, Anji Liu, Guy Van den Broeck, Yitao Liang
Abstract:
Abstract:Masked generative models (MGMs) have emerged as a powerful framework for image synthesis, combining parallel decoding with strong bidirectional context modeling. However, generating high-quality samples typically requires many iterative decoding steps, resulting in high inference costs. A straightforward way to speed up generation is by decoding more tokens in each step, thereby reducing the total number of steps. However, when many tokens are decoded simultaneously, the model can only estimate the univariate marginal distributions independently, failing to capture the dependency among them. As a result, reducing the number of steps significantly compromises generation fidelity. In this work, we introduce ReCAP (Reused Context-Aware Prediction), a plug-and-play module that accelerates inference in MGMs by constructing low-cost steps via reusing feature embeddings from previously decoded context tokens. ReCAP interleaves standard full evaluations with lightweight steps that cache and reuse context features, substantially reducing computation while preserving the benefits of fine-grained, iterative generation. We demonstrate its effectiveness on top of three representative MGMs (MaskGIT, MAGE, and MAR), including both discrete and continuous token spaces and covering diverse architectural designs. In particular, on ImageNet256 class-conditional generation, ReCAP achieves up to 2.4$\times$ faster inference than the base model with minimal performance drop, and consistently delivers better efficiency–fidelity trade-offs under various generation settings.



Paperid:1709
Authors:David Bruns-Smith, Zhongming Xie, Avi Feller
Abstract:
Multiaccuracy, developed in fair machine learning, provides a framework for reducing predictive bias uniformly over subpopulations defined by an auditing class. Recent work shows that multiaccurate estimators trained only on source data can remain low-bias under unknown covariate shifts—a property known as "Universal Adaptability". Building on this, we show that when the auditing class is a Hilbert space, a simple multiaccurate ridge-boosting estimator is numerically equivalent to an ``automatic debiased machine learning'' estimator. As a result, it inherits semiparametric efficiency and optimal asymptotic variance. In this setting, the multiaccurate estimator is not only robust under unknown shifts --- it is also efficient, even without access to data from the target distribution.



Paperid:1710
Authors:Javier E. Santos, Agnese Marcato, Roman Colman, NIcholas Lubbers, Yen Ting Lin
Abstract:
Generative diffusion models have achieved remarkable success in producing high-quality images. However, because these models typically operate in continuous intensity spaces—diffusing independently per pixel and color channel—they are fundamentally ill-suited for applications where quantities such as particle counts or material units are inherently discrete and governed by strict conservation laws like mass preservation, which limits their applicability in scientific workflows. To address this limitation, we propose Discrete Spatial Diffusion (DSD), a framework based on a continuous-time, discrete-state jump stochastic process that operates directly in discrete spatial domains while strictly preserving mass in both forward and reverse diffusion processes. By using spatial diffusion to achieve mass preservation, we introduce stochasticity naturally through a discrete formulation. We demonstrate the expressive flexibility of DSD by performing image synthesis, class conditioning, and image inpainting across widely-used image benchmarks, with the ability to condition on image intensity. Additionally, we highlight its applicability to domain-specific scientific data for materials microstructure, bridging the gap between diffusion models and mass-conditioned scientific applications.



Paperid:1711
Authors:Yuxuan Zhu, Tengjun Jin, Yada Pruksachatkun, Andy Zhang, Shu Liu, Sasha Cui, Sayash Kapoor, Shayne Longpre, Kevin Meng, Rebecca Weiss, Fazl Barez, Rahul Gupta, Jwala Dhamala, Jacob Merizian, Mario Giulianelli, Harry Coppock, Cozmin Ududec, Antony Kellermann, Jasjeet Sekhon, Jacob Steinhardt, Sarah Schwettmann, Arvind Narayanan, Matei A Zaharia, Ion Stoica, Percy Liang, Daniel Kang
Abstract:
Benchmarks are essential tools for quantitatively tracking progress in AI. As AI agents become increasingly capable of solving complex, real-world tasks, researchers and practitioners have introduced agentic benchmarks that evaluate agents based on task outcomes. However, we show that using outcome-based design suboptimally can misrepresent the true capabilities of agents. For example, SWE-bench-Verified uses insufficient test cases, while τ -bench counts empty responses as successful. Such flaws can lead to under- or overestimation agents’ performance by up to 100% in relative terms. To address this issue, we introduce the Agentic Benchmark Checklist (ABC), a set of benchmark development guidelines that we synthesized from our benchmark-building experience, a survey of best practices, and previously reported issues. When applied to CVE-bench, a benchmark with a particularly complex evaluation design, ABC reduces the performance overestimation by 33%.



Paperid:1712
Authors:Jisu Nam, Soowon Son, Dahyun Chung, Jiyoung Kim, Siyoon Jin, Junhwa Hur, Seungryong Kim
Abstract:
Recent advancements in video diffusion models based on Diffusion Transformers (DiTs) have achieved remarkable success in generating temporally coherent videos. Yet, a fundamental question persists: how do these models internally establish and represent temporal correspondences across frames? We introduce DiffTrack, the first quantitative analysis framework designed to answer this question. DiffTrack constructs a dataset of prompt-generated video with pseudo ground-truth tracking annotations and proposes novel evaluation metrics to systematically analyze how each component within the full 3D attention mechanism of DiTs (e.g., representations, layers, and timesteps) contributes to establishing temporal correspondences. Our analysis reveals that query-key similarities in specific (but not all) layers play a critical role in temporal matching, and that this matching becomes increasingly prominent throughout denoising. We demonstrate practical applications of DiffTrack in zero-shot point tracking, where it achieves state-of-the-art performance compared to existing vision foundation and self-supervised video models. Further, we extend our findings to motion-enhanced video generation with a novel guidance method that improves temporal consistency of generated videos without additional training. We believe our work offers crucial insights into the inner workings of video DiTs and establishes a foundation for further research and applications leveraging their temporal understanding.



Paperid:1713
Authors:Kianté Brantley, Mingyu Chen, Zhaolin Gao, Jason Lee, Wen Sun, Wenhao Zhan, Xuezhou Zhang
Abstract:
Abstract:Reinforcement learning (RL) has emerged as a powerful tool for fine-tuning large language models (LLMs) to improve complex reasoning abilities. However, state-of-the-art policy optimization methods often suffer from high computational overhead and memory consumption, primarily due to the need for multiple generations per prompt and the reliance on critic networks or advantage estimates of the current policy. In this paper, we propose $A^\star$-PO, a novel two-stage policy optimization framework that directly approximates the optimal advantage function and enables efficient training of LLMs for reasoning tasks. In the first stage, we leverage offline sampling from a reference policy to estimate the optimal value function $V^{\star}$, eliminating the need for costly online value estimation. In the second stage, we perform on-policy updates using a simple least-squares regression loss with only a single generation per prompt. Theoretically, we establish performance guarantees and prove that the KL-regularized RL objective can be optimized without requiring complex exploration strategies. Empirically, $A^\star$-PO achieves competitive performance across a wide range of mathematical reasoning benchmarks, while reducing training time by up to 2$\times$ and peak memory usage by over 30\% compared to PPO, GRPO, and REBEL.



Paperid:1714
Authors:Guiyao Tie, Zenghui Yuan, Zeli Zhao, Chaoran Hu, Tianhe Gu, Ruihang Zhang, Sizhe Zhang, Junran Wu, Xiaoyue Tu, Ming Jin, Qingsong Wen, Lixing Chen, Pan Zhou, Lichao Sun
Abstract:
Self-correction of large language models (LLMs) emerges as a critical component for enhancing their reasoning performance. Although various self-correction methods have been proposed, a comprehensive evaluation of these methods remains largely unexplored, and the question of whether LLMs can truly correct themselves is a matter of significant interest and concern. In this study, we introduceCorrectBench, a benchmark developed to evaluate the effectiveness of self-correction strategies, including intrinsic, external, and fine-tuned approaches, across three tasks: commonsense reasoning, mathematical reasoning, and code generation. Our findings reveal that: 1) Self-correction methods can improve accuracy, especially for complex reasoning tasks; 2) Mixing different self-correction strategies yields further improvements, though it reduces efficiency; 3) Reasoning LLMs (e.g., DeepSeek-V3) have limited optimization under additional self-correction methods and have high time costs. Interestingly, a comparatively simple chain-of-thought (CoT) baseline demonstrates competitive accuracy and efficiency. These results underscore the potential of self-correction to enhance LLM's reasoning performance while highlighting the ongoing challenge of improving their efficiency. Consequently, we advocate for further research focused on optimizing the balance between reasoning capabilities and operational efficiency.



Paperid:1715
Authors:Nathan Stromberg, Christos Thrampoulidis, Lalitha Sankar
Abstract:
While machine learning models become more capable in discriminative tasks at scale, their ability to overcome biases introduced by training data has come under increasing scrutiny. Previous results suggest that there are two extremes of parameterization with very different behaviors: the population (underparameterized) setting where loss weighting is optimal and the separable overparameterized setting where loss weighting is ineffective at ensuring equal performance across classes. This work explores the regime of last layer retraining (LLR) in which the unseen limited (retraining) data is frequently inseparable and the model proportionately sized, falling between the two aforementioned extremes. We show, in theory and practice, that loss weighting is still effective in this regime, but that these weightsmusttake into account the relative overparameterization of the model.



Paperid:1716
Authors:Animesh Gupta, Jay Parmar, Ishan Rajendrakumar Dave, Mubarak Shah
Abstract:
Composed Video Retrieval (CoVR) retrieves a target video given a query video and a modification text describing the intended change. Existing CoVR benchmarks emphasize appearance shifts or coarse event changes and therefore do not test the ability to capture subtle, fast-paced temporal differences. We introduce TF-CoVR, the first large-scale benchmark dedicated to temporally fine-grained CoVR. TF-CoVR focuses on gymnastics and diving and provides 1.8 M triplets drawn from FineGym and FineDiving. Previous CoVR benchmarks focusing on temporal aspect, link each query to a single target segment taken from the same video, limiting practical usefulness. In TF-CoVR, we instead construct eachpair by prompting an LLM with the label differences between clips drawn from different videos; every pair is thus associated with multiple valid target videos (3.9 on average), reflecting real-world tasks such as sports-highlight generation. To model these temporal dynamics we propose TF-CoVR-Base, a concise two-stage training framework: (i) pre-train a video encoder on fine-grained action classification to obtain temporally discriminative embeddings; (ii) align the composed query with candidate videos using contrastive learning. We conduct the first comprehensive study of image, video, and general multimodal embedding (GME) models on temporally fine-grained composed retrieval in both zero-shot and fine-tuning regimes. On TF-CoVR, TF-CoVR-Base improves zero-shot mAP@50 from 5.92 (LanguageBind) to 7.51, and after fine-tuning raises the state of the art from 19.83 to 25.82.



Paperid:1717
Authors:Zhiyuan Fan, Arnab Maiti, Lillian Ratliff, Kevin Jamieson, Gabriele Farina
Abstract:
Abstract:In this paper, we study the classical Hedge algorithm in combinatorial settings. In each round, the learner selects a vector $\mathbf{x}_t$ from a set $\mathcal{X} \subseteq$ {$0,1$}$^d$, observes a full loss vector $\mathbf{y}_t \in \mathbb{R}^d$, and incurs a loss $\langle \mathbf{x}_t, \mathbf{y}_t \rangle \in [-1,1]$. This setting captures several important problems, including extensive-form games, resource allocation, $m$-sets, online multitask learning, and shortest-path problems on directed acyclic graphs (DAGs). It is well known that Hedge achieves a regret of $\mathcal{O}\big(\sqrt{T \log |\mathcal{X}|}\big)$ after $T$ rounds of interaction. In this paper, we ask whether Hedge is optimal across all combinatorial settings. To that end, we show that for any $\mathcal{X} \subseteq$ {$0,1$}$^d$, Hedge is near-optimal—specifically, up to a $\sqrt{\log d}$ factor—by establishing a lower bound of $\Omega\big(\sqrt{T \log(|\mathcal{X}|)/\log d}\big)$ that holds for any algorithm. We then identify a natural class of combinatorial sets—namely, $m$-sets with $\log d \leq m \leq \sqrt{d}$—for which this lower bound is tight, and for which Hedge is provably suboptimal by a factor of exactly $\sqrt{\log d}$. At the same time, we show that Hedge is optimal for online multitask learning, a generalization of the classical $K$-experts problem. Finally, we leverage the near-optimality of Hedge to establish the existence of a near-optimal regularizer for online shortest-path problems in DAGs—a setting that subsumes a broad range of combinatorial domains. Specifically, we show that the classical Online Mirror Descent (OMD) algorithm, when instantiated with the dilated entropy regularizer, is iterate-equivalent to Hedge, and therefore inherits its near-optimal regret guarantees for DAGs.



Paperid:1718
Authors:Jiahuan Wang, Yuxin Chen, Jun Yu, Guangming Lu, Wenjie Pei
Abstract:
Adapting pretrained diffusion-based generative models for text-driven image editing with negligible tuning overhead has demonstrated remarkable potential. A classical adaptation paradigm, as followed by these methods, first infers the generative trajectory inversely for a given source image by image inversion, then performs image editing along the inferred trajectory guided by the target text prompts. However, the performance of image editing is heavily limited by the approximation errors introduced during image inversion by diffusion models, which arise from the absence of exact supervision in the intermediate generative steps. To circumvent this issue, we investigate the parameter-efficient adaptation of VQ-based generative models for image editing, and leverage their inherent characteristic that the exact intermediate quantized representations of a source image are attainable, enabling more effective supervision for precise image inversion. Specifically, we propose \emph{EditInfinity}, which adapts \emph{Infinity}, a binary-quantized generative model, for image editing. We propose an efficient yet effective image inversion mechanism that integrates text prompting rectification and image style preservation, enabling precise image reconstruction. Furthermore, we devise a holistic smoothing strategy which allows our \emph{EditInfinity} to perform image editing with high fidelity to source images and precise semantic alignment to the text prompts. Extensive experiments on the PIE-Bench benchmark across "add", "change", and "remove" editing operations, demonstrate the superior performance of our model compared to state-of-the-art diffusion-based baselines. Code will be released.



Paperid:1719
Authors:Yiwen Wu, Yuyang Chen, Ye Xia, Yao Zhao, Jingya Wang, Xuming He, Hao GENG, Jingyi Yu
Abstract:
Manhattan representations, defined by axis-aligned, orthogonal structures, are widely used in vision, robotics, and semiconductor design for their geometric regularity and algorithmic simplicity. In integrated circuit (IC) design, Manhattan geometry is key for routing, design rule checking, and lithographic manufacturability. However, as feature sizes shrink, optical system distortions lead to inconsistency between intended layout and printed wafer. Although Inverse Lithography Technology(ILT) is proposed to compensates these effects, learning-based ILT methods, while achieving high simulation fidelity, often generate curvilinear masks on continuous pixel grids, violating Manhattan constraints. Therefore, we propose TokMan, the first framework to formulate mask optimization as a discrete, structure-aware sequence modeling task. Our method leverages a Diffusion Transformer to tokenize layouts into discrete geometric primitives with polygon-wise dependencies and denoise Manhattan-aligned point sequences corrupted by optical proximity effects, while ensuring binary, manufacturable masks. Trained with self-supervised lithographic feedback through differentiable simulation and refined with ILT post-processing, TokMan achieves state-of-the-art fidelity, runtime efficiency, and strict manufacturing compliance on a large-scale dataset of IC layouts.



Paperid:1720
Authors:Sophia Sun, Rose Yu
Abstract:
Conformal prediction has been explored as a general and efficient way to provide uncertainty quantification for time series. However, current methods struggle to handle time series data with change points — sudden shifts in the underlying data-generating process. In this paper, we propose a novel Conformal Prediction for Time-series with Change points (CPTC) algorithm, addressing this gap by integrating a model to predict the underlying state with online conformal prediction to model uncertainties in non-stationary time series. We prove CPTC's validity and improved adaptivity in the time series setting under minimum assumptions, and demonstrate CPTC's practical effectiveness on 6 synthetic and real-world datasets, showing improved validity and adaptivity compared to state-of-the-art baselines.



Paperid:1721
Authors:Chong You, Rajesh Jayaram, Ananda Theertha Suresh, Robin Nittka, Felix Yu, Sanjiv Kumar
Abstract:
Dual encoder (DE) models, where a pair of matching query and document are embedded into similar vector representations, are widely used in information retrieval due to their simplicity and scalability. However, the Euclidean geometry of the embedding space limits the expressive power of DEs, which may compromise their quality. This paper investigate such limitations in the context of hierarchical retrieval (HR), where the document set has a hierarchical structure and the matching documents for a query are all of its ancestors. We first prove that DEs are feasible for HR as long as the embedding dimension is linear in the depth of the hierarchy and logarithmic in the number of documents. Then we study the problem of learning such embeddings in a standard retrieval setup where DEs are trained on samples of matching query and document pairs. Our experiments reveal a lost-in-the-long-distance phenomenon, where retrieval accuracy degrades for documents further away in the hierarchy. To address this, we introduce a pretrain-finetune recipe that significantly improves long-distance retrieval without sacrificing performance on closer documents. We experiment on a realistic hierarchy from WordNet for retrieving documents at various levels of abstraction, and show that pretrain-finetune boosts the recall on long distance pairs from 19\% to 76\%. Finally, we demonstrate that our method improves retrieval of relevant products on a shopping queries dataset.



Paperid:1722
Authors:Shaoshu Yang, Yingya Zhang, Ran He
Abstract:
Video inpainting and editing have long been challenging tasks in the video generation community, requiring extensive computational resources and large datasets to train models with satisfactory performance. Recent breakthroughs in large-scale video foundation models have greatly enhanced text-to-video generation capabilities. This naturally leads to the idea of leveraging the prior knowledge from these powerful generators to facilitate video inpainting and editing. In this work, we investigate the feasibility of employing pre-trained text-to-video foundation models for high-quality video inpainting and editing without additional training. Specifically, we introduce a model-agnostic denoising sampler that optimizes the trajectory by maximizing the log-likelihood expectation conditioned on the known video segments. To enable efficient dynamic object removal and replacement, we propose a latent mask fuser that performs accurate video masking directly in latent space, eliminating the need for explicit VAE decoding and encoding. We implement our approach in widely-used foundation generators such as CogVideoX and HunyuanVideo, demonstrating the model-agnostic nature of our sampler. Comprehensive quantitative and qualitative evaluations confirm that our method achieves outstanding video inpainting and editing performance in a plug-and-play fashion.



Paperid:1723
Authors:Cristiano Migali, Marco Mussi, Gianmarco Genalti, Alberto Maria Metelli
Abstract:
Abstract:*Restless* Multi-Armed Bandits (MABs) are a general framework designed to handle real-world decision-making problems where the expected rewards evolve over time, such as in recommender systems and dynamic pricing. In this work, we investigate from a theoretical standpoint two well-known structured subclasses of restless MABs: the *rising* and the *rising concave* settings, where the expected reward of each arm evolves over time following an unknown *non-decreasing* and a *non-decreasing concave* function, respectively. By providing a novel methodology of independent interest for general restless bandits, we establish new lower bounds on the expected cumulative regret for both settings. In the rising case, we prove a lower bound of order $\Omega(T^{2/3})$, matching known upper bounds for restless bandits; whereas, in the rising concave case, we derive a lower bound of order $\Omega(T^{3/5})$, proving for the first time that this setting is provably more challenging than stationary MABs. Then, we introduce Rising Concave Budgeted Exploration (RC-BE($\alpha$)), a new regret minimization algorithm designed for the rising concave MABs. By devising a novel proof technique, we show that the expected cumulative regret of RC-BE($\alpha$) is in the order of $\widetilde{\mathcal{O}}(T^{7/11})$. These results collectively make a step towards closing the gap in rising concave MABs, positioning them between stationary and general restless bandit settings in terms of statistical complexity.



Paperid:1724
Authors:Elia Torre, Michele Viscione, Lucas Pompe, Benjamin F. Grewe, Valerio Mante
Abstract:
Recurrent neural networks (RNNs) provide a powerful approach in neuroscience to infer latent dynamics in neural populations and to generate hypotheses about the neural computations underlying behavior. However, past work has focused on relatively simple, input-driven, and largely deterministic behaviors - little is known about the mechanisms that would allow RNNs to generate the richer, spontaneous, and potentially stochastic behaviors observed in natural settings. Modeling with Hidden Markov Models (HMMs) has revealed a segmentation of natural behaviors into discrete latent states with stochastic transitions between them, a type of dynamics that may appear at odds with the continuous state spaces implemented by RNNs. Here we first show that RNNs can replicate HMM emission statistics and then reverse-engineer the trained networks to uncover the mechanisms they implement. In the absence of inputs, RNN activity collapses towards a single, stable fixed point. When driven by stochastic input, trajectories instead move along a noise-sustained limit cycle. Rotation along the cycle leads to a change in the emission probabilities, and is governed by transitions between regions of slow, noisy dynamics that are connected by fast, deterministic dynamics. The RNNs learn highly structured connectivity, with only a few 'kick neurons' initiating transitions between slow regions. This mechanism emerges through training, triggered by a Hopf bifurcation that shifts RNNs into a regime of stochastic resonance to implement probabilistic computations. This work demonstrates the potential of RNNs in modeling computations underlying spontaneous behaviors and leads to novel hypotheses about the role of stochastic resonance in biological neural networks.



Paperid:1725
Authors:Jungsoo Lee, Janghoon Cho, Hyojin Park, Munawar Hayat, Kyuwoong Hwang, Fatih Porikli, Sungha Choi
Abstract:
Despite their consistent performance improvements, cross-modal retrieval models (e.g., CLIP) show degraded performances with retrieving keys composed of fused image-text modality (e.g., Wikipedia pages with both images and text). To address this critical challenge, multimodal retrieval has been recently explored to develop a unified single retrieval model capable of retrieving keys across diverse modality combinations. A common approach involves constructing new composed sets of image-text triplets (e.g., retrieving a pair of image and text given a query image). However, such an approach requires careful curation to ensure the dataset quality and fails to generalize to unseen modality combinations. To overcome these limitations, this paper proposes Generalized Contrastive Learning (GCL), a novel loss formulation that improves multimodal retrieval performance without the burdensome need for new dataset curation. Specifically, GCL operates by enforcing contrastive learning across all modalities within a mini-batch, utilizing existing image-caption paired datasets to learn a unified representation space. We demonstrate the effectiveness of GCL by showing consistent performance improvements on off-the-shelf multimodal retrieval models (e.g., VISTA, CLIP, and TinyCLIP) using the M-BEIR, MMEB, and CoVR benchmarks.



Paperid:1726
Authors:Xuannan Liu, Zekun Li, Zheqi He, Peipei Li, shuhan xia, Xing Cui, Huaibo Huang, Xi Yang, Ran He
Abstract:
The increasing deployment of Large Vision-Language Models (LVLMs) raises safety concerns under potential malicious inputs. However, existing multimodal safety evaluations primarily focus on model vulnerabilities exposed by static image inputs, ignoring the temporal dynamics of video that may induce distinct safety risks. To bridge this gap, we introduce Video-SafetyBench, the first comprehensive benchmark designed to evaluate the safety of LVLMs under video-text attacks. It comprises 2,264 video-text pairs spanning 48 fine-grained unsafe categories, each pairing a synthesized video with either a harmful query, which contains explicit malice, or a benign query, which appears harmless but triggers harmful behavior when interpreted alongside the video. To generate semantically accurate videos for safety evaluation, we design a controllable pipeline that decomposes video semantics into subject images (what is shown) and motion text (how it moves), which jointly guide the synthesis of query-relevant videos. To effectively evaluate uncertain or borderline harmful outputs, we propose RJScore, a novel LLM-based metric that incorporates the confidence of judge models and human-aligned decision threshold calibration. Extensive experiments show that benign-query video composition achieves average attack success rates of 67.2%, revealing consistent vulnerabilities to video-induced attacks. We believe Video-SafetyBench will catalyze future research into video-based safety evaluation and defense strategies.



Paperid:1727
Authors:whie jung, Dong Hoon Lee, Seunghoon Hong
Abstract:
Recent disentangled representation learning (DRL) methods heavily rely on factor-specific strategies—either learning objectives for attributes or model architectures for objects—to embed inductive biases.Such divergent approaches lead to significant overhead when novel factors of variation do not align with prior assumptions,e.g., statistical independence or spatial exclusivity, or when multiple factors coexist, as practitioners must redesign architectures or objectives.To address this, we propose acompositional bias, a modular inductive bias decoupled from both objectives and architectures.Specifically, we implement compositional bias by formulating DRL as maximizing compositionality. Given two sets of latent representations from different images, we create a composite representation by exchanging random subsets of latents, then maximize the likelihood and compositional consistency of the resulting composite image. We derivemixing strategiesbased on the distinct compositional structures of attribute and object factors, to embed factor-specific biases in a modular way.Under this general framework, simply adjusting the mixing strategy enables disentanglement of attributes, objects and even both without modifying the objectives or architectures.Extensive experiments demonstrate that our method competitive performance in both attribute and object disentanglement, and uniquely achieves joint disentanglement of global style and objects.



Paperid:1728
Authors:Ping Zhang, Zheda Mai, Quang-Huy Nguyen, Wei-Lun (Harry) Chao
Abstract:
Semi-supervised learning (SSL) enhances model performance by leveraging abundant unlabeled data alongside limited labeled data. As vision foundation models (VFMs) become central to modern vision applications, this paper revisits SSL in the context of these powerful pre-trained models. We conduct a systematic study on tasks where frozen VFMs underperform and reveal several key insights when fine-tuning them. First, parameter-efficient fine-tuning (PEFT) using only labeled data often surpasses traditional SSL methods---even without access to unlabeled data. Second, pseudo-labels generated by PEFT models offer valuable supervisory signals for unlabeled data, and different PEFT techniques yield complementary pseudo-labels. These findings motivate a simple yet effective SSL baseline for the VFM era: \emph{ensemble pseudo-labeling across diverse PEFT methods and VFM backbones}. Extensive experiments validate the effectiveness of this approach, offering actionable insights into SSL with VFMs and paving the way for more scalable and robust semi-supervised learning in the foundation model era.



Paperid:1729
Authors:Hao Shao, Shulun Wang, Yang Zhou, Guanglu Song, Dailan He, ZHUOFAN ZONG, Shuo Qin, Yu Liu, Hongsheng Li
Abstract:
Video face swapping has seen increasing adoption in diverse applications, yet existing methods primarily trained on static images struggle to address temporal consistency and complex real-world scenarios. To overcome these limitations, we propose the first video face swapping framework, VividFace, a robust and high-fidelity diffusion-based framework. VividFace employs a novel hybrid training strategy that leverages abundant static image data alongside temporal video sequences, enabling it to effectively model temporal coherence and identity consistency in videos.Central to our approach is a carefully designed diffusion model integrated with a specialized VAE, capable of processing image-video hybrid data efficiently. To further enhance identity and pose disentanglement, we introduce and release the Attribute-Identity Disentanglement Triplet (AIDT) dataset, comprising a large-scale collection of triplets where each set contains three face images—two sharing the same pose and two sharing the same identity. Augmented comprehensively with occlusion scenarios, AIDT significantly boosts the robustness of VividFace against occlusions.Moreover, we incorporate advanced 3D reconstruction techniques as conditioning inputs to address significant pose variations effectively. Extensive experiments demonstrate that VividFace achieves state-of-the-art performance in identity preservation, temporal consistency, and visual realism, surpassing existing methods while requiring fewer inference steps. Our framework notably mitigates common challenges such as temporal flickering, identity loss, and sensitivity to occlusions and pose variations.The AIDT dataset, source code, and pre-trained weights will be released to support future research. The code and weights are available in the supplementary materials.



Paperid:1730
Authors:Artin Tajdini, Jonathan Scarlett, Kevin Jamieson
Abstract:
Abstract:We study stochastic linear bandits with heavy-tailed rewards, where the rewards have a finite $(1+\epsilon)$-absolute central moment bounded by $\upsilon$ for some $\epsilon \in (0,1]$. We improve both upper and lower bounds on the minimax regret compared to prior work. When $\upsilon = \mathcal{O}(1)$, the best prior known regret upper bound is $\tilde{O}(d T^{\frac{1}{1+\epsilon}})$. While a lower with the same scaling has been given, it relies on a construction using $\upsilon = d$, and adapting the construction to the bounded-moment regime with $\upsilon = \mathcal{O}(1)$ yields only a $\Omega(d^{\frac{\epsilon}{1+\epsilon}} T^{\frac{1}{1+\epsilon}})$ lower bound. This matches the known rate for multi-armed bandits and is generally loose for linear bandits, in particular being $\sqrt{d}$ below the optimal rate in the finite-variance case ($\epsilon = 1$).We propose a new elimination-based algorithm guided by experimental design, which achieves regret $\tilde{\mathcal{O}}(d^{\frac{1+3\epsilon}{2(1+\epsilon)}} T^{\frac{1}{1+\epsilon}})$, thus improving the dependence on $d$ for all $\epsilon \in (0,1)$ and recovering a known optimal result for $\epsilon = 1$. We also establish a lower bound of $\Omega(d^{\frac{2\epsilon}{1+\epsilon}} T^{\frac{1}{1+\epsilon}})$, which strictly improves upon the multi-armed bandit rate and highlights the hardness of heavy-tailed linear bandit problems. For finite action sets of size $n$, we derive upper and lower bounds of$\tilde{\mathcal{O}}(\sqrt d (\log n)^{\frac{\epsilon}{1+\epsilon}}T^{\frac{1}{1+\epsilon}})$ and$\tilde\Omega(d^{\frac{\epsilon}{1+\epsilon}}(\log n)^{\frac{\epsilon}{1+\epsilon}} T^{\frac{1}{1+\epsilon}})$, respectively.Finally, we provide action-set-dependent regret upper bounds and show that for some geometries, such as $l_p$-norm balls for $p \le 1 + \epsilon$, we can further reduce the dependence on $d$.



Paperid:1731
Authors:Fan Zhang, Tianyu Liu, Zhihong Zhu, Hao Wu, Haixin Wang, Donghao Zhou, Yefeng Zheng, Kun Wang, Xian Wu, Pheng-Ann Heng
Abstract:
Abstract:Recent studies have demonstrated the feasibility of modeling single-cell data as natural languages and the potential of leveraging powerful large language models (LLMs) for understanding cell biology. However, a comprehensive evaluation of LLMs' performance on language-driven single-cell analysis tasks still remains unexplored. Motivated by this challenge, we introduce CellVerse, a unified language-centric question-answering benchmark that integrates four types of single-cell multi-omics data and encompasses three hierarchical levels of single-cell analysis tasks: cell type annotation (cell-level), drug response prediction (drug-level), and perturbation analysis (gene-level). Going beyond this, we systematically evaluate the performance across 14 open-source and closed-source LLMs ranging 160M $\rightarrow$ 671B on CellVerse. Remarkably, the experimental results reveal: (1) Existing specialist models (C2S-Pythia) fail to make reasonable decisions across all sub-tasks within CellVerse, while generalist models such as Qwen, Llama, GPT, and DeepSeek family models exhibit preliminary understanding capabilities within the realm of cell biology. (2) The performance of current LLMs falls short of expectations and has substantial room for improvement. Notably, in the widely studied drug response prediction task, none of the evaluated LLMs demonstrate significant performance improvement over random guessing. CellVerse offers the first large-scale empirical demonstration that significant challenges still remain in applying LLMs to cell biology. By introducing CellVerse, we lay the foundation for advancing cell biology through natural languages and hope this paradigm could facilitate next-generation single-cell analysis. Project Page: https://cellverse-cuhk.github.io



Paperid:1732
Authors:Matthew Dutson, Nathan Labiosa, Yin Li, Mohit Gupta
Abstract:
Frame-based neural networks often exhibit temporal inconsistency, i.e., flickering or incoherent predictions, when applied to videos. This problem is amplified when the input frames contain time-varying corruptions such as noise. Many prior works propose stable video-optimized networks for individual tasks or groups of related tasks. In this work, we introduce a general approach for adapting frame-wise networks for stable inference on video data. We describe a class of stabilization adapters that can be inserted into virtually any architecture and a resource-efficient training process that can be performed with a frozen base network. We introduce a unified conceptual framework for describing temporal stability and corruption robustness, centered on a proposed accuracy-stability-robustness loss. We analyze the theoretical properties of this loss, identifying the conditions where it produces well-behaved training. Our experiments validate our approach on several vision tasks including denoising (NAFNet), image enhancement (HDRNet), and monocular depth (Depth Anything v2). Our method improves temporal stability and robustness against a range of image corruptions, while generally preserving or improving task accuracy.



Paperid:1733
Authors:Sujun Tang, Christopher Priebe, Rohan Mahapatra, Lianhui Qin, Hadi Esmaeilzadeh
Abstract:
While model serving has unlocked unprecedented capabilities, the high cost of serving large-scale models continues to be a significant barrier to widespread accessibility and rapid innovation. Compiler optimizations have long driven substantial performance improvements, but existing compilers struggle with neural workloads due to the exponentially large and highly interdependent space of possible transformations. Although existing stochastic search techniques can be effective, they are often sample-inefficient and fail to leverage the structural context underlying compilation decisions. We set out to investigate the research question of whether reasoning with Large Language Models (LLMs)—without any retraining —can leverage the context-aware decision space of compiler optimization to significantly improve sample efficiency. To that end, we introduce a novel compilation framework (dubbed REASONING COMPILER) that formulates optimization as a sequential, context-aware decision process, guided by Large Language Models and structured Monte Carlo tree search (MCTS). The LLM acts as a proposal mechanism, suggesting hardware-aware transformations that reflect the current program state and accumulated performance feedback. Monte Carlo tree search (MCTS) incorporates the LLM-generated proposals to balance exploration and exploitation, facilitating structured, context-sensitive traversal of the expansive compiler optimization space. By achieving substantial speedups with markedly fewer samples than leading neural compilers, our approach demonstrates the potential of LLM-guided reasoning to transform the landscape of compiler optimization.



Paperid:1734
Authors:Ryan Thompson, Edwin Bonilla, Robert Kohn
Abstract:
Directed acyclic graph (DAG) learning is a central task in structure discovery and causal inference. Although the field has witnessed remarkable advances over the past few years, it remains statistically and computationally challenging to learn a single (point estimate) DAG from data, let alone provide uncertainty quantification. We address the difficult task of quantifying graph uncertainty by developing a Bayesian variational inference framework based on novel, provably valid distributions that have support directly on the space of sparse DAGs. These distributions, which we use to define our prior and variational posterior, are induced by a projection operation that maps an arbitrary continuous distribution onto the space of sparse weighted acyclic adjacency matrices. While this projection is combinatorial, it can be solved efficiently using recent continuous reformulations of acyclicity constraints. We empirically demonstrate that our method, \texttt{ProDAG}, can outperform state-of-the-art alternatives in both accuracy and uncertainty quantification.



Paperid:1735
Authors:Hany Abdulsamad, Sahel Mohammad Iqbal, Simo Sarkka
Abstract:
Optimal decision-making under partial observability requires agents to balance reducing uncertainty (exploration) against pursuing immediate objectives (exploitation). In this paper, we introduce a novel policy optimization framework for continuous partially observable Markov decision processes (POMDPs) that explicitly addresses this challenge. Our method casts policy learning as probabilistic inference in a non-Markovian Feynman--Kac model that inherently captures the value of information gathering by anticipating future observations, without requiring extrinsic exploration bonuses or handcrafted heuristics. To optimize policies under this model, we develop a nested sequential Monte Carlo~(SMC) algorithm that efficiently estimates a history-dependent policy gradient under samples from the optimal trajectory distribution induced by the POMDP. We demonstrate the effectiveness of our algorithm across standard continuous POMDP benchmarks, where existing methods struggle to act under uncertainty.



Paperid:1736
Authors:Baoqi Pei, Yifei Huang, Jilan Xu, Yuping He, Guo Chen, Fei Wu, Jiangmiao Pang, Yu Qiao
Abstract:
Egocentric video reasoning centers on an unobservable agent behind the camera who dynamically shapes the environment, requiring inference of hidden intentions and recognition of fine-grained interactions. This core challenge limits current multimodal large language models (MLLMs), which excel at visible event reasoning but lack embodied, first-person understanding. To bridge this gap, we introduce EgoThinker, a novel framework that endows MLLMs with robust egocentric reasoning capabilities through spatio-temporal chain-of-thought supervision and a two-stage learning curriculum. First, we introduce EgoRe-5M, a large-scale egocentric QA dataset constructed from 13M diverse egocentric video clips. This dataset features multi-minute segments annotated with detailed CoT rationales and dense hand–object grounding. Second, we employ SFT on EgoRe-5M to instill reasoning skills, followed by reinforcement fine-tuning (RFT) to further enhance spatio-temporal localization. Experimental results show that EgoThinker outperforms existing methods across multiple egocentric benchmarks, while achieving substantial improvements in fine-grained spatio-temporal localization tasks.



Paperid:1737
Authors:Jaa-Yeon Lee, ByungHee Cha, Jeongsol Kim, Jong Chul Ye
Abstract:
While recent advancements in generative modeling have significantly improved text-image alignment, some residual misalignment between text and image representations still remains. Some approaches address this issue by fine-tuning models in terms of preference optimization, etc., which require tailored datasets. Orthogonal to these methods, we revisit the challenge from the perspective of representation alignment—an approach that has gained popularity with the success of REPresentation Alignment (REPA). We first argue that conventional text-to-image (T2I) diffusion models, typically trained on paired image and text data (i.e., positive pairs) by minimizing score matching or flow matching losses, is suboptimal from the standpoint of representation alignment. Instead, a better alignment can be achieved through contrastive learning that leverages existing dataset as both positive and negative pairs. To enable efficient alignment with pretrained models, we propose SoftREPA—a lightweight contrastive fine-tuning strategy that leverages soft text tokens for representation alignment. This approach improves alignment with minimal computational overhead by adding fewer than 1M trainable parameters to the pretrained model. Our theoretical analysis demonstrates that our method explicitly increases the mutual information between text and image representations, leading to enhanced semantic consistency. Experimental results across text-to-image generation and text-guided image editing tasks validate the effectiveness of our approach in improving the semantic consistency of T2I generative models.



Paperid:1738
Authors:Tianyu Hu, Zhen Tan, Song Wang, Huaizhi Qu, Tianlong Chen
Abstract:
With advancements in reasoning capabilities, Large Language Models (LLMs) are increasingly employed for automated judgment tasks.While LLMs-as-Judges offer promise in automating evaluations, current approaches often rely on simplistic aggregation methods (e.g., majority voting), which can fail even when individual agents provide correct answers.To address this, we propose a multi-agent debate judge framework where agents collaboratively reason and iteratively refine their responses. We formalize the debate process mathematically, analyzing agent interactions and proving that debate amplifies correctness compared to static ensembles.To enhance efficiency, we introduce a stability detection mechanism that models judge consensus dynamics via a time-varying Beta-Binomial mixture, with adaptive stopping based on distributional similarity (Kolmogorov-Smirnov test).This mechanism models the judges' collective correct rate dynamics using a time-varying mixture of Beta-Binomial distributions and employs an adaptive stopping criterion based on distributional similarity (Kolmogorov-Smirnov statistic). Experiments across multiple benchmarks and models demonstrate that our framework improves judgment accuracy over majority voting while maintaining computational efficiency.



Paperid:1739
Authors:Qiao Gu, Yuanliang Ju, Shengxiang Sun, Igor Gilitschenski, Haruki Nishimura, Masha Itkina, Florian Shkurti
Abstract:
Abstract:While vision-language-action models (VLAs) have shown promising robotic behaviors across a diverse set of manipulation tasks, they achieve limited success rates when deployed on novel tasks out-of-the-box. To allow these policies to safely interact with their environments, we need a failure detector that gives a timely alert such that the robot can stop, backtrack, or ask for help. However, existing failure detectors are trained and tested only on one or a few specific tasks, while VLAs require the detector to generalize and detect failures also in unseen tasks and novel environments. In this paper, we introduce the multitask failure detection problem and propose SAFE, a failure detector for generalist robot policies such as VLAs. We analyze the VLA feature space and find that VLAs have sufficient high-level knowledge about task success and failure, which is generic across different tasks. Based on this insight, we design SAFE to learn from VLA internal features and predict a single scalar indicating the likelihood of task failure. SAFE is trained on both successful and failed rollouts, and is evaluated on unseen tasks. SAFE is compatible with different policy architectures. We test it on OpenVLA, $\pi_0$, and $\pi_0$-FAST in both simulated and real-world environments extensively. We compare SAFE with diverse baselines and show that SAFE achieves state-of-the-art failure detection performance and the best trade-off between accuracy and detection time using conformal prediction.



Paperid:1740
Authors:Reza Ramezanpour, Victor Manuel Tenorio Gomez, Antonio G. Marques, Ashutosh Sabharwal, Santiago Segarra
Abstract:
Graphons, as limit objects of dense graph sequences, play a central role in the statistical analysis of network data. However, existing graphon estimation methods often struggle with scalability to large networks and resolution-independent approximation, due to their reliance on estimating latent variables or costly metrics such as the Gromov-Wasserstein distance. In this work, we propose a novel, scalable graphon estimator that directly recovers the graphon via moment matching, leveraging implicit neural representations (INRs). Our approach avoids latent variable modeling by training an INR--mapping coordinates to graphon values--to match empirical subgraph counts (i.e., moments) from observed graphs.This direct estimation mechanism yields a polynomial-time solution and crucially sidesteps the combinatorial complexity of Gromov-Wasserstein optimization. Building on foundational results, we establish a theoretical guarantee: when the observed subgraph motifs sufficiently represent those of the true graphon (a condition met with sufficiently large or numerous graph samples), the estimated graphon achieves a provable upper bound in cut distance from the ground truth. Additionally, we introduce MomentMixup, a data augmentation technique that performs mixup in the moment space to enhance graphon-based learning. Our graphon estimation method achieves strong empirical performance--demonstrating high accuracy on small graphs and superior computational efficiency on large graphs--outperforming state-of-the-art scalable estimators in 75\% of benchmark settings and matching them in the remaining cases. Furthermore, MomentMixup demonstrated improved graph classification accuracy on the majority of our benchmarks.



Paperid:1741
Authors:Yichuan Cao, Yibo Miao, Yinpeng Dong, Xiao-Shan Gao
Abstract:
Text-to-image (T2I) models raise ethical and safety concerns due to their potential to generate inappropriate or harmful images. Evaluating these models’ security through red-teaming is vital, yet white-box approaches are limited by their need for internal access, complicating their use with closed-source models. Moreover, existing black-box methods often assume knowledge about the model’s specific defense mechanisms, limiting their utility in real-world commercial API scenarios. A significant challenge is how to evade unknown and diverse defense mechanisms. To overcome this difficulty, we propose a novel Rule-based Preference modeling Guided Red-Teaming (RPG-RT), which iteratively employs LLM to modify prompts to query and leverages feedback from T2I systems for fine-tuning the LLM. RPG-RT treats the feedback from each iteration as a prior, enabling the LLM to dynamically adapt to unknown defense mechanisms. Given that the feedback is often labeled and coarse-grained, making it difficult to utilize directly, we further propose rule-based preference modeling, which employs a set of rules to evaluate desired or undesired feedback, facilitating finer-grained control over the LLM’s dynamic adaptation process. Extensive experiments on nineteen T2I systems with varied safety mechanisms, three online commercial API services, and T2V models verify the superiority and practicality of our approach.



Paperid:1742
Authors:Hieu Vu, Tan Nguyen
Abstract:
Controlling specific behaviors in large language models while preserving their general capabilities is a central challenge for safe and reliable artificial intelligence (AI) deployment. Current steering methods, such as vector addition and directional ablation, are constrained within a two-dimensional subspace defined by the activation and feature direction, making them sensitive to chosen parameters and potentially affecting unrelated features due to unintended interactions in activation space. We introduce Angular Steering, a novel and flexible method for behavior modulation that operates by rotating activations within a fixed two-dimensional subspace. By formulating steering as a geometric rotation toward or away from a target behavior direction, Angular Steering provides continuous, fine-grained control over behaviors such as refusal and compliance. We demonstrate this method using refusal steering as a use case. Additionally, we propose Adaptive Angular Steering, a selective variant that rotates only activations aligned with the target feature, further enhancing stability and coherence. Angular Steering generalizes existing addition and orthogonalization techniques under a unified geometric rotation framework, simplifying parameter selection and maintaining model stability across a broader range of adjustments. Experiments across multiple model families and sizes show that Angular Steering achieves robust behavioral control without degrading general language modeling performance, underscoring its flexibility, generalization, and robustness compared to prior approaches.



Paperid:1743
Authors:Qinfan Xiao, Ziyun Cui, Chi Zhang, SiQi Chen, Wen Wu, Andrew Thwaites, Alexandra Woolgar, Bowen Zhou, Chao Zhang
Abstract:
Electroencephalography (EEG) and magnetoencephalography (MEG) measure neural activity non-invasively by capturing electromagnetic fields generated by dendritic currents. Although rooted in the same biophysics, EEG and MEG exhibit distinct signal patterns, further complicated by variations in sensor configurations across modalities and recording devices.Existing approaches typically rely on separate, modality- and dataset-specific models, which limits the performance and cross-domain scalability.This paper proposes BrainOmni, the first brain foundation model that generalises across heterogeneous EEG and MEG recordings. To unify diverse data sources, we introduce BrainTokenizer, the first tokeniser that quantises spatiotemporal brain activity into discrete representations. Central to BrainTokenizer is a novel Sensor Encoder that encodes sensor properties such as spatial layout, orientation, and type, enabling compatibility across devices and modalities.Building upon the discrete representations, BrainOmni learns unified semantic embeddings of brain signals by self-supervised pretraining. To the best of our knowledge, it is the first foundation model to support both EEG and MEG signals, as well as the first to incorporate large-scale MEG pretraining.A total of 1,997 hours of EEG and 656 hours of MEG data are curated and standardised from publicly available sources for pretraining.Experiments show that BrainOmni outperforms both existing foundation models and state-of-the-art task-specific models on a range of downstream tasks. It also demonstrates strong generalisation to unseen EEG and MEG devices. Further analysis reveals that joint EEG-MEG (EMEG) training yields consistent improvements across both modalities. Code and model checkpoints will be released upon acceptance.



Paperid:1744
Authors:Zhening Huang, Xiaoyang Wu, Fangcheng Zhong, Hengshuang Zhao, Matthias Niessner, Joan Lasenby
Abstract:
We propose LiteReality, a novel pipeline that converts RGB-D scans of indoor environments into compact, realistic, and interactive 3D virtual replicas. LiteReality not only reconstructs scenes that visually resemble reality but also supports key features essential for graphics pipelines—such as object individuality, articulation, high-quality physically based rendering materials, and physically based interaction. At its core, LiteReality first performs scene understanding and parses the results into a coherent 3D layout and objects, with the help of a structured scene graph. It then reconstructs the scene by retrieving the most visually similar 3D artist-crafted models from a curated asset database. Later, the Material Painting module enhances the realism of retrieved objects by recovering high-quality, spatially varying materials. Finally, the reconstructed scene is integrated into a simulation engine with basic physical properties applied to enable interactive behavior. The resulting scenes are compact, editable, and fully compatible with standard graphics pipelines, making them suitable for applications in AR/VR, gaming, robotics, and digital twins. In addition, LiteReality introduces a training-free object retrieval module that achieves state-of-the-art similarity performance, as benchmarked on the Scan2CAD dataset, along with a robust Material Painting module capable of transferring appearances from images of any style to 3D assets—even in the presence of severe misalignment, occlusion, and poor lighting. We demonstrate the effectiveness of LiteReality on both real-life scans and public datasets.



Paperid:1745
Authors:Darshan Prabhu, Preethi Jyothi
Abstract:
Abstract:Attention-based encoder decoder models are the backbone of state-of-the-art architectures for automatic speech recognition (ASR). These models combine a powerful encoder that extracts rich acoustic features with a decoder that autoregressively produces the ASR output. The decoder handles two critical tasks: (1) building rich text-only context and (2) merging acoustic information from the encoder to ensure the predictions remain faithful to the audio. We observe a systematic pattern across the attention distributions of decoder layers in prior architectures: the initial layers direct most attention towards building textual context, while the latter layers focus largely on merging acoustic and textual information to make accurate predictions. Leveraging this key insight, we propose **BlockDecoder**, a novel decoder architecture comprising two distinct components: a **TextEncoder** that is purely text-based, and a **Merger** that autoregressively generates tokens in blockwise fashion by combining information from the audio encoder and **TextEncoder**. These two components of **BlockDecoder** collectively result in substantial latency gains. Across diverse datasets, languages and speech tasks, we demonstrate that our proposed **BlockDecoder** achieves a significant speedup ($\sim 2$x) compared to traditional decoders, without any degradation in performance.



Paperid:1746
Authors:Huimin Wu, Xiaojian (Shawn) Ma, Haozhe Zhao, Yanpeng Zhao, Qing Li
Abstract:
Text-guided image editing involves modifying a source image based on a language instruction and, typically, requires changes to only small local regions. However, existing approaches generate the entire target image rather than selectively regenerate only the intended editing areas. This results in (1) unnecessary computational costs and (2) a bias toward reconstructing non-editing regions, which compromises the quality of the intended edits. To resolve these limitations, we propose to formulate image editing as Next Editing-token Prediction (NEP) based on autoregressive image generation, where only regions that need to be edited are regenerated, thus avoiding unintended modification to the non-editing areas. To enable any-region editing, we propose to pre-train an any-order autoregressive text-to-image (T2I) model. Once trained, it is capable of zero-shot image editing and can be easily adapted to NEP for image editing, which achieves a new state-of-the-art on widely used image editing benchmarks. Moreover, our model naturally supports test-time scaling (TTS) through iteratively refining its generation in a zero-shot manner.



Paperid:1747
Authors:Yifan Wang, Wuliang Huang, Yufan Wen, Shunning Liu, Chun Yuan
Abstract:
The interactive task of composed image retrieval aims to retrieve the most relevant images with the bi-modal query, consisting of a reference image and a modification sentence. Despite significant efforts to bridge the heterogeneous gap within the bi-modal query and leverage contrastive learning to reduce the disparity between positive and negative triplets, prior methods often fail to ensure reliable matching due to aleatoric and epistemic uncertainty. Specifically, the aleatoric uncertainty stems from underlying semantic correlations within candidate instances and annotation noise, and the epistemic uncertainty is usually caused by overconfidence in dominant semantic categories. In this paper, we propose Robust UNcertainty Calibration (RUNC) to quantify the uncertainty and calibrate the imbalanced semantic distribution. To mitigate semantic ambiguity in similarity distribution between fusion queries and targets, RUNC maximizes the matching evidence by utilizing a high-order conjugate prior distribution to fit the semantic covariances in candidate samples. With the estimated uncertainty coefficient of each candidate, the target distribution is calibrated to encourage balanced semantic alignment. Additionally, we minimize the ambiguity in the fusion evidence when forming the unified query by incorporating orthogonal constraints on explicit textual embeddings and implicit queries, to reduce the representation redundancy. Extensive experiments and ablation analysis on benchmark datasets FashionIQ and CIRR verify the robustness of RUNC in predicting reliable retrieval results from a large image gallery.



Paperid:1748
Authors:DongWon Kim, Sungdong Lee, Joong-Ho (Johann) Won
Abstract:
Abstract:We develop a scalable second-order algorithm for a recently proposed $\ell_1$-regularized pseudolikelihood-based partial correlation network estimation framework. While the latter method admits statistical guarantees and is inherently scalable compared to likelihood-based methods such as graphical lasso, the currently available implementations rely only on first-order information and require thousands of iterations to obtain reliable estimates even on high-performance supercomputers.In this paper, we further investigate the inherent scalability of the framework and propose locally and globally convergent semismooth Newton methods. Despite the nonsmoothness of the problem, these second-order algorithms converge at a asymptotically quadratic rate, and require only a few tens of iterations in practice. Each iteration reduces to solving linear systems of small dimensions or linear complementary problems of smaller dimensions, making the computation also suitable for less powerful computing environments.Experiments on both simulated and real-world genomic datasets demonstrate the superior convergence behavior and computational efficiency of the proposed algorithm, which position our method as a promising tool for massive-scale network analysis sought for in, e.g., modern multi-omics research.



Paperid:1749
Authors:Yunrui Guan, Krishnakumar Balasubramanian, Shiqian Ma
Abstract:
We introduce the \textit{Riemannian Proximal Sampler}, a method for sampling from densities defined on Riemannian manifolds. The performance of this sampler critically depends on two key oracles: the \textit{Manifold Brownian Increments (MBI)} oracle and the \textit{Riemannian Heat-kernel (RHK)} oracle. We establish high-accuracy sampling guarantees for the Riemannian Proximal Sampler, showing that generating samples with (\varepsilon)-accuracy requires (\mathcal{O}(\log(1/\varepsilon))) iterations in Kullback-Leibler divergence assuming access to exact oracles and (\mathcal{O}(\log^2(1/\varepsilon))) iterations in the total variation metric assuming access to sufficiently accurate inexact oracles. Furthermore, we present practical implementations of these oracles by leveraging heat-kernel truncation and Varadhan’s asymptotics. In the latter case, we interpret the Riemannian Proximal Sampler as a discretization of the entropy-regularized Riemannian Proximal Point Method on the associated Wasserstein space. We provide preliminary numerical results that illustrate the effectiveness of the proposed methodology.



Paperid:1750
Authors:Phuc Tran, Nisheeth K. Vishnoi
Abstract:
Abstract:Low-rank pseudoinverses are widely used to approximate matrix inverses in scalable machine learning, optimization, and scientific computing. However, real-world matrices are often observed with noise, arising from sampling, sketching, and quantization. The spectral-norm robustness of low-rank inverse approximations remains poorly understood. We systematically study the spectral-norm error $\| \tilde{A}_p^{-1} - A_p^{-1} \|$ for an $n\times n$ symmetric matrix $A$, where $A_p^{-1}$ denotes the best rank-\(p\) approximation of $A^{-1}$, and $\tilde{A} = A + E$ is a noisy observation. Under mild assumptions on the noise, we derive sharp non-asymptotic perturbation bounds that reveal how the error scales with the eigengap, spectral decay, and noise alignment with low-curvature directions of $A$. Our analysis introduces a novel application of contour integral techniques to the \emph{non-entire} function $f(z) = 1/z$, yielding bounds that improve over naive adaptations of classical full-inverse bounds by up to a factor of $\sqrt{n}$. Empirically, our bounds closely track the true perturbation error across a variety of real-world and synthetic matrices, while estimates based on classical results tend to significantly overpredict. These findings offer practical, spectrum-aware guarantees for low-rank inverse approximations in noisy computational environments.



Paperid:1751
Authors:Boyan Gao, Xin Wang, Yibo Yang, David Clifton
Abstract:
Large Language Models (LLMs) have demonstrated remarkable performance in real-world applications. However, adapting LLMs to novel tasks via fine-tuning often requires substantial training data and computational resources that are impractical in few-shot scenarios. Existing approaches, such as in-context learning and Parameter-Efficient Fine-Tuning (PEFT), face key limitations: in-context learning introduces additional inference computational overhead with limited performance gains, while PEFT models are prone to overfitting on the few demonstration examples. In this work, we reinterpret the forward pass of LLMs as an optimization process, a sequence of preconditioned gradient descent steps refining internal representations. Based on this connection, we propose Optimization-Inspired Few-Shot Adaptation (OFA), integrating a parameterization that learns preconditioners without introducing additional trainable parameters, and an objective that improves optimization efficiency by learning preconditioners based on a convergence bound, while simultaneously steering the optimization path toward the flat local minimum. Our method overcomes both issues of ICL-based and PEFT-based methods, and demonstrates superior performance over the existing methods on a variety of few-shot adaptation tasks in experiments.



Paperid:1752
Authors:Sihan Zeng, Benjamin Patrick Evans, Sujay Bhatt, Leo Ardon, Sumitra Ganesh, Alec Koppel
Abstract:
We study policy optimization in Stackelberg mean field games (MFGs), a hierarchical framework for modeling the strategic interaction between a single leader and an infinitely large population of homogeneous followers. The objective can be formulated as a structured bi-level optimization problem, in which the leader needs to learn a policy maximizing its reward, anticipating the response of the followers. Existing methods for solving these (and related) problems often rely on restrictive independence assumptions between the leader’s and followers’ objectives, use samples inefficiently due to nested-loop algorithm structure, and lack finite-time convergence guarantees. To address these limitations, we propose AC-SMFG, a single-loop actor-critic algorithm that operates on continuously generated Markovian samples. The algorithm alternates between (semi-)gradient updates for the leader, a representative follower, and the mean field, and is simple to implement in practice. We establish the finite-time and finite-sample convergence of the algorithm to a stationary point of the Stackelberg objective. To our knowledge, this is the first Stackelberg MFG algorithm with non-asymptotic convergence guarantees. Our key assumption is a "gradient alignment" condition, which requires that the full policy gradient of the leader can be approximated by a partial component of it, relaxing the existing leader-follower independence assumption. Simulation results in a range of well-established economics environments demonstrate that AC-SMFG outperforms existing multi-agent and MFG learning baselines in policy quality and convergence speed.



Paperid:1753
Authors:Shyam Venkatasubramanian, Vahid Tarokh
Abstract:
Accelerating model convergence in resource-constrained environments is essential for fast and efficient neural network training. This work presents learn2mix, a new training strategy that adaptively adjusts class proportions within batches, focusing on classes with higher error rates. Unlike classical training methods that use static class proportions, learn2mix continually adapts class proportions during training, leading to faster convergence. Empirical evaluations on benchmark datasets show that neural networks trained with learn2mix converge faster than those trained with existing approaches, achieving improved results for classification, regression, and reconstruction tasks under limited training resources and with imbalanced classes. Our empirical findings are supported by theoretical analysis.



Paperid:1754
Authors:Konstantinos Kontras, Thomas Strypsteen, Christos Chatzichristos, Paul Pu Liang, Matthew Blaschko, Maarten De Vos
Abstract:
Multimodal learning holds the promise for richer information extraction by capturing dependencies across data sources. Yet, current training methods often underperform due to modality competition, a phenomenon where modalities contend for training resources, leaving some underoptimized. This raises a pivotal question: how can we address training imbalances, ensure adequate optimization across all modalities, and achieve consistent performance improvements as we transition from unimodal to multimodal data? This paper proposes the Multimodal Competition Regularizer (MCR), inspired by a mutual information (MI) decomposition designed to prevent the adverse effects of competition in multimodal training. Our key contributions are: 1) A game-theoretic framework that adaptively balances modality contributions by encouraging each to maximize its informative role in the final prediction. 2) Refining lower and upper bounds for each MI term to enhance the extraction of both task-relevant unique and shared information across modalities. 3) Proposing latent space permutations for conditional MI estimation, significantly improving computational efficiency. MCR outperforms all previously suggested training strategies and simple baselines, demonstrating that training modalities jointly lead to important performance gains on synthetic and large real-world datasets.



Paperid:1755
Authors:Junhyung Lyle Kim, Brandon Augustino, Dylan Herman, Enrico Fontana, Jacob Watkins, Shouvanik Chakrabarti, Marco Pistoia
Abstract:
We study quantum algorithms based on quantum (sub)gradient estimation using noisy function evaluation oracles, and demonstrate the first dimension independent query complexities (up to poly-logarithmic factors) for zeroth-order convex optimization in both smooth and nonsmooth settings. Interestingly, only using noisy function evaluation oracles, we match the first-order query complexities of classical gradient descent, thereby exhibiting exponential separation between quantum and classical zeroth-order optimization. We then generalize these algorithms to work in non-Euclidean settings by using quantum (sub)gradient estimation to instantiate mirror descent and its variants, including dual averaging and mirror prox. By leveraging a connection between semidefinite programming and eigenvalue optimization, we use our quantum mirror descent method to give a new quantum algorithm for solving semidefinite programs, linear programs and zero-sum games. For various parameter regimes, our algorithm for solving zero-sum games is faster than existing state of the art classical and quantum approaches.



Paperid:1756
Authors:Haonan Dong, Wenhao Zhu, Guojie Song, Liang Wang
Abstract:
Abstract:Low-Rank Adaptation (LoRA) is a widely adopted parameter-efficient fine-tuning (PEFT) method validated across NLP and CV domains. However, LoRA faces an inherent low-rank bottleneck: narrowing its performance gap with full fine-tuning requires increasing the rank of its parameter matrix, resulting in significant parameter overhead. Recent linear LoRA variants have attempted to enhance expressiveness by introducing additional linear mappings; however, their composition remains inherently linear and fails to fundamentally improve LoRA’s representational capacity. To address this limitation, we propose \ourmethod, which incorporates an Adaptive Nonlinear Layer (ANL) between two linear projectors to capture \emph{fixed} and \emph{learnable} nonlinearities. This combination forms an {\fontfamily{lmtt}\selectfont \textbf{MLP-like structure}} with a compressed rank, enabling flexible and precise approximation of diverse target functions while theoretically guaranteeing lower approximation errors and bounded gradients. Extensive experiments on 22 datasets and 6 pretrained models demonstrate that \ourmethod: (\textbf{I}) not only matches or surpasses full fine-tuning performance with only $6.18\%\sim25\%$ of LoRA’s parameters but also (\textbf{II}) outperforms state-of-the-art PEFT methods by up to $10.88\%$ in both NLP and CV tasks, and \textbf{(III)} exhibits robust performance across various rank configurations.



Paperid:1757
Authors:Charlie Tan, Majdi Hassan, Leon Klein, Saifuddin Syed, Dominique Beaini, Michael Bronstein, Alexander Tong, Kirill Neklyudov
Abstract:
Efficient equilibrium sampling of molecular conformations remains a core challenge in computational chemistry and statistical inference. Classical approaches such as molecular dynamics or Markov chain Monte Carlo inherently lack amortization; the computational cost of sampling must be paid in-full for each system of interest. The widespread success of generative models has inspired interest into overcoming this limitation through learning sampling algorithms. Despite performing on par with conventional methods when trained on a single system, learned samplers have so far demonstrated limited ability to transfer across systems. We prove that deep learning enables the design of scalable and transferable samplers by introducing Ensemble, a 200 million parameter all-atom transferable normalizing flow trained on a corpus of peptide molecular dynamics trajectories up to 8 residues in length. Ensemble draws zero-shot uncorrelated proposal samples for arbitrary peptide systems, achieving the previously intractable transferability across sequence length, whilst retaining the efficient likelihood evaluation of normalizing flows. Through extensive empirical evaluation we demonstrate the efficacy of Ensemble as a proposal for a variety of sampling algorithms, finding a simple importance sampling-based finetuning procedure to achieve superior performance to established methods such as sequential Monte Carlo. We open-source the Ensemble codebase, model weights, and training dataset, to further stimulate research into amortized sampling methods and finetuning objectives.



Paperid:1758
Authors:Runheng Liu, Heyan Huang, Xingchen Xiao, Zhijing Wu
Abstract:
Large language models (LLMs) have demonstrated remarkable capabilities across various tasks. However, their ability to generate human-like text has raised concerns about potential misuse. This underscores the need for reliable and effective methods to detect LLM-generated text. In this paper, we propose IRM, a novel zero-shot approach that leverages Implicit Reward Models for LLM-generated text detection. Such implicit reward models can be derived from publicly available instruction-tuned and base models. Previous reward-based method relies on preference construction and task-specific fine-tuning. In comparison, IRM requires neither preference collection nor additional training.We evaluate IRM on the DetectRL benchmark and demonstrate that IRM can achieve superior detection performance, outperforms existing zero-shot and supervised methods in LLM-generated text detection.



Paperid:1759
Authors:Pingbang Hu, Joseph Melkonian, Weijing Tang, Han Zhao, Jiaqi Ma
Abstract:
Gradient-based data attribution methods, such as influence functions, are critical for understanding the impact of individual training samples without requiring repeated model retraining. However, their scalability is often limited by the high computational and memory costs associated with per-sample gradient computation. In this work, we proposeGraSS, a novel gradient compression algorithm and its variantsFactGraSSfor linear layers specifically, that explicitly leverage the inherent sparsity of per-sample gradients to achieve sub-linear space and time complexity. Extensive experiments demonstrate the effectiveness of our approach, achieving substantial speedups while preserving data influence fidelity. In particular,FactGraSSachieves up to 165% faster throughput on billion-scale models compared to the previous state-of-the-art baselines.



Paperid:1760
Authors:Masih Aminbeidokhti, Subhankar Roy, Eric Granger, Elisa Ricci, Marco Pedersoli
Abstract:
Abstract:Real-world datasets typically exhibit long-tailed (LT) distributions, where a few head classes dominate and many tail classes are severely underrepresented. While recent work shows that parameter-efficient fine-tuning (PEFT) methods like LoRA and AdaptFormer preserve tail-class performance on foundation models such as CLIP, we find that they do so at the cost of head-class accuracy. We identify the head-tail ratio, the proportion of head to tail classes, as a crucial but overlooked factor influencing this trade-off. Through controlled experiments on CIFAR100 with varying imbalance ratio ($\rho$) and head-tail ratio ($\eta$), we show that PEFT excels in tail-heavy scenarios but degrades in more balanced and head-heavy distributions. To overcome these limitations, we propose LT-Soups, a two-stage model soups framework designed to generalize across diverse LT regimes. In the first stage, LT-Soups averages models fine-tuned on balanced subsets to reduce head-class bias; in the second, it fine-tunes only the classifier on the full dataset to restore head-class accuracy. Experiments across six benchmark datasets show that LT-Soups achieves superior trade-offs compared to both PEFT and traditional model soups across a wide range of imbalance regimes.



Paperid:1761
Authors:Jin Zhou, Kaiwen Wang, Jonathan Chang, Zhaolin Gao, Nathan Kallus, Kilian Weinberger, Kianté Brantley, Wen Sun
Abstract:
Abstract:Reinforcement learning (RL) post-training is crucial for LLM alignment and reasoning, but existing policy-based methods, such as PPO and DPO, can fall short of fixing shortcuts inherited from pre-training. In this work, we introduce $Q\sharp$, a value-based algorithm for KL-regularized RL that guides the reference policy using the optimal regularized $Q$ function. We propose to learn the optimal $Q$ function using distributional RL on an aggregated online dataset. Unlike prior value-based baselines that guide the model using unregularized $Q$-values, our method is theoretically principled and provably learns the optimal policy for the KL-regularized RL problem. Empirically, $Q\sharp$ outperforms prior baselines in math reasoning benchmarks while maintaining a smaller KL divergence to the reference policy. Theoretically, we establish a reduction from KL-regularized RL to no-regret online learning, providing the first bounds for deterministic MDPs under only realizability. Thanks to distributional RL, our bounds are also variance-dependent and converge faster when the reference policy has small variance. In sum, our results highlight $Q\sharp$ as an effective approach for post-training LLMs, offering both improved performance and theoretical guarantees.



Paperid:1762
Authors:seil kang, Woojung Han, Dayun Ju, Seong Jae Hwang
Abstract:
Starting from flow- and diffusion-based transformers, Multi-modal Diffusion Transformers (MM-DiTs) have reshaped text-to-vision generation, gaining acclaim for exceptional visual fidelity. As these models advance, users continually push the boundary with imaginative or rare prompts, which advanced models still falter in generating, since their concepts are often too scarce to leave a strong imprint during pre-training. In this paper, we propose a simple yet effective intervention that surfaces rare semantics inside MM-DiTs without additional training steps, data, denoising-time optimization, or reliance on external modules (e.g., LLMs). In particular, the joint-attention mechanism intrinsic to MM-DiT sequentially updates text embeddings alongside image embeddings throughout transformer blocks. We find that by mathematically expanding representational basins around text token embeddings viavariance scale-upbefore the joint-attention blocks, rare semantics clearly emerge in MM-DiT’s outputs. Furthermore, our results generalize effectively across text-to-vision tasks, including text-to-image, text-to-video, and text-driven image editing. All the source codes will be made available to the public.



Paperid:1763
Authors:Stephen Zhang, Suryanarayana Maddu, Xiaojie Qiu, Victor Chardès
Abstract:
Time-resolved single-cell omics data offers high-throughput, genome-wide measurements of cellular states, which are instrumental to reverse-engineer the processes underpinning cell fate. Such technologies are inherently destructive, allowing only cross-sectional measurements of the underlying stochastic dynamical system. Furthermore, cells may divide or die in addition to changing their molecular state. Collectively these present a major challenge to inferring realistic biophysical models. We present a novel approach, unbalanced probability flow inference, that addresses this challenge for biological processes modelled as stochastic dynamics with growth. By leveraging a Lagrangian formulation of the Fokker-Planck equation, our method accurately disentangles drift from intrinsic noise and growth. We showcase the applicability of our approach through evaluation on a range of simulated and real single-cell RNA-seq datasets. Comparing to several existing methods, we find our method achieves higher accuracy while enjoying a simple two-step training scheme.



Paperid:1764
Authors:Anson Lei, Bernhard Schölkopf, Ingmar Posner
Abstract:
Capturing the interactions between entities in a structured way plays a central role in world models that flexibly adapt to changes in the environment. Recent works motivate the benefits of models that explicitly represent the structure of interactions and formulate the problem as discovering local causal structures. In this work, we demonstrate that reliably capturing these relationships in complex settings remains challenging. To remedy this shortcoming, we postulate that sparsity is a critical ingredient for the discovery of such local structures. To this end we present the SPARse TrANsformer World model (SPARTAN), a Transformer-based world model that learns context-dependent interaction structures between entities in a scene. By applying sparsity regularisation on the attention patterns between object-factored tokens, SPARTAN learns sparse, context-dependent interaction graphs that accurately predict future object states. We further extend our model to adapt to sparse interventions with unknown targets on the dynamics of the environment. This results in a highly interpretable world model that can efficiently adapt to changes. Empirically, we evaluate SPARTAN against the current state-of-the-art in object-centric world models on observation-based environments and demonstrate that our model can learn local causal graphs that accurately reflects the underlying interactions between objects and achieve significantly improved few-shot adaptation to dynamics changes as well as robustness against distractors.



Paperid:1765
Authors:Hongyu Qu, Xiangbo Shu, Jianan Wei, Yazhou Yao, Wenguan Wang, Jinhui Tang
Abstract:
Abstract:Current 3D gaze estimation methods struggle to generalize across diverse data domains, primarily due to $\textbf{i)}$ the scarcity of annotated datasets, and $\textbf{ii)}$ the insufficient diversity of labeled data. In this work, we present OmniGaze, a semi-supervised framework for 3D gaze estimation, which utilizes large-scale unlabeled data collected from diverse and unconstrained real-world environments to mitigate domain bias and generalize gaze estimation in the wild. First, we build a diverse collection of unlabeled facial images, varying in facial appearances, background environments, illumination conditions, head poses, and eye occlusions. In order to leverage unlabeled data spanning a broader distribution, OmniGaze adopts a standard pseudo-labeling strategy and devises a reward model to assess the reliability of pseudo labels. Beyond pseudo labels as 3D direction vectors, the reward model also incorporates visual embeddings extracted by an off-the-shelf visual encoder and semantic cues from gaze perspective generated by prompting a Multimodal Large Language Model to compute confidence scores.Then, these scores are utilized to select high-quality pseudo labels and weight them for loss computation.Extensive experiments demonstrate that OmniGaze achieves state-of-the-art performance on five datasets under both in-domain and cross-domain settings. Furthermore, we also evaluate the efficacy of OmniGaze as a scalable data engine to build a foundation model for gaze estimation, which exhibits robust zero-shot generalization performance on four unseen datasets. The source code will be released.



Paperid:1766
Authors:Kwangryeol Park, Jaeho Kim, Seulki Lee
Abstract:
Long-term Time Series Forecasting (LTSF) tasks, which leverage the current data sequence as input to predict the future sequence, have become increasingly crucial in real-world applications such as weather forecasting and planning of electricity consumption. However, state-of-the-art LTSF models often fail to achieve prediction output alignment for the same timestamps across lagged input sequences. Instead, these models exhibit low output alignment, resulting in fluctuation in prediction outputs for the same timestamps, undermining the model's reliability. To address this, we propose AliO (Align Outputs), a novel approach designed to improve the output alignment of LTSF models by reducing the discrepancies between prediction outputs for the same timestamps in both the time and frequency domains. To measure output alignment, we introduce a new metric, TAM (Time Alignment Metric), which quantifies the alignment between prediction outputs, whereas existing metrics such as MSE only capture the distance between prediction outputs and ground truths. Experimental results show that AliO effectively improves the output alignment, i.e., up to 58.2\% in TAM, while maintaining or enhancing the forecasting performance (up to 27.5\%). This improved output alignment increases the reliability of the LTSF models, making them more applicable in real-world scenarios. The code implementation is on an anonymous GitHub repository.



Paperid:1767
Authors:Joseph Wilson, Chris van der Heide, Liam Hodgkinson, Fred Roosta
Abstract:
While neural networks have demonstrated impressive performance across various tasks, accurately quantifying uncertainty in their predictions is essential to ensure their trustworthiness and enable widespread adoption in critical systems. Several Bayesian uncertainty quantification (UQ) methods exist that are either cheap or reliable, but not both. We propose a post-hoc, sampling-based UQ method for overparameterized networks at the end of training. Our approach constructs efficient and meaningful deep ensembles by employing a (stochastic) gradient-descent sampling process on appropriately linearized networks. We demonstrate that our method effectively approximates the posterior of a Gaussian Process using the empirical Neural Tangent Kernel. Through a series of numerical experiments, we show that our method not only outperforms competing approaches in computational efficiency--often reducing costs by multiple factors--but also maintains state-of-the-art performance across a variety of UQ metrics for both regression and classification tasks.



Paperid:1768
Authors:Jun-Peng Jiang, Yu Xia, Hai-Long Sun, Shiyin Lu, Qingguo Chen, Weihua Luo, Kaifu Zhang, De-Chuan Zhan, Han-Jia Ye
Abstract:
Abstract:Tabular reasoning involves multi-step information extraction and logical inference over tabular data. While recent advances have leveraged large language models (LLMs) for reasoning over structured tables, such high-quality textual representations are often unavailable in real-world settings, where tables typically appear as images. In this paper, we tackle the task of tabular reasoning from table images, leveraging privileged structured information available during training to enhance multimodal large language models (MLLMs). The key challenges lie in the complexity of accurately aligning structured information with visual representations, and in effectively transferring structured reasoning skills to MLLMs despite the input modality gap. To address these, we introduce TabUlar Reasoning with Bridged infOrmation (Turbo), a new framework for multimodal tabular reasoning with privileged structured tables. Turbo benefits from a structure-aware reasoning trace generator based on DeepSeek-R1, contributing to high-quality modality-bridged data. On this basis, Turbo repeatedly generates and selects the advantageous reasoning paths, further enhancing the model's tabular reasoning ability. Experimental results demonstrate that, with limited ($9$k) data, Turbo achieves state-of-the-art performance ($+7.2\%$ vs. previous SOTA) across multiple datasets.



Paperid:1769
Authors:Yihan Wang, Andrew Bai, Nanyun Peng, Cho-Jui Hsieh
Abstract:
Pre-trained Large Language Models (LLMs) require post-training methods such as supervised fine-tuning (SFT) on instruction-response pairs to enable instruction following. However, this process can cause forgetting in capabilities learned during pre-training. In this paper, we investigate the loss of context awareness after SFT, where context awareness is defined as the ability to extract and understand information from user-provided context. % and respond accordingly. Surprisingly, we discovered that the loss of context awareness occurs in instruction fine-tuned LLMs when the chat template is applied to input prompts. We identify that the performance decline is associated with a bias toward different roles learned during conversational instruction fine-tuning. The bias can be traced to training samples where the assistant response minimally relies on the user-provided instruction. Based on these observations, we propose a metric to identify context-dependent examples from general instruction fine-tuning datasets. We then apply conditional instruction fine-tuning with a context-dependency indicator, enabling the model to preserve context awareness after SFT. Experiments on four context-dependent downstream tasks and three pre-trained LLMs of different sizes show that our method effectively mitigates the loss of context awareness without compromising general instruction-following capabilities.



Paperid:1770
Authors:Jinouwen Zhang, Junjie Ren, Ma Qianhong, Jianyu Wu, Aobo Yang, Yan Lu, Lu Chen, Hairun Xie, Jing Wang, Miao Zhang, Wanli Ouyang, SHIXIANG TANG
Abstract:
Aircraft manufacturing is the jewel in the crown of industry, in which generating high-fidelity airfoil geometries with controllable and editable representations remains a fundamental challenge. Existing deep learning methods, which typically rely on predefined parametric representations (e.g., Bézier curves) or discrete point sets, face an inherent trade-off between expressive power and resolution adaptability.To tackle this challenge, we introduce FuncGenFoil, a novel function-space generative model that directly reconstructs airfoil geometries as function curves. Our method inherits the advantages of arbitrary-resolution sampling and smoothness from parametric functions, as well as the strong expressiveness of discrete point-based representations.Empirical evaluations demonstrate that FuncGenFoil improves upon state-of-the-art methods in airfoil generation, achieving a relative 74.4% reduction in label error and a 23.2% increase in diversity on the AF-200K dataset. Our results highlight the advantages of function-space modeling for aerodynamic shape optimization, offering a powerful and flexible framework for high-fidelity airfoil design.



Paperid:1771
Authors:Zhimeng Guo, Minhao Cheng
Abstract:
The rapid advancement of Large Language Models (LLMs) in code generation has raised significant attribution and intellectual property concerns. Code watermarking offers a potential solution but faces unique challenges due to programming languages' strict syntactic constraints and semantic requirements.To address these challenges, we introduce ACW (AST-guided Code Watermarking), a novel adaptive framework that leverages Abstract Syntax Tree (AST) analysis during training to learn watermark embedding strategies. Our framework identifies substitutable code components and strategically biases token selections to embed watermarks. We also propose a novel sampling scheme that distributes tokens between green/red lists according to semantic context, ensuring statistical distinguishability while preserving code functionality. Extensive experiments demonstrate that ACW achieves a significant improvement in watermark detection accuracy compared to existing methods, with negligible impact on code functionality. This adaptive framework offers a promising solution for effective and practical code watermarking in the age of LLMs. Our code is available at: https://anonymous.4open.science/r/acw-36F3/.



Paperid:1772
Authors:Haibo Jin, Peiyan Zhang, Man Luo, Haohan Wang
Abstract:
Large Language Models (LLMs) have shown remarkable progress across domains, yet their ability to perform inductive reasoning—inferring latent rules from sparse examples—remains limited. It is often assumed that chain-of-thought (CoT) prompting, as used in Large Reasoning Models (LRMs), enhances such reasoning. We investigate this assumption with creating four controlled, diagnostic game-based tasks—chess, Texas Hold’em, dice games, and blackjack—with hidden human-defined rules. We find that CoT reasoning can degrade inductive performance, with LRMs often underperforming their non-reasoning counterparts.To explain this, we present a theoretical framework that reveals how reasoning steps can amplify error through three failure modes: incorrect sub-task decomposition, incorrect sub-task solving, and incorrect final answer summarization. Based on our theoretical and empirical analysis, we introduce structured interventions that adapt CoT generation according to our identified failure types. These interventions improve inductive accuracy without retraining. Our findings suggest that effective (CoT) reasoning depends not only on taking more steps but also on ensuring those steps are well-structured.



Paperid:1773
Authors:Heiko Hoppe, Léo Baty, Louis Bouvier, Axel Parmentier, Maximilian Schiffer
Abstract:
Reinforcement learning (RL) is increasingly applied to real-world problems involving complex and structured decisions, such as routing, scheduling, and assortment planning. These settings challenge standard RL algorithms, which struggle to scale, generalize, and exploit structure in the presence of combinatorial action spaces. We propose Structured Reinforcement Learning (SRL), a novel actor-critic framework that embeds combinatorial optimization layers into the actor neural network. We enable end-to-end learning of the actor via Fenchel-Young losses and provide a geometric interpretation of SRL as a primal-dual algorithm in the dual of the moment polytope. Across six environments with exogenous and endogenous uncertainty, SRL matches or surpasses the performance of unstructured RL and imitation learning on static tasks and improves over these baselines by up to 92\% on dynamic problems, with improved stability and convergence speed.



Paperid:1774
Authors:Zijie Zhao, Zhongyue Zhao, Kaixuan Xu, Yuqian Fu, Jiajun Chai, Yuanheng Zhu, Dongbin Zhao
Abstract:
Multi-task multi-agent reinforcement learning (MT-MARL) aims to develop a single model capable of solving a diverse set of tasks. However, existing methods often fall short due to the substantial variation in optimal policies across tasks, making it challenging for a single policy model to generalize effectively. In contrast, we find that many tasks exhibitbounded similarityin their underlying dynamics—highly similar within certain groups (e.g., door-open/close) diverge significantly between unrelated tasks (e.g., door-open \& object-catch). To leverage this property, we reconsider the role of modularity in multi-task learning, and proposeM3W, a novel approach that applies mixture-of-experts (MoE) to world model instead of policy, enabling both learning and planning. For learning, it uses a SoftMoE-based dynamics model alongside a SparseMoE-based predictor to facilitate knowledge reuse across similar tasks while avoiding gradient conflicts across dissimilar tasks. For planning, it evaluates and optimizes actions using the predicted rollouts from the world model, without relying directly on a explicit policy model, thereby overcoming the limitations of policy-centric methods. As the first MoE-based multi-task world model, M3W demonstrates superior performance, sample efficiency, and multi-task adaptability, as validated on Bi-DexHands with 14 tasks and MA-Mujoco with 24 tasks. The demos and anonymous code are available at \url{https://sites.google.com/view/m3w-marl}.



Paperid:1775
Authors:Justin Wong, Yury Orlovskiy, Alexander Shypula, Michael Luo, Sanjit Seshia, Joseph Gonzalez
Abstract:
Generating diverse responses from large language models (LLMs) is crucial for applications such as planning/search and synthetic data generation, where diversity provides distinct answers across generations. Previous approaches rely solely on increasing the temperature, sacrificing quality. Furthermore, the model's next-token probabilities may not be representative of the true answer distribution. To combat these challenges, we propose SimpleStrat, an alternative that uses the language model itself to partition the space into strata from which to sample. To measure resampling diversity, we introduce CoverageQA, a dataset of underspecified questions with multiple equally plausible answers. We propose measuring resampling diversity as the KL Divergence between the output distribution and uniform distribution over valid ground truth answers and use recall as an alternative when assessing proprietary models. On CoverageQA, SimpleStrat improves diversity across all temperatures showing orthogonal benefits. Quantifiably, we achieves as much as 2X the recall compared to GPT-4o, and an average reduction in KL divergence by 0.36 compared on Llama 3. Furthermore, we show that SimpleStrat achieves more resampling diversity at temperature T=0 than scaling temperature to T=1 on creative writing, an open-ended domain.



Paperid:1776
Authors:George Tsoukalas, Rahul Saha, Amitayush Thakur, Sabrina Reguyal, Swarat Chaudhuri
Abstract:
We take two key steps in automating the open-ended discovery of new mathematical theories, a grand challenge in artificial intelligence. First, we introduce Fermat, a reinforcement learning (RL) environment that models concept discovery and theorem-proving using a set of symbolic actions, opening up a range of RL problems relevant to theory discovery. Second, we explore a specific problem through Fermat: automatically scoring theinterestingnessof mathematical objects. We investigate evolutionary algorithms for synthesizing nontrivial interestingness measures. In particular, we introduce an LLM-based evolutionary algorithm that features function abstraction, leading to notable improvements in learning elementary number theory over hard-coded baselines.



Paperid:1777
Authors:Nicola Dall'Asen, Xiaofeng Zhang, Reyhane Askari Hemmat, Melissa Hall, Jakob Verbeek, Adriana Romero-Soriano, Michal Drozdzal
Abstract:
Conditional image generative models hold considerable promise to produce infinite amounts of synthetic training data. Yet, recent progress in generation quality has come at the expense of generation diversity, limiting the utility of these models as a source of synthetic training data. Although guidance-based approaches have been introduced to improve the utility of generated data by focusing on quality or diversity, the (implicit or explicit) utility functions oftentimes disregard the potential distribution shift between synthetic and real data. In this work, we introduce Chamfer Guidance: a training-free guidance approach which leverages a handful of real exemplar images to characterize the quality and diversity of synthetic data. We show that by leveraging the proposed Chamfer Guidance, we can boost the diversity of the generations w.r.t. a dataset of real images while maintaining or improving the generation quality on ImageNet-1k and standard geo-diversity benchmarks. Our approach achieves state-of-the-art few-shot performance with as little as 2 exemplar real images, obtaining 96.4% in terms of precision, and 86.4% in terms of distributional coverage, which increase to 97.5% and 92.7%, respectively, when using 32 real images. We showcase the benefits of the Chamfer Guidance generation by training downstream image classifiers on synthetic data, achieving accuracy boost of up to 12% over the baselines. Furthermore, our approach does not require using the unconditional model, and thus obtains a 31% reduction in FLOPs w.r.t. classifier-free-guidance-based approaches at sampling time.



Paperid:1778
Authors:Samantha Chen, Oren Ciolli, Anastasios Sidiropoulos, Yusu Wang
Abstract:
Abstract:Neural networks offer a promising data-driven approach to tackle computationally challenging optimization problems. In this work, we introduce neural approximation frameworks for a family of geometric "extent measure" problems, including shape-fitting descriptors (e.g. minimum enclosing ball or annulus). Central to our approach is the \textit{alignment} of our neural model with a new variant of the classical $\varepsilon$-kernel technique from computational geometry. In particular, we develop a new relaxed-$\varepsilon$-kernel theory that maintains the approximation guarantees of the classical $\varepsilon$-kernels but with the crucial benefit that it can be easily implemented with \textit{bounded model complexity} (i.e, constant number of parameters) by the simple SumFormer neural network. This leads to a simple neural model to approximate objects such as the directional width of any input point set, and empirically shows excellent out-of-distribution generalization. Many geometric extent measures, such as the minimum enclosing spherical shell, cannot be directly captured by $\varepsilon$-kernels. To this end, we show that an encode-process-decode framework with our kernel approximating NN used as the ``process'' module can approximate such extent measures, again, with bounded model complexity where parameters scale only with the approximation error $\varepsilon$ and not the size of the input set. Empirical results on diverse point‐cloud datasets demonstrate the practical performance of our models.



Paperid:1779
Authors:Victor Dheur, Souhaib Ben Taieb
Abstract:
Reliably characterizing the full conditional distribution of a multivariate response variable given a set of covariates is crucial for trustworthy decision-making. However, misspecified or miscalibrated multivariate models may yield a poor approximation of the joint distribution of the response variables, leading to unreliable predictions and suboptimal decisions. Furthermore, standard recalibration methods are primarily limited to univariate settings, while conformal prediction techniques, despite generating multivariate prediction regions with coverage guarantees, do not provide a full probability density function. We address this gap by first introducing a novel notion of latent calibration, which assesses probabilistic calibration in the latent space of a conditional normalizing flow. Second, we propose latent recalibration (LR), a novel post-hoc model recalibration method that learns a transformation of the latent space with finite-sample bounds on latent calibration. Unlike existing methods, LR produces a recalibrated distribution with an explicit multivariate density function while remaining computationally efficient. Extensive experiments on both tabular and image datasets show that LR consistently improves latent calibration error and the negative log-likelihood of the recalibrated models.



Paperid:1780
Authors:Yuxiang Zhang, Zhengxu Yu, Weihang Pan, Zhongming Jin, Qiang Fu, Binbin Lin, Deng Cai, Jieping Ye
Abstract:
Emerging reasoning LLMs such as OpenAI-o1 and DeepSeek-R1 have achieved strong performance on complex reasoning tasks by generating long chain-of-thought (CoT) traces. However, these long CoTs result in increased token usage, leading to higher inference latency and memory consumption. As a result, balancing accuracy and reasoning efficiency has become essential for deploying reasoning LLMs in practical applications. Existing long-to-short (Long2Short) methods aim to reduce inference length but often sacrifice accuracy, revealing a need for an approach that maintains performance while lowering token costs. To address this efficiency-accuracy tradeoff, we propose TokenSqueeze, a novel Long2Short method that condenses reasoning paths while preserving performance and relying exclusively on self-generated data. First, to prevent performance degradation caused by excessive compression of reasoning depth, we propose to selectively identify self-generated samples that preserve the model's competency while minimizing token usage. To further optimize the linguistic expression without altering the underlying reasoning paths, we introduce a distribution-aligned linguistic refinement method that enhances the clarity and conciseness of the reasoning path while preserving its logical integrity. Comprehensive experimental results demonstrated the effectiveness of TokenSqueeze in reducing token usage while maintaining accuracy. Notably, DeepSeek‑R1‑Distill‑Qwen‑7B fine-tuned by using our proposed method achieved a 50\% average token reduction while preserving accuracy on the MATH500 benchmark. TokenSqueeze exclusively utilizes the model's self-generated data, enabling efficient and high-fidelity reasoning without relying on manually curated short-answer datasets across diverse applications. Our code is available at https://xxxxx/TokenSqueeze.



Paperid:1781
Authors:Abhijnan Nath, Nikhil Krishnaswamy
Abstract:
Large Language Models (LLMs) are increasingly bring deployed in agentic settings where they act as collaborators with humans. Therefore, it is increasingly important to be able to evaluate their abilities to collaborate effectively in multi-turn, multi-party tasks. In this paper, we build on the AI-alignment and “safe interruptability” literature to offer novel theoretical insights on LLM-agent collaborative behavior between a collaborator agent and a intervention agent, where the goal is to learn an ideal “partner-aware” collaborator that increases the group’s common-ground (CG)—alignment on task-relevant propositions—by intelligently collecting information provided in interventions by a partner agent. We show how LLM agents trained using standard RLHF and related approaches are naturally inclined to ignore possibly well-meaning interventions, which makes increasing group common ground non-trivial in this setting. We employ a two-player Modified-action MDP to examine on this suboptimal behavior of standard AI agents, and then propose Interruptible Collaborative Roleplayer (ICR)—a novel “partner-aware” learning algorithm to train CG-optimal collaborators. Experiments on multiple collaborative task environments show that ICR, on average, is more capable of promoting successful CG convergence and exploring more diverse solutions in such tasks.



Paperid:1782
Authors:Yushan Jiang, Wenchao Yu, Geon Lee, Dongjin Song, Kijung Shin, Wei Cheng, Yanchi Liu, Haifeng Chen
Abstract:
Time series analysis provides essential insights for real-world system dynamics and informs downstream decision-making, yet most existing methods often overlook the rich contextual signals present in auxiliary modalities. To bridge this gap, we introduce TimeXL, a multi-modal prediction framework that integrates a prototype-based time series encoder with three collaborating Large Language Models (LLMs) to deliver more accurate predictions and interpretable explanations. First, a multi-modal prototype-based encoder processes both time series and textual inputs to generate preliminary forecasts alongside case-based rationales. These outputs then feed into a prediction LLM, which refines the forecasts by reasoning over the encoder's predictions and explanations. Next, a reflection LLM compares the predicted values against the ground truth, identifying textual inconsistencies or noise. Guided by this feedback, a refinement LLM iteratively enhances text quality and triggers encoder retraining. This closed-loop workflow---prediction, critique (reflect), and refinement---continuously boosts the framework's performance and interpretability. Empirical evaluations on four real-world datasets demonstrate that TimeXL achieves up to 8.9\% improvement in AUC and produces human-centric, multi-modal explanations, highlighting the power of LLM-driven reasoning for time series prediction.



Paperid:1783
Authors:Bargav Jayaraman, Virendra Marathe, Hamid Mozaffari, William Shen, Krishnaram Kenthapadi
Abstract:
In enterprise settings, organizational data is segregated, siloed and carefully protected by elaborate access control frameworks. These access control structures can completely break down if an LLM fine-tuned on the siloed data serves requests, for downstream tasks, from individuals with disparate access privileges. We propose Permissioned LLMs (PermLLM), a new class of LLMs that superimpose the organizational data access control structures on query responses they generate. We formalize abstractions underpinning the means to determine whether access control enforcement happens correctly over LLM query responses. Our formalism also introduces a novel metric, called access advantage, to evaluate the efficacy of a PermLLM mechanism. We introduce three novel PermLLM mechanisms that build on Parameter Efficient Fine-Tuning to achieve the desired access control. We furthermore present two instantiations of access advantage–(i) Domain Distinguishability Index (DDI) based on Membership Inference Attacks, and (ii) Utility Gap Index (UGI) based on LLM utility evaluation. We demonstrate the efficacy of our PermLLM mechanisms through extensive experiments on four public datasets in addition to evaluating the validity of DDI and UGI metrics themselves for quantifying access control in LLMs.



Paperid:1784
Authors:Yinjie Wang, Ling Yang, Ye Tian, Ke Shen, Mengdi Wang
Abstract:
Mathematical reasoning in large language models has been successfully incentivized through reinforcement learning with verifiable rewards, leading to improved one-shot precision. In this work, we turn our focus to the coding domain. Beyond one-shot precision, we highlight unit test generation as another key factor for enhancing coding ability, since accurate unit tests are essential for enabling self-checking and self-correction during inference.Traditional approaches for fine-tuning LLMs on unit test generation rely heavily on ground-truth code solutions in the training data. We propose CURE, a novel reinforcement learning framework with a dedicated reward design that co-evolves coding and unit test generation capabilities based on their interaction outcomes—without any ground-truth code as supervision. This approach enables flexible and scalable training and allows the unit tester to learn directly from the coder’s mistakes.Through extensive evaluations, we demonstrate that our CURE models, derived from base models of varying sizes, excel in both code generation and unit test generation. They naturally extend to downstream tasks such as test-time scaling—achieving a 6.2\% improvement over the base model—and agentic unit test generation, with a 25.1\% improvement. Our 4B model consistently outperforms Qwen3-4B while achieving 64.8\% inference efficiency in unit test generation. Notably, we also find that the CURE model can serve as an effective reward model for reinforcement learning on base models, even in the absence of any labeled supervision.



Paperid:1785
Authors:Yiming Wang, Pei Zhang, Jialong Tang, Hao-Ran Wei, Baosong Yang, Rui Wang, Chenshu Sun, Feitong Sun, Jiran Zhang, Junxuan Wu, Qiqian Cang, Yichang Zhang, Fei Huang, Junyang Lin, Fei Huang, Jingren Zhou
Abstract:
In this paper, we introducePolyMath, a multilingual mathematical reasoning benchmark covering 18 languages and 4 easy-to-hard difficulty levels. Our benchmark ensures difficulty comprehensiveness, language diversity, and high-quality translation, making it a highly discriminative multilingual mathematical benchmark in the era of reasoning LLMs.We conduct a comprehensive evaluation for advanced LLMs and find that even Qwen-3-235B-A22B-Thinking and Gemini-2.5-pro, achieve only 54.6 and 52.2 benchmark scores, with about 40% accuracy under the highest level.From a language perspective, our benchmark reveals several key challenges of LLMs in multilingual reasoning:(1) Reasoning performance varies widely across languages for current LLMs;(2) Input-output language consistency is low in reasoning LLMs and may be correlated with performance;(3) The thinking length differs significantly by language for current LLMs.Additionally, we demonstrate that controlling the output language in the instructions has the potential to affect reasoning performance, especially for some low-resource languages, suggesting a promising direction for improving multilingual capabilities in LLMs.



Paperid:1786
Authors:Tobias Wegel, Geelon So, Junhyung Park, Fanny Yang
Abstract:
Abstract:In multi-objective learning, a set of $K$ learning objectives $\mathcal{E}_1,\ldots, \mathcal{E}_K$ must be solved at once by a single model $g \in \mathcal{G}$. Although good trade-offs may only be available in a complex model class, it comes with a statistical cost: we show that, in the worst-case, the amount of _labeled_ data available from each task constrains how complex $\mathcal{G}$ can be, even if the learner has access to the individual Bayes optimal models and infinite amounts of unlabeled data. At the same time, for an important class of problems, we show that the complexity of $\mathcal{G}$ can come into play only in terms of _unlabeled_ data. We establish sample complexity upper bounds, showing when and how unlabeled data can significantly alleviate the need for labeled data. These rates are achieved by a simple, semi-supervised method via pseudo-labeling.



Paperid:1787
Authors:Hossein Adeli, Sun Minni, Nikolaus Kriegeskorte
Abstract:
A major goal of neuroscience is to understand brain computations during visual processing in naturalistic settings. A dominant approach is to use image-computable deep neural networks trained with different task objectives as a basis for linear encoding models. However, in addition to requiring tuning a large number of parameters, the linear encoding approach ignores the structure of the feature maps both in the brain and the models. Recently proposed alternatives have focused on decomposing the linear mapping to spatial and feature components but focus on finding static receptive fields for units that are applicable only in early visual areas. In this work, we employ the attention mechanism used in the transformer architecture to study how retinotopic visual features can be dynamically routed to category-selective areas in high-level visual processing. We show that this computational motif is significantly more powerful than alternative methods in predicting brain activity during natural scene viewing, across different feature basis models and modalities. We also show that this approach is inherently more interpretable, without the need to create importance maps, by interpreting the attention routing signal for different high-level categorical areas. Our approach proposes a mechanistic model of how visual information from retinotopic maps can be routed based on the relevance of the input content to different category-selective regions.



Paperid:1788
Authors:Yashas Annadani, Syrine Belakaria, Stefano Ermon, Stefan Bauer, Barbara Engelhardt
Abstract:
Offline multi-objective optimization aims to identify Pareto-optimal solutions given a dataset of designs and their objective values. In this work, we propose a preference-guided diffusion model that generates Pareto-optimal designs by leveraging a classifier-based guidance mechanism. Our guidance classifier is a preference model trained to predict the probability that one design dominates another, directing the diffusion model toward optimal regions of the design space. Crucially, this preference model generalizes beyond the training distribution, enabling the discovery of Pareto-optimal solutions outside the observed dataset. We introduce a novel diversity-aware preference guidance, augmenting Pareto dominance preference with diversity criteria. This ensures that generated solutions are optimal and well-distributed across the objective space, a capability absent in prior generative methods for offline multi-objective optimization. We evaluate our approach on various continuous offline multi-objective optimization tasks and find that it consistently outperforms other inverse/generative approaches while remaining competitive with forward/surrogate-based optimization methods. Our results highlight the effectiveness of classifier-guided diffusion models in generating diverse and high-quality solutions that approximate the Pareto front well.



Paperid:1789
Authors:Benedetta Liberatori, Alessandro Conti, Lorenzo Vaquero, Yiming Wang, Elisa Ricci, Paolo Rota
Abstract:
What does it mean for two videos to be similar? Videos may appear similar when judged by the actions they depict, yet entirely different if evaluated based on the locations where they were filmed. While humans naturally compare videos by taking different aspects into account, this ability has not been thoroughly studied and presents a challenge for models that often depend on broad global similarity scores. Large Multimodal Models (LMMs) with video understanding capabilities open new opportunities for leveraging natural language in comparative video tasks. We introduce Concept-based Video Similarity estimation (ConViS), a novel task that compares pairs of videos by computing interpretable similarity scores across a predefined set of key semantic concepts. ConViS allows for human-like reasoning about video similarity and enables new applications such as concept-conditioned video retrieval. To support this task, we also introduce ConViS-Bench, a new benchmark comprising carefully annotated video pairs spanning multiple domains. Each pair comes with concept-level similarity scores and textual descriptions of both differences and similarities. Additionally, we benchmark several state-of-the-art models on ConViS, providing insights into their alignment with human judgments. Our results reveal significant performance differences on ConViS, indicating that some concepts present greater challenges for estimating video similarity. We believe that ConViS-Bench will serve as a valuable resource for advancing research in language-driven video understanding.



Paperid:1790
Authors:Dongkwan Lee, JunHoo Lee, Nojun Kwak
Abstract:
We introduce the Deep Edge Filter, a novel approach that applies high-pass filtering to deep neural network features to improve model generalizability. Our method is motivated by our hypothesis that neural networks encode task-relevant semantic information in high-frequency components while storing domain-specific biases in low-frequency components of deep features. By subtracting low-pass filtered outputs from original features, our approach isolates generalizable representations while preserving architectural integrity. Experimental results across diverse domains such as Vision, Text, 3D, and Audio demonstrate consistent performance improvements regardless of model architecture and data modality. Analysis reveals that our method induces feature sparsification and effectively isolates high-frequency components, providing empirical validation of our core hypothesis.



Paperid:1791
Authors:Ruilin Luo, Zhuofan Zheng, Yifan Wang, Xinzhe Ni, Zicheng Lin, Songtao Jiang, Yiyao Yu, Chufan Shi, Ruihang Chu, Lei Wang, Jin zeng, Yujiu Yang
Abstract:
Process Reward Models (PRMs) have shown promise in enhancing the mathematical reasoning capabilities of Large Language Models (LLMs) through Test-Time Scaling (TTS). However, their integration into multimodal reasoning remains largely unexplored. In this work, we take the first step toward unlocking the potential of PRMs in multimodal mathematical reasoning. We identify three key challenges: (i) the scarcity of high-quality reasoning data constrains the capabilities of foundation Multimodal Large Language Models (MLLMs), which imposes further limitations on the upper bounds of TTS and reinforcement learning (RL); (ii) a lack of automated methods for process labeling within multimodal contexts persists; (iii) the employment of process rewards in unimodal RL faces issues like reward hacking, which may extend to multimodal scenarios. To address these issues, we introduce URSA, a three-stage Unfolding multimodal pRocess-Supervision Aided training framework. We first construct MMathCoT-1M, a high-quality large-scale multimodal Chain-of-Thought (CoT) reasoning dataset, to build a stronger math reasoning foundation MLLM, URSA-8B. Subsequently, we go through an automatic process to synthesize process supervision data, which emphasizes both logical correctness and perceptual consistency. We introduce DualMath-1.1M to facilitate the training of URSA-8B-RM. Finally, we propose Process-Supervised Group-Relative-Policy-Optimization (PS-GRPO), pioneering a multimodal PRM-aided online RL method that outperforms vanilla GRPO. With PS-GRPO application, URSA-8B-PS-GRPO outperforms Gemma3-12B and GPT-4o by 8.4% and 2.7% on average across 6 benchmarks.



Paperid:1792
Authors:Sofiane ENNADIR, Levente Zólyomi, Oleg Smirnov, Tianze Wang, John Pertoft, Filip Cornell, Lele Cao
Abstract:
Abstract:Transformer models have become the dominant backbone for sequence modeling, leveraging self-attention to produce contextualized token representations. These are typically aggregated into fixed-size vectors via pooling operations for downstream tasks. While much of the literature has focused on attention mechanisms, the role of pooling remains underexplored despite its critical impact on model behavior. In this paper, we introduce a theoretical framework that rigorously characterizes the expressivity of Transformer-based models equipped with widely used pooling methods by deriving closed-form bounds on their representational capacity and the ability to distinguish similar inputs. Our analysis extends to different variations of attention formulations, demonstrating that these bounds hold across diverse architectural variants. We empirically evaluate pooling strategies across tasks requiring both $\textit{global}$ and $\textit{local}$ contextual understanding, spanning three major modalities: computer vision, natural language processing, and time-series analysis. Results reveal consistent trends in how pooling choices affect accuracy, sensitivity, and optimization behavior. Our findings unify theoretical and empirical perspectives, providing practical guidance for selecting or designing pooling mechanisms suited to specific tasks. This work positions pooling as a key architectural component in Transformer models and lays the foundation for more principled model design beyond attention alone. We make source code available to support reproducibility of results.



Paperid:1793
Authors:Meyer Scetbon, Chao Ma, Wenbo Gong, Ted Meeds
Abstract:
Abstract:Training large language models (LLMs) typically relies on adaptive optimizers like Adam (Kingma & Ba, 2015), which store additional state information to accelerate convergence but incur significant memory overhead. Recent efforts, such as SWAN (Ma et al., 2024), address this by eliminating the need for optimizer states while achieving performance comparable to Adam via a multi-step preprocessing procedure applied to instantaneous gradients. Motivated by the success of SWAN, we introduce a novel framework for designing stateless optimizers that normalizes stochastic gradients according to multiple norms. To achieve this, we propose a simple alternating scheme to enforce the normalization of gradients w.r.t these norms. We show that our procedure can produce, up to an arbitrary precision, a fixed-point of the problem, and that SWAN is a particular instance of our approach with carefully chosen norms, providing a deeper understanding of its design. Using our principled perspective, we develop of a more efficient, scalable, and practical stateless optimizer. Our algorithm relaxes the properties of SWAN, significantly reducing its computational cost while retaining its memory efficiency, making it applicable to training large-scale models. Experiments on pretraining LLaMA models with up to 1B parameters demonstrate a $3$× speedup over Adam with significantly reduced memory requirements.



Paperid:1794
Authors:Lehan He, Zeren Chen, Zhelun Shi, Tianyu Yu, Lu Sheng, Jing Shao
Abstract:
Abstract:The success of Direct Preference Optimization (DPO) in mitigating hallucinations in Vision Language Models (VLMs) critically hinges on the true reward gaps within preference pairs. However, current methods, typically relying on ranking or rewriting strategies, often struggle to optimize these reward gaps in a systematic way during data curation. A core difficulty lies in precisely characterizing and strategically manipulating the overall reward gap configuration, that is, the deliberate design of how to shape these reward gaps within each preference pair across the data. To address this, we introduce Topic-level Preference Rewriting (TPR), a novel framework designed for the systematic optimization of reward gap configuration. Through selectively replacing semantic topics within VLM responses with model’s own resampled candidates for targeted rewriting, TPR can provide topic-level control over fine-grained semantic details. This precise control enables advanced data curation strategies, such as progressively adjusting the difficulty of rejected responses, thereby sculpting an effective reward gap configuration that guides the model to overcome challenging hallucinations. Comprehensive experiments demonstrate TPR achieves state-of-the-art performance on multiple hallucination benchmarks, outperforming previous methods by an average of $\sim$20%. Notably, it significantly reduces hallucinations by up to 93% on ObjectHal-Bench, and also exhibits superior data efficiency towards robust and cost-effective VLM alignment.



Paperid:1795
Authors:Michal Nazarczuk, Sibi Catley-Chandar, Thomas Tanay, Zhensong Zhang, Greg Slabaugh, Eduardo Pérez-Pellitero
Abstract:
Dynamic Novel View Synthesis aims to generate photorealistic views of moving subjects from arbitrary viewpoints. This task is particularly challenging when relying on monocular video, where disentangling structure from motion is ill-posed and supervision is scarce. We introduce Video Diffusion-Aware Reconstruction (ViDAR), a novel 4D reconstruction framework that leverages personalised diffusion models to synthesise a pseudo multi-view supervision signal for training a Gaussian splatting representation. By conditioning on scene-specific features, ViDAR recovers fine-grained appearance details while mitigating artefacts introduced by monocular ambiguity. To address the spatio-temporal inconsistency of diffusion-based supervision, we propose a diffusion-aware loss function and a camera pose optimisation strategy that aligns synthetic views with the underlying scene geometry. Experiments on DyCheck, a challenging benchmark with extreme viewpoint variation, show that ViDAR outperforms all state-of-the-art baselines in visual quality and geometric consistency. We further highlight ViDAR’s strong improvement over baselines on dynamic regions and provide a new benchmark to compare performance in reconstructing motion-rich parts of the scene.



Paperid:1796
Authors:Jiawei Lian, Jianhong Pan, Lefan Wang, Yi Wang, Shaohui Mei, Lap-Pui Chau
Abstract:
Large Language Models (LLMs) increasingly employ alignment techniques to prevent harmful outputs. Despite these safeguards, attackers can circumvent them by crafting prompts that induce LLMs to generate harmful content. Current methods typically target exact affirmative responses, suffering from limited convergence, unnatural prompts, and high computational costs. We introduce semantic representation attacks, a novel paradigm that fundamentally reconceptualizes adversarial objectives against aligned LLMs. Rather than targeting exact textual patterns, our approach exploits the semantic representation space that can elicit diverse responses that share equivalent harmful meanings. This innovation resolves the inherent trade-off between attack effectiveness and prompt naturalness that plagues existing methods. Our Semantic Representation Heuristic Search (SRHS) algorithm efficiently generates semantically coherent adversarial prompts by maintaining interpretability during incremental search. We establish rigorous theoretical guarantees for semantic convergence and demonstrate that SRHS achieves unprecedented attack success rates (89.4% averaged across 18 LLMs, including 100% on 11 models) while significantly reducing computational requirements. Extensive experiments show that our method consistently outperforms existing approaches.



Paperid:1797
Authors:Pengxiang Li, Shilin Yan, Jiayin Cai, Renrui Zhang, Ruichuan An, Ziyu Guo, Xiaowei Gao
Abstract:
Classifier-Free Guidance (CFG) significantly enhances controllability in generative models by interpolating conditional and unconditional predictions. However, standard CFG often employs a static unconditional input, which can be suboptimal for iterative generation processes where model uncertainty varies dynamically. We introduce Adaptive Classifier-Free Guidance (A-CFG), a novel method that tailors the unconditional input by leveraging the model's instantaneous predictive confidence. At each step of an iterative (masked) diffusion language model, A-CFG identifies tokens in the currently generated sequence for which the model exhibits low confidence. These tokens are temporarily re-masked to create a dynamic, localized unconditional input. This focuses CFG's corrective influence precisely on areas of ambiguity, leading to more effective guidance. We integrate A-CFG into a state-of-the-art masked diffusion language model and demonstrate its efficacy. Experiments on diverse language generation benchmarks show that A-CFG yields substantial improvements over standard CFG, achieving, for instance, a 3.9 point gain on GPQA. Our work highlights the benefit of dynamically adapting guidance mechanisms to model uncertainty in iterative generation.



Paperid:1798
Authors:Ziyuan He, Yuxuan Wang, Jiaqi Li, Kexin Liang, Muhan Zhang
Abstract:
Large language models (LLMs) are equipped with increasingly extended context windows recently, yet their long context understanding capabilities over long dependency tasks remains fundamentally limited and underexplored. This gap is especially significant in many real-world long-context applications that were rarely benchmarked. In this paper, we introduce \textbf{LooGLE v2}, a novel benchmark designed to evaluate LLMs' long context ability in real-world applications and scenarios. Our benchmark consists of automatically collected real-world long texts, ranging from 16k to 2M tokens, encompassing domains in law, finance, game and code. Accordingly, we delicately design 10 types of domain-specific long-dependency tasks and generate 1,934 QA instances with various diversity and complexity in a scalable data curation pipeline for further practical needs. We make comprehensive assessment on 6 locally deployed and 4 API-based LLMs. The evaluation results show that even the the best-performing model achieves only a 59.2\% overall score on our benchmark. Despite the extensive context windows, popular LLMs are only capable of understanding a much shorter length of context than they claim to be, revealing significant limitations in their ability to handle real-world tasks with long-range dependencies and highlighting substantial room for model improvement in practical long-context understanding.



Paperid:1799
Authors:Hong Huang, Jinhai Yang, Yuan Chen, Jiaxun Ye, Dapeng Wu
Abstract:
Federated Learning (FL) enables collaborative model training across distributed clients without data sharing, but its high computational and communication demands strain resource-constrained devices. While existing methods use dynamic pruning to improve efficiency by periodically adjusting sparse model topologies while maintaining sparsity, these approaches suffer from issues such asgreedy adjustments,unstable topologies, andcommunication inefficiency, resulting in less robust models and suboptimal performance under data heterogeneity and partial client availability. To address these challenges, we proposeFederatedRobust pruning via combinatorialThompsonSampling (FedRTS),a novel framework designed to develop robust sparse models. FedRTS enhances robustness and performance through its Thompson Sampling-based Adjustment (TSAdj) mechanism, which uses probabilistic decisions informed by stable and farsighted information, instead of deterministic decisions reliant on unstable and myopic information in previous methods. Extensive experiments demonstrate that FedRTS achieves state-of-the-art performance in computer vision and natural language processing tasks while reducing communication costs, particularly excelling in scenarios with heterogeneous data distributions and partial client participation. Our codes are available at: https://anonymous.4open.science/r/FedRTS-4567/.



Paperid:1800
Authors:Arjun Karuvally, Pichsinee Lertsaroj, Terrence Sejnowski, Hava Siegelmann
Abstract:
Abstract:The energy paradigm, exemplified by Hopfield networks, offers a principled framework for memory in neural systems by interpreting dynamics as descent on an energy surface. While powerful for static associative memories, it falls short in modeling sequential memory, where transitions between memories are essential. We introduce the Exponential Dynamic Energy Network (EDEN), a novel architecture that extends the energy paradigm to temporal domains by evolving the energy function over multiple timescales. EDEN combines a static high-capacity energy network with a slow, asymmetrically interacting modulatory population, enabling robust and controlled memory transitions. We formally derive short-timescale energy functions that govern local dynamics and use them to analytically compute memory escape times, revealing a phase transition between static and dynamic regimes. The analysis of capacity for EDEN shows that it achieves exponential sequence memory capacity $\mathcal{O}(\gamma^N)$, outperforming the linear capacity $\mathcal{O}(N)$ of conventional models. Furthermore, EDEN’s dynamics resemble the activity of time and ramping cells observed in human episodic memory, grounding its biological relevance. By unifying static and sequential memory within a dynamic energy framework, EDEN offers a scalable and interpretable model for high-capacity temporal memory in both artificial and biological systems.



Paperid:1801
Authors:Yoshihiro Yamada
Abstract:
Abstract:Transformers have driven remarkable breakthroughs in natural language processing and computer vision, yet their standard attention mechanism still imposes $O(N^2)$ complexity, hindering scalability to longer sequences. We introduce Circular-convolutional ATtention (CAT), a Fourier-based approach that efficiently applies circular convolutions to reduce complexity without sacrificing representational power. CAT achieves $O(N \log N)$ computations, requires fewer learnable parameters by streamlining fully connected layers, and introduces no heavier operations, resulting in consistent accuracy improvements and about a 10\% speedup in naive PyTorch implementations. Based on the engineering-isomorphic transformer framework, CAT's design not only offers practical efficiency and ease of implementation, but also provides insights to guide the development of future high-performance Transformer architectures. Finally, our ablation studies highlight the key conditions underlying CAT’s success, shedding light on broader principles for scalable attention mechanisms.



Paperid:1802
Authors:Mingjie Li, Wai Man Si, Michael Backes, Yang Zhang, Yisen Wang
Abstract:
Despite the impressive performance of general-purpose large language models (LLMs), they often require fine-tuning or post-training to excel at specific tasks. For instance, large reasoning models (LRMs), such as the DeepSeek-R1 series, demonstrate strong reasoning capabilities after post-training different general large language models on diverse chain-of-thought (CoT) datasets. However, this additional training frequently comes at the cost of reduced safety, as the fine-tuned or post-trained models tend to exhibit more harmful behaviors compared with the regular LLMs before post-training or fine-tuning, potentially leading to harmful outcomes due to their enhanced capabilities. Taking LRMs as an example, we first investigate the underlying cause of this safety degradation in this paper. Our analysis reveals that post-training can mask the original safety mechanisms of the base LLM, while over-amplifying representations related to their post-training ability. But luckily, we also find that LRMs' safety mechanisms still exist instead of being removed during their post-training. Based on these findings, we propose a lightweight and cost-effective solution called SafeReAct that restores the suppressed safety behaviors by aligning with LoRA adapters on a few layers. Experiments on four state-of-the-art LRMs show that our method significantly improves safety on harmful prompts without compromising reasoning performance. Besides LRMs, additional results on other domain-specific LLMs, like medical models, further confirm the generality and effectiveness of our approach.



Paperid:1803
Authors:Noa Shoham, Ron Dorfman, Shalev Shaer, Kfir Y. Levy, Yaniv Romano
Abstract:
Prediction-Powered Inference (PPI) is a recently proposed statistical inference technique for parameter estimation that leverages pseudo-labels on both labeled and unlabeled data to construct an unbiased, low-variance estimator.In this work, we extend its core idea to semi-supervised learning (SSL) for model training, introducing a novel unbiased gradient estimator. This extension addresses a key challenge in SSL: while unlabeled data can improve model performance, its benefit heavily depends on the quality of pseudo-labels. Inaccurate pseudo-labels can introduce bias, leading to suboptimal models.To balance the contributions of labeled and pseudo-labeled data, we utilize an interpolation parameter and tune it on the fly, alongside the model parameters, using a one-dimensional online learning algorithm. We verify the practical advantage of our approach through experiments on both synthetic and real datasets, demonstrating improved performance over classic SSL baselines and PPI methods that tune the interpolation parameter offline.



Paperid:1804
Authors:Yuchen Zhu, Wei Guo, Jaemoo Choi, Guan-Horng Liu, Yongxin Chen, Molei Tao
Abstract:
Abstract:We study the problem of learning a neural sampler to generate samples from discrete state spaces where the target probability mass function $\pi\propto\exp(U)$ is known up to a normalizing constant, which is an important task in fields such as statistical physics, machine learning, combinatorial optimization, etc. To better address this challenging task when the state space has a large cardinality and the distribution is multi-modal, we propose $\textbf{M}$asked $\textbf{D}$iffusion $\textbf{N}$eural $\textbf{S}$ampler ($\textbf{MDNS}$), a novel framework for training discrete neural samplers by aligning two path measures through a family of learning objectives, theoretically grounded in the stochastic optimal control of the continuous-time Markov chains. We validate the efficiency and scalability of MDNS through extensive experiments on various distributions with distinct statistical properties, where MDNS learns to accurately sample from the target distributions despite the extremely high problem dimensions and outperforms other learning-based baselines by a large margin. A comprehensive study of ablations and extensions is also provided to demonstrate the efficacy and potential of the proposed framework.



Paperid:1805
Authors:Ziming Wei, Bingqian Lin, Zijian Jiao, Yunshuang Nie, Liang Ma, Yuecheng Liu, Yuzheng Zhuang, Xiaodan Liang
Abstract:
Spatial Planning is a crucial part in the field of spatial intelligence, which requires the understanding and planning about object arrangements in space perspective. AI agents with the spatial planning ability can better adapt to various real-world applications, including robotic manipulation, automatic assembly, urban planning etc. Recent works have attempted to construct benchmarks for evaluating the spatial intelligence of Multimodal Large Language Models (MLLMs). Nevertheless, these benchmarks primarily focus on spatial reasoning based on typical Visual Question-Answering (VQA) forms, which suffers from the gap between abstract spatial understanding and concrete task execution. In this work, we take a step further to build a comprehensive benchmark called MineAnyBuild, aiming to evaluate the spatial planning ability of open-world AI agents in the Minecraft game. Specifically, MineAnyBuild requires an agent to generate executable architecture building plans based on the given multi-modal human instructions. It involves 4,000 curated spatial planning tasks and also provides a paradigm for infinitely expandable data collection by utilizing rich player-generated content. MineAnyBuild evaluates spatial planning through four core supporting dimensions: spatial understanding, spatial reasoning, creativity, and spatial commonsense. Based on MineAnyBuild, we perform a comprehensive evaluation for existing MLLM-based agents, revealing the severe limitations but enormous potential in their spatial planning abilities. We believe our MineAnyBuild will open new avenues for the evaluation of spatial intelligence and help promote further development for open-world AI agents capable of spatial planning.



Paperid:1806
Authors:Zhixin Xie, Xurui Song, Jun Luo
Abstract:
Despite substantial efforts in safety alignment, recent research indicates that Large Language Models (LLMs) remain highly susceptible to jailbreak attacks. Among these attacks, finetuning-based ones that compromise LLMs' safety alignment via fine-tuning stand out due to their stable jailbreak performance. In particular, a recent study indicates that fine-tuning with as few as 10 harmful question-answer (QA) pairs can lead to successful jailbreaking across various harmful questions. However, such \textit{malicious fine-tuning} attacks are readily detectable and hence thwarted by moderation models. In this paper, we demonstrate that LLMs can be jailbroken by fine-tuning with only 10 \textit{benign} QA pairs; our attack exploits the increased sensitivity of LLMs to fine-tuning data after being \textit{overfitted}. Specifically, our fine-tuning process starts with overfitting an LLM via fine-tuning with benign QA pairs involving identical \textit{refusal} answers. Further fine-tuning is then performed with standard benign answers, causing the overfitted LLM to forget the refusal attitude and thus provide compliant answers regardless of the harmfulness of a question. We implement our attack on the ten LLMs and compare it with five existing baselines. Experiments demonstrate that our method achieves significant advantages in both attack effectiveness and attack stealth. Our findings expose previously unreported security vulnerabilities in current LLMs and provide a new perspective on understanding how LLMs' security is compromised, even with benign fine-tuning. Our code is available at https://anonymous.4open.science/r/ten_benign-EBF2/readme.md.



Paperid:1807
Authors:Niklas Lauffer, Ameesh Shah, Micah Carroll, Sanjit Seshia, Stuart J Russell, Michael Dennis
Abstract:
Adversarial optimization algorithms that explicitly search for flaws in agents' policies have been successfully applied to finding robust and diverse policies in the context of multi-agent learning. However, the success of adversarial optimization has been largely limited to zero-sum settings because its naive application in cooperative settings leads to a critical failure mode: agents are irrationally incentivized toself-sabotage, blocking the completion of tasks and halting further learning. To address this, we introduceRationality-preserving Policy Optimization (RPO), a formalism for adversarial optimization that avoids self-sabotage by ensuring agents remainrational—that is, their policies are optimal with respect to some possible partner policy. To solve RPO, we developRational Policy Gradient (RPG), which trains agents to maximize their own reward in a modified version of the original game in which we useopponent shapingtechniques to optimize the adversarial objective. RPG enables us to extend a variety of existing adversarial optimization algorithms that, no longer subject to the limitations of self-sabotage, can find adversarial examples, improve robustness and adaptability, and learn diverse policies. We empirically validate that our approach achieves strong performance in several popular cooperative and general-sum environments.



Paperid:1808
Authors:Sharvaree Vadgama, Mohammad Islam, Domas Buracas, Christian A Shewmake, Artem Moskalev, Erik Bekkers
Abstract:
In this work, we explore the trade-offs of explicit structural priors, particularly group-equivariance. We address this through theoretical analysis and a comprehensive empirical study. To enable controlled and fair comparisons, we introduce \texttt{Rapidash}, a unified group convolutional architecture that allows for different variants of equivariant and non-equivariant models. Our results suggest that more constrained equivariant models outperform less constrained alternatives when aligned with the geometry of the task, and increasing representation capacity does not fully eliminate performance gaps. We see improved performance of models with equivariance and symmetry-breaking through tasks like segmentation, regression, and generation across diverse datasets. Explicit \textit{symmetry breaking} via geometric reference frames consistently improves performance, while \textit{breaking equivariance} through geometric input features can be helpful when aligned with task geometry. Our results provide task-specific performance trends that offer a more nuanced way for model selection.



Paperid:1809
Authors:Madhur Panwar, Gail Weiss, Navin Goyal, Antoine Bosselut
Abstract:
Abstract:Memorization in language models is a critical yet poorly understood phenomenon. In this work, we investigate memorization in transformer-based language models by analyzing their training dynamics over multiple epochs. We find that memorization is neither a constant accumulation of sequences nor simply dictated by the recency of exposure to these sequences. Instead, much like generalization, memorization appears to be driven by pattern recognition. Tracking memorization dynamics in mixed datasets, we observe that models memorize different sub-datasets in distinct bursts, suggesting that each subset is associated with unique underlying patterns, and that the model prefers to learn these patterns in a predictable order. While easily learnable patterns tend to support generalization on unseen data, more complex patterns do not. Furthermore, in datasets with weak or absent patterns, models may delay memorization while seeking them, a behavior we term $\textit{overthinking}$. Our results show that the subset of sequences memorized by a model over time is not arbitrary, and give insights into the internal processes a model goes through during training.



Paperid:1810
Authors:Patrick Lutz, Aditya Gangrade, Hadi Daneshmand, Venkatesh Saligrama
Abstract:
A transformer is merely a stack of learned data–to–data maps—yet those maps can hide rich algorithms. We train a linear, attention-only transformer on millions of masked-block completion tasks: each prompt is a masked low-rank matrix whose missing block may be (i) a scalar prediction target or (ii) an unseen kernel slice for Nyström extrapolation. The model sees only input–output pairs and a mean-squared loss; it is given no normal equations, no handcrafted iterations, and no hint that the tasks are related. Surprisingly, after training, algebraic unrolling reveals the same parameter-free update rule across all three resource regimes (full visibility, bandwidth-limited heads, rank-limited attention). We prove that this rule achieves second-order convergence on full-batch problems, cuts distributed iteration complexity, and remains accurate with compute-limited attention. Thus, a transformer trained solely to patch missing blocks implicitly discovers a unified, resource-adaptive iterative solver spanning prediction, estimation, and Nyström extrapolation—highlighting a powerful capability of in-context learning.



Paperid:1811
Authors:Ortal Senouf, Antoine Wehenkel, Cédric Vincent-Cuaz, Emmanuel Abbe, Pascal Frossard
Abstract:
Simulation-based inference (SBI) of latent parameters in physical systems is often hindered by model misspecification--the mismatch between simulated and real-world observations caused by inherent modeling simplifications. RoPE, a recent SBI approach, addresses this challenge through a two-stage domain transfer process that combines semi-supervised calibration with optimal transport (OT)-based distribution alignment. However, RoPE operates in a fully transductive setting, requiring access to a batch of test samples at inference time, which limits scalability and generalization. We propose a fully inductive and amortized SBI framework that integrates calibration and distributional alignment into a single, end-to-end trainable model. Our method leverages mini-batch OT with a closed-form coupling to align real and simulated observations that correspond to the same latent parameters, using both paired calibration data and unpaired samples. A conditional normalizing flow is then trained to approximate the OT-induced posterior, enabling efficient inference without simulation access at test time. Across a range of synthetic and real-world benchmarks--including complex medical biomarker estimation--our approach matches or exceeds the performance of RoPE, while offering improved scalability and applicability in challenging, misspecified environments.



Paperid:1812
Authors:Marzi Heidari, Hanping Zhang, Yuhong Guo
Abstract:
The challenge of learning with noisy labels is significant in machine learning, as it can severely degrade the performance of prediction models if not addressed properly. This paper introduces a novel framework that conceptualizes noisy label correction as a reinforcement learning (RL) problem. The proposed approach, Reinforcement Learning for Noisy Label Correction (RLNLC), defines a comprehensive state space representing data and their associated labels, an action space that indicates possible label corrections, and a reward mechanism that evaluates the efficacy of label corrections. RLNLC learns a deep feature representation based policy network to perform label correction through reinforcement learning, utilizing an actor-critic method.The learned policy is subsequently deployed to iteratively correct the noisy training labels and support prediction model training. The effectiveness of RLNLC is demonstrated through extensive experiments on multiple benchmark datasets, where it consistently outperforms existing state-of-the-art techniques for learning from noisy labels.



Paperid:1813
Authors:Amit Peleg, Naman Deep Singh, Matthias Hein
Abstract:
Vision-language models like CLIP have demonstrated remarkable zero-shot capabilities in classification and retrieval.However, these models often struggle with compositional reasoning – the ability to understand the relationships between concepts. A recent benchmark, SugarCrepe++, reveals that previous works on improving compositionality have mainly improved lexical sensitivity but neglected semantic understanding. In addition, downstream retrieval performance often deteriorates, although one would expect that improving compositionality should enhance retrieval. In this work, we introduce CLIC (Compositionally-aware Learning in CLIP), a fine-tuning method based on a novel training technique combining multiple images and their associated captions. CLIC improves compositionality across architectures as well as differently pre-trained CLIP models, both in terms of lexical and semantic understanding, and achieves consistent gains in retrieval performance. This even applies to the recent CLIPS, which achieves SOTA retrieval performance. Nevertheless, the short fine-tuning with CLIC leads to an improvement in retrieval and to the best compositional CLIP model on SugarCrepe++.



Paperid:1814
Authors:Penghui Qi, Zichen Liu, Tianyu Pang, Chao Du, Wee Sun Lee, Min Lin
Abstract:
Scaling test-time compute is crucial for enhancing the reasoning capabilities of large language models (LLMs). Existing approaches typically employ reinforcement learning (RL) to maximize a verifiable reward obtained at the end of reasoning traces. However, such methods optimize only the final performance under a large and fixed token budget, which hinders efficiency in both training and deployment. In this work, we present a novel framework,AnytimeReasoner, to optimizeanytime reasoning performance, which aims to improve token efficiency and the flexibility of reasoning under varying token budget constraints. To achieve this, we truncate the complete thinking process to fit within sampled token budgets from a prior distribution, compelling the model to summarize the optimal answer for each truncated thinking for verification. This introduces verifiable dense rewards into the reasoning process, facilitating more effective credit assignment in RL optimization. We then optimize the thinking and summary policies in a decoupled manner to maximize the cumulative reward. Additionally, we introduce a novel variance reduction technique,Budget Relative Policy Optimization (BRPO), to enhance the robustness and efficiency of the learning process when reinforcing the thinking policy. Empirical results in mathematical reasoning tasks demonstrate that our method consistently outperforms GRPO across all thinking budgets under various prior distributions, enhancing both training and token efficiency.



Paperid:1815
Authors:Deepan Adak, Yogesh Rawat, Shruti Vyas
Abstract:
Molecular property prediction is a fundamental task in computational chemistry with critical applications in drug discovery and materials science. While recent works have explored Large Language Models (LLMs) for this task, they primarily rely on textual molecular representations such as SMILES/SELFIES, which can be ambiguous and structurally uninformative. In this work, we introduce MolVision, a novel approach that leverages Vision-Language Models (VLMs) by integrating both molecular structure images and textual descriptions to enhance property prediction. We construct a benchmark spanning nine diverse datasets, covering both classification and regression tasks. Evaluating nine different VLMs in zero-shot, few-shot, and fine-tuned settings, we find that visual information improves prediction performance, particularly when combined with efficient fine-tuning strategies such as LoRA. Our results reveal that while visual information alone is insufficient, multimodal fusion significantly enhances generalization across molecular properties. Adaptation of vision encoder for molecular images in conjunction with LoRA further improves the performance. The code and data is available at : https://chemvision.github.io/chemvision/.



Paperid:1816
Authors:Victor Geadah, Amin Nejatbakhsh, David Lipshutz, Jonathan Pillow, Alex Williams
Abstract:
Neural population activity exhibits complex, nonlinear dynamics, varying in time, over trials, and across experimental conditions. Here, we developConditionally Linear Dynamical System(CLDS) models as a general-purpose method to characterize these dynamics. These models use Gaussian Process priors to capture the nonlinear dependence of circuit dynamics on task and behavioral variables. Conditioned on these covariates, the data is modeled with linear dynamics. This allows for transparent interpretation and tractable Bayesian inference. We find that CLDS models can perform well even in severely data-limited regimes (e.g. one trial per condition) due to their Bayesian formulation and ability to share statistical power across nearby task conditions. In example applications, we apply CLDS to model thalamic neurons that nonlinearly encode heading direction and to model motor cortical neurons during a cued-reaching task.



Paperid:1817
Authors:Esther Rodriguez, Monica Welfert, Samuel McDowell, Nathan Stromberg, Julian Camarena, Lalitha Sankar
Abstract:
Diffusion models have achieved impressive performance in generating high-quality and diverse synthetic data. However, their success typically assumes a class-balanced training distribution. In real-world settings, multi-class data often follow a long-tailed distribution, where standard diffusion models struggle—producing low-diversity and lower-quality samples for underrepresented (tail) classes. While this degradation is well-documented, its underlying cause remains poorly understood. In this work, we investigate the behavior of diffusion models trained on long-tailed datasets and identify a key issue: the latent representations (from the bottleneck layer of the U-Net) for tail class subspaces exhibit significant overlap with those of head classes, leading to feature borrowing and poor generation quality. Importantly, we show that this is not merely due to limited data per class, but that the relative class imbalance significantly contributes to this phenomenon. To address this, we proposeCOntrastiveRegularization forAligningLatents (CORAL), a contrastive latent alignment framework that leverages supervised contrastive losses to encourage well-separated latent class representations. Experiments demonstrate that CORAL significantly improves both the diversity and visual quality of samples generated for tail classes relative to state-of-the-art methods.



Paperid:1818
Authors:Minki Kang, Jongwon Jeong, Seanie Lee, Jaewoong Cho, Sung Ju Hwang
Abstract:
Large language models (LLMs) excel at complex reasoning tasks but remain computationally expensive, limiting their practical deployment.To address this, recent works have focused on distilling reasoning capabilities into smaller language models (sLMs) using chain-of-thought (CoT) traces from teacher LLMs.However, this approach struggles in scenarios requiring rare factual knowledge or precise computation, where sLMs often hallucinate due to limited capability.In this work, we propose Agent Distillation, a framework for transferring not only reasoning capability but full task-solving behavior from LLM-based agents into sLMs with retrieval and code tools. We improve agent distillation along two complementary axes: (1) we introduce a prompting method called first-thought prefix to enhance the quality of teacher-generated trajectories; and (2) we propose a self-consistent action generation for improving test-time robustness of small agents.We evaluate our method on eight reasoning tasks across factual and mathematical domains, covering both in-domain and out-of-domain generalization.Our results show that sLMs as small as 0.5B, 1.5B, 3B parameters can achieve performance competitive with next-tier larger 1.5B, 3B, 7B models fine-tuned using CoT distillation, demonstrating the potential of agent distillation for building practical, tool-using small agents.



Paperid:1819
Authors:Davin Hill, Brian Hill, Aria Masoomi, Vijay Nori, Robert Tillman, Jennifer Dy
Abstract:
Sequential deep learning models excel in domains with temporal or sequential dependencies, but their complexity necessitates post-hoc feature attribution methods for understanding their predictions. While existing techniques quantify feature importance, they inherently assume fixed feature ordering — conflating the effects of (1) feature values and (2) their positions within input sequences. To address this gap, we introduce OrdShap, a novel attribution method that disentangles these effects by quantifying how a model's predictions change in response to permuting feature position. We establish a game-theoretic connection between OrdShap and Sanchez-Bergantiños values, providing a theoretically grounded approach to position-sensitive attribution. Empirical results from health, natural language, and synthetic datasets highlight OrdShap's effectiveness in capturing feature value and feature position attributions, and provide deeper insight into model behavior.



Paperid:1820
Authors:Liming Liu, Zixuan Zhang, Simon Du, Tuo Zhao
Abstract:
Recent advances in deep learning optimization have unveiled two intriguing phenomena under large learning rates: Edge of Stability (EoS) and Progressive Sharpening (PS), challenging classical Gradient Descent (GD) analyses. Current research approaches, using either generalist frameworks or minimalist examples, face significant limitations in explaining these phenomena. This paper advances the minimalist approach by introducing a two-layer network with a two-dimensional input, where one dimension is relevant to the response and the other is irrelevant. Through this model, we rigorously prove the existence of progressive sharpening and self-stabilization under large learning rates, and establish non-asymptotic analysis of the training dynamics and sharpness along the entire GD trajectory. Besides, we connect our minimalist example to existing works by reconciling the existence of a well-behaved "stable set" between minimalist and generalist analyses, and extending the analysis of Gradient Flow Solution sharpness to our two-dimensional input scenario. These findings provide new insights into the EoS phenomenon from both parameter and input data distribution perspectives, potentially informing more effective optimization strategies in deep learning practice.



Paperid:1821
Authors:Seungwoo Kim, Khai Loong Aw, Klemen Kotar, Cristobal Eyzaguirre, Wanhee Lee, Yunong Liu, Jared Watrous, Stefan Stojanov, Juan Carlos Niebles, Jiajun Wu, Daniel Yamins
Abstract:
Extracting dense motion (optical flow) from videos remains a core computer-vision problem. Motivated by the recent success of large general-purpose models, we ask whether frozen self-supervised video world models trained only to predict future frames can be prompted, without fine-tuning, to output flow. Prior attempts to read out depth or illumination from video generators required fine-tuning; that strategy is ill-suited for flow, where labeled data are scarce and synthetic datasets suffer from a sim-to-real gap. We study several popular generative model architectures and find that successful zero-shot flow extraction requires three model properties: (1) distributional prediction of future frames (avoiding blurry or noisy outputs); (2) factorized latents that treat each spatio-temporal patch independently; and (3) random-access decoding that can condition on any subset of future pixels. These criteria are met by the recently introduced Local Random Access Sequence (LRAS) architecture. Building on LRAS, we propose KL-tracing: a procedure for injecting a small, local perturbation into the first frame, rolling out the model one step, and computing the Kullback–Leibler divergence between perturbed and unperturbed predictive distributions. The KL peak traces the displacement field, yielding optical flow in a single forward pass. Our method outperforms state-of-the-art models on real-world TAP-Vid DAVIS dataset (16.6% relative improvement for endpoint error) and synthetic TAP-Vid Kubric (4.7% relative improvement), despite being trained on real-world videos. Our results indicate that prompting controllable, self-supervised world models is a scalable and effective alternative to supervised or photometric-loss approaches for high-quality optical flow.



Paperid:1822
Authors:Mingyu Chen, Yiding Chen, Wen Sun, Xuezhou Zhang
Abstract:
Reinforcement Learning from Human Feedback (RLHF) has emerged as a pivotal technique for large language model (LLM) alignment.This paper studies the setting of online RLHF and focuses on improving its sample efficiency.All existing algorithms for online RLHF, whether doing passive exploration or active exploration, suffer from a sample complexity that scales exponentially with the range of the reward function.This statistical inefficiency hinders their effectiveness in scenarios with heavily skewed preferences, e.g. questions with objectively correct answers.To address this, we introduce Self-Exploring Preference-Incentive Online Preference Optimization (SE-POPO), an online RLHF algorithm that for the first time achieves a sample complexity that scales polynomially with the reward range, answering an open problem raised by Xie et al. [2024].Theoretically, we demonstrate that the sample complexity of SE-POPO dominates that of existing exploration algorithms. Empirically, our systematic evaluation confirms that SE-POPO is more sample-efficient than both exploratory and non-exploratory baselines, in two primary application scenarios of RLHF as well as on public benchmarks, marking a significant step forward in RLHF algorithm design.



Paperid:1823
Authors:Bardienus Duisterhof, Jan Oberst, Bowen Wen, Stan Birchfield, Deva Ramanan, Jeffrey Ichnowski
Abstract:
3D shape completion has broad applications in robotics, digital twin reconstruction, and extended reality (XR). Although recent advances in 3D object and scene completion have achieved impressive results, existing methods lack 3D consistency, are computationally expensive, and struggle to capture sharp object boundaries. Our work (RaySt3R) addresses these limitations by recasting 3D shape completion as a novel view synthesis problem. Specifically, given a single RGB-D image, and a novel viewpoint (encoded as a collection of query rays),we train a feedforward transformer to predict depth maps, object masks, and per-pixel confidence scores for those query rays. RaySt3R fuses these predictions across multiple query views to reconstruct complete 3D shapes. We evaluate RaySt3R on synthetic and real-world datasets, and observe it achieves state-of-the-art performance,outperforming the baselines on all datasets by up to 44% in 3D chamfer distance.



Paperid:1824
Authors:Yichong Lu, Yuzhuo Tian, Zijin Jiang, Yikun Zhao, Yuanbo Yang, Hao Ouyang, Haoji Hu, Huimin Yu, Yujun Shen, Yiyi Liao
Abstract:
Humans intuitively perceive object shape and orientation from a single image, guided by strong priors about canonical poses. However, existing 3D generative models often produce misaligned results due to inconsistent training data, limiting their usability in downstream tasks. To address this gap, we introduce the task of orientation-aligned 3D object generation: producing 3D objects from single images with consistent orientations across categories. To facilitate this, we construct Objaverse-OA, a dataset of 14,832 orientation-aligned 3D models spanning 1,008 categories. Leveraging Objaverse-OA, we fine-tune two representative 3D generative models based on multi-view diffusion and 3D variational autoencoder frameworks to produce aligned objects that generalize well to unseen objects across various categories. Experimental results demonstrate the superiority of our method over post-hoc alignment approaches. Furthermore, we showcase downstream applications enabled by our aligned object generation, including zero-shot object orientation estimation via analysis-by-synthesis and efficient arrow-based object insertion.



Paperid:1825
Authors:Ruoyu Wang, Beier Zhu, Junzhi Li, Liangyu Yuan, Chi Zhang
Abstract:
Diffusion-based generative processes, grounded in differential equation solving, frequently require striking a balance between computational speed and output quality. Our theoretical investigation of prevalent solving approaches - ordinary differential equations (ODE) and stochastic differential equations (SDE) solvers - uncovers distinct limitations: ODE solvers exhibit irreducible gradient error accumulation from deterministic path dependence, while SDE methods suffer amplified discretization errors when step counts are reduced. Building upon this insight, we introduce AdaSDE, a novel single-step SDE solver that aims to unify the efficiency of ODEs with the error resilience of SDEs. At the core of our design is a learnable parameter obtained through lightweight tuning, which dynamically regulates the error correction strength to accelerate diffusion sampling. Notably, our framework can be integrated with numerous solvers to enhance their capabilities through lightweight parameter tuning. Extensive experiments demonstrate state-of-the-art performance: At 5 NFE, AdaSDE achieves FID scores of 4.79 on CIFAR-10, 8.91 on FFHQ 64×64 and 6.96 on LSUN Bedroom.



Paperid:1826
Authors:Xiangning Yu, Zhuohan Wang, Linyi Yang, Haoxuan Li, Anjie Liu, Xiao Xue, Jun Wang, Mengyue Yang
Abstract:
Chain-of-Thought (CoT) prompting plays an indispensable role in endowing large language models (LLMs) with complex reasoning capabilities. However, CoT currently faces two fundamental challenges: (1) Sufficiency, which ensures that the generated intermediate inference steps comprehensively cover and substantiate the final conclusion; and (2) Necessity, which identifies the inference steps that are truly indispensable for the soundness of the resulting answer. We propose a causal framework that characterizes CoT reasoning through the dual lenses of sufficiency and necessity. Incorporating causal Probability of Sufficiency and Necessity allows us not only to determine which steps are logically sufficient or necessary to the prediction outcome, but also to quantify their actual influence on the final reasoning outcome under different intervention scenarios, thereby enabling the automated addition of missing steps and the pruning of redundant ones. Extensive experimental results on various mathematical and commonsense reasoning benchmarks confirm substantial improvements in reasoning efficiency and reduced token usage without sacrificing accuracy. Our work provides a promising direction for improving LLM reasoning performance and cost-effectiveness. The code will be publicly available upon acceptance at: https://anonymous.4open.science/r/causalmath-1CEF.



Paperid:1827
Authors:Chris Cundy, Adam Gleave
Abstract:
As AI systems become more capable, deceptive behaviors can undermine evaluation and mislead users at deployment.Recent work has shown that lie detectors can accurately classify deceptive behavior, but they are not typically used in the training pipeline due to concerns around contamination and objective hacking.We examine these concerns by incorporating a lie detector into the labelling step of LLM post-training and evaluating whether the learned policy is genuinely more honest, or instead learns to fool the lie detector while remaining deceptive.Using DolusChat, a novel 65k-example dataset with paired truthful/deceptive responses, we identify three key factors that determine the honesty of learned policies: amount of exploration during preference learning, lie detector accuracy, and KL regularization strength. We find that preference learning with lie detectors and GRPO can lead to policies which evade lie detectors, with deception rates of over 85\%. However, if the lie detector true positive rate (TPR) or KL regularization is sufficiently high, GRPO learns honest policies.In contrast, off-policy algorithms (DPO) consistently lead to deception rates under 25\% for realistic TPRs.Our results illustrate a more complex picture than previously assumed: depending on the context, lie-detector-enhanced training can be a powerful tool for scalable oversight, or a counterproductive method encouraging undetectable misalignment.



Paperid:1828
Authors:Jonathan Schmidt, Simon Giebenhain, Matthias Niessner
Abstract:
We introduceBecomingLit, a novel method for reconstructing relightable, high-resolution head avatars that can be rendered from novel viewpoints at interactive rates. Therefore, we propose a new low-cost light stage capture setup, tailored specifically towards capturing faces. Using this setup, we collect a novel dataset consisting of diverse multi-view sequences of numerous subjects under varying illumination conditions and facial expressions. By leveraging our new dataset, we introduce a new relightable avatar representation based on 3D Gaussian primitives that we animate with a parametric head model and an expression-dependent dynamics module. We propose a new hybrid neural shading approach, combining a neural diffuse BRDF with an analytical specular term. Our method reconstructs disentangled materials from our dynamic light stage recordings and enables all-frequency relighting of our avatars with both point lights and environment maps. In addition, our avatars can easily be animated and controlled from monocular videos. We validate our approach in extensive experiments on our dataset, where we consistently outperform existing state-of-the-art methods in relighting and reenactment by a significant margin.



Paperid:1829
Authors:Kyriakos Lotidis, Panayotis Mertikopoulos, Nicholas Bambos, Jose Blanchet
Abstract:
In this paper, we examine the convergence landscape of multi-agent learning in continuous games under uncertainty.In more detail, we consider two stochastic models of regularized learning—in continuous and discrete time respectively—and we set out to characterize the long-run behavior of the induced sequence of play.In stark contrast to deterministic, full-information models of learning (or models with a vanishing learning rate), we show that, in general, the resulting dynamics do not converge.In lieu of this, we ask instead which actions are played more often in the long run, and by how much.To that end, we show that, despite wandering away from equilibrium infinitely often, the dynamics of regularized learning in strongly monotone games always return to its vicinity in finite time (which we estimate), and their long-run distribution is sharply concentrated around a neighborhood thereof.We quantify the degree of this concentration, and we show that these favorable properties may all break down if the underlying game is not strongly monotone—underscoring in this way the limits of regularized learning in the presence of persistent randomness and uncertainty.



Paperid:1830
Authors:Bolin Wang, Tong Wei, Jiang-Xin Shi, Yu-Feng Li, Min-Ling Zhang
Abstract:
Recognizing out-of-distribution (OOD) samples is essential for deploying robust machine learning systems in the open-world environments. Conventional OOD detection approaches rely on feature representations from the final layer of neuron networks, often neglecting the rich information encapsulated in shallow layers. Leveraging the strengths of transformer-based architectures, we introduce an adaptive fusion module, which dynamically assigns importance weights to representations learned by each Transformer layer and detects OOD samples using the Mahalanobis distance. Compared to existing approaches, our method enables a lightweight fine-tuning of pre-trained models, and retains all feature representations that are beneficial to the OOD detection. We also thoroughly study various parameter-efficient fine-tuning strategies. Our experiments show the benefit of using shallow features, and demonstrate the influence of different Transformer layers. We fine-tune pre-trained models in both class-balanced and long-tailed in-distribution classification tasks, and show that our method achieves state-of-the-art OOD detection performance averaged across nine OOD datasets. The source code is provided in the supplementary material.



Paperid:1831
Authors:Xinran Han, Ko Nishino, Todd Zickler
Abstract:
Perceiving the shape and material of an object from a single image is inherently ambiguous, especially when lighting is unknown and unconstrained. Despite this, humans can often disentangle shape and material, and when they are uncertain, they often move their head slightly or rotate the object to help resolve the ambiguities. Inspired by this behavior, we introduce a novel conditional denoising-diffusion model that generates samples of shape-and-material maps from a given short input video of an object undergoing differential motions. Our parameter-efficient architecture allows training directly in pixel-space and joint, generative disentanglement of multiple object attributes simultaneously. Trained on a modest number of synthetic object-motion videos with supervision on shape and material, the model exhibits compelling emergent properties: for static observations, it produces diverse, multimodal predictions of plausible shape-and-material maps that capture the inherent ambiguities; and when objects move, the distributions quickly converge to more accurate explanations. Meanwhile, it produces high-quality shape-and-material estimates on less ambiguous, real-world objects. By moving beyond single view to continuous observations, our work suggests an avenue of generative perception for improving visual reasoning for physically-embodied systems.



Paperid:1832
Authors:Fan LIU, Zherui Yang, Cancheng Liu, Tianrui Song, Xiaofeng Gao, Hao Liu
Abstract:
Mathematical modeling is a cornerstone of scientific discovery and engineering practice, enabling the translation of real-world problems into formal systems across domains such as physics, biology, and economics. Unlike mathematical reasoning, which assumes a predefined formulation, modeling requires open-ended problem analysis, abstraction, and principled formalization. While Large Language Models (LLMs) have shown strong reasoning capabilities, they fall short in rigorous model construction, limiting their utility in real-world problem-solving. To this end, we formalize the task of LLM-powered real-world mathematical modeling, where agents must analyze problems, construct domain-appropriate formulations, and generate complete end-to-end solutions.We introduce MM-Bench, a curated benchmark of 111 problems from the Mathematical Contest in Modeling (MCM/ICM), spanning the years 2000 to 2025 and across ten diverse domains such as physics, biology, and economics. To tackle this task, we propose MM-Agent, an expert-inspired framework that decomposes mathematical modeling into four stages: open-ended problem analysis, structured model formulation, computational problem solving, and report generation.Experiments on MM-Bench show that MM-Agent significantly outperforms baseline agents, achieving an 11.88\% improvement over human expert solutions while requiring only 15 minutes and \$0.88 per task using GPT-4o. Furthermore, under official MCM/ICM protocols, MM-Agent assisted two undergraduate teams in winning the Finalist Award (\textbf{top 2.0\% among 27,456 teams}) in MCM/ICM 2025, demonstrating its practical effectiveness as a modeling copilot.



Paperid:1833
Authors:Yeonjun In, Kanghoon Yoon, Sukwon Yun, Kibum Kim, Sungchul Kim, Chanyoung Park
Abstract:
In real-world applications, node features in graphs often contain noise from various sources, leading to significant performance degradation in GNNs. Although several methods have been developed to enhance robustness, they rely on the unrealistic assumption that noise in node features is independent of the graph structure and node labels, thereby limiting their applicability. To this end, we introduce a more realistic noise scenario, dependency-aware noise on graphs (DANG), where noise in node features create a chain of noise dependencies that propagates to the graph structure and node labels. We propose a novel robust GNN, DA-GNN, which captures the causal relationships among variables in the data generating process (DGP) of DANG using variational inference. In addition, we present new benchmark datasets that simulate DANG in real-world applications, enabling more practical research on robust GNNs. Extensive experiments demonstrate that DA-GNN consistently outperforms existing baselines across various noise scenarios, including both DANG and conventional noise models commonly considered in this field.



Paperid:1834
Authors:Haoran Xu, Liyuan Mao, Hui Jin, Weinan Zhang, Xianyuan Zhan, Amy Zhang
Abstract:
The practical use of reinforcement learning (RL) requires handling diverse settings, including online, offline, and offline-to-online learning. Instead of developing separate algorithms for each setting, we propose Uni-IVR, a unified model-free RL framework that addresses all these scenarios within a single formulation. Uni-IVR builds on the Implicit Value Regularization (IVR) framework and generalizes its dataset behavior constraint to the constraint w.r.t a reference policy, yielding an unified value learning objective for general settings. The reference policy is chosen to be the target policy in the online setting and the behavior policy in the offline setting. Using an iteratively refined behavior policy solves the over-constrained problem of directly applying IVR in the online setting, it provides an implicit trust-region style update through the value function while being off-policy. Uni-IVR also introduces an unified policy extraction objective that estimates in-sample policy gradient using only actions from the reference policy. This supports various policy classes and theoretically guaranntees less value estimation error and larger performance improvement over the reference policy. We evaluate Uni-IVR on a range of standard RL benchmarks across online, offline, and offline-to-online settings. In online RL, Uni-IVR achieves higher sample efficiency than both off-policy methods without trust-region updates and on-policy methods with trust-region updates. In offline RL, Uni-IVR retains the benefits of in-sample learning while outperforming IVR through better policy extraction. In offline-to-online RL, Uni-IVR beats both constraint-based methods and unconstrained approaches by effectively balancing stability and adaptability.



Paperid:1835
Authors:Shijun Liang, Ismail Alkhouri, Siddhant Gautam, Qing Qu, Saiprasad Ravishankar
Abstract:
Abstract:Recent advances in data-centric deep generative models have led to significant progress in solving inverse imaging problems. However, these models (e.g., diffusion models (DMs)) typically require large amounts of fully sampled (clean) training data, which is often impractical in medical and scientific settings such as dynamic imaging. On the other hand, training-data-free approaches like the Deep Image Prior (DIP) do not require clean ground-truth images but suffer from noise overfitting and can be computationally expensive as the network parameters need to be optimized for each measurement set independently. Moreover, DIP-based methods often overlook the potential of learning a prior using a small number of sub-sampled measurements (or degraded images) available during training. In this paper, we propose **UGoDIT**—an **U**nsupervised **G**r**o**up **DI**P with **T**ransferable weights—designed for the low-data regime where only a very small number, $M$, of sub-sampled measurement vectors are available during training. Our method learns a set of transferable weights by optimizing a shared encoder and $M$ disentangled decoders. At test time, we reconstruct the unseen degraded image using a DIP network, where part of the parameters are fixed to the learned weights, while the remaining are optimized to enforce measurement consistency. We evaluate \our on both medical (multi-coil MRI) and natural (super resolution and non-linear deblurring) image recovery tasks under various settings. Compared to recent standalone DIP methods, \our provides accelerated convergence and notable improvement in reconstruction quality. Furthermore, our method achieves performance competitive with SOTA DM-based and supervised approaches, despite not requiring large amounts of clean training data.



Paperid:1836
Authors:Shaoang Li, Jian Li
Abstract:
Abstract:Non-stationary multi-armed bandits (nsMAB) enable agents to adapt to changing environments by incorporating mechanisms to detect and respond to shifts in reward distributions, making them well-suited for dynamic settings. However, existing approaches typically assume that reward feedback is available at every round—an assumption that overlooks many real-world scenarios where feedback is limited. In this paper, we take a significant step forward by introducing a new model of *constrained feedback in non-stationary multi-armed bandits* (ConFee-nsMAB), where the availability of reward feedback is restricted. We propose the first prior-free algorithm—that is, one that does not require prior knowledge of the degree of non-stationarity—that achieves near-optimal dynamic regret in this setting. Specifically, our algorithm attains a dynamic regret of $\tilde {\mathcal{O}}({K^{1/3} V_T^{1/3} T }/{ B^{1/3}})$, where $T$ is the number of rounds, $K$ is the number of arms, $B$ is the query budget, and $V_T$ is the variation budget capturing the degree of non-stationarity.



Paperid:1837
Authors:Jiabei Cheng, Changxi Chi, Jingbo Zhou, Hongyi Xin, Jun Xia
Abstract:
Reliable prediction of single-cell responses is crucial for both high-throughput screening and advancing biological understanding. Recent in-silico models have made much progress in prediction accuracy, but uncertainty metrics are rarely reported. When scientists apply these models to study specific perturbations, knowing how trustworthy individual predictions are is more practical than averaged accuracy.Therefore, we introduce \textbf{PRESCRIBE} (Predicting Single-Cell Responses with Bayesian Estimation), a novel model that not only predicts transcriptomic outcome but provides uncertainty estimates for each perturbation. PRESCRIBE employs the Natural Posterior Network to quantify uncertainty by modeling the energy distance between perturbed and control cellular states, while accounting for generalization difficulty. Our analysis demonstrates that PRESCRIBE effectively estimates prediction confidence, correlates well with its prediction accuracy, and enables the filtering of untrustworthy results. In practical applications, PRESCRIBE steadily achieves accuracy improvements of over 3\% against comparable baselines.



Paperid:1838
Authors:Yuezhou Ma, Haixu Wu, Hang Zhou, Huikun Weng, Jianmin Wang, Mingsheng Long
Abstract:
Physics sensing plays a central role in many scientific and engineering domains, which inherently involves two coupled tasks: reconstructing dense physical fields from sparse observations and optimizing scattered sensor placements to observe maximum information. While deep learning has made rapid advances in sparse-data reconstruction, existing methods generally omit optimization of sensor placements, leaving the mutual enhancement between reconstruction and placement on the shelf. To change this suboptimal practice, we propose PhySense, a synergistic two-stage framework that learns to jointly reconstruct physical fields and to optimize sensor placements, both aiming for accurate physics sensing. The first stage involves a flow-based generative model enhanced by cross-attention to adaptively fuse sparse observations. Leveraging the reconstruction feedback, the second stage performs sensor placement via projected gradient descent to satisfy spatial constraints. We further prove that the learning objectives of the two stages are consistent with classical variance-minimization principles, providing theoretical guarantees. Extensive experiments across three challenging benchmarks, especially a 3D geometry dataset, indicate PhySense achieves state-of-the-art physics sensing accuracy and discovers informative sensor placements previously unconsidered.



Paperid:1839
Authors:Dongwei Jiang, Bowei Zhang, Andrew Wang, Nicholas Andrews, Daniel Khashabi
Abstract:
Recent studies have shown LLMs possess \textit{some} ability to improve their responses when given external feedback.However, it remains unclear how effectively and thoroughly these models can incorporate extrinsic feedback. In an ideal scenario, if LLMs receive near-perfect and complete feedback, we would expect them to \emph{fully} integrate the feedback and change their incorrect answers to correct ones.In this paper, we systematically investigate LLMs' ability to incorporate feedback by designing a controlled experimental environment. For each problem, a solver model attempts a solution, then a feedback generator with access to \textit{near-complete} ground-truth answers produces targeted feedback, after which the solver tries again. We evaluate this pipeline across a diverse range of tasks, including mathematical reasoning, knowledge reasoning, complex scientific reasoning, and general multi-domain evaluations with state-of-the-art language models as solver and feedback generator. Surprisingly, even under these near-ideal conditions, solver models consistently show resistance to feedback, a limitation that we term RIGID THINKING. To mitigate this limitation, we experiment with sampling-based strategies like progressive temperature increases and explicit rejection of previously attempted incorrect answers, which yield improvements but still fail to help models achieve target performance. We perform a rigorous exploration of potential causes of RIGID THINKING, ruling out factors such as model overconfidence and data familiarity.



Paperid:1840
Authors:Kele Shao, Keda TAO, Can Qin, Haoxuan You, Yang Sui, Huan Wang
Abstract:
Video large language models (video LLMs) excel at video comprehension but face significant computational inefficiency due to redundant video tokens. Existing token pruning methods offer solutions. However, approaches operating within the LLM (inner-LLM pruning), such as FastV, incur intrinsic computational overhead in shallow layers. In contrast, methods performing token pruning before the LLM (outer-LLM pruning) primarily address spatial redundancy within individual frames or limited temporal windows, neglecting the crucial global temporal dynamics and correlations across longer video sequences. This leads to sub-optimal spatio-temporal reduction and does not leverage video compressibility fully. Crucially, the synergistic potential and mutual influence of combining these strategies remain unexplored. To further reduce redundancy, we introduce HoliTom, a novel training-free holistic token merging framework. HoliTom employs outer-LLM pruning through global redundancy-aware temporal segmentation, followed by spatial-temporal merging to reduce visual tokens by over 90%, significantly alleviating the LLM's computational burden. Complementing this, we introduce a robust inner-LLM token similarity-based merging approach, designed for superior performance and compatibility with outer-LLM pruning. Evaluations demonstrate our method's promising efficiency-performance trade-off on LLaVA-OneVision-7B, reducing computational costs to 6.9% of FLOPs while maintaining 99.1% of the original performance. Furthermore, we achieve a 2.28× reduction in Time-To-First-Token (TTFT) and a 1.32× acceleration in decoding throughput, highlighting the practical benefits of our integrated pruning approach for efficient video LLMs inference.



Paperid:1841
Authors:Komal Kumar, Rao Anwer, Fahad Shahbaz Khan, Salman Khan, Ivan Laptev, Hisham Cholakkal
Abstract:
Efficient fine-tuning of pre-trained Text-to-Image (T2I) models involves adjusting the model to suit a particular task or dataset while minimizing computational resources and limiting the number of trainable parameters. However, it often faces challenges in striking a trade-off between aligning with the target distribution: learning a novel concept from a limited image for personalization and retaining the flexibility needed for unifying multiple tasks, all while maintaining editability (aligning with a variety of prompts or in-context generation). In this work, we introduce DEFT, Decompositional Efficient Fine-Tuning, an efficient fine-tuning framework that adapts a pre-trained weight matrix by decomposing its update into two components with two trainable matrices: (1) a projection onto the complement of a low-rank subspace spanned by a low-rank matrix, and (2) a low-rank update. The single trainable low-rank matrix defines the subspace, while the other trainable low-rank matrix enables flexible parameter adaptation within that subspace. We conducted extensive experiments on the Dreambooth and Dreambench Plus datasets for personalization, the InsDet dataset for object and scene adaptation, and the VisualCloze dataset for a universal image generation framework through visual in-context learning with both Stable Diffusion and a unified model. Our results demonstrated state-of-the-art performance, highlighting the emergent properties of efficient fine-tuning. Our code will be publicly available on GitHub.



Paperid:1842
Authors:Yongliang Wu, Zonghui Li, Xinting Hu, Xinyu Ye, Xianfang Zeng, Gang Yu, Wenbo Zhu, Bernt Schiele, Ming-Hsuan Yang, Xu Yang
Abstract:
Recent advances in multi-modal generative models have enabled significant progress in instruction-based image editing. However, while these models produce visually plausible outputs, their capacity for knowledge-based reasoning editing tasks remains under-explored. In this paper, We introduce KRIS-Bench (Knowledge-based Reasoning in Image-editing Systems Benchmark), a diagnostic benchmark designed to assess models through a cognitively informed lens. Drawing from educational theory, KRIS-Bench categorizes editing tasks across three foundational knowledge types: Factual, Conceptual, and Procedural. Based on this taxonomy, we design 22 representative tasks spanning 7 reasoning dimensions and release 1,267 high-quality annotated editing instances. To support fine-grained evaluation, we propose a comprehensive protocol that incorporates a novel Knowledge Plausibility metric, enhanced by knowledge hints and calibrated through human studies. Empirical results on nine state-of-the-art models reveal significant gaps in reasoning performance, highlighting the need for knowledge-centric benchmarks to advance the development of intelligent image editing systems.



Paperid:1843
Authors:Nolan Dey, Bin Zhang, Lorenzo Noci, Mufan Li, Blake Bordelon, Shane Bergsma, Cengiz Pehlevan, Boris Hanin, Joel Hestness
Abstract:
We study compute efficiency of LLM training when using different parameterizations, i.e., rules for adjusting model and optimizer hyperparameters (HPs) as model size changes. Some parameterizations fail to transfer optimal base HPs (such as learning rate) across changes in model depth, requiring practitioners to either re-tune these HPs as they scale up (expensive), or accept sub-optimal training when re-tuning is prohibitive. Even when they achieve HP transfer, we develop theory to show parameterizations may still exist in the lazy learning regime where layers learn only features close to their linearization, preventing effective use of depth and nonlinearity. Finally, we identify and adopt the parameterization we call CompleteP that achieves both depth-wise HP transfer and non-lazy learning in all layers. CompleteP enables a wider range of model width/depth ratios to remain compute-efficient, unlocking shapes better suited for different hardware settings and operational contexts. Moreover, CompleteP enables 12-34% compute efficiency improvements over the prior state-of-the-art.



Paperid:1844
Authors:Yi Hu, Shijia Kang, Haotong Yang, Haotian Xu, Muhan Zhang
Abstract:
Length generalization—the ability to solve problems longer than those seen during training—remains a critical challenge for large language models (LLMs). Previous work modifies positional encodings (PEs) and data formats to improve length generalization on specific symbolic tasks such as addition and sorting. However, these approaches are fundamentally limited to special tasks, often degrading general language performance. Furthermore, they are typically evaluated on small transformers trained from scratch on single tasks and can cause performance drop when applied during post-training stage of practical LLMs with general capabilities. Hu et al., (2024) proposed Rule-Following Fine-Tuning (RFFT) to improve length generalization in the post-training stage of LLMs. Despite its compatibility with practical models and strong performance, RFFT is proposed for single tasks too, requiring re-training for each individual task with extensive examples. In this paper, we study length generalization in multi-task settings and proposeMeta Rule-Following Fine-Tuning (Meta-RFFT), the first framework enabling robustcross-tasklength generalization. As our first contribution, we construct a large length generalization dataset containing86 tasksspanning code execution, number processing, symbolic and logical reasoning tasks, beyond the common addition or multiplication tasks. Secondly, we show that cross-task length generalization is possible with Meta-RFFT—after training on a large number of tasks and instances, the models achieve remarkable length generalization ability onunseentasks withminimal fine-tuning or one-shot prompting. For example, after fine-tuning on 1 to 5 digit addition, our 32B modelachieves 95% accuracy on 30 digit addition, significantly outperforming the state-of-the-art reasoning models (DeepSeek-R1-671B: 72%; QwQ-32B: 32%), despite never seeing this task during RF-pretraining.



Paperid:1845
Authors:Brij Malhotra, Shivvrat Arya, Tahrima Rahman, Vibhav Gogate
Abstract:
We introduce learning to condition (L2C), a scalable, data-driven framework for accelerating Most Probable Explanation (MPE) inference in Probabilistic Graphical Models (PGMs)—a fundamentally intractable problem. L2C trains a neural network to score variable-value assignments based on their utility for conditioning, given observed evidence. To facilitate supervised learning, we develop a scalable data generation pipeline that extracts training signals from the search traces of existing MPE solvers. The trained network serves as a heuristic that integrates with search algorithms, acting as a conditioning strategy prior to exact inference or as a branching and node selection policy within branch-and-bound solvers. We evaluate L2C on challenging MPE queries involving high-treewidth PGMs. Experiments show that our learned heuristic significantly reduces the search space while maintaining or improving solution quality over state-of-the-art methods.



Paperid:1846
Authors:Jörg Franke, Urs Spiegelhalter, Marianna Nezhurina, Jenia Jitsev, Frank Hutter, Michael Hefenbrock
Abstract:
The successful training of deep neural networks requires addressing challenges such as overfitting, numerical instabilities leading to divergence, or the increasing variance in the residual stream. A common solution is applying regularization and normalization techniques that usually involve tuning of additional hyperparameters. An alternative is to force all parameters and representations to lie on a hypersphere. This removes the need for regularization and increases convergence speed, but comes with additional costs. In this work, we propose a more holistic, approximate normalization via simple scalar multiplications motivated by the tight concentration of the norms of high-dimensional random vectors. Additionally, instead of applying strict normalization for the parameters, we constrain their norm. These modifications remove the need for weight decay and learning rate warm-up as well, but do not increase the total number of normalization layers. Our experiments with transformer architectures show up to 40% faster convergence compared to GPT models with QK normalization, at only 3% additional runtime cost. When deriving scaling laws, we found that our method enables training with larger batch sizes while preserving the favorable scaling characteristics of classic GPT architectures.



Paperid:1847
Authors:Shiva Sreeram, Alaa Maalouf, Pratyusha Sharma, Daniela Rus
Abstract:
Recently, Sharma et al. (2024) suggested a method called LAyer- SElective-Rank reduction (LASER) which demonstrated that pruning high‑order components of carefully chosen LLM’s weight matrices can boost downstream accuracy—without any gradient‑based fine‑tuning. Yet LASER’s exhaustive, per‑matrix search (each requiring full‑dataset forward passes) makes it impractical for rapid deployment. We demonstrate that this overhead can be removed and find that: (i) Only a small, carefully chosen subset of matrices needs to be inspected—eliminating the layer‑by‑layer sweep, (ii) The gradient of each matrix’s singular values pinpoints which matrices merit reduction, (iii) Increasing the factorization search space by allowing matrices rows to cluster around multiple subspaces and then decomposing each cluster separately further reduces overfitting on the original training data and further lifts accuracy by up to 24.6 percentage points, and finally, (iv) we discover that evaluating on just 100 samples rather than the full training data—both for computing the indicative gradients and for measuring the final accuracy—suffices to further reduce the search time; we explain that as adaptation to downstream tasks is dominated by prompting style, not dataset size. As a results, we show that combining these findings yields a fast and robust adaptation algorithm for downstream tasks. Overall, with a single gradient step on 100 examples and a quick scan of the top candidate layers and factorization techniques, we can adapt LLMs to new datasets—entirely without fine‑tuning.



Paperid:1848
Authors:Pengyun Yue, Xuanlin Yang, Mingqing Xiao, Zhouchen Lin
Abstract:
Abstract:Zeroth-order Optimization (ZO) has received wide attention in machine learning, especially when computing full gradient is expensive or even impossible. Recently, ZO has emerged as an important paradigm for memory-efficient fine-tuning of large language models (LLMs), circumventing the memory overhead of backpropagation. However, existing ZO gradient estimators exhibit dimension-dependent variance scaling as $\Theta(d)$, leading to dimension-dependent convergence rates which is prohibitive for large-scale LLM parameters. To address this problem, we present a Pseudo-Zeroth-Order (PZO) framework for optimizing composite objective functions, especially large-scale models: $\min_{\mathbf{x}\in \mathcal{X}} \mathbb{E}_\mathbf{z} g\circ h(\mathbf{x};\mathbf{z})$, where $h$ represents complex, high-dimensional representations and $g$ is a task-specific loss. While existing zeroth-order methods estimate gradients with final loss functions, our PZO algorithm estimate the Jacobian matrix of $h(\mathbf{x})$ with the model output $\mathbf{o}= h(\mathbf{x})$, and the gradient of the loss function on model output $\mathbf{e} = \nabla_{\mathbf{o}} g(\mathbf{o})$. Moreover, we apply exponential moving average on Jacobian estimators to reduce the variance. Experimental results demonstrate that PZO outperforms MeZO and MeZO-SVRG in classification, multiple choice and generation tasks in both full-parameter and PEFT fine-tuning settings by boosting convergence in the early stages of training. With the sliding window technique, our PZO only introduced a small dimension-independent memory overhead, which enables efficient scaling of the model size.



Paperid:1849
Authors:Mehran Shakerinava, Siamak Ravanbakhsh, Adam Oberman
Abstract:
Abstract:Recent work has formalized the reward hypothesis through the lens of expected utility theory, by interpreting reward as utility. Hausner's foundational work showed that dropping the continuity axiom leads to a generalization of expected utility theory where utilities are *lexicographically* ordered *vectors* of arbitrary dimension. In this paper, we extend this result by identifying a simple and practical condition under which preferences cannot be represented by scalar rewards, necessitating a 2-dimensional reward function. We provide a full characterization of such reward functions, as well as the general $d$-dimensional case, in Markov Decision Processes (MDPs) under a memorylessness assumption on preferences. Furthermore, we show that optimal policies in this setting retain many desirable properties of their scalar-reward counterparts, while in the Constrained MDP (CMDP) setting -- another common multiobjective setting -- they do not.



Paperid:1850
Authors:Byeonghu Na, Minsang Park, Gyuwon Sim, Donghyeok Shin, HeeSun Bae, Mina Kang, Se Jung Kwon, Wanmo Kang, Il-chul Moon
Abstract:
Text-to-image diffusion models rely on text embeddings from a pre-trained text encoder, but these embeddings remain fixed across all diffusion timesteps, limiting their adaptability to the generative process. We propose Diffusion Adaptive Text Embedding (DATE), which dynamically updates text embeddings at each diffusion timestep based on intermediate perturbed data. We formulate an optimization problem and derive an update rule that refines the text embedding at each sampling step to improve alignment and preference between the mean predicted image and the text. This allows DATE to dynamically adapts the text conditions to the reverse-diffused images throughout diffusion sampling without requiring additional model training. Through theoretical analysis and empirical results, we show that DATE maitains the generative capability of the model while providing superior text-image alignment over fixed text embeddings across various tasks, including multi-concept generation and text-guided image editing.



Paperid:1851
Authors:Jasraj Singh, Keyue Jiang, Brooks Paige, Laura Toni
Abstract:
Message Passing Neural Networks (MPNNs) are a class of Graph Neural Networks (GNNs) that propagate information across the graph via local neighborhoods. The scheme gives rise to two key challenges: over-smoothing and over-squashing. While several Dropout-style algorithms, such as DropEdge and DropMessage, have successfully addressed over-smoothing, their impact on over-squashing remains largely unexplored. This represents a critical gap in the literature, as failure to mitigate over-squashing would make these methods unsuitable for long-range tasks – the intended use case of deep MPNNs. In this work, we study the aforementioned algorithms, and closely related edge-dropping algorithms – DropNode, DropAgg and DropGNN – in the context of over-squashing. We present theoretical results showing that DropEdge-variants reduce sensitivity between distant nodes, limiting their suitability for long-range tasks. To address this, we introduce DropSens, a sensitivity-aware variant of DropEdge that explicitly controls the proportion of information lost due to edge-dropping, thereby increasing sensitivity to distant nodes despite dropping the same number of edges. Our experiments on long-range synthetic and real-world datasets confirm the predicted limitations of existing edge-dropping and feature dropping methods. Moreover, DropSens consistently outperforms graph rewiring techniques designed to mitigate over-squashing, suggesting that simple, targeted modifications can substantially improve a model’s ability to capture long-range interactions. Our conclusions highlight the need to re-evaluate and re-design existing methods for training deep GNNs, with a renewed focus on modelling long-range interactions.



Paperid:1852
Authors:Max Torop, Aria Masoomi, Masih Eskandar, Jennifer Dy
Abstract:
Abstract:A variety of recent methods guide large language model outputs via the inference-time addition of *steering vectors* to residual-stream or attention-head representations. In contrast, we propose to inject steering vectors directly into the query and value representation spaces within attention heads. We provide evidence that a greater portion of these spaces exhibit high linear discriminability of concepts --a key property motivating the use of steering vectors-- than attention head outputs. We analytically characterize the effect of our method, which we term *DISentangled COmmunication (DISCO) Steering*, on attention head outputs. Our analysis reveals that DISCO disentangles a strong but underutilized baseline, steering attention inputs, which implicitly modifies queries and values in a rigid manner. In contrast, DISCO's direct modulation of these components enables more granular control. We find that DISCO achieves superior performance over a number of steering vector baselines across multiple datasets on LLaMA 3.1 8B and Gemma 2 9B, with steering efficacy scoring up to $19.1$% higher than the runner-up. Our results support the conclusion that the query and value spaces are powerful building blocks for steering vector methods. We will make our code publicly available upon acceptance.



Paperid:1853
Authors:Azim Ospanov, Farzan Farnia, Roozbeh Yousefzadeh
Abstract:
We perform a thorough analysis of the formal and informal statements in the miniF2F benchmark from the perspective of an AI system that is tasked to participate in a math Olympiad consisting of the problems in miniF2F. In such setting, the model has to read and comprehend the problems in natural language, formalize them in Lean language, then proceed with proving the problems, and it will get credit for each problem if the formal proof corresponds to the original informal statement presented to the model. Our evaluation results reveal that the best accuracy of such pipeline can be about 36\% using the SoTA models in the literature, considerably lower than the individual SoTA accuracies, 97\% and 69\% reported in the autoformalization and theorem proving literature. Analyzing the failure modes, we trace back a considerable portion of this drop to discrepancies between the formal and informal statements for more than half of the problems in miniF2F. We proceed with correcting all the errors, discrepancies and simplifications in formal and informal statements, and present theminiF2F-v2with fully verified formal and informal statements and proofs. Evaluating the full theorem proving pipeline onminiF2F-v2leads to the best accuracy of 57\%, a significant improvement from the 36\% on miniF2F-v1, yet indicating considerable misalignment between the autoformalization models and theorem provers. Our deep analysis suggests that a higher quality benchmark can help the community better evaluate progress in the field of formal reasoning and also better diagnose the failure and success modes of autoformalization and theorem proving models. We release our dataset,miniF2F-v2, and all the corresponding formal proofs to the public.



Paperid:1854
Authors:Yichuan Ma, Linyang Li, Yongkang Chen, Peiji Li, Jiasheng Ye, Qipeng Guo, Dahua Lin, Kai Chen
Abstract:
Large language models (LLMs) have demonstrated exceptional performance in reasoning tasks such as mathematics and coding, matching or surpassing human capabilities. However, these impressive reasoning abilities face significant challenges in specialized domains. Taking Go as an example, although AlphaGo has established the high performance ceiling of AI systems in Go, mainstream LLMs still struggle to reach even beginner-level proficiency, let alone perform natural language reasoning. This performance gap between general-purpose LLMs and domain experts is significantly limiting the application of LLMs on a wider range of domain-specific tasks. In this work, we aim to bridge the divide between LLMs' general reasoning capabilities and expert knowledge in domain-specific tasks. We perform mixed fine-tuning with structured Go expertise and general long Chain-of-Thought (CoT) reasoning data as a cold start, followed by reinforcement learning to integrate expert knowledge in Go with general reasoning capabilities. Through this methodology, we present LoGos, a powerful LLM that not only maintains outstanding general reasoning abilities, but also conducts Go gameplay in natural language, demonstrating effective strategic reasoning and accurate next-move prediction. LoGos achieves performance comparable to human professional players, substantially surpassing all existing LLMs. Through this work, we aim to contribute insights on applying general LLM reasoning capabilities to specialized domains. We will release the first large-scale Go dataset for LLM training, the first LLM Go evaluation benchmark, and the first general LLM that reaches human expert-level performance in Go.



Paperid:1855
Authors:Chenheng Zhang, Tianqi Du, Jizhe Zhang, Mingqing Xiao, Yifei Wang, Yisen Wang, Zhouchen Lin
Abstract:
Conventional research on large language models (LLMs) has primarily focused on refining output distributions, while paying less attention to the decoding process that transforms these distributions into final responses. Recent advances, such as scaling the computation of inference time with reward models, have underscored the importance of decoding, but these methods often suffer from high computational costs and limited applicability.In this paper, we revisit LLM decoding through the lens of recommender systems, conceptualizing the decoding process as analogous to the ranking stage in recommendation pipelines. From this perspective, we observe that both traditional decoding methods and reward models exhibit clear limitations such as redundancy.Motivated by this insight, we propose Language System, a novel framework that introduces a lightweight ranker to rerank candidate responses using features extracted by the base model. Experiments across a wide range of tasks show that Language System achieves performance comparable to large-scale reward models, while requiring only <0.5M additional parameters, significantly reducing the computational overhead during both training and inference stages. This highlights the efficiency and effectiveness of our method, showcasing its potential to fully unlock the capabilities of LLMs.



Paperid:1856
Authors:Matthias Schultheis, Jana-Sophie Schönfeld, Constantin Rothkopf, Heinz Koeppl
Abstract:
Human behavior is characterized by continuous learning to reduce uncertainties about the world in pursuit of goals. When trying to understand such behavior from observations, it is essential to account for this adaptive nature and reason about the uncertainties that may have led to seemingly suboptimal decisions. Nevertheless, most inverse approaches to sequential decision-making focus on inferring cost functions underlying stationary behavior or are limited to low-dimensional tasks. In this paper, we address this gap by considering the problem of inferring an agent's knowledge about the environment based on a given trajectory. We assume that the agent aims to reach a goal in an environment they only partially know, and integrates new information into their plan as they act. We propose a Bayesian approach to infer their latent knowledge state, leveraging an approximate navigation model that optimistically incorporates partial information while accounting for uncertainty. By combining sample-based Bayesian inference with dynamic graph algorithms, we achieve an efficient method for computing posterior beliefs about the agent's knowledge. Empirical validation using simulated behavioral data and human data from an online experiment demonstrates that our model effectively captures human navigation under uncertainty and reveals interpretable insights into their environmental knowledge.



Paperid:1857
Authors:Gleb Rodionov, Roman Garipov, Alina Shutova, George Yakushev, Erik Schultheis, Vage Egiazarian, Anton Sinitsin, Denis Kuznedelev, Dan Alistarh
Abstract:
Large Language Models (LLMs) have demonstrated the ability to tackle increasingly complex tasks through advanced reasoning, long-form content generation, and tool use. Solving these tasks often involves long inference-time computations. In human problem solving, a common strategy to expedite work is collaboration: by dividing the problem into sub-tasks, exploring different strategies concurrently, etc. Recent research has shown that LLMs can also operate in parallel by implementing explicit cooperation frameworks, such as voting mechanisms or the explicit creation of independent sub-tasks that can be executed in parallel. However, each of these frameworks may not be suitable for all types of tasks, which can hinder their applicability. In this work, we propose a different design approach: we run LLM "workers" in parallel , allowing them to synchronize via a concurrently-updated attention cache and prompt these workers to decide how best to collaborate. Our approach allows the instances to come up with their own collaboration strategy for the problem at hand, all the while "seeing" each other's partial progress in the concurrent cache. We implement this approach via Hogwild! Inference: a parallel LLM inference engine where multiple instances of the same LLM run in parallel with the same attention cache, with "instant" access to each other's generated tokens. Hogwild! inference takes advantage of Rotary Position Embeddings (RoPE) to avoid recomputation while improving parallel hardware utilization. We find that modern reasoning-capable LLMs can perform inference with shared Key-Value cache out of the box, without additional fine-tuning.



Paperid:1858
Authors:Abhinav Chakraborty, Arnab Auddy, T. Tony Cai
Abstract:
Abstract:We consider the problem of estimating the correlation of two random variables $X$ and $Y$, where the pairs $(X,Y)$ are not observed together, but are instead separated co-ordinate-wise at two servers: server 1 contains all the $X$ observations, and server 2 contains the corresponding $Y$ observations. In this vertically distributed setting, we assume that each server has its own privacy constraints, owing to which they can only share suitably privatized statistics of their own component observations. We consider differing privacy budgets $(\varepsilon_1,\delta_1)$ and $(\varepsilon_2,\delta_2)$ for the two servers and determine the minimax optimal rates for correlation estimation allowing for both non-interactive and interactive mechanisms. We also provide correlation estimators that achieve these rates and further develop inference procedures, namely, confidence intervals, for the estimated correlations. Our results are characterized by an interesting rate in terms of the sample size $n$, $\varepsilon_1$, $\varepsilon_2$, which is strictly slower than the usual central privacy estimation rates. More interestingly, we find that the interactive mechanism is always better than its non-interactive counterpart whenever the two privacy budgets are different. Results from extensive numerical experiments support our theoretical findings.



Paperid:1859
Authors:Han Xu, Yuyang Li, Yunfei Deng, Jiayi Ma, Guangcan Liu
Abstract:
Most image fusion methods are designed for ideal scenarios and struggle to handle noise. Existing noise-aware fusion methods are supervised and heavily rely on constructed paired data, limiting performance and generalization. This paper proposes a novel unsupervised noisy visible and infrared image fusion method, comprising two key modules. First, when only noisy source images are available, a convolutional low-rank optimization module decomposes clean components based on convolutional low-rank priors, guiding subsequent optimization. The unsupervised approach eliminates data dependency and enhances generalization across various and variable noise. Second, a unified network jointly realizes denoising and fusion. It consists of both intra-modal recovery and intra-modal recovery and fusion, also with a convolutional low-rankness loss for regularization. By exploiting the commonalities of denoising and fusion, the joint framework significantly reduces network complexity while expanding functionality. Extensive experiments validate the effectiveness and generalization of the proposed method for image fusion under various and variable noise conditions. The code will be publicly available.



Paperid:1860
Authors:Kaixuan Yao, Zhuo Li, Jianqing Liang, Jiye Liang, Ming Li, Feilong Cao
Abstract:
Hypergraphs offer a natural paradigm for modeling complex systems with multi-way interactions. Hypergraph neural networks (HGNNs) have demonstrated remarkable success in learning from such higher-order relational data. While such higher-order modeling enhances relational reasoning, the effectiveness of hypergraph learning remains bottlenecked by two persistent challenges: the scarcity of labeled data inherent to complex systems, and the vulnerability to structural noise in real-world interaction patterns. Traditional data augmentation methods, though successful in Euclidean and graph-structured domains, struggle to preserve the intricate balance between node features and hyperedge semantics, often disrupting the very group-wise interactions that define hypergraph value. To bridge this gap, we present HyperMixup, a hypergraph-aware augmentation framework that preserves higher-order interaction patterns through structure-guided feature mixing. Specifically, HyperMixup contains three critical components: 1) Structure-aware node pairing guided by joint feature-hyperedge similarity metrics, 2) Context-enhanced hierarchical mixing that preserves hyperedge semantics through dual-level feature fusion, and 3) Adaptive topology reconstruction mechanisms that maintain hypergraph consistency while enabling controlled diversity expansion. Theoretically, we establish that our method induces hypergraph-specific regularization effects through gradient alignment with hyperedge covariance structures, while providing robustness guarantees against combined node-hyperedge perturbations. Comprehensive experiments across diverse hypergraph learning tasks demonstrate consistent performance improvements over state-of-the-art baselines, with particular effectiveness in low-label regimes. The proposed framework advances hypergraph representation learning by unifying data augmentation with higher-order topological constraints, offering both practical utility and theoretical insights for relational machine learning.



Paperid:1861
Authors:Yifan Xie, Mingyang Li, Shoujie Li, Xingting Li, Guangyu Chen, Fei Ma, Fei Yu, Wenbo Ding
Abstract:
Tactile perception is essential for embodied agents to understand the physical attributes of objects that cannot be determined through visual inspection alone. While existing methods have made progress in visual and language modalities for physical understanding, they fail to effectively incorporate tactile information that provides crucial haptic feedback for real-world interaction. In this paper, we present VTV-LLM, the first multi-modal large language model that enables universal Visuo-Tactile Video (VTV) understanding, bridging the gap between tactile perception and natural language. To address the challenges of cross-sensor and cross-modal integration, we contribute VTV150K, a comprehensive dataset comprising 150,000 video frames from 100 diverse objects captured across three different tactile sensors (GelSight Mini, DIGIT, and Tac3D), annotated with four fundamental tactile attributes (hardness, protrusion, elasticity, and friction). We develop a novel three-stage training paradigm that includes VTV enhancement for robust visuo-tactile representation, VTV-text alignment for cross-modal correspondence, and text prompt finetuning for natural language generation. Our framework enables sophisticated tactile reasoning capabilities including feature assessment, comparative analysis, and scenario-based decision-making. Extensive experimental evaluations demonstrate that VTV-LLM achieves superior performance in tactile reasoning tasks, establishing a foundation for more intuitive human-machine interaction in tactile domains.



Paperid:1862
Authors:Hilal Asi, Vitaly Feldman, Hannah Keller, Guy Rothblum, Kunal Talwar
Abstract:
We revisit the problem of secure aggregation of high-dimensional vectors in a two-server system such as Prio. These systems are typically used to aggregate vectors such as gradients in private federated learning, where the aggregate itself is protected via noise addition to ensure differential privacy. Existing approaches require communication scaling with the dimensionality, and thus limit the dimensionality of vectors one can efficiently process in this setup. We propose PREAMBLE: {\bf Pr}ivate {\bf E}fficient {\bf A}ggregation {\bf M}echanism via {\bf BL}ock-sparse {\bf E}uclidean Vectors. PREAMBLE builds on an extension of distributed point functions that enables communication- and computation-efficient aggregation of {\em block-sparse vectors}, which are sparse vectors where the non-zero entries occur in a small number of clusters of consecutive coordinates. We show that these block-sparse DPFs can be combined with random sampling and privacy amplification by sampling results, to allow asymptotically optimal privacy-utility trade-offs for vector aggregation, at a fraction of the communication cost. When coupled with recent advances in numerical privacy accounting, our approach incurs a negligible overhead in noise variance, compared to the Gaussian mechanism used with Prio.



Paperid:1863
Authors:Yichen Li, Yuying Wang, Jiahua Dong, Haozhao Wang, Yining Qi, Rui Zhang, Ruixuan Li
Abstract:
Federated Continual Learning (FCL) aims to enable sequential privacy-preserving model training on streams of incoming data that vary in edge devices by preserving previous knowledge while adapting to new data. Current FCL literature focuses on restricted data privacy and access to previously seen data while imposing no constraints on the training overhead. This is unreasonable for FCL applications in real-world scenarios, where edge devices are primarily constrained by resources such as storage, computational budget, and label rate. We revisit this problem with a large-scale benchmark and analyze the performance of state-of-the-art FCL approaches under different resource-constrained settings. Various typical FCL techniques and six datasets in two incremental learning scenarios (Class-IL and Domain-IL) are involved in our experiments. Through extensive experiments amounting to a total of over 1,000+ GPU hours, we find that, under limited resource-constrained settings, existing FCL approaches, with no exception, fail to achieve the expected performance. Our conclusions are consistent in the sensitivity analysis. This suggests that most existing FCL methods are particularly too resource-dependent for real-world deployment. Moreover, we study the performance of typical FCL techniques with resource constraints and shed light on future research directions in FCL.



Paperid:1864
Authors:feilong tang, xiangan, Haolin Yang, Yin Xie, Kaicheng Yang, Ming Hu, Zheng Cheng, Xingyu Zhou, Zimin Ran, Imran Razzak, Ziyong Feng, Behzad Bozorgtabar, Jiankang Deng, Zongyuan Ge
Abstract:
Despite the impressive progress of recent pretraining methods on multimodal tasks, existing methods are inherently biased towards either spatial modeling (e.g., CLIP) or temporal modeling (e.g., V-JEPA), limiting their joint capture of spatial details and temporal dynamics. To this end, we propose UniViT, a cluster-driven unified self-supervised learning framework that effectively captures the structured semantics of both image spatial content and video temporal dynamics through event-level and object-level clustering and discrimination. Specifically, we leverage offline clustering to generate semantic clusters across both modalities. For videos, multi-granularity event-level clustering progressively expands from single-event to structured multi-event segments, capturing coarse-to-fine temporal semantics; for images, object-level clustering captures fine-grained spatial semantics. However, while global clustering provides semantically consistent clusters, it lacks modeling of structured semantic relations (e.g., temporal event structures). To address this, we introduce a contrastive objective that leverages these semantic clusters as pseudo-label supervision to explicitly enforce structural constraints, including temporal event relations and spatial object co-occurrences, capturing structured semantics beyond categories. Meanwhile, UniViT jointly embeds structured object-level and event-level semantics into a unified representation space. Furthermore, UniViT introduces two key components: (i) Unified Rotary Position Embedding integrates relative positional embedding with frequency-aware dimension allocation to support position-invariant semantic learning and enhance the stability of structured semantics in the discrimination stage; and (ii) Variable Spatiotemporal Streams adapt to inputs of varying frame lengths, addressing the rigidity of conventional fixed-input approaches. Extensive experiments across varying model scales demonstrate that UniViT achieves state-of-the-art performance on linear probing, attentive probing, question answering, and spatial understanding tasks.



Paperid:1865
Authors:Yi-Kai Zhang, Shiyin Lu, Qingguo Chen, Weihua Luo, De-Chuan Zhan, Han-Jia Ye
Abstract:
Recently, test-time scaling of Large Language Models (LLMs) has emerged as a practical alternative to parameter and data scaling for reasoning tasks. Complex tasks often require large-scale parameters or advanced reasoning approaches like Chain-of-Thought (CoT) or RL-based LLMs, while more economical LLMs can handle simpler tasks. Routing an LLM tailored to domain and task complexity ensures flexibility and efficiency. We introduce EcoRouter, which routes the most suitable LLM based on throughput and reasoning process length. It constructs historical interactions between models and instructions as "preference scores" in a recommendation system and uses refined cold-start data for model representations. EcoRouter generalizes to unseen out-of-domain instructions and LLMs without fine-tuning. We also define the Pareto frontier-based metric under user-defined cost thresholds to balance cost and accuracy. We evaluate EcoRouter on a benchmark of over 50 LLMs, including DeepSeek-R1 and its distilled extensions, and three inference modes: self-consistency, CoT, and Tree-of-Thought (ToT). Our approach reduces unit costs by over 38% while maintaining performance, outperforming the average-optimal model for each dataset at the same cost level.



Paperid:1866
Authors:Salim I. Amoukou, Nicolas Brunel
Abstract:
Abstract:We analyze two widely used local attribution methods, Local Shapley Values and LIME, which aim to quantify the contribution of a feature value $x_i$ to a specific prediction $f(x_1, \dots, x_p)$. Despite their widespread use, we identify fundamental limitations in their ability to reliably detect locally important features, even under ideal conditions with exact computations and independent features. We argue that a sound local attribution method should not assign importance to features that neither influence the model output (e.g., features with zero coefficients in a linear model) nor exhibit statistical dependence with functionality-relevant features. We demonstrate that both Local SV and LIME violate this fundamental principle. To address this, we propose R-LOCO (Regional Leave Out COvariates), which bridges the gap between local and global explanations and provides more accurate attributions. R-LOCO segments the input space into regions with similar feature importance characteristics. It then applies global attribution methods within these regions, deriving an instance's feature contributions from its regional membership. This approach delivers more faithful local attributions while avoiding local explanation instability and preserving instance-specific detail often lost in global methods.



Paperid:1867
Authors:Xiaoyang Yi, Yang Liu, Binhan Yang, Jian Zhang
Abstract:
Federated continual learning (FCL) aims to maintain the model's performance on old tasks (i.e., stability) while enhancing its ability to acquire knowledge from current tasks (i.e., plasticity). With the development of pre-trained models (PTMs), fine-tuning PTMs on clients has become a promising approach to leveraging their extensive knowledge in FCL. In this paper, we propose MultiFCL, a novel FCL framework that fine-tunes PTMs to adapt to FCL while preserving their strong generalization capabilities. Specifically, to ensure the stability, MultiFCL introduces lightweight adapters for task adaption, which are subsequently frozen to prevent catastrophic forgetting. Moreover, by utilizing the semantic features of old tasks, MultiFCL performs multi-modal initialization of new task class prototypes. To enhance the plasticity, MultiFCL employs a multi-expert training mechanism that integrates multi-scale feature learning with multi-teacher dynamic self-distillation. Through intra-client and inter-client expert communication, MultiFCL facilitates cross-task and cross-client knowledge fusion. Experimental results demonstrate that MultiFCL achieves state-of-the-art performance across multiple datasets and settings, showcasing its effectiveness in FCL scenarios.



Paperid:1868
Authors:Zhilong Zhang, Yuxuan Song, Yichun Wang, Jingjing Gong, Hanlin Wu, Dongzhan Zhou, Hao Zhou, Wei-Ying Ma
Abstract:
Geometric molecule generative models have found expanding applications across various scientific domains, but their generation inefficiency has become a critical bottleneck. Through a systematic investigation of the generative trajectory, we discover a unique challenge for molecule geometric graph generation: generative models require determining the permutation order of atoms in the molecule before refining its atomic feature values. Based on this insight, we decompose the generation process into permutation phase and adjustment phase, and propose a geometric-informed prior and consistency parameter objective to accelerate each phase. Extensive experiments demonstrate that our approach achieves competitive performance with approximately 10 sampling steps, 7.5 × faster than previous state-of-the-art models and approximately 100 × faster than diffusion-based models, offering a significant step towards scalable molecular generation.



Paperid:1869
Authors:Yuchen Liang, Yingbin Liang, Lifeng LAI, Ness Shroff
Abstract:
Abstract:Discrete diffusion models have recently gained significant prominence in applications involving natural language and graph data. A key factor influencing their effectiveness is the efficiency of discretized samplers. Among these, $\tau$-leaping samplers have become particularly popular due to their empirical success. However, existing theoretical analyses of $\tau$-leaping often rely on somewhat restrictive and difficult-to-verify regularity assumptions, and their convergence bounds contain quadratic dependence on the vocabulary size. In this work, we introduce a new analytical approach for discrete diffusion models that removes the need for such assumptions. For the standard $\tau$-leaping method, we establish convergence guarantees in KL divergence that scale linearly with vocabulary size, improving upon prior results with quadratic dependence. Our approach is also more broadly applicable: it provides the first convergence guarantees for other widely used samplers, including the Euler method and Tweedie $\tau$-leaping. Central to our approach is a novel technique based on differential inequalities, offering a more flexible alternative to the traditional Girsanov change-of-measure methods. This technique may also be of independent interest for the analysis of other stochastic processes.



Paperid:1870
Authors:Niklas Kemper, Tom Wollschläger, Stephan Günnemann
Abstract:
Expressivity theory, characterizing which graphs a GNN can distinguish, has become the predominant framework for analyzing GNNs, with new models striving for higher expressivity. However, we argue that this focus is misguided: First, higher expressivity is not necessary for most real-world tasks as these tasks rarely require expressivity beyond the basic WL test. Second, expressivity theory's binary characterization and idealized assumptions fail to reflect GNNs' practical capabilities.To overcome these limitations, we propose Message Passing Complexity (MPC): a continuous measure that quantifies the difficulty for a GNN architecture to solve a given task through message passing.MPC captures practical limitations like over-squashing while preserving the theoretical impossibility results from expressivity theory, effectively narrowing the gap between theory and practice.Through extensive validation on fundamental GNN tasks, we show that MPC's theoretical predictions correlate with empirical performance, successfully explaining architectural successes and failures. Thereby, MPC advances beyond expressivity theory to provide a more powerful framework for understanding and developing GNN architectures.



Paperid:1871
Authors:Theodoros Kouzelis, Efstathios Karypidis, Ioannis Kakogeorgiou, Spyridon Gidaris, Nikos Komodakis
Abstract:
Latent diffusion models (LDMs) dominate high-quality image generation, yet integrating representation learning with generative modeling remains a challenge. We introduce a novel generative image modeling framework that seamlessly bridges this gap by leveraging a diffusion model to jointly model low-level image latents (from a variational autoencoder) and high-level semantic features (from a pretrained self-supervised encoder like DINO). Our latent-semantic diffusion approach learns to generate coherent image-feature pairs from pure noise, significantly enhancing both generative quality and training efficiency, all while requiring only minimal modifications to standard Diffusion Transformer architectures. By eliminating the need for complex distillation objectives, our unified design simplifies training and unlocks a powerful new inference strategy: Representation Guidance, which leverages learned semantics to steer and refine image generation. Evaluated in both conditional and unconditional settings, our method delivers substantial improvements in image quality and training convergence speed, establishing a new direction for representation-aware generative modeling.



Paperid:1872
Authors:Yuxian Gu, Qinghao Hu, Haocheng Xi, Junyu Chen, Shang Yang, Song Han, Han Cai
Abstract:
We present JetLM, a new family of hybrid-architecture language models, which matches or exceeds the accuracy of leading full-attention models while significantly improving generation throughput. JetLM is developed using Post Neural Architecture Search (PostNAS), a novel neural architecture exploration pipeline that enables efficient model design. Unlike prior approaches, PostNAS begins with a pre-trained full-attention model and freezes its MLP weights, allowing efficient exploration of attention block designs. The pipeline includes four key components: (1) learning optimal full-attention layer placement and elimination, (2) linear attention block selection, (3) designing new attention blocks, and (4) performing hardware-aware hyperparameter search. Our JetLM-2B model achieves comparable or superior accuracy to Qwen2.5, Gemma3, and Llama3.2 across a comprehensive suite of benchmarks while delivering up to 13x generation throughput speedup and 4.7x prefilling speedup. We will release our models and code upon publication.



Paperid:1873
Authors:Timo Stoll, Luis Müller, Christopher Morris
Abstract:
Graph Transformers (GTs) have shown strong empirical performance, yet current architectures vary widely in their use of attention mechanisms, positional embeddings (PEs), and expressivity. Existing expressivity results are often tied to specific design choices and lack comprehensive empirical validation on large-scale data. This leaves a gap between theory and practice, preventing generalizable insights that exceed particular application domains. Here, we propose the Generalized-Distance Transformer (GDT), a GT architecture using standard attention that incorporates many advancements for GTs from recent years, and develop a fine-grained understanding of the GDT's representation power in terms of attention and PEs. Through extensive experiments, we identify design choices that consistently perform well across various applications, tasks, and model scales, demonstrating strong performance in a few-shot transfer setting without fine-tuning. Our evaluation covers over eight million graphs with roughly 270M tokens across diverse domains, including image-based object detection, molecular property prediction, code summarization, and out-of-distribution algorithmic reasoning. We distill our theoretical and practical findings into several generalizable insights about effective GT design, training, and inference.



Paperid:1874
Authors:Ziyang Zheng, Kezhi Li, Zhengyuan Shi, Qiang Xu
Abstract:
Subgraph matching in logic circuits is foundational for numerous Electronic Design Automation (EDA) applications, including datapath optimization, arithmetic verification, and hardware trojan detection. However, existing techniques rely primarily on structural graph isomorphism and thus fail to identify function-related subgraphs when synthesis transformations substantially alter circuit topology. To overcome this critical limitation, we introduce the concept of functional subgraph matching, a novel approach that identifies whether a given logic function is implicitly present within a larger circuit, irrespective of structural variations induced by synthesis or technology mapping. Specifically, we propose a two-stage multi-modal framework: (1) learning robust functional embeddings across AIG and post-mapping netlists for functional subgraph detection, and (2) identifying fuzzy boundaries using a graph segmentation approach. Evaluations on standard benchmarks (ITC99, OpenABCD, ForgeEDA) demonstrate significant performance improvements over existing structural methods, with average 93.8% accuracy in functional subgraph detection and a dice score of 91.3% in fuzzy boundary identification.



Paperid:1875
Authors:Zechun Liu, Changsheng Zhao, Hanxian Huang, Sijia Chen, Jing Zhang, Jiawei Zhao, Scott Roy, Lisa Jin, Yunyang Xiong, Yangyang Shi, Lin Xiao, Yuandong Tian, Bilge Soran, Raghuraman Krishnamoorthi, Tijmen Blankevoort, Vikas Chandra
Abstract:
The optimal bit-width for achieving the best trade-off between quantized model size and accuracy has been a subject of ongoing debate. While some advocate for 4-bit quantization, others propose that 1.58-bit offers superior results. However, the lack of a cohesive framework for different bits has left such conclusions relatively tenuous. We present ParetoQ, the first unified framework that facilitates rigorous comparisons across 1-bit, 1.58-bit, 2-bit, 3-bit, and 4-bit quantization settings. Our findings reveal a notable learning transition between 2 and 3 bits: For 3-bits and above, the fine-tuned models stay close to their original pre-trained distributions, whereas for learning 2-bit networks or below, the representations change drastically. By optimizing training schemes and refining quantization functions, ParetoQ surpasses all previous methods tailored to specific bit widths. Remarkably, our ParetoQ ternary 600M-parameter model even outperforms the previous SoTA ternary 3B-parameter model in accuracy, using only one-fifth of the parameters. Extensive experimentation shows that ternary, 2-bit, and 3-bit quantization maintains comparable performance in the size-accuracy trade-off and generally exceeds 4-bit and binary quantization. Considering hardware constraints, 2-bit quantization offers promising potential for memory reduction and speedup.



Paperid:1876
Authors:Tzu-Han Hsu, Arshia Rafieioskouei, Borzoo Bonakdarpour
Abstract:
Abstract:Reward shaping in multi-agent reinforcement learning (MARL) for complex tasks remains a significant challenge. Existing approaches often fail to find optimal solutions or cannot efficiently handle such tasks. We propose HYPRL, a specification-guided reinforcement learning framework that learns control policies w.r.t. hyperproperties expressed in HyperLTL. Hyperproperties constitute a powerful formalism for specifying objectives and constraints over sets of execution traces across agents. To learn policies that maximize the satisfaction of a HyperLTL formula $\varphi$, we apply Skolemization to manage quantifier alternations and define quantitative robustness functions to shape rewards over execution traces of a Markov decision process with unknown transitions. A suitable RL algorithm is then used to learn policies that collectively maximize the expected reward and, consequently, increase the probability of satisfying $\varphi$. We evaluate HYPRL on a diverse set of benchmarks, including safety-aware planning, Deep Sea Treasure, and the Post Correspondence Problem. We also compare with specification-driven baselines to demonstrate the effectiveness and efficiency of HYPRL.



Paperid:1877
Authors:Long Ma, Fangwei Zhong, Yizhou Wang
Abstract:
Abstract:Modern causal language models, followed by rapid developments in discrete diffusion models, can now produce a wide variety of interesting and useful content. However, these families of models are predominantly trained to output tokens with a fixed (left-to-right) or random order, which may deviate from the logical order in which tokens are generated originally. In this paper, we observe that current causal and diffusion models encounter difficulties in problems that require adaptive token generation orders to solve tractably, which we characterize with the $\mathcal{V}$-information framework. Motivated by this, we propose Reinforced Context Order Recovery (ReCOR), a reinforcement-learning-based framework to extract adaptive, data-dependent token generation orders from text data without annotations. Self-supervised by token prediction statistics, ReCOR estimates the hardness of predicting every unfilled token and adaptively selects the next token during both training and inference. Experiments on challenging reasoning and planning datasets demonstrate the superior performance of ReCOR compared with baselines, sometimes outperforming oracle models supervised with the ground-truth order.



Paperid:1878
Authors:Juan Rodriguez, Haotian Zhang, Abhay Puri, Rishav Pramanik, Aarash Feizi, Pascal Wichmann, Arnab Kumar Mondal, Mohammad R. Samsami, Rabiul Awal, Perouz Taslakian, Spandana Gella, Sai Rajeswar Mudumba, David Vazquez, Chris Pal, Marco Pedersoli
Abstract:
Scalable Vector Graphics (SVG) offer a powerful format for representing visual designs as interpretable code. Recent advances in vision-language models (VLMs) have enabled high-quality SVG generation by framing the problem as a code generation task and leveraging large-scale pretraining. VLMs are particularly suitable for this task as they capture both global semantics and fine-grained visual patterns, while transferring knowledge across vision, natural language, and code domains. However, existing VLM approaches often struggle to produce faithful and efficient SVGs because they never observe the rendered images during training. Although differentiable rendering for autoregressive SVG code generation remains unavailable, rendered outputs can still be compared to original inputs, enabling evaluative feedback suitable for reinforcement learning (RL). In this work, we introduce RLVG, a reinforcement learning approach for SVG generation with autoregressive VLMs. Given an input image, the model generates SVG rollouts that are rendered and compared to the original image to compute a reward. This visual fidelity feedback guides the model toward producing more accurate, efficient, and semantically coherent SVGs. RLVG significantly outperforms supervised fine-tuning, addressing common failure modes and enabling precise, high-quality SVG generation with strong structural understanding and generalization.



Paperid:1879
Authors:Zhiheng Liu, Xueqing Deng, Shoufa Chen, Angtian Wang, Qiushan Guo, Mingfei Han, Zeyue Xue, Mengzhao Chen, Ping Luo, Linjie Yang
Abstract:
Generative video modeling has made significant strides, yet ensuring structural and temporal consistency over long sequences remains a challenge. Current methods predominantly rely on RGB signals, leading to accumulated errors in object structure and motion over extended durations. To address these issues, we introduce WorldWeaver, a robust framework for long video generation that jointly models RGB frames and perceptual conditions within a unified long-horizon modeling scheme. Our training framework offers three key advantages. First, by jointly predicting perceptual conditions and color information from a unified representation, it significantly enhances temporal consistency and motion dynamics. Second, by leveraging depth cues, which we observe to be more resistant to drift than RGB, we construct a memory bank that preserves clearer contextual information, improving quality in long-horizon video generation. Third, we employ segmented noise scheduling for training prediction groups, which further mitigates drift and reduces computational cost. Extensive experiments on both diffusion and rectified flow-based models demonstrate the effectiveness of WorldWeaver in reducing temporal drift and improving the fidelity of generated videos.



Paperid:1880
Authors:Enqi Liu, Liyuan Pan, Yan Yang, Yiran Zhong, Zhijing Wu, Xinxiao Wu, Liu Liu
Abstract:
Fine-grained video action recognition can be formulated as a video–text matching problem. Previous approaches primarily rely on global video semantics to consolidate video embeddings, often leading to misaligned video–text pairs due to inaccurate atomic-level action understanding. This inaccuracy arises due to i) videos with distinct global semantics may share similar atomic actions or visual appearances, and ii) atomic actions can be momentary, gradual, or not directly aligned with overarching video semantics. Inspired by storyboarding, where a script is segmented into individual shots, we propose a multi-granularity framework, SFAR. SFAR generates fine-grained descriptions of common atomic actions for each global semantic using a large language model. Unlike existing works that refine global semantics with auxiliary video frames, SFAR introduces a filtering metric to ensure correspondence between the descriptions and the global semantics, eliminating the need for direct video involvement and thereby enabling more nuanced recognition of subtle actions. By leveraging both global semantics and fine-grained descriptions, our SFAR effectively identifies prominent frames within videos, thereby improving the accuracy of embedding aggregation. Extensive experiments on various video action recognition datasets demonstrate the competitive performance of our SFAR in supervised, few-shot, and zero-shot settings.



Paperid:1881
Authors:Joohwan Ko, Justin Domke
Abstract:
Variational inference often struggles with the posterior geometry exhibited by complex hierarchical Bayesian models. Recent advances in flow‐based variational families and Variationally Inferred Parameters (VIP) each address aspects of this challenge, but their formal relationship is unexplored. Here, we prove that the combination of VIP and a full-rank Gaussian can be represented exactly as a forward autoregressive flow augmented with a translation term and input from the model’s prior. Guided by this theoretical insight, we introduce the Model‐Informed Flow (MIF) architecture, which adds the necessary translation mechanism, prior information, and hierarchical ordering. Empirically, MIF delivers tighter posterior approximations and matches or exceeds state‐of‐the‐art performance across a suite of hierarchical and non‐hierarchical benchmarks.



Paperid:1882
Authors:Xiao Cao, Beibei Lin, Bo Wang, Zhiyong Huang, Robby Tan
Abstract:
Image-based 3D texture transfer from a single 2D reference image enables practical customization of 3D object appearances with minimal manual effort.Adapted 2D editing and text-driven 3D editing approaches can serve this purpose. However, 2D editing typically involves frame-by-frame manipulation, often resulting in inconsistencies across views, while text-driven 3D editing struggles to preserve texture characteristics from reference images.To tackle these challenges, we introduce 3D-OTT, a 3D Object Texture Transfer method based on a single reference image, integrating: 1) progressive generation, 2) view-consistency gradient guidance, and 3) prompt-tuned gradient guidance.To ensure view consistency, progressive generation starts by transferring texture from the reference image and gradually propagates it to adjacent views.View-consistency gradient guidance further reinforces coherence by conditioning the generation model on feature differences between consistent and inconsistent outputs.To preserve texture characteristics, prompt-tuning-based gradient guidance learns a token that describes differences between original and reference textures, guiding the transfer for faithful texture preservation across views.Overall, 3D-OTT combines these strategies to achieve effective texture transfer while maintaining structural coherence across viewpoints.Extensive qualitative and quantitative evaluations confirm that our three components enable convincing and effective 2D-to-3D texture transfer. Code will be available upon acceptance.



Paperid:1883
Authors:Zishen Zhang, Xiangzhe Kong, Wenbing Huang, Yang Liu
Abstract:
Designing protein binders targeting specific sites, which requires to generate realistic and functional interaction patterns, is a fundamental challenge in drug discovery. Current structure-based generative models are limited in generating nterfaces with sufficient rationality and interpretability. In this paper, we proposeRetrieval-AugmentedDiffusion forAligned interface(RADiAnce), a new framework that leverages known interfaces to guide the design of novel binders. By unifying retrieval and generation in a shared contrastive latent space, our model efficiently identifies relevant interfaces for a given binding site and seamlessly integrates them through a conditional latent diffusion generator, enabling cross-domain interface transfer. Extensive exeriments show thatRADiAncesignificantly outperforms baseline models across multiple metrics, including binding affinity and recovery of geometries and interactions.Additional experimental results validate cross-domain generalization, demonstrating that retrieving interfaces from diverse domains, such as peptides, antibodies, and protein fragments, enhances the generation performance of binders for other domains. Our work establishes a new paradigm for protein binder design that successfully bridges retrieval-based knowledge and generative AI, opening new possibilities for drug discovery.



Paperid:1884
Authors:Félix Chavelli, Paul Boniol, Michaël Thomazo
Abstract:
Time series segmentation is a fundamental task in analyzing temporal data across various domains, from human activity recognition to energy monitoring. While numerous state-of-the-art methods have been developed to tackle this problem, the evaluation of their performance remains critically limited. Existing measures predominantly focus on change point accuracy or rely on point-based metrics such as Adjusted Rand Index (ARI), which fail to capture the quality of the detected segments, ignore the nature of errors, and offer limited interpretability. In this paper, we address these shortcomings by introducing two novel evaluation measures: WARI (Weighted Adjusted Rand Index), a temporal extension of ARI that accounts for the position of segmentation errors, and SMS (State Matching Score), a fine-grained metric that identifies and scores four distinct and fundamental types of segmentation errors while allowing error-specific weighting. We empirically validate WARI and SMS on synthetic and real-world benchmarks, showing that they not only provide a more accurate assessment of segmentation quality but also uncover insights, such as error provenance and type, that are inaccessible with traditional measures.



Paperid:1885
Authors:Kuan Wei Huang, Brandon Li, Bharath Hariharan, Noah Snavely
Abstract:
Geometric models like DUSt3R have shown great advances in understanding the geometry of a scene from pairs of photos. However, they fail when the inputs are from vastly different viewpoints (e.g., aerial vs. ground) or modalities (e.g., photos vs. abstract drawings) due to the vast differences in viewpoint or style compared to what was observed during training. This paper addresses a challenging version of this problem: predicting correspondences between ground-level photos and floor plans. Current datasets for joint photo--floor plan reasoning are limited, either lacking in varying modalities (VIGOR) or lacking in correspondences (WAFFLE). To address these limitations, we introduce a new dataset, C3, created by first reconstructing a number of scenes in 3D from Internet photo collections via structure from motion, then manually registering the reconstructions to floor plans gathered from the Internet, from which we can derive correspondence between images and floor plans. C3 contains 91K paired floor plans and photos across 574 scenes with 155M pixel-level correspondences. We find that state-of-the-art correspondence models struggle on this task. By training on our new data, we can improve on the best performing method by 34\% in RMSE. However, we also identify open challenges in cross-modal geometric reasoning that our dataset aims to help address.



Paperid:1886
Authors:Xiuwen Fang, Mang Ye
Abstract:
Federated Learning enables collaborative training across decentralized edge devices while preserving data privacy. However, fine-tuning large-scale pre-trained models in federated learning is hampered by substantial communication overhead and client resource limitations. Parameter-efficient fine-tuning methods like Low-Rank Adaptation (LoRA) reduce resource demands but suffer from aggregation discrepancies and heightened vulnerability to label noise, particularly in heterogeneous federated settings. In this paper, we introduce RFedLR, a robust federated PEFT framework designed to overcome these challenges. RFedLR integrates two key components: (1) Sensitivity-aware robust tuning, which identifies and selectively updates noise-sensitive parameters to bolster local robustness against label noise, and (2) Adaptive federated LoRA aggregation, which dynamically weights and aggregates LoRA updates based on their importance and stability to minimize bias and noise propagation. Comprehensive experimental validation shows RFedLR outperforms existing methods, achieving superior accuracy and robustness in noisy federated scenarios. Our code and models will be released.



Paperid:1887
Authors:Tien Manh Luong, Khai Nguyen, Dinh Phung, Reza Haffari, Lizhen Qu
Abstract:
Abstract:Audio captioning systems face a fundamental challenge: teacher-forcing training creates exposure bias that leads to caption degeneration during inference. While contrastive methods have been proposed as solutions, they typically fail to capture the crucial temporal relationships between acoustic and linguistic modalities. We address this limitation by introducing the unbiased sliced Wasserstein RBF (USW-RBF) kernel with rotary positional embedding, specifically designed to preserve temporal information across modalities. Our approach offers both theoretical and practical advantages: the kernel enables efficient stochastic gradient optimization with a provable Monte Carlo approximation error rate of $\mathcal{O}(L^{-1/2})$, making it computationally feasible for real-world applications. Building on this foundation, we develop a complete audio captioning framework that integrates stochastic decoding to further mitigate caption degeneration. Extensive experiments on AudioCaps and Clotho datasets demonstrate that our method significantly improves caption quality, lexical diversity, and text-to-audio retrieval accuracy. Furthermore, we demonstrate the generalizability of our USW-RBF kernel by applying it to audio reasoning tasks, where it enhances the reasoning capabilities of large audio language models on the CompA-R in terms of correctness and quality. Our kernel also improves the reasoning accuracy of the MMAU-test-mini benchmarks by $4\\%$. These results establish our approach as a powerful and generalizable solution for cross-modal alignment challenges in audio-language tasks.



Paperid:1888
Authors:Pengyue Jia, Seongheon Park, Song Gao, Xiangyu Zhao, Sharon Li
Abstract:
Worldwide image geolocalization—the task of predicting GPS coordinates from images taken anywhere on Earth—poses a fundamental challenge due to the vast diversity in visual content across regions. While recent approaches adopt a two-stage pipeline of retrieving candidates and selecting the best match, they typically rely on simplistic similarity heuristics and point-wise supervision, failing to model spatial relationships among candidates. In this paper, we proposeGeoRanker, a distance-aware ranking framework that leverages large vision-language models to jointly encode query–candidate interactions and predict geographic proximity. In addition, we introduce amulti-order distance lossthat ranks both absolute and relative distances, enabling the model to reason over structured spatial relationships. To support this, we curate GeoRanking, the first dataset explicitly designed for geographic ranking tasks with multimodal candidate information. GeoRanker achieves state-of-the-art results on two well-established benchmarks (IM2GPS3K and YFCC4K), significantly outperforming current best methods. We also release our code, checkpoint, and dataset online for ease of reproduction.



Paperid:1889
Authors:Ilgee Hong, Changlong Yu, Liang Qiu, Weixiang Yan, Zhenghao Xu, Haoming Jiang, Qingru Zhang, Qin Lu, Xin Liu, Chao Zhang, Tuo Zhao
Abstract:
Reinforcement learning from human feedback (RLHF) has become a powerful post-training paradigm for aligning large language models with human preferences. A core challenge in RLHF is constructing accurate reward signals, where the conventional Bradley-Terry reward models (BT RMs) often suffer from sensitivity to data size and coverage, as well as vulnerability to reward hacking. Generative reward models (GenRMs) offer a more robust alternative by generating chain-of-thought (CoT) rationales followed by a final reward. However, existing GenRMs rely on shallow, vertically scaled reasoning, limiting their capacity to handle nuanced or complex (e.g., reasoning-intensive) tasks. Moreover, their pairwise preference outputs are incompatible with standard RLHF algorithms that require pointwise reward signals. In this work, we introduce Think-RM, a training framework that enables long-horizon reasoning in GenRMs by modeling an internal thinking process. Rather than producing structured, externally provided rationales, Think-RM generates flexible, self-guided reasoning traces that support advanced capabilities such as self-reflection, hypothetical reasoning, and divergent reasoning. To elicit these reasoning abilities, we first warm-up the models by supervised fine-tuning (SFT) over long CoT data. We then further improve the model's long-horizon abilities by rule-based reinforcement learning (RL). In addition, we propose a novel pairwise RLHF pipeline that directly optimizes policies using pairwise preference rewards, eliminating the need for pointwise reward conversion and enabling more effective use of Think-RM outputs. Experiments show that Think-RM achieves state-of-the-art results on RM-Bench, outperforming both BT RM and vertically scaled GenRM by 8\%. When combined with our pairwise RLHF pipeline, it demonstrates superior end-policy performance compared to traditional approaches. This depth-oriented approach not only broadens the GenRM design space but also establishes a new paradigm for preference-based policy optimization in RLHF.



Paperid:1890
Authors:Yongqi Huang, Jitao Zhao, Dongxiao He, Xiaobao Wang, Yawen Li, Yuxiao Huang, Di Jin, Zhiyong Feng
Abstract:
Graph Prompt Learning (GPL) has emerged as a promising paradigm that bridges graph pretraining models and downstream scenarios, mitigating label dependency and the misalignment between upstream pretraining and downstream tasks. Although existing GPL studies explore various prompt strategies, their effectiveness and underlying principles remain unclear. We identify two critical limitations: (1) Lack of consensus on underlying mechanisms: Despite current GPLs have advanced the field, there is no consensus on how prompts interact with pretrained models, as different strategies intervene at varying spaces within the model, i.e., input-level, layer-wise, and representation-level prompts. (2) Limited scenario adaptability: Most methods fail to generalize across diverse downstream scenarios, especially under data distribution shifts (e.g., homophilic-to-heterophilic graphs). To address these issues, we theoretically analyze existing GPL approaches and reveal that representation-level prompts essentially function as fine-tuning a simple downstream classifier, proposing that graph prompt learning should focus on unleashing the capability of pretrained models, and the classifier adapts to downstream scenarios. Based on our findings, we propose UniPrompt, a novel GPL method that adapts any pretrained models, unleashing the capability of pretrained models while preserving the structure of the input graph. Extensive experiments demonstrate that our method can effectively integrate with various pretrained models and achieve strong performance across in-domain and cross-domain scenarios.



Paperid:1891
Authors:Yizhao Gao, Zhichen Zeng, DaYou Du, Shijie Cao, Peiyuan Zhou, Jiaxing Qi, Junjie Lai, Hayden Kwok-Hay So, Ting Cao, Fan Yang, Mao Yang
Abstract:
Attention is the cornerstone of modern Large Language Models (LLMs). Yet its quadratic complexity hinders efficiency and scalability, especially for long-context processing. A promising approach is to leverage sparsity in attention. However, existing sparsity-based solutions predominantly rely on predefined patterns or heuristics at the attention head level, struggling to adapt dynamically to different contexts efficiently.We propose SeerAttention, a simple yet effective attention mechanism that directly learns the block-level attention sparsity from the LLM itself.Inspired by the gating mechanism in Mixture of Experts (MoE), SeerAttention augments the conventional attention with alearnable gatethatselectively activates important blockswithin the attention map.Specifically, the gate first pools the query (Q) and key (K) tensors along the sequence dimension and processes them through learnable linear layers. The resulting matrices are then multiplied together to produce the gating scores, which are used to predict block-level attention sparsity.Combined with our block-sparse FlashAttention kernel, SeerAttention can achieve significant speedup on GPUs. When applied to pre-trained LLMs, SeerAttention only requires training the gate parameters in a lightweight self-distillation manner, allowing rapid convergence.Our evaluation results demonstrate that SeerAttention achieves better model accuracy and lower latency for long-context pre-filling compared to prior methods.



Paperid:1892
Authors:Qingyue Yang, Jie Wang, Xing Li, Zhihai Wang, Chen Chen, Lei Chen, Xianzhi Yu, Wulong Liu, Jianye Hao, Mingxuan Yuan, Bin Li
Abstract:
Abstract:With the development of large language models (LLMs), efficient inference through Key-Value (KV) cache compression has attracted considerable attention, especially for long-context generation.To compress the KV cache, recent methods identify critical KV tokens through static modeling of attention scores. However, these methods often struggle to accurately determine critical tokens as they neglect the *temporal patterns* in attention scores, resulting in a noticeable degradation in LLM performance. To address this challenge, we propose **AttentionPredictor**, which is the **first learning-based KV cache compression and critical token identification method with direct attention pattern prediction**. Specifically, AttentionPredictor learns a lightweight, unified convolution model to dynamically capture spatiotemporal patterns and predict the next-token attention scores. An appealing feature of AttentionPredictor is that it accurately predicts the attention score and shares the unified prediction model, which consumes negligible memory, among all transformer layers. Moreover, we propose a cross-token critical cache prefetching framework that hides the token estimation time overhead to accelerate the decoding stage. By retaining most of the attention information, AttentionPredictor achieves **13$\times$** KV cache compression and **5.6$\times$** speedup with comparable LLM performance, significantly outperforming the state-of-the-arts.



Paperid:1893
Authors:Zhuo Han, Yi Feng, Yi Yang, Wanhong Huang, Ding Xuxing, Chuanyi Li, Jidong Ge, Vincent Ng
Abstract:
Legal Judgment Prediction (LJP) seeks to predict case outcomes given available case information, offering practical value for both legal professionals and laypersons. However, a key limitation of existing LJP models is their limited adaptability to statutory revisions. Current SOTA models are neither designed nor evaluated for statutory revisions. To bridge this gap, we introduce LawShift, a benchmark dataset for evaluating LJP under statutory revisions. Covering 31 fine-grained change types, LawShift enables systematic assessment of SOTA models' ability to handle legal changes. We evaluate five representative SOTA models on LawShift, uncovering significant limitations in their response to legal updates. Our findings show that model architecture plays a critical role in adaptability, offering actionable insights and guiding future research on LJP in dynamic legal contexts.



Paperid:1894
Authors:Payam Dibaeinia, Chris German, Suyash Shringarpure, Adam Auton, Aly A. Khan
Abstract:
Abstract:Predicting disease risk from DNA presents an unprecedented emerging challenge as biobanks approach population scale sizes ($N>10^6$ individuals) with ultra-high-dimensional features ($L>10^5$ genotypes). Current methods, often linear and reliant on summary statistics, fail to capture complex genetic interactions and discard valuable individual-level information. We introduce PRSformer, a scalable deep learning architecture designed for end-to-end, multitask disease prediction directly from million-scale individual genotypes. PRSformer employs neighborhood attention, achieving linear $O(L)$ complexity per layer, making Transformers tractable for genome-scale inputs. Crucially, PRSformer utilizes a stacking of these efficient attention layers, progressively increasing the effective receptive field to model local dependencies (e.g., within linkage disequilibrium blocks) before integrating information across wider genomic regions. This design, tailored for genomics, allows PRSformer to learn complex, potentially non-linear and long-range interactions directly from raw genotypes. We demonstrate PRSformer's effectiveness using a unique large private cohort ($N \approx 5$M) for predicting 18 autoimmune diseases using $L \approx 140$k variants. PRSformer significantly outperforms highly optimized linear models trained on the same individual-level data and state-of-the-art summary-statistic-based methods (LDPred2) derived from the same cohort, quantifying the benefits of non-linear modeling and multitask learning at scale. Furthermore, experiments reveal that the advantage of non-linearity emerges primarily at large sample sizes ($N > 1$M), and a multi-ancestry trained PRSformer shows improved generalization to diverse populations. PRSformer establishes a new framework for deep learning in population-scale genomics and offers insights into modeling ultra-long genomic sequences.



Paperid:1895
Authors:Mikey Shechter, Yair Carmon
Abstract:
We introduce Filter Like You Test (FLYT), an algorithm for curating large-scale vision-language datasets thatlearnsthe usefulness of each data point as a pretraining example. FLYT trains a scoring model that learns to weigh each example's features using gradient signals from downstream tasks training sets. Based on FLYT, we implement Mixing-FLYT (M-FLYT), which takes the per-example scores generated by different scoring methods as features, and learns to unify them into a single score. FLYT naturally produces a distribution over the training examples, which we leverage through Soft Cap Sampling (SCS), a strategy for obtaining a filtered pretraining dataset from per-example probabilities that samples examples while preventing over-representation through a repetition penalty. Using these methods, we achieve 40.1\% ImageNet zero-shot accuracy on the DataComp medium scale filtering benchmark, a 2\% absolute accuracy increase over all previous results and a 5.5\% increase over results that---like us---use only public resources. Our approach also yields 37.7\% on the average of 38 DataComp evaluation tasks, outperforming previous public-resource approaches by 0.4\%.



Paperid:1896
Authors:Di Wen, Yu Wang, Zhigang Wu, Zhaocheng He, Zhe Wu, Zheng Qingfang
Abstract:
Vehicle trajectory prediction serves as a critical enabler for autonomous navigation and intelligent transportation systems. While existing approaches predominantly focus on temporal pattern extraction and vehicle-environment interaction modeling, they exhibit a fundamental limitation in addressing trajectory heterogeneity originating from human driving styles. This oversight constrains prediction reliability in complex real-world scenarios.To bridge this gap, we propose the Driving-Style-Adaptive (\underline{\textbf{DSA}}) framework, which establishes the first systematic integration of heterogeneous driving behaviors into trajectory prediction models. Specifically, our framework employs a set of basis functions tailored to each driving style to approximate the trajectory patterns. By dynamically combining and adaptively adjusting the degree of these basis functions, DSA not only enhances prediction accuracy but also provides \textbf{explanations} insights into the prediction process. Extensive experiments on public real-world datasets demonstrate that the DSA framework outperforms state-of-the-art methods.



Paperid:1897
Authors:Xiaodong Wang, Zijun He, Ping Wang, Lishun Wang, Yanan Hu, Xin Yuan
Abstract:
In coded aperture snapshot spectral imaging (CASSI), the captured measurement entangles spatial and spectral information, posing a severely ill-posed inverse problem for hyperspectral images (HSIs) reconstruction. Moreover, the captured radiance inherently depends on scene illumination, making it difficult to recover the intrinsic spectral reflectance that remains invariant to lighting conditions. To address these challenges, we propose a \textbf{chromaticity-intensity decomposition framework}, which disentangles an HSI into a spatially smooth intensity map and a spectrally variant chromaticity cube. The chromaticity encodes lighting-invariant reflectance, enriched with high-frequency spatial details and local spectral sparsity. Building on this decomposition, we develop \textbf{CIDNet}—a Chromaticity-Intensity Decomposition unfolding network within a dual-camera CASSI system. CIDNet integrates a hybrid spatial-spectral Transformer tailored to reconstruct fine-grained and sparse spectral chromaticity and a degradation-aware, spatially-adaptive noise estimation module that captures anisotropic noise across iterative stages. Extensive experiments on both synthetic and real-world CASSI datasets demonstrate that our method achieves superior performance in both spectral and chromaticity fidelity. Code and models will be publicly available.



Paperid:1898
Authors:Divin Yan, Shengzhong Zhang, Bisheng Li, Menglin Yang, Chen Yang, Min Zhou, Weiyang Ding, Yutong Xie, zengfeng Huang
Abstract:
Class imbalance in graph data poses a major challenge for effective node classification, particularly in semi-supervised settings. In this work, we formally introduce the concept of geometric imbalance, characterizing how class imbalance induces geometric ambiguity among minority nodes in the embedding space. We provide a rigorous theoretical analysis and propose a unified framework to explicitly mitigate geometric imbalance through pseudo-label alignment, node reordering, and ambiguity filtering. Extensive experiments on diverse benchmarks show that our approach consistently outperforms existing methods, especially under severe class imbalance. Our findings offer new theoretical insights and practical tools for robust semi-supervised imbalanced node classification. The detailed code implementation of this work is provided in the supplementary material.



Paperid:1899
Authors:Hongchi Xia, Chih-Hao Lin, Hao-Yu Hsu, Quentin Leboutet, Katelyn Gao, Michael Paulitsch, Benjamin Ummenhofer, Shenlong Wang
Abstract:
Digitizing the physical world into accurate simulation‑ready virtual environments offers significant opportunities in a variety of fields such as augmented and virtual reality, gaming, and robotics. However, current 3D reconstruction and scene-understanding methods commonly fall short in one or more critical aspects, such as geometry completeness, object interactivity, physical plausibility, photorealistic rendering, or realistic physical properties for reliable dynamic simulation. To address these limitations, we introduce HoloScene, a novel interactive 3D reconstruction framework that simultaneously achieves these requirements. HoloScene leverages a comprehensive interactive scene-graph representation, encoding object geometry, appearance, and physical properties alongside hierarchical and inter-object relationships. Reconstruction is formulated as an energy-based optimization problem, integrating observational data, physical constraints, and generative priors into a unified, coherent objective. Optimization is efficiently performed via a hybrid approach combining sampling-based exploration with gradient-based refinement. The resulting digital twins exhibit complete and precise geometry, physical stability, and realistic rendering from novel viewpoints. Evaluations conducted on multiple benchmark datasets demonstrate superior performance, while practical use-cases in interactive gaming and real-time digital-twin manipulation illustrate HoloScene's broad applicability and effectiveness.



Paperid:1900
Authors:Sehyun Park, Jongjin Lee, Yunseop Shin, Ilsang Ohn, Yongdai Kim
Abstract:
Deep ensembles deliver state-of-the-art, reliable uncertainty quantification, but their heavy computational and memory requirements hinder their practical deployments to real applications such as on-device AI. Knowledge distillation compresses an ensemble into small student models, but existing techniques struggle to preserve uncertainty partly because reducing the size of DNNs typically results in variation reduction. To resolve this limitation, we introduce a new method of distribution distillation (i.e. compressing a teacher ensemble into a student distribution instead of a student ensemble) called Gaussian distillation, which estimates the distribution of a teacher ensemble through a special Gaussian process called the deep latent factor model (DLF) by treating each member of the teacher ensemble as a realization of a certain stochastic process. The mean and covariance functions in the DLF model are estimated stably by use of the expectation-maximization (EM) algorithm. By using multiple benchmark datasets, we demonstrate that the proposed Gaussian distillation outperforms existing baselines. In addition, we illustrate that Gaussian distillation works well for fine-tuning of LLMs and distribution shift problems.



Paperid:1901
Authors:Naifu Xue, Zhaoyang Jia, Jiahao Li, Bin Li, Yuan Zhang, Yan Lu
Abstract:
Abstract:While recent diffusion-based generative image codecs have shown impressive performance, their iterative sampling process introduces unpleasing latency. In this work, we revisit the design of a diffusion-based codec and argue that multi-step sampling is not necessary for generative compression. Based on this insight, we propose OneDC, a One-step Diffusion-based generative image Codec—that integrates a latent compression module with a one-step diffusion generator. Recognizing the critical role of semantic guidance in one-step diffusion, we propose using the hyperprior as a semantic signal, overcoming the limitations of text prompts in representing complex visual content. To further enhance the semantic capability of the hyperprior, we introduce a semantic distillation mechanism that transfers knowledge from a pretrained generative tokenizer to the hyperprior codec. Additionally, we adopt a hybrid pixel- and latent-domain optimization to jointly enhance both reconstruction fidelity and perceptual realism. Extensive experiments demonstrate that OneDC achieves SOTA perceptual quality even with one-step generation, offering over 40% bitrate reduction and 20$\times$ faster decoding compared to prior multi-step diffusion-based codecs. Code will be released later.



Paperid:1902
Authors:Artun Saday, Y. Cahit Yıldırım, Cem Tekin
Abstract:
Abstract:We address the problem of Gaussian Process (GP) optimization in the presence of unknown and potentially varying adversarial perturbations. Unlike traditional robust optimization approaches that focus on maximizing performance under worst-case scenarios, we consider a robust satisficing objective, where the goal is to consistently achieve a predefined performance threshold $\tau$, even under adversarial conditions. We propose two novel algorithms based on distinct formulations of robust satisficing, and show that they are instances of a general robust satisficing framework. Further, each algorithm offers different guarantees depending on the nature of the adversary. Specifically, we derive two regret bounds: one that is sublinear over time, assuming certain conditions on the adversary and the satisficing threshold $\tau$, and another that scales with the perturbation magnitude but requires no assumptions on the adversary. Through extensive experiments, we demonstrate that our approach outperforms the established robust optimization methods in achieving the satisficing objective, particularly when the ambiguity set of the robust optimization framework is inaccurately specified.



Paperid:1903
Authors:Wanjia Zhao, Mert Yuksekgonul, Shirley Wu, James Zou
Abstract:
Multi-agent AI systems powered by large language models (LLMs) are increasingly applied to solve complex tasks. However, these systems often rely on fragile, manually designed prompts and heuristics, making optimization difficult. A key challenge in optimizing multi-agent systems is acquiring suitable training data for specialized agents. We introduce SiriuS, a self-improving, reasoning-driven optimization framework for multi-agent systems. Central to our approach is the construction of an experience library: a repository of high-quality reasoning trajectories. The library is built by retaining reasoning steps that lead to successful outcomes, providing a robust training set for optimizing multi-agent system. Additionally, we introduce a library augmentation procedure that refines unsuccessful trajectories, further enriching the library. SiriuS boosts performance by 2.86% to 21.88% on reasoning and biomedical QA and enhances agent negotiation in competitive settings. Our results show that SiriuS enhances multi-agent performance while generating reusable data for self-correction and self-play enhancement in the future.



Paperid:1904
Authors:Jingjing Chang, Yixiao Fang, Peng Xing, Shuhan Wu, Wei Cheng, Rui Wang, Xianfang Zeng, Gang Yu, Hai-Bao Chen
Abstract:
Abstract:Text-to-image (T2I) models have garnered significant attention for generating high-quality images aligned with text prompts. However, rapid T2I model advancements reveal limitations in early benchmarks, lacking comprehensive evaluations, especially for text rendering and style. Notably, recent state-of-the-art models, with their rich knowledge modeling capabilities, show potential in reasoning-driven image generation, yet existing evaluation systems have not adequately addressed this frontier. To systematically address these gaps, we introduce $\textbf{OneIG-Bench}$, a meticulously designed comprehensive benchmark framework for fine-grained evaluation of T2I models across multiple dimensions, including subject-element alignment, text rendering precision, reasoning-generated content, stylization, and diversity. By structuring the evaluation, this benchmark enables in-depth analysis of model performance, helping researchers and practitioners pinpoint strengths and bottlenecks in the full pipeline of image generation. Our codebase and dataset are now publicly available to facilitate reproducible evaluation studies and cross-model comparisons within the T2I research community.



Paperid:1905
Authors:Stephen Obadinma, Xiaodan Zhu
Abstract:
Robust verbal confidence generated by large language models (LLMs) is crucial for the deployment of LLMs to ensure transparency, trust, and safety in human-AI interactions across many high-stakes applications. In this paper, we present the first comprehensive study on the robustness of verbal confidence under adversarial attacks. We introduce a novel framework for attacking verbal confidence scores through both perturbation and jailbreak-based methods, and show that these attacks can significantly jeopardize verbal confidence estimates and lead to frequent answer changes. We examine a variety of prompting strategies, model sizes, and application domains, revealing that current confidence elicitation methods are vulnerable and that commonly used defence techniques are largely ineffective or counterproductive. Our findings underscore the urgent need to design more robust mechanisms for confidence expression in LLMs, as even subtle semantic-preserving modifications can lead to misleading confidence in responses.



Paperid:1906
Authors:Yarden As, Chengrui Qu, Benjamin Unger, Dong Ho Kang, Max van der Hart, Laixi Shi, Stelian Coros, Adam Wierman, Andreas Krause
Abstract:
Safety remains a major concern for deploying reinforcement learning (RL) in real-world applications. Simulators provide safe, scalable training environments, but the inevitablesim-to-real gapintroduces additional safety concerns, as policies must satisfy constraints in real-world conditions that differ from simulation. To address this challenge, robust safe RL techniques offer principled methods, but are often incompatible with standard scalable training pipelines. In contrast,domain randomization, a simple and popular sim-to-real technique, stands out as a promising alternative, although it often results in unsafe behaviors in practice. We present SPiDR, short for Sim-to-real via Pessimistic Domain Randomization—a scalable algorithm with provable guarantees for safe sim-to-real transfer. SPiDR uses domain randomization to incorporate the uncertainty about the sim-to-real gap into the safety constraints, making it flexible and highly compatible with existing training pipelines. Through extensive experiments on sim-to-sim benchmarks and two distinct real-world robotic platforms, we demonstrate that SPiDR effectively ensures safety under sim-to-real gaps while maintaining strong performance.



Paperid:1907
Authors:Qijiong Liu, Jieming Zhu, Lu Fan, Kun Wang, Hengchang Hu, Wei Guo, Yong Liu, Xiao-Ming Wu
Abstract:
Integrating large language models (LLMs) into recommender systems has created new opportunities for improving recommendation quality. However, a comprehensive benchmark is needed to thoroughly evaluate and compare the recommendation capabilities of LLMs with traditional recommender systems. In this paper, we introduce \recbench{}, which systematically investigates various item representation forms (including unique identifier, text, semantic embedding, and semantic identifier) and evaluates two primary recommendation tasks, i.e., click-through rate prediction (CTR) and sequential recommendation (SeqRec). Our extensive experiments cover up to 17 large models and are conducted across five diverse datasets from fashion, news, video, books, and music domains. Our findings indicate that LLM-based recommenders outperform conventional recommenders, achieving up to a 5% AUC improvement in CTR and up to a 170% NDCG@10 improvement in SeqRec. However, these substantial performance gains come at the expense of significantly reduced inference efficiency, rendering LLMs impractical as real-time recommenders. We have released our code and data to enable other researchers to reproduce and build upon our experimental results.



Paperid:1908
Authors:Carl Hvarfner, David Eriksson, Eytan Bakshy, Maximilian Balandat
Abstract:
Bayesian Optimization (BO) is a widely used method for optimizing expensive black-box functions, relying on probabilistic surrogate models such as Gaussian Processes (GPs). The quality of the surrogate model is crucial for good optimization performance, especially in the few-shot setting where only a small number of batches of points can be evaluated. In this setting, the initialization plays a critical role in shaping the surrogate's predictive quality and guiding subsequent optimization. Despite this, practitioners typically rely on (quasi-)random designs to cover the input space. However, such approaches neglect two key factors: (a) random designs may not be space-filling, and (b) efficient hyperparameter learning during initialization is essential for high-quality prediction, which may conflict with space-filling designs. To address these limitations, we propose Hyperparameter-Informed Predictive Exploration (HIPE), a novel acquisition strategy that balances space-filling exploration with hyperparameter learning using information-theoretic principles. We derive a closed-form expression for HIPE in the GP setting and demonstrate its effectiveness through extensive experiments in active learning and few-shot BO. Our results show that HIPE outperforms standard initialization strategies in terms of predictive accuracy, hyperparameter identification, and optimization performance, particularly in large-batch, few-shot settings relevant to many real-world BO applications.



Paperid:1909
Authors:Ashwinee Panda, Vatsal Baherwani, Zain Sarwar, Benjamin Thérien, Sambit Sahu, Tom Goldstein, Supriyo Chakraborty
Abstract:
Mixture of Experts (MoE) pretraining is more scalable than dense Transformer pretraining, because MoEs learn to route inputs to a sparse set of their feedforward parameters. However, this means that MoEs only receive a sparse backward update, leading to training instability and suboptimal performance. We present a lightweight approximation method that gives the MoE router a dense gradient update while continuing to sparsely activate its parameters. Our method, which we refer to as Default MoE, substitutes missing expert activations with default outputs consisting of an exponential moving average of expert outputs previously seen over the course of training. This allows the router to receive signals from every expert for each token, leading to significant improvements in training performance. Our Default MoE outperforms standard TopK routing in a variety of settings without requiring significant computational overhead.



Paperid:1910
Authors:GUOGUO AI, Hezhe Qiao, Hui Yan, Guansong Pang
Abstract:
Semi-supervised graph anomaly detection (GAD) utilizes a small set of labeled normal nodes to identify abnormal nodes from a large set of unlabeled nodes in a graph. Current methods in this line posit that 1) normal nodes share a similar level of homophily and 2) the labeled normal nodes can well represent the homophily patterns in the normal class. However, this assumption often does not hold well since normal nodes in a graph can exhibit diverse homophily in real-world GAD datasets. In this paper, we propose RHO, namely Robust Homophily Learning, to adaptively learn such homophily patterns. RHO consists of two novel modules, adaptive frequency response filters (AdaFreq) and graph normality alignment (GNA). AdaFreq learns a set of adaptive spectral filters that capture different frequency components of the labeled normal nodes with varying homophily in the channel-wise and cross-channel views of node attributes. GNA is introduced to enforce consistency between the channel-wise and cross-channel homophily representations to robustify the normality learned by the filters in the two views. Experiments on eight real-world GAD datasets show that RHO can effectively learn varying, often under-represented, homophily in the small normal node set and substantially outperforms state-of-the-art competing methods. Code is available at \url{https://anonymous.4open.science/r/RHO-5C57}.



Paperid:1911
Authors:Kyuyoung Kim, Hyunjun Jeon, Jinwoo Shin
Abstract:
Abstract:Large language models (LLMs) are increasingly used in sensitive domains, where their ability to infer personal data from seemingly benign text poses emerging privacy risks. While recent LLM-based anonymization methods help mitigate such risks, they often rely on proprietary models (e.g., GPT-4), raising concerns about cost and the potential exposure of sensitive data to untrusted external systems. To address this, we introduce $\textit{SElf-refining Anonymization with Language model}$ (SEAL), a novel distillation framework for training small language models (SLMs) to perform effective anonymization without relying on external costly models at inference time. We leverage adversarial interactions between an LLM anonymizer and an inference model to collect trajectories of anonymized texts and inferred attributes, which are used to distill anonymization, adversarial inference, and utility evaluation capabilities into SLMs via supervised fine-tuning and preference learning. The resulting models learn to both anonymize text and critique their outputs, enabling iterative improvement of anonymization quality via self-refinement. Experiments on SynthPAI, a dataset of synthetic personal profiles and text comments, demonstrate that SLMs trained with SEAL achieve substantial improvements in anonymization capabilities. Notably, 8B models attain a privacy-utility trade-off comparable to that of the GPT-4 anonymizer and, with self-refinement, even surpass it in terms of privacy. These results show the effectiveness of our adversarial distillation framework in training SLMs as efficient anonymizers. To facilitate further research, we release the full dataset used in our experiments.



Paperid:1912
Authors:Deming Li, Kaiwen Jiang, Yutao Tang, Ravi Ramamoorthi, Rama Chellappa, Cheng Peng
Abstract:
In-the-wild photo collections often contain limited volumes of imagery and exhibit multiple appearances, e.g., taken at different times of day or seasons, posing significant challenges to scene reconstruction and novel view synthesis. Although recent adaptations of Neural Radiance Field (NeRF) and 3D Gaussian Splatting (3DGS) have improved in these areas, they tend to oversmooth and are prone to overfitting. In this paper, we present MS-GS, a novel framework designed with \textbf{M}ulti-appearance capabilities in \textbf{S}parse-view scenarios using 3D\textbf{GS}. To address the lack of support due to sparse initializations, our approach is built on the geometric priors elicited from monocular depth estimations. The key lies in extracting and utilizing local semantic regions with a Structure-from-Motion (SfM) points anchored algorithm for reliable alignment and geometry cues. Then, to introduce multi-view constraints, we propose a series of geometry-guided supervision at virtual views in a fine-grained and coarse scheme to encourage 3D consistency and reduce overfitting. We also introduce a dataset and an in-the-wild experiment setting to set up more realistic benchmarks. We demonstrate that MS-GS achieves photorealistic renderings under various challenging sparse-view and multi-appearance conditions and outperforms existing approaches significantly across different datasets.



Paperid:1913
Authors:Alberto Sinigaglia, Davide Sartor, Marina Ceccon, Gian Antonio Susto
Abstract:
Fair classification is a critical challenge that has gained increasing importance due to international regulations and its growing use in high-stakes decision-making settings.Existing methods often rely on adversarial learning or distribution matching across sensitive groups; however, adversarial learning can be unstable, and distribution matching can be computationally intensive.To address these limitations, we propose a novel approach based on the characteristic function distance. Our method ensures that the learned representation contains minimal sensitive information while maintaining high effectiveness for downstream tasks. By utilizing characteristic functions, we achieve a more stable and efficient solution compared to traditional methods.Additionally, we introduce a simple relaxation of the objective function that guarantees fairness in common classification models with no performance degradation. Experimental results on benchmark datasets demonstrate that our approach consistently matches or achieves better fairness and predictive accuracy than existing methods. Moreover, our method maintains robustness and computational efficiency, making it a practical solution for real-world applications.



Paperid:1914
Authors:Jiahui Wang, Chao Chen, Jiacheng Xu, Zongzhang Zhang, Yang Yu
Abstract:
Visual reinforcement learning has shown promise in various real-world applications. However, deploying policies in complex real-world environments with visual perturbations remains a significant challenge. We notice that humans tend to filter information at the object level prior to decision-making, facilitating efficient skill transfer across different contexts. Inspired by this, we introduce Focus-Then-Reuse (FTR), a method utilizing a novel object selection mechanism to focus on task-relevant objects, and directly reuse the simulation-trained policy on them. The training of the object selection mechanism integrates prior knowledge from a vision-language model and feedback from the environment. Experimental results on challenging tasks based on DeepMind Control Suite and Franka Emika Robotics demonstrate that FTR enables rapid adaptation in visual perturbation environments and achieves state-of-the-art performance.



Paperid:1915
Authors:EN YU, Jie Lu, Xiaoyu Yang, Guangquan Zhang, Zhen Fang
Abstract:
Modern machine learning models struggle to maintain performance in dynamic environments where temporal distribution shifts, \emph{i.e., concept drift}, are prevalent. Temporal Domain Generalization (TDG) seeks to enable model generalization across evolving domains, yet existing approaches typically assume smooth incremental changes, struggling with complex real-world drifts involving long-term structure (incremental evolution/periodicity) and local uncertainties. To overcome these limitations, we introduce FreKoo, which tackles these challenges via a novel frequency-domain analysis of parameter trajectories. It leverages the Fourier transform to disentangle parameter evolution into distinct spectral bands. Specifically, low-frequency component with dominant dynamics are learned and extrapolated using the Koopman operator, robustly capturing diverse drift patterns including both incremental and periodicity. Simultaneously, potentially disruptive high-frequency variations are smoothed via targeted temporal regularization, preventing overfitting to transient noise and domain uncertainties. In addition, this dual spectral strategy is rigorously grounded through theoretical analysis, providing stability guarantees for the Koopman prediction, a principled Bayesian justification for the high-frequency regularization, and culminating in a multiscale generalization bound connecting spectral dynamics to improved generalization. Extensive experiments demonstrate FreKoo's significant superiority over SOTA TDG approaches, particularly excelling in real-world streaming scenarios with complex drifts and uncertainties.



Paperid:1916
Authors:Daniel Jarne Ornia, Nicholas Bishop, Joel Dyer, Wei-Chen Lee, Anisoara Calinescu, Doyne Farmer, Michael Wooldridge
Abstract:
Advanced reasoning models with agentic capabilities (AI agents) are deployed to interact with humans and to solve sequential decision‑making problems under (often approximate) utility functions and internal models. When such problems have resource or failure constraints where action sequences may be forcibly terminated once resources are exhausted, agents face implicit trade‑offs that reshape their utility-driven (rational) behaviour. Additionally, since these agents are typically commissioned by a human principal to act on their behalf, asymmetries in constraint exposure can give rise to previously unanticipated misalignment between human objectives and agent incentives. We formalise this setting through a survival bandit framework, provide theoretical and empirical results that quantify the impact of survival‑driven preference shifts, identify conditions under which misalignment emerges and propose mechanisms to mitigate the emergence of risk-seeking or risk-averse behaviours. As a result, this work aims to increase understanding and interpretability of emergent behaviours of AI agents operating under such survival pressure, and offer guidelines for safely deploying such AI systems in critical resource‑limited environments.



Paperid:1917
Authors:Yunwei Ren, Eshaan Nichani, Denny Wu, Jason Lee
Abstract:
Abstract:We study the complexity of online stochastic gradient descent (SGD) for learning a two-layer neural network with $P$ neurons on isotropic Gaussian data: $f_*(\boldsymbol{x}) = \sum_{p=1}^P a_p\cdot \sigma(\langle\boldsymbol{x},\boldsymbol{v_p}^{\star}\rangle)$, $\boldsymbol{x} \sim \mathcal{N}(0,\boldsymbol{I_d})$, where the activation $\sigma:\mathbb{R}\to\mathbb{R}$ is an even function with information exponent $k>2$ (defined as the lowest degree in the Hermite expansion), $\lbrace\boldsymbol{v_p}^\star\rbrace_{p\in[P]}\subset \mathbb{R}^d$ are orthonormal signal directions, and the non-negative second-layer coefficients satisfy $\sum_{p} a_p^2=1$. We focus on the challenging extensive-width regime $P\gg 1$ and permit diverging condition number in the second-layer, covering as a special case the power-law scaling $a_p\asymp p^{-\beta}$ where $\beta\in\mathbb{R}_{\ge 0}$. We provide a precise analysis of SGD dynamics for the training of a student two-layer network to minimize the mean squared error (MSE) objective, and explicitly identify sharp transition times to recover each signal direction. In the power-law setting, we characterize scaling law exponents for the MSE loss with respect to the number of training samples and SGD steps, as well as the number of parameters in the student neural network. Our analysis entails that while the learning of individual teacher neurons exhibits abrupt transitions, the juxtaposition of $P\gg 1$ emergent learning curves at different timescales leads to a smooth scaling law in the cumulative objective.



Paperid:1918
Authors:Xiangyu Hong, Che Jiang, Kai Tian, Biqing Qi, Youbang Sun, Ning Ding, Bowen Zhou
Abstract:
Attributing the behavior of Transformer models to internal computations is a central challenge in mechanistic interpretability. We introduce DePass, a unified framework for feature attribution based on a single decomposed forward pass. DePass decomposes hidden states into customized additive components, then propagates them with attention scores and MLP's activations fixed. It achieves faithful, fine-grained attribution without requiring auxiliary training. We validate DePass across token-level, model component-level, and subspace-level attribution tasks, demonstrating its effectiveness and fidelity. Our experiments highlight its potential to attribute information flow between arbitrary components of a Transformer model. We hope DePass serves as a foundational tool for broader applications in interpretability.



Paperid:1919
Authors:Haoran Zhang, Yang Xu, Xuchuang Wang, Hao-Xu Chen, Hao Qiu, Lin Yang, Yang Gao
Abstract:
Abstract:In this work, we study the federated multi-armed bandit (FMAB) problem, where a set of distributed agents collaboratively aim to minimize cumulative regret while interacting with a shared set of arms. Unlike traditional centralized bandit models, agents in FMAB settings are connected via a communication graph and cannot share data freely due to bandwidth limitations or privacy constraints. This raises a fundamental challenge: how to achieve optimal learning performance under stringent communication budgets.We propose a novel communication-efficient algorithm that decouples the learning process into two phases: one for eliminating suboptimal arms through early and frequent communication of key decisions, and another for refining global estimates using buffered, quantized, and differentially transmitted statistics. By carefully balancing the communication frequency and precision of shared information, our algorithm achieves the optimal individual regret bound $O(N^{-1}\log T)$ while significantly reducing the total number of communication rounds and transmitted bits.Theoretically, we derive tight upper bounds on both individual cumulative regret and group regret, and prove that our method asymptotically matches the lower bound of regret in federated settings. Experimental results on synthetic data validate the effectiveness of the proposed approach in various graph topologies and under heterogeneous feedback.



Paperid:1920
Authors:Vinh Tong, Trung-Dung Hoang, Anji Liu, Guy Van den Broeck, Mathias Niepert
Abstract:
In domains such as molecular and protein generation, physical systems exhibit inherent symmetries that are critical to model. Two main strategies have emerged for learning invariant distributions: designing equivariant network architectures and using data augmentation to approximate equivariance. While equivariant architectures preserve symmetry by design, they often involve greater complexity and pose optimization challenges. Data augmentation, on the other hand, offers flexibility but may fall short in fully capturing symmetries. Our framework enhances both approaches by reducing training variance and providing a provably lower-variance gradient estimator. We achieve this by interpreting data augmentation as a Monte Carlo estimator of the training gradient and applying Rao–Blackwellization. This leads to more stable optimization, faster convergence, and reduced variance, all while requiring only a single forward and backward pass per sample. We also present a practical implementation of this estimator—incorporating the loss and sampling procedure—through a method we call Orbit Diffusion. Theoretically, we guarantee that our loss admits equivariant minimizers. Empirically, Orbit Diffusion achieves state-of-the-art results on GEOM-QM9 for molecular conformation generation, improves crystal structure prediction, and advances text-guided crystal generation on the Perov-5 and MP-20 benchmarks. Additionally, it enhances protein designability in protein structure generation.



Paperid:1921
Authors:Milena Rmus, Akshay Kumar Jagadish, Marvin Mathony, Tobias Ludwig, Eric Schulz
Abstract:
Computational cognitive models, which formalize theories of cognition, enable researchers to quantify cognitive processes and arbitrate between competing theories by fitting models to behavioral data. Traditionally, these models are handcrafted, which requires significant domain knowledge, coding expertise, and time investment. However, recent advances in machine learning offer solutions to these challenges. In particular, Large Language Models (LLMs) have demonstrated remarkable capabilities for in-context pattern recognition, leveraging knowledge from diverse domains to solve complex problems, and generating executable code that can be used to facilitate the generation of cognitive models. Building on this potential, we introduce a pipeline for Guided generation of Computational Cognitive Models (GeCCo). Given task instructions, participant data, and a template function, GeCCo prompts an LLM to propose candidate models, fits proposals to held-out data, and iteratively refines them based on feedback constructed from their predictive performance. We benchmark this approach across four different cognitive domains -- decision making, learning, planning, and memory -- using three open-source LLMs, spanning different model sizes, capacities, and families. On four human behavioral data sets, the LLM generated models that consistently matched or outperformed the best domain-specific models from the cognitive science literature. To validate these findings, we performed control experiments that investigated (1) the contribution of the different LLM features (model size, model family, capacities); (2) the causal role of different prompt components; (3) the effect of data contamination; (4) the ability to recover ground truth models from simulated data; and (5) the total explainable variance in human behavior captured by LLM-generated models. Taken together, our results suggest that LLMs can rapidly generate cognitive models with conceptually plausible theories that rival -- or even surpass -- the best models from the literature across diverse task domains.



Paperid:1922
Authors:Yasmine Briefs, Markus Bläser
Abstract:
Abstract:Linear structural causal models (SCMs) are used to analyze the relationships between random variables. Directed edges represent direct causal effects and bidirected edges represent hidden confounders. Generically identifying the causal parameters from observed correlations between the random variables is an open problem in causality. Gupta and Bl\"aser solve the case of SCMs in which the directed edges form a tree by giving a randomized polynomial time algorithm with running time $O(n^6)$. We present an improved algorithm with running time $O(n^3 \log^2 n)$ and demonstrate its feasibility by providing an implementation that outperforms existing state-of-the-art implementations.



Paperid:1923
Authors:Berk Tinaz, Zalan Fabian, Mahdi Soltanolkotabi
Abstract:
Diffusion models have become the go-to method for text-to-image generation, producing high-quality images from pure noise. However, the inner workings of diffusion models is still largely a mystery due to their black-box nature and complex, multi-step generation process. Mechanistic interpretability techniques, such as Sparse Autoencoders (SAEs), have been successful in understanding and steering the behavior of large language models at scale. However, the great potential of SAEs has not yet been applied toward gaining insight into the intricate generative process of diffusion models. In this work, we leverage the SAE framework to probe the inner workings of a popular text-to-image diffusion model, and uncover a variety of human-interpretable concepts in its activations. Interestingly, we find thateven before the first reverse diffusion stepis completed, the final composition of the scene can be predicted surprisingly well by looking at the spatial distribution of activated concepts. Moreover, going beyond correlational analysis, we design intervention techniques aimed at manipulating image composition and style, and demonstrate that (1) in early stages of diffusion image composition can be effectively controlled, (2) in the middle stages image composition is finalized, however stylistic interventions are effective, and (3) in the final stages only minor textural details are subject to change.



Paperid:1924
Authors:Guanghui Min, Yinhan He, Chen Chen
Abstract:
Computational epidemiology is crucial in understanding and controlling infectious diseases, as highlighted by large-scale outbreaks such as COVID-19. Given the inherent uncertainty and variability of disease spread, Monte Carlo (MC) simulations are widely used to predict infection peaks, estimate reproduction numbers, and evaluate the impact of non-pharmaceutical interventions (NPIs). While effective, MC-based methods require numerous runs to achieve statistically reliable estimates and variance, which suffer from high computational costs. In this work, we present a unified theoretical framework for analyzing disease spread dynamics on both directed and undirected contact networks, and propose an algorithm,RAPID, that significantly improves computational efficiency. Our contributions are threefold. First, we derive a variance lower bound for MC estimates and identify the key factors influencing estimation variance. Second, we provide a theoretical analysis of the probabilistic disease spread process using linear approximations and derive the convergence conditions under non-reinfection epidemic models. Finally, we conduct extensive experiments on six real-world datasets, demonstrating our method's effectiveness and robustness in estimating the nodes' final state distribution. Specifically, our proposed method consistently produces accurate estimates aligned with results from a large number of MC simulations, while maintaining a runtime comparable to a single MC simulation. Our code and datasets are available at https://anonymous.4open.science/r/rapid-8080.



Paperid:1925
Authors:Romy Luo, Zihui (Sherry) Xue, Alex Dimakis, Kristen Grauman
Abstract:
Video reasoning, the task of enabling machines to infer from dynamic visual content through multi-step logic, is crucial for advanced AI. While the Chain-of-Thought (CoT) mechanism has enhanced reasoning in text-based tasks, its application to video understanding remains underexplored. This paper presents a systematic analysis revealing that CoT often degrades performance in video reasoning, generating verbose but misleading internal monologues, and leading to hallucinated visual details and overridden correct intuitions—a phenomenon we term "visual thinking drift." We explain this drift through a Bayesian lens, positing that CoT traces often diverge from actual visual evidence, instead amplifying internal biases or language priors, causing models to storytell rather than engage in grounded reasoning. To counteract this, we introduce Visual Evidence Reward (VER), a novel reinforcement learning framework that explicitly rewards the generation of reasoning traces that are verifiably grounded in visual evidence. Comprehensive evaluation across 10 diverse video understanding benchmarks demonstrates that our Video-VER model consistently achieves top performance. Our work sheds light on the distinct challenges of video-centric reasoning and encourages the development of AI that robustly grounds its inferences in visual evidence---for large multimodal models that not only "think before answering", but also "see while thinking".



Paperid:1926
Authors:Shuguang Wang, Qian Zhou, Kui Wu, Dapeng Wu, Wei-Bin Lee, Jianping Wang
Abstract:
Safety validation, which assesses the safety of an autonomous system's motion planning decisions, is critical for the safe deployment of autonomous vehicles. Existing input validation techniques from other machine learning domains, such as image classification, face unique challenges in motion planning due to its contextual properties, including complex inputs and one-to-many mapping. Furthermore, current output validation methods in autonomous driving primarily focus on open-loop trajectory prediction, which is ill-suited for the closed-loop nature of motion planning. We introduce REDOUBT, the first systematic safety validation framework for autonomous vehicle motion planning that employs a duo mechanism, simultaneously inspecting input distributions and output uncertainty. REDOUBT identifies previously overlooked unsafe modes arising from the interplay of In-Distribution/Out-of-Distribution (OOD) scenarios and certain/uncertain planning decisions. We develop specialized solutions for both OOD detection via latent flow matching and decision uncertainty estimation via an energy-based approach. Our extensive experiments demonstrate that both modules outperform existing approaches, under both open-loop and closed-loop evaluation settings.



Paperid:1927
Authors:Tan Pan, Kaiyu Guo, Dongli Xu, Zhaorui Tan, Chen Jiang, Deshu Chen, Xin Guo, Brian Lovell, LIMEI HAN, Yuan Cheng, Mahsa Baktashmotlagh
Abstract:
The generalization ability of deep learning has been extensively studied in supervised settings, yet it remains less explored in unsupervised scenarios. Recently, the Unsupervised Domain Generalization (UDG) task has been proposed to enhance the generalization of models trained with prevalent unsupervised learning techniques, such as Self-Supervised Learning (SSL). UDG confronts the challenge of distinguishing semantics from variations without category labels. Although some recent methods have employed domain labels to tackle this issue, such domain labels are often unavailable in real-world contexts. In this paper, we address these limitations by formalizing UDG as the task of learning a Minimal Sufficient Semantic Representation: a representation that (i) preserves all semantic information shared across augmented views (sufficiency), and (ii) maximally removes information irrelevant to semantics (minimality). We theoretically ground these objectives from the perspective of information theory, demonstrating that optimizing representations to achieve sufficiency and minimality directly reduces out-of-distribution risk. Practically, we implement this optimization through Minimal-Sufficient UDG (MS-UDG), a learnable model by integrating (a) an InfoNCE-based objective to achieve sufficiency; (b) two complementary components to promote minimality: a novel semantic-variation disentanglement loss and a reconstruction-based mechanism for capturing adequate variation. Empirically, MS-UDG sets a new state-of-the-art on popular unsupervised domain-generalization benchmarks, consistently outperforming existing SSL and UDG methods, without category or domain labels during representation learning.



Paperid:1928
Authors:Xiangfei Qiu, Xingjian Wu, Hanyin Cheng, Xvyuan Liu, Chenjuan Guo, Jilin Hu, Bin Yang
Abstract:
Time series forecasting holds significant value in various domains such as economics, traffic, energy, and AIOps, as accurate predictions facilitate informed decision-making. However, the existing Mean Squared Error (MSE) loss function sometimes fails to accurately capture the seasonality or trend within the forecasting horizon, even when decomposition modules are used in the forward propagation to model the trend and seasonality separately. To address these challenges, we propose a simple yet effective Decomposition-Based Loss function called DBLoss. This method uses exponential moving averages to decompose the time series into seasonal and trend components within the forecasting horizon, and then calculates the loss for each of these components separately, followed by weighting them. As a general loss function, DBLoss can be combined with any deep learning forecasting model. Extensive experiments demonstrate that DBLoss significantly improves the performance of state-of-the-art models across diverse real-world datasets and provides a new perspective on the design of time series loss functions.



Paperid:1929
Authors:Jingru Jia, Zehua Yuan, Junhao Pan, Paul McNamara, Deming Chen
Abstract:
What does it truly mean for a language model to “reason” strategically, and can scaling up alone guarantee intelligent, context-aware decisions? Strategic decision-making requires adaptive reasoning, where agents anticipate and respond to others’ actions under uncertainty. Yet, most evaluations of large language models (LLMs) for strategic decision-making often rely heavily on Nash Equilibrium (NE) benchmarks, overlook reasoning depth, and fail to reveal the mechanisms behind model behavior. To address this gap, we introduce a behavioral game-theoretic evaluation framework that disentangles intrinsic reasoning from contextual influence. Using this framework, we evaluate 22 state-of-the-art LLMs across diverse strategic scenarios. We find models like GPT-o3-mini, GPT-o1, and DeepSeek-R1 lead in reasoning depth. Through thinking chain analysis, we identify distinct reasoning styles—such as maximin or belief-based strategies—and show that longer reasoning chains do not consistently yield better decisions. Furthermore, embedding demographic personas reveals context-sensitive shifts: some models (e.g., GPT-4o, Claude-3-Opus) improve when assigned female identities, while others (e.g., Gemini 2.0) show diminished reasoning under minority sexuality personas. These findings underscore that technical sophistication alone is insufficient; alignment with ethical standards, human expectations, and situational nuance is essential for the responsible deployment of LLMs in interactive settings.



Paperid:1930
Authors:Guangchen (Eric) Lan, Huseyin A. Inan, Sahar Abdelnabi, Janardhan Kulkarni, Lukas Wutschitz, Reza Shokri, Christopher Brinton, Robert Sim
Abstract:
Abstract:As the era of autonomous agents making decisions on behalf of users unfolds, ensuring contextual integrity (CI) -- what is the appropriate information to share while carrying out a certain task -- becomes a central question to the field. We posit that CI demands a form of reasoning where the agent needs to reason about the context in which it is operating.To test this, we first prompt LLMs to reason explicitly about CI when deciding what information to disclose. We then extend this approach by developing a reinforcement learning (RL) framework that further instills in models the reasoning necessary to achieve CI.Using a synthetic, automatically created, dataset of only $\sim700$ examples but with diverse contexts and information disclosure norms, we show that our method substantially reduces inappropriate information disclosure while maintaining task performance across multiple model sizes and families. Importantly, improvements transfer from this synthetic dataset to established CI benchmarks such as PrivacyLens that has human annotations and evaluates privacy leakage of AI assistants in actions and tool calls.



Paperid:1931
Authors:Amin Parchami-Araghi, Sukrut Rao, Jonas Fischer, Bernt Schiele
Abstract:
Abstract:Deep networks have shown remarkable performance across a wide range of tasks, yet getting a global concept-level understanding of how they function remains a key challenge. Many post-hoc concept-based approaches have been introduced to understand their workings, yet have been shown to be not always faithful to the model. Further, they make restrictive assumptions on the concepts a model learns, such as class-specificity and small spatial extent, or evaluate assuming concept alignment to human expectations. In this work, we put emphasis on the faithfulness of such concept-based explanations and propose a new model with model-inherent mechanistic concept-explanations. Our concepts are shared across classes and, from any layer, their contribution to the logit and their input-visualization can be faithfully traced. We also leverage foundation models to propose a new concept-consistency metric C$^2$-Score that can be used across different concept-based methods. We show that, compared to prior work, our concepts are quantitatively more consistent and users find our concepts to be more interpretable, all while retaining competitive ImageNet performance.



Paperid:1932
Authors:Fan Wu, Xinyu Yan, Jiabei Liu, Wei Lim
Abstract:
Abstract:Federated Multi-Task Learning (FL-MTL) enables clients with heterogeneous data to collaboratively train models capable of handling multiple downstream tasks. However, FL-MTL faces key challenges, including statistical heterogeneity, task interference, and the need to balance local learning with global knowledge sharing. Traditional methods like FedAvg struggle in such settings due to the lack of explicit mechanisms to address these issues.In this paper, we propose FedRAM, a three-step framework that progressively updates two scalar hyperparameters: the task importance weight and the client aggregation coefficient. FedRAM introduces a reference-proxy-agent strategy, where the proxy model serves as an intermediate between the local reference model and the global agent model. This design reduces the need for repeated local training while preserving local performance.Extensive experiments on six real-world FL-MTL benchmarks show that FedRAM improves performance by at least 3$\%$ over the most baseline on both in-domain and out-of-domain tasks, while reducing computational cost by 15$\times$. These results make FedRAM a robust and practical solution for large-scale FL-MTL applications.The code is available at \url{https://anonymous.4open.science/r/FedRAM-D206}.



Paperid:1933
Authors:Linyang He, Tianjun Zhong, Richard Antonello, Gavin Mischler, Micah Goldblum, Nima Mesgarani
Abstract:
Conventional brain encoding analysis using language models that feeds whole hidden states can be biased toward shallow lexical cues. Here we present a residual-layer disentangling method that extracts four nearly orthogonal vectors from a language model, respectively containing information corresponding to lexicon, syntax, meaning, and reasoning. We first probe the model to locate the layers where each linguistic feature is maximal, then strip lower-level feature incrementally. Applying bootstrap-ridge encoding to natural-speech ECoG yields three insights: 1) Our residual pipeline isolates a reasoning embedding with unique predictive value, possible only because the latest large language models exhibit emergent reasoning behavior. 2) Apparent high-level predictive performance in conventional analyses is largely attributable to recycled shallow information, rather than genuine deep processing. 3) The reasoning embedding reveals distinct spatiotemporal brain activation patterns, including recruitment of frontal and visual regions beyond classical language areas, suggesting a potential neural substrate for high-level reasoning. Together, our approach removes shallow bias, aligns distinct transformer strata with brain hierarchies, and provides the first brain-relevant representation of reasoning.



Paperid:1934
Authors:Jiaru Zou, Yikun Ban, Zihao Li, Yunzhe Qi, Ruizhong Qiu, Ling Yang, Jingrui He
Abstract:
Large language models are typically adapted to downstream tasks through supervised fine-tuning on domain-specific data. While standard fine-tuning focuses on minimizing generation loss to optimize model parameters, we take a deeper step by retaining and leveraging the model’s own learning signals, analogous to how human learners reflect on past mistakes to improve future performance. We first introduce the concept of Mistake Log to systematically track the model’s learning behavior and recurring errors throughout fine-tuning. Treating the original transformer-based model as the Pilot, we correspondingly design a Copilot model to refine the Pilot’s inference performance via logits rectification. We name the overall Pilot-Copilot framework the Transformer Copilot, which introduces (i) a novel Copilot model design, (ii) a joint training paradigm where the Copilot continuously learns from the evolving Mistake Log alongside the Pilot, and (iii) a fused inference paradigm where the Copilot rectifies the Pilot’s logits for enhanced generation. We provide both theoretical and empirical analyses on our new learning framework. Experiments on 12 benchmarks spanning commonsense, arithmetic, and recommendation tasks demonstrate that Transformer Copilot consistently improves performance by up to 34.5%, while introducing marginal computational overhead to Pilot models and exhibiting strong scalability and transferability.



Paperid:1935
Authors:Shaowei Zhang, Jiahan Cao, Dian Cheng, Xunlan Zhou, Shenghua Wan, Le Gan, De-Chuan Zhan
Abstract:
Effective denoising is critical for managing complex visual inputs contaminated with noisy distractors in model-based reinforcement learning (RL). Current methods often oversimplify the decomposition of observations by neglecting the conditional dependence between task-relevant and task-irrelevant components given an observation. To address this limitation, we introduce CsDreamer, a model-based RL approach built upon the world model of Collider-structure Recurrent State-Space Model (CsRSSM). CsRSSM incorporates colliders to comprehensively model the denoising inference process and explicitly capture the conditional dependence. Furthermore, it employs a decoupling regularization to balance the influence of this conditional dependence. By accurately inferring a task-relevant state space, CsDreamer improves learning efficiency during rollouts. Experimental results demonstrate the effectiveness of CsRSSM in extracting task-relevant information, leading to CsDreamer outperforming existing approaches in environments characterized by complex noise interference.



Paperid:1936
Authors:Xiyuan Zhang, Danielle Maddix, Junming Yin, Nick Erickson, Abdul Fatir Ansari, Boran Han, Shuai Zhang, Leman Akoglu, Christos Faloutsos, Michael Mahoney, Tony Hu, Huzefa Rangwala, George Karypis, Yuyang (Bernie) Wang
Abstract:
Since the seminal work of TabPFN, research on tabular foundation models (TFMs) based on in-context learning (ICL) has challenged long-standing paradigms in machine learning. Without seeing any real-world data, models pretrained on purely synthetic datasets generalize remarkably well across diverse datasets, often using only a moderate number of in-context examples. This shifts the focus in tabular machine learning from model architecture design to the design of synthetic datasets, or, more precisely, the prior distributions that generate them. Yet the guiding principles for prior design remain poorly understood. This work marks the first attempt to address the gap. We systematically investigate and identify the key properties of synthetic priors that allow pretrained TFMs to generalize well. Based on these insights, we introduce Mitra, a TFM trained on a curated mixture of synthetic priors selected for their diversity, distinctiveness, and performance on real-world tabular data. Mitra consistently outperforms state-of-the-art TFMs, such as TabPFNv2 and TabICL, across both classification and regression benchmarks, with better sample efficiency.



Paperid:1937
Authors:Zhipeng Zhou, Ziqiao Meng, Pengcheng Wu, Peilin Zhao, Chunyan Miao
Abstract:
Nowadays deep models are required to be versatile due to the increasing realistic needs. Multi-task learning (MTL) offers an efficient way for this purpose to learn multiple tasks simultaneously with a single model. However, prior MTL solutions often focus on resolving conflicts and imbalances during optimization, which may not outperform simple linear scalarization strategies~\citep{xin2022current}. Instead of altering the optimization trajectory, this paper leverages mode connectivity to efficiently approach the Pareto front and identify the desired trade-off point. Unlike Pareto Front Learning (PFL), which aims to align with the entire Pareto front, we focus on effectively and efficiently exploring optimal trade-offs. However, three challenges persist: (1) the low-loss path can neither fully traverse trade-offs nor align with user preference due to its randomness, (2) commonly adopted Bézier curves in mode connectivity are ill-suited to navigating the complex loss landscapes of deep models, and (3) poor scalability to large-scale task scenarios. To address these challenges, we adopt non-uniform rational B-Splines (NURBS) to model mode connectivity, allowing for more flexible and precise curve optimization. Additionally, we introduce an order-aware objective to explore task loss trade-offs and employ a task grouping strategy to enhance scalability under massive task scenarios. Extensive experiments on key MTL datasets demonstrate that our proposed method,EXTRA(EXplore TRAde-offs), effectively identifies the desired point on the Pareto front and achieves state-of-the-art performance.EXTRAis also validated as a plug-and-play solution for mainstream MTL approaches.



Paperid:1938
Authors:Wenyu Mao, Jiancan Wu, Guoqing Hu, Zhengyi Yang, Wei Ji, Xiang Wang
Abstract:
Diffusion models have emerged as a powerful paradigm for generative sequential recommendation, which typically generate next items to recommend guided by user interaction histories with a multi-step denoising process. However, the multi-step process relies on discrete approximations, introducing discretization error that creates a trade-off between computational efficiency and recommendation effectiveness.To address this trade-off, we propose TA-Rec, a two-stage framework that achieves one-step generation by smoothing the denoising function during pretraining while alleviating trajectory deviation by aligning with user preferences during fine-tuning. Specifically, to improve the efficiency without sacrificing the recommendation performance, TA-Rec pretrains the denoising model with Temporal Consistency Regularization (TCR), enforcing the consistency between the denoising results across adjacent steps. Thus, we can smooth the denoising function to map the noise as oracle items in one step with bounded error. To further enhance effectiveness, TA-Rec introduces Adaptive Preference Alignment (APA) that aligns the denoising process with user preference adaptively based on preference pair similarity and timesteps. Extensive experiments prove that TA-Rec’s two-stage objective effectively mitigates the discretization errors-induced trade-off, enhancing both efficiency and effectiveness of diffusion-based recommenders.



Paperid:1939
Authors:Philipp Bomatter, Henry Gouk
Abstract:
The application of machine learning (ML) to electroencephalography (EEG) has great potential to advance both neuroscientific research and clinical applications. However, the generalisability and robustness of EEG-based ML models often hinge on the amount and diversity of training data. It is common practice to split EEG recordings into small segments, thereby increasing the number of samples substantially compared to the number of individual recordings or participants. We conceptualise this as a multi-level data generation process and investigate the scaling behaviour of model performance with respect to the overall sample size and the participant diversity through large-scale empirical studies. We then use the same framework to investigate the effectiveness of different ML strategies designed to address limited data problems: data augmentations and self-supervised learning. Our findings show that model performance scaling can be severely constrained by participant distribution shifts and provide actionable guidance for data collection and ML research. The code for our experiments is publicly available online.



Paperid:1940
Authors:Hyunmo Kang, Abdulkadir Canatar, SueYeon Chung
Abstract:
Understanding representational similarity between neural recordings and computational models is essential for neuroscience, yet remains challenging to measure reliably due to the constraints on the number of neurons that can be recorded simultaneously. In this work, we apply tools from Random Matrix Theory to investigate how such limitations affect similarity measures, focusing on Centered Kernel Alignment (CKA) and Canonical Correlation Analysis (CCA). We propose an analytical framework for representational similarity analysis that relates measured similarities to the spectral properties of the underlying representations. We demonstrate that neural similarities are systematically underestimated under finite neuron sampling, mainly due to eigenvector delocalization. To overcome this, we introduce a denoising method to infer population-level similarity, enabling accurate analysis even with small neuron samples. Theoretical predictions are validated on synthetic and real datasets, offering practical strategies for interpreting neural data under finite sampling constraints.



Paperid:1941
Authors:Xiaosen Wang, Shaokang Wang, Zhijin Ge, Yuyang Luo, Shudong Zhang
Abstract:
Large Vision-Language Models (VLMs) have achieved remarkable success in understanding complex real-world scenarios and supporting data-driven decision-making processes. However, VLMs exhibit significant vulnerability against adversarial examples, either text or image, which can lead to various adversarial outcomes, e.g., jailbreaking, hijacking, and hallucination, etc. In this work, we empirically and theoretically demonstrate that VLMs are particularly susceptible to image-based adversarial examples, where imperceptible perturbations can precisely manipulate each output token. To this end, we propose a novel attack called Vision-language model Manipulation Attack (VMA), which integrates first-order and second-order momentum optimization techniques with a differentiable transformation mechanism to effectively optimize the adversarial perturbation. Notably, VMA can be a double-edged sword: it can be leveraged to implement various attacks, such as jailbreaking, hijacking, privacy breaches, Denial-of-Service, and the generation of sponge examples, etc, while simultaneously enabling the injection of watermarks for copyright protection. Extensive empirical evaluations substantiate the efficacy and generalizability of VMA across diverse scenarios and datasets.



Paperid:1942
Authors:Kun-Yu Lin, Hongjun Wang, Weining Ren, Kai Han
Abstract:
This work introduces panoptic captioning, a novel task striving to seek the minimum text equivalence of images. We take the first step towards panoptic captioning by formulating it as a task of generating a comprehensive textual description for an image, which encapsulates all entities, their respective locations and attributes, relationships among entities, as well as global image state. Through an extensive evaluation, our work reveals that state-of-the-art Multi-modal Large Language Models (MLLMs) have limited performance in solving panoptic captioning. To address this, we propose an effective data engine named PancapEngine to produce high-quality data and a novel method named PancapChain to improve panoptic captioning. Specifically, our PancapEngine first detects diverse categories of entities in images by an elaborate detection suite, and then generates required panoptic captions using entity-aware prompts. Additionally, our PancapChain explicitly decouples the challenging panoptic captioning task into multiple stages and generates panoptic captions step by step. More importantly, we contribute a comprehensive metric named PancapScore and a human-curated test set for reliable model evaluation. Experiments show that our PancapChain-13B model can beat state-of-the-art open-source MLLMs like InternVL-2.5-78B and even surpass proprietary models like GPT-4o and Gemini-2.0-Pro, demonstrating the effectiveness of our data engine and method.



Paperid:1943
Authors:Lara Magdalena Lazier, Aritra Dhar, Vasilije Stambolic, Lukas Cavigelli
Abstract:
Corporate LLMs are gaining traction for efficient knowledge dissemination and management within organizations. However, as current LLMs are vulnerable to leaking sensitive information, it has proven difficult to apply them in settings where strict access control is necessary.To this end, we design AC-LoRA, an end-to-end system for access control-aware corporate LLM chatbots that maintains a strong information isolation guarantee.AC-LoRA maintains separate LoRA adapters for permissioned datasets, along with the document embedding they are finetuned on.AC-LoRA retrieves a precise set of LoRA adapters based on the similarity score with the user query and their permission.This similarity score is later used to merge the responses if more than one LoRA is retrieved, without requiring any additional training for LoRA routing.We provide an end-to-end prototype of AC-LoRA, evaluate it on two datasets, and show that AC-LoRA matchesor even exceeds the performance of state-of-the-art LoRA mixing techniques while providing strong isolation guarantees.Furthermore, we show that AC-LoRA design can be directly applied to different modalities.



Paperid:1944
Authors:Ofir Schlisselberg, Tal Lancewicki, Peter Auer, Yishay Mansour
Abstract:
Abstract:We study the multi-armed bandit problem with adversarially chosen delays in the Best-of-Both-Worlds (BoBW) framework, which aims to achieve near-optimal performance in both stochastic and adversarial environments. While prior work has made progress toward this goal, existing algorithms suffer from significant gaps to the known lower bounds, especially in the stochastic settings. Our main contribution is a new algorithm that, up to logarithmic factors, matches the known lower bounds in each setting individually. In the adversarial case, our algorithm achieves regret of $\widetilde{O}(\sqrt{KT} + \sqrt{D})$, which is optimal up to logarithmic terms, where $T$ is the number of rounds, $K$ is the number of arms, and $D$ is the cumulative delay. In the stochastic case, we provide a regret bound which scale as $\sum_{i:\Delta_i>0}(\log T/\Delta_i) + \frac{1}{K}\sum \Delta_i \sigma_{max}$, where $\Delta_i$ is the suboptimality gap of arm $i$ and $\sigma_{\max}$ is the maximum number of missing observations. To the best of our knowledge, this is the first BoBW algorithm to simultaneously match the lower bounds in both stochastic and adversarial regimes. Moreover, even beyond the BoBW setting, our stochastic regret bound is the first to match the known lower bound under adversarial delays, improving the second term over the best known result by a factor of $K$.



Paperid:1945
Authors:Leqi Shen, Guoqiang Gong, Tao He, Yifeng Zhang, Pengzhang Liu, Sicheng Zhao, guiguang ding
Abstract:
Abstract:Video Large Language Models have demonstrated strong video understanding capabilities, yet their practical deployment is hindered by substantial inference costs caused by redundant video tokens. Existing pruning techniques fail to fully exploit the spatiotemporal redundancy inherent in video data. To bridge this gap, we perform a systematic analysis of video redundancy from two perspectives: temporal context and visual context. Leveraging these insights, we propose Dynamic Density Pruning for Fast Video LLMs termed FastVID.Specifically, FastVID dynamically partitions videos into temporally ordered segments to preserve temporal structure and applies a density-based token pruning strategy to maintain essential visual information.Our method significantly reduces computational overhead while maintaining temporal and visual integrity. Extensive evaluations show that FastVID achieves state-of-the-art performance across various short- and long-video benchmarks on leading Video LLMs, including LLaVA-OneVision and LLaVA-Video.Notably, on LLaVA-OneVision-7B, FastVID effectively prunes $\textbf{90.3}$\% of video tokens, reduces FLOPs to $\textbf{8.3}$\%, and accelerates the prefilling stage by $\textbf{7.1}\times$, while maintaining $\textbf{98.0}$\% of the original accuracy. Our code will be publicly released.



Paperid:1946
Authors:Tim Genewein, Kevin Li, Jordi Grau-Moya, Anian Ruoss, Laurent Orseau, Marcus Hutter
Abstract:
Prompting is one of the main ways to adapt a pretrained model to target tasks. Besides manually constructing prompts, many prompt optimization methods have been proposed in the literature. Method development is mainly empirically driven, with less emphasis on a conceptual understanding of prompting. In this paper we discuss how optimal prompting can be understood through a Bayesian view, which also implies some fundamental limitations of prompting that can only be overcome by tuning weights. The paper explains in detail how meta-trained neural networks behave as Bayesian predictors over the pretraining distribution, whose hallmark feature is rapid in-context adaptation. Optimal prompting can be studied formally as conditioning these Bayesian predictors, yielding criteria for target tasks where optimal prompting is and is not possible. We support the theory with educational experiments on LSTMs and Transformers, where we compare different versions of prefix-tuning and different weight-tuning methods. We also confirm that soft prefixes, which are sequences of real-valued vectors outside the token alphabet, can lead to very effective prompts for trained and even untrained networks by manipulating activations in ways that are not achievable by hard tokens. This adds an important mechanistic aspect beyond the conceptual Bayesian theory.



Paperid:1947
Authors:Mingqi Wu, Qiang Sun, Archer Yang
Abstract:
Real-world high-dimensional data often hide low-dimensional signals beneath dominant structured backgrounds—from batch effects in genomics to scanner artifacts in imaging. We introduce PCA++, a contrastive PCA that enforces a hard uniformity constraint—requiring projected features to have identity covariance—to robustly isolate shared signal subspaces under strong background noise. PCA++ admits a closed-form solution via a generalized eigendecomposition with low-rank truncation for stability and enjoys rigorous non-asymptotic and high-dimensional guarantees under a linear contrastive factor model. In the fixed aspect-ratio regime, we derive exact BBP‐style limits for subspace error; in the growing-spike regime, we obtain a simple closed-form bias that vanishes as the signal strengthens. Empirically, PCA++ outperforms standard PCA and alignment-only contrastive PCA on simulations, corrupted-MNIST, and single-cell transcriptomics, reliably recovering condition-invariant structure. More broadly, we clarify uniformity’s role in contrastive learning—showing that explicit feature dispersion defends against structured noise and enhances robustness.



Paperid:1948
Authors:Stephen Zhang, Mustafa Khan, Vardan Papyan
Abstract:
Large language models (LLMs) often concentrate their attention on a few specific tokens referred to asattention sinks. Common examples include the first token, a prompt-independent sink, and punctuation tokens, which are prompt-dependent. While the tokens causing the sinks often lack direct semantic meaning, the presence of the sinks is critical for model performance, particularly under model compression and KV-caching. Despite their ubiquity, the function, semantic role, and origin of attention sinks—especially those beyond the first token—remain poorly understood. In this work, we conduct a comprehensive investigation demonstrating that attention sinks:catcha sequence of tokens,tagthem using a common direction in embedding space, andreleasethem back into the residual stream, where tokens are later retrieved based on the tags they have acquired. Probing experiments reveal these tags carry semantically meaningful information, such as the truth of a statement. These findings extend to reasoning models, where the mechanism spans more heads and explains greater variance in embeddings, or recent models with query-key normalization, where sinks remain just as prevalent. To encourage future theoretical analysis, we introduce a minimal problem which can be solved through the 'catch, tag, release' mechanism, and where it emerges through training.



Paperid:1949
Authors:Nathanael Jo, Ashia Wilson, Kathleen Creel, Manish Raghavan
Abstract:
Firms' algorithm development practices are often homogeneous Whether firms train algorithms on similar data or rely on similar pre-trained models, the result is correlated predictions. In the context of personalized pricing, correlated algorithms can be viewed as a means to collude among competing firms, but whether or not this conduct is legal depends on the mechanisms of achieving collusion. We investigate the precise mechanisms through a formal game-theoretic model. Indeed, we find that (1) higher correlation diminishes consumer welfare and (2) as consumers become more price sensitive, firms are increasingly incentivized to compromise on the accuracy of their predictions in exchange for coordination. We demonstrate our theoretical results in a stylized empirical study where two firms compete using personalized pricing algorithms. Our results demonstrate a new mechanism for achieving collusion through correlation, which allows us to analyze its legal implications. Correlation through algorithms is a new frontier of anti-competitive behavior that is largely unconsidered by US antitrust law.



Paperid:1950
Authors:Dongyi Wang, Yuanwei Jiang, Zhenyi Zhang, Xiang Gu, Peijie Zhou, Jian Sun
Abstract:
Learning the underlying dynamics of single cells from snapshot data has gained increasing attention in scientific and machine learning research. The destructive measurement technique and cell proliferation/death result in unpaired and unbalanced data between snapshots, making the learning of the underlying dynamics challenging. In this paper, we propose joint Velocity-Growth Flow Matching (VGFM), a novel paradigm that jointly learns state transition and mass growth of single-cell populations via flow matching. VGFM builds an ideal single-cell dynamics containing velocity of state and growth of mass, driven by a presented two-period dynamic understanding of the static semi-relaxed optimal transport, a mathematical tool that seeks the coupling between unpaired and unbalanced data. To enable practical usage, we approximate the ideal dynamics using neural networks, forming our joint velocity and growth matching framework. A distribution fitting loss is also employed in VGFM to further improve the fitting performance for snapshot data. Extensive experimental results on both synthetic and real datasets demonstrate that VGFM can capture the underlying biological dynamics accounting for mass and state variations over time, outperforming existing approaches for single-cell dynamics modeling.



Paperid:1951
Authors:Yu Zhou, Xingyu Wu, Jibin Wu, Liang Feng, KC Tan
Abstract:
Model merging is a technique that combines multiple large pretrained models into a single model, enhancing performance and broadening task adaptability without original data or additional training. However, most existing model merging methods focus primarily on exploring the parameter space, merging models with identical architectures. Despite its potential, merging in the architecture space remains in its early stages due to the vast search space and challenges related to layer compatibility. This paper designs a hierarchical model merging framework named HM3, formulating a bilevel multi-objective model merging problem across both parameter and architecture spaces. At the parameter level, HM3 integrates existing merging methods to quickly identify optimal parameters. Based on these, an actor-critic strategy with efficient policy discretization is employed at the architecture level to explore inference paths with Markov property in the layer-granularity search space for reconstructing these optimal models. By training reusable policy and value networks, HM3 learns Pareto optimal models to provide customized solutions for various tasks. Experimental results on language and vision tasks demonstrate that HM3 outperforms methods focusing solely on the parameter or architecture space. The code will be made publicly available after the review process is complete.



Paperid:1952
Authors:Vahid Balazadeh, Hamidreza Kamkari, Valentin Thomas, Junwei Ma, Bingru Li, Jesse Cresswell, Rahul Krishnan
Abstract:
Causal effect estimation from observational data is fundamental across various applications. However, selecting an appropriate estimator from dozens of specialized methods demands substantial manual effort and domain expertise. We present CausalPFN, a single transformer thatamortizesthis workflow: trained once on a large library of simulated data-generating processes that satisfy ignorability, it infers causal effects for new observational datasets out-of-the-box. CausalPFN combines ideas from Bayesian causal inference with the large-scale training protocol of prior-fitted networks (PFNs), learning to map raw observations directly to causal-effects without any task-specific adjustment. Our approach achieves superior average performance on heterogeneous and average treatment effect estimation benchmarks (IHDP, Lalonde, ACIC). Moreover, it shows competitive performance for real-world policy making on uplift modeling tasks. CausalPFN provides calibrated uncertainty estimates to support reliable decision-making based on Bayesian principles. This ready-to-use model does not require any further training or fine-tuning and takes a step toward automated causal inference.



Paperid:1953
Authors:Dafeng Zhang
Abstract:
In the field of computer vision, the task of stereo image super-resolution (StereoSR) has garnered significant attention due to its potential applications in augmented reality, virtual reality, and autonomous driving. Traditional Transformer-based models, while powerful, often suffer from attention noise, leading to suboptimal reconstruction issues in super-resolved images. This paper introduces DIFFSSR, a novel neural network architecture designed to address these challenges. We introduce the Diff Cross Attention Block (DCAB) and the Sliding Stereo Cross-Attention Module (SSCAM) to enhance feature integration and mitigate the impact of attention noise. The DCAB differentiates between relevant and irrelevant context, amplifying attention to important features and canceling out noise. The SSCAM, with its sliding window mechanism and disparity-based attention, adapts to local variations in stereo images, preserving details, and addressing the performance degradation due to misalignment of horizontal epipolar lines in stereo images. Extensive experiments on benchmark datasets demonstrate that DIFFSSR outperforms state-of-the-art methods, including NAFSSR and SwinFIRSSR, in terms of both quantitative metrics and visual quality.



Paperid:1954
Authors:Xingbo Fu, Zhenyu Lei, Zihan Chen, Binchi Zhang, Chuxu Zhang, Jundong Li
Abstract:
Graph Neural Networks (GNNs) have revolutionized the field of graph learning by learning expressive graph representations from massive graph data.As a common pattern to train powerful GNNs, the "pre-training, adaptation" scheme first pre-trains GNNs over unlabeled graph data and subsequently adapts them to specific downstream tasks. In the adaptation phase, graph prompting is an effective strategy that modifies input graph data with learnable prompts while keeping pre-trained GNN models frozen. Typically, existing graph prompting studies mainly focus onfeature-orientedmethods that apply graph prompts to node features or hidden representations. However, these studies often achieve suboptimal performance, as they consistently overlook the potential oftopology-orientedprompting, which adapts pre-trained GNNs by modifying the graph topology. In this study, we conduct a pioneering investigation of graph prompting in terms of graph topology. We propose the firstGraphTopology-OrientedPrompting (GraphTOP) framework to effectively adapt pre-trained GNN models for downstream tasks. More specifically, we reformulate topology-oriented prompting as an edge rewiring problem within multi-hop local subgraphs and relax it into the continuous probability space through reparameterization while ensuring tight relaxation and preserving graph sparsity. Extensive experiments on five graph datasets under four pre-training strategies demonstrate that our proposed GraphTOP outshines six baselines on multiple node classification datasets. Our code is available at https://anonymous.4open.science/r/GraphTOP-9678.



Paperid:1955
Authors:Alina Ene, Ta Duy Nguyen, Adrian Vladu
Abstract:
Abstract:In this paper, we study the problem $\min_{x\in R^{d},Nx=v}\sum\_{i=1}^{n}f((Ax-b)_{i})$for a quasi-self-concordant function $f: R\to R$, where $A,N$ are$n\times d$ and $m\times d$ matrices, $b,v$ are vectors of length$n$ and $m$ with $n\ge d.$ We show an algorithm based on a trust-regionmethod with an oracle that can be implemented using $\widetilde{O}(d^{1/3})$linear system solves, improving the $\widetilde{O}(n^{1/3})$ oracleby [Adil-Bullins-Sachdeva, NeurIPS 2021]. Our implementation ofthe oracle relies on solving the overdetermined $\ell\_{\infty}$-regressionproblem $\min\_{x\in R^{d},Nx=v}\|Ax-b\|\_{\infty}$. We provide analgorithm that finds a $(1+\epsilon)$-approximate solution to thisproblem using $O((d^{1/3}/\epsilon^{2/3}+1/\epsilon^{2})\log(n/\epsilon))$linear system solves. This algorithm leverages $\ell\_{\infty}$ Lewisweight overestimates and achieves this iteration complexity via asimple lightweight IRLS approach, inspired by the work of [Ene-Vladu,ICML 2019]. Experimentally, we demonstrate that our algorithm significantlyimproves the runtime of the standard CVX solver.



Paperid:1956
Authors:Ruiqi Wang, Hao Zhang
Abstract:
We present an open-vocabulary and zero-shot method for arbitrary referring expression segmentation (RES), targeting input expressions that are more general than what prior works were designed to handle. Specifically, our inputs encompass both object- and part-level labels as well as implicit references pointing to properties or qualities of object/part function, design, style, material, etc. Our model, coined RESAnything, leverages Chain-of-Thoughts (CoT) reasoning, where the key idea is attribute prompting. We generate detailed descriptions of object/part attributes including shape, color, and location for potential segment proposals through systematic prompting of a large language model (LLM), where the proposals are produced by a foundational image segmentation model. Our approach encourages deep reasoning about object or part attributes related to function, style, design, etc., enabling the system to handle implicit queries without any part annotations for training or fine-tuning. As the first zero-shot and LLM-based RES method, RESAnything achieves clearly superior performance among zero-shot methods on traditional RES benchmarks and significantly outperforms existing methods on challenging scenarios involving implicit queries and complex part-level relations. Finally, we contribute a new benchmark dataset to offer ~3K carefully curated RES instances to assess part-level, arbitrary RES solutions.



Paperid:1957
Authors:Alvaro Arroyo, Alessio Gravina, Benjamin Gutteridge, Federico Barbero, Claudio Gallicchio, Xiaowen Dong, Michael Bronstein, Pierre Vandergheynst
Abstract:
Graph Neural Networks (GNNs) are models that leverage the graph structure to transmit information between nodes, typically through the message-passing operation. While widely successful, this approach is well-known to suffer from representational collapse as the number of layers increases and insensitivity to the information contained at distant and poorly connected nodes. In this paper, we present a unified view of on the appearance of these issues through the lens of vanishing gradients, using ideas from linear control theory for our analysis. We propose an interpretation of GNNs as recurrent models and empirically demonstrate that a simple state-space formulation of an GNN effectively alleviates these issues at no extra trainable parameter cost. Further, we show theoretically and empirically that (i) Traditional GNNs are by design prone to extreme gradient vanishing even after few layers; (ii) Feature collapse is directly related to the mechanism causing vanishing gradients; (iii) Long-range modeling is most easily achieved by a combination of graph rewiring and vanishing gradient mitigation. We believe our work will help bridge the gap between the recurrent and graph neural network literature and will unlock the design of new deep and performant GNNs.



Paperid:1958
Authors:Sana Tonekaboni, Lena Stempfle, Adibvafa Fallahpour, Walter Gerych, Marzyeh Ghassemi
Abstract:
Foundation models trained on large-scale de-identified electronic health records (EHRs) hold promise for clinical applications. However, their capacity to memorize patient information raises important privacy concerns. In this work, we introduce a suite of black-box evaluation tests to assess memorization risks in foundation models trained on structured EHR data. Our framework includes methods for probing memorization at both the embedding and generative levels, and distinguishes between generalization and harmful memorization in clinically relevant settings. We contextualize memorization in terms of its potential to compromise patient privacy, particularly for vulnerable subgroups. We validate our approach on a publicly available EHR foundation model and release an open-source toolkit to facilitate reproducible and collaborative privacy assessments in healthcare AI.



Paperid:1959
Authors:Yihan Zhu, Gang Liu, Eric Inae, Tengfei Luo, Meng Jiang
Abstract:
Polymers are large macromolecules composed of repeating structural units known as monomers and are widely applied in fields such as energy storage, construction, medicine, and aerospace. However, existing graph neural network methods, though effective for small molecules, only model the single unit of polymers and fail to produce consistent vector representations for the true polymer structure with varying numbers of units. To address this challenge, we introduce Graph Repetition Invariance (GRIN), a novel method to learn polymer representations that are invariant to the number of repeating units in their graph representations. GRIN integrates a graph-based maximum spanning tree alignment with repeat-unit augmentation to ensure structural consistency. We provide theoretical guarantees for repetition‐invariance from both model and data perspectives, demonstrating that three repeating units are the minimal augmentation required for optimal invariant representation learning. GRIN outperforms state-of-the-art baselines on both homopolymer and copolymer benchmarks, learning stable, repetition-invariant representations that generalize effectively to polymer chains of unseen sizes.



Paperid:1960
Authors:Weijie Wang, Donny Y. Chen, Zeyu Zhang, Duochao Shi, Akide Liu, Bohan Zhuang
Abstract:
Abstract:Feed-forward 3D Gaussian Splatting (3DGS) models have recently emerged as a promising solution for novel view synthesis, enabling one-pass inference without the need for per-scene 3DGS optimization. However, their scalability is fundamentally constrained by the limited capacity of their encoders, leading to degraded performance or excessive memory consumption as the number of input views increases. In this work, we analyze feed-forward 3DGS frameworks through the lens of the Information Bottleneck principle and introduce ZPressor, a lightweight architecture-agnostic module that enables efficient compression of multi-view inputs into a compact latent state $Z$ that retains essential scene information while discarding redundancy. Concretely, ZPressor enables existing feed-forward 3DGS models to scale to over 100 input views at 480P resolution on an 80GB GPU, by partitioning the views into anchor and support sets and using cross attention to compress the information from the support views into anchor views, forming the compressed latent state $Z$. We show that integrating ZPressor into several state-of-the-art feed-forward 3DGS models consistently improves performance under moderate input views and enhances robustness under dense view settings on two large-scale benchmarks DL3DV-10K and RealEstate10K.



Paperid:1961
Authors:Edward Hu, Jie Wang, Xingfang Yuan, Fiona Luo, Muyao Li, Gaspard Lambrechts, Oleh Rybkin, Dinesh Jayaraman
Abstract:
A robot's instantaneous sensory observations do not always reveal all task-relevant state information. Under such partial observability, optimal behavior typically involves explicitly acting to gain the missing information. Today's standard robot learning techniques struggle to produce such active perception behaviors. We propose a simple real-world robot learning recipe to efficiently train active perception policies. Our approach, asymmetric advantage weighted regression (AAWR), exploits access to "privileged" extra sensors at training time. The privileged sensors enable training high-quality privileged value functions that aid in estimating the advantage of the target policy. Bootstrapping from a small number of potentially suboptimal demonstrations and an easy-to-obtain coarse policy initialization, AAWR quickly acquires active perception behaviors and boosts task performance. In evaluations on 8 manipulation tasks on 3 robots spanning varying degrees of partial observability, AAWR synthesizes reliable active perception behaviors that outperform all prior approaches. When initialized with a "generalist" robot policy that struggles with active perception tasks, AAWR efficiently generates information-gathering behaviors that allow it to operate under severe partial observability for manipulation tasks.



Paperid:1962
Authors:Fabio Brau, Maura Pintor, Antonio Cinà, Raffaele Mura, Luca Scionis, Luca Oneto, Fabio Roli, Battista Biggio
Abstract:
Ensemble-based black-box transfer attacks optimize adversarial examples on a set of surrogate models, claiming to reach high success rates by querying the (unknown) target model only a few times. In this work, we show that prior evaluations are systematically biased, as such methods are tested only under overly optimistic scenarios, without considering (i) how the choice of surrogate models influences transferability, (ii) how they perform against robust target models, and (iii) whether querying the target to refine the attack is really required.To address these gaps, we introduce TransferBench, a framework for evaluating ensemble-based black-box transfer attacks under more realistic and challenging scenarios than prior work. Our framework considers 17 distinct settings on CIFAR-10 and ImageNet, including diverse surrogate-target combinations, robust targets, and comparisons to baseline methods that do not use any query-based refinement mechanism. Our findings reveal that existing methods fail to generalize to more challenging scenarios, and that query-based refinement offers little to no benefit, contradicting prior claims. These results highlight that building reliable and query-efficient black-box transfer attacks remains an open challenge. We release our benchmark and evaluation code at: https://github.com/pralab/transfer-bench.



Paperid:1963
Authors:Wentao Deng, Jiahuan Pei, Zhiwei Xu, Zhaochun Ren, Zhumin Chen, Pengjie Ren
Abstract:
A multi-agent system (MAS) enhances its capacity to solve complex natural language processing (NLP) tasks through collaboration among multiple agents, where consensus-seeking serves as a fundamental mechanism.However, existing consensus-seeking approaches typically rely on voting mechanisms to judge consensus, overlooking contradictions in system-internal beliefs that destabilize the consensus.Moreover, these methods often involve agents updating their results through indiscriminate collaboration with every other agent.Such uniform interaction fails to identify the optimal collaborators for each agent, hindering the emergence of a stable consensus.To address these challenges, we provide a theoretical framework for selecting optimal collaborators that maximize consensus stability.Based on the theorems, we propose the Belief-Calibrated Consensus Seeking (BCCS) framework to facilitate stable consensus via selecting optimal collaborators and calibrating the consensus judgment by system-internal beliefs.Experimental results on the MATH and MMLU benchmark datasets demonstrate that the proposed BCCS framework outperforms the best existing results by 2.23\% and 3.95\% of accuracy on challenging tasks, respectively.Our code and data are available at https://anonymous.4open.science/r/BCCS-EB58.



Paperid:1964
Authors:Gilad Yehudai, Noah Amsel, Joan Bruna
Abstract:
Abstract:It is understood that different neural network architectures are suited to different tasks, but is there always a single best architecture for a given task? We compare the expressive power of transformers, RNNs, and transformers with chain of thought tokens on a simple and natural class of tasks we term Compositional Reasoning Questions (CRQ). This family captures multi-step problems with tree-like compositional structure, such as evaluating Boolean formulas. We prove that under standard hardness assumptions, *none* of these three architectures is capable of solving CRQs unless some hyperparameter (depth, embedding dimension, and number of chain of thought tokens, respectively) grows with the size of the input. We then provide constructions for solving CRQs with each architecture. For transformers, our construction uses depth that is logarithmic in the problem size. For RNNs, logarithmic embedding dimension is necessary and sufficient, so long as the inputs are provided in a certain order. For transformers with chain of thought, our construction uses $n$ CoT tokens. These results show that, while CRQs are inherently hard, there are several different ways for language models to overcome this hardness. Even for a single class of problems, each architecture has strengths and weaknesses, and none is strictly better than the others.



Paperid:1965
Authors:Finlay Hudson, James Gardner, William Smith
Abstract:
Humans excel at constructing panoramic mental models of their surroundings, maintaining object permanence and inferring scene structure beyond visible regions. In contrast, current artificial vision systems struggle with persistent, panoramic understanding, often processing scenes egocentrically on a frame-by-frame basis. This limitation is pronounced in the Track Any Point (TAP) task, where existing methods fail to track 2D points outside the field of view. To address this, we introduce TAP-Vid 360, a novel task that requires predicting the 3D direction to queried scene points across a video sequence, even when far outside the narrow field of view of the observed video. This task fosters learning allocentric scene representations without needing dynamic 4D ground truth scene models for training. Instead, we exploit 360 videos as a source of supervision, resampling them into narrow field-of-view perspectives while computing ground truth directions by tracking points across the full panorama using a 2D pipeline. We introduce a new dataset and benchmark, TAP360-10k comprising 10k perspective videos with ground truth directional point tracking. Our baseline adapts CoTracker v3 to predict per-point rotations for direction updates, outperforming existing TAP and TAP-Vid 3D methods.



Paperid:1966
Authors:Jingjun Yang, Liangwei Fan, Jinpu Zhang, Xiangkai Lian, Hui Shen, Dewen Hu
Abstract:
The integration of image and event streams offers a promising approach for achieving robust visual object tracking in complex environments. However, current fusion methods achieve high performance at the cost of significant computational overhead and struggle to efficiently extract the sparse, asynchronous information from event streams, failing to leverage the energy-efficient advantages of event-driven spiking paradigms. To address this challenge, we propose the first fully Spiking Frame-Event Tracking framework called SpikeFET. This network achieves synergistic integration of convolutional local feature extraction and Transformer-based global modeling within the spiking paradigm, effectively fusing frame and event data. To overcome the degradation of translation invariance caused by convolutional padding, we introduce a Random Patchwork Module (RPM) that eliminates positional bias through randomized spatial reorganization and learnable type encoding while preserving residual structures. Furthermore, we propose a Spatial-Temporal Regularization (STR) strategy that overcomes similarity metric degradation from asymmetric features by enforcing spatio-temporal consistency among temporal template features in latent space. Extensive experiments across multiple benchmarks demonstrate that the proposed framework achieves superior tracking accuracy over existing methods while significantly reducing power consumption, attaining an optimal balance between performance and efficiency. The code will be released.



Paperid:1967
Authors:Avrajit Ghosh, Bai Cong, Rio Yokota, Saiprasad Ravishankar, Rongrong Wang, Molei Tao, Mohammad Emtiyaz Khan, Thomas Möllenhoff
Abstract:
The performance of Variational Learning (VL) for deep neural networks has consistently been improving over the years and is now at par with the standard optimizers. Part of its empirical success can be explained by theories such as PAC-Bayes bounds, minimum description length and marginal likelihood, but there are few tools to unravel the implicit regularization in play. Here, we use the Edge of Stability (EoS) to understand the implicit regularization of VL. EoS has previously been used to show that gradient descent can find flat solutions and we extend this result to VL to show that it can find even flatter solutions. This is obtained by simply controlling the posterior covariance and the number of Monte Carlo samples from the posterior. These results are derived in a similar fashion as the standard EoS literature for deep learning by first deriving a result for a quadratic problem and then extending it to general loss functions. We empirically validate these findings on a wide variety of large networks, such as ResNet and ViT, to find that the theoretical results closely match the empirical ones. Ours is the first work to use EoS for VL and show its effectiveness for deep learning.



Paperid:1968
Authors:Noam Razin, Zixuan Wang, Hubert Strauss, Stanley Wei, Jason Lee, Sanjeev Arora
Abstract:
The success of Reinforcement Learning from Human Feedback (RLHF) critically depends on the quality of the reward model. However, while this quality is primarily evaluated through accuracy, it remains unclear whether accuracy fully captures what makes a reward model an effective teacher. We address this question from an optimization perspective. First, we prove that regardless of how accurate a reward model is, if it induces low reward variance, then the RLHF objective suffers from a flat landscape. Consequently, even a perfectly accurate reward model can lead to extremely slow optimization, underperforming less accurate models that induce higher reward variance. We additionally show that a reward model that works well for one language model can induce low reward variance, and thus a flat objective landscape, for another. These results establish a fundamental limitation of evaluating reward models solely based on accuracy or independently of the language model they guide. Experiments using models of up to 8B parameters corroborate our theory, demonstrating the interplay between reward variance, accuracy, and reward maximization rate. Overall, our findings highlight that beyond accuracy, a reward model needs to induce sufficient variance for efficient optimization.



Paperid:1969
Authors:Vijay Viswanathan, Yanchao Sun, Xiang Kong, Meng Cao, Graham Neubig, Tongshuang Wu
Abstract:
Language models must be adapted to understand and follow user instructions. Reinforcement learning is widely used to facilitate this— typically using fixed criteria such as "helpfulness" and "harmfulness". In our work, we instead propose using flexible, instruction-specific criteria as a means of broadening the impact that reinforcement learning can have in eliciting instruction following. We propose "Reinforcement Learning from Checklist Feedback" (RLCF). From instructions, we extract checklists and evaluate how well responses satisfy each item—using both AI judges and specialized verifier programs—then combine these scores to compute rewards for RL. We compare RLCF with other alignment methods applied to a state-of-the-art instruction following model (Qwen2.5-7B-Instruct) — RLCF is the only method to improve on every benchmark, including a 4 point increase in hard satisfaction rate on FollowBench and a 3 point boost in win rate on Arena-Hard. These results establish checklist feedback as a key tool for improving language models' support of queries that express a multitude of needs. We will release our models and our dataset of checklists, "WildChecklists", to the public.



Paperid:1970
Authors:Julian Minder, Clément Dumas, Caden Juang, Bilal Chughtai, Neel Nanda
Abstract:
Model diffing is the study of how fine-tuning changes a model's representations and internal algorithms. Many behaviors of interest are introduced during fine-tuning, and model diffing offers a promising lens to interpret such behaviors. Crosscoders are a recent model diffing method that learns a shared dictionary of interpretable concepts represented as latent directions in both the base and fine-tuned models, allowing us to track how concepts shift or emerge during fine-tuning. Notably, prior work has observed concepts with no direction in the base model, and it was hypothesized that these model-specific latents were concepts introduced during fine-tuning.However, we identify two issues which stem from the crosscoders L1 training loss that can misattribute concepts as unique to the fine-tuned model, when they really exist in both models. We develop Latent Scaling to flag these issues by more accurately measuring each latent's presence across models.In experiments comparing Gemma 2 2B base and chat models, we observe that the standard crosscoder suffers heavily from these issues. Building on these insights, we train a crosscoder with BatchTopK loss and show that it substantially mitigates these issues, finding more genuinely chat-specific and highly interpretable concepts. We recommend practitioners adopt similar techniques.Using the BatchTopK crosscoder, we successfully identify a set of chat-specific latents that are both interpretable and causally effective, representing concepts such as false information and personal question, along with multiple refusal-related latents that show nuanced preferences for different refusal triggers. Overall, our work advances best practices for the crosscoder-based methodology for model diffing and demonstrates that it can provide concrete insights into how chat-tuning modifies model behavior.



Paperid:1971
Authors:Zihan Zhou, Muhammad Qasim Elahi, Murat Kocaoglu
Abstract:
A fundamental challenge in the empirical sciences involves uncovering causal structure through observation and experimentation. Causal discovery entails linking the conditional independence (CI) invariances in observational data to their corresponding graphical constraints via d-separation. In this paper, we consider a general setting where we have access to data from multiple experimental distributions resulting from hard interventions, as well as potentially from an observational distribution. By comparing different interventional distributions, we propose a set of graphical constraints that are fundamentally linked to Pearl's do-calculus within the framework of hard interventions. These graphical constraints associate each graphical structure with a set of interventional distributions that are consistent with the rules of do-calculus. We characterize the interventional equivalence class of causal graphs with latent variables and introduce a graphical representation that can be used to determine whether two causal graphs are interventionally equivalent, i.e., whether they are associated with the same family of hard interventional distributions, where the elements of the family are indistinguishable using the invariances from do-calculus. We also propose a learning algorithm to integrate multiple datasets from hard interventions, introducing new orientation rules. The learning objective is a tuple of augmented graphs which entails a set of causal graphs. We also prove the soundness of the proposed algorithm.



Paperid:1972
Authors:Soyoung Yoon, Gyuwan Kim, Gyu-Hwung Cho, seung-won hwang
Abstract:
Listwise reranking with large language models (LLMs) enhances top-ranked results in retrieval-based applications.Due to the limit in context size and high inference cost of long context, reranking is typically performed over a fixed size of small subsets, with the final ranking aggregated from these partial results.This fixed computation disregards query difficulty and document distribution, leading to inefficiencies. We propose AcuRank, an adaptive reranking framework that dynamically adjusts both the amount and target of computation based on uncertainty estimates over document relevance. Using a Bayesian TrueSkill model, we iteratively refine relevance estimates until reaching sufficient confidence levels, and our explicit modeling of ranking uncertainty enables principled control over reranking behavior and avoids unnecessary updates to confident predictions.Results on the TREC-DL and BEIR benchmarks show that our method consistently achieves a superior accuracy–efficiency trade-off and scales better with compute than fixed-computation baselines.These results highlight the effectiveness and generalizability of our method across diverse retrieval tasks and LLM-based reranking models.



Paperid:1973
Authors:Eric Lei, Hamed Hassani, Shirin Saeedi Bidokhti
Abstract:
Recent efforts in neural compression have focused on the rate-distortion-perception (RDP) tradeoff, where the perception constraint ensures the source and reconstruction distributions are close in terms of a statistical divergence. Theoretical work on RDP describes properties of RDP-optimal compressors without providing constructive and low complexity solutions. While classical rate distortion theory shows that optimal compressors should efficiently pack space, RDP theory additionally shows that infinite randomness shared between the encoder and decoder may be necessary for RDP optimality. In this paper, we propose neural compressors that are low complexity and benefit from high packing efficiency through lattice coding and shared randomness through shared dithering over the lattice cells. For two important settings, namely infinite shared and zero shared randomness, we analyze the RDP tradeoff achieved by our proposed neural compressors and show optimality in both cases. Experimentally, we investigate the roles that these two components of our design, lattice coding and randomness, play in the performance of neural compressors on synthetic and real-world data. We observe that performance improves with more shared randomness and better lattice packing.



Paperid:1974
Authors:Yiheng Lin, Zaiwei Chen, Christopher Yeh, Adam Wierman
Abstract:
We study the value of stochastic predictions in online optimal control with random disturbances. Prior work provides performance guarantees based on prediction error but ignores the stochastic dependence between predictions and disturbances. We introduce a general framework modeling their joint distribution and define ``prediction power'' as the control cost improvement from the optimal use of predictions compared to ignoring the predictions. In the time-varying Linear Quadratic Regulator (LQR) setting, we derive a closed-form expression for prediction power and discuss its mismatch with prediction accuracy and connection with online policy optimization. To extend beyond LQR, we study general dynamics and costs. We establish a lower bound of prediction power under two sufficient conditions that generalize the properties of the LQR setting, characterizing the fundamental benefit of incorporating stochastic predictions. We apply this lower bound to non-quadratic costs and show that even weakly dependent predictions yield significant performance gains.



Paperid:1975
Authors:Xiong Peng, Bo Han, Fengfei Yu, Tongliang Liu, Feng Liu, Mingyuan Zhou
Abstract:
Model inversion attacks (MIAs) aim to reconstruct class-representative samples from trained models. Recent generative MIAs utilize generative adversarial networks to learn image priors that guide the inversion process, yielding reconstructions with high visual quality and strong fidelity to the private data. To explore the reason behind their effectiveness, we begin by examining the gradients of inversion loss w.r.t. synthetic inputs, and find that these gradients are surprisingly noisy. Further analysis shows that generative model inversion approaches implicitly denoise the gradients by projecting them onto the tangent space of the generator manifold—filtering out directions that deviate from the manifold structure while preserving informative components aligned with it. Our empirical measurements show that, in models trained with standard supervision, loss gradients exhibit large angular deviations from the data manifold, indicating poor alignment with class-relevant directions. This observation motivates our central hypothesis: models become more vulnerable to MIAs when their loss gradients align more closely with the generator manifold. We validate this hypothesis by designing a novel training objective that explicitly promotes such alignment. Building on this insight, we further introduce a training-free approach to enhance gradient–manifold alignment during inversion, leading to consistent improvements over state-of-the-art generative MIAs.



Paperid:1976
Authors:Jang-Hyun Kim, Jinuk Kim, Sangwoo Kwon, Jae Lee, Sangdoo Yun, Hyun Oh Song
Abstract:
Abstract:Transformer-based large language models (LLMs) cache context as key-value (KV) pairs during inference. Longer context increases KV cache sizes, leading to significant memory overhead and higher attention computation latency. This paper introduces KVzip, a query-agnostic KV cache eviction method enabling effective reuse of compressed context caches across different queries. KVzip quantifies the importance of each KV pair using the underlying LLM to reconstruct original contexts from cached KV pairs, subsequently evicting pairs with lower importance. Extensive empirical evaluations demonstrate that KVzip reduces KV cache size by $3$-$4\times$ and decreases FlashAttention latency by approximately $2\times$, without performance degradation in question-answering, retrieval, mathematical reasoning, and code comprehension tasks. Evaluations include state-of-the-art models such as LLaMA3.1-8B, Qwen2.5-14B, and Gemma3-12B, with context lengths reaching up to 170K tokens. KVzip significantly outperforms existing KV eviction methods, which suffer performance losses even at a 90\% cache budget ratio under multi-query scenarios.



Paperid:1977
Authors:Zixun Fang, Kai Zhu, Zhiheng Liu, Yu Liu, Wei Zhai, Yang Cao, Zheng-Jun Zha
Abstract:
Panoramic video generation aims to synthesize 360-degree immersive videos, holding significant importance in the fields of VR, world models, and spatial intelligence. Existing works fail to synthesize high-quality panoramic videos due to the inherent modality gap between panoramic data and perspective data, which constitutes the majority of the training data for modern diffusion models. In this paper, we propose a novel framework utilizing pretrained perspective video models for generating panoramic videos. Specifically, we design a novel panorama representation named ViewPoint map, which possesses global spatial continuity and fine-grained visual details simultaneously. With our proposed Pano-Perspective attention mechanism, the model benefits from pretrained perspective priors and captures the panoramic spatial correlations of the ViewPoint map effectively. Extensive experiments demonstrate that our method can synthesize highly dynamic and spatially consistent panoramic videos, achieving state-of-the-art performance and surpassing previous methods.



Paperid:1978
Authors:Bowen Dong, Minheng Ni, Zitong Huang, Guanglei Yang, Wangmeng Zuo, Lei Zhang
Abstract:
Multimodal hallucination in multimodal large language models (MLLMs) restricts the correctness of MLLMs. However, multimodal hallucinations are multi-sourced and arise from diverse causes. Existing benchmarks fail to adequately distinguish between perception-induced hallucinations and reasoning-induced hallucinations. This failure constitutes a significant issue and hinders the diagnosis of multimodal reasoning failures within MLLMs. To address this, we propose the MIRAGE benchmark, which isolates reasoning hallucinations by constructing questions where input images are correctly perceived by MLLMs yet reasoning errors persist. MIRAGE introduces multi-granular evaluation metrics: accuracy, factuality, and LLMs hallucination score for hallucination quantification. Our analysis reveals strong correlations between question types and specific hallucination patterns, particularly systematic failures of MLLMs in spatial reasoning involving complex relationships (\emph{e.g.}, complex geometric patterns across images). This highlights a critical limitation in the reasoning capabilities of current MLLMs and provides targeted insights for hallucination mitigation on specific types. To address these challenges, we propose Logos, a method that combines curriculum reinforcement fine-tuning to encourage models to generate logic-consistent reasoning chains by stepwise reducing learning difficulty, and collaborative hint inference to reduce reasoning complexity. Logos establishes a baseline on MIRAGE, and reduces the logical hallucinations in original base models. MIRAGE will be publicly available.



Paperid:1979
Authors:Wei Wang, Haifeng Xia, Chao Huang, Zhengming Ding, Cong Wang, Haojie Li, Xiaochun Cao
Abstract:
In domain adaption (DA), joint maximum mean discrepancy (JMMD), as a famous distribution-distance metric, aims to measure joint probability distribution difference between the source domain and target domain, while it is still not fully explored and especially hard to be applied into a subspace-learning framework as its empirical estimation involves a tensor-product operator whose partial derivative is difficult to obtain. To solve this issue, we deduce a concise JMMD based on the Representer theorem that avoids the tensor-product operator and obtains two essential findings. First, we reveal the uniformity of JMMD by proving that previous marginal, class conditional, and weighted class conditional probability distribution distances are three special cases of JMMD with different label reproducing kernels. Second, inspired by graph embedding, we observe that the similarity weights, which strengthen the intra-class compactness in the graph of Hilbert Schmidt independence criterion (HSIC), take opposite signs in the graph of JMMD, revealing why JMMD degrades the feature discrimination. This motivates us to propose a novel loss JMMD-HSIC by jointly considering JMMD and HSIC to promote discrimination of JMMD. Extensive experiments on several cross-domain datasets could demonstrate the validity of our revealed theoretical results and the effectiveness of our proposed JMMD-HSIC.



Paperid:1980
Authors:Sangkyu Lee, Janghoon Han, Hosung Song, Stanley Jungkyu Choi, Honglak Lee, Youngjae Yu
Abstract:
Abstract:Direct Preference Optimization (DPO) demonstrates the advantage of aligning a large language model with human preference using only an offline dataset. However, DPO has the limitation that the KL penalty, which prevents excessive deviation from the reference model, is static throughout the training process. Several methods claim to change this static KL penalty of DPO into a dynamic one, but no approach can adaptively assign different KL penalties for each preference pair. In this paper, we propose $\varepsilon$-Direct Preference Optimization ($\varepsilon$-DPO), which allows adaptive control of the KL penalty strength $\beta$ for each preference pair. Specifically, $\varepsilon$-DPO adaptively controls $\beta$ for each preference pair based on the monotonicity of logits as a preference model under the perturbation of $\beta$ during training. This is equivalent to adjusting the KL penalty by checking whether the change in training-time temperature can lead to better preference confidence as preference models by simply reusing the logit of the current policy and the reference policy. Experimental results show that the simple criterion of $\varepsilon$-DPO for KL penalty relaxation significantly improves DPO compared to most existing direct alignment algorithms on general chatbot benchmarks and reveal that this adaptive KL penalty control criterion can reflect confusion as a preference model and provide an efficient KL trade-off, highlighting the significance of instance-level adaptive KL penalty control in DPO.



Paperid:1981
Authors:Shaowei Liu, David Yao, Saurabh Gupta, Shenlong Wang
Abstract:
Today, people can easily record memorable moments, ranging from concerts, sports events, lectures, family gatherings, and birthday parties with multiple consumer cameras. However, synchronizing these cross‑camera streams remains challenging. Existing methods assume controlled settings, specific targets, manual correction, or costly hardware. We present Visual Sync, an optimization framework based on multi‑view dynamics that aligns unposed, unsynchronized videos at millisecond accuracy. Our key insight is that any moving 3D point, when co‑visible in two cameras, obeys epipolar constraints once properly synchronized. To exploit this, Visual Sync leverages off‑the‑shelf 3D reconstruction, feature matching, and dense tracking to extract tracklets, relative poses, and cross‑view correspondences. It then jointly minimizes the epipolar error to estimate each camera’s time offset. Experiments on four diverse, challenging datasets show that Visual Sync outperforms baseline methods, achieving an average synchronization error below 130 ms.



Paperid:1982
Authors:Pedram Khorsandi, Rushil Gupta, Mehrnaz Mofakhami, Simon Lacoste-Julien, Gauthier Gidel
Abstract:
Performative prediction is a framework accounting for the shift in the data distribution induced by the prediction of a model deployed in the real world. Ensuring convergence to a stable solution—one at which the post‑deployment data distribution no longer changes—is crucial in settings where model predictions can influence future data. This paper, for the first time, extends the Repeated Risk Minimization (RRM) algorithm class by utilizing historical datasets from previous retraining snapshots, yielding a class of algorithms that we call Affine Risk Minimizers that converges to a performatively stable point for a broader class of problems. We introduce a new upper bound for methods that use only the final iteration of the dataset and prove for the first time the tightness of both this new bound and the previous existing bounds within the same regime. We also prove that our new algorithm class can surpass the lower bound for standard RRM, thus breaking the prior lower bound, and empirically observe faster convergence to the stable point on various performative prediction benchmarks. We offer at the same time the first lower bound analysis for RRM within the class of Affine Risk Minimizers, quantifying the potential improvements in convergence speed that could be achieved with other variants in our scheme.



Paperid:1983
Authors:Yigit Korkmaz, Urvi Bhuwania, Ayush Jain, Erdem Bıyık
Abstract:
Value-based algorithms are a cornerstone of off-policy reinforcement learning due to their simplicity and training stability. However, their use has traditionally been restricted to discrete action spaces, as they rely on estimating Q-values for individual state-action pairs. In continuous action spaces, evaluating the Q-value over the entire action space becomes computationally infeasible. To address this, actor-critic methods are typically employed, where a critic is trained on off-policy data to estimate Q-values, and an actor is trained to maximize the critic's output. Despite their popularity, these methods often suffer from instability during training. In this work, we propose a purely value-based framework for continuous control that revisits structural maximization of Q-functions, introducing a set of key architectural and algorithmic choices to enable efficient and stable learning. We evaluate the proposed actor-free Q-learning approach on a range of standard simulation tasks, demonstrating performance and sample-efficiency on par with state-of-the-art baselines, without the cost of learning a separate actor. Particularly, in environments with constrained action spaces, where the value functions are typically non-smooth, our method with structural maximization outperforms traditional actor-critic methods with gradient-based maximization.



Paperid:1984
Authors:Xavier Gonzalez, Leo Kozachkov, Kenneth Clarkson, David Zoltowski, Scott Linderman
Abstract:
Abstract:The rise of parallel computing hardware has made it increasingly important to understand which nonlinear state space models can be efficiently parallelized. Recent advances have shown that evaluating a state space model can be recast as solving a parallelizable optimization problem, and sometimes this approach yields dramatic speed-ups in evaluation time. However, the factors that govern the difficulty of these optimization problems remain unclear, limiting the larger adoption of the technique. In this work, we establish a precise relationship between the dynamics of a nonlinear system and the conditioning of its corresponding optimization formulation. We show that the predictability of a system, defined as the degree to which small perturbations in state influence future behavior, directly governs the number of optimization steps required for evaluation. In predictable systems, the state trajectory can be computed in $\mathcal{O}((\log T)^2)$ time, where $T$ is the sequence length, a major improvement over the conventional sequential approach. In contrast, chaotic or unpredictable systems exhibit poor conditioning, with the consequence that parallel evaluation converges too slowly to be useful. Importantly, our theoretical analysis demonstrates that for predictable systems, the optimization problem is always well-conditioned, whereas for unpredictable systems, the conditioning degrades exponentially as a function of the sequence length. We validate our claims through extensive experiments, providing practical guidance on when nonlinear dynamical systems can be efficiently parallelized, and highlighting predictability as a key design principle for parallelizable models.



Paperid:1985
Authors:Xiaoyu Kong, Junguang Jiang, Bin Liu, Ziru Xu, Han Zhu, Jian Xu, Bo Zheng, Jiancan Wu, Xiang Wang
Abstract:
The core task of recommender systems is to learn user preferences from historical user-item interactions. With the rapid development of large language models (LLMs), recent research has explored leveraging the reasoning capabilities of LLMs to enhance rating prediction tasks. However, existing distillation-based methods suffer from limitations such as the teacher model's insufficient recommendation capability, costly and static supervision, and superficial transfer of reasoning ability. To address these issues, this paper proposes RecZero, a reinforcement learning (RL)-based recommendation paradigm that abandons the traditional multi-model and multi-stage distillation approach. Instead, RecZero trains a single LLM through pure RL to autonomously develop reasoning capabilities for rating prediction. RecZero consists of two key components: (1) "Think-before-Recommendation" prompt construction, which employs a structured reasoning template to guide the model in step-wise analysis of user interests, item features, and user-item compatibility; and (2) rule-based reward modeling, which adopts group relative policy optimization (GRPO) to compute rewards for reasoning trajectories and optimize the LLM. Additionally, the paper explores a hybrid paradigm, RecOne, which combines supervised fine-tuning with RL, initializing the model with cold-start reasoning samples and further optimizing it with RL. Experimental results demonstrate that RecZero and RecOne significantly outperform existing baseline methods on multiple benchmark datasets, validating the superiority of the RL paradigm in achieving autonomous reasoning-enhanced recommender systems.



Paperid:1986
Authors:Gengyu Wang, Runze Zhang, Zhongzhi Zhang
Abstract:
Public opinion governance in social networks is critical for public health campaigns, political elections, and commercial marketing. In this paper, we addresse the problem of maximizing overall opinion in social networks by strategically modifying the internal opinions of key nodes. Traditional matrix inversion methods suffer from prohibitively high computational costs, prompting us to propose two efficient sampling-based algorithms. Furthermore, we develop a deterministic asynchronous algorithm that exactly identifies the optimal set of nodes through asynchronous update operations and progressive refinement, ensuring both efficiency and precision. Extensive experiments on real-world datasets demonstrate that our methods outperform baseline approaches. Notably, our asynchronous algorithm delivers exceptional efficiency and accuracy across all scenarios, even in networks with tens of millions of nodes.



Paperid:1987
Authors:Tongyao Zhu, Qian Liu, Haonan Wang, Shiqi Chen, Xiangming Gu, Tianyu Pang, Min-Yen Kan
Abstract:
Recent advancements in LLM pretraining have featured ever-expanding context windows to process longer sequences. However, our controlled study reveals that models pretrained with shorter context windows consistently outperform their long-context counterparts under a fixed token budget. This finding motivates us to explore an optimal context window scheduling strategy to better balance long-context capability with pretraining efficiency. To this end, we propose SkyLadder, a simple yet effective approach that implements a short-to-long context window transition. SkyLadder preserves strong standard benchmark performance, while matching or exceeding baseline results on long-context tasks. Through extensive experiments, we pretrain 1B-parameter models (up to 32K context) and 3B-parameter models (8K context) on 100B tokens, demonstrating that SkyLadder yields consistent gains of up to 3.7% on common benchmarks, while achieving up to 22% faster training speeds compared to baselines.



Paperid:1988
Authors:Yunhong Min, Daehyeon Choi, Kyeongmin Yeo, Jihyun Lee, Minhyuk Sung
Abstract:
We introduce ORIGEN, the first zero-shot method for 3D orientation grounding in text-to-image generation across multiple objects and diverse categories. While previous work on spatial grounding in image generation has mainly focused on 2D positioning, it lacks control over 3D orientation.To address this, we propose a reward-guided sampling approach using a pretrained discriminative model for 3D orientation estimation and a one-step text-to-image generative flow model. While gradient-ascent-based optimization is a natural choice for reward-based guidance, it struggles to maintain image realism. Instead, we adopt a sampling-based approach using Langevin dynamics, which extends gradient ascent by simply injecting random noise—requiring just a single additional line of code.Additionally, we introduce adaptive time rescaling based on the reward function to accelerate convergence. Our experiments show that \textsc{Origen} outperforms both training-based and test-time guidance methods across quantitative metrics and user studies.



Paperid:1989
Authors:Robi Bhattacharjee, Geelon So, Sanjoy Dasgupta
Abstract:
Abstract:In the adaptive sampling model of online learning, future prediction tasks can depend on the past. At every round, an adversary selects an instance to test the learner, and after a prediction is made, a noisy label is drawn from an underlying conditional label distribution and is observed by both learner and adversary. A learner is consistent if it eventually performs no worse than the Bayes optimal predictor. In particular, we study the $k_n$-nearest neighbor learner (Fix and Hodges 1951). In the worst-case setting, the $k_n$-nearest neighbor rule fails to learn because an adaptive process can generate spurious patterns out of noise; however, under the mild assumption that the process generating the instances is uniformly absolutely continuous and that the underlying conditional label distribution is continuous, we show that $k_n$-nearest neighbor rule is consistent.



Paperid:1990
Authors:Manish Shetty, Naman Jain, Jinjian Liu, Vijay Kethanaboyina, Koushik Sen, Ion Stoica
Abstract:
Developing high-performance software is a complex task that requires specialized expertise. We introduce GSO, a benchmark for evaluating language models' capabilities in developing high-performance software.We develop an automated pipeline that generates and executes performance tests to analyze repository commit histories to identify 105 challenging optimization tasks across 10 codebases, spanning diverse domains and programming languages.An agent is provided with a codebase and performance test as a precise specification, and tasked to improve the runtime efficiency, which is measured against the expert developer optimization.Our quantitative evaluation reveals that leading SWE-Agents struggle significantly, achieving less than 5% success rate, with limited improvements even with inference-time scaling.Our qualitative analysis identifies key failure modes, including difficulties with low-level languages, practicing lazy optimization strategies, and challenges in accurately localizing bottlenecks.We release the code and artifacts of our benchmark along with agent trajectories to enable future research.



Paperid:1991
Authors:Yu-Jie Zhang, Sheng-An Xu, Peng Zhao, Masashi Sugiyama
Abstract:
Abstract:We study the generalized linear bandit (GLB) problem, a contextual multi-armed bandit framework that extends the classical linear model by incorporating a non-linear link function, thereby modeling a broad class of reward distributions such as Bernoulli and Poisson. While GLBs are widely applicable to real-world scenarios, their non-linear nature introduces significant challenges in achieving both computational and statistical efficiency. Existing methods typically trade off between two objectives, either incurring high per-round costs for optimal regret guarantees or compromising statistical efficiency to enable constant-time updates. In this paper, we propose a jointly efficient algorithm that attains a nearly optimal regret bound with $\mathcal{O}(1)$ time and space complexities per round. The core of our method is a tight confidence set for the online mirror descent (OMD) estimator, which is derived through a novel analysis that leverages the notion of mix loss from online prediction. The analysis shows that our OMD estimator, even with its one-pass updates, achieves statistical efficiency comparable to maximum likelihood estimation, thereby leading to a jointly efficient optimistic method.



Paperid:1992
Authors:Emile Anand, Ishani Karmarkar, Guannan Qu
Abstract:
Abstract:Designing efficient algorithms for multi-agent reinforcement learning (MARL) is fundamentally challenging because the size of the joint state and action spaces grows exponentially in the number of agents. These difficulties are exacerbated when balancing sequential global decision-making with local agent interactions. In this work, we propose a new algorithm $\texttt{SUBSAMPLE-MFQ}$ ($\textbf{Subsample}$-$\textbf{M}$ean-$\textbf{F}$ield-$\textbf{Q}$-learning) and a decentralized randomized policy for a system with $n$ agents. For any $k\leq n$, our algorithm learns a policy for the system in time polynomial in $k$. We prove that this learned policy converges to the optimal policy on the order of $\tilde{O}(1/\sqrt{k})$ as the number of subsampled agents $k$ increases. In particular, this bound is independent of the number of agents $n$.



Paperid:1993
Authors:Eunji Kim, Sriya Mantena, Weiwei Yang, Chandan Singh, Sungroh Yoon, Jianfeng Gao
Abstract:
While large transformer models excel in predictive performance, their lack of interpretability restricts their usefulness in high-stakes domains. To remedy this, we propose the Generalized Induction-Head Model (GIM), an interpretable model for next-token prediction inspired by the observation of “induction heads” in LLMs. GIM is a retrieval-based module that identifies similar sequences in the input context by combining exact n-gram matching and fuzzy matching based on a neural similarity metric. We evaluate GIM in two settings: language modeling and fMRI response prediction. In language modeling, GIM improves next-token prediction by up to 25%p over interpretable baselines, significantly narrowing the gap with black-box LLMs. In an fMRI setting, GIM improves neural response prediction by 20% and offers insights into the language selectivity of the brain. GIM represents a significant step toward uniting interpretability and performance across domains.



Paperid:1994
Authors:Yang Cao, Xiaoyu Li, Zhao Song, Chiwun Yang
Abstract:
Abstract:We consider the task of interpolating a band–limited signal from a small collection of noisy time-domain samples. Exploiting a new analytic framework for hierarchical frequency decomposition that performs systematic noise cancellation, we give the first polynomial-time algorithm with a provable $(3+\sqrt{2}+\epsilon)$-approximation guarantee for continuous interpolation. Our method breaks the long-standing $C > 100$ barrier set by the best previous algorithms, sharply reducing the gap to optimal recovery and establishing a new state of the art for high-accuracy band–limited interpolation. We also give a refined ``shrinking-range'' variant that achieves a $(\sqrt{2}+\varepsilon+c)$-approximation on any sub-interval $(1-c)T$ for some $c \in (0,1)$, which gives even higher interpolation accuracy.



Paperid:1995
Authors:Ho Yin Au, Jie Chen, Junkun Jiang, Jingyu Xiang
Abstract:
Recent research on motion generation has shown significant progress in generating semantically aligned motion with singular semantics. However, when employing these models to create composite sequences containing multiple semantically generated motion clips, they often struggle to preserve the continuity of motion dynamics at the transition boundaries between clips, resulting in awkward transitions and abrupt artifacts. To address these challenges, we present Compositional Phase Diffusion, which leverages the Semantic Phase Diffusion Module (SPDM) and Transitional Phase Diffusion Module (TPDM) to progressively incorporate semantic guidance and phase details from adjacent motion clips into the diffusion process. Specifically, SPDM and TPDM operate within the latent motion frequency domain established by the pre-trained Action-Centric Motion Phase Autoencoder (ACT-PAE). This allows them to learn semantically important and transition-aware phase information from variable-length motion clips during training. Experimental results demonstrate the competitive performance of our proposed framework in generating compositional motion sequences that align semantically with the input conditions, while preserving phase transitional continuity between preceding and succeeding motion clips. Additionally, motion inbetweening task is made possible by keeping the phase parameter of the input motion sequences fixed throughout the diffusion process, showcasing the potential for extending the proposed framework to accommodate various application scenarios.



Paperid:1996
Authors:Weiqing He, Xiang Li, Tianqi Shang, Li Shen, Weijie Su, Qi Long
Abstract:
Large language models (LLMs) raise concerns about content authenticity and integrity because they can generate human-like text at scale. Text watermarks, which embed detectable statistical signals into generated text, offer a provable way to verify content origin. Many detection methods rely on pivotal statistics that are i.i.d. under human-written text, making goodness-of-fit (GoF) tests a natural tool for watermark detection. However, GoF tests remain largely underexplored in this setting.In this paper, we systematically evaluate eight GoF tests across three popular watermarking schemes, using three open-source LLMs, two datasets, various generation temperatures, and multiple post-editing methods. We find that general GoF tests can improve both the detection power and robustness of watermark detectors. Notably, we observe that text repetition, common in low-temperature settings, gives GoF tests a unique advantage not exploited by existing methods.Our results highlight that classic GoF tests are a simple yet powerful and underused tool for watermark detection in LLMs.



Paperid:1997
Authors:Kevin Aydin, Honghao Lin, David Woodruff, Peilin Zhong
Abstract:
Abstract:$k$-anonymity is a widely-used privacy-preserving concept that ensures each record in a dataset is indistinguishable from at least $k-1$ other records. In this paper, we revisit $k$-anonymity by suppression and give an $O(k)$-approximation algorithm with a nearly-linear runtime of $\tilde{O}(nd + n^{1+1/C^2}/k^{1/C^2})$ for an arbitrary constant $C$, where $n$ is the number of records and $d$ is the number of attributes. Previous algorithms with provable guarantees either (1) achieve the same $O(k)$ approximation ratio but require at least $O(n^2 k)$ runtime, or (2) provide a better $O(\log k)$ approximation ratio at the cost of an impractical $O(n^{2k})$ worst-case runtime for general $d$ and $k$. Our algorithm extends to the Massively Parallel Computation (MPC) model, where it can be adapted into an MPC algorithm requiring $\tilde{O}(\log^{1+\epsilon} n)$ rounds and total space $O(n^{1+1/C^2}(d+k))$. Empirical evaluations on real-world datasets demonstrate that our algorithm performs comparably to existing fast heuristic approaches.Although~\citep{PS07} introduced improvements to achieve more practical runtimes for their $O(\log k)$-approximation algorithm, its worst-case runtime remains $O(n^{2k})$. A natural question arises: can we develop an algorithm with an $o(k)$ approximation ratio and a polynomial runtime? We investigate the single-point $k$-anonymity problem, where the goal is to select $k-1$ additional records to make a given record indistinguishable. Surprisingly, assuming the dense vs random conjecture in complexity theory, we show that for $n = k^c$, no algorithm can achieve a $k^{1 - O(1/c)}$ approximation in $\mathrm{poly}(k)$ time. This provides evidence of the inherent hardness of the $k$-anonymity problem.



Paperid:1998
Authors:Zhengren Wang, Rui ling, Chufan Wang, Yongan Yu, Sizhe Wang, Zhiyu li, Feiyu Xiong, Wentao Zhang
Abstract:
Modern code generation has made significant strides in functional correctness and execution efficiency. However, these systems often overlook a critical dimension in real-world software development: \textit{maintainability}. To handle dynamic requirements with minimal rework, we propose \textbf{MaintainCoder} as a pioneering solution. It integrates the Waterfall model, design patterns, and multi-agent collaboration to systematically enhance cohesion, reduce coupling, achieving clear responsibility boundaries and better maintainability. We also introduce \textbf{MaintainBench}, a benchmark comprising requirement changes and novel dynamic metrics on maintenance efforts. Experiments demonstrate that existing code generation methods struggle to meet maintainability standards when requirements evolve. In contrast, MaintainCoder improves dynamic maintainability metrics by more than 60\% with even higher correctness of initial codes. Furthermore, while static metrics fail to accurately reflect maintainability and even contradict each other, our proposed dynamic metrics exhibit high consistency. Our work not only provides the foundation for maintainable code generation, but also highlights the need for more realistic and comprehensive code generation research.



Paperid:1999
Authors:Runyang You, Yongqi Li, Xinyu Lin, Xin Zhang, Wenjie Wang, Wenjie Li, Liqiang Nie
Abstract:
Abstract:Recent advances in Large Recommender Models extend LLMs for recommendation tasks via encoding or item generation,while reasoning ability of LLM is typically utilized as an external module to produce extra inputs or features for conventional architectures, resulting in misaligned optimization and underutilized LLM capacity. We propose $\text{R}^2\text{ec}$, a unified large recommender model that interleaves reasoning and recommendation in a single autoregressive pass. With a prompt encoding user history, the model generates a reasoning chain, whose hidden state is used to score items—seamlessly fusing generative reasoning and discriminative ranking.To train this architecture, we introduce a policy optimization framework that treats reasoning and item prediction as sequential actions in one trajectory, enabling joint optimization of reasoning and recommendation.Experiments on three public datasets show that $\text{R}^2\text{ec}$ significantly outperforms existing LLM-based baselines, achieving up to 50\% gains in NDCG@5.



Paperid:2000
Authors:Xue zhucun, Jiangning Zhang, Teng Hu, Haoyang He, Yinan Chen, Yuxuan Cai, Yabiao Wang, Chengjie Wang, Yong Liu, Xiangtai Li, Dacheng Tao
Abstract:
The quality of the video dataset (image quality, resolution, and fine-grained caption) greatly influences the performance of the video generation model. %The growing demand for video applications sets higher requirements for high-quality video generation models. %For example, the generation of movie-level Ultra-High Definition (UHD) videos and the creation of 4K short video content. %However, the existing public datasets cannot support related research and applications. %In this paper, we first propose a high-quality open-sourced UHD-4K (22.4\% of which are 8K) text-to-video dataset named UltraVideo, which contains a wide range of topics (more than 100 kinds), and each video has 9 structured captions with one summarized caption (average of 824 words). %Specifically, we carefully design a highly automated curation process with four stages to obtain the final high-quality dataset: \textit{i)} collection of diverse and high-quality video clips. \textit{ii)} statistical data filtering. \textit{iii)} model-based data purification. \textit{iv)} generation of comprehensive, structured captions. %In addition, we expand WAN to UltraWAN-1K/-4K, which can natively generate high-quality 1K/4K videos with more consistent text controllability, demonstrating the effectiveness of our data curation.%We believe that this work can make a significant contribution to future research on UHD video generation. UltraVideo dataset and UltraWAN models are available at \href{https://zhangzjn.github.io/projects/UltraVideo}{project page}.



Paperid:2001
Authors:Oleksii Furman, Ulvi Movsum-zada, Patryk Marszałek, Maciej Zieba, Marek Śmieja
Abstract:
Counterfactual explanations play a pivotal role in explainable artificial intelligence (XAI) by offering intuitive, human-understandable alternatives that elucidate machine learning model decisions. Despite their significance, existing methods for generating counterfactuals often require constant access to the predictive model, involve computationally intensive optimization for each instance, and lack the flexibility to adapt to new user-defined constraints without retraining. In this paper, we propose DiCoFlex, a novel model-agnostic, conditional generative framework that produces multiple diverse counterfactuals in a single forward pass. Leveraging conditional normalizing flows trained solely on labeled data, DiCoFlex addresses key limitations by enabling real-time, user-driven customization of constraints such as sparsity and actionability at inference time. Extensive experiments on standard benchmark datasets show that DiCoFlex outperforms existing methods in terms of validity, diversity, proximity, and constraint adherence, making it a practical and scalable solution for counterfactual generation in sensitive decision-making domains.



Paperid:2002
Authors:Huikang Su, Dengyun Peng, Zifeng Zhuang, Yuhan Liu, Qiguang Chen, Donglin Wang, Qinghe Lli
Abstract:
Offline safe reinforcement learning aims to learn policies that satisfy predefined safety constraints from static datasets. Existing sequence-model-based methods condition action generation on symmetric input tokens for return-to-go and cost-to-go, neglecting their intrinsic asymmetry: RTG serves as a flexible performance target, while CTG should represent a rigid safety boundary. This symmetric conditioning leads to unreliable constraint satisfaction, especially when encountering out-of-distribution cost trajectories. To address this, we propose Boundary-to-Region (B2R), a framework that enables asymmetric conditioning through cost signal realignment . B2R redefines CTG as a boundary constraint under a fixed safety budget, unifying the cost distribution of all feasible trajectories while preserving reward structures. Combined with rotary positional embeddings , it enhances exploration within the safe region. Experimental results show that B2R satisfies safety constraints in 35 out of 38 safety-critical tasks while achieving superior reward performance over baseline methods. This work highlights the limitations of symmetric token conditioning and establishes a new theoretical and practical approach for applying sequence models to safe RL.



Paperid:2003
Authors:Senne Berden, Ali Mahmutoğulları, Dimos Tsouros, Tias Guns
Abstract:
Mathematical optimization is a fundamental tool for decision-making in a wide range of applications. However, in many real-world scenarios, the parameters of the optimization problem are not known a priori and must be predicted from contextual features. This gives rise to predict-then-optimize problems, where a machine learning model predicts problem parameters that are then used to make decisions via optimization. A growing body of work on decision-focused learning (DFL) addresses this setting by training models specifically to produce predictions that maximize downstream decision quality, rather than accuracy. While effective, DFL is computationally expensive, because it requires solving the optimization problem with the predicted parameters at each loss evaluation. In this work, we address this computational bottleneck for linear optimization problems, a common class of problems in both DFL literature and real-world applications. We propose a solver-free training method that exploits the geometric structure of linear optimization to enable efficient training with minimal degradation in solution quality. Our method is based on the insight that a solution is optimal if and only if it achieves an objective value that is at least as good as that of its adjacent vertices on the feasible polytope. Building on this, our method compares the estimated quality of the ground-truth optimal solution with that of its precomputed adjacent vertices, and uses this as loss function. Experiments demonstrate that our method significantly reduces computational cost while maintaining high decision quality.



Paperid:2004
Authors:Weiwei Sun, Haokun Liu, Nikhil Kandpal, Colin Raffel, Yiming Yang
Abstract:
Training data attribution (TDA) methods aim to measure how training data impacts a model's predictions. While gradient-based attribution methods, such as influence functions, offer theoretical rigor, their computational costs make them impractical for large-scale applications. Representation-based attribution methods are more efficient, relying on similarity computations between examples in some representation space, but they often lack task-aware and model-specific optimization, limiting their accuracy. To address these challenges, we propose AirRep, a novel representation-based approach that enhances representation quality through task-driven optimization of a representation encoding model.Furthermore, we extend this method beyond single-sample attribution using an attention-based pooling mechanism to effectively estimate the collective influence of groups of samples.Experiments on instruction tuning of large language models demonstrate that AirRep achieves performance on par with state-of-the-art gradient-based approaches while being nearly two orders of magnitude more efficient. Further analysis highlights its robustness, including generalization to new data and new TDA tasks.



Paperid:2005
Authors:Tsubasa Masumura, Masato Taki
Abstract:
Associative memory models based on Hopfield networks and self-attention based on key-value mechanisms have been popular approaches in the study of memory mechanisms in deep learning. It has been pointed out that the state update rule of the modern Hopfield network (MHN) in the adiabatic approximation is in agreement with the self-attention layer of Transformer. In this paper, we go beyond this approximation and investigate the relationship between MHN and self-attention. Our results show that the correspondence between Hopfield networks and Transformers can be established in a more generalized form by adding a new variable, the hidden state derived from the MHN, to self-attention. This new attention mechanism, modern Hopfield attention (MHA), allows the inheritance of attention scores from the input layer of the Transformer to the output layer, which greatly improves the nature of attention weights. In particular, we show both theoretically and empirically that MHA hidden states significantly improve serious problem of deep Transformers known as rank collapse and token uniformity. We also confirm that MHA can systematically improve accuracy without adding training parameters to the Vision Transformer or GPT. Our results provide a new case in which Hopfield networks can be a useful perspective for improving the Transformer architecture.



Paperid:2006
Authors:Haiduo Huang, Jiangcheng Song, Yadong Zhang, Pengju Ren
Abstract:
Recent advances in knowledge distillation have emphasized the importance of decoupling different knowledge components. While existing methods utilize momentum mechanisms to separate task-oriented and distillation gradients, they overlook the inherent conflict between target-class and non-target-class knowledge flows. Furthermore, low-confidence dark knowledge in non-target classes introduces noisy signals that hinder effective knowledge transfer. To address these limitations, we propose DeepKD, a novel training framework that integrates dual-level decoupling with adaptive denoising. First, through theoretical analysis of gradient signal-to-noise ratio (GSNR) characteristics in task-oriented and non-task-oriented knowledge distillation, we design independent momentum updaters for each component to prevent mutual interference. We observe that the optimal momentum coefficients for task-oriented gradient (TOG), target-class gradient (TCG), and non-target-class gradient (NCG) should be positively related to their GSNR. Second, we introduce a dynamic top-k mask (DTM) mechanism that gradually increases K from a small initial value to incorporate more non-target classes as training progresses, following curriculum learning principles. The DTM jointly filters low-confidence logits from both teacher and student models, effectively purifying dark knowledge during early training. Extensive experiments on CIFAR-100, ImageNet, and MS-COCO demonstrate DeepKD's effectiveness.



Paperid:2007
Authors:Ziqiang Li, Jiazhen Yan, Ziwen He, Kai Zeng, Weiwei Jiang, Lizhi Xiong, Zhangjie Fu
Abstract:
The rapid advancement of generative models, such as GANs and Diffusion models, has enabled the creation of highly realistic synthetic images, raising serious concerns about misinformation, deepfakes, and copyright infringement. Although numerous Artificial Intelligence Generated Image (AIGI) detectors have been proposed, often reporting high accuracy, their effectiveness in real-world scenarios remains questionable. To bridge this gap, we introduce AIGIBench, a comprehensive benchmark designed to rigorously evaluate the robustness and generalization capabilities of state-of-the-art AIGI detectors. AIGIBench simulates real-world challenges through four core tasks: multi-source generalization, robustness to image degradation, sensitivity to data augmentation, and impact of test-time pre-processing. It includes 23 diverse fake image subsets that span both advanced and widely adopted image generation techniques, along with real-world samples collected from social media and AI art platforms. Extensive experiments on 11 advanced detectors demonstrate that, despite their high reported accuracy in controlled settings, these detectors suffer significant performance drops on real-world data, limited benefits from common augmentations, and nuanced effects of pre-processing, highlighting the need for more robust detection strategies. By providing a unified and realistic evaluation framework, AIGIBench offers valuable insights to guide future research toward dependable and generalizable AIGI detection.



Paperid:2008
Authors:Siyuan Huang, Liliang Chen, Pengfei Zhou, Shengcong Chen, Yue Liao, Zhengkai Jiang, Yue Hu, Peng Gao, Hongsheng Li, Maoqing Yao, Guanghui Ren
Abstract:
We introduce EnerVerse, a generative robotics foundation model that constructs and interprets embodied spaces. EnerVerse employs a chunk-wise autoregressive video diffusion framework to predict future embodied spaces from instructions, enhanced by a sparse context memory for long-term reasoning. To model the 3D robotics world, we adopt a multi-view video representation, providing rich perspectives to address challenges like motion ambiguity and 3D grounding. Additionally, EnerVerse-D, a data engine pipeline combining generative modeling with 4D Gaussian Splatting, forms a self-reinforcing data loop to reduce the sim-to-real gap. Leveraging these innovations, EnerVerse translates 4D world representations into physical actions via a policy head (EnerVerse-A), achieving state-of-the-art performance in both simulation and real-world tasks.



Paperid:2009
Authors:Ivan Evtimov, Arman Zharmagambetov, Aaron Grattafiori, Chuan Guo, Kamalika Chaudhuri
Abstract:
Autonomous UI agents powered by AI have tremendous potential to boost human productivity by automating routine tasks such as filing taxes and paying bills. However, a major challenge in unlocking their full potential is security, which is exacerbated by the agent's ability to take action on their user's behalf. Existing tests for prompt injections in web agents either over-simplify the threat by testing unrealistic scenarios or giving the attacker too much power, or look at single-step isolated tasks. To more accurately measure progress for secure web agents, we introduce WASP – a new publicly available benchmark for end-to-end evaluation of Web Agent Security against Prompt Injection attacks. Evaluating with WASP shows that even top-tier AI models, including those with advanced reasoning capabilities, can be deceived by simple, low-effort human-written injections in very realistic scenarios. Our end-to-end evaluation reveals a previously unobserved insight: while attacks partially succeed in up to 86% of the case, even state-of-the-art agents often struggle to fully complete the attacker goals – highlighting the current state of security by incompetence.



Paperid:2010
Authors:Sungik Choi, Moontae Lee, Hankook Lee
Abstract:
Abstract:AI-generated image detection has become crucial with the rapid advancement of vision-generative models. Instead of training detectors tailored to specific datasets, we study a training-free approach leveraging self-supervised models without requiring prior data knowledge. These models, pre-trained with augmentations like $\texttt{RandomResizedCrop}$, learn to produce consistent representations across varying resolutions. Motivated by this, we propose $\textbf{WaRPAD},$ a training-free AI-generated image detection algorithm based on self-supervised models. Since neighborhood pixel differences in images are highly sensitive to resizing operations, WaRPAD first defines a base score function that quantifies the sensitivity of image embeddings to perturbations along high-frequency directions extracted via Haar wavelet decomposition. To simulate robustness against cropping augmentation, we rescale each image to a multiple of the model’s input size, divide it into smaller patches, and compute the base score for each patch. The final detection score is then obtained by averaging the scores across all patches. We validate WaRPAD on real datasets of diverse resolutions and domains, and images generated by 23 different generative models. Our method consistently achieves competitive performance and demonstrates strong robustness to test-time corruptions. Furthermore, as invariance to $\texttt{RandomResizedCrop}$ is a common training scheme across self-supervised models, we show that WaRPAD is applicable across self-supervised models.



Paperid:2011
Authors:Gianluca Galletti, Fabian Paischer, William Hornsby, Paul Setinek, Lorenzo Zanisi, Naomi Carey, Stanislas Pamela, Johannes Brandstetter
Abstract:
Nuclear fusion plays a pivotal role in the quest for reliable and sustainable energy production. A major roadblock to viable fusion power is understanding plasma turbulence, which significantly impairs plasma confinement, and is crucial for modelling plasma scenarios for next-generation reactor design. Plasma turbulence is governed by the nonlinear gyrokinetic equation, which tracks a 5D distribution function over time, but its high computational cost leads to the use of reduced models for turbulent transport. However, these models omit nonlinear effects unique to the full 5D dynamics. To this end, we introduce GyroSwin -- the first 5D neural surrogate for modelling nonlinear gyrokinetic simulations that can capture physical phenomena neglected by reduced models. GyroSwin (i) extends a hierarchical Vision Transformer to 5D and is trained on the 5D distribution function for the adiabatic electron approximation, (ii) introduces latent cross-attention and integration modules for 3D-5D interactions between potential fields and the distribution function, and (iii) performs channelwise mode separation based on nonlinear physics. Experimentally, we show that GyroSwin captures nonlinear physics such as zonal flows and significantly improves flux predictions over reduced numerical models in most cases. GyroSwin has the potential to drastically reduce the complexity of plasma turbulence simulation in 5D.



Paperid:2012
Authors:Xu Shi, Rufeng Xiao, Rujun Jiang
Abstract:
Abstract:Existing methods for solving Riemannian bilevel optimization (RBO) problems require prior knowledge of the problem's first- and second-order information and curvature parameter of the Riemannian manifold to determine step sizes, which poses practical limitations when these parameters are unknown or computationally infeasible to obtain. In this paper, we introduce the Adaptive Riemannian Hypergradient Descent (AdaRHD) algorithm for solving RBO problems. To our knowledge, AdaRHD is the first method to incorporate a fully adaptive step size strategy that eliminates the need for problem-specific parameters in RBO problem resolution. We prove that AdaRHD achieves an $\mathcal{O}(1/\epsilon)$ iteration complexity for finding an $\epsilon$-stationary point, thus matching the complexity of existing non-adaptive methods. Furthermore, we demonstrate that substituting exponential mappings with retraction mappings maintains the same complexity bound. Experiments demonstrate that AdaRHD achieves comparable performance to existing non-adaptive approaches while exhibiting greater robustness.



Paperid:2013
Authors:Beier Luo, Shuoyuan Wang, Sharon Li, Hongxin Wei
Abstract:
Abstract:Post-training of large language models is essential for adapting pre-trained language models (PLMs) to align with human preferences and downstream tasks. While PLMs typically exhibit well-calibrated confidence, post-trained language models (PoLMs) often suffer from over-confidence, assigning high confidence to both correct and incorrect outputs, which can undermine reliability in critical applications.A major obstacle in calibrating PoLMs is the scarcity of labeled data for individual downstream tasks. To address this, we propose Disagreement-Aware Confidence Alignment (DACA), a novel unsupervised method to optimize the parameters (e.g., temperature $\tau$) in post-hoc confidence calibration. Our method is motivated by the under-confidence issue caused by prediction disagreement between the PLM and PoLM while aligning their confidence via temperature scaling. Theoretically, the PLM's confidence underestimates PoLM's prediction accuracy on disagreement examples, causing a larger $\tau$ and producing under-confident predictions. DACA mitigates this by selectively using only agreement examples for calibration, effectively decoupling the influence of disagreement.In this manner, our method avoids an overly large $\tau$ in temperature scaling caused by disagreement examples, improving calibration performance.Extensive experiments demonstrate the effectiveness of our method, improving the average ECE of open-sourced and API-based LLMs (e.g. GPT-4o) by up to 15.08$\%$ on common benchmarks.



Paperid:2014
Authors:Jaeseung Heo, Kyeongheung Yun, Seokwon Yoon, MoonJeong Park, Jungseul Ok, Dongwoo Kim
Abstract:
Understanding how individual edges influence the behavior of graph neural networks (GNNs) is essential for improving their interpretability and robustness. Graph influence functions have emerged as promising tools to efficiently estimate the effects of edge deletions without retraining. However, existing influence prediction methods rely on strict convexity assumptions, exclusively consider the influence of edge deletions while disregarding edge insertions, and fail to capture changes in message propagation caused by these modifications. In this work, we propose a proximal Bregman response function specifically tailored for GNNs, relaxing the convexity requirement and enabling accurate influence prediction for standard neural network architectures. Furthermore, our method explicitly accounts for message propagation effects and extends influence prediction to both edge deletions and insertions in a principled way. Experiments with real-world datasets demonstrate accurate influence predictions for different characteristics of GNNs. We further demonstrate that the influence function is versatile in applications such as graph rewiring and adversarial attacks.



Paperid:2015
Authors:Patryk Marszałek, Tomasz Kuśmierczyk, Witold Wydmański, Jacek Tabor, Marek Śmieja
Abstract:
Clustering tabular data remains a significant open challenge in data analysis and machine learning. Unlike for image data, similarity between tabular records often varies across datasets, making the definition of clusters highly dataset-dependent. Furthermore, the absence of supervised signals complicates hyperparameter tuning in deep learning clustering methods, frequently resulting in unstable performance. To address these issues and minimize the need for per-dataset tuning, we adopt an emerging approach in deep learning: zero-shot learning. We propose ZEUS, a self-contained model capable of clustering new datasets without any additional training or fine-tuning. It operates by decomposing complex datasets into meaningful components that can then be clustered effectively. Thanks to pre-training on synthetic datasets generated from a latent-variable prior, it generalizes across various datasets without requiring user intervention. To the best of our knowledge, ZEUS is the first zero-shot method capable of generating embeddings for tabular data in a fully unsupervised manner. Experimental results demonstrate that it performs on par with or better than traditional clustering algorithms and recent deep learning-based methods, while being significantly faster and more user-friendly.



Paperid:2016
Authors:Yangge Li, Chenxi Ji, Xiangru Zhong, Huan Zhang, Sayan Mitra
Abstract:
The rendering process, which generates 2D images from 3D scene representations, has been extensively studied, yet the impact of camera pose and scene uncertainty on rendered outputs and downstream tasks remains underexplored. We proposeAbstract Rendering, a framework that computes provable bounds on all images rendered under continuously varying camera poses and scene configurations. The resultingAbstract Image, defined by constraints over the image matrix, enables rigorous uncertainty propagation through rendering and downstream neural networks, offering a principled way to certify models under realistic 3D semantic perturbations—beyond traditional pixel-level noise models. Our approach propagates camera pose uncertainty through each rendering step using efficient piecewise linear bounds, including custom abstractions for three rendering-specific operations—matrix inversion, sorting-based aggregation, and cumulative product summation—not supported by standard tools. Our implementation, AbstractRender, targets two state-of-the-art photorealistic scene representations—Gaussian Splatting and Neural Radiance Fields (NeRF)—and scales to complex scenes with up to 1M Gaussians. Our computed provable image bounds achieves up to 3% over-approximation error comparing with sampling results (baseline) and enables certified robustness for downstream tasks, including classification (ResNet), object detection (YOLO), and pose estimation (GATENet) under camera pose uncertainty.



Paperid:2017
Authors:Andong Wang, Yuning Qiu, Haonan Huang, Zhong Jin, Guoxu Zhou, Qibin Zhao
Abstract:
Despite the growing success of transform-based tensor models such as the t-product, their underlying geometric principles remain poorly understood. Classical differential geometry, built on real-valued function spaces, is not well suited to capture the algebraic and spectral structure induced by transform-based tensor operations. In this work, we take an initial step toward a geometric framework for tensors equipped with tube-wise multiplication via orthogonal transforms. We introduce the notion of smooth t-manifolds, defined as topological spaces locally modeled on structured tensor modules over a commutative t-scalar ring. This formulation enables transform-consistent definitions of geometric objects, including metrics, gradients, Laplacians, and geodesics, thereby bridging discrete and continuous tensor settings within a unified algebraic-geometric perspective. On this basis, we develop a statistical procedure for testing whether tensor data lie near a low-dimensional t-manifold, and provide nonasymptotic guarantees for manifold fitting under noise. We further establish approximation bounds for tensor neural networks that learn smooth functions over t-manifolds, with generalization rates determined by intrinsic geometric complexity. This framework offers a theoretical foundation for geometry-aware learning in structured tensor spaces and supports the development of models that align with transform-based tensor representations.



Paperid:2018
Authors:Xiaoyan Bai, Ike Peng, Aditya Singh, Chenhao Tan
Abstract:
Consider this prompt "Draw a unicorn with two horns". Should large language models (LLMs) recognize that a unicorn has only one horn by definition and ask users for clarifications, or proceed to generate something anyway? We introduceconcept incongruenceto capture such phenomena where concept boundaries clash with each other, either in user prompts or in model representations, often leading to under-specified or mis-specified behaviors. In this work, we take the first step towards defining and analyzing model behavior under concept incongruence. Focusing on temporal boundaries in the Role-Play setting, we propose three behavioral metrics---abstention rate, conditional accuracy, and answer rate---to quantify model behavior under incongruence due to the role's death. We show that models fail to abstain after death and suffer from an accuracy drop compared to the Non-Role-Play setting. Through probing experiments, we identify two main causes: (i) unreliable encoding of the "death" state across different years, leading to unsatisfactory abstention behavior, and (ii) role playing causes shifts in the model’s temporal representations, resulting in accuracy drops. We leverage these insights to improve consistency in the model's abstention and answer behaviors. Our findings suggest that concept incongruence leads to unexpected model behaviors and point to future directions on improving model behavior under concept incongruence.



Paperid:2019
Authors:Jae-Hong Lee
Abstract:
Foundation models based on deep neural networks hold strong potential for cross-domain adaptability. However, this potential is often impeded in online machine learning (OML) settings, where the breakdown of the independent and identically distributed (i.i.d.) assumption leads to unstable adaptation. While recent advances in test-time adaptation (TTA) have addressed aspects of this challenge under unsupervised learning, most existing methods focus exclusively on unsupervised objectives and overlook the risks posed by non-i.i.d. environments and the resulting dynamics of model parameters. In this work, we present a probabilistic framework that models the adaptation process using stochastic differential equations, enabling a principled analysis of parameter distribution dynamics over time. Within this framework, we find that the log-variance of the parameter transition distribution aligns closely with an inverse-gamma distribution under stable and high-performing adaptation conditions. Motivated by this insight, we propose Structured Inverse-Gamma Model Alignment (SIGMA), a novel algorithm that dynamically regulates parameter evolution to preserve inverse-gamma alignment throughout adaptation. Extensive experiments across diverse models, datasets, and adaptation scenarios show that SIGMA consistently enhances the performance of state-of-the-art TTA methods, highlighting the critical role of parameter dynamics in ensuring robust adaptation.



Paperid:2020
Authors:Zongyuan Li, Yanan Ni, Runnan Qi, Chang Lu, Lumin Jiang, Xu Xiaojie, Xiangbei Liu, Pengfei Li, Yunzheng Guo, Zhe Ma, Huanyu Li, wu hui, Xian Guo, Kuihua Huang, Xuebo Zhang
Abstract:
The tremendous potential has been demonstrated by large language models (LLMs) in intelligent decision-making problems, with unprecedented capabilities shown across diverse applications ranging from gaming AI systems to complex strategic planning frameworks. However, the StarCraft II platform, which has been widely adopted for validating decision-making algorithms in the past decade, has not yet provided substantial support for this emerging domain. To address issues that LLMs cannot interface with the hundreds of actions of the pysc2 backend and the lack of native support for multi-agent (MA) collaboration, we propose the LLM-PySC2 environment. This is the first environment that offers LLMs the complete pysc2 action space with sufficient multi-modal information and game Wiki knowledge. With an asynchronous query architecture, the environment efficiently interacts with LLMs that maintain a constant latency regardless of the scale of the agents' population. In the experiments, we evaluated LLMs' decision-making performance in both the macro-decision and micro-operation scenarios, with traditional StarCraft II Multi-Agent Challenge (SMAC) tasks and a series of new proposed. Results indicate that LLMs possess the potential to achieve victories in complex scenarios but cannot constantly generate correct decisions, especially in the recovered pysc2 action space and MA settings. Without task-relevant instructions, the pre-trained models suffer from issues such as hallucinations and inefficient collaboration. Our findings suggest that StarCraft II still challenges in the era of large models, revealing that there is a lot to do to develop an advanced LLM decision-making system, and the proposed LLM-PySC2 environment will support future development of LLM-based decision-making solutions.



Paperid:2021
Authors:Ziheng Cheng, Yixiao Huang, Hui Xu, Somayeh Sojoudi, Xuandong Zhao, Dawn Song, Song Mei
Abstract:
Text-to-Image (T2I) models have achieved remarkable success in generating visual content from text inputs. Although multiple safety alignment strategies have been proposed to prevent harmful outputs, they often lead to overly cautious behavior---rejecting even benign prompts---a phenomenon known as \textit{over-refusal} that reduces the practical utility of T2I models. Despite over-refusal having been observed in practice, there is no large-scale benchmark that systematically evaluates this phenomenon for T2I models. In this paper, we present an automatic workflow to construct synthetic evaluation data, resulting in OVERT (\textbf{OVE}r-\textbf{R}efusal evaluation on \textbf{T}ext-to-image models), the first large-scale benchmark for assessing over-refusal behaviors in T2I models. OVERT includes 4,600 seemingly harmful but benign prompts across nine safety-related categories, along with 1,785 genuinely harmful prompts (OVERT-unsafe) to evaluate the safety–utility trade-off. Using OVERT, we evaluate several leading T2I models and find that over-refusal is a widespread issue across various categories (Figure 1), underscoring the need for further research to enhance the safety alignment of T2I models without compromising their functionality. As a preliminary attempt to reduce over-refusal, we explore prompt rewriting; however, we find it often compromises faithfulness to the meaning of the original prompts.Finally, we demonstrate the flexibility of our generation framework in accommodating diverse safety requirements by generating customized evaluation data adapting to user-defined policies.



Paperid:2022
Authors:Parjanya Prashant, Kaustubh Ponkshe, Babak Salimi
Abstract:
Abstract:As language models scale, their performance improves dramatically across a wide range of tasks, but so does their tendency to memorize and regurgitate parts of their training data verbatim. This tradeoff poses serious legal, ethical, and safety concerns, especially in real-world deployments. Existing mitigation techniques, such as differential privacy or model unlearning, often require retraining or access to internal weights making them impractical for most users. In this work, we introduce TokenSwap, a lightweight, post-hoc defense designed for realistic settings where the user can only access token-level outputs. Our key insight is that while large models are necessary for high task performance, small models (e.g., DistilGPT-2) are often sufficient to assign fluent, grammatically plausible probabilities to common function words - and crucially, they memorize far less. By selectively swapping token probabilities between models, TokenSwap preserves the capabilities of large models while reducing their propensity for verbatim reproduction. Evaluations on Pythia-6.9B and Llama-3-8B show up to a 10$\times$ drop in exact memorization with negligible task degradation. Our method offers a practical, accessible solution for mitigating memorized generation in deployed LLMs.



Paperid:2023
Authors:Benjamin Cookson, Nisarg Shah, Ziqi Yu
Abstract:
Proportional fairness criteria inspired by democratic ideals of proportional representation have received growing attention in the clustering literature. Prior work has investigated them in two separate paradigms. Chen et al. [ICML 2019] studycentroid clustering, in which each data point's loss is determined by its distance to a representative point (centroid) chosen in its cluster. Caragiannis et al. [NeurIPS 2024] studynon-centroid clustering, in which each data point's loss is determined by its maximum distance to any other data point in its cluster. We generalize both paradigms to introducesemi-centroid clustering, in which each data point's loss is a combination of its centroid and non-centroid losses, and study two proportional fairness criteria---the core and, its relaxation, fully justified representation (FJR). Our main result is a novel algorithm which achieves a constant approximation to the core, in polynomial time, even when the distance metrics used for centroid and non-centroid loss measurements are different. We also derive improved results for more restricted loss functions and the weaker FJR criterion, and establish lower bounds in each case.



Paperid:2024
Authors:Chaoyou Fu, Haojia Lin, Xiong Wang, yifan zhang, Yunhang Shen, Xiaoyu Liu, Haoyu Cao, Zuwei Long, Heting Gao, Ke Li, Long MA, Xiawu Zheng, Rongrong Ji, Xing Sun, Caifeng Shan, Ran He
Abstract:
Recent Multimodal Large Language Models (MLLMs) have typically focused on integrating visual and textual modalities, with less emphasis placed on the role of speech in enhancing interaction. However, speech plays a crucial role in multimodal dialogue systems, and implementing high-performance in both vision and speech tasks remains a significant challenge due to the fundamental modality differences. In this paper, we propose a carefully designed multi-stage training methodology that progressively trains LLM to understand both visual and speech information, ultimately enabling fluent vision and speech interaction. Our approach not only preserves strong vision-language capacity, but also enables efficient speech-to-speech dialogue capabilities without separate ASR and TTS modules, significantly accelerating multimodal end-to-end response speed. By comparing our method against state-of-the-art counterparts across benchmarks for image, video, and speech tasks, we demonstrate that our model is equipped with both strong visual and speech capabilities, making near real-time vision and speech interaction.



Paperid:2025
Authors:Ruolin Meng, Ming-Yu Chung, Dhanajit Brahma, Ricardo Henao, Lawrence Carin
Abstract:
We consider inferring the causal effect of a treatment (intervention) on an outcome of interest in situations where there is potentially an unobserved confounder influencing both the treatment and the outcome. This is achievable by assuming access to two separate sets of control (proxy) measurements associated with treatment and outcomes, which are used to estimate treatment effects through a function termed thecausal bridge(CB). We present a new theoretical perspective, associated assumptions for when estimating treatment effects with the CB is feasible, and a bound on the average error of the treatment effect when the CB assumptions are violated. From this new perspective, we then demonstrate how coupling the CB with an autoencoder architecture allows for the sharing of statistical strength between observed quantities (proxies, treatment, and outcomes), thus improving the quality of the CB estimates. Experiments on synthetic and real-world data demonstrate the effectiveness of the proposed approach in relation to the state-of-the-art methodology for proxy measurements.



Paperid:2026
Authors:Akhiad Bercovich, Mohammed Dabbah, Omri Puny, Ido Galil, Amnon Geifman, Yonatan Geifman, Izhak Golan, Ehud Karpas, Itay Levy, Zach Moshe, Najeeb Nabwani, Tomer Ronen, Itamar Schen, Ido Shahaf, Oren Tropp, Ran Zilberstein, Ran El-Yaniv
Abstract:
Abstract:We introduce \textit{FFN Fusion}, an architectural optimization technique that reduces sequential computation in large language models by identifying and exploiting natural opportunities for parallelization. Our key insight is that sequences of Feed-Forward Network (FFN) layers, particularly those remaining after the removal of specific attention layers, can often be parallelized with minimal accuracy impact. We develop a principled methodology for identifying and fusing such sequences, transforming them into parallel operations that significantly reduce inference latency while preserving model behavior. Applying these techniques to Llama-3.1-405B-Instruct, we create a 253B model (253B-Base), an efficient and soon-to-be publicly available model that achieves a 1.71$\times$ speedup in inference latency and 35$\times$ lower per-token cost while maintaining strong performance across benchmarks. Most intriguingly, we find that even full transformer blocks containing both attention and FFN layers can sometimes be parallelized, suggesting new directions for neural architecture design.



Paperid:2027
Authors:Yunfeng Liao, Yangxin Wu, Xiucheng Li
Abstract:
Recent advances in graph neural network (GNN)-based neural operators have demonstrated significant progress in solving partial differential equations (PDEs) by effectively representing computational meshes. However, most existing approaches overlook the intrinsic physical and topological meaning of higher-order elements in the mesh, which are closely tied to differential forms. In this paper, we propose a higher-order GNN framework that incorporates higher-order interactions based on discrete and finite element exterior calculus. The time-independent boundary value problems (BVPs) in electromagnetism are instantiated to illustrate the proposed framework. It can be easily generalized to other PDEs that admit differential form formulations. Moreover, the novel physics-informed loss terms, integrated form estimators, and theoretical support are derived correspondingly. Experiments show that our proposed method outperforms the existing neural operators by large margins on BVPs in electromagnetism. Our code is available at https://anonymous.4open.science/r/NeurIPS2025-18F5.



Paperid:2028
Authors:Yihong Dong, Ge Li, Xue Jiang, Yongding Tao, Kechi Zhang, Lecheng Wang, Hao Zhu, Huanyu Liu, jiazheng ding, Jia Li, Jinliang Deng, Hong Mei
Abstract:
Periodicity, as one of the most important basic characteristics, lays the foundation for facilitating structured knowledge acquisition and systematic cognitive processes within human learning paradigms. However, the potential flaws of periodicity modeling in Transformer affect the learning efficiency and establishment of underlying principles from data for large language models (LLMs) built upon it. In this paper, we demonstrate that integrating effective periodicity modeling can improve the learning efficiency and performance of LLMs. We introduce FANformer, which adapts Fourier Analysis Network (FAN) into attention mechanism to achieve efficient periodicity modeling, by modifying the feature projection process of attention mechanism. Extensive experimental results on language modeling show that FANformer consistently outperforms Transformer when scaling up model size and training tokens, underscoring its superior learning efficiency. Our pretrained FANformer-1B exhibits marked improvements on downstream tasks compared to open-source LLMs with similar model parameters or training tokens. Moreover, we reveal that FANformer exhibits superior ability to learn and apply rules for reasoning compared to Transformer. The results position FANformer as an effective and promising architecture for advancing LLMs.



Paperid:2029
Authors:Andy Zhou, Kevin Wu, Francesco Pinto, Zhaorun Chen, Yi Zeng, Yu Yang, Shuang Yang, Sanmi Koyejo, James Zou, Bo Li
Abstract:
As large language models (LLMs) become increasingly capable, security and safety evaluation are crucial. While current red teaming approaches have made strides in assessing LLM vulnerabilities, they often rely heavily on human input and lack comprehensive coverage of emerging attack vectors. This paper introduces AutoRedTeamer, a novel framework for fully automated, end-to-end red teaming against LLMs. AutoRedTeamer combines a multi-agent architecture with a memory-guided attack selection mechanism to enable continuous discovery and integration of new attack vectors. The dual-agent framework consists of a red teaming agent that can operate from high-level risk categories alone to generate and execute test cases, and a strategy proposer agent that autonomously discovers and implements new attacks by analyzing recent research. This modular design allows AutoRedTeamer to adapt to emerging threats while maintaining strong performance on existing attack vectors. We demonstrate AutoRedTeamer’s effectiveness across diverse evaluation settings, achieving 20% higher attack success rates on HarmBench against Llama-3.1-70B while reducing computational costs by 46% compared to existing approaches. AutoRedTeamer also matches the diversity of human-curated benchmarks in generating test cases, providing a comprehensive, scalable, and continuously evolving framework for evaluating the security of AI systems.



Paperid:2030
Authors:Youming Tao, Zuyuan Zhang, Dongxiao Yu, Xiuzhen Cheng, Falko Dressler, Di Wang
Abstract:
We investigate the problem of finding second-order stationary points (SOSP) in differentially private (DP) stochastic non-convex optimization. Existing methods suffer from two key limitations: \textbf{(i)} inaccurate convergence error rate due to overlooking gradient variance in the saddle point escape analysis, and \textbf{(ii)} dependence on auxiliary private model selection procedures for identifying DP-SOSP, which can significantly impair utility, particularly in distributed settings. To address these issues, we propose a generic perturbed stochastic gradient descent (PSGD) framework built upon Gaussian noise injection and general gradient oracles. A core innovation of our framework is using model drift distance to determine whether PSGD escapes saddle points, ensuring convergence to approximate local minima without relying on second-order information or additional DP-SOSP identification. By leveraging the adaptive DP-SPIDER estimator as a specific gradient oracle, we develop a new DP algorithm that rectifies the convergence error rates reported in prior work. We further extend this algorithm to distributed learning with arbitrarily heterogeneous data, providing the first formal guarantees for finding DP-SOSP in such settings. Our analysis also highlights the detrimental impacts of private selection procedures in distributed learning under high-dimensional models, underscoring the practical benefits of our design. Numerical experiments on real-world datasets validate the efficacy of our approach.



Paperid:2031
Authors:Xi Zhang, Xiaolin Wu, Jiamang Wang, Weisi Lin
Abstract:
Large Language Models (LLMs) have demonstrated remarkable capabilities but typically require extensive computational resources and memory for inference. Post-training quantization (PTQ) can effectively reduce these demands by storing weights in lower bit-width formats. However, standard uniform quantization often leads to notable performance degradation, particularly in low-bit scenarios. In this work, we introduce a Grouped Lattice Vector Quantization (GLVQ) framework that assigns each group of weights a customized lattice codebook, defined by a learnable generation matrix. To address the non-differentiability of the quantization process, we adopt Babai rounding to approximate nearest-lattice-point search during training, which enables stable optimization of the generation matrices. Once trained, decoding reduces to a simple matrix-vector multiplication, yielding an efficient and practical quantization pipeline. Experiments on multiple benchmarks show that our approach achieves a better trade-off between model size and accuracy compared to existing post-training quantization baselines, highlighting its effectiveness in deploying large models under stringent resource constraints.



Paperid:2032
Authors:Leander Diaz-Bone, Marco Bagatella, Jonas Hübotter, Andreas Krause
Abstract:
Sparse-reward reinforcement learning (RL) can model a wide range of highly complex tasks.Solving sparse-reward tasks is RL's core premise — requiring efficient exploration coupled with long-horizon credit assignment — and overcoming these challenges is key for building self-improving agents with superhuman ability.We argue that solving complex and high-dimensional tasks requires solving simpler tasks that arerelevantto the target task.In contrast, most prior work designs strategies for selecting exploratory tasks with the objective of solvinganytask, making exploration of challenging high-dimensional, long-horizon tasks intractable.We find that the sense of direction, necessary for effective exploration, can be extracted from existing reinforcement learning algorithms, without needing any prior information.Based on this finding, we propose a method fordirectedsparse-reward goal-conditionedvery long-horizonRL(DISCOVER), which selects exploratory goals in the direction of the target task.We connect DISCOVER to principled exploration in bandits, formally bounding the time until the target task becomes achievable in terms of the agent's initial distance to the target, but independent of the volume of the space of all tasks.Empirically, we perform a thorough evaluation in high-dimensional simulated environments. We find that the directed goal selection of DISCOVER solves exploration problems that are beyond the reach of prior state-of-the-art exploration methods in RL.



Paperid:2033
Authors:Giacomo Camposampiero, Pietro Barbiero, Michael Hersche, Roger Wattenhofer, Abbas Rahimi
Abstract:
Compositional generalization—a key open challenge in modern machine learning—requires models to predict unknown combinations of known concepts. However, assessing compositional generalization remains a fundamental challenge due to the lack of standardized evaluation protocols and the limitations of current benchmarks, which often favor efficiency over rigor. At the same time, general-purpose vision architectures lack the necessary inductive biases, and existing approaches to endow them compromise scalability. As a remedy, this paper introduces: 1) a rigorous evaluation framework that unifies and extends previous approaches while reducing computational requirements from combinatorial to constant; 2) an extensive and modern evaluation on the status of compositional generalization in supervised vision backbones, training more than 5000 models; 3) Attribute Invariant Networks, a class of models establishing a new Pareto frontier in compositional generalization, achieving a 23.43% accuracy improvement over baselines while reducing parameter overhead from 600% to 16% compared to fully disentangled counterparts.



Paperid:2034
Authors:Yi He, Xingyu Zhou
Abstract:
Policy optimization (PO) is a cornerstone of modern reinforcement learning (RL), with diverse applications spanning robotics, healthcare, and large language model training. The increasing deployment of PO in sensitive domains, however, raises significant privacy concerns. In this paper, we initiate a theoretical study of differentially private policy optimization, focusing explicitly on its sample complexity. We first formalize an appropriate definition of differential privacy (DP) tailored to PO, addressing the inherent challenges arising from on-policy learning dynamics and the subtlety involved in defining the unit of privacy. We then systematically analyze the sample complexity of widely-used PO algorithms, including policy gradient (PG), natural policy gradient (NPG) and more, under DP constraints and various settings, via a unified framework. Our theoretical results demonstrate that privacy costs can often manifest as lower-order terms in the sample complexity, while also highlighting subtle yet important observations in private PO settings. These offer valuable practical insights for privacy-preserving PO algorithms.



Paperid:2035
Authors:Roey Magen, Gal Vardi
Abstract:
Abstract:Transformers have demonstrated impressive in-context learning (ICL) capabilities, raising the question of whether they can serve as metalearners that adapt to new tasks using only a small number of in-context examples, without any further training. While recent theoretical work has studied transformers' ability to perform ICL, most of these analyses do not address the formal metalearning setting, where the objective is to solve a collection of related tasks more efficiently than would be possible by solving each task individually. In this paper, we provide the first theoretical analysis showing that a simplified transformer architecture trained via gradient descent can act as a near-optimal metalearner in a linear classification setting. We consider a natural family of tasks where each task corresponds to a class-conditional Gaussian mixture model, with the mean vectors lying in a shared $k$-dimensional subspace of $\mathbb{R}^d$. After training on a sufficient number of such tasks, we show that the transformer can generalize to a new task using only $\widetilde{O}(k / \widetilde{R}^4)$ in-context examples, where $\widetilde{R}$ denotes the signal strength at test time. This performance (almost) matches that of an optimal learner that knows exactly the shared subspace and significantly outperforms any learner that only has access to the in-context data, which requires $\Omega(d / \widetilde{R}^4)$ examples to generalize.



Paperid:2036
Authors:Sung-Hoon Yoon, Minghan Li, Gaspard Beaudouin, Congcong Wen, Muhammad Azhar, Mengyu Wang
Abstract:
Abstract:Rectified flow models have become a $\textit{de facto}$ standard in image generation due to their stable sampling trajectories and high-fidelity outputs. Despite their strong generative capabilities, they face critical limitations in image editing tasks: inaccurate inversion processes for mapping real images back into the latent space, and gradient entanglement issues during editing often result in outputs that do not faithfully reflect the target prompt. Recent efforts have attempted to directly map source and target distributions via ODE-based approaches without inversion; however, these methods still yield suboptimal editing quality. In this work, we propose a flow decomposition-and-aggregation framework built upon an inversion-free formulation to address these limitations. Specifically, we semantically decompose the target prompt into multiple sub-prompts, compute an independent flow for each, and aggregate them to form a unified editing trajectory. While we empirically observe that decomposing the original flow enhances diversity in the target space, generating semantically aligned outputs still requires consistent guidance toward the full target prompt. To this end, we design a cosine similarity-based soft aggregation mechanism inspired by gradient conflict resolution in multi-task learning. This approach adaptively weights the sub-target velocity fields, suppressing semantic redundancy while emphasizing distinct directions, thereby preserving both diversity and consistency in the final edited output. Experimental results demonstrate that our method outperforms existing zero-shot editing approaches in terms of semantic fidelity and attribute disentanglement.



Paperid:2037
Authors:Tianxu Li, Kun Zhu
Abstract:
In cooperative Multi-Agent Reinforcement Learning (MARL), agents that share policy network parameters often learn similar behaviors, which hinders effective exploration and can lead to suboptimal cooperative policies. Recent advances have attempted to promote multi-agent diversity by leveraging the Wasserstein distance to increase policy differences. However, these methods cannot effectively encourage diverse policies due to ineffective Wasserstein distance caused by the policy similarity. To address this limitation, we propose Wasserstein Contrastive Diversity (WCD) exploration, a novel approach that promotes multi-agent diversity by maximizing the Wasserstein distance between the trajectory distributions of different agents in a latent representation space. To make the Wasserstein distance meaningful, we propose a novel next-step prediction method based on Contrastive Predictive Coding (CPC) to learn distinguishable trajectory representations. Additionally, we introduce an optimized kernel-based method to compute the Wasserstein distance more efficiently. Since the Wasserstein distance is inherently defined for two distributions, we extend it to support multiple agents, enabling diverse policy learning. Empirical evaluations across a variety of challenging multi-agent tasks demonstrate that WCD outperforms existing state-of-the-art methods, delivering superior performance and enhanced exploration.



Paperid:2038
Authors:Jiachen Yao, Abbas Mammadov, Julius Berner, Gavin Kerrigan, Jong Chul Ye, Kamyar Azizzadenesheli, Animashree Anandkumar
Abstract:
We propose a general framework for conditional sampling in PDE-based inverse problems, targeting the recovery of whole solutions from extremely sparse or noisy measurements.This is accomplished by a function-space diffusion model and plug-and-play guidance for conditioning.Our method first trains an unconditional discretization-agnostic denoising model using neural operator architectures. At inference, we refine the samples to satisfy sparse observation data via a gradient-based guidance mechanism.Through rigorous mathematical analysis, we extend Tweedie's formula to infinite-dimensional Hilbert spaces, providing the theoretical foundation for our posterior sampling approach.Our method (FunDPS) accurately captures posterior distributions in function spaces under minimal supervision and severe data scarcity. Across five PDE tasks with only 3\% observation, our method achieves an average 32\% accuracy improvement over state-of-the-art fixed-resolution diffusion baselines while reducing sampling steps by 4x. Furthermore, multi-resolution fine-tuning ensures strong cross-resolution generalizability. To the best of our knowledge, this is the first diffusion-based framework to operate independently of discretization, offering a practical and flexible solution for forward and inverse problems in the context of PDEs.



Paperid:2039
Authors:Aruna Gauba, Irene Pi, Yunze Man, Ziqi Pang, Vikram Adve, Yu-Xiong Wang
Abstract:
We presentAgMMU, a challenging real‑world benchmark for evaluating and advancing vision-language models (VLMs) in the knowledge‑intensive domain of agriculture. Unlike prior datasets that rely on crowdsourced prompts, AgMMU is distilled from 116,231 authentic dialogues between everyday growers and USDA-authorized Cooperative Extension experts. Through a three‑stage pipeline: automated knowledge extraction, QA generation, and human verification, we construct (i) AgMMU, an evaluation set of 746 multiple‑choice questions (MCQs) and 746 open‑ended questions (OEQs), and (ii) AgBase, a development corpus of 57,079 multimodal facts covering five high-stakes agricultural topics: insect identification, species identification, disease categorization, symptom description, and management instruction. AgMMU has three key advantages:-Authentic \& Expert‑Verified: All facts, images, and answers originate from real farmer and gardener inquiries answered by credentialed specialists, ensuring high‑fidelity agricultural knowledge.-Complete Development Suite: AgMMU uniquely couples a dual‑format evaluation benchmark (MCQ and OEQ) with AgBase, a large‑scale training set, enabling both rigorous assessment and targeted improvement of VLMs.-Knowledge‑intensive Challenge: Our tasks demand the synergy of nuanced visual perception and domain expertise, exposing fundamental limitations of current general‑purpose models and charting a path toward robust, application‑ready agricultural AI.Benchmarking 12 leading VLMs reveals pronounced gaps in fine‑grained perception and factual grounding. Open‑sourced models trail after proprietary ones by a wide margin. Simple fine‑tuning on AgBase boosts open-sourced model performance on challenging OEQs for up to 11.6\% on average, narrowing this gap and also motivating future research to propose better strategies in knowledge extraction and distillation from AgBase. We hope AgMMU stimulates research on domain‑specific knowledge integration and trustworthy decision support in agriculture AI development.



Paperid:2040
Authors:Fudong Lin, Xu Yuan
Abstract:
The imbalance (or long-tail) is the nature of many real-world data distributions, which often induces the undesirable bias of deep classification models toward frequent classes, resulting in poor performance for tail classes. In this paper, we propose a novel two-stage learning approach to mitigate such a majority-biased tendency while preserving valuable information within datasets. Specifically, the first stage proposes a new representation learning technique from the information theory perspective. This approach is theoretically equivalent to minimizing intra-class distance, yielding an effective and well-separated feature space. The second stage develops a novel undersampling strategy that selects mathematically informative instances, able to rectify majority-biased decision boundaries without compromising a model’s overall performance. As a result, our approach achieves the state-of-the-art performance across various long-tailed benchmark datasets, validated via extensive experiments. Our code is available at https://anonymous.4open.science/r/BNS_IPDPP.



Paperid:2041
Authors:Vivek Ramanujan, Kushal Tirumala, Armen Aghajanyan, Luke Zettlemoyer, Ali Farhadi
Abstract:
Abstract:Current image generation methods are based on a two-stage training approach. In stage 1, an auto-encoder is trained to compress an image into a latent space; in stage 2, a generative model is trained to learn a distribution over that latent space. This reveals a fundamental trade-off, do we compress more aggressively to make the latent distribution easier for the stage 2 model to learn even if it makes reconstruction worse? We study this problem in the context of discrete, auto-regressive image generation. Through the lens of scaling laws, we show that smaller stage 2 models can benefit from more compressed stage 1 latents even if reconstruction performance worsens, demonstrating that generation modeling capacity plays a role in this trade-off. Diving deeper, we rigorously study the connection between compute scaling and the stage 1 rate-distortion trade-off. Next, we introduce Causally Regularized Tokenization (CRT), which uses knowledge of the stage 2 generation modeling procedure to embed useful inductive biases in stage 1 latents. This regularization improves stage 2 generation performance better by making the tokens easier to model without affecting the stage 1 compression rate and marginally affecting distortion: we are able to improve compute efficiency 2-3$\times$ over baseline. Finally, we use CRT with further optimizations to the visual tokenizer setup to result in a generative pipeline that matches LlamaGen-3B generation performance (2.18 FID) with half the tokens per image (256 vs. 576) and a fourth the total model parameters (775M vs. 3.1B) while using the same architecture and inference procedure.



Paperid:2042
Authors:Ling Tang, YueFeng Chen, Hui Xue', Quanshi Zhang
Abstract:
This paper proves a new fingerprinting method to embed the ownership information into a deep neural network (DNN) with theoretically guaranteed robustness to fine-tuning. Specifically, we prove that when the input feature of a convolutional layer only contains low-frequency components, specific frequency components of the convolutional filter will not be changed by gradient descent during the fine-tuning process, where we propose a revised Fourier transform to extract frequency components from the convolutional filter. Additionally, we also prove that these frequency components are equivariant to weight scaling and weight permutations. In this way, we design a fingerprint module to embed the fingerprint information into specific frequency components of convolutional filters. Preliminary experiments demonstrate the effectiveness of our method.



Paperid:2043
Authors:Yifeng He, Luning Yang, Christopher Gonzalo, Hao Chen
Abstract:
Abstract:Large Language Models (LLMs) are increasingly integrated into the software engineering ecosystem.Their test-time compute reasoning capabilities promise significant potential in understanding program logic and semantics beyond mere token recognition. However, current benchmarks evaluating reasoning LLMs for code lack a formal, program-centric deductive framework for the soundness of evaluation, incompetent in assessing of whether models genuinely reason about program semantics or merely associate superficial connections between natural language and code tokens. % There is limited discussion on addressing the semantic gap between natural languages and programming languages when evaluating reasoning LLMs for code.To bridge this gap, we introduce TF-Bench, a benchmark designed to evaluate LLM reasoning based on type inference in System F, a task we refer to as \emph{program semantics reasoning}. By employing verified transformations to remove semantically irrelevant natural language,we construct TF-Bench_pure, a purely semantics-driven variant of TF-Bench. Our analysis reveals substantial limitations in state-of-the-art LLMs, with the best-performing LLM (Claude-3.7-sonnet) achieving only $55.85\%$ accuracy on TF-Bench_pure. Additionally, we propose two novel metrics to assess the robustness and effectiveness of extended reasoning,underscoring the critical limitation in current LLM capabilities and highlighting essential directions for future research.



Paperid:2044
Authors:Jiahang Zhang, Lilang Lin, Shuai Yang, Jiaying Liu
Abstract:
3D human motion-text retrieval is essential for accurate motion understanding, targeted at cross-modal alignment learning. Existing methods typically align the global motion-text concepts directly, suffering from sub-optimal generalization due to the uncertainty of correspondence learning between multiple motion concepts coupled in a single motion/text sequence. Therefore, we study the explicit fine-grained concept decomposition for alignment learning and present a novel framework, Structural Generative Augmentation for 3D Human Motion Retrieval (SGAR), to enable generation-augmented retrieval. Specifically, relying on the strong priors of existing large language model (LLM) assets, we effectively decompose human motions structurally into subtler semantic units, \ie, body parts, for fine-grained motion modeling. Based on this, we develop part-mixture learning to better decouple the local motion concept learning, boosting part-level alignment. Moreover, a directional relation alignment strategy exploiting the correspondence between full-body and part motions is incorporated to regularize feature manifold for better consistency. Extensive experiments on three benchmarks, including motion-text retrieval as well as recognition and generation applications, demonstrate the superior performance and promising transferability of our method.



Paperid:2045
Authors:Tony Yousefnezhad
Abstract:
Task-based functional magnetic resonance imaging (fMRI) provides invaluable insights into human cognition but faces critical hurdles—low signal-to-noise ratio, high dimensionality, limited sample sizes, and costly data acquisition—that are amplified when integrating datasets across subjects or sites. This paper introduces orthogonal contrastive learning (OCL), a unified multi-representation framework for multi-subject fMRI analysis that aligns neural responses without requiring temporal preprocessing or uniform time-series lengths across subjects or sites. OCL employs two identical encoders: an online network trained with a contrastive loss that pulls together same-stimulus responses and pushes apart different-stimulus responses, and a target network whose weights track the online network via exponential moving average to stabilize learning. Each OCL network layer combines QR decomposition for orthogonal feature extraction, locality-sensitive hashing (LSH) to produce compact subject-specific signatures, positional encoding to embed temporal structure alongside spatial features, and a transformer encoder to generate discriminative, stimulus-aligned embeddings. We further enhance OCL with an unsupervised pretraining stage on fMRI-like synthetic data and demonstrate a transfer-learning workflow for multi-site studies. Across extensive experiments on multi-subject and multi-site fMRI benchmarks, OCL consistently outperforms state-of-the-art alignment and analysis methods in both representation quality and downstream classification accuracy.



Paperid:2046
Authors:Chengzhuo Tong, Ziyu Guo, Renrui Zhang, Wenyu Shan, Xinyu Wei, Zhenghao Xing, Hongsheng Li, Pheng-Ann Heng
Abstract:
Recent advancements underscore the significant role of Reinforcement Learning (RL) in enhancing the Chain-of-Thought (CoT) reasoning capabilities of large language models (LLMs). Two prominent RL algorithms, Direct Preference Optimization (DPO) and Group Relative Policy Optimization (GRPO), are central to these developments, showcasing different pros and cons. Autoregressive image generation, also interpretable as a sequential CoT reasoning process, presents unique challenges distinct from LLM-based CoT reasoning. These encompass ensuring text-image consistency, improving image aesthetic quality, and designing sophisticated reward models, rather than relying on simpler rule-based rewards. While recent efforts have extended RL to this domain, these explorations typically lack an in-depth analysis of the domain-specific challenges and the characteristics of different RL strategies. To bridge this gap, we provide the first comprehensive investigation of the GRPO and DPO algorithms in autoregressive image generation, evaluating theirin-domainperformance andout-of-domaingeneralization, while scrutinizing the impact ofdifferent reward modelson their respective capabilities. Our findings reveal that GRPO and DPO exhibit distinct advantages, and crucially, that reward models possessing stronger intrinsic generalization capabilities potentially enhance the generalization potential of the applied RL algorithms. Furthermore, we systematically explorethree prevalent scaling strategiesto enhance both their in-domain and out-of-domain proficiency, deriving unique insights into efficiently scaling performance for each paradigm. We hope our study paves a new path for inspiring future work on developing more effective RL algorithms to achieve robust CoT reasoning in the realm of autoregressive image generation.



Paperid:2047
Authors:Ermis Soumalias, Michael Curry, Sven Seuken
Abstract:
The next frontier of online advertising is revenue generation from LLM-generated content. We consider a setting where advertisers aim to influence the responses of an LLM, while platforms seek to maximize advertiser value and ensure user satisfaction. The challenge is that advertisers' preferences generally conflict with those of the user, and advertisers may misreport their preferences. To address this, we introduce MOSAIC, an auction mechanism that ensures that truthful reporting is a dominant strategy for advertisers and that aligns the utility of each advertiser with their contribution to social welfare. Importantly, the mechanism operates without LLM fine-tuning or access to model weights and provably converges to the output of the optimally fine-tuned LLM as computational resources increase. Additionally, it can incorporate contextual information about advertisers, which significantly improves social welfare. Via experiments with publicly available LLMs, we show that MOSAIC leads to high advertiser value and platform revenue with low computational costs. While our motivating application is online advertising, our mechanism can be applied in any setting with monetary transfers, making it a general-purpose solution for truthfully aggregating the preferences of self-interested agents over LLM-generated replies.



Paperid:2048
Authors:Bingquan Dai, Luo Li, Jie Wang, Qihong Tang, Xinyu Lian, Hao Xu, Minghan Qin, Xudong XU, Bo Dai, Haoqian Wang, Zhaoyang Lyu, Jiangmiao Pang
Abstract:
Reconstructing 3D objects into editable programs is pivotal for applications like reverse engineering and shape editing. However, existing methods often rely on limited domain-specific languages (DSLs) and small-scale datasets, restricting their ability to model complex geometries and structures. To address these challenges, we introduce MeshLLM, a novel framework that reconstructs complex 3D objects from point clouds into editable Blender Python scripts. We develop a comprehensive set of expressive Blender Python APIs capable of synthesizing intricate geometries. Leveraging these APIs, we construct a large-scale paired object-code dataset, where the code for each object is decomposed into distinct semantic parts. Subsequently, we train a multimodal large language model (LLM) that translates 3D point cloud into executable Blender Python scripts. Our approach not only achieves superior performance in shape-to-code reconstruction tasks but also facilitates intuitive geometric and topological editing through convenient code modifications. Furthermore, our code-based representation enhances the reasoning capabilities of LLMs in 3D shape understanding tasks. Together, these contributions establish MeshLLM as a powerful and flexible solution for programmatic 3D shape reconstruction and understanding.



Paperid:2049
Authors:Alex Costanzino, Woody Bayliss, Juil Sock, Marc Gorriz Blanch, Danijela Horak, Ivan Laptev, Philip Torr, Fabio Pizzati
Abstract:
Changing facial expressions, gestures, or background details may dramatically alter the meaning conveyed by an image.Notably, recent advances in diffusion models greatly improve the quality of image manipulation while also opening the door to misuse.Identifying changes made to authentic images, thus, becomes an important task, constantly challenged by new diffusion-based editing tools.To this end, we propose a novel approach for ReliAble iDentification of inpainted AReas (RADAR).RADAR builds on existing foundation models and combines features from different image modalities.It also incorporates an auxiliary contrastive loss that helps to isolate manipulated image patches.We demonstrate these techniques to significantly improve both the accuracy of our method and its generalisation to a large number of diffusion models.To support realistic evaluation, we further introduce MIBench, a new comprehensive benchmark, with images tampered by 28 diffusion models. Our experiments show that RADAR achieves excellent results, outperforming the state-of-the-art in detecting and localising image edits made by both seen and unseen diffusion models. Our code, data and models will be publicly available.



Paperid:2050
Authors:Ningfeng Yang, Tor Aamodt
Abstract:
Abstract:We study the problem of training neural networks with quantized parameters. Learning low-precision quantized parameters by enabling computation of gradients via the Straight-Through Estimator (STE) can be challenging. While the STE enables back-propagation, which is a first-order method, recent works have explored the use of zeroth-order (ZO) gradient descent for fine-tuning. We note that the STE provides high-quality biased gradients and ZO gradients are unbiased but can be expensive. We thus propose First-Order-Guided Zeroth-Order Gradient Descent (FOGZO) that reduces STE bias while reducing computations relative to ZO methods. Empirically we show FOGZO improves the tradeoff between quality and training time in quantization aware training (QAT). Specifically, versus STE at the same number of iterations we show a 1-8\% accuracy improvement for DeiT Tiny/Small, 1-2\% accuracy improvement on Resnet 18/50, and 1-22 perplexity point improvement for LLaMA 9m/20m. We also show that for the same loss FOGZO yields a 796$\times$ reduction in computation versus n-SPSA for a 2-layer MLP on MNIST.



Paperid:2051
Authors:Minghao Yin, Yukang Cao, Kai Han
Abstract:
We present WUKONG, a novel training-free framework for high-fidelity textured 3D morphing that takes a pair of source and target prompts (text or images) as input. Unlike conventional methods -- which rely on manual correspondence matching and deformation trajectory estimation (limiting generalization and requiring costly preprocessing) -- WUKONG leverages the generative prior of flow-based transformers to produce high-fidelity 3D transitions with rich texture details. To ensure smooth shape transitions, we exploit the inherent continuity of flow-based generative processes and formulate morphing as an optimal transport barycenter problem. We further introduce a sequential initialization strategy to prevent abrupt geometric distortions and preserve identity coherence. For faithful texture preservation, we propose a similarity-guided semantic consistency mechanism that selectively retains high-frequency details and enables precise control over blending dynamics. This avoids common artifacts like oversmoothing while maintaining semantic fidelity. Through extensive quantitative and qualitative evaluations, we demonstrate that WUKONG significantly outperforms state-of-the-art methods, achieving superior results across diverse geometry and texture variations.



Paperid:2052
Authors:Lingzhuang Meng, Mingwen Shao, Yuanjian Qiao, Xiang Lv
Abstract:
3D editing with Gaussian splatting is exciting in creating realistic content, but it also poses abuse risks for generating malicious 3D content. Existing 2D defense approaches mainly focus on adding perturbations to single image to resist malicious image editing. However, there remain two limitations when applied directly to 3D scenes: (1) These methods fail to reflect 3D spatial correlations, thus protecting ineffectively under multiple viewpoints. (2) Such pixel-level perturbation is easily eliminated during the iterations of 3D editing, leading to failure of protection. To address the above issues, we propose a novel Defense framework against malicious 3D Editing for Gaussian splatting (DEGauss) for robustly disrupting the trajectory of 3D editing in multi-views. Specifically, to enable the effectiveness of perturbation across various views, we devise a view-focal gradient fusion mechanism that dynamically emphasizes the contributions of the most challenging views to adaptively optimize 3D perturbations. Furthermore, we design a dual discrepancy optimization strategy that both maximize the semantic deviation and the edit direction deviation of the guidance conditions to stably disrupt the editing trajectory. Benefiting from the collaborative designs, our method achieves effective resistance to 3D editing from various views while preserving photorealistic rendering quality. Extensive experiments demonstrate that our DEGauss not only performs excellent defense in different scenes, but also exhibits strong generalization across various state-of-the-art 3D editing pipelines.



Paperid:2053
Authors:Zijian Feng, Hanzhang Zhou, Zixiao Zhu, Tianjiao Li, Chua Deryl, Lee Onn Mak, Gee Ng, Kezhi Mao
Abstract:
Pruning is a widely used technique to reduce the size and inference cost of large language models (LLMs), but it often causes performance degradation. To mitigate this, existing restoration methods typically employ parameter-efficient fine-tuning (PEFT), such as LoRA, to recover the pruned model's performance. However, most PEFT methods are designed for dense models and overlook the distinct properties of pruned models, often resulting in suboptimal recovery. In this work, we propose a targeted restoration strategy for pruned models that restores performance while preserving their low cost and high efficiency. We observe that pruning-induced information loss is reflected in attention activations, and selectively reintroducing components of this information can significantly recover model performance. Based on this insight, we introduce RestoreLCC (Restoring Pruned LLMs via Lost Component Compensation), a plug-and-play method that contrastively probes critical attention heads via activation editing, extracts lost components from activation differences, and finally injects them back into the corresponding pruned heads for compensation and recovery. RestoreLCC is compatible with structured, semi-structured, and unstructured pruning schemes. Extensive experiments demonstrate that RestoreLCC consistently outperforms state-of-the-art baselines in both general and task-specific performance recovery, without compromising the sparsity or inference efficiency of pruned models.



Paperid:2054
Authors:Zhuo Chen, Oriol Comas, Zhuotao Jin, Di Luo, Marin Soljacic
Abstract:
This paper rigorously establishes abipartitemutual information scaling law in natural language that fundamentally governs its long-range dependencies, providing the key to understandinglong-context language modeling. We demonstrate that the bipartite mutual information captures multi-token interactions in a manner that is distinct from, and scales independently of, conventional two-point mutual information, offering a more comprehensive characterization of the dependencies crucial for modeling long sequences. Leveraging this scaling law, we formulate theLong-contextLanguageModeling (L²M) condition, which establishes a minimum requirement on the scaling of a model's latent state size for storing past information necessary for effective long-context modeling. Our theoretical framework and its predictions are validated through experiments on both transformer and state-space models. This work provides a crucial theoretical foundation for advancing the design of large language models toward effective longer context capabilities.



Paperid:2055
Authors:Yutong Feng, Xu Liu, Yutong Xia, Yuxuan Liang
Abstract:
Accurately modeling complex dynamic spatio-temporal systems requires capturing flow-mediated interdependencies and context-sensitive interaction dynamics. Existing methods, predominantly graph-based or attention-driven, rely on similarity-driven connectivity assumptions, neglecting asymmetric flow exchanges that govern system evolution. We propose Spatio-Temporal Flow, a physics-inspired paradigm that explicitly models dynamic node couplings through quantifiable flow transfers governed by conservation principles. Building on this, we design FlowNet, a novel architecture leveraging flow tokens as information carriers to simulate source-to-destination transfers via Flow Allocation Modules, ensuring state redistribution aligns with physical laws. FlowNet dynamically adjusts the interaction radius through an Adaptive Spatial Masking module, suppressing irrelevant noise while enabling context-aware propagation. A cascaded architecture enhances scalability and nonlinear representation capacity. Experiments demonstrate that FlowNet significantly outperforms existing SOTA approaches on seven metrics in the modeling of three real-world systems, validating its efficiency and physical interpretability. We establish a principled methodology for modeling complex systems through spatio-temporal flow interactions.



Paperid:2056
Authors:Changxin Ke, Rui Zhang, Shuo Wang, Li Ding, Guangli Li, Yuanbo Wen, Shuoming Zhang, Ruiyuan Xu, Jin Qin, Jiaming Guo, Chenxi Wang, Ling Li, Qi Guo, Yunji Chen
Abstract:
The rise of GPU-based high-performance computing (HPC) has driven the widespread adoption of parallel programming models such as CUDA. Yet, the inherent complexity of parallel programming creates a demand for the automated sequential-to-parallel approaches. However, data scarcity poses a significant challenge for machine learning-based sequential-to-parallel code translation. Although recent back-translation methods show promise, they still fail to ensure functional equivalence in the translated code. In this paper, we propose a novel Mutual-Supervised Learning (MSL) framework for sequential-to-parallel code translation to address the functional equivalence issue. MSL consists of two models, a Translator and a Tester. Through an iterative loop consisting of Co-verify and Co-evolve steps, the Translator and the Tester mutually generate data for each other and improve collectively. The Tester generates unit tests to verify and filter functionally equivalent translated code, thereby evolving the Translator, while the Translator generates translated code as augmented input to evolve the Tester. Experimental results demonstrate that MuSL significantly enhances the performance of the base model: when applied to Qwen2.5-Coder, it not only improves Pass@1 by up to 28.91\% and boosts Tester performance by 68.90\%, but also outperforms the previous state-of-the-art method CodeRosetta by 1.56 and 6.92 in BLEU and CodeBLEU scores, while achieving performance comparable to DeepSeek-R1 and GPT-4.1.



Paperid:2057
Authors:Hui Li, Huafeng Liu, Shuyang Lin, Jingyue Shi, Yiran Fu, Liping Jing
Abstract:
Neural Processes (NPs) combine the adaptability of neural networks with the efficiency of meta-learning, offering a powerful framework for modeling stochastic processes. However, existing methods focus on empirical performance while lacking a rigorous theoretical understanding of generalization. To address this, we propose an information-theoretic framework to analyze the generalization bounds of NPs, introducing dynamical stability regularization to minimize sharpness and improve optimization dynamics. Additionally, we show how noise-injected parameter updates complement this regularization. The proposed approach, applicable to a wide range of NP models, is validated through experiments on classic benchmarks, including 1D regression, image completion, Bayesian optimization, and contextual bandits. The results demonstrate tighter generalization bounds and superior predictive performance, establishing a principled foundation for advancing generalizable NP models.



Paperid:2058
Authors:Xi Liu, Chaoyi Zhou, Nanxuan Zhao, Siyu Huang
Abstract:
Differentiable vector graphics (VGs) are widely used in image vectorization and vector synthesis, while existing representations are costly to optimize and struggle to achieve high-quality rendering results for high-resolution images. This work introduces a new differentiable VG representation, dubbed Bézier Splatting, that enables fast yet high-fidelity VG rasterization. Bézier Splatting samples 2D Gaussians along Bézier curves, which naturally provide positional gradients at object boundaries. Thanks to the efficient splatting-based differentiable rasterizer, Bézier Splatting achieves 30× and 150× faster per forward and backward rasterization step for open curves compared to DiffVG. Additionally, we introduce an adaptive pruning and densification strategy that dynamically adjusts the spatial distribution of curves to escape local minima, further improving VG quality. Furthermore, our new VG representation supports conversion to standard XML-based SVG format, enhancing interoperability with existing VG tools and pipelines. Experimental results show that Bézier Splatting significantly outperforms existing methods with better visual fidelity and significant optimization speedup.



Paperid:2059
Authors:Minghe Gao, Zhongqi Yue, Wenjie Yan, Yihao Hu, Wei Ji, Siliang Tang, Jun Xiao, Tat-Seng Chua, Yueting Zhuang, Juncheng Li
Abstract:
The rapid development of Multimodal Large Language Models (MLLMs) poses increasing demands on the diversity and complexity of multimodal datasets. Yet manual annotation pipelines can no longer keep pace. Existing augmentation methods often follow fixed rules and lack verifiable control over sample diversity and reasoning complexity. To address this, we introduce Scalable COunterfactual Program Evolution (SCOPE), a framework that uses symbolic Visual Programming to guide program evolution via counterfactual reasoning. SCOPE performs the three steps of counterfactual inference: (1) Abduction, by generating verifiable programs to model reasoning associations; (2) Action, by intervening on program structure along three axes—reasoning path, visual context, and cross-instance composition; and (3) Prediction, by categorizing evolved instances by difficulty, structure, and input multiplicity. Based on this process, we build SCOPE-Train and SCOPE-Test, evolving benchmarks with expert validation. To support training, we propose MAP, a curriculum learning strategy that aligns model capacity with sample difficulty. Experiments show that SCOPE improves reasoning performance, exposes model blind spots, and enhances visual dialog capabilities.



Paperid:2060
Authors:Xiaohu Huang, Jingjing Wu, Qunyi Xie, Kai Han
Abstract:
Recent advances in scene understanding have leveraged multimodal large language models (MLLMs) for 3D reasoning by capitalizing on their strong 2D pretraining. However, the lack of explicit 3D data during MLLM pretraining limits 3D representation capability. In this paper, we investigate the 3D-awareness of MLLMs by evaluating multi-view correspondence and reveal a strong positive correlation between the quality of 3D-aware representation and downstream task performance. Motivated by this, we propose 3DRS, a framework that enhances MLLM 3D representation learning by introducing supervision from pretrained 3D foundation models. Our approach aligns MLLM visual features with rich 3D knowledge distilled from 3D models, effectively improving scene understanding. Extensive experiments across multiple benchmarks and MLLMs—including visual grounding, captioning, and question answering—demonstrate consistent performance gains. Code will be released to facilitate future research.



Paperid:2061
Authors:Ali Shehper, Anibal Medina-Mardones, Lucas Fagan, Bartłomiej Lewandowski, Angus Gruen, Yang Qiu, Piotr Kucharski, Zhenghan Wang, Sergei Gukov
Abstract:
Using a long-standing conjecture from combinatorial group theory, we explore, from multiple perspectives, the challenges of finding rare instances carrying disproportionately high rewards. Based on lessons learned in the context defined by the Andrews--Curtis conjecture, we analyze how reinforcement learning agents handle problems of varying hardness. We also address many mathematical questions as a part of our study. Notably, we demonstrate the length reducibility of all but two presentations in the Akbulut--Kirby series (1981), and resolve various potential counterexamples in the Miller--Schupp series (1991), including three infinite subfamilies.



Paperid:2062
Authors:Jing Zuo, Luoping Cui, Chuang Zhu, Yonggang Qi
Abstract:
The diffusion inversion problem seeks to recover the latent generative trajectory of a diffusion model given a real image. Faithful inversion is critical for ensuring consistency in diffusion-based image editing. Prior works formulate this task as a fixed-point problem and solve it using numerical methods. However, achieving both accuracy and efficiency remains challenging, especially for few-step models and novel samples. In this paper, we proposePreciseInv, a general-purpose learning framework that enables fast and faithful inversion in as few as two inference steps. Unlike root-finding approaches, we reformulate inversion as an optimization problem and introduce a dynamic programming-inspired strategy to recursively estimate a parameterized sequence of noise embeddings. This design leverages the smoothness of the diffusion latent space for accurate gradient-based optimization and ensures memory efficiency via recursive subproblem construction. Extensive experiments on COCO 2017, DarkFace, and a stylized cartoon dataset show thatPreciseInvachieves state-of-the-art performance in reconstruction quality and inference speed. Improvements are especially notable for few-step models and under distribution shifts. Moreover, precise inversion yields substantial gains in editing consistency for text-driven image manipulation tasks.



Paperid:2063
Authors:Chenxi Xie, Minghan Li, Shuai Li, Yuhui Wu, Qiaosi Yi, Lei Zhang
Abstract:
Leveraging the powerful generation capability of large-scale pretrained text-to-image models, training-free methods have demonstrated impressive image editing results. Conventional diffusion-based methods, as well as recent rectified flow (RF)-based methods, typically reverse synthesis trajectories by gradually adding noise to clean images, during which the noisy latent at the current timestep is used to approximate that at the next timesteps, introducing accumulated drift and degrading reconstruction accuracy. Considering the fact that in RF the noisy latent is estimated through direct interpolation between Gaussian noises and clean images at each timestep, we propose Direct Noise Alignment (DNA), which directly refines the desired Gaussian noise in the noise domain, significantly reducing the error accumulation in previous methods. Specifically, DNA estimates the velocity field of the interpolated noised latent at each timestep and adjusts the Gaussian noise by computing the difference between the predicted and expected velocity field. We validate the effectiveness of DNA and reveal its relationship with existing RF-based inversion methods. Additionally, we introduce a Mobile Velocity Guidance (MVG) to control the target prompt-guided generation process, balancing image background preservation and target object editability. DNA and MVG collectively constitute our proposed method, namely DNAEdit. Finally, we introduce DNA-Bench, a long-prompt benchmark, to evaluate the performance of advanced image editing models. Experimental results demonstrate that our DNAEdit achieves superior performance to state-of-the-art text-guided editing methods. Our code, model, and benchmark will be made publicly available.



Paperid:2064
Authors:Zeju Qiu, Simon Buchholz, Tim Xiao, Maximilian Dax, Bernhard Schölkopf, Weiyang Liu
Abstract:
While Large language models (LLMs) are driving the rapid advancement of artificial intelligence, effectively and reliably training these large models remains one of the field's most significant challenges. To address this challenge, we propose POET, a novel reparameterized training algorithm that uses orthogonal equivalence transformation to optimize neurons. Specifically, POET reparameterizes each neuron with two learnable orthogonal matrices and a fixed random weight matrix. Because of its provable preservation of spectral properties of weight matrices, POET can stably optimize the objective function with improved generalization. We further develop efficient approximations that make POET flexible and scalable for training large-scale neural networks. Extensive experiments validate the effectiveness and scalability of POET in training LLMs.



Paperid:2065
Authors:Xia Jiang, Yaoxin Wu, Minshuo Li, Zhiguang Cao, Yingqian Zhang
Abstract:
Combinatorial optimization (CO) problems, central to decision-making scenarios like logistics and manufacturing, are traditionally solved using problem-specific algorithms requiring significant domain expertise. While large language models (LLMs) have shown promise in automating CO problem solving, existing approaches rely on intermediate steps such as code generation or solver invocation, limiting their generality and accessibility. This paper introduces a novel framework that empowers LLMs to serve as end-to-end CO solvers by directly mapping natural language problem descriptions to solutions. We propose a two-stage training strategy: supervised fine-tuning (SFT) imparts LLMs with solution construction patterns from domain-specific solvers, while a feasibility-and-optimality-aware reinforcement learning (FOARL) process explicitly mitigates constraint violations and refines solution quality. Evaluation across seven NP-hard CO problems shows that our method achieves a high feasibility rate and reduces the average optimality gap to 1.03–8.20% by tuning a 7B-parameter LLM, surpassing both general-purpose LLMs (e.g., GPT-4o), reasoning models (e.g., DeepSeek-R1), and domain-specific heuristics. Our method establishes a unified language-based pipeline for CO without extensive code execution or manual architectural adjustments for different problems, offering a general and language-driven alternative to traditional solver design while maintaining relative feasibility guarantees.



Paperid:2066
Authors:Kechi Zhang, Ge Li, Jia Li, Huangzhao Zhang, Yihong Dong, Jia Li, Jingjing Xu, Zhi Jin
Abstract:
Abstract:The Transformer architecture has emerged as a landmark advancement within the broad field of artificial intelligence, effectively catalyzing the advent of large language models (LLMs).However, despite its remarkable capabilities and the substantial progress it has facilitated, the Transformer architecture still has some limitations.One such intrinsic limitation is its inability to effectively capture the Chomsky hierarchy, such as regular expressions or deterministic context-free grammars.Drawing inspiration from pushdown automata, which efficiently resolve deterministic context-free grammars using stacks, we propose StackTrans to address the aforementioned issue within LLMs.Unlike previous approaches that modify the attention computation, StackTrans explicitly incorporates hidden state stacks between Transformer layers. This design maintains compatibility with existing frameworks like flash-attention. Specifically, our design features stack operations -- such as pushing and popping hidden states -- that are differentiable and can be learned in an end-to-end manner.Our comprehensive evaluation spans benchmarks for both Chomsky hierarchies and large-scale natural languages. Across these diverse tasks, StackTrans consistently outperforms standard Transformer models and other baselines. We have successfully scaled StackTrans up from 360M to 7B parameters. In particular, our from-scratch pretrained model StackTrans-360M outperforms several larger open-source LLMs with 2–3$\times$ more parameters, showcasing its superior efficiency and reasoning capability.



Paperid:2067
Authors:Jiaqi Chen, Bang Zhang, Ruotian Ma, Peisong Wang, Xiaodan Liang, Zhaopeng Tu, Xiaolong Li, Kwan-Yee K. Wong
Abstract:
Evaluating the step-by-step reliability of large language model (LLM) reasoning, such as Chain-of-Thought, remains challenging due to the difficulty and cost of obtaining high-quality step-level supervision. In this paper, we introduce Self-Play Critic (SPC), a novel approach where a critic model evolves its ability to assess reasoning steps through adversarial self-play games, eliminating the need for manual step-level annotation. SPC involves fine-tuning two copies of a base model to play two roles, namely a "sneaky generator" that deliberately produces erroneous steps designed to be difficult to detect, and a "critic" that analyzes the correctness of reasoning steps. These two models engage in an adversarial game in which the generator aims to fool the critic, while the critic model seeks to identify the generator's errors. Using reinforcement learning based on the game outcomes, the models iteratively improve; the winner of each confrontation receives a positive reward and the loser receives a negative reward, driving continuous self-evolution. Experiments on three reasoning process benchmarks (ProcessBench, PRM800K, DeltaBench) demonstrate that our SPC progressively enhances its error detection capabilities (e.g., accuracy increases from 70.8% to 77.7% on ProcessBench) and surpasses strong baselines, including distilled R1 model. Furthermore, SPC can guide the test-time search of diverse LLMs and significantly improve their mathematical reasoning performance on MATH500 and AIME2024, surpassing those guided by state-of-the-art process reward models.



Paperid:2068
Authors:Benjamin Matthias Ruppik, Julius von Rohrscheidt, Carel van Niekerk, Michael Heck, Renato Vukovic, Shutong Feng, Hsien-chin Lin, Nurul Lubis, Bastian Rieck, Marcus Zibrowius, Milica Gasic
Abstract:
Understanding the internal mechanisms of large language models (LLMs) remains a challenging and complex endeavor. Even fundamental questions, such as how fine-tuning affects model behavior, often require extensive empirical evaluation. In this paper, we introduce a novel perspective based on the geometric properties of contextual latent embeddings to study the effects of training and fine-tuning. To that end, we measure the local dimensions of a contextual language model's latent space and analyze their shifts during training and fine-tuning.We show that the local dimensions provide insights into the model's training dynamics and generalization ability.Specifically, the mean of the local dimensions predicts when the model’s training capabilities are exhausted, as exemplified in a dialogue state tracking task, overfitting, as demonstrated in an emotion recognition task, and grokking, as illustrated with an arithmetic task.Furthermore, our experiments suggest a practical heuristic: reductions in the mean local dimension tend to accompany and predict subsequent performance gains.Through this exploration, we aim to provide practitioners with a deeper understanding of the implications of fine-tuning on embedding spaces, facilitating informed decisions when configuring models for specific applications. The results of this work contribute to the ongoing discourse on the interpretability, adaptability, and generalizability of LLMs by bridging the gap between intrinsic model mechanisms and geometric properties in the respective embeddings.



Paperid:2069
Authors:Hugo Cui, Cengiz Pehlevan, Yue Lu
Abstract:
In this manuscript, we analyze a solvable model of flow or diffusion-based generative model. We consider the problem of learning a model parametrized by a two-layer auto-encoder, trained with online stochastic gradient descent, on a high-dimensional target density with an underlying low-dimensional manifold structure. We derive a tight asymptotic characterization of low-dimensional projections of the distribution of samples generated by the learned model, ascertaining in particular its dependence on the number of training samples. Building on this analysis, we discuss how mode collapse can arise, and lead to model collapse when the generative model is re-trained on generated synthetic data.



Paperid:2070
Authors:Jian Yao, Ran Cheng, Xingyu Wu, Jibin Wu, KC Tan
Abstract:
The reasoning capabilities of large language models (LLMs) have advanced rapidly, particularly following the release of DeepSeek-R1, which has inspired a surge of research into data quality and reinforcement learning (RL) algorithms. Despite the pivotal role diversity plays in RL, its influence on LLM reasoning remains largely underexplored. To bridge this gap, this work presents a systematic investigation into the impact of diversity in RL-based training for LLM reasoning, and proposes a novel diversity-aware policy optimization method. Across evaluations on 12 LLMs, we observe a strong positive correlation between the solution diversity and potential@k (a novel metric quantifying an LLM’s reasoning potential) in high-performing models. This finding motivates our method to explicitly promote diversity during RL training. Specifically, we design a token-level diversity and reformulate it into a practical objective, then we selectively apply it to positive samples. Integrated into the R1-zero training framework, our method achieves a 3.5\% average improvement across four mathematical reasoning benchmarks, while generating more diverse and robust solutions.



Paperid:2071
Authors:Jiang Qin, Alexandra Gomez-Villa, Senmao Li, Shiqi Yang, Yaxing Wang, KAI WANG, Joost van de Weijer
Abstract:
Recent advances in Text-to-Image (T2I) diffusion models have transformed image generation, enabling significant progress in stylized generation using only a few style reference images. However, current diffusion-based methods struggle with \textit{fine-grained} style customization due to challenges in controlling multiple style attributes, such as color and texture. This paper introduces the first tuning-free approach to achieve free-lunch color-texture disentanglement in stylized T2I generation, addressing the need for independently controlled style elements for the Disentangled Stylized Image Generation (DisIG) problem. Our approach leverages the \textit{Image-Prompt Additivity} property in the CLIP image embedding space to develop techniques for separating and extracting Color-Texture Embeddings (CTE) from individual color and texture reference images. To ensure that the color palette of the generated image aligns closely with the color reference, we apply a whitening and coloring transformation to enhance color consistency. Additionally, to prevent texture loss due to the signal-leak bias inherent in diffusion training, we introduce a noise term that preserves textural fidelity during the Regularized Whitening and Coloring Transformation (RegWCT). Through these methods, our Style Attributes Disentanglement approach (SADis) delivers a more precise and customizable solution for stylized image generation. Experiments on images from the WikiArt and StyleDrop datasets demonstrate that, both qualitatively and quantitatively, SADis surpasses state-of-the-art stylization methods in the DisIG task.



Paperid:2072
Authors:Csaba Dékány, Stefan Balauca, Dimitar I. Dimitrov, Robin Staab, Martin Vechev
Abstract:
Despite recent efforts in Large Language Models (LLMs) safety and alignment, current adversarial attacks on frontier LLMs are still able to force harmful generations consistently. Although adversarial training has been widely studied and shown to significantly improve the robustness of traditional machine learning models, its strengths and weaknesses in the context of LLMs are less understood. Specifically, while existing discrete adversarial attacks are effective at producing harmful content, training LLMs with concrete adversarial prompts is often computationally expensive, leading to reliance on continuous relaxations. As these relaxations do not correspond to discrete input tokens, such latent training methods often leave models vulnerable to a diverse set of discrete attacks. In this work, we aim to bridge this gap by introducing MixAT, a novel method that combines stronger discrete and faster continuous attacks during training. We rigorously evaluate MixAT across a wide spectrum of state‑of‑the‑art attacks, proposing theAt Least One Attack Success Rate(ALO‑ASR) metric to capture the worst‑case vulnerability of models. We show MixAT achieves substantially better robustness (ALO‑ASR <  20%) compared to prior defenses (ALO‑ASR > 50%), while maintaining a runtime comparable to methods based on continuous relaxations. We further analyze MixAT in realistic deployment settings, exploring how chat templates, quantization, low-rank adapters, and temperature affect both adversarial training and evaluation, revealing additional blind spots in current methodologies. Our results demonstrate that MixAT’s discrete-continuous defense offers a principled and superior robustness-accuracy tradeoff with minimal computational overhead, highlighting its promise for building safer LLMs.



Paperid:2073
Authors:Hude Liu, Jerry Yao-Chieh Hu, Zhao Song, Han Liu
Abstract:
Abstract:We establish the universal approximation capability of single-layer, single-head self- and cross-attention mechanisms with minimal attached structures. Our key insight is to interpret single-head attention as an input domain-partition mechanism that assigns distinct values to subregions. This allows us to engineer the attention weights such that this assignment imitates the target function. Building on this, we prove that a single self-attention layer, preceded by sum-of-linear transformations, is capable of approximating any continuous function on a compact domain under the $L_\infty$-norm. Furthermore, we extend this construction to approximate any Lebesgue integrable function under $L_p$-norm for $1\leq p <\infty$. Lastly, we also extend our techniques and show that, for the first time, single-head cross-attention achieves the same universal approximation guarantees.



Paperid:2074
Authors:LinshuangDiao, Sensen Song, Dayong Ren, Yurong Qian
Abstract:
State Space Models (SSMs) like PointMamba provide efficient feature extraction for point cloud self-supervised learning with linear complexity, surpassing Transformers in computational efficiency. However, existing PointMamba-based methods rely on complex token ordering and random masking, disrupting spatial continuity and local semantic correlations. We propose ZigzagPointMamba to address these challenges. The key to our approach is a simple zigzag ordering strategy that globally sequences point cloud tokens, enhancing spatial continuity by preserving the proximity of spatially adjacent point tokens. Yet, random masking impairs local semantic modeling in self-supervised learning. To overcome this, we introduce a Semantic-Siamese Masking Strategy (SMS), which masks semantically similar tokens to facilitate reconstruction by integrating local features of original and similar tokens, thus overcoming dependence on isolated local features and enabling robust global semantic modeling. Our pre-training ZigzagPointMamba weights significantly boost downstream tasks, achieving a 1.59% mIoU gain on ShapeNetPart for part segmentation, a 0.4% higher accuracy on ModelNet40 for classification, and 0.19%, 1.22%, and 0.72% higher accuracies respectively for the classification tasks on the OBJ-BG, OBJ-ONLY, and PB-T50-RS subsets of ScanObjectNN. Code is available at https://anonymous.4open.science/r/ZigzagPointMamba-1800/.



Paperid:2075
Authors:Chuandong Liu, Huijiao Wang, Lei YU, Gui-Song Xia
Abstract:
Recent advances in 3D Gaussian Splatting have shown remarkable potential for novel view synthesis. However, most existing large-scale scene reconstruction methods rely on the divide-and-conquer paradigm, which often leads to the loss of global scene information and requires complex parameter tuning due to scene partitioning and local optimization. To address these limitations, we propose MixGS, a novel holistic optimization framework for large-scale 3D scene reconstruction. MixGS models the entire scene holistically by integrating camera pose and Gaussian attributes into a view-aware representation, which is decoded into fine-detailed Gaussians. Furthermore, a novel mixing operation combines decoded and original Gaussians to jointly preserve global coherence and local fidelity. Extensive experiments on large-scale scenes demonstrate that MixGS achieves state-of-the-art rendering quality and competitive speed, while significantly reducing computational requirements, enabling large-scale scene reconstruction training on a single 24GB VRAM GPU. The code will be released publicly.



Paperid:2076
Authors:Zhongzheng Qiao, Chenghao Liu, Yiming Zhang, Ming Jin, Quang Pham, Qingsong Wen, Ponnuthurai Suganthan, Xudong Jiang, Savitha Ramasamy
Abstract:
Time series foundation models (TSFMs) demonstrate impressive zero-shot performance for time series forecasting. However, an important yet underexplored challenge is how to effectively finetune TSFMs on specific downstream tasks. While naive finetuning can yield performance gains, we argue that it falls short of fully leveraging TSFMs' capabilities, often resulting in overfitting and suboptimal performance. Given the diverse temporal patterns across sampling scales and the inherent multi-scale forecasting capabilities of TSFMs, we adopt a causal perspective to analyze finetuning process, through which we highlight the critical importance of explicitly modeling multiple scales and reveal the shortcomings of naive approaches. Focusing on \textit{encoder-based} TSFMs, we propose \textbf{M}ulti\textbf{\textsc{s}}cale \textbf{\textsc{f}}ine\textbf{\textsc{t}}uning (\textbf{MSFT}), a simple yet general framework that explicitly integrates multi-scale modeling into the finetuning process. Experimental results on three different backbones (\moirai, \moment\ and \units) demonstrate that TSFMs finetuned with MSFT not only outperform naive and typical parameter efficient finetuning methods but also surpass state-of-the-art deep learning methods. Codes are available at \magenta{\url{https://github.com/anonymousauthors1818/MSFT}}.



Paperid:2077
Authors:Wenhao Yang, Lin Li, Xiaohui Tao, Kaize Shi
Abstract:
The imperative of user privacy protection and regulatory compliance necessitates sensitive data removal in model training, yet this process often induces distributional shifts that undermine model performance-particularly in out-of-distribution (OOD) scenarios. We propose a novel data removal approach that enhances deep predictive models through factor decorrelation and loss perturbation. Our approach introduces: (1) a discriminative-preserving factor decorrelation module employing dynamic adaptive weight adjustment and iterative representation updating to reduce feature redundancy and minimize inter-feature correlations. (2) a smoothed data removal mechanism with loss perturbation that creates information-theoretic safeguards against data leakage during removal operations. Extensive experiments on five benchmark datasets show that our approach outperforms other baselines and consistently achieves high predictive accuracy and robustness even under significant distribution shifts. The results highlight its superior efficiency and adaptability in both in-distribution and out-of-distribution scenarios.



Paperid:2078
Authors:Bettina Messmer, Vinko Sabolčec, Martin Jaggi
Abstract:
Dataset curation has become a basis for strong large language model (LLM) performance.While various rule-based filtering heuristics exist for English and multilingual datasets, model-based filtering techniques have primarily focused on English.To address the disparity stemming from limited research on non-English languages, we develop a model-based filtering framework for multilingual datasets that aims to identify a diverse set of structured and knowledge-rich samples.Our approach emphasizes transparency, simplicity, and efficiency, leveraging Transformer- and FastText-based classifiers to ensure the broad accessibility of our technique and data.We conduct comprehensive ablation studies on the FineWeb-2 web crawl dataset across diverse language families, scripts, and resource availability to demonstrate the effectiveness of our method.Training a 1B-parameter Llama model for 70B and 119B tokens, our approach can match the baseline MMLU score with as little as 15\% of the training tokens, while also improving across other benchmarks and mitigating the curse of multilinguality.These findings provide strong evidence for the generalizability of our approach to other languages. As a result, we extend our framework to 20 languages for which we release the refined pretraining datasets.



Paperid:2079
Authors:Roy Xie, Junlin Wang, Paul Rosu, Chunyuan Deng, Bolun Sun, Zihao Lin, Bhuwan Dhingra
Abstract:
Abstract:Large language models (LLMs) process entire input contexts indiscriminately, which is inefficient in cases where theinformation required to answer a query is localized within the context.We present dynamic context cutoff, a human-inspired method enabling LLMs to self-terminate processing upon acquiring sufficient task-relevant information. Through analysis of model internals, we discover that specific attention heads inherently encode ``sufficiency signals'' – detectable through lightweight classifiers – that predict when critical information has been processed. This reveals a new efficiency paradigm: models' internal understanding naturally dictates processing needs rather than external compression heuristics. Comprehensive experiments across six QA datasets (up to 40K tokens) with three model families (LLaMA/Qwen/Mistral, 1B-70B) demonstrate $3.4$\% accuracy improvement while achieving $1.33\times$ token reduction on average. Furthermore, our method demonstrates better performance with the same rate of token reduction compared to other context efficiency methods.Additionally, we observe an emergent scaling phenomenon: while smaller models require probing for sufficiency detection, larger models exhibit intrinsic self-assessment capabilities through prompting.



Paperid:2080
Authors:Junjie Xing, Yeye He, Mengyu Zhou, Haoyu Dong, Shi Han, Lingjiao Chen, Dongmei Zhang, Surajit Chaudhuri, H. V. Jagadish
Abstract:
Tables and table-based use cases play a crucial role in many important real-world applications, such as spreadsheets, databases, and computational notebooks, which traditionally require expert-level users like data engineers, data analysts, and database administrators to operate. Although LLMs have shown remarkable progress in working with tables (e.g., in spreadsheet and database copilot scenarios), comprehensive benchmarking of such capabilities remains limited. In contrast to an extensive and growing list of NLP benchmarks, evaluations of table-related tasks are scarce, and narrowly focus on tasks like NL-to-SQL and Table-QA, overlooking the broader spectrum of real-world tasks that professional users face. This gap limits our understanding and model progress in this important area.In this work, we introduce MMTU, a large-scale benchmark with over 30K questions across 25 real-world table tasks, designed to comprehensively evaluate models ability to understand, reason, and manipulate real tables at the expert-level. These tasks are drawn from decades’ worth of computer science research on tabular data, with a focus on complex table tasks faced by professional users. We show that MMTU require a combination of skills -- including table understanding, reasoning, and coding -- that remain challenging for today's frontier models, where even frontier reasoning models like OpenAI o4-mini and DeepSeek r1 score only 61%, suggesting significant room for improvement. We highlight key findings in our evaluation using MMTU and hope this benchmark drives further advances in understanding and developing foundation models for structured data processing and analysis.



Paperid:2081
Authors:Hanwen Liu, Yadong Mu
Abstract:
This work introduces Clipout, a method for removing a target concept in pre-trained text-to-image models. By randomly clipping units from the learned data embedding and using a contrastive objective, models are encouraged to differentiate these clipped embedding vectors. Our goal is to remove private, copyrighted, inaccurate, or harmful concepts from trained models without the need for retraining. This is achieved by considering only negative samples and generating them in a bootstrapping-like manner, requiring minimal prior knowledge. Additionally, theoretical analyses are provided to further understand our proposed Clipout. Extensive experiments on text-to-image show that Clipout is simple yet highly effective and efficient compared with previous state-of-the-art approaches.



Paperid:2082
Authors:Siqiao Mu, Diego Klabjan
Abstract:
Abstract:Machine unlearning algorithms aim to efficiently remove data from a model without retraining it from scratch, in order to remove corrupted or outdated data or respect a user's "right to be forgotten." Certified machine unlearning is a strong theoretical guarantee based on differential privacy that quantifies the extent to which an algorithm erases data from the model weights. In contrast to existing works in certified unlearning for convex or strongly convex loss functions, or nonconvex objectives with limiting assumptions, we propose the first, first-order, black-box (i.e., can be applied to models pretrained with vanilla gradient descent) algorithm for unlearning on general nonconvex loss functions, which unlearns by ``rewinding" to an earlier step during the learning process before performing gradient descent on the loss function of the retained data points. We prove $(\epsilon, \delta)$ certified unlearning and performance guarantees that establish the privacy-utility-complexity tradeoff of our algorithm, and we prove generalization guarantees for nonconvex functions that satisfy the Polyak-Lojasiewicz inequality. Finally, we demonstrate the superior performance of our algorithm compared to existing methods, within a new experimental framework that more accurately reflects unlearning user data in practice.



Paperid:2083
Authors:Jiandong Shao, Yao Lu, Jianfei Yang
Abstract:
Large Language Models (LLMs) exhibit impressive performance on complex reasoning tasks, yet they frequently fail on basic numerical problems, producing incorrect outputs. Inspired by Benford’s Law---a statistical pattern where lower digits occur more frequently as leading digits---we hypothesize that the long-tailed digit distributions in web-collected corpora may be learned by LLMs during pretraining, leading to biased numerical generation. To investigate the hypothesis, we first examine whether digits frequencies in pretraining corpus (OLMo2) follows Benford's law. We then construct an evaluation benchmark with uniformly distributed ground-truth digits across seven numerical reasoning tasks. Our evaluation results demonstrate that leading open-source LLMs show a consistent pattern of digit bias that resembles Benford's law. Through logit-lens tracing and neuron-level dissection, we identify that this bias arises predominantly from a small subset of highly digit-selective feed-forward network (FFN) neurons in the deeper layers. Finally, we demonstrate that pruning these neurons mitigates imbalanced overgeneration and partially corrects erroneous outputs, providing causal evidence that fine-grained pretraining digit bias can propagate into model behavior. Our findings reveal a fundamental connection between corpus-level statistics and symbolic failure modes in LLMs, offering a new lens for diagnosing and mitigating hallucinations in numerical tasks.



Paperid:2084
Authors:HAORAN YANG, Xiaohui Chen, Chuan-Xian Ren
Abstract:
Aiming to generalize the well-trained gaze estimation model to new target domains, Cross-domain Gaze Estimation (CDGE) is developed for real-world application scenarios. Existing CDGE methods typically extract the domain-invariant features to mitigate domain shift in feature space, which is proved insufficient by Generalized Label Shift (GLS) theory. In this paper, we introduce a novel GLS perspective to CDGE and modelize the cross-domain problem by label and conditional shift problem. A GLS correction framework is presented and a feasible realization is proposed, in which a importance reweighting strategy based on truncated Gaussian distribution is introduced to overcome the continuity challenges in label shift correction. To embed the reweighted source distribution to conditional invariant learning, we further derive a probability-aware estimation of conditional operator discrepancy. Extensive experiments on standard CDGE tasks with different backbone models validate the superior generalization capability across domain and applicability on various models of proposed method.



Paperid:2085
Authors:Kai Jiang, Zhengyan Shi, Dell Zhang, Hongyuan Zhang, Xuelong Li
Abstract:
Class Incremental Learning (CIL) aims to continuously learn new categories while retaining the knowledge of old ones. Pre-trained models (PTMs) show promising capabilities in CIL. However, existing approaches that apply lightweight fine-tuning to backbones still induce parameter drift, thereby compromising the generalization capability of pre-trained models. Parameter drift can be conceptualized as a form of noise that obscures critical patterns learned for previous tasks. However, recent researches have shown that noise is not always harmful. For example, the large number of visual patterns learned from pre-training can be easily abused by a single task, and introducing appropriate noise can suppress some low-correlation features, thus leaving a margin for future tasks. To this end, we propose learning beneficial noise for CIL guided by information theory and propose Mixture of Noise (MiN), aiming to mitigate the degradation of backbone generalization from adapting new tasks. Specifically, task-specific noise is learned from high-dimension features of new tasks. Then, a set of weights is adjusted dynamically for optimal mixture of different task noise. Finally, MiN embeds the beneficial noise into the intermediate features to mask the response of inefficient patterns. Extensive experiments on six benchmark datasets demonstrate that MiN achieves state-of-the-art performance in most incremental settings, with particularly outstanding results in 50-steps incremental settings. This shows the significant potential for beneficial noise in continual learning. Anonymous code link is available at https://anonymous.4open.science/r/MiN-16097.



Paperid:2086
Authors:Rubo Wang, Xingyu Gao, Peilin Zhao
Abstract:
Abstract:With the development of biotechnology, RNA therapies have shown great potential. However, different from proteins, the sequences corresponding to a single RNA three-dimensional structure are more abundant. Most of the existing RNA design methods merely take into account the secondary structure of RNA, or are only capable of generating a limited number of candidate sequences. To address the limitations of existing methods, we propose a geometric-algebra-enhanced $\textbf{B}$ayesian $\textbf{F}$low $\textbf{N}$etwork for the inverse design of $\textbf{R}$NA, called $\textbf{RBFN}$. RBFN uses a Bayesian flow network to model the distribution of nucleotide sequences in RNA, enabling the generation of more reasonable RNA sequences. Meanwhile, considering the more flexible characteristics of RNA conformations, we utilize geometric algebra to enhance the modeling ability of the RNA three-dimensional structure, facilitating a better understanding of RNA structural properties. In addition, due to the scarcity of RNA structures and the limitation that there are only four types of nucleic acids, we propose a new time-step distribution sampling to address the scarcity of RNA structure data and the relatively small number of nucleic acid types. Evaluation on the single-state fixed-backbone re-design benchmark and multi-state fixed-backbone benchmark indicates that RBFN can outperform existing RNA design methods in various RNA design tasks, enabling effective RNA sequence design.



Paperid:2087
Authors:Paul Brunzema, Sebastian Trimpe
Abstract:
We introduce BayeSQP, a novel algorithm for general black-box optimization that merges the structure of sequential quadratic programming with concepts from Bayesian optimization. BayeSQP employs second-order Gaussian process surrogates for both the objective and constraints to jointly model the function values, gradients, and Hessian from only zero-order information. At each iteration, a local subproblem is constructed using the GP posterior estimates and solved to obtain a search direction. Crucially, the formulation of the subproblem explicitly incorporates uncertainty in both the function and derivative estimates, resulting in a tractable second-order cone program for high probability improvements under model uncertainty. A subsequent one-dimensional line search via constrained Thompson sampling selects the next evaluation point. Empirical results show that BayeSQP outperforms state-of-the-art methods in specific high-dimensional settings. Our algorithm offers a principled and flexible framework that bridges classical optimization techniques with modern approaches to black-box optimization.



Paperid:2088
Authors:Daniel Ovalle, Lorenz Biegler, Ignacio Grossmann, Carl Laird, Mateo Dulce Rubio
Abstract:
Abstract:We propose Conformal Mixed-Integer Constraint Learning (C-MICL), a novel framework that provides probabilistic feasibility guarantees for data-driven constraints in optimization problems. While standard Mixed-Integer Constraint Learning methods often violate the true constraints due to model error or data limitations, our C-MICL approach leverages conformal prediction to ensure feasible solutions are ground-truth feasible. This guarantee holds with probability at least $1{-}\alpha$, under a conditional independence assumption. The proposed framework supports both regression and classification tasks without requiring access to the true constraint function, while avoiding the scalability issues associated with ensemble-based heuristics. Experiments on real-world applications demonstrate that C-MICL consistently achieves target feasibility rates, maintains competitive objective performance, and significantly reduces computational cost compared to existing methods. Our work bridges mathematical optimization and machine learning, offering a principled approach to incorporate uncertainty-aware constraints into decision-making with rigorous statistical guarantees.



Paperid:2089
Authors:Jingbo Yang, Bairu Hou, Wei Wei, Yujia Bao, Shiyu Chang
Abstract:
We describe KVLink, an approach for efficient key-value (KV) cache reuse in large language models (LLMs). In many LLM applications, different inputs can share overlapping context, such as the same retrieved document appearing in multiple queries. However, the LLMs still need to encode the entire context for each query, leading to redundant computation. In this paper, we investigate a new strategy to eliminate such inefficiency, where the KV cache of each document is precomputed independently. During inference, the KV caches of retrieved documents are concatenated, allowing the model to reuse cached representations instead of recomputing them. To mitigate the performance degradation when using KV caches computed independently for each document, KVLink introduces two key techniques: adjusting positional embeddings of the KV cache at inference to match the global position after concatenation, and using trainable special tokens to restore self-attention across independently encoded documents. Experiments across 7 datasets demonstrate that KVLink improves question answering accuracy by an average of 4% over state-of-the-art methods. Furthermore, by leveraging precomputed KV caches, our approach reduces time-to-first-token by up to 96% compared to standard LLM inference, making it a scalable and efficient solution for context reuse. Additionally, KVLink can be combined with KV cache compression to further save cache loading and storage overhead while outperforming the baselines. Code is available in the supplementary material.



Paperid:2090
Authors:Niklas Freymuth, Tobias Würth, Nicolas Schreiber, Balázs Gyenes, Andreas Boltres, Johannes Mitsch, Aleksandar Taranovic, Tai Hoang, Philipp Dahlinger, Philipp Becker, Luise Kärger, Gerhard Neumann
Abstract:
The cost and accuracy of simulating complex physical systems using the Finite Element Method (FEM) scales with the resolution of the underlying mesh. Adaptive meshes improve computational efficiency by refining resolution in critical regions, but typically require task-specific heuristics or cumbersome manual design by a human expert. We propose Adaptive Meshing By Expert Reconstruction (AMBER), a supervised learning approach to mesh adaptation. Starting from a coarse mesh, AMBER iteratively predicts the sizing field, i.e., a function mapping from the geometry to the local element size of the target mesh, and uses this prediction to produce a new intermediate mesh using an out-of-the-box mesh generator. This process is enabled through a hierarchical graph neural network, and relies on data augmentation by automatically projecting expert labels onto AMBER-generated data during training. We evaluate AMBER on 2D and 3D datasets, including classical physics problems, mechanical components, and real-world industrial designs with human expert meshes. AMBER generalizes to unseen geometries and consistently outperforms multiple recent baselines, including ones using Graph and Convolutional Neural Networks, and Reinforcement Learning-based approaches.



Paperid:2091
Authors:Yaming Yang, Ziyu Zheng, Weigang Lu, Zhe Wang, Xinyan Huang, Wei Zhao, Ziyu Guan
Abstract:
Heterogeneous hypergraph is a kind of structural data that contains multiple types of nodes and multiple types of hyperedges. Each hyperedge type corresponds to a specific multi-ary relation (called hyper-relation) among subsets of nodes, which goes beyond traditional pair-wise relations in simple graphs. Existing representation learning methods for heterogeneous hypergraphs typically learn embeddings for nodes and hyperedges based on graph neural networks. Although achieving promising performance, they are still limited in capturing more complex structural features and richer semantics conveyed by the composition of various hyper-relations. To fill this research gap, in this work, we propose the concept of hyper-meta-path for heterogeneous hypergraphs, which is defined as the composition of a sequence of hyper-relations. Besides, we design an attention-based heterogeneous hypergraph neural network (HHNN) to automatically learn the importance of hyper-meta-paths. By exploiting useful ones, HHNN is able to capture more complex structural features to boost the model's performance, as well as leverage their conveyed semantics to improve the model's interpretability. Extensive experiments show that HHNN can achieve significantly better performance than state-of-the-art baselines, and the discovered hyper-meta-paths bring good interpretability for the model predictions. To facilitate the reproducibility of this work, we provide our dataset as well as anonymized source code at: https://anonymous.4open.science/r/HHNN.



Paperid:2092
Authors:Qiusheng Huang, Yuan Niu, Xiaohui Zhong, AnboyuGuo, Lei Chen, dianjun zhang, Xuefeng Zhang, Hao Li
Abstract:
Accurate, high-resolution ocean forecasting is crucial for maritime operations and environmental monitoring. While traditional numerical models are capable of producing sub-daily, eddy-resolving forecasts, they are computationally intensive and face challenges in maintaining accuracy at fine spatial and temporal scales. In contrast, recent data-driven approaches offer improved computational efficiency and emerging potential, yet typically operate at daily resolution and struggle with sub-daily predictions due to error accumulation over time.We introduce FuXi-Ocean, the first data-driven global ocean forecasting model achieving six-hourly predictions at eddy-resolving 1/12° spatial resolution, reaching depths of up to 1500 meters. The model architecture integrates a context-aware feature extraction module with a predictive network employing stacked attention blocks. The core innovation is the Mixture-of-Time (MoT) module, which adaptively integrates predictions from multiple temporal contexts by learning variable-specific reliability, mitigating cumulative errors in sequential forecasting. Through comprehensive experimental evaluation, FuXi-Ocean demonstrates superior skill in predicting key variables, including temperature, salinity, and currents, across multiple depths.



Paperid:2093
Authors:Sameera Ramasinghe, Thalaiyasingam Ajanthan, Hadi Mohaghegh Dolatabadi, Gil Avraham, Violetta Shevchenko, Yan Zuo, Chamin Hewa Koneputugodage, Alexander Long
Abstract:
Pretraining language models with extended context windows enhances their ability to leverage rich information during generation. Existing methods split input sequences into chunks, broadcast them across multiple devices, and compute attention block by block which incurs significant communication overhead. While feasible in high-speed clusters, these methods are impractical for decentralized training over low-bandwidth connections. We propose a compression method for communication-efficient context parallelism in decentralized settings, achieving a remarkable compression rate of over 95% with negligible overhead and no loss in convergence. Our key insight is to exploit the intrinsic low-rank structure of activation outputs by dynamically constraining them to learned mixtures of subspaces via efficient reparameterizations. We demonstrate scaling billion-parameter decentralized models to context lengths exceeding 100K tokens on networks as slow as 300Mbps, matching the wall-clock convergence speed of centralized models on 100Gbps interconnects.



Paperid:2094
Authors:Renpu Liu, Jing Yang
Abstract:
Large language models (LLMs) exhibit impressive in‑context learning (ICL) capabilities, yet the quality of their predictions is fundamentally limited by the few costly labeled demonstrations that can fit into a prompt. Meanwhile, there exist vast and continuously growing amounts of unlabeled data that may be closely related to the ICL task. How to utilize such unlabeled data to provably enhance the performance of ICL thus becomes an emerging fundamental question. In this work, we propose a novel {\it augmented ICL} framework, in which the prompt includes a small set of labeled examples alongside a block of unlabeled inputs. We focus on the multi-class linear classification setting and demonstrate that, with chain-of-thought (CoT) prompting, a multi-layer transformer can effectively emulate an expectation–maximization (EM) algorithm. This enables the transformer to implicitly extract useful information from both labeled and unlabeled data, leading to provable improvements in ICL accuracy. Moreover, we show that such a transformer can be trained via teacher forcing, with its parameters converging to the desired solution at a sub-linear rate. Experiments demonstrate that the augmented ICL framework consistently outperforms conventional few-shot ICL, providing empirical support for our theoretical findings. To the best of our knowledge, this is the first theoretical study on the impact of unlabeled data on the ICL performance of transformers.



Paperid:2095
Authors:Camila Kolling, Vy Vo, Mariya Toneva
Abstract:
Associative learning--forming links between co-occurring items--is fundamental to human cognition, reshaping internal representations in complex ways. Competing hypotheses propose that associative learning in the brain either integrates similar items to promote generalization, differentiates them to minimize interference, or--according to a more intricate theory--follows a non-monotonic pattern where moderately similar items differentiate while highly similar or dissimilar items integrate or remain stable. Testing these hypotheses in biological systems is challenging, but large language models (LLMs) offer a scalable alternative. Leveraging LLMs' in-context learning (ICL), we adapt a cognitive neuroscience associative learning paradigm and investigate representational changes across six LLMs. Our findings align best with the Non-Monotonic Plasticity Hypothesis (NMPH), showing that moderately similar items differentiate after learning.However, we find that similarity to the model’s broader prior knowledge plays an important role by amplifying differentiation. Our findings position LLMs not only as powerful tools for studying representational dynamics in human-like learning systems, but also as accessible and general computational models for generating new hypotheses about the principles underlying memory reorganization in the brain.



Paperid:2096
Authors:Lingshun Kong, Jiawei Zhang, Dongqing Zou, Fu Lee Wang, SIJIE REN, Xiaohe Wu, Jiangxin Dong, Jinshan Pan
Abstract:
Diffusion models have achieved significant progress in image generation and the pre-trained Stable Diffusion (SD) models are helpful for image deblurring by providing clear image priors. However, directly using a blurry image or a pre-deblurred one as a conditional control for SD will either hinder accurate structure extraction or make the results overly dependent on the deblurring network. In this work, we propose a Latent Kernel Prediction Network (LKPN) to achieve robust real-world image deblurring. Specifically, we co-train the LKPN in the latent space with conditional diffusion. The LKPN learns a spatially variant kernel to guide the restoration of sharp images in the latent space. By applying element-wise adaptive convolution (EAC), the learned kernel is utilized to adaptively process the blurry feature, effectively preserving the information of the blurry input. This process thereby more effectively guides the generative process of SD, enhancing both the deblurring efficacy and the quality of detail reconstruction. Moreover, the results at each diffusion step are utilized to iteratively estimate the kernels in LKPN to better restore the sharp latent by EAC in the subsequent step. This iterative refinement enhances the accuracy and robustness of the deblurring process. Extensive experimental results demonstrate that the proposed method outperforms state-of-the-art image deblurring methods on both benchmark and real-world images.



Paperid:2097
Authors:Zhongyu Xia, Jishuo Li, Zhiwei Lin, Xinhao Wang, Yongtao Wang, Ming-Hsuan Yang
Abstract:
Open-world perception aims to develop a model adaptable to novel domains and various sensor configurations and can understand uncommon objects and corner cases. However, current research lacks sufficiently comprehensive open-world 3D perception benchmarks and robust generalizable methodologies. This paper introduces OpenAD, the first real open-world autonomous driving benchmark for 3D object detection. OpenAD is built upon a corner case discovery and annotation pipeline that integrates with a multimodal large language model (MLLM). The proposed pipeline annotates corner case objects in a unified format for five autonomous driving perception datasets with 2000 scenarios. In addition, we devise evaluation methodologies and evaluate various open-world and specialized 2D and 3D models. Moreover, we propose a vision-centric 3D open-world object detection baseline and further introduce an ensemble method by fusing general and specialized models to address the issue of lower precision in existing open-world methods for the OpenAD benchmark. We host an online challenge on EvalAI. Data, toolkit codes, and evaluation codes are available at https://github.com/VDIGPKU/OpenAD.



Paperid:2098
Authors:Dongyang Fan, Vinko Sabolčec, Martin Jaggi
Abstract:
Large Language Models (LLMs) are commonly pretrained on vast corpora of text without utilizing contextual metadata such as source, quality, or topic, leading to a context-free learning paradigm. While recent studies suggest that adding metadata like URL information as context (i.e., auxiliary inputs not used in the loss calculation) can improve training efficiency and downstream performance, they offer limited understanding of which types of metadata are truly effective and under what conditions. In this work, we conduct a systematic evaluation and find that not all metadata types contribute equally. Only URL context speeds up training, whereas quality scores and topic/format domain information offer no clear benefit. Furthermore, the improved downstream performances of URL conditioning emerge only when longer prompts are used at inference time. In addition, we demonstrate that context-aware pretraining enables more controllable generation than context-free pretraining, in a classifier-free guidance fashion. Although topic and format metadata do not accelerate training, they are effective for steering outputs, offering human-interpretable control over generation.



Paperid:2099
Authors:Lingdong Kong, Dongyue Lu, Alan Liang, Rong Li, Yuhao Dong, Tianshuai Hu, Lai Xing Ng, Wei Tsang Ooi, Benoit Cottereau
Abstract:
Event cameras offer microsecond-level latency and robustness to motion blur, making them ideal for understanding dynamic environments. Yet, connecting these asynchronous streams to human language remains an open challenge. We introduce Talk2Event, the first large-scale benchmark for language-driven object grounding in event-based perception. Built from real-world driving data, Talk2Event provides over 30,000 validated referring expressions, each enriched with four grounding attributes -- appearance, status, relation to viewer, and relation to other objects -- bridging spatial, temporal, and relational reasoning. To fully exploit these cues, we propose EventRefer, an attribute-aware grounding framework that dynamically fuses multi-attribute representations through a Mixture of Event-Attribute Experts (MoEE). Our method adapts to different modalities and scene dynamics, achieving consistent gains over state-of-the-art baselines in event-only, frame-only, and event-frame fusion settings. We hope our dataset and approach will establish a foundation for advancing multimodal, temporally-aware, and language-driven perception in real-world robotics and autonomy.



Paperid:2100
Authors:Milad Sefidgaran, Kimia Nadjahi, Abdellatif Zaidi
Abstract:
Abstract:In this paper, we leverage stochastic projection and lossy compression to establish new conditional mutual information (CMI) bounds on the generalization error of statistical learning algorithms. It is shown that these bounds are generally tighter than existing ones. In particular, we prove that for certain problem instances for which existing MI and CMI bounds where shown recently in Attias et al. [2024] and Livni [2023] to become vacuous or fail to describe the right generalization behavior, our bounds yield suitable generalization guarantees of the order of $\mathcal{O}(1/\sqrt{n})$, where $n$ is the size of the training dataset. Furthermore, we use our bounds to investigate the problem of data "memorization" raised in those works, and which asserts that there are learning problem instances for which for any learning algorithm there exist distributions under which the algorithm "memorizes'' the training dataset. We show that for every learning algorithm and every data distribution, there exists an algorithm that does *not* memorize and which yields comparable generalization error. Finally, we apply our approach to derive a novel generalization bound for random subspace algorithms and use it in the context of SGD and SGLD optimization to establish more explicit bounds.



Paperid:2101
Authors:Zidi Xiong, Shan Chen, Zhenting Qi, Himabindu Lakkaraju
Abstract:
Large Reasoning Models (LRMs) have significantly enhanced their capabilities in complex problem-solving by introducing a thinking draft that enables multi-path Chain-of-Thought explorations before producing final answers. Ensuring the faithfulness of these intermediate reasoning processes is crucial for reliable monitoring, interpretation, and effective control. In this paper, we propose a systematic counterfactual intervention framework to rigorously evaluatethinking draft faithfulness. Our approach focuses on two complementary dimensions:(1) Intra-Draft Faithfulness, which assesses whether individual reasoning steps causally influence subsequent steps and the final draft conclusion through counterfactual step insertions; and(2) Draft-to-Answer Faithfulness, which evaluates whether final answers are logically consistent with and dependent on the thinking draft, by perturbing the draft’s concluding logic.We conduct extensive experiments across six state-of-the-art LRMs. Our findings show that current LRMs demonstrate selective faithfulness to intermediate reasoning steps and frequently fail to faithfully align with the draft conclusions.These results underscore the need for more faithful and interpretable reasoning in advanced LRMs.



Paperid:2102
Authors:Jerome Quenum, Wen-Han Hsieh, Tsung-Han (Patrick) Wu, Ritwik Gupta, Trevor Darrell, David Chan
Abstract:
Segmentation models can recognize a pre-defined set of objects in images. However, segmentation models capable of "reasoning" over complex user queries that implicitly refer to multiple objects of interest remain underexplored, especially in the geospatial domain. Recent advances in "reasoning segmentation"---generating segmentation masks from complex, implicit query text---demonstrate the potential of vision-language models (VLMs) to reason across an open domain of objects. Yet, our experiments reveal that these models struggle when applied to the unique challenges of remote-sensing imagery. To address this gap, we introduce a new dataset which consists of: GRES, a curated geospatial reasoning-segmentation dataset with 27,615 annotations across 9,205 images, and PreGRES, a collection of existing datasets to make up a large-scale multimodal pretraining corpus with over 1M question-answer pairs across 119,279 images. We propose an initial benchmark model, LISAt, a VLM for geospatial analysis that can describe complex remote-sensing scenes, answer detailed queries, and segment objects based on natural-language prompts. LISAt establishes a strong initial geospatial benchmark, outperforming prior foundation models such as RS-GPT4V by 10.04\% (BLEU-4) on visual description tasks and surpassing open-domain models on geospatial reasoning segmentation by 143.36\% (gIoU). Our model, dataset, and code are available on our project page: https://lisat-bair.github.io/LISAt/.



Paperid:2103
Authors:睿涵 杨, yikai zhang, Chen, Xintao Wang, Jiangjie Chen, Siyu Yuan, Deqing Yang, Yanghua Xiao
Abstract:
Large language models (LLMs) have enabled agents to perform complex reasoning and decision-making through free-form language interactions. However, in open-ended language action environments (e.g., negotiation or question-asking games), the action space can be formulated as a joint distribution over tokens, resulting in an extremely large and combinatorial action space. Sampling actions in such a space can lead to extreme reward sparsity, which brings large reward variance, hindering effective reinforcement learning (RL). To address this, we proposeARIA, a method thatAggregatesRewards inIntention space to enable efficient and effective languageAgents training. ARIA aims to project natural language actions from the high-dimensional joint token distribution space into a low-dimensional intention space, where semantically similar actions are clustered and assigned shared rewards. This intention-aware reward aggregation reduces reward variance by densifying reward signals, fostering efficient and effective policy optimization. Extensive experiments demonstrate that ARIA not only significantly reduces gradient variance, but also delivers substantial performance gains of average 9.95% across four downstream tasks (e.g., negotiation and text-based games), consistently outperforming strong offline and online RL baselines.



Paperid:2104
Authors:Xiaochen Shang, Pengwei Luo, Xinning Wang, Jiayue Zhao, Huilin Ge, Bo Dong, Xin Yang
Abstract:
Unmanned aerial vehicles could accurately accomplish complex navigation and obstacle avoidance tasks under external control. However, enabling unmanned aerial vehicles (UAVs) to rely solely on onboard computation and sensing for real-time navigation and dynamic obstacle avoidance remains a significant challenge due to stringent latency and energy constraints. Inspired by the efficiency of biological systems, we propose a fully neuromorphic framework achieving end-to-end obstacle avoidance during navigation with an overall latency of just 2.3 milliseconds. Specifically, our bio-inspired approach enables accurate moving object detection and avoidance without requiring target recognition or trajectory computation. Additionally, we introduce the first monocular event-based pose correction dataset with over 50,000 paired and labeled event streams. We validate our system on an autonomous quadrotor using only onboard resources, demonstrating reliable navigation and avoidance of diverse obstacles moving at speeds up to 10 m/s.



Paperid:2105
Authors:Mengchen Zhao, Yifan Gao, Yaqing Hou, Xiangyang Li, Pengjie Gu, Zhenhua Dong, Ruiming Tang, Yi Cai
Abstract:
Recommendation models are predominantly trained using implicit user feedback, since explicit feedback is often costly to obtain. However, implicit feedback, such as clicks, does not always reflect users' real preferences. For example, a user might click on a news article because of its attractive headline, but end up feeling uncomfortable after reading the content. In the absence of explicit feedback, such erroneous implicit signals may severely mislead recommender systems. In this paper, we propose MTRec, a novel sequential recommendation framework designed to align with real user preferences by uncovering their internal satisfaction on recommended items. Specifically, we introduce a mental reward model to quantify user satisfaction and propose a distributional inverse reinforcement learning approach to learn it. The learned mental reward model is then used to guide recommendation models to better align with users’ real preferences. Our experiments show that MTRec brings significant improvements to a variety of recommendation models. We also deploy MTRec on an industrial short video platform and observe a 7\% increase in average user viewing time.



Paperid:2106
Authors:Avery Hee-Woon Ryoo, Nanda H Krishna, Ximeng Mao, Mehdi Azabou, Eva Dyer, Matthew Perich, Guillaume Lajoie
Abstract:
Real-time decoding of neural spiking data is a core aspect of neurotechnology applications such as brain-computer interfaces, where models are subject to strict latency constraints. Traditional methods, including simple recurrent neural networks, are fast and lightweight but are less equipped for generalization to unseen data. In contrast, recent Transformer-based approaches leverage large-scale neural datasets to attain strong generalization performance. However, these models typically have much larger computational requirements and are not suitable for settings requiring low latency or limited memory. To address these shortcomings, we present POSSM, a novel architecture that combines individual spike tokenization and an input cross-attention module with a recurrent state-space model (SSM) backbone, thereby enabling (1) fast and causal online prediction on neural activity and (2) efficient generalization to new sessions, individuals, and tasks through multi-dataset pre-training. We evaluate our model's performance in terms of decoding accuracy and inference speed on monkey reaching datasets, and show that it extends to clinical applications, namely handwriting and speech decoding. Notably, we demonstrate that pre-training on monkey motor-cortical recordings improves decoding performance on the human handwriting task, highlighting the exciting potential for cross-species transfer. In all of these tasks, we find that POSSM achieves comparable decoding accuracy with state-of-the-art Transformers, at a fraction of the inference cost. These results suggest that hybrid SSMs may be the key to bridging the gap between accuracy, inference speed, and generalization when training neural decoders for real-time, closed-loop applications.



Paperid:2107
Authors:Rustem Islamov, Niccolò Ajroldi, Antonio Orvieto, Aurelien Lucchi
Abstract:
Abstract:Modern optimization algorithms that incorporate momentum and adaptive step-size offer improved performance in numerous challenging deep learning tasks. However, their effectiveness is often highly sensitive to the choice of hyperparameters, especially the step-size. Tuning these parameters is often difficult, resource-intensive, and time-consuming. Therefore, recent efforts have been directed toward enhancing the stability of optimizers across a wide range of hyperparameter choices [Schaipp et al., 2024]. In this paper, we introduce an algorithm that matches the performance of state-of-the-art optimizers while improving stability to the choice of the step-size hyperparameter through a novel adaptation of the ${\sf NGN}$ step-size method [Orvieto and Xiao, 2024]. Specifically, we propose a momentum-based version ${\sf NGN}$-${\sf M}$ that attains the standard convergence rate of $\mathcal{O}(1/\sqrt{K})$ under less restrictive assumptions, without the need for interpolation condition or assumptions of bounded stochastic gradients or iterates, in contrast to previous approaches. Additionally, we empirically demonstrate that the combination of the ${\sf NGN}$ step-size with momentum results in enhanced robustness to the choice of the step-size hyperparameter while delivering performance that is comparable to or surpasses other state-of-the-art optimizers.



Paperid:2108
Authors:Yuhan Zhang, Long Zhuo, Ziyang Chu, Tong Wu, Zhibing Li, Liang Pan, Dahua Lin, Ziwei Liu
Abstract:
Despite rapid advances in 3D content generation, quality assessment for the generated 3D assets remains challenging.Existing methods mainly rely on image-based metrics and operate solely at the object level, limiting their ability to capture spatial Despite rapid advances in 3D content generation, quality assessment for the generated 3D assets remains challenging.Existing methods mainly rely on image-based metrics and operate solely at the object level, limiting their ability to capture spatial coherence, material authenticity, and high-fidelity local details.1) To address these challenges, we introduce Hi3DEval, a hierarchical evaluation framework tailored for 3D generative content. It combines both object-level and part-level evaluation, enabling holistic assessments across multiple dimensions as well as fine-grained quality analysis. Additionally, we extend texture evaluation beyond aesthetic appearance by explicitly assessing material realism, focusing on attributes such as albedo, saturation, and metallicness. 2) To support this framework, we construct Hi3DBench, a large-scale dataset comprising diverse 3D assets and high-quality annotations, accompanied by a reliable multi-agent annotation pipeline.We further propose a 3D-aware automated scoring system based on hybrid 3D representations. Specifically, we leverage video-based representations for object-level and material-subject evaluations to enhance modeling of spatio-temporal consistency and employ pretrained 3D features for part-level perception.Extensive experiments demonstrate that our approach outperforms existing image-based metrics in modeling 3D characteristics and achieves superior alignment with human preference, providing a scalable alternative to manual evaluations.



Paperid:2109
Authors:Byung-Kwan Lee, Ryo Hachiuma, Yong Man Ro, Frank Wang, Yueh-Hua Wu
Abstract:
Vision-Language Models (VLMs) have achieved remarkable progress, yet their large scale often renders them impractical for resource-constrained environments. This paper introduces Unified Reinforcement and Imitation Learning (RIL), a novel and efficient training algorithm designed to create powerful, lightweight VLMs. RIL distinctively combines the strengths of reinforcement learning with adversarial imitation learning. This enables smaller student VLMs not only to mimic the sophisticated text generation of large teacher models but also to systematically improve their generative capabilities through reinforcement signals. Key to our imitation framework is a LLM-based discriminator that adeptly distinguishes between student and teacher outputs, complemented by guidance from multiple large teacher VLMs to ensure diverse learning. This unified learning strategy, leveraging both reinforcement and imitation, empowers student models to achieve significant performance gains, making them competitive with leading closed-source VLMs. Extensive experiments on diverse vision-language benchmarks demonstrate that RIL significantly narrows the performance gap with state-of-the-art open- and closed-source VLMs and, in several instances, surpasses them.



Paperid:2110
Authors:Yujin Jeong, Arnas Uselis, Seong Joon Oh, Anna Rohrbach
Abstract:
Understanding visual scenes is fundamental to human intelligence. While discriminative models have significantly advanced computer vision, they often struggle with compositional understanding. In contrast, recent generative text-to-image diffusion models excel at synthesizing complex scenes, suggesting inherent compositional capabilities.Building on this, zero-shot diffusion classifiers have been proposed to repurpose diffusion models for discriminative tasks. While prior work offered promising results in discriminative compositional scenarios, these results remain preliminary due to a small number of benchmarks and a relatively shallow analysis of conditions under which the models succeed. To address this, we present a comprehensive study of the discriminative capabilities of diffusion classifiers on a wide range of compositional tasks. Specifically, our study covers three diffusion models (SD 1.5, 2.0, and, for the first time, 3-m) spanning 10 datasets and over 30 tasks. Further, we shed light on the role that target dataset domains play in respective performance; to isolate the domain effects, we introduce a new diagnostic benchmark Self-Bench comprised of images created by diffusion models themselves. Finally, we explore the importance of timestep weighting and uncover a relationship between domain gap and timestep sensitivity, particularly for SD3-m.To sum up, diffusion classifiers understand compositionality, but conditions apply!



Paperid:2111
Authors:Chengye Wang, Yuyuan Li, XiaoHua Feng, Chaochao Chen, Xiaolin Zheng, Jianwei Yin
Abstract:
Although Multimodal Large Language Models (MLLMs) have advanced numerous fields, their training on extensive multimodal datasets introduces significant privacy concerns, prompting the necessity for efficient unlearning methods.However, current multimodal unlearning approaches often directly adapt techniques from unimodal contexts, largely overlooking the critical issue of modality alignment, i.e., consistently removing knowledge across both unimodal and multimodal settings. To close this gap, we introduce UMU-bench, a unified benchmark specifically targeting modality misalignment in multimodal unlearning. UMU-bench consists of a meticulously curated dataset featuring 653 individual profiles, each described with both unimodal and multimodal knowledge.Additionally, novel tasks and evaluation metrics focusing on modality alignment are introduced, facilitating a comprehensive analysis of unimodal and multimodal unlearning effectiveness. Through extensive experimentation with state-of-the-art unlearning algorithms on UMU-bench, we demonstrate prevalent modality misalignment issues in existing methods. These findings underscore the critical need for novel multimodal unlearning approaches explicitly considering modality alignment.



Paperid:2112
Authors:Shengda Fan, Xin Cong, Zhong Zhang, Yuepeng Fu, Yesai Wu, Hao Wang, Xinyu Zhang, Enrui Hu, Yankai Lin
Abstract:
Recent efforts to employ large language models (LLMs) as agents have demonstrated promising results in a wide range of multi-step agent tasks. However, existing agents lack an effective experience reuse approach to leverage historical completed tasks. In this paper, we propose a novel experience reuse framework MetaFlowLLM, which constructs a hierarchical experience tree from historically completed tasks. Each node in this experience tree is presented as a MetaFlow which contains static execution workflow and subtask required by agents to complete dynamically. Then, we propose a Hierarchical MetaFlow Merging algorithm to construct the hierarchical experience tree. When accomplishing a new task, MetaFlowLLM can first retrieve the most relevant MetaFlow node from the experience tree and then execute it accordingly. To effectively generate valid MetaFlows from historical data, we further propose a reinforcement learning pipeline to train the MetaFlowGen. Extensive experimental results on AppWorld and WorkBench demonstrate that integrating with MetaFlowLLM, existing agents (e.g., ReAct, Reflexion) can gain substantial performance improvement with reducing execution costs. Notably, MetaFlowLLM achieves an average success rate improvement of 32.3% on AppWorld and 6.2% on WorkBench, respectively.



Paperid:2113
Authors:Sascha Xu, Nils Philipp Walter, Jilles Vreeken
Abstract:
Interpretable machine learning is essential in high-stakes domains like healthcare. Rule lists are apopular choice due to their transparency and accuracy, but learning them effectively remains a challenge.Existing methods require feature pre-discretization, constrain rule complexity or ordering, or struggleto scale. We present NeuRules, a novel end-to-end framework that overcomes these limitations. At itscore, NeuRules transforms the inherently combinatorial task of rule list learning into a differentiableoptimization problem, enabling gradient-based learning. It simultaneously discovers feature conditions,assembles them into conjunctive rules, and determines their order—without pre-processing or manualconstraints. A key contribution here is a gradient shaping technique that steers learning toward sparserules with strong predictive performance. To produce ordered lists, we introduce a differentiable relaxationthat, through simulated annealing, converges to a strict rule list. Extensive experiments showthat NeuRules consistently outperforms combinatorial and neural baselines on binary as well asmulti-class classification tasks across a wide range of datasets.



Paperid:2114
Authors:Luyu Chen, Zeyu Zhang, Haoran Tan, Quanyu Dai, Yang Hao, Zhenhua Dong, Xu Chen
Abstract:
LLMs have emerged as powerful evaluators in the LLM-as-a-Judge paradigm, offering significant efficiency and flexibility compared to human judgments. However, previous methods primarily rely on single-point evaluations, overlooking the inherent diversity and uncertainty in human evaluations. This approach leads to information loss and decreases the reliability of evaluations. To address this limitation, we propose a novel training framework that explicitly aligns the LLM-generated judgment distribution with empirical human distributions. Specifically, we propose a distributional alignment objective based on KL divergence, combined with an auxiliary cross-entropy regularization to stabilize the training process. Furthermore, considering that empirical distributions may derive from limited human annotations, we incorporate adversarial training to enhance model robustness against distribution perturbations. Extensive experiments across various LLM backbones and evaluation tasks demonstrate that our framework significantly outperforms existing closed-source LLMs and conventional single-point alignment methods, with improved alignment quality, evaluation accuracy, and robustness.



Paperid:2115
Authors:Kwan Ho Ryan Chan, Yuyan Ge, Edgar Dobriban, Hamed Hassani, Rene Vidal
Abstract:
A significant use case of instruction-finetuned Large Language Models (LLMs) is to solve question-answering tasks interactively. In this setting, an LLM agent is tasked with making a prediction by sequentially querying relevant information from the user, as opposed to a single-turn conversation. This paper explores sequential querying strategies that aim to minimize the expected number of queries. One such strategy is Information Pursuit (IP), a greedy algorithm that at each iteration selects the query that maximizes information gain or equivalently minimizes uncertainty. However, obtaining accurate estimates of mutual information or conditional entropy for LLMs is very difficult in practice due to over- or under-confident LLM probabilities, which leads to suboptimal query selection and predictive performance. To better estimate the uncertainty at each iteration, we proposeConformal Information Pursuit (C-IP), an alternative approach to sequential information gain based on conformal prediction sets. More specifically, C-IP leverages a relationship between prediction sets and conditional entropy at each iteration to estimate uncertainty based on the average size of conformal prediction sets. In contrast to conditional entropy, we find that conformal prediction sets are a distribution-free and robust method of measuring uncertainty. Experiments with 20 Questions show that C-IP obtains better predictive performance and shorter query-answer chains compared to previous approaches to IP and uncertainty-based chain-of-thought methods. Furthermore, extending to an interactive medical setting between a doctor and a patient on the MediQ dataset, C-IP achieves competitive performance with direct single-turn prediction while offering greater interpretability.



Paperid:2116
Authors:Zujin Guo, Size Wu, Zhongang Cai, Wei Li, Chen Change Loy
Abstract:
Existing interpolation methods use pre‑trained video diffusion priors to generate intermediate frames between sparsely sampled keyframes. In the absence of 3D geometric guidance, these methods struggle to produce plausible results for complex, articulated human motions and offer limited control over the synthesized dynamics. In this paper, we introduce PoseFuse3D Keyframe Interpolator (PoseFuse3D-KI), a novel framework that integrates 3D human guidance signals into the diffusion process for Controllable Human-centric Keyframe Interpolation (CHKI). To provide rich spatial and structural cues for interpolation, our PoseFuse3D, a 3D‑informed control model, features a novel SMPL‑X encoder that encodes and aggregates 3D geometry and shape into the 2D latent conditioning space, alongside a fusion network that integrates these 3D cues with 2D pose embeddings. For evaluation, we build CHKI-Video, a new dataset annotated with both 2D poses and 3D SMPL‑X parameters. We show that PoseFuse3D-KI consistently outperforms state-of-the-art baselines on CHKI-Video, achieving a 9\% improvement in PSNR and a 38\% reduction in LPIPS. Comprehensive ablations demonstrate that our PoseFuse3D model improves interpolation fidelity.



Paperid:2117
Authors:Ansel Blume, Jeonghwan Kim, Hyeonjeong Ha, Elen Chatikyan, Xiaomeng Jin, Khanh Nguyen, Nanyun Peng, Kai-Wei Chang, Derek Hoiem, Heng Ji
Abstract:
Real-world objects are composed of distinctive, object-specific parts. Identifying these parts is key to performing fine-grained, compositional reasoning—yet, large multimodal models (LMMs) struggle to perform this seemingly straightforward task. In this work, we introduce PARTONOMY, an LMM benchmark designed for pixel-level part grounding. We construct PARTONOMY from existing part datasets and our own rigorously annotated set of images, encompassing 862 parts and 5346objects for evaluation. Unlike existing datasets that simply ask models to identify generic parts, PARTONOMY utilizes highly technical concepts and challenges models to compare objects’ parts, consider part-whole relationships, and justify textual predictions with visual segmentations. Our experiments demonstrate significant limitations in state-of-the-art LMMs (e.g., LISA-13B achieves only 5.9% gIoU), highlighting a critical gap in their part grounding abilities. We note that existing segmentation-enabled LMMs (segmenting LMMs) have two key architectural shortcomings: they use special [SEG] tokens not seen during pretraining which induce distribution shift, and they discard predicted segmentations instead of using past predictions to guide future ones. To address these deficiencies, we train several part-centric LMMs and propose PLUM, a novel segmenting LMM that utilizes span tagging instead of segmentation tokens and that conditions on prior predictions in a feedback loop. We find that pretrained PLUM dominates existing segmenting LMMs on reasoning segmentation, VQA, and visual hallucination benchmarks. In addition, PLUM finetuned on our proposed Explanatory Part Segmentation task is competitive with segmenting LMMs trained on significantly more segmentation data. Our work opens up new avenues towards enabling fine-grained, grounded visual understanding in LMMs.



Paperid:2118
Authors:Namkyeong Lee, Yunhak Oh, Heewoong Noh, Gyoung S. Na, Minkai Xu, Hanchen Wang, Tianfan Fu, Chanyoung Park
Abstract:
Molecular Relational Learning (MRL) is a rapidly growing field that focuses on understanding the interaction dynamics between molecules, which is crucial for applications ranging from catalyst engineering to drug discovery. Despite recent progress, earlier MRL approaches are limited to using only the 2D topological structure of molecules, as obtaining the 3D interaction geometry remains prohibitively expensive.This paper introduces a novel 3D geometric pre-training strategy for MRL (3DMRL) that incorporates a 3D virtual interaction environment, overcoming the limitations of costly traditional quantum mechanical calculation methods. With the constructed 3D virtual interaction environment, 3DMRL trains 2D MRL model to learn the global and local 3D geometric information of molecular interaction.Extensive experiments on various tasks using real-world datasets, including out-of-distribution and extrapolation scenarios, demonstrate the effectiveness of 3DMRL, showing up to a 24.93% improvement in performance across 40 tasks.Our code is publicly available at https://anonymous.4open.science/r/3DMRL-5436.



Paperid:2119
Authors:Belinda Mo, Kyssen Yu, Joshua Kazdan, Proud Mpala, Lisa Yu, Charilaos Kanatsoulis, Sanmi Koyejo
Abstract:
Recent interest in building foundation models for knowledge graphs has highlighted a fundamental challenge: knowledge graph data is scarce. The best-known knowledge graphs are primarily human-labeled, created by pattern-matching, or extracted using early NLP techniques. While human-generated knowledge graphs are in short supply, automatically extracted ones are of questionable quality. We present KGGen, a novel text-to-knowledge-graph generator that uses language models to extract high-quality graphs from plain text with a novel entity resolution approach that clusters related entities, significantly reducing the sparsity problem that plagues existing extractors. Unlike other KG generators, KGGen clusters and deduplicates related entities to reduce sparsity in extracted KGs. Along with KGGen, we release Measure of Information in Nodes and Edges (MINE), the first benchmark to test an extractor's ability to produce a useful KG from plain text. We benchmark our new tool against leading existing generators such as Microsoft's GraphRAG; we achieve comparable retrieval accuracy on the generated graphs and better information retention. Moreover, our graphs exhibit more concise and generalizable entities and relations. KGGen is available as a Python library (pip install [redacted]), making it accessible to everyone.



Paperid:2120
Authors:Eleanor Spens, Neil Burgess, Tim Behrens
Abstract:
Brains reorganise knowledge offline to improve future behaviour, with 'replay' involved in consolidating memories, abstracting patterns from experience, and simulating new scenarios. However, there are few models of how the brain might orchestrate these processes, and of when different types of replay might be useful. Here we propose a framework in which a meta-controller learns to coordinate offline learning of a lower-level agent or model in 'sleep' phases to maximise reward in an 'awake' phase. The agent selects among several actions, such as learning from recent memories in a hippocampal store, abstracting patterns from memories into a 'world model', and learning from generated data. In addition, the agent learns to estimate the value of each episode, enabling the prioritisation of past events in memory replay, or of new simulations in generative replay. Using image classification, maze solving, and relational inference tasks, we show that the meta-controller learns an adaptive curriculum for offline learning. This lays the groundwork for normative predictions about replay in a range of experimental neuroscience tasks.



Paperid:2121
Authors:Frédéric Lin, Biruk Abere Ambaw, Adrian Popescu, Hejer AMMAR, Romaric Audigier, Hervé Le Borgne
Abstract:
AI systems must adapt to the evolving visual landscape, especially in domains where object appearance shifts over time. While prior work on time-aware vision models has primarily addressed commonsense-level categories, we introduce Car Models in Time (CaMiT). This fine-grained dataset captures the temporal evolution of this representative subset of technological artifacts. CaMiT includes 787K labeled samples of 190 car models (2007–2023) and 5.1M unlabeled samples (2005–2023), supporting supervised and self-supervised learning. We show that static pretraining on in-domain data achieves competitive performance with large-scale generalist models, offering a more resource-efficient solution. However, accuracy degrades when testing a year's models backward and forward in time. To address this, we evaluate CaMiT in a time-incremental classification setting, a realistic continual learning scenario with emerging, evolving, and disappearing classes. We investigate two mitigation strategies: time-incremental pretraining, which updates the backbone model, and time-incremental classifier learning, which updates the final classification layer, with positive results in both cases. Finally, we introduce time-aware image generation by consistently using temporal metadata during training. Results indicate improved realism compared to standard generation. CaMiT provides a rich resource for exploring temporal adaptation in a fine-grained visual context for discriminative and generative AI systems.



Paperid:2122
Authors:Sungho Jeon, Xinyue Ma, Kwang In Kim, Myeongjae Jeon
Abstract:
Recent rehearsal-free continual learning (CL) methods guided by prompts achieve strong performance on vision tasks with non-stationary data but remain resource-intensive, hindering real-world deployment. We introduce resource-efficient prompting (REP), which improves the computational and memory efficiency of prompt-based rehearsal-free methods while minimizing accuracy trade-offs. Our approach employs swift prompt selection to refine input data using a carefully provisioned model and introduces adaptive token merging AToM and adaptive layer dropping ALD for efficient prompt updates. AToM and ALD selectively skip data and model layers while preserving task-specific features during the learning of new tasks. Extensive experiments on multiple image classification datasets demonstrate REP’s superior resource efficiency over state-of-the-art prompt-based CL methods.



Paperid:2123
Authors:Fabian Gröger, Shuo Wen, Huyen Le, Maria Brbic
Abstract:
Abstract:Multimodal models have demonstrated powerful capabilities in complex tasks requiring multimodal alignment including zero-shot classification and cross-modal retrieval. However, existing models typically rely on millions of paired multimodal samples, which are prohibitively expensive or infeasible to obtain in many domains.In this work, we explore the feasibility of building multimodal models with limited amount of paired data by aligning pretrained unimodal foundation models. We show that high-quality alignment is possible with as few as tens of thousands of paired samples—less than $1\%$ of the data typically used in the field. To achieve this, we introduce STRUCTURE, an effective regularization technique that preserves the neighborhood geometry of the latent space of unimodal encoders. Additionally, we show that aligning last layers is often suboptimal and demonstrate the benefits of aligning the layers with the highest representational similarity across modalities. These two components can be readily incorporated into existing alignment methods, yielding consistent gains across 24 zero-shot classification and retrieval benchmarks, with average relative improvement of $51.6\%$ in classification and $91.8\%$ in retrieval tasks. Our results highlight the effectiveness and broad applicability of our framework for limited-sample multimodal learning and offer a promising path forward for resource-constrained domains.



Paperid:2124
Authors:Maojiang Su, Jerry Yao-Chieh Hu, Yi-Chen Lee, Ning Zhu, Jui-Hui Chung, Shang Wu, Zhao Song, Minshuo Chen, Han Liu
Abstract:
Abstract:Flow matching is an emerging generative modeling framework that learns continuous-time dynamics to map noise into data.To enhance expressiveness and sampling efficiency, recent works have explored incorporating high-order trajectory information. Despite the empirical success, a holistic theoretical foundation is still lacking. We present a unified framework for standard and high-order flow matching that incorporates trajectory derivatives up to an arbitrary order $K$. A key innovation is a high-order marginalization technique that converts the intractable $K$-order loss into a simple conditional regression with exact gradients.We establish sharp statistical rates of the $K$-order flow matching implemented with transformer networks. With $n$ samples, flow matching estimates nonparametric distributions at a rate $\tilde{O}(n^{-\Theta(1/d )})$, matching minimax lower bounds up to logarithmic factors.



Paperid:2125
Authors:Shuhai Zhang, ZiHao Lian, Jiahao Yang, Daiyuan Li, Guoxuan Pang, Feng Liu, Bo Han, Shutao Li, Mingkui Tan
Abstract:
AI-generated videos have achieved near-perfect visual realism (e.g., Sora), urgently necessitating reliable detection mechanisms. However, detecting such videos faces significant challenges in modeling high-dimensional spatiotemporal dynamics and identifying subtle anomalies that violate physical laws. In this paper, we propose a physics-driven AI-generated video detection paradigm based on probability flow conservation principles. Specifically, we propose a statistic called Normalized Spatiotemporal Gradient (NSG), which quantifies the ratio of spatial probability gradients to temporal density changes, explicitly capturing deviations from natural video dynamics. Leveraging pre-trained diffusion models, we develop an NSG estimator through spatial gradients approximation and motion-aware temporal modeling without complex motion decomposition while preserving physical constraints. Building on this, we propose an NSG-based video detection method (NSG-VD) that computes the Maximum Mean Discrepancy (MMD) between NSG features of the test and real videos as a detection metric. Last, we derive an upper bound of NSG feature distances between real and generated videos, proving that generated videos exhibit amplified discrepancies due to distributional shifts. Extensive experiments confirm that NSG-VD outperforms state-of-the-art baselines by 16.00\% in Recall and 10.75\% in F1-Score, validating the superior performance of NSG-VD.



Paperid:2126
Authors:Haoyang Fang, Boran Han, Nick Erickson, Xiyuan Zhang, Su Zhou, Anirudh Dagar, Jiani Zhang, Caner Turkmen, Tony Hu, Huzefa Rangwala, Ying Nian Wu, Yuyang (Bernie) Wang, George Karypis
Abstract:
Existing AutoML systems have advanced the automation of machine learning (ML); however, they still require substantial manual configuration and expert input, particularly when handling multimodal data. We introduce MLZero, a novel multi-agent framework powered by Large Language Models (LLMs) that enables end-to-end ML automation across diverse data modalities with minimal human intervention. A cognitive perception module is first employed, transforming raw multimodal inputs into perceptual context that effectively guides the subsequent workflow. To address key limitations of LLMs, such as hallucinated code generation and outdated API knowledge, we enhance the iterative code generation process with semantic and episodic memory. MLZero demonstrates superior performance on MLE-Bench Lite, outperforming all competitors in both success rate and solution quality, securing six gold medals. Furthermore, when evaluated on our Multimodal AutoML Agent Benchmark, which includes 25 more challenging tasks spanning diverse data modalities, MLZero outperforms the competing methods by a large margin with a success rate of 0.92 (+263.6\%) and an average rank of 2.28. Our approach maintains its robust effectiveness even with a compact 8B LLM, outperforming full-size systems from existing solutions.



Paperid:2127
Authors:Fei Ye, Yongcheng Zhong, Qihe Liu, Adrian G. Bors, Jingling sun, Rongyao Hu, shijie zhou
Abstract:
Plasticity and stability denote the ability to assimilate new tasks while preserving previously acquired knowledge, representing two important concepts in continual learning. Recent research addresses stability by leveraging pre-trained models to provide informative representations, yet the efficacy of these methods is highly reliant on the choice of the pre-trained backbone, which may not yield optimal plasticity. This paper addresses this limitation by introducing a streamlined and potent framework that orchestrates multiple different pre-trained backbones to derive semantically rich multi-source representations. We propose an innovative Multi-Scale Interaction and Dynamic Fusion (MSIDF) technique to process and selectively capture the most relevant parts of multi-source features through a series of learnable attention modules, thereby helping to learn better decision boundaries to boost performance. Furthermore, we introduce a novel Multi-Level Representation Optimization (MLRO) strategy to adaptively refine the representation networks, offering adaptive representations that enhance plasticity. To mitigate over-regularization issues, we propose a novel Adaptive Regularization Optimization (ARO) method to manage and optimize a switch vector that selectively governs the updating process of each representation layer, which promotes the new task learning. The proposed MLRO and ARO approaches are collectively optimized within a unified optimization framework to achieve an optimal trade-off between plasticity and stability. Our extensive experimental evaluations reveal that the proposed framework attains state-of-the-art performance.



Paperid:2128
Authors:whie jung, Semin Kim, Junee Kim, Seunghoon Hong
Abstract:
Human intelligence effortlessly interprets visual scenes along a rich spectrum of semantic dimensions. However, existing approaches for language-grounded visual concept learning are limited to a few predefined primitive axes such as color and shape, and are explored in synthetic datasets.In this work, we propose a scalable framework that adaptively identifies image-related concept axes and grounds visual concepts along these axes in real-world scenes. Leveraging a pretrained vision-language model with our simple universal prompting strategy, our framework identifies a diverse image-related axes without requiring any prior knowledge. Our universal concept encoder then adaptively binds visual features to the discovered axes without introducing additional model parameters per concept. To ground visual concepts along discovered axes, we maximize the compositional consistency of concept representations, which ensures each axis to be independently manipulated without affecting other axes.We demonstrate the effectiveness of our framework on CelebA-HQ and AFHQ datasets, achieving superior editing capabilities across diverse concepts and strong compositional generalization compared to existing visual concept learning method and text-based editing methods.



Paperid:2129
Authors:Huaying Yuan, Jian Ni, Zheng Liu, Yueze Wang, Junjie Zhou, Zhengyang Liang, Bo Zhao, Zhao Cao, Ji-Rong Wen, Zhicheng Dou
Abstract:
Accurately locating key moments within long videos is crucial for solving long video understanding (LVU) tasks. However, existing benchmarks are either severely limited in terms of video length and task diversity, or they focus solely on the end-to-end LVU performance, making them inappropriate for evaluating whether key moments can be accurately accessed. To address this challenge, we propose \textbf{MomentSeeker}, a novel benchmark for long-video moment retrieval, distinguished by the following features. First, it is created based on long and diverse videos, averaging over 1200 seconds in duration and collected from various domains, e.g., movie, anomaly, egocentric, sports, etc. Second, it covers a variety of real-world tasks, such as action recognition, object localization, and causal reasoning. Third, it incorporates rich forms of queries, including text-only queries, image-conditioned queries, and video-conditioned queries. On top of MomentSeeker, we conduct comprehensive experiments for both generation-based approaches (directly using MLLMs) and retrieval-based approaches (leveraging video retrievers). Our results reveal the significant challenges in long-video moment retrieval in terms of accuracy and efficiency, despite improvements from the latest long-video MLLMs and task-specific fine-tuning. We have publicly released MomentSeeker\footnote{https://yhy-2000.github.io/MomentSeeker/} to facilitate future research in this area.



Paperid:2130
Authors:Haoguang Lu, Jiacheng Chen, Zhenguo Yang, Aurele Gnanha, Fu Lee Wang, Qing Li, Xudong Mao
Abstract:
Recent advancements in text-guided image editing have achieved notable success by leveraging natural language prompts for fine-grained semantic control. However, certain editing semantics are challenging to specify precisely using textual descriptions alone. A practical alternative involves learning editing semantics from paired examples. Existing methods using paired images typically handle only simple semantic changes or require extensive training on large-scale datasets. In this paper, we introduce PairEdit, a novel visual editing method designed to effectively learn complex editing semantics from a limited number of image pairs or even a single image pair. We propose a guidance-based noise prediction that explicitly models semantic variations within paired images through the guidance direction term. Moreover, we introduce a content-preserving noise schedule to facilitate more effective semantic learning. We also propose optimizing distinct LoRAs to disentangle the learning of semantic variations and content. Extensive qualitative and quantitative evaluations demonstrate that PairEdit successfully learns intricate semantics while significantly improving content consistency compared to baseline methods. Code will be made publicly available.



Paperid:2131
Authors:Ziyu Xiang, Kenna Ashen, Xiaofeng Qian, Xiaoning Qian
Abstract:
Symbolic regression (SR) offers interpretable over `black-box' solutions by providing analytical expressions, enabling the identification of key functional dependencies in complex systems. In this paper, we identify two key limiting factors in existing SR methods: (a) redundant representations arising from the inability to capture expression equivalences and higher-order operand relations, which breaks permutation invariance and limits encoder effectiveness; and (b) sparse rewards caused by incomplete incorporation of constraints evaluated on full expressions (e.g. constant fitting for unit consistency and/or physical law verification) during the sampling process. To address these challenges, we propose an efficient graph-based SR method,GSR, which preserves the permutation invariance by encoding expression equivalences and higher-order structural features on an expression graph (EG), and mitigates reward sparsity via employing hybrid neural-guided Monte-Carlo tree search (hnMCTS) for complete constraint integration and improved exploration-exploitation trade-off. Experiments on synthetic and real-world scientific datasets demonstrate the efficiency and accuracy of GSR in discovering underlying expressions and adhering to physical laws, offering practical solutions for scientific discovery.



Paperid:2132
Authors:Jiaqi Li, Zhipeng Lou, Johannes Schmidt-Hieber, Wei Biao Wu
Abstract:
Abstract:Stochastic Gradient Descent (SGD) and its Ruppert–Polyak averaged variant (ASGD) lie at the heart of modern large-scale learning, yet their theoretical properties in high-dimensional settings are rarely understood. In this paper, we provide rigorous statistical guarantees for constant learning-rate SGD and ASGD in high-dimensional regimes. Our key innovation is to transfer powerful tools from high-dimensional time series to online learning. Specifically, by viewing SGD as a nonlinear autoregressive process and adapting existing coupling techniques, we prove the geometric-moment contraction of high-dimensional SGD for constant learning rates, thereby establishing asymptotic stationarity of the iterates. Building on this, we derive the $q$-th moment convergence of SGD and ASGD for any $q\ge2$ in general $\ell^s$-norms, and, in particular, the $\ell^{\infty}$-norm that is frequently adopted in high-dimensional sparse or structured models. Furthermore, we provide sharp high-probability concentration analysis which entails the probabilistic bound of high-dimensional ASGD. Beyond closing a critical gap in SGD theory, our proposed framework offers a novel toolkit for analyzing a broad class of high-dimensional learning algorithms.



Paperid:2133
Authors:Piotr Gaiński, Oussama Boussif, Andrei Rekesh, Dmytro Shevchuk, Ali Parviz, Mike Tyers, Robert Batey, Michał Koziarski
Abstract:
Template-based molecular generation offers a promising avenue for drug design by ensuring generated compounds are synthetically accessible through predefined reaction templates and building blocks. In this work, we tackle three core challenges in template-based GFlowNets: (1) minimizing synthesis cost, (2) scaling to large building block libraries, and (3) effectively utilizing small fragment sets. We proposeRecursive Cost Guidance, a backward policy framework that employs auxiliary machine learning models to approximate synthesis cost and viability. This guidance steers generation toward low-cost synthesis pathways, significantly enhancing cost-efficiency, molecular diversity, and quality, especially when paired with anExploitation Penaltythat balances the trade-off between exploration and exploitation. To enhance performance in smaller building block libraries, we develop aDynamic Librarymechanism that reuses intermediate high-reward states to construct full synthesis trees. Our approach establishes state-of-the-art results in template-based molecular generation.



Paperid:2134
Authors:Ruicheng Wang, Sicheng Xu, Yue Dong, Yu Deng, Jianfeng Xiang, Zelong Lv, Guangzhong Sun, Xin Tong, Jiaolong Yang
Abstract:
We propose MoGe-2, an advanced open-domain geometry estimation model that recovers a metric-scale 3D point map of a scene from a single image. Our method builds upon the recent monocular geometry estimation approach, MoGe, which predicts affine-invariant point maps with unknown scales. We explore effective strategies to extend MoGe for metric geometry prediction without compromising the relative geometry accuracy provided by the affine-invariant point representation. Additionally, we discover that noise and errors in real data diminish fine-grained detail in the predicted geometry. We address this by developing a data refinement approach that filters and completes real data using sharp synthetic labels, significantly enhancing the granularity of the reconstructed geometry while maintaining the overall accuracy. We train our model on a large corpus of mixed datasets and conducted comprehensive evaluations, demonstrating its superior performance in achieving accurate relative geometry, precise metric scale, and fine-grained detail recovery -- capabilities that no previous methods have simultaneously achieved.



Paperid:2135
Authors:Subhojyoti Khastagir, KISHALAY DAS, Pawan Goyal, Seung-Cheol Lee, Satadeep Bhattacharjee, niloy ganguly
Abstract:
Recent advances in generative modeling have shown significant promise in designing novel periodic crystal structures. Existing approaches typically rely on either large language models (LLMs) or equivariant denoising models, each with complementary strengths: LLMs excel at handling discrete atomic types but often struggle with continuous features such as atomic positions and lattice parameters, while denoising models are effective at modeling continuous variables but encounter difficulties in generating accurate atomic compositions. To bridge this gap, we propose \our{}, a hybrid framework that integrates an LLM with a diffusion model to leverage their complementary strengths for crystal material generation. During sampling, CrysLLMGen first employs a fine-tuned LLM to produce an intermediate representation of atom types, atomic coordinates, and lattice structure. While retaining the predicted atom types, it passes the atomic coordinates and lattice structure to a pre-trained equivariant diffusion model for refinement. Our framework outperforms state-of-the-art generative models across several benchmark tasks and datasets. Specifically, CrysLLMGen not only achieves a balanced performance in terms of structural and compositional validity but also generates 73% and 40% more stable materials compared to LLM-based and denoising-based models, respectively. Furthermore, CrysLLMGen exhibits strong conditional generation capabilities, effectively producing materials that satisfy user-defined constraints.



Paperid:2136
Authors:Amit Attia, Matan Schliserman, Uri Sherman, Tomer Koren
Abstract:
Abstract:We study population convergence guarantees of stochastic gradient descent (SGD) for smooth convex objectives in the interpolation (low noise) regime. The behavior of the last iterate of SGD in this setting---particularly with large (constant) stepsizes---has received growing attention in recent years due to implications for the training of over-parameterized models, analyzing forgetting in continual learning, and understanding the convergence of the randomized Kaczmarz method for solving linear systems. We establish that after $T$ steps of SGD on $\beta$-smooth convex loss functions with stepsize $\eta \leq 1/\beta$, the last iterate exhibits expected excess risk $\smash{ \widetilde O(1/(\eta T^{1-\beta\eta/2}) + \eta T^{\beta\eta/2} \sigma_\star^2)}$, where $\sigma_\star^2$ denotes the variance of the stochastic gradients at the optimum. In particular, for a well-tuned stepsize we obtain a near optimal $\smash{\widetilde O(1/T + \sigma_\star/\sqrt T)}$ rate for the last iterate, extending the results of Varre et al. (2021) beyond least squares regression; and when $\sigma_\star=0$ we obtain a rate of $\smash{O(1/\sqrt T)}$ with $\eta=1/\beta$, improving upon the best-known $\smash{O(T^{-1/4})}$ rate recently established by Evron et al. (2025) in the special case of realizable linear regression.



Paperid:2137
Authors:Yining Shi, Kun Jiang, Qiang Meng, Ke Wang, JiaBao Wang, Wenchao Sun, Tuopu Wen, mengmeng yang, diange yang
Abstract:
World models are critical for autonomous driving to simulate environmental dynamics and generate synthetic data.Existing methods struggle to disentangle ego-vehicle motion (perspective shifts) from scene evolvement (agent interactions), leading to suboptimal predictions. Instead, we propose to separate environmental changes from ego-motion by leveraging the scene-centric coordinate systems. In this paper, we introduce COME: a framework that integrates scene-centric forecasting Control into the Occupancy world ModEl. Specifically, COME first generates ego-irrelevant, spatially consistent future features through a scene-centric prediction branch, which are then converted into scene condition using a tailored ControlNet. These condition features are subsequently injected into the occupancy world model, enabling more accurate and controllable future occupancy predictions. Experimental results on the nuScenes-Occ3D dataset show that COME achieves consistent and significant improvements over state-of-the-art (SOTA) methods across diverse configurations, including different input sources (ground-truth, camera-based, fusion-based occupancy) and prediction horizons (3s and 8s). For example, under the same settings, COME achieves 26.3% better mIoU metric than DOME and 23.7% better mIoU metric than UniScene. These results highlight the efficacy of disentangled representation learning in enhancing spatio-temporal prediction fidelity for world models. Code will be released.



Paperid:2138
Authors:Chen Fan, Mark Schmidt, Christos Thrampoulidis
Abstract:
Different gradient-based methods for optimizing overparameterized models can all achieve zero training error yet converge to distinctly different solutions inducing different generalization properties. We provide the first complete characterization of implicit optimization bias for p-norm normalized steepest descent (NSD) and momentum steepest descent (NMD) algorithms in multi-class linear classification with cross-entropy loss. Our key theoretical contribution is proving that these algorithms converge to solutions maximizing the margin with respect to the classifier matrix's p-norm, with established convergence rates. These results encompass important special cases including Spectral Descent and Muon, which we show converge to max-margin solutions with respect to the spectral norm. A key insight of our contribution is that the analysis of general entry-wise and Schatten p-norms can be reduced to the analysis of NSD/NMD with max-norm by exploiting a natural ordering property between all p-norms relative to the max-norm and its dual sum-norm. For the specific case of descent with respect to the max-norm, we further extend our analysis to include preconditioning, showing that Adam converges to the matrix's max-norm solution. Our results demonstrate that the multi-class linear setting, which is inherently richer than the binary counterpart, provides the most transparent framework for studying implicit biases of matrix-parameter optimization algorithms.



Paperid:2139
Authors:Steven An
Abstract:
Programmatic Weak Supervision (PWS) aims to reduce the cost of constructing large high quality labeled datasets often used in training modern machine learning models. A major component of the PWS pipeline is the label model, which amalgamates predictions from multiple noisy weak supervision sources, i.e. label functions (LFs), to label datapoints. While most label models are either probabilistic or adversarial, a recently proposed label model achieves strong empirical performance without falling into either camp. That label model constructs a polytope of plausible labelings using the LF predictions and outputs the "center" of that polytope as its proposed labeling. In this paper, we attempt to theoretically study that strategy by proposing Bayesian Balsubramani-Freund (BBF), a label model that implicitly constructs a polytope of plausible labelings and selects a labeling in its interior. We show an assortment of statistical results for BBF: log-concavity of its posterior, its form of solution, consistency, and rates of convergence. Extensive experiments compare our proposed method against eleven baseline label models over eleven datasets. BBF compares favorably to other Bayesian label models and label models that don't use datapoint features -- matching or exceeding their performance on eight out of eleven datasets.



Paperid:2140
Authors:Mingxiao Li, Mang Ning, Marie-Francine Moens
Abstract:
Text-to-image generation models have made significant progress in producing high-quality images from textual descriptions, yet they continue to struggle with maintaining subject consistency across multiple images, a fundamental requirement for visual storytelling. Existing methods attempt to address this by either fine-tuning models on large-scale story visualization datasets, which is resource-intensive, or by using training-free techniques that share information across generations, which still yield limited success. In this paper, we introduce a novel training-free sampling strategy called Zigzag Sampling with Asymmetric Prompts and Visual Sharing to enhance subject consistency in visual story generation. Our approach proposes a zigzag sampling mechanism that alternates between asymmetric prompting to retain subject characteristics, while a visual sharing module transfers visual cues across generated images to %further enforce consistency. Experimental results, based on both quantitative metrics and qualitative evaluations, demonstrate that our method significantly outperforms previous approaches in generating coherent and consistent visual stories.



Paperid:2141
Authors:Weizhe Yuan, Jane Yu, Song Jiang, Karthik Padthe, Yang Li, Dong Wang, Ilia Kulikov, Kyunghyun Cho, Yuandong Tian, Jason Weston, Xian Li
Abstract:
Scaling reasoning capabilities beyond traditional domains such as math and coding is hindered by the lack of diverse and high-quality questions. To overcome this limitation, we introduce a scalable approach for generating diverse and challenging reasoning questions, accompanied by reference answers. We present NaturalReasoning, a comprehensive dataset comprising 2.8 million questions that span multiple domains, including STEM fields (e.g., Physics, Computer Science), Economics, Social Sciences, and more. We demonstrate the utility of the questions in NaturalReasoning through knowledge distillation experiments which show that NaturalReasoning can effectively elicit and transfer reasoning capabilities from a strong teacher model. Furthermore, we demonstrate that NaturalReasoning is also effective for unsupervised self-training using external reward models or self-rewarding.



Paperid:2142
Authors:Jiajun Luo, Yicheng Xiao, Jianru Xu, Yangxiu You, Rongwei Lu, Chen Tang, Jingyan Jiang, Zhi Wang
Abstract:
Abstract:Diffusion models produce realistic images and videos but require substantial computational resources, necessitating multi-accelerator parallelism for real-time deployment. However, parallel inference introduces significant communication overhead from exchanging large activations between devices, limiting efficiency and scalability. We present CompactFusion, a compression framework that significantly reduces communication while preserving generation quality. Our key observation is that diffusion activations exhibit strong temporal redundancy—adjacent steps produce highly similar activations, saturating bandwidth with near-duplicate data carrying little new information. To address this inefficiency, we seek a more compact representation that encodes only the essential information. CompactFusion achieves this via Residual Compression that transmits only compressed residuals (step-wise activation differences). Based on empirical analysis and theoretical justification, we show that it effectively removes redundant data, enabling substantial data reduction while maintaining high fidelity. We also integrate lightweight error feedback to prevent error accumulation. CompactFusion establishes a new paradigm for parallel diffusion inference, delivering lower latency and significantly higher generation quality than prior methods. On 4$\times$L20, it achieves $3.0\times$ speedup while greatly improving fidelity. It also uniquely supports communication-heavy strategies like sequence parallelism on slow networks, achieving $6.7\times$ speedup over prior overlap-based method. CompactFusion applies broadly across diffusion models and parallel settings, and integrates easily without requiring pipeline rework. Portable implementation demonstrated on xDiT is publicly available at https://anonymous.4open.science/r/CompactFusion.



Paperid:2143
Authors:Tianjun Yao, Haoxuan Li, Yongqiang Chen, Tongliang Liu, Le Song, Eric Xing, Zhiqiang Shen
Abstract:
Abstract:Graph Neural Networks (GNNs) often encounter significant performance degradation under distribution shifts between training and test data, hindering their applicability in real-world scenarios. Recent studies have proposed various methods to address the out-of-distribution (OOD) generalization challenge, with many methods in the graph domain focusing on directly identifying an invariant subgraph that is predictive of the target label. However, we argue that identifying the edges from the invariant subgraph directly is challenging and error-prone, especially when some spurious edges exhibit strong correlations with the targets. In this paper, we propose $\texttt{PrunE}$, the first pruning-based graph OOD method that eliminates spurious edges to improve OOD generalizability. By pruning spurious edges, \mine retains the invariant subgraph more comprehensively, which is critical for OOD generalization. Specifically, $\texttt{PrunE}$ employs two regularization terms to prune spurious edges: 1) _graph size constraint_ to exclude uninformative spurious edges, and 2) _$\epsilon$-probability alignment_ to further suppress the occurrence of spurious edges. Through theoretical analysis and extensive experiments, we show that $\texttt{PrunE}$ achieves superior OOD performance and outperforms previous state-of-the-art methods significantly.



Paperid:2144
Authors:Beibei Lin, Tingting Chen, Robby Tan
Abstract:
Reference-driven image completion, which restores missing regions in a target view using additional images, is particularly challenging when the target view differs significantly from the references. Existing generative methods rely solely on diffusion priors and, without geometric cues such as camera pose or depth, often produce misaligned or implausible content. We propose \textbf{GeoDiff}, a novel framework that incorporates explicit 3D structural guidance to enforce geometric consistency in the completed regions, setting it apart from prior image-only approaches. GeoDiff introduces two key ideas: conditioning the diffusion process on projected point clouds to infuse geometric information, and applying target-aware masking to guide the model toward relevant reference cues. The framework features a dual-branch diffusion architecture. One branch synthesizes the missing regions from the masked target, while the other extracts geometric features from the projected point cloud. Joint self-attention across branches ensures coherent and accurate completion. To address regions visible in references but absent in the target, we project the target view into each reference to detect occluded areas, which are then masked during training. This target-aware masking directs the model to focus on useful cues, enhancing performance in difficult scenarios. To our knowledge, GeoDiff is the first to tightly couple explicit 3D geometry with diffusion-based image completion in a unified framework. Experiments show that GeoDiff achieves a 17.1% PSNR improvement over state-of-the-art methods, significantly boosting geometric accuracy while maintaining high visual quality.



Paperid:2145
Authors:Kaito Takanami, Takashi Takahashi, Ayaka Sakata
Abstract:
Self-distillation (SD), a technique where a model improves itself using its own predictions, has attracted attention as a simple yet powerful approach in machine learning. Despite its widespread use, the mechanisms underlying its effectiveness remain unclear. In this study, we investigate the efficacy of hyperparameter-tuned multi-stage SD with a linear classifier for binary classification on noisy Gaussian mixture data. For the analysis, we employ the replica method from statistical physics. Our findings reveal that the primary driver of SD's performance improvement is denoising through hard pseudo-labels, with the most notable gains observed in moderately sized datasets. We also identify two practical heuristics to enhance SD: early stopping that limits the number of stages, which is broadly effective, and bias parameter fixing, which helps under label imbalance. To empirically validate our theoretical findings derived from our toy model, we conduct additional experiments on CIFAR-10 classification using pretrained ResNet backbone. These results provide both theoretical and practical insights, advancing our understanding and application of SD in noisy settings.



Paperid:2146
Authors:En Yu, Kangheng Lin, Liang Zhao, jisheng yin, Yana Wei, Yuang Peng, Haoran Wei, Jianjian Sun, Chunrui Han, Zheng Ge, Xiangyu Zhang, Daxin Jiang, Jingyu Wang, Wenbing Tao
Abstract:
Inspired by the success of DeepSeek-R1, we explore the potential of rule-based reinforcement learning (RL) in MLLM post-training for perception policy learning. While promising, our initial experiments reveal that incorporating a thinking process through RL does not consistently lead to performance gains across all visual perception tasks. This leads us to delve into the essential role of RL in the context of visual perception. In this work, we return to the fundamentals and explore the effects of RL on different perception tasks. We observe that the perceptual perplexity is a major factor in determining the effectiveness of RL. We also observe that reward design plays a crucial role in further approaching the upper limit of model perception. To leverage these findings, we propose Perception-R1, a scalable RL framework using GRPO during MLLM post-training. With a standard Qwen2-VL-2B-Instruct, Perception-R1 achieves +4.2% on RefCOCO+, +17.9% on PixMo-Count, +4.2% on PageOCR, and notably, 31.9% AP on COCO2017 val for the first time, establishing a strong baseline for perception policy learning.



Paperid:2147
Authors:Xiao Yu, Yan Fang, Yao Zhao, Yunchao Wei
Abstract:
Abstract:Audio-visual event parsing plays a crucial role in understanding multimodal video content, but existing methods typically rely on offline processing of entire videos with huge model sizes, limiting their real-time applicability. We introduce Online Audio-Visual Event Parsing (On-AVEP), a novel paradigm for parsing audio, visual, and audio-visual events by sequentially analyzing incoming video streams. The On-AVEP task necessitates models with two key capabilities: (1) Accurate online inference, to effectively distinguish events with unclear and limited context in online settings, and (2) Real-time efficiency, to balance high performance with computational constraints. To cultivate these, we propose the $\textbf{Pre}$dictive $\textbf{F}$uture $\textbf{M}$odeling (PreFM) framework featured by (a) predictive multimodal future modeling to infer and integrate beneficial future audio-visual cues, thereby enhancing contextual understanding and (b) modality-agnostic robust representation along with focal temporal prioritization to improve precision and generalization. Extensive experiments on the UnAV-100 and LLP datasets show PreFM significantly outperforms state-of-the-art methods by a large margin with significantly fewer parameters, offering an insightful approach for real-time multimodal video understanding.



Paperid:2148
Authors:Hanjun Luo, Shenyu Dai, Chiming Ni, Xinfeng Li, Guibin Zhang, Kun Wang, Tongliang Liu, Hanan Salam
Abstract:
Despite the rapid advancement of LLM-based agents, the reliable evaluation of their safety and security remains a significant challenge. Existing rule-based or LLM-based evaluators often miss dangers in agents' step-by-step actions, overlook subtle meanings, fail to see how small issues compound, and get confused by unclear safety or security rules. To overcome this evaluation crisis, we introduce AgentAuditor, a universal, training-free, memory-augmented reasoning framework that empowers LLM evaluators to emulate human expert evaluators. AgentAuditor constructs an experiential memory by having an LLM adaptively extract structured semantic features (e.g., scenario, risk, behavior) and generate associated chain-of-thought reasoning traces for past interactions. A multi-stage, context-aware retrieval-augmented generation process then dynamically retrieves the most relevant reasoning experiences to guide the LLM evaluator's assessment of new cases. Moreover, we developed ASSEBench, the first benchmark designed to check how well LLM-based evaluators can spot both safety risks and security threats. ASSEBench comprises 2293 meticulously annotated interaction records, covering 15 risk types across 29 application scenarios. A key feature of ASSEBench is its nuanced approach to ambiguous risk situations, employing "Strict" and "Lenient" judgment standards. Experiments demonstrate that AgentAuditor not only consistently improves the evaluation performance of LLMs across all benchmarks but also sets a new state-of-the-art in LLM-as-a-judge for agent safety and security, achieving human-level accuracy. Our work is openly accessible.



Paperid:2149
Authors:Zibo Zhou, Yue Hu, Lingkai Zhang, Zonglin Li, Siheng Chen
Abstract:
Zero-shot object navigation (ZSON) allows robots to find target objects in unfamiliar environments using natural language instructions, without relying on pre-built maps or task-specific training. Recent general-purpose models, such as large language models (LLMs) and vision-language models (VLMs), equip agents with semantic reasoning abilities to estimate target object locations in a zero-shot manner. However, these models often greedily select the next goal without maintaining a global understanding of the environment and are fundamentally limited in the spatial reasoning necessary for effective navigation. To overcome these limitations, we propose a novel 3D voxel-based belief map that estimates the target’s prior presence distribution within a voxelized 3D space. This approach enables agents to integrate semantic priors from LLMs and visual embeddings with hierarchical spatial structure, alongside real-time observations, to build a comprehensive 3D global posterior belief of the target’s location. Building on this 3D voxel map, we introduce BeliefMapNav, an efficient navigation system with two key advantages: i) grounding LLM semantic reasoning within the 3D hierarchical semantics voxel space for precise target position estimation, and ii) integrating sequential path planning to enable efficient global navigation decisions. Experiments on HM3D, MP3D, and HSSD benchmarks show that BeliefMapNav achieves state-of-the-art (SOTA) Success Rate (SR) and Success weighted by Path Length (SPL), with a notable46.4\%SPL improvement over the previous best SR method, validating its effectiveness and efficiency. We will release the code of BeliefMapNav.



Paperid:2150
Authors:Peijie Wang, Chao Yang, Zhong-Zhi Li, Fei yin, Dekang Ran, Mi Tian, Zhilong Ji, Jinfeng Bai, Cheng-lin Liu
Abstract:
Geometry is a fundamental branch of mathematics and plays a crucial role in evaluating the reasoning capabilities of multimodal large language models (MLLMs). However, existing multimodal mathematics benchmarks mainly focus on plane geometry and largely ignore solid geometry, which requires spatial reasoning and is more challenging than plane geometry. To address this critical gap, we introduceSolidGeo, the first large-scale benchmark specifically designed to evaluate the performance of MLLMs on mathematical reasoning tasks in solid geometry. SolidGeo consists of 3,113 real-world K–12 and competition-level problems, each paired with visual context and annotated with difficulty levels and fine-grained solid geometry categories. Our benchmark covers a wide range of 3D reasoning subjects such as projection, unfolding, spatial measurement, and spatial vector, offering a rigorous testbed for assessing solid geometry. Through extensive experiments, we observe that MLLMs encounter substantial challenges in solid geometry math tasks, with a considerable performance gap relative to human capabilities on SolidGeo. Moreover, we analyze the performance, inference effiency and error patterns of various models, offering insights into the solid geometric mathematical reasoning capabilities of MLLMs. We hope SolidGeo serves as a catalyst for advancing MLLMs toward deeper geometric reasoning and spatial intelligence. The dataset is released at https://huggingface.co/datasets/HarryYancy/SolidGeo/



Paperid:2151
Authors:Martin Ondrus, Ivor Cribben, Yang Feng
Abstract:
Networks have been extensively used in and provided novel insights in a wide variety of research areas. However, many real-world systems are in fact a ``network of networks'', or a multilayer network, which interact as components of a larger multimodal system. A major difficulty in this multilayer framework is the estimation of interlayer edges or connections. In this work, we propose a new estimation method, called multilayer sparse low-rank inverse covariance estimation (multiSLICE), which estimates the interlayer edges. multiSLICE bridges latent variable Gaussian graphical methods with multilayer networks, offering a flexible framework for modeling processes with irregular sampling and heterogeneous graph structures. We develop an efficient computational algorithm to compute this estimator and establish theoretical conditions for the recoverability of the joint space and how inter-layer interactions influence joint parameter estimation, and we provide theoretical bounds on their relationships. Finally, we rigorously evaluate our method on both simulated and multimodal neuroimaging data, demonstrating improvements over state-of-the-art approaches. All experiments are available in an Anonymous Github.



Paperid:2152
Authors:Xinyu Zhu, Mengzhou Xia, Zhepei Wei, Wei-Lin Chen, Danqi Chen, Yu Meng
Abstract:
Abstract:Reinforcement learning with verifiable rewards (RLVR) is a promising approach for training language models (LMs) on reasoning tasks that elicit emergent long chains of thought (CoTs).Unlike supervised learning, it updates the model using both correct and incorrect samples via policy gradients.To better understand its mechanism, we decompose the learning signal into positive and negative sample reinforcement.Surprisingly, we find that training with only negative samples---without reinforcing correct responses---can be highly effective: it consistently improves performance over the base model across the entire Pass@$k$ spectrum, often matching or surpassing PPO and GRPO.In contrast, reinforcing only correct responses improves Pass@$1$ but degrades performance at higher $k$, due to reduced diversity.These inference-scaling trends highlight the critical yet underappreciated role of negative sample reinforcement (NSR) in improving LM reasoning. Through gradient analysis, we show that it works by suppressing incorrect generations and redistributing probability mass toward other plausible candidates, guided by the model's prior beliefs.It refines the model's existing knowledge, without learning new behaviors from scratch. Building on this insight, we propose a simple variant of the RL objective that upweights NSR, and show that it consistently improves overall Pass@$k$ performance on MATH, AIME 2025, and AMC23.



Paperid:2153
Authors:Ayan Sengupta, Siddhant Chaudhary, Tanmoy Chakraborty
Abstract:
Abstract:Key-value (KV) cache compression has emerged as a critical technique for reducing the memory and latency overhead of autoregressive language models during inference. Prior approaches predominantly rely on query-key attention scores to rank and evict cached tokens, assuming that attention intensity correlates with semantic importance. However, this heuristic overlooks the contribution of value vectors, which directly influence the attention output. In this paper, we propose CurDKV, a novel, value-centric KV compression method that selects keys and values based on leverage scores computed from CUR matrix decomposition. Our approach approximates the dominant subspace of the attention output $\mathrm{softmax}(QK^\top)V$, ensuring that the retained tokens best preserve the model’s predictive behavior. Theoretically, we show that attention score approximation does not guarantee output preservation, and demonstrate that CUR-based selection minimizes end-to-end attention reconstruction loss. Empirically, CurDKV achieves up to $9.6$\% higher accuracy than state-of-the-art methods like SnapKV and ChunkKV under aggressive compression budgets on LLaMA and Mistral, while maintaining compatibility with FlashAttention and Grouped Query Attention. In addition to improved accuracy, CurDKV reduces generation latency by up to 40\% at high compression, offering a practical speed-accuracy tradeoff.



Paperid:2154
Authors:David A. R. Robin, Killian Bakong, Kevin Scaman
Abstract:
In the context of smooth stochastic optimization with first order methods, we introduce the stability ratio of gradient estimates, as a measure of local relative noise level, from zero for pure noise to one for negligible noise. We show that a schedule-free variant (Stab-SGD) of stochastic gradient descent obtained by just shrinking the learning rate by the stability ratio achieves real adaptivity to noise levels (i.e. without tuning hyperparameters to the gradient's variance), with all key properties of a good schedule-free algorithm: neither plateau nor explosion at intialization, and no saturation of the loss.We believe this theoretical development reveals the importance of estimating the local stability ratio in the construction of well-behaved (last-iterate) schedule-free algorithms, particularly when hyperparameter-tuning budgets are a small fraction of the total budget since noise-adaptivity and cheaper horizon-free tuning are most crucial in this regime.



Paperid:2155
Authors:Hyunseung Kim, Chiho Choi, Srikanth Malla, Sai Padmanabhan, Saurabh Bagchi, Joon Hee Choi
Abstract:
Text-guided image editing has been allowing users to transform and synthesize images through natural language instructions, offering considerable flexibility. However, most existing image editing models naively attempt to follow all user instructions, even if those instructions are inherently infeasible or contradictory, often resulting in nonsensical output. To address these challenges, we propose a context-aware method for image editing named as CAMILA (Context-Aware Masking for Image Editing with Language Alignment). CAMILA is designed to validate the contextual coherence between instructions and the image, ensuring that only relevant edits are applied to the designated regions while ignoring non-executable instructions. For comprehensive evaluation of this new method, we constructed datasets for both single- and multi-instruction image editing, incorporating the presence of infeasible requests. Our method achieves better performance and higher semantic alignment than state-of-the-art models, demonstrating its effectiveness in handling complex instruction challenges while preserving image integrity.



Paperid:2156
Authors:Zihao Fu, Ryan Brown, Shun Shao, Kai Rawal, Eoin Delaney, Chris Russell
Abstract:
Text-to-image diffusion models, such as Stable Diffusion, have demonstrated remarkable capabilities in generating high-quality and diverse images from natural language prompts. However, recent studies reveal that these models often replicate and amplify societal biases, particularly along demographic attributes like gender and race. In this paper, we introduce, a post-hoc debiasing framework that operates on prompt embeddings to mitigate such biases without retraining or modifying the underlying diffusion model. Our method integrates Fair Principal Component Analysis to project CLIP-based embeddings into a subspace that minimizes group-specific information while preserving semantic content. We further enhance debiasing effectiveness through empirical noise injection and propose a unified cross-demographic projection method that enables simultaneous debiasing across multiple demographic attributes. Extensive experiments across gender, race, and intersectional settings demonstrate that FairImagen significantly improves fairness with a moderate trade-off in image quality and prompt fidelity. Our framework outperforms existing post-hoc methods and offers a simple, scalable, and model-agnostic solution for equitable text-to-image generation.



Paperid:2157
Authors:Qiyi Wang, Ying Shen, Senda Chen
Abstract:
Video Temporal Grounding (VTG) aims to localize relevant segments in untrimmed videos based on natural language queries and has seen notable progress in recent years.However, most existing methods suffer from two critical limitations. First, they are prone to learning superficial co-occurrence patterns—such as associating specific objects or phrases with certain events—induced by dataset biases, which ultimately degrades their semantic understanding abilities.Second, they typically assume that relevant segments always exist in the video, an assumption misaligned with real-world scenarios where queried content may be absent.Fortunately, causal inference offers a natural solution to the above-mentioned issues by disentangling dataset-induced biases and enabling counterfactual reasoning about query relevance. To this end, we propose CausalVTG, a novel framework that explicitly integrates causal reasoning into VTG. Specifically, we introduce a causality-aware disentangled encoder (CADE) based on front-door adjustment to mitigate confounding biases in visual and textual modalities. To better capture temporal granularity, we design a multi-scale temporal perception module (MSTP) that reconstructs query-conditioned video features at multiple resolutions. Additionally, a counterfactual contrastive learning objective is employed to help the model discern whether a query is truly grounded in a video.Extensive experiments on five widely-used benchmarks demonstrate that CausalVTG outperforms state-of-the-art methods, achieving higher localization precision under stricter IoU thresholds and more accurately identifying whether a query is truly grounded in the video. These results demonstrate both the effectiveness and generalizability of proposed CausalVTG.



Paperid:2158
Authors:Shivam Agarwal, Zimin Zhang, Lifan Yuan, Jiawei Han, Hao Peng
Abstract:
Entropy minimization (EM) trains the model to concentrate even more probability mass on its most confident outputs. We show that this simple objective alone, without any labeled data, can substantially improve large language models’ (LLMs) performance on challenging math, physics, and coding tasks. We explore three approaches: (1) EM-FT minimizes token-level entropy similarly to instruction finetuning, but on unlabeled outputs drawn from the model; (2) EM-RL: reinforcement learning with negative entropy as the only reward to maximize; (3) EM-INF: inference-time logit adjustment to reduce entropy without any training data or parameter updates. On Qwen-7B, EM-RL, without any labeled data, achieves comparable or better performance than strong RL baselines such as GRPO and RLOO that are trained on 60K labeled examples. Furthermore, EM-INF enables Qwen-32B to match or exceed the performance of proprietary models like GPT-4o, Claude 3 Opus, and Gemini 1.5 Pro on the challenging SciCode benchmark, while being 3x more efficient than self-consistency and sequential refinement. Our findings reveal that many pretrained LLMs possess previously underappreciated reasoning capabilities that can be effectively elicited through entropy minimization alone, without any labeled data or even any parameter updates.



Paperid:2159
Authors:Chenlong Zhang, Zhuoran Jin, Hongbang Yuan, Jiaheng Wei, Tong Zhou, Kang Liu, Jun Zhao, Yubo Chen
Abstract:
Abstract:The widespread deployment of Large Language Models (LLMs) trained on massive, uncurated corpora has raised growing concerns about the inclusion of sensitive, copyrighted, or illegal content. This has led to increasing interest in LLM unlearning: the task of selectively removing specific information from a model without retraining from scratch or degrading overall utility.However, existing methods often rely on large-scale forget and retain datasets, and suffer from unnatural responses, poor generalization, or catastrophic utility loss.In this work, we propose $\textbf{R}$einforcement $\textbf{U}$n$\textbf{LE}$arning ($\textbf{RULE}$), an efficient framework that formulates unlearning as a refusal boundary optimization problem. RULE is trained with a small portion of forget set and synthesized boundary queries, using a verifiable reward function that encourages safe refusal on forget-related queries while preserving helpful responses on permissible inputs.We provide both theoretical and empirical evidence demonstrating the effectiveness of RULE in achieving targeted unlearning without compromising model utility. Experimental results show that, with only 12\% forget set and 8\% synthesized boundary data, RULE outperforms existing baselines by up to $17.4\%$ forget quality and $16.3\%$ naturalness response while maintaining general utility, achieving $\textit{forget-retain Pareto Optimality}$. Remarkably, we further observe that RULE improves the $\textit{naturalness}$ of model outputs, enhances training $\textit{efficiency}$, and exhibits strong $\textit{generalization ability}$, generalizing refusal behavior to semantically related but unseen queries.



Paperid:2160
Authors:Jiahao Lu, Weitao Xiong, Jiacheng Deng, Peng Li, Tianyu Huang, Zhiyang Dou, Cheng Lin, Sai-Kit Yeung, Yuan Liu
Abstract:
Monocular 3D tracking aims to capture the long-term motion of pixels in 3D space from a single monocular video and has witnessed rapid progress in recent years. However, we argue that the existing monocular 3D tracking methods still fall short in separating the camera motion from foreground dynamic motion and cannot densely track newly emerging dynamic subjects in the videos. To address these two limitations, we propose TrackingWorld, a novel pipeline for dense 3D tracking of almost all pixels within a world-centric 3D coordinate system. First, we introduce a tracking upsampler that efficiently lifts the arbitrary sparse 2D tracks into dense 2D tracks. Then, to generalize the current tracking methods to newly emerging objects, we apply the upsampler to all frames and reduce the redundancy of 2D tracks by eliminating the tracks in overlapped regions. Finally, we present an efficient optimization-based framework to back-project dense 2D tracks into world-centric 3D trajectories by estimating the camera poses and the 3D coordinates of these 2D tracks. Extensive evaluations on both synthetic and real-world datasets demonstrate that our system achieves accurate and dense 3D tracking in a world-centric coordinate frame.



Paperid:2161
Authors:Rui Liu, Yu Shen, Peng Gao, Pratap Tokekar, Ming Lin
Abstract:
Multi-modal learning has become a crucial technique for improving the performance of machine learning applications across domains such as autonomous driving, robotics, and perception systems. However, in certain scenarios, particularly in resource-constrained environments, some modalities available during training may be absent during inference. While existing frameworks effectively utilize multiple data sources during training and enable inference with reduced modalities, they are primarily designed for single-agent settings. This poses a critical limitation in dynamic environments such as connected autonomous vehicles (CAV), where incomplete data coverage can lead to decision-making blind spots. Conversely, some works explore multi-agent collaboration but without addressing missing modality at test time. To overcome these limitations, we propose Collaborative Auxiliary Modality Learning (CAML), a novel multi-modal multi-agent framework that enables agents to collaborate and share multi-modal data during training, while allowing inference with reduced modalities during testing. Experimental results in collaborative decision-making for CAV in accident-prone scenarios demonstrate that CAML achieves up to a 58.1% improvement in accident detection. Additionally, we validate CAML on real-world aerial-ground robot data for collaborative semantic segmentation, achieving up to a 10.6% improvement in mIoU.



Paperid:2162
Authors:Wentao Xu, Huiqiao Fu, Haoyu Dong, Zhehao Zhou, Chunlin Chen
Abstract:
Training Reinforcement Learning (RL) controllers in simulation offers cost-efficiency and safety advantages. However, the resultant policies often suffer significant performance degradation during real-world deployment due to the reality gap. Previous works like System Identification (Sys-Id) have attempted to bridge this discrepancy by improving simulator fidelity, but encounter challenges including the collapse of high-dimensional parameter identification, low identification accuracy, and unstable convergence dynamics. To address these challenges, we propose a novel Sys-Id framework that combines Diffusion Evolution with Adversarial Learning (DEAL) to iteratively infer physical parameters with limited real-world data, which makes the state transitions between simulation and reality as similar as possible. Specifically, our method iteratively refines physical parameters through a dual mechanism: a discriminator network evaluates the similarity of state transitions between parameterized simulations and target environment as fitness guidance, while diffusion evolution adaptively modulates noise prediction and denoising processes to optimize parameter distributions. We validate DEAL in both simulated and real-world environments. Compared to baseline methods, DEAL demonstrates state-of-the-art stability and identification accuracy in high-dimensional parameter identification tasks, and significantly enhances sim-to-real transfer performance while requiring minimal real-world data.



Paperid:2163
Authors:Dexiong Chen, Markus Krimmel, Karsten Borgwardt
Abstract:
We introduce AutoGraph, a scalable autoregressive model for attributed graph generation using decoder-only transformers. By flattening graphs into random sequences of tokens through a reversible process, AutoGraph enables modeling graphs as sequences without relying on additional node features that are expensive to compute, in contrast to diffusion-based approaches. This results in sampling complexity and sequence lengths that scale optimally linearly with the number of edges, making it scalable and efficient for large, sparse graphs. A key success factor of AutoGraph is that its sequence prefixes represent induced subgraphs, creating a direct link to sub-sentences in language modeling. Empirically, AutoGraph achieves state-of-the-art performance on synthetic and molecular benchmarks, with up to 100x faster generation and 3x faster training than leading diffusion models. It also supports substructure-conditioned generation without fine-tuning and shows promising transferability, bridging language modeling and graph generation to lay the groundwork for graph foundation models.



Paperid:2164
Authors:Zhangyun Wang, Ni Ding, Aniket Mahanti
Abstract:
Pre-trained Vision-Language Models (VLMs) exhibit significant vulnerability to imperceptible adversarial perturbations. Current advanced defense strategies typically employ adversarial prompt tuning to improve the adversarial robustness of VLMs, which struggle to simultaneously maintain generalization across both natural and adversarial examples under different benchmarks and downstream tasks. We propose a Collaborative Adversarial Prompt Tuning (CoAPT) approach from pre-trained models to target robust models. Specifically, we adopt an improved fast total variation (TV) technique to suppress and eliminate high-frequency details from images while preserving edge structures, thereby disrupting the adversarial perturbation space. Subsequently, guided by the high-level image and text representations in the latent space of the pre-trained VLMs, the corrupted natural features are restored while inheriting the superior generalization capability. Compared to existing state-of-the-art methods, CoAPT achieves a superior trade-off between robustness and generalization. Across four benchmark tasks, CoAPT improves robustness by 28.37\% (few-shot-16), 31.67\% (base-to-novel), 20.01\% (zero-shot), and 17.37\% (out-of-distribution), while maintaining an average increase of 9.83\% in natural accuracy.



Paperid:2165
Authors:Tianyi Bai, Zengjie Hu, Fupeng Sun, Qiu Jiantao, Yizhen Jiang, Guangxin He, Bohan Zeng, Conghui He, Binhang Yuan, Wentao Zhang
Abstract:
Multi-modal large language models (MLLMs) have achieved remarkable capabilities by integrating visual perception with language understanding, enabling applications such as image-grounded dialogue, visual question answering, and scientific analysis. However, most MLLMs adopt a static inference paradigm, encoding the entire image into fixed visual tokens upfront, which limits their ability to iteratively refine understanding or adapt to context during inference. This contrasts sharply with human perception, which is dynamic, selective, and feedback-driven.In this work, we introduce a novel framework for inference-time visual token scaling that enables MLLMs to perform iterative, verifier-guided reasoning over visual content. We formulate the problem as a Markov Decision Process, involving a reasoner that proposes visual actions and a verifier—trained via multi-step Direct Preference Optimization (DPO)—that evaluates these actions and determines when reasoning should terminate. To support this, we present a new dataset, VTS, comprising supervised reasoning trajectories (VTS-SFT) and preference-labeled reasoning comparisons (VTS-DPO).Our method significantly outperforms existing approaches across diverse visual reasoning benchmarks, offering not only improved accuracy but also more interpretable and grounded reasoning processes. These results demonstrate the promise of dynamic inference mechanisms for enabling fine-grained, context-aware visual reasoning in next-generation MLLMs.



Paperid:2166
Authors:Juha Harviainen, Frank Sommer, Manuel Sorge
Abstract:
Abstract:Algorithms for learning decision trees often include heuristic local-search operations such as (1) adjusting the threshold of a cut or (2) also exchanging the feature of that cut. We study minimizing the number of classification errors by performing a fixed number of a single type of these operations. Although we discover that the corresponding problems are NP-complete in general, we provide a comprehensive parameterized-complexity analysis with the aim of determining those properties of the problems that explain the hardness and those that make the problems tractable. For instance, we show that the problems remain hard for a small number $d$ of features or small domain size $D$ but the combination of both yields fixed-parameter tractability. That is, the problems are solvable in $(D + 1)^{2d} \cdot |\mathcal{I}|^{O(1)}$ time, where $|\mathcal{I}|$ is the size of the input. We also provide a proof-of-concept implementation of this algorithm and report on empirical results.



Paperid:2167
Authors:Johanna Vielhaben, Dilyara Bareeva, Jim Berend, Wojciech Samek, Nils Strodthoff
Abstract:
Measuring the alignment between representations lets us understand similarities between the feature spaces of different models, such as Vision Transformers trained under diverse paradigms. However, traditional measures for representational alignment yield only scalar values that obscure how these spaces agree in terms of learned features. To address this, we combine alignment analysis with concept discovery, allowing a fine-grained breakdown of alignment into individual concepts. This approach reveals both universal concepts across models and each representation’s internal concept structure. We introduce a new definition of concepts as non-linear manifolds, hypothesizing they better capture the geometry of the feature space. A sanity check demonstrates the advantage of this manifold-based definition over linear baselines for concept-based alignment. Finally, our alignment analysis of four different ViTs shows that increased supervision tends to reduce semantic organization in learned representations.



Paperid:2168
Authors:Hui Chen, Miao Xiong, Yujie Lu, Wei Han, Ailin Deng, Yufei He, Jiaying Wu, Yibo Li, Yue Liu, Bryan Hooi
Abstract:
Recent advancements in AI agents have demonstrated their growing potential to drive and support scientific discovery. In this work, we introduce \textbf{MLR-Bench}, a comprehensive benchmark for evaluating AI agents on open-ended machine learning research. MLR-Bench includes three key components: (1) 201 research tasks sourced from NeurIPS, ICLR, and ICML workshops covering diverse ML topics; (2) \textbf{MLR-Judge}, an automated evaluation framework combining LLM-based reviewers with carefully designed review rubrics to assess research quality; and (3) \textbf{MLR-Agent}, a modular agent scaffold capable of completing research tasks through four stages: idea generation, proposal formulation, experimentation, and paper writing. Our framework supports both \emph{stepwise} assessment across these distinct research stages, and \emph{end-to-end} evaluation of the final research paper. We then use MLR-Bench to evaluate six frontier LLMs and an advanced coding agent, finding that while LLMs are effective at generating coherent ideas and well-structured papers, current coding agents frequently (e.g., in 80\% of the cases) produce fabricated or invalidated experimental results—posing a major barrier to scientific reliability. We validate MLR-Judge through human evaluation, showing high agreement with expert reviewers, supporting its potential as a scalable tool for research evaluation. We open-source MLR-Bench to help the community benchmark, diagnose, and improve AI research agents toward trustworthy and transparent scientific discovery.



Paperid:2169
Authors:Zahra Rahimi Afzal, Tara Esmaeilbeig, Mojtaba Soltanalian, Mesrob I Ohannessian
Abstract:
Abstract:Parameter-Efficient Fine-Tuning (PEFT) is a popular class of techniques that strive to adapt large models in a scalable and resource-efficient manner. Yet, the mechanisms underlying their training performance and generalization remain underexplored. In this paper, we provide several insights into such fine-tuning, through the lens of linearization. Fine-tuned models are often implicitly encouraged to remain close to the pretrained model. By making this explicit, using an $l_2$-distance inductive bias in parameter space, we show that fine-tuning dynamics become equivalent to learning with the positive-definite neural tangent kernel (NTK). We specifically analyze how close the fully linear and the linearized fine-tuning optimizations are, based on the strength of the regularization. This allows us to be pragmatic about how good a model linearization is when fine-tuning large language models (LLMs). When linearization is a good model, our findings reveal a strong correlation between the eigenvalue spectrum of the NTK and the performance of model adaptation. Motivated by this, we give spectral perturbation bounds on the NTK induced by the choice of layers selected for fine-tuning. We empirically validate our theory on Low Rank Adaptation (LoRA) on LLMs. These insights not only characterize fine-tuning, but also have the potential to enhance PEFT techniques, paving the way to better informed and more nimble adaptation in LLMs.



Paperid:2170
Authors:HaiYang Li, Liao Yu, Qiang Yu, Yunliang Zang
Abstract:
Biological circuits have evolved to incorporate various modules that perform similar functions. In the fly olfactory circuit, both lateral inhibition and neuronal spike frequency adaptation are thought to enhance pattern separation for odor learning. However, it remains unclear whether these mechanisms serve redundant or distinct roles in odor learning. This study presents a model of the fly olfactory circuit to investigate odor identification under varying noisy conditions that simulate complex environments. Our results show that lateral inhibition enhances odor discrimination primarily in low-noise scenarios, while spike frequency adaptation becomes essential in noisier conditions. Both mechanisms lead to inconsistent changes in intra-class similarity and inter-class dissimilarity by shaping the representation of noisy odors. This work illustrates that seemingly redundant modules in biological circuits may be crucial for optimal learning in complex environments.



Paperid:2171
Authors:Zhizhong Li, Sina Sajadmanesh, Jingtao Li, Lingjuan Lyu
Abstract:
Abstract:Low-rank adaptation (LoRA) has been widely adopted as a parameter-efficient technique for fine-tuning large-scale pre-trained models. However, it still lags behind full fine-tuning in performance, partly due to its insufficient exploitation of the geometric structure underlying low-rank manifolds. In this paper, we introduce a geometry-aware extension of LoRA that uses a three-factor decomposition $USV^\top$, separating the adapter's input and output subspaces $V$ and $U$ from the scaling component $S$, in the spirit of singular value decomposition (SVD). Our method constrains $U$ and $V$ to lie on the Stiefel manifold, ensuring their orthonormality throughout training. To optimize on the Stiefel manifold, we employ a flexible geometric optimization design that converts any Euclidean optimizer to a Riemannian optimizer via a modular interface. This enables principled and stable subspace learning while remaining compatible with existing fine-tuning pipelines. Empirical results across a wide range of downstream tasks, including commonsense reasoning, math and code generation, image classification, and image generation, demonstrate the superior performance of our approach against the vanilla LoRA and recent state-of-the-art variants.



Paperid:2172
Authors:Francesco Bacchiocchi, Matteo Castiglioni, Roberto Colomboni, Alberto Marchesi
Abstract:
Abstract:Online learning algorithms for designing optimal bilateral trade mechanisms have recently received significant attention. This paper addresses a key inefficiency in prior two-bit feedback models, which synchronously query both the buyer and the seller for their willingness to trade. This approach is inherently inefficient as it offers a trade to the seller even if the buyer rejects the offer. We propose an asynchronous mechanism that queries the seller only if the buyer has already accepted the offer. Consequently, the mechanism receives one bit of feedback from the buyer and a "censored" bit from the seller---a signal richer than the standard one-bit (trade/no-trade) feedback, but less informative than the two-bit model. Assuming independent valuations with bounded densities---the same distributional conditions underlying the two-bit results of Cesa-Bianchi et al. [2024a]---we design an algorithm that achieves $\tilde{O}(T^{2/3})$ regret against the best fixed price in hindsight. This matches the lower bound for the strictly richer two-bit model, showing that our mechanism elicits the minimal feedback necessary to attain optimal rates.



Paperid:2173
Authors:Di He, Ajay Jaiswal, Songjun Tu, Li Shen, Ganzhao Yuan, Shiwei Liu, Lu Yin
Abstract:
Weight decay is a standard regularization technique for training large language models (LLMs). While it is common to assign a uniform decay rate to every layer, this approach overlooks the structural diversity of LLMs and the varying spectral properties across modules. In this paper, we introduce AlphaDecay, a simple yet effective method that adaptively assigns different weight decay strengths to each module of an LLM. Our approach is guided by Heavy-Tailed Self-Regularization (HT-SR) theory, which analyzes the empirical spectral density (ESD) of weight correlation matrices to quantify “heavy-tailedness.” Modules exhibiting more pronounced heavy-tailed ESDs, reflecting stronger feature learning, are assigned weaker decay, while modules with lighter-tailed spectra receive stronger decay. Our method leverages tailored weight decay assignments to balance the module-wise differences in spectral properties, leading to improved performance. Extensive pre-training tasks with various model sizes from 60M to 1B demonstrate that {\method} achieves better perplexity and generalization than conventional uniform decay and other adaptive decay baselines.



Paperid:2174
Authors:Hainan Fang, Yuanbo Wen, Jun Bi, Yihan Wang, Tonghui He, Yanlin Tang, Di Huang, Jiaming Guo, Rui Zhang, Qi Guo, Yunji Chen
Abstract:
Compilers, while essential, are notoriously complex systems that demand prohibitively expensive human expertise to develop and maintain. The recent advancements in Large Language Models (LLMs) offer a compelling new paradigm: Neural Compilation, which could potentially simplify compiler development for new architectures and facilitate the discovery of innovative optimization techniques. However, several critical obstacles impede its practical adoption. Firstly, a significant lack of dedicated benchmarks and robust evaluation methodologies hinders objective assessment and tracking of progress in the field. Secondly, systematically enhancing the reliability and performance of LLM-generated assembly remains a critical challenge. Addressing these challenges, this paper introduces NeuComBack, a novel benchmark dataset specifically designed for IR-to-assembly compilation. Leveraging this dataset, we first define a foundational Neural Compilation workflow and conduct a comprehensive evaluation of the capabilities of recent frontier LLMs on Neural Compilation, establishing new performance baselines. We further propose a self-evolving prompt optimization method that enables LLMs to iteratively evolve their internal prompt strategies by extracting insights from prior self-debugging traces, thereby enhancing their neural compilation capabilities.Experiments demonstrate that our method significantly improves both the functional correctness and the performance of LLM-generated assembly code. Compared to baseline prompts, the functional correctness rates improved from 44% to 64% on x8664 and from 36% to 58% on aarch64, respectively. More significantly, among the 16 correctly generated x8664 programs using our method, 14 (87.5%) surpassed clang-O3 performance. These consistent improvements across diverse architectures (x86_64 and aarch64) and program distributions (NeuComBack L1 and L2) validate our method's superiority over conventional approaches and its potential for broader adoption in low-level neural compilation.



Paperid:2175
Authors:Qi Kuang, Jiayi Wang, Fan Zhou, Zhengling Qi
Abstract:
Abstract:We consider off-policy evaluation (OPE) in Partially Observable Markov Decision Processes (POMDPs) with unobserved confounding. Recent advances have introduced bridge-function to circumvent unmeasured confounding and develop estimators for the policy value, yet the statistical error bounds of them related to the length of horizon $T$ and the size of the state-action space $|\mathcal{O}||\mathcal{A}|$ remain largely unexplored. In this paper, we systematically investigate finite-sample error bounds of OPE estimators in finite-horizon tabular confounded POMDPs. Specifically, we show that under certain rank conditions, the estimation error for policy value can achieve a rate of $\mathcal{O}(T^{1.5}/\sqrt{n})$, excluding the cardinality of observation space $|\mathcal{O}|$ and action space $|\mathcal{A}|$. With an additional and mild condition on the concentrability coefficients in confounded POMDPs, the rate of estimation error can be improved to $\mathcal{O}(T/\sqrt{n})$. We also show that for fully history-dependent policy, the estimation error scales as $\mathcal{O}\big(T/\sqrt{n}(|\mathcal{O}| |\mathcal{A}|)^{\frac{T}{2}}\big)$, highlighting the exponential error dependence introduced by history-based proxies to infer hidden states. Furthermore, when the target policy is memoryless policy, the error bound improves to $\mathcal{O}\big(T/\sqrt{n}\sqrt{|\mathcal{O}| |\mathcal{A}|}\big)$, which matches the optimal rate known for tabular MDPs. To the best of our knowledge, this is the first work to provide a comprehensive finite-sample analysis of OPE in confounded POMDPs.



Paperid:2176
Authors:Jinwoo Park, Seunggeun Cho, Dongsu Han
Abstract:
Large language models (LLMs) power many modern applications, but serving them at scale remains costly and resource-intensive. Current server-centric systems overlook consumer-grade GPUs at the edge. We introduce SpecEdge, an edge-assisted inference framework that splits LLM workloads between edge and server GPUs using a speculative decoding scheme, exchanging only token outputs over the network. SpecEdge employs proactive edge drafting to overlap edge token creation with server verification and pipeline-aware scheduling that interleaves multiple user requests to increase server-side throughput. Experiments show SpecEdge enhances overall cost efficiency by1.91×through achieving2.22xserver throughput, and reduces inter token latency by11.24\%compared to a server-only baseline, introducing a scalable, cost-effective paradigm for LLM serving.



Paperid:2177
Authors:Yehjin Shin, Jeongwhan Choi, Seojin Kim, Noseong Park
Abstract:
Recently, convolutional filters have been increasingly adopted in sequential recommendation for their ability to capture local sequential patterns. However, most of these models complement convolutional filters with self-attention. This is because convolutional filters alone, generally fixed filters, struggle to capture global interactions necessary for accurate recommendation. We propose \textbf{T}ime-\textbf{V}ariant Convolutional Filters for Sequential \textbf{Rec}ommendation (TV-Rec), a model inspired by graph signal processing, where time-variant graph filters dynamically adapt to temporal positions in user sequences. By replacing both fixed kernels and self-attention with time-variant filters, TV-Rec achieves higher expressive power and better captures complex interaction patterns in user behavior. This design not only eliminates the need for self-attention but also reduces computation while accelerating inference. Extensive experiments on six public benchmarks show that TV-Rec outperforms ten state-of-the-art baselines by an average of 7.49\%.



Paperid:2178
Authors:Ori Lifschitz, Tali Treibitz, Dan Rosenbaum
Abstract:
We address the problem of looking into the water from the air, where we seek to remove image distortions caused by refractions at the water surface. Our approach is based on modeling the different water surface structures at various points in time, assuming the underlying image is constant. To this end, we propose a model that consists of two neural-field networks. The first network predicts the height of the water surface at each spatial position and time, and the second network predicts the image color at each position. Using both networks, we reconstruct the observed sequence of images and can therefore use unsupervised training. We show that using implicit neural representations with periodic activation functions (SIREN) leads to effective modeling of the surface height spatio-temporal signal and its derivative, as required for image reconstruction. Using both simulated and real data we show that our method outperforms the latest unsupervised image restoration approach. In addition, it provides an estimate of the water surface.



Paperid:2179
Authors:Sarit Khirirat, Abdurakhmon Sadiev, Artem Riabinin, Eduard Gorbunov, Peter Richtarik
Abstract:
Abstract:We provide the first proof of convergence for normalized error feedback algorithms across a wide range of machine learning problems. Despite their popularity and efficiency in training deep neural networks, traditional analyses of error feedback algorithms rely on the smoothness assumption that does not capture the properties of objective functions in these problems. Rather, these problems have recently been shown to satisfy generalized smoothness assumptions, and the theoretical understanding of error feedback algorithms under these assumptions remains largely unexplored. Moreover, to the best of our knowledge, all existing analyses under generalized smoothness either i) focus on single-node settings or ii) make unrealistically strong assumptions for distributed settings, such as requiring data heterogeneity, and almost surely bounded stochastic gradient noise variance. In this paper, we propose distributed error feedback algorithms that utilize normalization to achieve the $\mathcal{O}(1/\sqrt{K})$ convergence rate for nonconvex problems under generalized smoothness. Our analyses apply for distributed settings without data heterogeneity conditions, and enable stepsize tuning that is independent of problem parameters. Additionally, we provide strong convergence guarantees of normalized error feedback algorithms for stochastic settings. Finally, we show that due to their larger allowable stepsizes, our new normalized error feedback algorithms outperform their non-normalized counterparts on various tasks, including the minimization of polynomial functions, logistic regression, and ResNet-20 training.



Paperid:2180
Authors:Zhaoyu Chen, HaiJing Guo, Kaixun Jiang, Jiyuan Fu, Xinyu Zhou, Dingkang Yang, Hao Tang, Bo Li, Wenqiang Zhang
Abstract:
Deep neural networks are vulnerable to adversarial examples that exhibit transferability across various models. Numerous approaches are proposed to enhance the transferability of adversarial examples, including advanced optimization, data augmentation, and model modifications. However, these methods still show limited transferability, partiovovocularly in cross-architecture scenarios, such as from CNN to ViT. To achieve high transferability, we propose a technique termed Spatial Adversarial Alignment (SAA), which employs an alignment loss and leverages a witness model to fine-tune the surrogate model. Specifically, SAA consists of two key parts: spatial-aware alignment and adversarial-aware alignment. First, we minimize the divergences of features between the two models in both global and local regions, facilitating spatial alignment. Second, we introduce a self-adversarial strategy that leverages adversarial examples to impose further constraints, aligning features from an adversarial perspective. Through this alignment, the surrogate model is trained to concentrate on the common features extracted by the witness model. This facilitates adversarial attacks on these shared features, thereby yielding perturbations that exhibit enhanced transferability. Extensive experiments on various architectures on ImageNet show that aligned surrogate models based on SAA can provide higher transferable adversarial examples, especially in cross-architecture attacks. Our source code is available at Supplementary Materials.



Paperid:2181
Authors:Yao Huang, Huanran Chen, Shouwei Ruan, Yichi Zhang, Xingxing Wei, Yinpeng Dong
Abstract:
Recent advances in Large Reasoning Models (LRMs) have demonstrated remarkable capabilities in solving complex tasks such as mathematics and coding. However, these models frequently exhibit a phenomenon known as overthinking during inference, characterized by excessive validation loops and redundant deliberation, leading to substantial computational overheads. In this paper, we aim to mitigate overthinking by investigating the underlying mechanisms from the perspective of mechanistic interpretability. We first showcase that the tendency of overthinking can be effectively captured by a single direction in the model’s activation space and the issue can be eased by intervening the activations along this direction. However, this efficacy soon reaches a plateau and even deteriorates as the intervention strength increases. We therefore systematically explore the activation space and find that the overthinking phenomenon is actually tied to a low-dimensional manifold, which indicates that the limited effect stems from the noises introduced by the high-dimensional steering direction. Based on this insight, we propose Manifold Steering, a novel approach that elegantly projects the steering direction onto the low-dimensional activation manifold given the theoretical approximation of the interference noise. Extensive experiments on DeepSeek-R1 distilled models validate that our method reduces output tokens by up to 71\% while maintaining and even improving the accuracy on several mathematical benchmarks. Our method also exhibits robust cross-domain transferability, delivering consistent token reduction performance in code generation and knowledge-based QA tasks.



Paperid:2182
Authors:Chunyuan Deng, Ruidi Chang, Hanjie Chen
Abstract:
Recent advancements in language models (LMs) gradually ushered in an era where post-training is crucial. Yet, post-training approaches such as supervised fine-tuning (SFT) do not guarantee effective use of knowledge acquired during pretraining. We therefore present InfoSteer, a lightweight method that encourages parametric information utilization in LMs during post-training. This is achieved via treating FFN layer as associate key-value memory, and promotes the use of stored memory vectors via forward-pass interventions or regularization during backpropagation. We find this simple guidance during post-training phase delivers consistent performance improvements across diverse model families--including Qwen, Gemma and Llama-spanning over 15 downstream tasks in both ID and OOD evaluations. Beyond performance gains, we also find that steered LMs can adaptively allocate information—placing more emphasis on generating semantically meaningful tokens, while using fewer resources on simple transition ones (e.g., , or and). Our work underscores that vanilla post-training does not fully leverage pre-training potential, and steering LMs in latent representation space offers a promising approach that enhances both performance and interpretability.



Paperid:2183
Authors:Yifan Shen, Yuanzhe Liu, Jingyuan Zhu, Xu Cao, Xiaofeng Zhang, Yixiao He, Wenming Ye, James Rehg, Ismini Lourentzou
Abstract:
Current Vision-Language Models (VLMs) struggle with fine-grained spatial reasoning, particularly when multi-step logic and precise spatial alignment are required. In this work, we introduce SpatialReasoner, a novel VLM designed to address these limitations. First, we propose Multi-LLM Guided Monte Carlo Tree Search (M3CTS) and Fine-Grained Spatial Rewards methods to construct a high-quality dataset. Second, we use fine-grained Direct Preference Optimization (fDPO) to train our model. fDPO introduces segment-specific preference granularity for descriptive grounding and logical reasoning, achieving an average improvement of 4.1% over standard DPO across spatial quality tasks, and a 9.0% boost in spatial quantity tasks. To address the scarcity of multi-step spatial reasoning data, M3CTS enables collaborative exploration of diverse reasoning paths, significantly enriching spatial comprehension and logical coherence. Empirical evaluations demonstrate that SpatialReasoner sets a new state-of-the-art on SpatialRGPT-Bench, outperforming the strongest baseline by 9.8% in average accuracy, while maintaining competitive performance on general vision-language tasks.



Paperid:2184
Authors:Yijia Dai, Zhaolin Gao, Yahya Sattar, Sarah Dean, Jennifer Sun
Abstract:
Hidden Markov Models (HMMs) are fundamental tools for modeling sequential data with latent states that follow Markovian dynamics. However, they present significant challenges in model fitting and computational efficiency on real-world datasets. In this work, we demonstrate that pre-trained large language models (LLMs) can effectively model data generated by HMMs through in-context learning (ICL) — their ability to learn patterns from examples within the input context.We evaluate LLMs' performance on diverse synthetic HMMs, showing that their prediction accuracy converges to the theoretical optimum. We discover novel scaling trends influenced by HMM properties and provide theoretical conjectures for these empirical observations. Furthermore, we present practical guidelines for scientists on using ICL as a diagnostic tool for complex data. Applied to real-world animal decision-making tasks, ICL achieves competitive performance with models designed by human experts. Our results demonstrate potential for advancing understanding of LLMs' capabilities while opening new avenues for scientific discovery of biological mechanisms and hidden structures in real-world phenomena.



Paperid:2185
Authors:Xiaolong Wang, Lixiang Ru, Ziyuan Huang, Kaixiang Ji, DanDan Zheng, Jingdong Chen, Jun Zhou
Abstract:
We propose a novel AutoRegressive Generation-based paradigm for image Segmentation (ARGenSeg), achieving multimodal understanding and pixel-level perception within a unified framework. Prior works integrating image segmentation into multimodal large language models (MLLMs) typically employ either boundary points representation or dedicated segmentation heads.These methods rely on discrete representations or semantic prompts fed into task-specific decoders, which limits the ability of the MLLM to capture fine-grained visual details.To address these challenges, we introduce a segmentation framework for MLLM based on image generation, which naturally produces dense masks for target objects.We leverage MLLM to output visual tokens and detokenize them into images using an universal VQ-VAE, making the segmentation fully dependent on the pixel-level understanding of the MLLM.To reduce inference latency, we employ a next-scale-prediction strategy to generate required visual tokens in parallel.Extensive experiments demonstrate that our method surpasses prior state-of-the-art approaches on multiple segmentation datasets with a remarkable boost in inference speed, while maintaining strong understanding capabilities.



Paperid:2186
Authors:Weijia Shi, Xiaochuang Han, Chunting Zhou, Weixin Liang, Victoria Lin, Luke Zettlemoyer, LILI YU
Abstract:
We present LMFusion, a framework for empowering pretrained text-only large language models (LLMs) with multimodal generative capabilities, enabling them to understand and generate both text and images in arbitrary sequences. LMFusion leverages existing Llama-3's weights for processing texts autoregressively while introducing additional and parallel transformer modules for processing images with diffusion. During training, the data from each modality is routed to its dedicated modules: modality-specific feedforward layers, query-key-value projections, and normalization layers process each modality independently, while the shared self-attention layers allow interactions across text and image features. By freezing the text-specific modules and only training the image-specific modules, LMFusion preserves the language capabilities of text-only LLMs while developing strong visual understanding and generation abilities. Compared to methods that pretrain multimodal generative models from scratch, our experiments demonstrate that, LMFusion improves image understanding by 20% and image generation by 3.6% using only 50% of the FLOPs while maintaining Llama-3's language capabilities. We also demonstrate that this framework can adapt existing vision-language models with multimodal generation ability. Overall, this framework not only leverages existing computational investments in text-only LLMs but also enables the parallel development of language and vision capabilities, presenting a promising direction for efficient multimodal model development.



Paperid:2187
Authors:HAORAN YANG, Chuan-Xian Ren
Abstract:
While data-driven methods such as neural operator have achieved great success in solving differential equations (DEs), they suffer from domain shift problems caused by different learning environments (with data bias or equation changes), which can be alleviated by transfer learning (TL). However, existing TL methods adopted in DEs problems lack either generalizability in general DEs problems or physics preservation during training. In this work, we focus on a general transfer learning method that adaptively correct the domain shift and preserve physical relation within the equation. Mathematically, we characterize the data domain as product distribution and the essential problems as distribution bias and operator bias. A Physics-preserved Optimal Tensor Transport (POTT) method that simultaneously admits generalizability to common DEs and physics preservation of specific problem is proposed to adapt the data-driven model to target domain, utilizing the pushforward distribution induced by the POTT map. Extensive experiments in simulation and real-world datasets demonstrate the superior performance, generalizability and physics preservation of the proposed POTT method.



Paperid:2188
Authors:Thomas Walton, Darin Tsui, Aryan Musharaf, Amirali Aghazadeh
Abstract:
Autoregressive models have transformed protein engineering by enabling the generation of novel protein sequences beyond those found in nature. However, their sequential inference introduces significant latency, limiting their utility in high-throughput protein screening. Speculative decoding accelerates generation by employing a lightweight draft model to sample tokens, which a larger target model then verifies and refines. Yet in protein sequence generation, draft models are typically agnostic to the structural and functional constraints of the target protein, leading to biologically implausible outputs and a shift in the likelihood distribution of generated sequences. We introduce SpecMER (Speculative Decoding via k-mer Guidance), a novel framework that incorporates biological, structural, and functional priors using k-mer motifs extracted from multiple sequence alignments. By scoring candidate sequences in parallel and selecting those most consistent with known biological patterns, SpecMER significantly improves sequence plausibility while retaining the efficiency of speculative decoding. SpecMER achieves 24–32% speedups over standard autoregressive decoding, along with higher acceptance rates and improved sequence likelihoods.



Paperid:2189
Authors:Wenchao Ma, Dario Kneubuehler, Maurice Chu, Ian Sachs, Haomiao Jiang, Sharon Huang
Abstract:
In this paper, we present RigAnyFace (RAF), a scalable neural auto-rigging framework for facial meshes of diverse topologies, including those with multiple disconnected components. RAF deforms a static neutral facial mesh into industry-standard FACS poses to form an expressive blendshape rig. Deformations are predicted by a triangulation-agnostic surface learning network augmented with our tailored architecture design to condition on FACS parameters and efficiently process disconnected components. For training, we curated a dataset of facial meshes, with a subset meticulously rigged by professional artists to serve as accurate 3D ground truth for deformation supervision. Due to the high cost of manual rigging, this subset is limited in size, constraining the generalization ability of models trained exclusively on it. To address this, we design a 2D supervision strategy for unlabeled neutral meshes without rigs. This strategy increases data diversity and allows for scaled training, thereby enhancing the generalization ability of models trained on this augmented data. Extensive experiments demonstrate that RAF is able to rig meshes of diverse topologies on not only our artist-crafted assets but also in-the-wild samples, outperforming previous works in accuracy and generalizability. Moreover, our method advances beyond prior work by supporting multiple disconnected components, such as eyeballs, for more detailed expression animation.



Paperid:2190
Authors:Junkang Wu, Kexin Huang, xue wang, Jinyang Gao, Bolin Ding, Jiancan Wu, Xiangnan He, Xiang Wang
Abstract:
Abstract:Preference learning has become a common approach in various recent methods for aligning large language models with human values. These methods optimize the preference margin between chosen and rejected responses, subject to certain constraints for avoiding over-optimization. In this paper, we report surprising empirical findings that simple ReLU activation can learn meaningful alignments even using \emph{none} of the following: (i) sigmoid-based gradient constraints, (ii) explicit regularization terms. Our experiments show that over-optimization does exist, but a threshold parameter $\gamma$ plays an essential role in preventing it by dynamically filtering training examples. We further provide theoretical analysis demonstrating that ReLU-based Preference Optimization (RePO) corresponds to the convex envelope of the 0-1 loss, establishing its fundamental soundness. Our ``RePO'' method achieves competitive or superior results compared to established preference optimization approaches. We hope this simple baseline will motivate researchers to rethink the fundamental mechanisms behind preference optimization for language model alignment.



Paperid:2191
Authors:Milad Afshari, Ramin Akbari, Vishnu Boddeti
Abstract:
Concept erasure aims to remove unwanted attributes, such as social or demographic factors, from learned representations, while preserving their task-relevant utility. Existing methods are vulnerable to nonlinear adversaries because they fail to capture the nonlinear statistical dependencies between learned representations and undesired attributes. Moreover, the progressive degradation of utility throughout the erasure process, i.e., the cost of erasure, remains underexplored. In this work, we introduce Obliviator, which captures nonlinear statistical dependencies between undesired attributes and learned representations. We propose a multi-step framework that gradually morphs the feature space to facilitate the adversarial optimization and enable smooth concept removal. Unlike prior methods, Obliviator guards against nonlinear adversaries while achieving state-of-the-art performance that generalizes well across representations from recent language models. Our method quantifies the cost of nonlinear concept erasure and, across all levels of sensitive attribute protection, achieves higher task performance, revealing an empirical upper bound on this trade-off.



Paperid:2192
Authors:Ahmed Heakl, Yahia Shaaban, Salem Lahlou, Martin Takac, Zangir Iklassov
Abstract:
Robust routing under uncertainty is central to real-world logistics, yet most benchmarks assume static, idealized settings. We present \texttt{SVRPBench}, the first open benchmark to capture high-fidelity stochastic dynamics in vehicle routing at urban scale. Spanning more than 500 instances with up to 1000 customers, it simulates realistic delivery conditions: time-dependent congestion, log-normal delays, probabilistic accidents, and empirically grounded time windows for residential and commercial clients. Our pipeline generates diverse, constraint-rich scenarios, including multi-depot and multi-vehicle setups. Benchmarking reveals that state-of-the-art RL solvers like POMO and AM degrade by over 20\% under distributional shift, while classical and metaheuristic methods remain robust. To enable reproducible research, we release the dataset (Huggingface) and evaluation suite (Github). SVRPBench challenges the community to design solvers that generalize beyond synthetic assumptions and adapt to real-world uncertainty.



Paperid:2193
Authors:Song Lai, Changyi Ma, Fei Zhu, Zhe Zhao, Xi Lin, GAOFENG MENG, Qingfu Zhang
Abstract:
Abstract:Online Continual Learning (OCL) requires models to learn sequentially from data streams with limited memory. Rehearsal-based methods, particularly Experience Replay (ER), are commonly used in OCL scenarios. This paper revisits ER through the lens of $\epsilon$-constraint optimization, revealing that ER implicitly employs a soft constraint on past task performance, with its weighting parameter post-hoc defining a slack variable. While effective, ER's implicit and fixed slack strategy has limitations: it can inadvertently lead to updates that negatively impact generalization, and its fixed trade-off between plasticity and stability may not optimally balance current streaming with memory retention, potentially overfitting to the memory buffer. To address these shortcomings, we propose the \textbf{G}radient-Guided \textbf{E}psilon \textbf{C}onstraint (\textbf{GEC}) method for online continual learning. GEC explicitly formulates the OCL update as an $\epsilon$-constraint optimization problem, which minimize the loss on the current task data and transform the stability objective as constraints and propose a gradient-guided method to dynamically adjusts the update direction based on whether the performance on memory samples violates a predefined slack tolerance $\bar{\varepsilon}$: if forgetting exceeds this tolerance, GEC prioritizes constraint satisfaction; otherwise, it focuses on the current task while controlling the rate of increase in memory loss. Empirical evaluations on standard OCL benchmarks would demonstrate GEC's ability to achieve a superior trade-off, leading to improved overall performance.



Paperid:2194
Authors:A Saif, Lisha Chen, Xiaodong Cui, Songtao Lu, Brian Kingsbury, Tianyi Chen
Abstract:
The need for training multilingual multi-task speech processing (MSP) models that perform both automatic speech recognition and speech-to-text translation is increasingly evident. However, a significant challenge arises from the conflicts among multiple objectives when using a single model. Multi-objective optimization can address this challenge by facilitating the optimization of multiple conflicting objectives and aligning the gradient updates in a common descent direction. While multi-objective optimization helps avoid conflicting gradient updates, a critical issue is that when there are many objectives, such as in MSP, it is often {\em difficult to find} a common descent direction. This leads to an important question: Is it more effective to separate highly conflicting objectives into different optimization levels or to keep them in a single level? To address this question, this paper investigates three multi-objective MSP formulations, which we refer to as \textbf{objective soup recipes}. These formulations apply multi-objective optimization at different optimization levels to mitigate potential conflicts among all objectives. To keep computation and memory overhead low, we incorporate a lightweight layer‑selection strategy that detects the most conflicting layers and uses only their gradients when computing the conflict‑avoidance direction. We conduct an extensive investigation using the CoVoST v2 dataset for combined multilingual ASR and ST tasks, along with the LibriSpeech and AISHELL-1 datasets for multilingual ASR, to identify highly conflicting objectives and determine the most effective training recipe among the three proposed multi-objective optimization algorithms.



Paperid:2195
Authors:Jixuan He, Chieh Lin, Lu Qi, Ming-Hsuan Yang
Abstract:
Motion is one of the key components in deformable 3D scenes. Generative video models allow users to animate static scenes with text prompts for novel motion, but when it comes to 4D reconstruction, such reanimations often fall apart. The generated videos often suffer from geometric artifacts, implausible motion, and occlusions, which hinder physically consistent 4D reanimation. In this work, we introduce \textbf{Restage4D}, a geometry-preserving pipeline for deformable scene reconstruction from a single edited video. Our key insight is to leverage the unedited original video as an additional source of supervision, allowing the model to propagate accurate structure into occluded and disoccluded regions.To achieve this, we propose a video-rewinding training scheme that temporally bridges the edited and original sequences via a shared motion representation. We further introduce an occlusion-aware ARAP regularization to preserve local rigidity, and a disocclusion backtracing mechanism that supplements missing geometry in the canonical space. Together, these components enable robust reconstruction even when the edited input contains hallucinated content or inconsistent motion.We validate Restage4D on DAVIS and PointOdyssey, demonstrating improved geometry consistency, motion quality, and 3D tracking performance. Our method not only preserves deformable structure under novel motion, but also automatically corrects errors introduced by generative models, bridging the gap between flexible video synthesis and physically grounded 4D reconstruction.



Paperid:2196
Authors:Wenkai Yang, Shuming Ma, Yankai Lin, Furu Wei
Abstract:
Recent studies have shown that making a model spend more time thinking through longer Chain of Thoughts (CoTs) enables it to gain significant improvements in complex reasoning tasks. While current researches continue to explore the benefits of increasing test-time compute by extending the CoT lengths of Large Language Models (LLMs), we are concerned about a potential issue hidden behind the current pursuit of test-time scaling: Would excessively scaling the CoT length actually bring adverse effects to a model's reasoning performance? Our explorations on mathematical reasoning tasks reveal an unexpected finding that scaling with longer CoTs can indeed impair the reasoning performance of LLMs in certain domains. Moreover, we discover that there exists an optimal scaled length distribution that differs across different domains. Based on these insights, we propose a Thinking-Optimal Scaling strategy. Our method first uses a small set of seed data with varying response length distributions to teach the model to adopt different reasoning efforts for deep thinking. Then, the model selects its shortest correct response under different reasoning efforts on additional problems for self-improvement. Our self-improved models built upon Qwen2.5-32B-Instruct outperform other distillation-based 32B o1-like models across various math benchmarks, and achieve performance on par with QwQ-32B-Preview.



Paperid:2197
Authors:Shubhankar Borse, Seokeon Choi, Sunghyun Park, Jeongho Kim, Shreya Kadambi, Risheek Garrepalli, Sungrack Yun, Munawar Hayat, Fatih Porikli
Abstract:
Generation of images containing multiple humans, performing complex actions, while preserving their facial identities, is a significant challenge. A major factor contributing to this is the lack of a a dedicated benchmark. To address this, we introduce MultiHuman-Testbench, a novel benchmark for rigorously evaluating generative models for multi-human generation. The benchmark comprises 1800 samples, including carefully curated text prompts, describing a range of simple to complex human actions. These prompts are matched with a total of 5,550 unique human face images, sampled uniformly to ensure diversity across age, ethnic background, and gender. Alongside captions, we provide human-selected pose conditioning images which accurately match the prompt. We propose a multi-faceted evaluation suite employing four key metrics to quantify face count, ID similarity, prompt alignment, and action detection. We conduct a thorough evaluation of a diverse set of models, including zero-shot approaches and training-based methods, with and without regional priors. We also propose novel techniques to incorporate image and region isolation using human segmentation and Hungarian matching, significantly improving ID similarity. Our proposed benchmark and key findings provide valuable insights and a standardized tool for advancing research in multi-human image generation.



Paperid:2198
Authors:Neil He, Rishabh Anand, Hiren Madhu, Ali Maatouk, Smita Krishnaswamy, Leandros Tassiulas, Menglin Yang, Rex Ying
Abstract:
Frontier large language models (LLMs) have shown great success in text modeling and generation tasks across domains. However, natural language exhibits inherent semantic hierarchies and nuanced geometric structure, which current LLMs do not capture completely owing to their reliance on Euclidean operations such as dot-products and norms. Furthermore, recent studies have shown that not respecting the underlying geometry of token embeddings leads to training instabilities and degradation of generative capabilities. These findings suggest that shifting to non-Euclidean geometries can better align language models with the underlying geometry of text. We thus propose to operate fully in \textit{Hyperbolic space}, known for its expansive, scale-free, and low-distortion properties. To this end, we introduce \textsc{HELM}, a family of \textbf{H}yp\textbf{E}rbolic Large \textbf{L}anguage \textbf{M}odels, offering a geometric rethinking of the Transformer-based LLM that addresses the representational inflexibility, missing set of necessary operations, and poor scalability of existing hyperbolic LMs. We additionally introduce a \textbf{Mi}xture-of-\textbf{C}urvature \textbf{E}xperts model, \textsc{HELM-MiCE}, where each expert operates in a distinct curvature space to encode more fine-grained geometric structure from text, as well as a dense model, \textsc{HELM-D}. For \textsc{HELM-MiCE}, we further develop hyperbolic Multi-Head Latent Attention (\textsc{HMLA}) for efficient, reduced-KV-cache training and inference. For both models, we further develop essential hyperbolic equivalents of rotary positional encodings and root mean square normalization. We are the first to train fully hyperbolic LLMs at billion-parameter scale, and evaluate them on well-known benchmarks such as MMLU and ARC, spanning STEM problem-solving, general knowledge, and commonsense reasoning. Our results show consistent gains from our \textsc{HELM} architectures – up to 4\% – over popular Euclidean architectures used in LLaMA and DeepSeek, highlighting the efficacy and enhanced reasoning afforded by hyperbolic geometry in large-scale language model pretraining.



Paperid:2199
Authors:Wei Zhou, Guoliang Li, Haoyu Wang, Yuxing Han, Xufei Wu, Fan Wu, Xuanhe Zhou
Abstract:
Large language models (LLMs) have shown increasing effectiveness in Text-to-SQL tasks. However, another closely related problem, Cross-System SQL Translation (a.k.a., SQL-to-SQL), which adapts a query written for one database system (e.g., MySQL) into its equivalent one for another system (e.g., ClickHouse), is of great practical importance but remains underexplored. Existing SQL benchmarks are not well-suited for SQL-to-SQL evaluation, which (1) focus on a limited set of database systems (often just SQLite) and (2) cannot capture many system-specific SQL dialects (e.g., customized functions, data types, and syntax rules). Thus, in this paper, we introduce PARROT, a Practical And Realistic BenchmaRk for CrOss-System SQL Translation. PARROT comprises 598 translation pairs from 38 open-source benchmarks and real-world business services, specifically prepared to challenge system-specific SQL understanding (e.g., LLMS achieve lower than 38.53% accuracy on average). We also provide multiple benchmark variants, including PARROT-Diverse with 28,003 translation (for extensive syntax testing) and PARROT-Simple with 5,306 representative samples (for focused stress testing), covering 22 production-grade database systems. To promote future research, we release a public leaderboard and source code at: https://code4db.github.io/parrot-bench/.



Paperid:2200
Authors:XUCHEN FENG, Siyu Liao
Abstract:
Abstract:Normalizing flows are deep generative models that achieve efficient likelihood estimation and sampling through invertible transformations. A key challenge is designing linear layers that enhance expressiveness while enabling efficient computation of the Jacobian determinant and inverse. In this work, we introduce a novel invertible linear layer based on the product of circulant and diagonal matrices. This decomposition provides a compact representation, reducing parameter complexity while approximating arbitrary linear transformations. Furthermore, leveraging the Fast Fourier Transform (FFT), our method reduces the time complexity of matrix inversion from $\mathcal{O}(n^{3})$ to $\mathcal{O}(mn \log n)$ and matrix log-determinant from $\mathcal{O}(n^{3})$ to $\mathcal{O}(mn)$, where $n$ is the input dimension. Building upon this, we introduce a novel normalizing flow model called Circulant-Diagonal Flow (CDFlow). Empirical results demonstrate that CDFlow excels in density estimation for natural image datasets and effectively models data with inherent periodicity. In terms of computational efficiency, our method speeds up the matrix inverse and log-determinant computations by $1.93\times$ and $3.22\times$, respectively, compared to the general dense matrix, when the number of channels is set to 96.



Paperid:2201
Authors:Andrea Dunn Beltran, Daniel Rho, Marc Niethammer, Roni Sengupta
Abstract:
Simultaneous Localization and Mapping (SLAM) systems typically assume static, distant illumination; however, many real-world scenarios, such as endoscopy, subterranean robotics, and search & rescue in collapsed environments, require agents to operate with a co-located light and camera in the absence of external lighting. In such cases, dynamic near-field lighting introduces strong, view-dependent shading that significantly degrades SLAM performance. We introduce Near-Field Lighting Bundle Adjustment Loss (NFL-BA) which explicitly models near-field lighting as a part of Bundle Adjustment loss and enables better performance for scenes captured with dynamic lighting. NFL-BA can be integrated into neural rendering-based SLAM systems with implicit or explicit scene representations. Our evaluations mainly focus on endoscopy procedure where SLAM can enable autonomous navigation, guidance to unsurveyed regions, blindspot detections, and 3D visualizations, which can significantly improve patient outcomes and endoscopy experience for both physicians and patients. Replacing Photometric Bundle Adjustment loss of SLAM systems with NFL-BA leads to significant improvement in camera tracking, 37% for MonoGS and 14% for EndoGSLAM, and leads to state-of-the-art camera tracking and mapping performance on the C3VD colonoscopy dataset. Further evaluation on indoor scenes captured with phone camera with flashlight turned on, also demonstrate significant improvement in SLAM performance due to NFL-BA.



Paperid:2202
Authors:Bohdan Turbal, Iryna Voitsitska, Lesia Semenova
Abstract:
Algorithmic recourse offers actionable explanations for model predictions by identifying minimal input changes to improve an outcome.A critical challenge is ensuring these explanations are robust, especially when the set of near-optimal models (the Rashomon set) is present, as multiple models can equally well explain the data. Such multiplicity means that recourse actions valid for one model might fail for another, making the explanations unreliable.This paper presents a novel framework for generating efficient and provably robust algorithmic recourse. By optimizing counterfactuals over an ellipsoidal approximation of the Rashomon set, our method finds explanations that are inherently robust to model multiplicity. We provide theoretical guarantees on the uniqueness, cost, and stability of the generated counterfactuals. Our approach produces explanations that are shorter and more robust than those from baselines, offering a practical pathway to more reliable recourse.



Paperid:2203
Authors:Gouki Minegishi, Hiroki Furuta, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo
Abstract:
Recent large-scale reasoning models have achieved state-of-the-art performance on challenging mathematical benchmarks, yet the internal mechanisms underlying their success remain poorly understood. In this work, we introduce the notion of a reasoning graph, extracted by clustering hidden‐state representations at each reasoning step, and systematically analyze three key graph-theoretic properties: cyclicity, diameter, and small-world index, across multiple tasks (GSM8K, MATH500, AIME~2024). Our findings reveal that distilled reasoning models (e.g., DeepSeek-R1-Distill-Qwen-32B) exhibit significantly more recurrent cycles (about 5 per sample), substantially larger graph diameters, and pronounced small-world characteristics (about 6x) compared to their base counterparts. Notably, these structural advantages grow with task difficulty and model capacity, with cycle detection peaking at the 14B scale and exploration diameter maximized in the 32B variant, correlating positively with accuracy. Furthermore, we show that supervised fine-tuning on an improved dataset systematically expands reasoning graph diameters in tandem with performance gains, offering concrete guidelines for dataset design aimed at boosting reasoning capabilities. By bridging theoretical insights into reasoning graph structures with practical recommendations for data construction, our work advances both the interpretability and the efficacy of large reasoning models.



Paperid:2204
Authors:Huiyang Shao, Xin Xia, Yuxi Ren, XING WANG, Xuefeng Xiao
Abstract:
Abstract:**Diffusion models** achieve remarkable generative performance but are hampered by slow, iterative inference. Model distillation seeks to train a fast student generator. **Variational Score Distillation (VSD)** offers a principled KL-divergence minimization framework for this task. This method cleverly avoids computing the teacher model's Jacobian, but its student gradient relies on the score of the student's own noisy marginal distribution, $\nabla\_{\mathbf{x}\_t} \log p\_{\phi,t}(\mathbf{x}\_t)$. VSD thus requires approximations, such as training an auxiliary network to estimate this score. These approximations can introduce biases, cause training instability, or lead to an incomplete match of the target distribution, potentially focusing on conditional means rather than broader distributional features.We introduce **VarFlow**, a method based on a **Score-Rule Variational Distillation (SRVD)** framework. VarFlow trains a one-step generator $g_{\phi}(\mathbf{z})$ by directly minimizing an energy distance (derived from the strictly proper energy score) between the student's induced noisy data distribution $p_{\phi,t}(\mathbf{x}_t)$ and the teacher's target noisy distribution $q_t(\mathbf{x}_t)$. This objective is estimated entirely using samples from these two distributions. Crucially, VarFlow bypasses the need to compute or approximate the intractable student score. By directly matching the full noisy marginal distributions, VarFlow aims for a more comprehensive and robust alignment between student and teacher, offering an efficient and theoretically grounded path to high-fidelity one-step generation.



Paperid:2205
Authors:Hsi-Che Lin, Yu-Chu Yu, Kai-Po Chang, Frank Wang
Abstract:
Open-source foundation models have seen rapid adoption and development, enabling powerful general-purpose capabilities across diverse domains. However, fine-tuning large foundation models for domain-specific or personalized tasks remains prohibitively expensive for most users due to the significant memory overhead beyond that of inference. We introduce EMLoC, an Emulator-based Memory-efficient fine-tuning framework with LoRA Correction, which enables model fine-tuning within the same memory budget required for inference. EMLoC constructs a task-specific light-weight emulator using activation-aware singular value decomposition (SVD) on a small downstream calibration set. Fine-tuning then is performed on this lightweight emulator via LoRA. To tackle the misalignment between the original model and the distilled emulator, we propose a novel compensation algorithm to correct the fine-tuned LoRA module, which thus can be merged into the original model for inference. EMLoC supports flexible compression ratios and standard training pipelines, making it adaptable to a wide range of applications. Extensive experiments demonstrate that EMLoC outperforms other baselines across multiple datasets and modalities. Moreover, without quantization, EMLoC enables fine-tuning of a 38B model on a single 24GB consumer GPU—bringing efficient and practical model adaptation to individual users.



Paperid:2206
Authors:Gaia Di Lorenzo, Federico Tombari, Marc Pollefeys, Daniel Barath
Abstract:
Learning effective multi-modal 3D representations of objects is essential for numerous applications, such as augmented reality and robotics. Existing methods often rely on task-specific embeddings that are tailored either for semantic understanding or geometric reconstruction. As a result, these embeddings typically cannot be decoded into explicit geometry and simultaneously reused across tasks.In this paper, we propose Object-X, a versatile multi-modal object representation framework capable of encoding rich object embeddings (e.g., images, point cloud, text) and decoding them back into detailed geometric and visual reconstructions. Object-X operates by geometrically grounding the captured modalities in a 3D voxel grid and learning an unstructured embedding fusing the information from the voxels with the object attributes. The learned embedding enables 3D Gaussian Splatting-based object reconstruction, while also supporting a range of downstream tasks, including scene alignment, single-image 3D object reconstruction, and localization.Evaluations on two challenging real-world datasets demonstrate that Object-X produces high-fidelity novel-view synthesis comparable to standard 3D Gaussian Splatting, while significantly improving geometric accuracy. Moreover, Object-X achieves competitive performance with specialized methods in scene alignment and localization.Critically, our object-centric descriptors require 3-4 orders of magnitude less storage compared to traditional image- or point cloud-based approaches, establishing Object-X as a scalable and highly practical solution for multi-modal 3D scene representation.



Paperid:2207
Authors:Mark Towers, Ariel Kwiatkowski, John Balis, Gianluca De Cola, Tristan Deleu, Manuel Goulão, Kallinteris Andreas, Markus Krimmel, Arjun KG, Rodrigo Perez-Vicente, J Terry, Andrea Pierré, Sander Schulhoff, Jun Jet Tai, Hannah Tan, Omar G. Younis
Abstract:
Reinforcement Learning (RL) is a continuously growing field that has the potential to revolutionize many areas of artificial intelligence. However, despite its promise, RL research is often hindered by the lack of standardization in environment and algorithm implementations. This makes it difficult for researchers to compare and build upon each other's work, slowing down progress in the field.Gymnasium is an open-source library that provides a standard API for RL environments, aiming to tackle this issue. Gymnasium's main feature is a set of abstractions that allow for wide interoperability between environments and training algorithms, making it easier for researchers to develop and test RL algorithms. In addition, Gymnasium provides a collection of easy-to-use environments, tools for easily customizing environments, and tools to ensure the reproducibility and robustness of RL research.Through this unified framework, Gymnasium significantly streamlines the process of developing and testing RL algorithms, enabling researchers to focus more on innovation and less on implementation details. By providing a standardized platform for RL research, Gymnasium helps to drive forward the field of reinforcement learning and unlock its full potential. Gymnasium is available online at \url{https://github.com/Farama-Foundation/Gymnasium}.



Paperid:2208
Authors:Umberto Cappellazzo, Minsu Kim, Pingchuan Ma, Honglie Chen, Stavros Petridis, Maja Pantic, Xubo Liu
Abstract:
Large language models (LLMs) have recently shown strong potential in audio-visual speech recognition (AVSR), but their high computational demands and sensitivity to token granularity limit their practicality in resource-constrained settings. Token compression methods can reduce inference cost, but they require fixing a compression rate in advance and produce a single fixed-length output, offering no flexibility to balance information density and efficiency at inference time. Matryoshka representation learning (MRL) addresses this by enabling a single model to operate across multiple token granularities, allowing compression rates to be adjusted dynamically. However, current MRL-based methods treat each scale independently during training, limiting cross-scale generalization, robustness at high compression, and interpretability. To overcome these limitations, we propose MoME (Mixture of Matryoshka Experts), a novel framework that integrates sparse Mixture-of-Experts (MoE) into MRL-based LLMs for AVSR. MoME augments a frozen LLM with top-k routed and shared experts, allowing dynamic capacity allocation across scales and modalities. A shared router promotes consistent expert activation across granularities, enabling compressed sequences to benefit from representations learned at lower compression. Experiments on LRS2 and LRS3 demonstrate that MoME achieves state-of-the-art performance across AVSR, ASR, and VSR tasks, while requiring significantly fewer parameters and maintaining robustness under noise. MoME unifies the adaptability of MRL with the efficiency of MoE, offering a scalable and interpretable solution for resource-aware speech recognition.



Paperid:2209
Authors:Ran Levinstein, Matan Tsipory, Itay Evron, Mark Kong, Deanna Needell, Daniel Soudry
Abstract:
We analyze task ordering strategies in continual learning for realizable linear regression.We focus on task orderings that greedily maximize dissimilarity between consecutive tasks, a concept briefly explored in prior work but still surrounded by open questions.Using tools from the Kaczmarz method literature, we formalize these orderings and develop both geometric and algebraic intuitions around them.We show empirically that, under random data, greedy orderings lead to faster convergence of the loss compared to random orderings.In a simplified setting, we prove bounds on the loss and establish optimality guarantees for greedy orderings.However, we also construct an adversarial task sequence that exploits high dimensionality to induce maximal forgetting under greedy orderings---an effect to which random orderings are notably more robust.Altogether, our findings advance the theoretical understanding of task orderings in continual learning, offer new insights into Kaczmarz methods, and provide a foundation for future research.



Paperid:2210
Authors:Sayantan Choudhury, Nicolas Loizou
Abstract:
Abstract:Introduced by Korpelevich in 1976, the extragradient method (EG) has become a cornerstone technique for solving min-max optimization problems and variational inequalities (VIs). Despite its longstanding presence and significant attention within the optimization community, most works focusing on understanding its convergence guarantees assume the strong $L$-Lipschitz condition. In this work, building on the proposed assumptions by Zhang et al. [2019] for minimization and Vankov et al. [2024] for VIs, we focus on the more relaxed $\alpha$-symmetric $(L_0, L_1)$-Lipschitz condition. This condition generalizes the standard Lipschitz assumption by allowing the Lipschitz constant to scale with the operator norm, providing a more refined characterization of problem structures in modern machine learning. Under the $\alpha$-symmetric $(L_0, L_1)$-Lipschitz condition, we propose a novel step size strategy for EG and establish sublinear convergence rates for monotone operators and linear convergence rates for strongly monotone operators. Additionally, we prove local convergence guarantees for weak Minty variational inequality problems. We supplement our analysis with experiments validating our theory and demonstrating the effectiveness and robustness of the proposed step sizes for EG.



Paperid:2211
Authors:Yongdong Luo, Xiawu Zheng, Guilin Li, Shukang Yin, Haojia Lin, Chaoyou Fu, Jinfa Huang, Jiayi Ji, Fei Chao, Jiebo Luo, Rongrong Ji
Abstract:
Existing large video-language models (LVLMs) struggle to comprehend long videos correctly due to limited context. To address this problem, fine-tuning long-context LVLMs and employing GPT-based agents have emerged as promising solutions. However, fine-tuning LVLMs would require extensive high-quality data and substantial GPU resources, while GPT-based agents would rely on proprietary models (e.g., GPT-4o). In this paper, we propose Video Retrieval-Augmented Generation (Video-RAG), a training-free and cost-effective pipeline that employs visually-aligned auxiliary texts to help facilitate cross-modality alignment while providing additional information beyond the visual content. Specifically, we leverage open-source external tools to extract visually-aligned information from pure video data (e.g., audio, optical character, and object detection), and incorporate the extracted information into an existing LVLM as auxiliary texts, alongside video frames and queries, in a plug-and-play manner. Our Video-RAG offers several key advantages: (i) lightweight with low computing overhead due to single-turn retrieval; (ii) easy implementation and compatibility with any LVLM; and (iii) significant, consistent performance gains across long video understanding benchmarks, including Video-MME, MLVU, and LongVideoBench. Notably, our model demonstrates superior performance over proprietary models like Gemini-1.5-Pro and GPT-4o when utilized with a 72B model.



Paperid:2212
Authors:Yuan Sun, Julio Contreras, Jorge Ortiz
Abstract:
Visual autoregressive modeling has recently demonstrated potential in image tasks by enabling coarse-to-fine, next-level prediction. Most indoor 3D occupancy prediction methods, however, continue to rely on dense voxel grids and convolution-heavy backbones, which incur high computational costs when applying such coarse-to-fine frameworks. In contrast, cost-efficient alternatives based on Gaussian representations—particularly in the context of multi-scale autoregression—remain underexplored. To bridge this gap, we propose DFGauss, a dynamic focused masking framework for multi-scale 3D Gaussian representation. Unlike conventional approaches that refine voxel volumes or 2D projections, DFGauss directly operates in the 3D Gaussian parameter space, progressively refining representations across resolutions under hierarchical supervision. Each finer-scale Gaussian is conditioned on its coarser-level counterpart, forming a scale-wise autoregressive process. To further enhance efficiency, we introduce an importance-guided refinement strategy that selectively propagates informative Gaussians across scales, enabling spatially adaptive detail modeling. Experiments on 3D occupancy benchmarks demonstrate that DFGauss achieves competitive performance, highlighting the promise of autoregressive modeling for scalable 3D occupancy prediction.



Paperid:2213
Authors:Ming Wen, Jiaqi Zhu, Yuedong Xu, Yipeng Zhou, DINGDING HAN
Abstract:
Large language models (LLMs) typically require fine-tuning for domain-specific tasks, and LoRA offers a computationally efficient approach by training low-rank adaptors. LoRA is also communication-efficient for federated LLMs when multiple users collaboratively fine-tune a global LLM model without sharing their proprietary raw data. However, even the transmission of local adaptors between a server and clients risks serious privacy leakage. Applying differential privacy (DP) to federated LoRA encounters a dilemma: adding noise to both adaptors amplifies synthetic noise on the model, while fixing one adaptor impairs the learnability of fine-tuning. In this paper, we propose FedASK (Differentially Private Federated Low Rank Adaptation with Double SKetching) , a novel federated LoRA framework to enable effective updating of both low-rank adaptor matrices with robust differential privacy. Inspired by randomized SVD, our key idea is a two-stage sketching pipeline. This pipeline first aggregates carefully sketched, privacy-preserving local updates, and then reconstructs the global matrices on the server to facilitate effective updating of both adaptors. We theoretically prove FedASK's differential privacy guarantee and its exact aggregation property. Comprehensive experiments demonstrate that FedASK consistently outperforms baseline methods across a variety of privacy settings and data distributions.



Paperid:2214
Authors:Jiyuan Liu, Xinwang Liu, Xinhang Wan, KE LIANG, Weixuan Liang, sihang zhou, Huijun Wu, Kehua Guo
Abstract:
Incomplete multi-view deep clustering is an emerging research hot-pot to incorporate data information of multiple sources or modalities when parts of them are missing. Most of existing approaches encode the available data observations into multiple view-specific latent representations and subsequently integrate them for the next clustering task. However, they ignore that the latent representations are unique to a fixed set of data samples in all views. Meanwhile, the pair-wise similarities of missing data observations are also failed to utilize in latent representation learning sufficiently, leading to unsatisfactory clustering performance. To address these issues, we propose an incomplete multi-view deep clustering method with data imputation and alignment. Assuming that each data sample corresponds to a same latent representation among all views, it projects the latent representations into feature spaces with neural networks. As a result, not only the available data observations are reconstructed, but also the missing ones can be imputed accordingly. Moreover, a linear alignment measurement of linear complexity is defined to compute the pair-wise similarities of all data observations, especially including those of the missing. By executing the above two procedures iteratively, the discriminative latent representations can be learned and used to group the data into categories with off-the-shelf clustering algorithms. In experiment, the proposed method is validated on a set of benchmark datasets and achieves state-of-the-art performances.



Paperid:2215
Authors:Wonje Choi, Jooyoung Kim, Honguk Woo
Abstract:
We address the challenge of adopting language models (LMs) for embodied tasks in dynamic environments, where online access to large-scale inference engines or symbolic planners is constrained due to latency, connectivity, and resource limitations. To this end, we present NeSyPr, a novel embodied reasoning framework that compiles knowledge via neurosymbolic proceduralization, thereby equipping LM-based agents with structured, adaptive, and timely reasoning capabilities. In NeSyPr, task-specific plans are first explicitly generated by a symbolic tool leveraging its declarative knowledge. These plans are then transformed into composable procedural representations that encode the plans' implicit production rules, enabling the resulting composed procedures to be seamlessly integrated into the LM's inference process. This neurosymbolic proceduralization abstracts and generalizes multi-step symbolic structured path-finding and reasoning into single-step LM inference, akin to human knowledge compilation. It supports efficient test-time inference without relying on external symbolic guidance, making it well suited for deployment in latency-sensitive and resource-constrained physical systems. We evaluate NeSyPr on the embodied benchmarks PDDLGym, VirtualHome, and ALFWorld, demonstrating its efficient reasoning capabilities over large-scale reasoning models and a symbolic planner, while using more compact LMs.



Paperid:2216
Authors:Malyaban Bal, Brian Matejek, Susmit Jha, Adam Cobb
Abstract:
Video Temporal Grounding (VTG) aims to retrieve precise temporal segments in a video conditioned on natural language queries. Unlike conventional neural frameworks that rely heavily on computationally expensive dense matrix multiplications, Spiking Neural Networks (SNNs)—previously underexplored in this domain—offer a unique opportunity to tackle VTG tasks through bio-plausible spike-based communication and an event-driven accumulation-based computational paradigm. We introduce SpikingVTG, a multi-modal spiking detection transformer, designed to harness the computational simplicity and sparsity of SNNs for VTG tasks. Leveraging the temporal dynamics of SNNs, our model introduces a Saliency Feedback Gating (SFG) mechanism that assigns dynamic saliency scores to video clips and applies multiplicative gating to highlight relevant clips while suppressing less informative ones. SFG enhances performance and reduces computational overhead by minimizing neural activity. We analyze the layer-wise convergence dynamics of SFG-enabled model and apply implicit differentiation at equilibrium to enable efficient, BPTT-free training. To improve generalization and maximize performance, we enable knowledge transfer by optimizing a Cos-L2 representation matching loss that aligns the layer-wise representation and attention maps of a non-spiking teacher with those of our student SpikingVTG. Additionally, we present Normalization-Free (NF)-SpikingVTG, which eliminates non-local operations like softmax and layer normalization, and an extremely quantized 1-bit (NF)-SpikingVTG variant for potential deployment on edge devices. Our models achieve competitive results on QVHighlights, Charades-STA, TACoS, and YouTube Highlights, establishing a strong baseline for multi-modal spiking VTG solutions.



Paperid:2217
Authors:Chaofan Lin, Jiaming Tang, Shuo Yang, Hanshuo Wang, Tian Tang, Boyu Tian, Ion Stoica, Song Han, Mingyu Gao
Abstract:
Abstract:Leveraging attention sparsity to accelerate long-context large language models (LLMs) has been of great importance recently. However, most existing sparse attention algorithms use a fixed budget of how many tokens to use in their computations. This simple static decision raises critical issues in real-world deployment because it fails to account for the dynamic nature of real-world scenarios, where the optimal balance between accuracy and efficiency can vary greatly. In this paper, we reveal a key insight that leveraging the idea of top-$p$ sampling (a.k.a., nucleus sampling) in sparse attention could enable efficient and adaptive budget decisions. Based on this, we propose Twilight, a framework that enhances any existing sparse attention algorithm with adaptive budget decision capabilities without sacrificing accuracy. Empirical results show that Twilight can adaptively prune up to 98% tokens with nearly no accuracy loss in both mid- and long-context scenarios, leading to a $1.4\times$ speedup over state-of-the-art sparse attention mechanisms.



Paperid:2218
Authors:Woojin Chung, Jeonghoon Kim
Abstract:
Abstract:Large language models are trained with tokenizers that map text to a fixed vocabulary, yet the resulting token distribution is highly imbalanced: a few words dominate the stream while most occur rarely. Recent practice favours ever-larger vocabularies, but it is unclear whether the benefit comes from better word segmentation or from amplifying this frequency skew. To this end, we perform a controlled study that scales the vocabulary of a constant-size Transformer from 24K to 196K symbols while holding data, compute and optimisation unchanged. Above 24K every common word is already a single token, so further growth only increases imbalance. Word-level loss decomposition shows that larger vocabularies reduce cross-entropy almost exclusively by lowering uncertainty on the ~$2,500$ most frequent words, even though loss on the rare tail rises. Constraining input and output embedding norms—thereby erasing the frequency signal—reverses the gain, directly demonstrating that the model exploits rather than suffers from imbalance. Because the same frequent words cover roughly $80\%$ of tokens in downstream benchmarks, this training advantage transfers intact. We further show that enlarging model parameters with a fixed tokenizer yields the same frequent-word benefit, revealing a shared mechanism behind vocabulary and model scaling. Our results recast “bigger vocabularies help” as “sharper frequency imbalance helps,” offering a simple, principled knob for tokenizer–model co-design and clarifying the loss dynamics that govern language-model scaling in pre-training.



Paperid:2219
Authors:Shutong Ding, Ke Hu, Shan Zhong, Haoyang Luo, Weinan Zhang, Jingya Wang, Jun Wang, Ye Shi
Abstract:
Recent advances in reinforcement learning (RL) have demonstrated the powerful exploration capabilities and multimodality of generative diffusion-based policies. While substantial progress has been made in offline RL and off-policy RL settings, integrating diffusion policies into on-policy frameworks like PPO remains underexplored. This gap is particularly significant given the widespread use of large-scale parallel GPU-accelerated simulators, such as IsaacLab, which are optimized for on-policy RL algorithms and enable rapid training of complex robotic tasks. A key challenge lies in computing state-action log-likelihoods under diffusion policies, which is straightforward for Gaussian policies but intractable for flow-based models due to irreversible forward-reverse processes and discretization errors (e.g., Euler-Maruyama approximations). To bridge this gap, we propose GenPO, a generative policy optimization framework that leverages exact diffusion inversion to construct invertible action mappings. GenPO introduces a novel doubled dummy action mechanism that enables invertibility via alternating updates, resolving log-likelihood computation barriers. Furthermore, we also use the action log-likelihood for unbiased entropy and KL divergence estimation, enabling KL-adaptive learning rates and entropy regularization in on-policy updates. Extensive experiments on eight IsaacLab benchmarks, including legged locomotion (Ant, Humanoid, Anymal-D, Unitree H1, Go2), dexterous manipulation (Shadow Hand), aerial control (Quadcopter), and robotic arm tasks (Franka), demonstrate GenPO’s superiority over existing RL baselines. Notably, GenPO is the first method to successfully integrate diffusion policies into on-policy RL, unlocking their potential for large-scale parallelized training and real-world robotic deployment.



Paperid:2220
Authors:Jiacheng Cui, Xinyue Bi, Yaxin Luo, Xiaohan Zhao, Jiacheng Liu, Zhiqiang Shen
Abstract:
Residual connection has been extensively studied and widely applied at the model architecture level. However, its potential in the more challenging data-centric approaches remains unexplored. In this work, we introduce the concept ofData Residual Matchingfor the first time, leveraging data-level skip connections to facilitate data generation and mitigate data information vanishing. This approach maintains a balance between newly acquired knowledge through pixel space optimization and existing core local information identification within raw data modalities, specifically for the dataset distillation task. Furthermore, by incorporating optimization-level refinements, our method significantly improves computational efficiency, achieving superior performance while reducing training time and peak GPU memory usage by 50\%. Consequently, the proposed methodFast andAccurateDataResidualMatching for Dataset Distillation (FADRM) establishes a new state-of-the-art, demonstrating substantial improvements over existing methods across multiple dataset benchmarks in both efficiency and effectiveness. For instance, with ResNet-18 as the student model and a 0.8\% compression ratio on ImageNet-1K, the method achieves 47.7\% test accuracy in single-model dataset distillation and 50.0\% in multi-model dataset distillation, surpassing RDED by +5.7\% and outperforming state-of-the-art multi-model approaches, EDC and CV-DD, by +1.4\% and +4.0\%.



Paperid:2221
Authors:Junyu Chen, Junzhuo Li, Zhen Peng, Wenjie Wang, Yuxiang Ren, Long Shi, Xuming Hu
Abstract:
Quantization and fine-tuning are crucial for deploying large language models (LLMs) on resource-constrained edge devices. However, fine-tuning quantized models presents significant challenges, primarily stemming from: First, the mismatch in data types between the low-precision quantized weights (e.g., 4-bit) and the high-precision adaptation weights (e.g., 16-bit). This mismatch limits the computational efficiency advantage offered by quantized weights during inference. Second, potential accuracy degradation when merging these high-precision adaptation weights into the low-precision quantized weights, as the adaptation weights often necessitate approximation or truncation. Third, as far as we know, no existing methods support the lossless merging of adaptation while adjusting all quantized weights. To address these challenges, we introduce lossless ternary adaptation for quantization-aware fine-tuning (LoTA-QAF). This is a novel fine-tuning method specifically designed for quantized LLMs, enabling the lossless merging of ternary adaptation weights into quantized weights and the adjustment of all quantized weights. LoTA-QAF operates through a combination of: i) A custom-designed ternary adaptation (TA) that aligns ternary weights with the quantization grid and uses these ternary weights to adjust quantized weights. ii) A TA-based mechanism that enables the lossless merging of adaptation weights. iii) Ternary signed gradient descent (t-SignSGD) for updating the TA weights. We apply LoTA-QAF to Llama-3.1/3.3 and Qwen-2.5 model families and validate its effectiveness on several downstream tasks. On the MMLU benchmark, our method effectively recovers performance for quantized models, surpassing 16-bit LoRA by up to 5.14\%. For task-specific fine-tuning, 16-bit LoRA achieves superior results, but LoTA-QAF still outperforms other methods. Code is available in the supplement.



Paperid:2222
Authors:Arian Mousakhan, Sudhanshu Mittal, Silvio Galesso, Karim Farid, Thomas Brox
Abstract:
Existing world models for autonomous driving struggle with long-horizon generation and generalization to challenging scenarios. In this work, we develop a model using simple design choices, and without additional supervision or sensors, such as maps, depth, or multiple cameras.We show that our model yields state-of-the-art performance, despite having only 469M parameters and being trained on 280h of video data. It particularly stands out in difficult scenarios like turning maneuvers and urban traffic. We test whether discrete token models possibly have advantages over continuous models based on flow matching. To this end, we set up a hybrid tokenizer that is compatible with both approaches and allows for a side-by-side comparison. Our study concludes in favor of the continuous autoregressive model, which is less brittle on individual design choices and more powerful than the model built on discrete tokens. We will open source our model upon publication.



Paperid:2223
Authors:Hemanth Saratchandran, Simon Lucey
Abstract:
We present a theoretical analysis of the Jacobian of a self-attention block within a transformer, showing that it is governed by the query, key, and value projections that define the attention mechanism. Leveraging this insight, we introduce a method that systematically alters the spectral properties of each self-attention layer to reduce the Jacobian’s condition number, thereby improving the overall conditioning of the self-attention layers within a transformer network. We empirically show that this improved Jacobian conditioning translates to enhanced performance in practice. Our approach is simple, broadly applicable, and can be easily integrated as a drop-in replacement for a wide range of existing attention mechanisms. We validate its effectiveness across diverse transformer architectures and tasks, demonstrating consistent improvements in performance.



Paperid:2224
Authors:Shriram Chennakesavalu, Frank Hu, Sebastian Ibarraran, Grant Rotskoff
Abstract:
Searching through chemical space is an exceptionally challenging problem because the number of possible molecules grows combinatorially with the number of atoms. Large, autoregressive models trained on databases of chemical compounds have yielded powerful generators, but we still lack robust strategies for generating molecules with desired properties. This molecular search problem closely resembles the "alignment" problem for large language models, though for many chemical tasks we have a specific and easily evaluable reward function. Here, we introduce an algorithm called energy rank alignment (ERA) that leverages an explicit reward function to produce a gradient-based objective that we use to optimize autoregressive policies. We show theoretically that this algorithm is closely related to proximal policy optimization (PPO) and direct preference optimization (DPO), but has a minimizer that converges to an ideal Gibbs-Boltzmann distribution with the reward playing the role of an energy function. Furthermore, this algorithm is highly scalable, does not require reinforcement learning, and performs well relative to DPO when the number of preference observations per pairing is small. We deploy this approach to align molecular transformers and protein language models to generate molecules and protein sequences, respectively, with externally specifiedproperties and find that it does so robustly, searching through diverse parts of chemical space.



Paperid:2225
Authors:Jie An, De Wang, Pengsheng Guo, Jiebo Luo, Alex Schwing
Abstract:
Recent work studying the generalization of diffusion models with UNet-based denoisers reveals inductive biases that can be expressed via geometry-adaptive harmonic bases. In practice, more recent denoising networks are often based on transformers, e.g., the diffusion transformer (DiT). This raises the question: do transformer-based denoising networks exhibit inductive biases that can also be expressed via geometry-adaptive harmonic bases? To our surprise, we find that this is not the case. This discrepancy motivates our search for the inductive bias that can lead to good generalization in DiT models. Investigating a DiT’s pivotal attention modules, we find that locality of attention maps in a DiT’s early layers are closely associated with generalization. To verify this finding, we modify the generalization of a DiT by restricting its attention windows. We inject local attention windows in early layers of a DiT and observe an improvement in generalization. Furthermore, we empirically find that both the placement and the effective attention size of these local attention windows are crucial factors. Experimental results on the CelebA, ImageNet, MSCOCO, and LSUN data show that strengthening the inductive bias of a DiT can improve both generalization and generation quality when less training data is available. Source code will be released publicly upon paper publication.



Paperid:2226
Authors:Weiqi Li, Xuanyu Zhang, Shijie Zhao, Yabin ZHANG, Junlin Li, Li zhang, Jian Zhang
Abstract:
Image quality assessment (IQA) focuses on the perceptual visual quality of images, playing a crucial role in downstream tasks such as image reconstruction, compression, and generation. The rapid advancement of multi-modal large language models (MLLMs) has significantly broadened the scope of IQA, moving toward comprehensive image quality understanding that incorporates content analysis, degradation perception, and comparison reasoning beyond mere numerical scoring. Previous MLLM-based methods typically either generate numerical scores lacking interpretability or heavily rely on supervised fine-tuning (SFT) using large-scale annotated datasets to provide descriptive assessments, limiting their flexibility and applicability. In this paper, we propose Q-Insight, a reinforcement learning-based model built upon group relative policy optimization (GRPO), which demonstrates strong visual reasoning capability for image quality understanding while requiring only a limited amount of rating scores and degradation labels. By jointly optimizing score regression and degradation perception tasks with carefully designed reward functions, our approach effectively exploits their mutual benefits for enhanced performance. Extensive experiments demonstrate that Q-Insight substantially outperforms existing state-of-the-art methods on both score regression and degradation perception tasks, while exhibiting impressive zero-shot generalization and superior comparison reasoning capability. The code and models will be made available.



Paperid:2227
Authors:En-Hao Gao, Cunjing Ge, Yuan Jiang, Zhi-Hua Zhou
Abstract:
Discovering symbolic Partial Differential Equation (PDE) from data is one of the most promising directions of modern scientific discovery. However, effectively constructing an expressive yet concise hypothesis space and accurately evaluating expression values remain challenging due to the exponential explosion with the spatial dimension and the noise in the measurements. To address these challenges, we propose the ABL-PDE approach that employs the Abductive Learning (ABL) framework to discover symbolic PDEs. By introducing a First-Order Logic (FOL) knowledge base, ABL-PDE can represent various PDEs, significantly constraining the hypothesis space without sacrificing expressive power, while also facilitating the incorporation of problem-specific knowledge. Furthermore, the proposed consistency optimization process establishes a synergistic interaction between the knowledge base and the neural network, achieving robust structure identification, accurate coefficient estimation, and enhanced stability against hyperparameter variation. Experimental results on three benchmarks across different noise levels demonstrate the state-of-the-art performance of our approach in PDE discovery.



Paperid:2228
Authors:Chang Wu, ZHIYUAN LIU, Wen Shu, Liang Wang, Yanchen Luo, Wenqiang Lei, Yatao Bian, Junfeng Fang, Xiang Wang
Abstract:
Masked graph modeling (MGM) is a promising approach for molecular representation learning (MRL). However, extending the success of re-mask decoding from 2D to 3D MGM is non-trivial, primarily due to two conflicting challenges: avoiding 2D structure leakage to the decoder, while still providing sufficient 2D context for reconstructing re-masked atoms. To address these challenges, we propose 3D-GSRD: a 3D Molecular Graph Auto-Encoder with Selective Re-mask Decoding. The core innovation of 3D-GSRD lies in its Selective Re-mask Decoding (SRD), which re-masks only 3D-relevant information from encoder representations while preserving the 2D graph structures. This SRD is synergistically integrated with a 3D Relational-Transformer (3D-ReTrans) encoder alongside a structure-independent decoder. We analyze that SRD, combined with the structure-independent decoder, enhances the encoder's role in MRL. Extensive experiments show that 3D-GSRD achieves strong downstream performance, setting a new state-of-the-art on 7 out of 8 targets in the widely used MD17 molecular property prediction benchmark. The code is released at https://anonymous.4open.science/r/3D-GSRD-4845.



Paperid:2229
Authors:Wufei Ma, Yu-Cheng Chou, Qihao Liu, Xingrui Wang, Celso de Melo, Jianwen Xie, Alan Yuille
Abstract:
Despite recent advances on multi-modal models, 3D spatial reasoning remains a challenging task for state-of-the-art open-source and proprietary models. Recent studies explore data-driven approaches and achieve enhanced spatial reasoning performance by fine-tuning models on 3D-related visual question-answering data. However, these methods typically perform spatial reasoning in an implicit manner and often fail on questions that are trivial to humans, even with long chain-of-thought reasoning. In this work, we introduce SpatialReasoner, a novel large vision-language model (LVLM) that addresses 3D spatial reasoning with explicit 3D representations shared between multiple stages--3D perception, computation, and reasoning. Explicit 3D representations provide a coherent interface that supports advanced 3D spatial reasoning and improves the generalization ability to novel question types. Furthermore, by analyzing the explicit 3D representations in multi-step reasoning traces of SpatialReasoner, we study the factual errors and identify key shortcomings of current LVLMs. Results show that our SpatialReasoner achieves improved performance on a variety of spatial reasoning benchmarks, outperforming Gemini 2.0 by 9.2% on 3DSRBench, and generalizes better when evaluating on novel 3D spatial reasoning questions. Our study bridges the 3D parsing capabilities of prior visual foundation models with the powerful reasoning abilities of large language models, opening new directions for 3D spatial reasoning.



Paperid:2230
Authors:Yize Sui, Jing Ren, Wenjing Yang, Ruochun Jin, Liyang Xu, Xiyao Liu, J Wang
Abstract:
LLM unlearning aims to remove sensitive or harmful information within the model, thus reducing the potential risk of generating unexpected information. However, existing Preference Optimization (PO)-based unlearning methods suffer two limitations. First, their rigid reward setting limits the effect of unlearning. Second, the lack of robustness causes unlearned information to reappear. To remedy these two weaknesses, we present a novel LLM unlearning optimization framework, namely Elastic Robust Unlearning (ERU), to efficiently and robustly remove specific knowledge from LLMs. We design the elastic reward setting instead of the rigid reward setting to enhance the unlearning performance. Meanwhile, we incorporate the refusal feature ablation into the unlearning process to trigger specific failure patterns for efficiently enhancing the robustness of the PO-based unlearning methods in multiple scenarios. Experimental results show that ERU can improve the unlearning effectiveness significantly while maintaining a high utility performance. Especially, on the WMDP-Bio benchmark, ERU shows a 9\% improvement over the second-best method, and maintains 83\% performance even under 1,000 sample fine-tuned retraining attacks, significantly better than the baseline method.



Paperid:2231
Authors:Zhepeng Cen, Yihang Yao, William Han, Zuxin Liu, DING ZHAO
Abstract:
Reinforcement fine-tuning (RFT) has emerged as a powerful post-training technique to incentivize the reasoning ability of large language models (LLMs). However, LLMs can respond very inconsistently to RFT: some show substantial performance gains, while others plateau or even degrade. To understand this divergence, we analyze the per-step influence of the RL objective and identify two key conditions for effective post-training: (1) RL-informative rollout accuracy, and (2) strong data co-influence, which quantifies how much the training data affects performance on other samples. Guided by these insights, we propose behavior injection, a task-agnostic data-augmentation scheme applied prior to RL. Behavior injection enriches the supervised finetuning (SFT) data by seeding exploratory and exploitative behaviors, effectively making the model more RL-ready. We evaluate our method across two reasoning benchmarks with multiple base models. The results demonstrate that our theoretically motivated augmentation can significantly increases the performance gain from RFT over the pre-RL model.



Paperid:2232
Authors:Ruijie Zhang, Ziyue Liu, Zhengyang Wang, Zheng Zhang
Abstract:
Abstract:Training foundation models such as ViTs and LLMs requires tremendous computing cost. Low-rank matrix or tensor factorization offers a parameter-efficient alternative, but often downgrades performance due to the restricted parameter space. In this work, we introduce ${\textbf{Latent Crossing (LaX)}}$ -- a simple yet effective plug-and-play module that enhances the capacity of low-rank models by enabling information flow across low-rank subspaces. We extensively validate the benefits of LaX on pre-training tasks with ViT-Base/Large and LLaMA-like models ranging from 60M to 1B parameters. LaX boosts low-rank model performance to match or exceed the full-rank baselines while using 2-3$\times$ fewer parameters. When equipped with low-rank adapters (i.e., LoRA) for fine-tuning LLaMA-7/13B, LaX consistently improves performance on arithmetic and common sense reasoning tasks with negligible cost.



Paperid:2233
Authors:Addison J. Wu, Ryan Liu, Kerem Oktar, Ted Sumers, Tom Griffiths
Abstract:
Abstract:Human communication is $\textit{motivated}$: people speak, write, and create content with a particular communicative intent in mind. As a result, the information large language models (LLMs) and associated AI agents receive is inherently biased by humans' intentions and incentives. People are remarkably attuned to navigating such biased information---we easily identify benevolent or self-serving motives in order to know what information to trust. For LLMs to be effective in the real world, they too must critically evaluate content by accounting for the motivations of the source: for example, discounting the claims made in a sales pitch. In this paper we undertake a comprehensive study of whether LLMs have this capacity of $\textit{motivational vigilance}$. We first use controlled experiments from cognitive science to identify that LLMs follow rational models of learning from motivated testimony, successfully discounting information from biased sources in a human-like manner. We then extend our evaluation to online recommendations, a more naturalistic reflection of LLM agents' information ecosystems. In these settings, we find that LLMs' inferences do not track the rational models' predictions nearly as closely---in part due to the presence of additional information that distract LLMs from vigilance-relevant considerations. Accordingly, a simple steering intervention that boosts the salience of intentions and incentives substantially increases the correspondence between LLMs and the rational model. These results suggest that LLMs possess a basic sensitivity to the motivations of others, but generalizing to novel real-world settings will require further improvements to these models.



Paperid:2234
Authors:Guozhen Zhu, Yuqian Hu, Weihang Gao, Wei-Hsiang Wang, Beibei Wang, K. Liu
Abstract:
WiFi sensing has emerged as a compelling contactless modality for human activity monitoring by capturing fine-grained variations in Channel State Information (CSI). Its ability to operate continuously and non-intrusively while preserving user privacy makes it particularly suitable for health monitoring. However, existing WiFi sensing systems struggle to generalize in real-world settings, largely due to datasets collected in controlled environments with homogeneous hardware and fragmented, session-based recordings that fail to reflect continuous daily activity.We present CSI-Bench, a large-scale, in-the-wild benchmark dataset collected using commercial WiFi edge devices across 26 diverse indoor environments with 35 real users. Spanning over 461 hours of effective data, CSI-Bench captures realistic signal variability under natural conditions. It includes task-specific datasets for fall detection, breathing monitoring, localization, and motion source recognition, as well as a co-labeled multitask dataset with joint annotations for user identity, activity, and proximity. To support the development of robust and generalizable models, CSI-Bench provides standardized evaluation splits and baseline results for both single-task and multi-task learning. CSI-Bench offers a foundation for scalable, privacy-preserving WiFi sensing systems in health and broader human-centric applications.



Paperid:2235
Authors:Boyuan Chen, Donghai Hong, Jiaming Ji, Jiacheng Zheng, Bowen Dong, Jiayi Zhou, Kaile Wang, Juntao Dai, Xuyao Wang, wenqi chen, Qirui Zheng, Wenxin Li, Sirui Han, Yike Guo, Yaodong Yang
Abstract:
As multimodal large models (MLLMs) continue to advance across challenging tasks, a key question emerges: \textbf{\textit{What essential capabilities are still missing? }}A critical aspect of human learning is continuous interaction with the environment -- not limited to language, but also involving multimodal understanding and generation.To move closer to human-level intelligence, models must similarly support \textbf{multi-turn}, \textbf{multimodal interaction}. In particular, they should comprehend interleaved multimodal contexts and respond coherently in ongoing exchanges.In this work, we present \textbf{an initial exploration} through the \textsc{InterMT} -- \textbf{the first preference dataset for \textit{multi-turn} multimodal interaction}, grounded in real human feedback. In this exploration, we particularly emphasize the importance of human oversight, introducing expert annotations to guide the process, motivated by the fact that current MLLMs lack such complex interactive capabilities. \textsc{InterMT} captures human preferences at both global and local levels into nine sub-dimensions, consists of 15.6k prompts, 52.6k multi-turn dialogue instances, and 32.4k human-labeled preference pairs. To compensate for the lack of capability for multi-modal understanding and generation, we introduce an agentic workflow that leverages tool-augmented MLLMs to construct multi-turn QA instances.To further this goal, we introduce \textsc{InterMT-Bench} to assess the ability ofMLLMs in assisting judges with multi-turn, multimodal tasks.We demonstrate the utility of \textsc{InterMT} through applications such as judge moderation and further reveal the \textit{multi-turn scaling law} of judge model.We hope the open-source of our data can help facilitate further research on aligning current MLLMs to the next step.



Paperid:2236
Authors:Olawale Salaudeen, Haoran Zhang, Kumail Alhamoud, Sara Beery, Marzyeh Ghassemi
Abstract:
Benchmarks for out‑of‑distribution (OOD) generalization frequently show a strong positive correlation between in‑distribution (ID) and OOD accuracy, termed "accuracy‑on‑the‑line." This pattern is often taken to imply that spurious correlations---correlations that improve ID but reduce OOD performance---are rare in practice. We find that this positive correlation is an artifact of aggregating heterogeneous OOD examples. Using a simple gradient‑based method, we partition each benchmark’s OOD split into semantically coherent subsets where accuracy on the line does not hold. Across six widely used distribution shift benchmarks, the method uncovers subsets, sometimes up to 77% of the usual OOD split, where higher ID accuracy predicts lower OOD accuracy. Our findings indicate that aggregate metrics can obscure important failure modes of OOD robustness. We release code and the identified subsets to facilitate further evaluation.



Paperid:2237
Authors:Lei Wang, Jieming Bian, Letian Zhang, Jie Xu
Abstract:
Abstract:Large Language Models (LLMs) have demonstrated impressive capabilities across various tasks, but fine-tuning them for domain-specific applications often requires substantial domain-specific data that may be distributed across multiple organizations. Federated Learning (FL) offers a privacy-preserving solution, but faces challenges with computational constraints when applied to LLMs. Low-Rank Adaptation (LoRA) has emerged as a parameter-efficient fine-tuning approach, though a single LoRA module often struggles with heterogeneous data across diverse domains. This paper addresses two critical challenges in federated LoRA fine-tuning: 1. determining the optimal number and allocation of LoRA experts across heterogeneous clients, and 2. enabling clients to selectively utilize these experts based on their specific data characteristics. We propose FedLEASE (Federated adaptive LoRA Expert Allocation and SElection), a novel framework that adaptively clusters clients based on representation similarity to allocate and train domain-specific LoRA experts. It also introduces an adaptive top-$M$ Mixture-of-Experts mechanism that allows each client to select the optimal number of utilized experts. Our extensive experiments on diverse benchmark datasets demonstrate that FedLEASE significantly outperforms existing federated fine-tuning approaches in heterogeneous client settings while maintaining communication efficiency.



Paperid:2238
Authors:Yeongmin Kim, HeeSun Bae, Byeonghu Na, Il-chul Moon
Abstract:
Abstract:Direct preference optimization (DPO) is widely used as a simple and stable method for aligning large language models (LLMs) with human preferences. This paper investigates a generalized DPO loss that enables a policy model to match the target policy from a likelihood ratio estimation perspective. The ratio of the target policy provides a unique identification of the policy distribution without relying on reward models or partition functions. This allows the generalized loss to retain both simplicity and theoretical guarantees, which prior work such as $f$-PO fails to achieve simultaneously. We propose \textit{Bregman preference optimization} (BPO), a generalized framework for ratio matching that provides a family of objective functions achieving target policy optimality. BPO subsumes DPO as a special case and offers tractable forms for all instances, allowing implementation with a few lines of code. We further develop scaled Basu's power divergence (SBA), a gradient scaling method that can be used for BPO instances. The BPO framework complements other DPO variants and is applicable to target policies defined by these variants. In experiments, unlike other probabilistic loss extensions such as $f$-DPO or $f$-PO, which exhibits a trade-off between generation fidelity and diversity, instances of BPO improve both win rate and entropy compared with DPO. When applied to Llama-3-Instruct-8B, BPO achieves state-of-the-art performance among Llama-3-8B backbones, with a 55.9\% length-controlled win rate on AlpacaEval2.



Paperid:2239
Authors:Yiyang Zhou, Yangfan He, Yaofeng Su, Siwei Han, Joel Jang, Gedas Bertasius, Mohit Bansal, Huaxiu Yao
Abstract:
Video understanding is fundamental to tasks such as action recognition, video reasoning, and robotic control. Early video understanding methods based on large vision-language models (LVLMs) typically adopt a single-pass reasoning paradigm without dynamic feedback, limiting the model’s capacity to self-correct and adapt in complex scenarios. Recent efforts have attempted to address this limitation by incorporating reward models and reinforcement learning to enhance reasoning, or by employing tool-agent frameworks. However, these approaches face several challenges, including high annotation costs, reward signals that fail to capture real-time reasoning states, and low inference efficiency. To overcome these issues, we propose ReAgent-V, a novel agentic video understanding framework that integrates efficient frame selection with real-time reward generation during inference. These reward signals not only guide iterative answer refinement through a multi-perspective reflection mechanism—adjusting predictions from conservative, neutral, and aggressive viewpoints—but also enable automatic filtering of high-quality data for supervised fine-tuning (SFT), direct preference optimization (DPO), and group relative policy optimization (GRPO). ReAgent-V is lightweight, modular, and extensible, supporting flexible tool integration tailored to diverse tasks. Extensive experiments on 12 datasets across three core applications—video understanding, video reasoning enhancement, and vision-language-action model alignment—demonstrate significant gains in generalization and reasoning, with improvements of up to 6.9%, 2.1%, and 9.8%, respectively, highlighting the effectiveness and versatility of the proposed framework.



Paperid:2240
Authors:Neal Mangaokar, Ashish Hooda, Zhuohang Li, Bradley Malin, Kassem Fawaz, Somesh Jha, Atul Prakash, Amrita Roy Chowdhury
Abstract:
Membership inference tests aim to determine whether a particular data point was included in a language model's training set. However, recent works have shown that such tests often fail under the strict definition of membership based on exact matching, and have suggested relaxing this definition to include semantic neighbors as members as well. In this work, we show that membership inference tests are stillunreliableunder this relaxation - it is possible to poison the training dataset in a way that causes the test to produce incorrect predictions for a target point. We theoretically reveal a trade-off between a test’s accuracy and its robustness to poisoning. We also present a concrete instantiation of this poisoning attack and empirically validate its effectiveness. Our results show that it can degrade the performance of existing tests to well below random.



Paperid:2241
Authors:Ebrahim Feghhi, Shreyas Kaasyap, Nima Hadidi, Jonathan Kao
Abstract:
Abstract:Speech neuroprostheses offer a path to restore communication in patients with severe paralysis by decoding speech directly from neural activity. To function seamlessly, speech neuroprostheses should ideally be (1) accurate, (2) capable of real-time decoding, (3) efficient in terms of memory and computation, and (4) robust to distribution shifts in neural actvity over time. Prior work on the Brain-to-Text Benchmark improved the accuracy of a baseline gated recurrent unit (GRU) decoder by integrating the outputs of an ensemble of GRUs with large language models. However, these gains come at the expense of criteria (2) and (3). In this work, we make holistic improvements to the baseline through two core contributions: masking contiguous chunks of neural activity during training, and replacing the GRU with a lightweight Transformer. Together, these improvements yield a $20$\% reduction in word error rate while substantially reducing memory requirements and per-epoch training times. Finally, leveraging the Transformer’s suitability for test-time adaptation, we show improved robustness to across-day distribution shifts by calibrating the decoder to make confident and consistent predictions across time-masked versions of the same input during inference.



Paperid:2242
Authors:He Chen, Jiajin Li, Anthony Man-Cho So
Abstract:
Solving bilevel optimization (BLO) problems to global optimality is generally intractable. A common alternative is to compute a hyper-stationary point—a stationary point of the hyper-objective function formed by minimizing/maximizing the upper-level function over the lower-level solution set. However, existing approaches either yield weak notions of stationarity or rely on restrictive assumptions to ensure the smoothness of hyper-objective functions. In this paper, we remove these impractical assumptions and show thatstrong(Clarke) hyper-stationarity is still computable even when the hyper-objective is nonsmooth. Our key tool is a new structural condition, calledset smoothness, which captures the variational relationship between the lower-level solution set and the upper-level variable. We prove that this condition holds for a broad class of BLO problems and ensures weak convexity (resp. concavity) of pessimistic (resp. optimistic) hyper-objective functions. Building on this, we show that a zeroth-order algorithm computes approximate Clarke hyper-stationary points with a non-asymptotic convergence guarantee. To the best of our knowledge, this is the first computational guarantee for Clarke-type stationarity for nonsmooth hyper-objective functions in BLO. Our developments, especially the set smoothness property, contribute to a deeper understanding of BLO computability and may inspire applications in other fields.



Paperid:2243
Authors:Xiang Li, Yong Tao, Siyuan Zhang, Siwei Liu, Zhitong Xiong, Chunbo Luo, Lu Liu, Mykola Pechenizkiy, Xiaoxiang Zhu, Tianjin Huang
Abstract:
Earth observation foundation models have shown strong generalization across multiple Earth observation tasks, but their robustness under real-world perturbations remains underexplored. To bridge this gap, we introduce REOBench, the first comprehensive benchmark for evaluating the robustness of Earth observation foundation models across six tasks and twelve types of image corruptions, including both appearance-based and geometric perturbations. To ensure realistic and fine-grained evaluation, our benchmark focuses on high-resolution optical remote sensing images, which are widely used in critical applications such as urban planning and disaster response. We conduct a systematic evaluation of a broad range of models trained using masked image modeling, contrastive learning, and vision-language pre-training paradigms. Our results reveal that (1) existing Earth observation foundation models experience significant performance degradation when exposed to input corruptions. (2) The severity of degradation varies across tasks, model architectures, backbone sizes, and types of corruption, with performance drop varying from less than 1% to over 20%. (3) Vision-language models show enhanced robustness, particularly in multimodal tasks. REOBench underscores the vulnerability of current Earth observation foundation models to real-world corruptions and provides actionable insights for developing more robust and reliable models.



Paperid:2244
Authors:junliang ye, Zhengyi Wang, Ruowen Zhao, Shenghao Xie, Jun Zhu
Abstract:
Recently, the powerful text-to-image capabilities of GPT-4o have led to growing appreciation for native multimodal large language models. However, its multimodal capabilities remain confined to images and text. Yet beyond images, the ability to understand and generate 3D content is equally crucial. To address this gap, we propose ShapeLLM-4o—a native 3D large language model capable of understanding and generating 3D assets and text in any sequence. First, we train a 3D vector-quantized variational autoencoder (VQ-VAE), which maps 3D objects into a discrete latent space to achieve efficient and accurate shape representation and reconstruction. Building upon the 3D-aware discrete tokens, we innovatively construct a large-scale continuous training dataset named 3D-Alpaca, encompassing generation, comprehension, and editing, thus providing rich resources for future research and training. Finally, by performing instruction-based training of the Qwen-2.5-vl-7B-Instruct model on the 3D-Alpaca dataset. Our work provides an effective attempt at extending multimodal models with basic 3D capabilities, which contributes to future research in 3D-native AI.



Paperid:2245
Authors:Kairun Wen, Yuzhihuang, Runyu Chen, Hui Zheng, Yunlong Lin, Panwang Pan, Chenxin Li, Wenyan Cong, Jian Zhang, Junbin Lu, Chenguo Lin, Dilin Wang, Zhicheng Yan, Hongyu Xu, Justin Theiss, Yue Huang, Xinghao Ding, Rakesh Ranjan, Zhiwen Fan
Abstract:
Understanding the dynamic physical world, characterized by its evolving 3D structure, real-world motion, and semantic content with textual descriptions, is crucial for human-agent interaction and enables embodied agents to perceive and act within real environments with human‑like capabilities. However, existing datasets are often derived from limited simulators or utilize traditional Structure-from-Motion for up-to-scale annotation and offer limited descriptive captioning, which restricts the capacity of foundation models to accurately interpret real-world dynamics from monocular videos, commonly sourced from the internet.To bridge these gaps, we introduce DynamicVerse, a physical‑scale, multimodal 4D modeling framework for real-world video. We employ large vision, geometric, and multimodal models to interpret metric-scale static geometry, real-world dynamic motion, instance-level masks, and holistic descriptive captions. By integrating window-based Bundle Adjustment with global optimization, our method converts long real-world video sequences into a comprehensive 4D multimodal format. DynamicVerse delivers a large-scale dataset consists of 100K+ video with 800K+ annotated masks and 10M+ frames from internet videos. Experimental evaluations on three benchmark tasks---video depth estimation, camera pose estimation, and camera intrinsics estimation---validate that our 4D modeling achieves superior performance in capturing physical-scale measurements with greater global accuracy than existing methods.



Paperid:2246
Authors:Nicolas Wolf, Leif Seute, Vsevolod Viliuga, Simon Wagner, Jan Stühmer, Frauke Gräter
Abstract:
Deep generative models have recently been proposed for sampling protein conformations from the Boltzmann distribution, as an alternative to often prohibitively expensive Molecular Dynamics simulations. However, current state-of-the-art approaches rely on fine-tuning pre-trained folding models and evolutionary sequence information, limiting their applicability and efficiency, and introducing potential biases. In this work, we propose a flow matching model for sampling protein conformations based solely on backbone geometry. We introduce a geometric encoding of the backbone equilibrium structure as input and propose to condition not only the flow but also the prior distribution on the respective equilibrium structure, eliminating the need for evolutionary information. The resulting model is orders of magnitudes faster than current state-of-the-art approaches at comparable accuracy, is transferable to multi-chain proteins, and can be trained from scratch in a few GPU days. In our experiments, we demonstrate that the proposed model achieves competitive performance with reduced inference time, across not only an established benchmark of naturally occurring proteins but also de novo proteins, for which evolutionary information is scarce.



Paperid:2247
Authors:Zonglin Lyu, Ming Li, Xinxin Liu, Chen Chen
Abstract:
Abstract:To enhance controllability in text-to-image generation, ControlNet introduces image-based control signals, while ControlNet++ improves pixel-level cycle consistency between generated images and the input control signal. To avoid the prohibitive cost of back-propagating through the sampling process, ControlNet++ optimizes only low-noise timesteps (e.g., $t < 200$) using a single-step approximation, which not only ignores the contribution of high-noise timesteps but also introduces additional approximation errors. A straightforward alternative for optimizing controllability across all timesteps is Direct Preference Optimization (DPO), a fine-tuning method that increases model preference for more controllable images ($I^{w}$) over less controllable ones ($I^{l}$). However, due to uncertainty in generative models, it is difficult to ensure that win--lose image pairs differ only in controllability while keeping other factors, such as image quality, fixed. To address this, we propose performing preference learning over control conditions rather than generated images. Specifically, we construct winning and losing control signals, $\mathbf{c}^{w}$ and $\mathbf{c}^{l}$, and train the model to prefer $\mathbf{c}^{w}$. This method, which we term \textit{Condition Preference Optimization} (CPO), eliminates confounding factors and yields a low-variance training objective. Our approach theoretically exhibits lower contrastive loss variance than DPO and empirically achieves superior results. Moreover, CPO requires less computation and storage for dataset curation. Extensive experiments show that CPO significantly improves controllability over the state-of-the-art ControlNet++ across multiple control types: over $10$% error rate reduction in segmentation, $70$-$80$% in human pose, and consistent $2$%-$5$% reductions in edge and depth maps. Here, the error rate is defined as the difference between evaluated controllability and oracle results.



Paperid:2248
Authors:Guang Yang, Jingwen Qiao, Tianpei Yang, Yanqing Wu, Jing Huo, Yang Gao, Xingguo Chen
Abstract:
Communication is one of the effective means to improve the learning of cooperative policy in multi-agent systems. However, in most real-world scenarios, lossy communication is a prevalent issue. Existing multi-agent reinforcement learning with communication, due to their limited scalability and robustness, struggles to apply to complex and dynamic real-world environments. To address these challenges, we propose a generalized communication-constrained model to uniformly characterize communication conditions across different scenarios. Based on this, we utilize it as a learning prior to distinguish between lossy and lossless messages for specific scenarios. Additionally, we decouple the impact of lossy and lossless messages on distributed decision-making, drawing on a dual mutual information estimatior, and introduce a communication-constrained multi-agent reinforcement learning framework, quantifying the impact of communication messages into the global reward. Finally, we validate the effectiveness of our approach across several communication-constrained benchmarks.



Paperid:2249
Authors:Shervin Khalafi, Ignacio Hounie, Dongsheng Ding, Alejandro Ribeiro
Abstract:
Diffusion models have become prevalent in generative modeling due to their ability to sample from complex distributions. To improve the quality of generated samples and their compliance with user requirements, two commonly used methods are: (i) Alignment, which involves fine-tuning a diffusion model to align it with a reward; and (ii) Composition, which combines several pre-trained diffusion models together, each emphasizing a desirable attribute in the generated outputs. However, trade-offs often arise when optimizing for multiple rewards or combining multiple models, as they can often represent competing properties. Existing methods cannot guarantee that the resulting model faithfully generates samples with all the desired properties. To address this gap, we propose a constrained optimization framework that unifies alignment and composition of diffusion models by enforcing that the aligned model satisfies reward constraints and/or remains close to each pre-trained model. We provide a theoretical characterization of the solutions to the constrained alignment and composition problems and develop a Lagrangian-based primal-dual training algorithm to approximate these solutions. Empirically, we demonstrate our proposed approach in image generation, applying it to alignment and composition, and show that our aligned or composed model satisfies constraints effectively.



Paperid:2250
Authors:Mintong Kang, Zhaorun Chen, Chejian Xu, Jiawei Zhang, Chengquan Guo, Minzhou Pan, Ivan Revilla, Yu Sun, Bo Li
Abstract:
As large language models (LLMs) become widespread across diverse applications, concerns about the security and safety of LLM interactions have intensified. Numerous guardrail models and benchmarks have been developed to ensure LLM content safety. However, existing guardrail benchmarks are often built upon ad hoc risk taxonomies that lack a principled grounding in standardized safety policies, limiting their alignment with real-world operational requirements. Moreover, they tend to overlook domain-specific risks, while the same risk category can carry different implications across different domains. To bridge these gaps, we introduce PolyGuard, the first massive multi-domain safety policy-grounded guardrail dataset. PolyGuard offers: (1) broad domain coverage across eight safety-critical domains, such as finance, law, and codeGen; (2) policy-grounded risk construction based on authentic, domain-specific safety guidelines; (3) diverse interaction formats, encompassing declarative statements, questions, instructions, and multi-turn conversations; (4) advanced benign data curation via detoxification prompting to challenge over-refusal behaviors; and (5) \textbf{attack-enhanced instances} that simulate adversarial inputs designed to bypass guardrails. Based on PolyGuard, we benchmark 19 advanced guardrail models and uncover a series of findings, such as: (1) All models achieve varied F1 scores, with many demonstrating high variance across risk categories, highlighting their limited domain coverage and insufficient handling of domain-specific safety concerns; (2) As models evolve, their coverage of safety risks broadens, but performance on common risk categories may decrease; (3) All models remain vulnerable to optimized adversarial attacks. The policy-grounded \dataset establishes the first principled and comprehensive guardrail benchmark. We believe that \dataset and the unique insights derived from our evaluations will advance the development of policy-aligned and resilient guardrail systems.



Paperid:2251
Authors:Hongshu Guo, Zeyuan Ma, Yining Ma, Xinglin Zhang, Wei-Neng Chen, Yue-Jiao Gong
Abstract:
Designing effective black‑box optimizers is hampered by limited problem-specific knowledge and manual tuning that spans months for almost every detail. In this paper, we present \textit{DesignX}, the first automated algorithm design framework that generates an effective optimizer specific to a given black-box optimization problem within seconds. Rooted in the first principles, we identify two key sub-tasks: 1) algorithm structure generation and 2) parameter tuning. To enable systematic construction, a comprehensive modular algorithmic space is first built, embracing hundreds of algorithm components collected from decades of research. We then introduce a dual-agent reinforcement learning system that collaborates on structural and parametric design through a novel cooperative training objective, enabling large-scale meta-training across 10k diverse instances. Remarkably, through days of autonomous learning, the DesignX-generated optimizers continuously surpass human-crafted optimizers by orders of magnitude, either on synthetic testbed or on realistic optimization scenarios such as Protein-docking, AutoML and UAV path planning. Further in-depth analysis reveals DesignX's capability to discover non-trivial algorithm patterns beyond expert intuition, which, conversely, provides valuable design insights for the optimization community. We provide DesignX's inference code at~\url{https://anonymous.4open.science/r/DesignX-AF70/}.



Paperid:2252
Authors:Bo Wang, Qinyuan Cheng, Runyu Peng, Rong Bao, Peiji Li, Qipeng Guo, Linyang Li, Zhiyuan Zeng, Yunhua Zhou, Xipeng Qiu
Abstract:
Post-training processes are essential phases in grounding pre-trained language models to real-world tasks, with learning from demonstrations or preference signals playing a crucial role in this adaptation. We present a unified theoretical framework bridging Supervised Fine-Tuning (SFT) and preference learning in Large Language Model (LLM) post-training. Through rigorous mathematical derivation, we demonstrate that both SFT and preference learning methods like Direct Preference Optimization (DPO) operate within the same optimal policy-reward subspace, with SFT representing a special case of implicit reward learning. Our analysis reveals a critical limitation in conventional SFT: the KL divergence term in distribution matching becomes constant with respect to the policy during optimization, failing to constrain model updates. To address this, we propose a simple yet effective learning rate reduction approach that yields significant performance improvements (up to \textbf{25\%} relative gain and \textbf{6\%} absolute win rate increase in instruction following tasks. Additionally, we derive alternative SFT objectives from various f-divergence functions that preserve the KL term during optimization, further enhancing post-DPO model performance. Finally, we extend the theoretical relationship between LLM logits and Q-functions from preference learning to the SFT context, providing mathematical derivations and experimental validation.



Paperid:2253
Authors:Qizhou Chen, Dakan Wang, Taolin Zhang, Zaoming Yan, Chengsong You, Chengyu Wang, Xiaofeng He
Abstract:
Model editing aims to enhance the accuracy and reliability of large language models (LLMs) by efficiently adjusting their internal parameters. Currently, most LLM editing datasets are confined to narrow knowledge domains and cover a limited range of editing evaluation. They often overlook the broad scope of editing demands and the diversity of ripple effects resulting from edits. In this context, we introduce UniEdit, a unified benchmark for LLM editing grounded in open-domain knowledge. First, we construct editing samples by selecting entities from 25 common disciplines across five major categories, utilizing the extensive triple knowledge available in open-domain knowledge graphs to ensure comprehensive coverage of the knowledge domains. To address the issues of generality and locality in editing, we design an Neighborhood Multi-hop Chain Sampling (NMCS) algorithm to sample subgraphs based on a given knowledge piece to entail comprehensive ripple effects to evaluate.Finally, we employ proprietary LLMs to convert the sampled knowledge subgraphs into natural language text, guaranteeing grammatical accuracy and syntactical diversity. Extensive statistical analysis confirms the scale, comprehensiveness, and diversity of our UniEdit benchmark. We conduct comprehensive experiments across multiple LLMs and editors, analyzing their performance to highlight strengths and weaknesses in editing across open knowledge domains and various evaluation criteria, thereby offering valuable insights for future research endeavors.



Paperid:2254
Authors:Zhe Kong, Feng Gao, Yong Zhang, Zhuoliang Kang, Xiaoming Wei, Xunliang Cai, Guanying Chen, Wenhan Luo
Abstract:
Audio-driven human animation methods, such as talking head and talking body generation, have made remarkable progress in generating synchronized facial movements and appealing visual quality videos. However, existing methods primarily focus on single human animation and struggle with multi-stream audio inputs, facing incorrect binding problems between audio and persons. Additionally, they exhibit limitations in instruction-following capabilities. To solve this problem, in this paper, we propose a novel task: Multi-Person Conversational Video Generation, and introduce a new framework, MultiTalk, to address the challenges during multi-person generation. Specifically, for audio injection, we investigate several schemes and propose the Label Rotary Position Embedding (L-RoPE) method to resolve the audio and person binding problem. Furthermore, during training, we observe that partial parameter training and multi-task training are crucial for preserving the instruction-following ability of the base model. MultiTalk achieves superior performance compared to other methods on several datasets, including talking head, talking body, and multi-person datasets, demonstrating the powerful generation capabilities of our approach.



Paperid:2255
Authors:Sungwoo Park
Abstract:
We introduce Neural Hamiltonian Diffusion, a unified framework for learning stochastic Hamiltonian dynamics on differentiable manifolds. While Hamiltonian Neural Networks (HNNs) model conservative systems in flat Euclidean space, they fail to account for geometric structure and intrinsic stochasticity. Conversely, diffusion models on Riemannian manifolds offer geometry-aware stochastic modeling but lack physical inductive biases. Our method parameterizes a Hamiltonian with a neural network and defines its dynamics as a stochastic differential equation on a (pseudo-)Riemannian manifold equipped with a Poisson structure. This enables physically consistent modeling of dynamics on curved, periodic, or causally structured spaces. We demonstrate that the proposed geometric dynamics generalizes existing approaches and applies to systems ranging from molecular dynamics to relativistic n-body problems.



Paperid:2256
Authors:Zeyuan Ma, Yue-Jiao Gong, Hongshu Guo, Wenjie Qiu, Sijie Ma, Hongqiao Lian, Jiajun Zhan, Kaixu Chen, Chen Wang, Zhiyang Huang, Zechuan Huang, Guojun Peng, Ran Cheng, Yining Ma
Abstract:
Abstract:Meta-Black-Box Optimization (MetaBBO) streamlines the automation of optimization algorithm design through meta-learning. It typically employs a bi-level structure: the meta-level policy undergoes meta-training to reduce the manual effort required in developing algorithms for low-level optimization tasks. The original MetaBox (2023) provided the first open-source framework for reinforcement learning-based single-objective MetaBBO. However, its relatively narrow scope no longer keep pace with the swift advancement in this field. In this paper, we introduce MetaBox-v2 (\url{https://github.com/MetaEvo/MetaBox}) as a milestone upgrade with four novel features: 1) a unified architecture supporting RL, evolutionary, and gradient-based approaches, by which we reproduce $23$ up-to-date baselines; 2) efficient parallelization schemes, which reduce the training/testing time by $10-40$x; 3) a comprehensive benchmark suite of $18$ synthetic/realistic tasks ($1900$+ instances) spanning single-objective, multi-objective, multi-model, and multi-task optimization scenarios; 4) plentiful and extensible interfaces for custom analysis/visualization and integrating to external optimization tools/benchmarks. To show the utility of MetaBox-v2, we carry out a systematic case study that evaluates the built-in baselines in terms of the optimization performance, generalization ability and learning efficiency. Valuable insights are concluded from thorough and detailed analysis for practitioners and those new to the field.



Paperid:2257
Authors:Narun Raman, Taylor Lundy, Thiago Amin, Kevin Leyton-Brown, Jesse Perla
Abstract:
Abstract:Large language models (LLMs) are increasingly being asked to make economically rational decisions and indeed are already being applied to economic tasks like stock picking and financial analysis. Existing LLM benchmarks tend to focus on specific applications, making them insufficient for characterizing economic reasoning more broadly. In previous work, we offered a blueprint for comprehensively benchmarking $\textit{strategic}$ decision-making Raman et al. 2024. However, this work did not engage with the even larger microeconomic literature on $\textit{non-strategic}$ settings. We address this gap here, taxonomizing microeconomic reasoning into $58$ distinct elements, each grounded in up to $10$ distinct domains, $5$ perspectives, and $3$ types. The generation of benchmark data across this combinatorial space is powered by a novel LLM-assisted data generation protocol that we dub auto-STEER, which generates a set of questions by adapting handwritten templates to target new domains and perspectives. By generating fresh questions for each element, auto-STEER induces diversity which could help to reduce the risk of data contamination. We use this benchmark to evaluate $27$ LLMs spanning a range of scales and adaptation strategies, comparing performance across multiple formats—multiple-choice and free-text question answering—and scoring schemes. Our results surface systematic limitations in current LLMs' ability to generalize economic reasoning across types, formats, and textual perturbations, and establish a foundation for evaluating and improving economic competence in foundation models.



Paperid:2258
Authors:Dor Tsur, Carol Long, Claudio Mayrink Verdun, Sajani Vithana, Hsiang Hsu, Richard Chen, haim permuter, Flavio Calmon
Abstract:
Large language model (LLM) watermarks enable authentication of text provenance, curb misuse of machine-generated text, and promote trust in AI systems. Current watermarks operate by changing the next-token predictions output by an LLM. The updated (i.e., watermarked) predictions depend on random side information produced, for example, by hashing previously generated tokens. LLM watermarking is particularly challenging in low-entropy generation tasks -- such as coding -- where next-token predictions are near-deterministic. In this paper, we propose an optimization framework for watermark design. Our goal is to understand how to most effectively use random side information in order to maximize the likelihood of watermark detection and minimize the distortion of generated text. Our analysis informs the design of two new watermarks: HeavyWater and SimplexWater. Both watermarks are tunable, gracefully trading-off between detection accuracy and text distortion. They can also be applied to any LLM and are agnostic to side information generation. We examine the performance of HeavyWater and SimplexWater through several benchmarks, demonstrating that they can achieve high watermark detection accuracy with minimal compromise of text generation quality, particularly in the low-entropy regime. Our theoretical analysis also reveals surprising new connections between LLM watermarking and coding theory.



Paperid:2259
Authors:Jack Cook, Danyal Akarca, Rui Costa, Jascha Achterberg
Abstract:
The brain is made up of a vast set of heterogeneous regions that dynamically organize into pathways as a function of task demands. Examples of such pathways can be seen in the interactions between cortical and subcortical networks during learning. This raises the question of how exactly brain regions organize into these dynamic groups. In this work, we use an extension of the Heterogeneous Mixture-of-Experts architecture, to show that heterogeneous regions do not form processing pathways by themselves, implying that the brain likely implements specific constraints which result in reliable formation of pathways. We identify three biologically relevant inductive biases that encourage pathway formation: a routing cost imposed on the use of more complex regions, a scaling factor that reduces this cost when task performance is low, and randomized expert dropout. When comparing our resulting Mixture-of-Pathways model with the brain, we observe that the artificial pathways match how the brain uses cortical and subcortical systems to learn and solve tasks of varying difficulty. In summary, we introduce a novel framework for investigating how the brain forms task-specific pathways through inductive biases which may make Mixture-of-Experts architectures in general more adaptive.



Paperid:2260
Authors:Lukas Miklautz, Chengzhi Shi, Theodoros Davarakis, Andrii Shkabrii, Prudence Lam, Claudia Plant, Jennifer Dy, Efstratios Ioannidis
Abstract:
We introduce H-SPLID, a novel algorithm for learning salient feature representations through the explicit decomposition of salient and non-salient features into separate spaces. We show that H-SPLID promotes learning low-dimensional, task-relevant features. We prove that the expected prediction deviation under input perturbations is upper-bounded by the dimension of the salient subspace and the Hilbert-Schmidt Independence Criterion (HSIC) between inputs and representations. This establishes a link between robustness and latent representation compression in terms of the dimensionality and information preserved. Empirical evaluations on image classification tasks show that models trained with H-SPLID primarily rely on salient input components, as indicated by reduced sensitivity to perturbations affecting non-salient features, such as image backgrounds.



Paperid:2261
Authors:Jiaming Ji, Xinyu Chen, Rui Pan, Han Zhu, Jiahao Li, Donghai Hong, Boyuan Chen, Jiayi Zhou, Kaile Wang, Juntao Dai, Chi-Min Chan, Sirui Han, Yike Guo, Yaodong Yang
Abstract:
Multimodal large language models (MLLMs) are essential for building general-purpose AI assistants; however, they pose increasing safety risks. How can we ensure safety alignment of MLLMs to prevent undesired behaviors? Going further, it is critical to explore how to fine-tune MLLMs to preserve capabilities while meeting safety constraints. Fundamentally, this challenge can be formulated as a min-max optimization problem. However, existing datasets have not yet disentangled single preference signals into explicit safety constraints, hindering systematic investigation in this direction. Moreover, it remains an open question whether such constraints can be effectively incorporated into the optimization process for multi-modal models. In this work, we present the first exploration of the Safe RLHF-V -- the first multimodal safety alignment framework. The framework consists of: (I) BeaverTails-V, the first open-source dataset featuring dual preference annotations for helpfulness and safety, supplemented with multi-level safety labels (minor, moderate, severe); (II) Beaver-Guard-V, a multi-level guardrail system to proactively defend against unsafe queries and adversarial attacks. Applying the guard model over five rounds of filtering and regeneration significantly enhances the precursor model’s overall safety by an average of 40.9%. (II) Based on dual preference, we initiate the first exploration of multi-modal safety alignment within a constrained optimization. Experimental results demonstrate that Safe RLHF effectively improves both model helpfulness and safety. Specifically, Safe RLHF-V enhances model safety by 34.2% and helpfulness by 34.3%.



Paperid:2262
Authors:Savya Khosla, Sethuraman T V, Barnett Lee, Alex Schwing, Derek Hoiem
Abstract:
We introduce the Region Encoder Network (REN), a fast and effective model for generating region-based image representations using point prompts. Recent methods combine class-agnostic segmenters (e.g., SAM) with patch-based image encoders (e.g., DINO) to produce compact and effective region representations, but they suffer from high computational cost due to the segmentation step. REN bypasses this bottleneck using a lightweight module that directly generates region tokens, enabling 60x faster token generation with 35x less memory, while also improving token quality. It uses a few cross-attention blocks that take point prompts as queries and features from a patch-based image encoder as keys and values to produce region tokens that correspond to the prompted objects. We train REN with three popular encoders—DINO, DINOv2, and OpenCLIP—and show that it can be extended to other encoders without dedicated training. We evaluate REN on semantic segmentation and retrieval tasks, where it consistently outperforms the original encoders in both performance and compactness, and matches or exceeds SAM-based region methods while being significantly faster. Notably, REN achieves state-of-the-art results on the challenging Ego4D VQ2D benchmark and outperforms proprietary LMMs on Visual Haystacks' single-needle challenge. We will release our models and code to support further research.



Paperid:2263
Authors:Mete Erdogan, Francesco Tonin, Volkan Cevher
Abstract:
The high inference demands of transformer-based Large Language Models (LLMs) pose substantial challenges in their deployment. To this end, we introduceNeural Block Linearization(NBL), a novel framework for accelerating transformer model inference by replacing self-attention layers with linear approximations derived from Linear Minimum Mean Squared Error estimators. NBL leverages Canonical Correlation Analysis to compute a theoretical upper bound on the approximation error. Then, we use this bound as a criterion for substitution, selecting the LLM layers with the lowest linearization error. NBL can be efficiently applied to pre-trained LLMs without the need for fine-tuning. In experiments, NBL achieves notable computational speed-ups while preserving competitive accuracy on multiple reasoning benchmarks. For instance, applying NBL to 12 self-attention layers inDeepSeek-R1-Distill-Llama-8Bincreases the inference speed by 32% with less than 1% accuracy trade-off, making it a flexible and promising solution to improve the inference efficiency of LLMs.



Paperid:2264
Authors:Jiyuan Shi, Xinzhe Liu, Dewei Wang, ouyang lu, Sören Schwertfeger, Chi Zhang, Fuchun Sun, Chenjia Bai, Xuelong Li
Abstract:
Humans exhibit diverse and expressive whole-body movements. However, attaining human-like whole-body coordination in humanoid robots remains challenging, as conventional approaches that mimic whole-body motions often neglect the distinct roles of upper and lower body. This oversight leads to computationally intensive policy learning and frequently causes robot instability and falls during real-world execution. To address these issues, we propose Adversarial Locomotion and Motion Imitation (ALMI), a novel framework that enables adversarial policy learning between upper and lower body. Specifically, the lower body aims to provide robust locomotion capabilities to follow velocity commands while the upper body tracks various motions. Conversely, the upper-body policy ensures effective motion tracking when the robot executes velocity-based movements. Through iterative updates, these policies achieve coordinated whole-body control, which can be extended to loco-manipulation tasks with teleoperation systems. Extensive experiments demonstrate that our method achieves robust locomotion and precise motion tracking in both simulation and on the full-size Unitree H1-2 robot. Additionally, we release a large-scale whole-body motion control dataset featuring high-quality episodic trajectories from MuJoCo simulations. The project page is https://almi-humanoid.github.io.



Paperid:2265
Authors:Haozhe Chi, Zhicheng Sun, Yang Jin, Yi Ma, Jing Wang, Yadong Mu
Abstract:
Flow-based generative models have gained popularity for image generation and editing. For instruction-based image editing, it is critical to ensure that modifications are confined to the targeted regions. Yet existing methods often fail to maintain consistency in non-targeted regions between the original / edited images. Our primary contribution is to identify the cause of this limitation as the error accumulation across individual editing steps and to address it by incorporating the historical editing trajectory. Specifically, we formulate image editing as a control problem and leverage the Kalman filter to integrate the historical editing trajectory. Our proposed algorithm, dubbed Kalman-Edit, reuses early-stage details from the historical trajectory to enhance the structural consistency of the editing results. To speed up editing, we introduce a shortcut technique based on approximate vector field velocity estimation. Extensive experiments on several datasets demonstrate its superior performance compared to previous state-of-the-art methods.



Paperid:2266
Authors:Sohyun Lee, Yeho Gwon, Lukas Hoyer, Suha Kwak
Abstract:
Improving robustness of the Segment Anything Model (SAM) to input degradations is critical for its deployment in high-stakes applications such as autonomous driving and robotics. Our approach to this challenge prioritizes three key aspects: first, parameter efficiency to maintain the inherent generalization capability of SAM; second, fine-grained and input-aware robustification to precisely address the input corruption; and third, adherence to standard training protocols for ease of training. To this end, we propose gated-rank adaptation (GaRA). GaRA introduces lightweight adapters into intermediate layers of the frozen SAM, where each adapter dynamically adjusts the effective rank of its weight matrix based on the input by selectively activating (rank-1) components of the matrix using a learned gating module. This adjustment enables fine-grained and input-aware robustification without compromising the generalization capability of SAM. Our model, GaRA-SAM, significantly outperforms prior work on all robust segmentation benchmarks. In particular, it surpasses the previous best IoU score by up to 21.3\%p on ACDC, a challenging real corrupted image dataset.



Paperid:2267
Authors:Fan Feng, Phillip Lippe, Sara Magliacane
Abstract:
Agents that can understand and reason about objects and their interactions offer significant potential for robust and generalizable policy learning. While most object-centric reinforcement learning methods focus on object-centric state representations that are factored based on individual objects, they neglect a structured representation of interactions. In this work, we propose Factored Interactive Object-Centric World Model (FIOC-WM), a unified framework for learning structured representations of objects and their interactions within a world model, which can model the dynamics of the environment in a more fine-grained way, enabling more efficient and generalizable policy learning. FIOC-WM first learns object-centric representations with state factorization and object interactions directly from pixels, leveraging pre-trained vision models. Then, we use the learned world model to decompose complex tasks into composable interactions, and learn a hierarchical policy where the high-level decides the interaction order and type, and the low-level learns how to invoke them. We demonstrate the efficacy of FIOC-WM in world model accuracy and policy learning across a set of robotic and embodied AI environments.



Paperid:2268
Authors:Jinlin Xiang, Minho Choi, Yubo Zhang, Zhihao Zhou, Arka Majumdar, Eli Shlizerman
Abstract:
Hybrid Optical Neural Networks (ONNs, typically consisting of an optical frontend and a digital backend) offer an energy-efficient alternative to fully digital deep networks for real-time, power-constrained systems. However, their adoption is limited by two main challenges: the accuracy gap compared to large-scale networks during training, and discrepancies between simulated and fabricated systems that further degrade accuracy. While previous work has proposed end-to-end optimizations for specific datasets (e.g., MNIST) and optical systems, these approaches typically lack generalization across tasks and hardware designs. To address these limitations, we propose a task-agnostic and hardware-agnostic pipeline that supports image classification and segmentation across diverse optical systems. To assist optical system design before training, we estimate achievable model accuracy based on user-specified constraints such as physical size and the dataset. For training, we introduce Neural Tangent Knowledge Distillation (NTKD), which aligns optical models with electronic teacher networks, thereby narrowing the accuracy gap. After fabrication, NTKD also guides fine-tuning of the digital backend to compensate for implementation errors. Experiments on multiple datasets (e.g., MNIST, CIFAR, Carvana Masking) and hardware configurations show that our pipeline consistently improves ONN performance and enables practical deployment in both pre-fabrication simulations and physical implementations.



Paperid:2269
Authors:Claudia Cuttano, Gabriele Trivigno, Giuseppe Averta, Carlo Masone
Abstract:
Few-shot segmentation aims to segment unseen object categories from just a handful of annotated examples. This requires mechanisms that can both identify semantically related objects across images and accurately produce segmentation masks. We note that Segment Anything 2 (SAM2), with its prompt-and-propagate mechanism, offers both strong segmentation capabilities and a built-in feature matching process. However, we show that its representations are entangled with task-specific cues optimized for object tracking, which impairs its use for tasks requiring higher level semantic understanding. Our key insight is that, despite its class-agnostic pretraining, SAM2 already encodes rich semantic structure in its features. We propose SANSA (Semantically AligNed SegmentAnything 2), a framework that makes this latent structure explicit, and repurposes SAM2 for few-shot segmentation through minimal task-specific modifications. SANSA achieves state-of-the-art performance on few-shot segmentation benchmarks specifically designed to assess generalization, outperforms generalist methods in the popular in-context setting, supports flexible promptable interaction via points, boxes, or scribbles, and remains significantly faster and more compact than prior approaches.



Paperid:2270
Authors:Wei Xu, Cheng Wang, Dingkang Liang, Zongchuang Zhao, Xingyu Jiang, Peng Zhang, Xiang Bai
Abstract:
Underwater exploration offers critical insights into our planet and attracts increasing attention for its broader applications in resource exploration, national security, etc. We study the underwater scene understanding methods, which aim to achieve automated underwater exploration. The underwater scene understanding task demands multi-task perceptions from multiple granularities. However, the absence of large-scale underwater multi-task instruction-tuning datasets hinders the progress of this research. To bridge this gap, we construct NautData, a dataset containing 1.45 M image-text pairs supporting eight underwater scene understanding tasks. It enables the development and thorough evaluations of the underwater scene understanding models. Underwater image degradation is a widely recognized challenge that interferes with underwater tasks. To improve the robustness of underwater scene understanding, we introduce physical priors derived from underwater imaging models and propose a plug-and-play vision feature enhancement (VFE) module, which explicitly restores clear underwater information. We integrate this module into renowned baselines LLaVA-1.5 and Qwen2.5-VL and build our underwater LMM, NAUTILUS. Experiments conducted on the NautData and public underwater datasets demonstrate the effectiveness of the VFE model, consistently improving the performance of both baselines on most supported tasks, thus ensuring the superiority of NAUTILUS across the majority of tasks.



Paperid:2271
Authors:Xuan Wang, Siyuan Liang, Dongping Liao, Han Fang, Aishan Liu, Xiaochun Cao, Yu-liang Lu, Ee-Chien Chang, Xitong Gao
Abstract:
Institutions with limited data and computing resources often outsource model training to third-party providers in a semi-honest setting, assuming adherence to prescribed training protocols with pre-defined learning paradigm (e.g., supervised or semi-supervised learning). However, this practice can introduce severe security risks, as adversaries may poison the training data to embed backdoors into the resulting model. Existing detection approaches predominantly rely on statistical analyses, which often fail to maintain universally accurate detection accuracy across different learning paradigms. To address this challenge, we propose a unified backdoor detection framework in the semi-honest setting that exploits cross-examination of model inconsistencies between two independent service providers. Specifically, we integrate central kernel alignment to enable robust feature similarity measurements across different model architectures and learning paradigms, thereby facilitating precise recovery and identification of backdoor triggers. We further introduce backdoor fine-tuning sensitivity analysis to distinguish backdoor triggers from adversarial perturbations, substantially reducing false positives. Extensive experiments demonstrate that our method achieves superior detection performance, improving accuracy by 4.4%, 1.7%, and 10.6% over SoTA baselines across supervised, semi-supervised, and autoregressive learning tasks, respectively. Notably, it is the first to effectively detect backdoors in multimodal large language models, further highlighting its broad applicability and advancing secure deep learning.



Paperid:2272
Authors:Victor-Alexandru Pădurean, Parameswaran Kamalaruban, Nachiket Kotalwar, Alkis Gotovos, Adish Singla
Abstract:
We study the problem of aligning a generative model's response with a user's preferences. Recent works have proposed several different formulations for personalized alignment; however, they either require a large amount of user preference queries or require that the preference be explicitly specified as a text input. In this paper, we propose a novel inference-time personalized alignment method, UserAlign, that elicits the user's preferences with a few queries as pairwise response comparisons. In particular, UserAlign builds on the theoretical framework of best-arm identification in logistic bandits and selects a personalized response from a fixed pool of the model's generated responses. The key idea is to consider the user's feedback consistent and noise-free, and incorporate it into the theoretical framework to identify the best response quickly. Experimental results across several tasks, involving personalized text and image generation, showcase the effectiveness of UserAlign in achieving personalized alignment.



Paperid:2273
Authors:Chenyang Jiang, Hang Zhao, Xinyu Zhang, Zhengcen Li, Qiben Shan, Shaocong Wu, Jingyong Su
Abstract:
Dataset distillation compresses large-scale datasets into compact, highly informative synthetic data, significantly reducing storage and training costs. However, existing research primarily focuses on balanced datasets and struggles to perform under real-world long-tailed distributions. In this work, we emphasize the critical role of soft labels in long-tailed dataset distillation and uncover the underlying mechanisms contributing to performance degradation.Specifically, we derive an imbalance-aware generalization bound for model trained on distilled dataset. We then identify two primary sources of soft-label bias, which originate from the teacher model and the distilled images, through systematic perturbation of the data imbalance levels.To address this, we propose ADSA, an Adaptive Soft-label Alignment module that calibrates the entangled biases. This lightweight module integrates seamlessly into existing distillation pipelines and consistently improves performance. On ImageNet-1k-LT with EDC and IPC=50, ADSA improves tail-class accuracy by up to 11.8\% and raises overall accuracy to 41.4\%. Extensive experiments demonstrate that ADSA provides a robust and generalizable solution under limited label budgets and across a range of distillation techniques.



Paperid:2274
Authors:Hong Wang, Haiyang Xin, Jie Wang, Xuanze Yang, Fei Zha, huanshuo dong, Yan Jiang
Abstract:
Pre-training has proven effective in addressing data scarcity and performance limitations in solving PDE problems with neural operators.However, challenges remain due to the heterogeneity of PDE datasets in equation types, which leads to high errors in mixed training. Additionally, dense pre-training models that scale parameters by increasing network width or depth incur significant inference costs.To tackle these challenges, we propose a novelMixture-of-ExpertsPre-trainingOperatorTransformer (MoE-POT), a sparse-activated architecture that scales parameters efficiently while controlling inference costs.Specifically, our model adopts a layer-wise router-gating network to dynamically select 4 routed experts from 16 expert networks during inference, enabling the model to focus on equation-specific features. Meanwhile, we also integrate 2 shared experts, aiming to capture common properties of PDE and reduce redundancy among routed experts. The final output is computed as the weighted average of the results from all activated experts.We pre-train models with parameters from 30M to 0.5B on 6 public PDE datasets.Our model with 90M activated parameters achieves up to a 40\% reduction in zero-shot error compared with existing models with 120M activated parameters.Additionally, we conduct interpretability analysis, showing that dataset types can be inferred from router-gating network decisions, which validates the rationality and effectiveness of the MoE architecture.



Paperid:2275
Authors:Daehee Lee, TaeYoon Kwack, Dongsu Lee, Wonje Choi, Honguk Woo
Abstract:
Skill Incremental Learning (SIL) is the process by which an embodied agent expands and refines its skill set over time by leveraging experience gained through interaction with its environment or by the integration of additional data. SIL facilitates efficient acquisition of hierarchical policies grounded in reusable skills for downstream tasks. However, as the skill repertoire evolves, it can disrupt compatibility with existing skill-based policies, limiting their reusability and generalization. In this work, we propose SIL-C, a novel framework that ensures skill-policy compatibility, allowing improvements in incrementally learned skills to enhance the performance of downstream policies without requiring policy re-training or structural adaptation. SIL-C employs a bilateral lazy learning-based mapping technique to dynamically align the subtask space referenced by policies with the skill space decoded into agent behaviors. This enables each subtask, derived from the policy's decomposition of a complex task, to be executed by selecting an appropriate skill based on trajectory distribution similarity. We evaluate SIL-C across diverse SIL scenarios and demonstrate that it maintains compatibility between evolving skills and downstream policies while ensuring efficiency throughout the learning process.



Paperid:2276
Authors:Guillaume Pourcel, Maxence Ernoult
Abstract:
Abstract:Deep State Space Models (SSMs) reignite physics-grounded compute paradigms, as RNNs could natively be embodied into dynamical systems. This calls for dedicated learning algorithms obeying to core physical principles, with efficient techniques to simulate these systems and guide their design. We propose \emph{Recurrent Hamiltonian Echo Learning} (RHEL), an algorithm which provably computes loss gradients as finite differences of physical trajectories of non-dissipative, \emph{Hamiltonian systems}. In ML terms, RHEL only requires three ``forward passes'' irrespective of model size, without explicit Jacobian computation, nor incurring any variance in the gradient estimation. Motivated by the physical realization of our algorithm, we first introduce RHEL in \emph{continuous time} and demonstrate its formal equivalence with the continuous adjoint state method. To facilitate the simulation of Hamiltonian systems trained by RHEL, we propose a \emph{discrete-time} version of RHEL which is equivalent to Backpropagation Through Time (BPTT) when applied to a class of recurrent modules which we call \emph{Hamiltonian Recurrent Units} (HRUs). This setting allows us to demonstrate the scalability of RHEL by generalizing these results to hierarchies of HRUs, which we call \emph{Hamiltonian SSMs} (HSSMs). We apply RHEL to train HSSMs with linear and nonlinear dynamics on a variety of time-series tasks ranging from mid-range to long-range classification and regression with sequence length reaching $\sim 50k$. We show that RHEL consistently matches the performance of BPTT across all models and tasks. This work opens new doors for the design of scalable, energy-efficient physical systems endowed with self-learning capabilities for sequence modelling.



Paperid:2277
Authors:Florian Graf, Paolo Pellizzoni, Martin Uray, Stefan Huber, Roland Kwitt
Abstract:
Abstract:We consider the problem of computing persistent homology (PH) for large-scale Euclidean point cloud data, aimed at downstream machine learning tasks, where the exponential growth of the most widely-used Vietoris-Rips complex imposes serious computational limitations. Although more scalable alternatives such as the Alpha complex or sparse Rips approximations exist, they often still result in a prohibitively large number of simplices. This poses challenges in the complex construction and in the subsequent PH computation, prohibiting their use on large-scale point clouds. To mitigate these issues, we introduce the *Flood complex*, inspired by the advantages of the Alpha and Witness complex constructions. Informally, at a given filtration value $r\geq 0$, the Flood complex contains all simplices from a Delaunay triangulation of a small subset of the point cloud $X$ that are fully covered by balls of radius $r$ emanating from $X$, a process we call *flooding*. Our construction allows for efficient PH computation, possesses several desirable theoretical properties, and is amenable to GPU parallelization. Scaling experiments on 3D point cloud data show that we can compute PH of up to dimension 2 on several millions of points. Importantly, when assessing object classification performance on both real-world and synthetic data, we provide evidence that this scaling capability is needed, especially if objects are geometrically or topologically complex, yielding performance superior to other PH-based methods and to PointNet-like models.



Paperid:2278
Authors:Fangzhou Wu, Sandeep Silwal
Abstract:
Abstract:Increasing demand for Large Language Models (LLMs) services imposes substantial deployment and computation costs on providers. LLM routing offers a cost-efficient solution by directing queries to the optimal LLM based on model and query features.However, existing works primarily focus on offline scenarios and struggle to adapt to online settings with high query volume and constrained token budgets.In this work, we introduce the first training-free algorithm for online routing scenarios.Our algorithm leverages approximate nearest neighbor search to efficiently estimate the features of queries and performs a one-time optimization over a small set of initial queries to learn a set of routing weights that guide future routing.We provide a theoretical guarantee that the algorithm achieves a competitive ratio of $1 - o(1)$ under natural assumptions, which is further validated by extensive experiments across 3 benchmark datasets and 8 baselines, showing an average improvement of 3.55$\times$ in performance, 1.85$\times$ in cost efficiency, and nearly 4.25$\times$ in throughput.



Paperid:2279
Authors:Junren Chen, Rui Chen, Gang Liang, Wei Wang, Junlong Cheng, zhanglei-scu, Liangyin Chen
Abstract:
Automated medical image segmentation based on neural networks is pivotal in promoting health equity. The attention mechanism increasingly serves as a key component in modern neural networks for medical image segmentation, as it enables the network to focus on regions of interest, achieving improved segmentation. However, current attention mechanisms confront a performance-complexity trade-off paradox: accuracy gains demand higher computational costs, while reducing complexity sacrifices model accuracy. Such a contradiction inherently restricts real-world deployment in resource-limited settings, thus exacerbating healthcare disparities. To overcome this dilemma, we propose a parameter-free Neighborhood Self-Dissimilarity Attention (NSDA), inspired by radiologists' diagnostic patterns of prioritizing regions exhibiting substantial differences during clinical image interpretation. Unlike pairwise-similarity-based self-attention mechanisms, NSDA constructs a size-adaptive local dissimilarity measure that quantifies element-neighborhood differences. This approach enhances segmentation accuracy without adding trainable parameters directly related to computational complexity. Experiments demonstrate the effectiveness and generalization of our method. This study advances attention mechanism design for medical image analysis by clinical prior knowledge, a parameter-free paradigm that has the potential to advance health equity in low-resource settings. The code is available upon publication.



Paperid:2280
Authors:Josip Jukić, Jan Šnajder
Abstract:
In-context learning (ICL) enables large language models to perform few-shot learning by conditioning on labeled examples in the prompt. Despite its flexibility, ICL suffers from instability – especially as prompt length increases with more demonstrations. To address this, we treat ICL as a source of weak supervision and propose a parameter-efficient method that disentangles demonstration-induced latent shifts from those of the query. An ICL-based teacher generates pseudo-labels on unlabeled queries, while a student predicts them using only the query input, updating a lightweight adapter. This captures demonstration effects in a compact, reusable form, enabling efficient inference while remaining composable with new demonstrations. Although trained on noisy teacher outputs, the student often outperforms its teacher through pseudo-label correction and coverage expansion, consistent with the weak-to-strong generalization effect. Empirically, our method improves generalization, stability, and efficiency across both in-domain and out-of-domain tasks, surpassing standard ICL and prior disentanglement methods.



Paperid:2281
Authors:Hossein Babaei, Mel White, Sina Alemohammad, Richard Baraniuk
Abstract:
State-Space Models (SSMs) have proven to be powerful tools for online function approximation and for modeling long-range dependencies in sequential data. While recent methods such as HiPPO have demonstrated strong performance using a few polynomial bases, they remain limited by their reliance on closed-form solutions for specific, well-behaved bases. The SaFARi framework generalizes this approach, enabling the construction of SSMs from arbitrary frames, including non-orthogonal and redundant ones, thus allowing an infinite diversity of possible "species'' within the SSM family. In this paper, we introduce WaLRUS (Wavelets for Long-range Representation Using SSMs), a new species of SaFARi built from Daubechies wavelet frames. We instantiate two variants, scaled-Walrus and translated-Walrus, and show that their multiresolution and localized nature offers significant advantages in representing non-smooth and transient signals. We compare Walrus to HiPPO-based models and demonstrate improved accuracy, better numerical properties, and more efficient implementations for online function approximation tasks.



Paperid:2282
Authors:Jiyuan Wang, Chunyu Lin, cheng guan, Lang Nie, Jing He, Haodong Li, Kang Liao, Yao Zhao
Abstract:
In this paper, we propose \textbf{Jasmine}, the first Stable Diffusion (SD)-based self-supervised framework for monocular depth estimation, which effectively harnesses SD’s visual priors to enhance the sharpness and generalization of unsupervised prediction. Previous SD-based methods are all supervised since adapting diffusion models for dense prediction requires high-precision supervision. In contrast, self-supervised reprojection suffers from inherent challenges (\textit{e.g.}, occlusions, texture-less regions, illumination variance), and the predictions exhibit blurs and artifacts that severely compromise SD's latent priors. To resolve this, we construct a novel surrogate task of mix-batch image reconstruction. Without any additional supervision, it preserves the detail priors of SD models by reconstructing the images themselves while preventing depth estimation from degradation. Furthermore, to address the inherent misalignment between SD's scale and shift invariant estimation and self-supervised scale-invariant depth estimation, we build the Scale-Shift GRU. It not only bridges this distribution gap but also isolates the fine-grained texture of SD output against the interference of reprojection loss. Extensive experiments demonstrate that Jasmine achieves SoTA performance on the KITTI benchmark and exhibits superior zero-shot generalization across multiple datasets.



Paperid:2283
Authors:Xiangyuan Peng, Xingsi Dong, Si Wu
Abstract:
We propose Grid-like Code Quantization (GCQ), a brain-inspired method for compressing observation–action sequences into discrete representations using grid-like patterns in attractor dynamics. Unlike conventional vector quantization approaches that operate on static inputs, GCQ performs spatiotemporal compression through an action-conditioned codebook, where codewords are derived from continuous attractor neural networks and dynamically selected based on actions. This enables GCQ to jointly compress space and time, serving as a unified world model. The resulting representation supports long-horizon prediction, planning, spatial reasoning, and inverse modeling. Experiments across diverse tasks demonstrate GCQ’s effectiveness in compact encoding and downstream performance. Our work offers both a computational tool for efficient sequence modeling and a theoretical perspective on the formation of grid-like codes in neural systems.



Paperid:2284
Authors:Kabir Dabholkar, Omri Barak
Abstract:
Many natural systems, including neural circuits involved in decision making, can be modeled as high-dimensional dynamical systems with multiple stable states. While existing analytical tools primarily describe behavior near stable equilibria, characterizing separatrices -- the manifolds that delineate boundaries between different basins of attraction -- remains challenging, particularly in high-dimensional settings. Here, we introduce a numerical framework leveraging Koopman Theory combined with Deep Neural Networks to effectively characterize separatrices. Specifically, we approximate Koopman Eigenfunctions (KEFs) associated with real positive eigenvalues, which vanish precisely at the separatrices. Utilizing these scalar KEFs, optimization methods efficiently locate separatrices even in complex systems. We demonstrate our approach on synthetic benchmarks, ecological network models, and recurrent neural networks trained on neuroscience-inspired tasks. Moreover, we illustrate the practical utility of our method by designing optimal perturbations that can shift systems across separatrices, enabling predictions relevant to optogenetic stimulation experiments in neuroscience.



Paperid:2285
Authors:Zixuan Hu, Li Shen, Zhenyi Wang, Yongxian Wei, Dacheng Tao
Abstract:
Harmful fine-tuning poses critical safety risks to fine-tuning-as-a-service for large language models. Existing defense strategies preemptively build robustness via attack simulation but suffer from fundamental limitations: (i) the infeasibility of performing attack simulation due to lacking prior knowledge about potential attack data, and (ii) limited adaptability to varying attack settings, as simulation fails to capture their variability and complexity. To address these challenges, we propose Bayesian Data Scheduler (BDS), an adaptive tuning-stage defense strategy with no need for attack simulation. BDS formulates harmful fine-tuning defense as a Bayesian inference problem, learning the posterior distribution of each data point's safety attribute, conditioned on the fine-tuning and alignment datasets. The fine-tuning process is then constrained by weighting data with their safety attributes sampled from the posterior, thus mitigating the influence of harmful data. By leveraging the post hoc nature of Bayesian inference, the posterior is conditioned on the fine-tuning dataset, enabling BDS to tailor its defense to the specific dataset, thereby achieving adaptive defense. Furthermore, we introduce a neural scheduler based on amortized Bayesian learning, enabling efficient transfer to new data without retraining. Comprehensive results across diverse attack and defense settings demonstrate the state-of-the-art performance of our approach.



Paperid:2286
Authors:Jianyu Wang, Yu Wang, Shengjie Zhao, Sifan Zhou
Abstract:
Voxel-based 3D object detectors have achieved remarkable performance in point cloud perception, yet their high computational and memory demands pose significant challenges for deployment on resource-constrained edge devices. Post-training quantization (PTQ) provides a practical means to compress models and accelerate inference; however, existing PTQ methods for point cloud detection are typically limited to INT8 and lack support for lower-bit formats such as INT4, which restricts their deployment potential. In this paper, we present Point4bit, the first general 4-bit PTQ framework tailored for voxel-based 3D object detectors. To tackle challenges in low-bit quantization, we propose two key techniques: (1) Foreground-aware Piecewise Activation Quantization (FA-PAQ), which leverages foreground structural cues to improve the quantization of sparse activations; and (2) Gradient-guided Key Weight Quantization (G-KWQ), which preserves task-critical weights through gradient-based analysis to reduce quantization-induced degradation. Extensive experiments demonstrate that Point4bit achieves INT4 quantization with minimal accuracy loss with less than 1.5\% accuracy drop. Moreover, we validate its generalization ability on point cloud classification and segmentation tasks, demonstrating broad applicability. Our method further advances the bit-width limitation of point cloud quantization to 4 bits, demonstrating strong potential for efficient deployment on resource-constrained edge devices.



Paperid:2287
Authors:Tuan Tran, Duy M. H. Nguyen, Hoai-Chau Tran, Michael Barz, Khoa D Doan, Roger Wattenhofer, Vien Ngo, Mathias Niepert, Daniel Sonntag, Paul Swoboda
Abstract:
In recent years, 3D point cloud transformers have emerged as powerful tools for various tasks such as semantic segmentation, object detection, and reconstruction. While these models have demonstrated remarkable performance, they are often computationally expensive due to the large number of tokens required to process 3D point clouds. Surprisingly, we observe that despite the high computational cost of these models, they actually require only a small fraction of the tokens to achieve their performance. This discovery challenges the conventional assumption that using more tokens directly improves performance. In this work, we demonstrate that with a proper token merging strategy for 3D point cloud transformers, estimating the importance of voxel tokens based on their spatial structures, we are able to reduce the token usage by up to 95-98\% with minimal performance degradation. Our findings highlight that many existing state-of-the-art models are unnecessarily token-heavy, and our method provides a more efficient way to handle tokenization without compromising accuracy. This phenomenon holds across multiple domains, including 3D semantic segmentation, point cloud reconstruction, and object detection, demonstrating the potential for scalability and efficiency in 3D vision tasks.



Paperid:2288
Authors:Yuta Oshima, Masahiro Suzuki, Yutaka Matsuo, Hiroki Furuta
Abstract:
The remarkable progress in text-to-video diffusion models enables photorealistic generations, although the contents of the generated video often include unnatural movement or deformation, reverse playback, and motionless scenes.Recently, an alignment problem has attracted huge attention, where we steer the output of diffusion models based on some quantity on the goodness of the content.Because there is a large room for improvement of perceptual quality along the frame direction, we should address which metrics we should optimize and how we can optimize them in the video generation.In this paper, we propose diffusion latent beam search with lookahead estimator, which can select a better diffusion latent to maximize a given alignment reward, at inference time.We then point out that the improvement of perceptual video quality considering the alignment to prompts requires reward calibration by weighting existing metrics.This is because when humans or vision language models evaluate outputs, many previous metrics to quantify the naturalness of video do not always correlate with evaluation.We demonstrate that our method improves the perceptual quality evaluated on the calibrated reward, VLMs, and human assessment, without model parameter update, and outputs the best generation compared to greedy search and best-of-N sampling under much more efficient computational cost.The experiments highlight that our method is beneficial to many capable generative models, and provide a practical guideline on that we should prioritize the inference-time compute allocation into lookahead steps for reward estimate more than search budget or denoising steps.



Paperid:2289
Authors:Lin Wu, Zhixiang Chen, Jianglin Lan
Abstract:
Generating realistic 3D human-object interactions (HOIs) remains a challenging task due to the difficulty of modeling detailed interaction dynamics. Existing methods treat human and object motions independently, resulting in physically implausible and causally inconsistent behaviors. In this work, we present HOI-Dyn, a novel framework that formulates HOI generation as a driver-responder system, where human actions drive object responses. At the core of our method is a lightweight transformer-based interaction dynamics model that explicitly predicts how objects should react to human motion. To further enforce consistency, we introduce a residual-based dynamics loss that mitigates the impact of dynamics prediction errors and prevents misleading optimization signals. The dynamics model is used only during training, preserving inference efficiency. Through extensive qualitative and quantitative experiments, we demonstrate that our approach not only enhances the quality of HOI generation but also establishes a feasible metric for evaluating the quality of generated interactions.



Paperid:2290
Authors:Xiuying Wei, Anunay Yadav, Razvan Pascanu, Caglar Gulcehre
Abstract:
Abstract:Transformers have become the cornerstone of modern large-scale language models; however, their dependence on softmax attention poses a major computational bottleneck, particularly in long-context settings. In this work, rather than following prevalent approaches such as linear attention (or SSMs) and local attention, we introduce an intermediate design called RAT between recurrence and attention mechanisms. It partitions the input into chunks, applies a simple linear recurrence within each chunk to capture local dependencies, and then performs softmax attention across chunks to model long-range interactions. By adjusting the size of the chunk, RAT enables flexible trade-offs, combining the strengths of RNN and attention. Empirically, with a chunk size of 16, the RAT layer achieves a $7\times$ improvement in training speed with 1M token sequences and $9\times$ in generation at 4K sequence length, while maintaining similar or sometimes even better accuracy compared to standard attention. We demonstrate this by training 1.3B parameter models from scratch and performing large-scale evaluations, including short- and long-context benchmarks, as well as supervised fine-tuning~(SFT). We further propose a hybrid architecture that interleaves RAT with local attention. By combining efficient long-range modeling with strong local interactions, this hybrid design not only improves inference speed and reduces cache memory usage compared to attention, but also consistently enhances performance, for example, achieving an average 1 point gain in commonsense reasoning tasks, up to 4 points on code tasks, and a 1 point Rouge-L increase in a summarization SFT task.



Paperid:2291
Authors:Hao Wei, Aleksandra Franz, Björn List, Nils Thuerey
Abstract:
Abstract:When simulating partial differential equations, hybrid solvers combine coarse numerical solvers with learned correctors. They promise accelerated simulations while adhering to physical constraints. However, as shown in our theoretical framework, directly applying learned corrections to solver outputs leads to significant autoregressive errors, which originate from amplified perturbations that accumulate during long-term rollouts, especially in chaotic regimes. To overcome this, we propose the Indirect Neural Corrector ($\mathrm{INC}$), which integrates learned corrections into the governing equations rather than applying direct state updates. Our key insight is that $\mathrm{INC}$ reduces the error amplification on the order of $\Delta t^{-1} + L$, where $\Delta t$ is the timestep and $L$ the Lipschitz constant. At the same time, our framework poses no architectural requirements and integrates seamlessly with arbitrary neural networks and solvers. We test $\mathrm{INC}$ in extensive benchmarks, covering numerous differentiable solvers, neural backbones, and test cases ranging from a 1D chaotic system to 3D turbulence. INC improves the long-term trajectory performance ($R^2$) by up to 158.7\%, stabilizes blowups under aggressive coarsening, and for complex 3D turbulence cases yields speed-ups of several orders of magnitude. INC thus enables stable, efficient PDE emulation with formal error reduction, paving the way for faster scientific and engineering simulations with reliable physics guarantees.



Paperid:2292
Authors:Yang Minghao, Zechen Bai, Jing Lin, Haoqian Wang, Alex Jinpeng Wang
Abstract:
Recent advances in visual tokenizers have demonstrated their effectiveness for multimodal large language models and autoregressive generative models. However, most existing visual tokenizers rely on a fixed downsampling rate at a given visual resolution, and consequently produce a constant number of visual tokens, ignoring the fact that visual information of varying complexity warrant different token budgets. Motivated by this observation, we propose an adaptive video tokenizer "VaporTok" with two core contributions:Probabilistic Taildrop: We introduce a novel taildrop mechanism that learns a truncation index sampling distribution conditioned on visual complexity of the video. During both training and inference, the decoder reconstructs videos at adaptive token lengths, allocating more tokens to complex videos and fewer to simpler ones. Parallel Sample GRPO with Vapor Reward: By leveraging the probability distribution produced by probabilistic taildrop, we reformulate the visual tokenization pipeline as a sequential decision process. To optimize this process, we propose a variant of GRPO and a composite reward encompassing token efficiency, reconstruction fidelity, and generative quality, thus enabling metrics-aware adaptive tokenization across diverse objectives. Extensive experiments on standard video generation benchmarks confirm our analysis, showing that our adaptive approach matches or outperforms fixed‐rate baselines and naive taildrop while using fewer tokens.



Paperid:2293
Authors:Zhifang Zhang, Shuo He, Haobo Wang, Bingquan Shen, Lei Feng
Abstract:
Multimodal contrastive learning models (e.g., CLIP) can learn high-quality representations from large-scale image-text datasets, while they exhibit significant vulnerabilities to backdoor attacks, raising serious safety concerns. In this paper, we reveal that CLIP's vulnerabilities primarily stem from its tendency to encode features beyond in-dataset predictive patterns, compromising its visual feature resistivity to input perturbations. This makes its encoded features highly susceptible to being reshaped by backdoor triggers. To address this challenge, we propose Repulsive Visual Prompt Tuning (RVPT), a novel defense approach that employs deep visual prompt tuning with a specially designed feature-repelling loss. Specifically, RVPT adversarially repels the encoded features from deeper layers while optimizing the standard cross-entropy loss, ensuring that only predictive features in downstream tasks are encoded, thereby enhancing CLIP’s visual feature resistivity against input perturbations and mitigating its susceptibility to backdoor attacks. Unlike existing multimodal backdoor defense methods that typically require the availability of poisoned data or involve fine-tuning the entire model, RVPT leverages few-shot downstream clean samples and only tunes a small number of parameters. Empirical results demonstrate that RVPT tunes only 0.27\% of the parameters in CLIP, yet it significantly outperforms state-of-the-art defense methods, reducing the attack success rate from 89.70\% to 2.76\% against the most advanced multimodal attacks on ImageNet and effectively generalizes its defensive capabilities across multiple datasets. Our code is available on https://anonymous.4open.science/r/rvpt-anonymous.



Paperid:2294
Authors:Farnam Mansouri, Shai Ben-David
Abstract:
PU (Positive Unlabeled) learning is a variant of supervised classification learning in which the only labels revealed to the learner are of positively labeled instances. PU learning arises in many real-world applications. Most existing work relies on the simplifying assumption that the positively labeled training data is drawn from the restriction of the data generating distribution to positively labeled instances and/or that the proportion of positively labeled points (a.k.a. the class prior) is known apriori to the learner. This paper provides a theoretical analysis of the statistical complexity of PU learning under a wider range of setups. Unlike most prior work, our study does not assume that the class prior is known to the learner. We prove upper and lower bounds on the required sample sizes (of both the positively labeled and the unlabeled samples).



Paperid:2295
Authors:Youjia Zhang, Youngeun Kim, Young-Geun Choi, Hongyeob Kim, Huiling Liu, Sungeun Hong
Abstract:
Test-time adaptation (TTA) enhances the zero-shot robustness under distribution shifts by leveraging unlabeled test data during inference. Despite notable advances, several challenges still limit its broader applicability. First, most methods rely on backpropagation or iterative optimization, which limits scalability and hinders real-time deployment. Second, they lack explicit modeling of class-conditional feature distributions. This modeling is crucial for producing reliable decision boundaries and calibrated predictions, but it remains underexplored due to the lack of both source data and supervision at test time. In this paper, we propose ADAPT, an Advanced Distribution-Aware and backPropagation-free Test-time adaptation method. We reframe TTA as a probabilistic inference task by modeling class-conditional likelihoods using gradually updated class means and a shared covariance matrix. This enables closed-form, training-free inference. To correct potential likelihood bias, we introduce lightweight regularization guided by CLIP priors and a historical knowledge bank. ADAPT requires no source data, no gradient updates, and no full access to target data, supporting both online and transductive settings. Extensive experiments across diverse benchmarks demonstrate that our method achieves state-of-the-art performance under a wide range of distribution shifts with superior scalability and robustness.



Paperid:2296
Authors:Haocheng Luo, Mehrtash Harandi, Dinh Phung, Trung Le
Abstract:
Sharpness‐aware minimization (SAM) has emerged as a highly effective technique for improving model generalization, but its underlying principles are not fully understood. We investigated the phenomenon known as m-sharpness, where the performance of SAM improves monotonically as the micro-batch size for computing perturbations decreases. Leveraging an extended Stochastic Differential Equation (SDE) framework, combined with an analysis of the structure of stochastic gradient noise (SGN), we precisely characterize the dynamics of various SAM variants. Our findings reveal that the stochastic noise introduced during SAM perturbations inherently induces a variance-based sharpness regularization effect. Motivated by our theoretical insights, we introduce Reweighted SAM, which employs sharpness-weighted sampling to mimic the generalization benefits of m-SAM while remaining parallelizable. Comprehensive experiments validate the effectiveness of our theoretical analysis and proposed method.



Paperid:2297
Authors:Inwoo Hwang, Yushu Pan, Elias Bareinboim
Abstract:
Understanding the predictions made by deep learning models remains a central challenge, especially in high-stakes applications. A promising approach is to equip models with the ability to answer counterfactual questions -- hypothetical ``what if?'' scenarios that go beyond the observed data and provide insight into a model reasoning. In this work, we introduce the notion of causal interpretability, which formalizes when counterfactual queries can be evaluated from a model and observational data. We analyze two common model classes -- blackbox and concept-based predictors -- and show that neither is causally interpretable in general. To address this gap, we develop a framework for building models that are causally interpretable by design. Specifically, we derive a complete graphical criterion that determines whether a given model architecture supports a given counterfactual query. This leads to a fundamental tradeoff between interpretability and predictive accuracy, which we characterize by identifying the unique maximal set of features that yields an interpretable model with maximal predictive expressiveness. Experiments corroborate the theoretical findings.



Paperid:2298
Authors:Haolei Xu, Yuchen Yan, Yongliang Shen, Wenqi Zhang, Guiyang Hou, Shengpei Jiang, Kaitao Song, Weiming Lu, Jun Xiao, Yueting Zhuang
Abstract:
Large language models (LLMs) have achieved remarkable progress on mathematical tasks through Chain-of-Thought (CoT) reasoning. However, existing mathematical CoT datasets often suffer fromThought Leapsdue to experts omitting intermediate steps, which negatively impacts model learning and generalization. We propose the CoT Thought Leap Bridge Task, which aims to automatically detect leaps and generate missing intermediate reasoning steps to restore the completeness and coherence of CoT. To facilitate this, we constructed a specialized training dataset calledScaleQM+, based on the structured ScaleQuestMath dataset, and trainedCoT-Bridgeto bridge thought leaps. Through comprehensive experiments on mathematical reasoning benchmarks, we demonstrate that models fine-tuned on bridged datasets consistently outperform those trained on original datasets, with improvements of up to +5.87\% on NuminaMath. Our approach effectively enhances distilled data (+3.02\%) and provides better starting points for reinforcement learning (+3.1\%), functioning as a plug-and-play module compatible with existing optimization techniques. Furthermore, CoT-Bridge demonstrate improved generalization to out-of-domain logical reasoning tasks, confirming that enhancing reasoning completeness yields broadly applicable benefits.



Paperid:2299
Authors:Woosung Kim, Jinho Lee, Jongmin Lee, Byung-Jun Lee
Abstract:
Multi-objective reinforcement learning (MORL) aims to optimize policies in the presence of conflicting objectives, where linear scalarization is commonly used to reduce vector-valued returns into scalar signals. While effective for certain preferences, this approach cannot capture fairness-oriented goals such as Nash social welfare or max-min fairness, which require nonlinear and non-additive trade-offs. Although several online algorithms have been proposed for specific fairness objectives, a unified approach for optimizing nonlinear welfare criteria in the offline setting—where learning must proceed from a fixed dataset—remains unexplored. In this work, we present FairDICE, the first offline MORL framework that directly optimizes nonlinear welfare objective. FairDICE leverages distribution correction estimation to jointly account for welfare maximization and distributional regularization, enabling stable and sample-efficient learning without requiring explicit preference weights or exhaustive weight search. Across multiple offline benchmarks, FairDICE demonstrates strong fairness-aware performance compared to existing baselines.



Paperid:2300
Authors:Yachao Liang, Min Yu, Gang Li, Jianguo Jiang, Fuqiang Du, Jingyuan Li, Lanchi Xie, Zhen Xu, Weiqing Huang
Abstract:
The rapid advancement of generative models has led to the widespread emergence of highly realistic synthetic images, making the detection of AI-generated content increasingly critical. In particular, diffusion models have recently achieved unprecedented levels of visual fidelity, further raising concerns. While most existing approaches rely on supervised learning, zero-shot detection methods have attracted growing interest due to their ability to bypass data collection and maintenance. Nevertheless, the performance of current zero-shot methods remains limited. In this paper, we introduce a novel zero-shot AI-generated image detection method. Unlike previous works that primarily focus on identifying artifacts in the final generated images, our work explores features within the image generation process that can be leveraged for detection. Specifically, we simulate the image sampling process via diffusion-based inversion and observe that the denoising outputs of generated images converge to the target image more rapidly than those of real images. Inspired by this observation, we compute the similarity between the original image and the outputs along the denoising trajectory, which is then used as an indicator of image authenticity.Since our method requires no training on any generated images, it avoids overfitting to specific generative models or dataset biases. Experiments across a wide range of generators demonstrate that our method achieves significant improvements over state-of-the-art supervised and zero-shot counterparts.



Paperid:2301
Authors:Muquan Li, Hang Gou, Dongyang Zhang, Shuang Liang, Xiurui Xie, Deqiang Ouyang, Ke Qin
Abstract:
The growing demand for efficient deep learning has positioned dataset distillation as a pivotal technique for compressing training dataset while preserving model performance. However, existing inner-loop optimization methods for dataset distillation typically rely on random truncation strategies, which lack flexibility and often yield suboptimal results. In this work, we observe that neural networks exhibit distinct learning dynamics across different training stages—early, middle, and late—making random truncation ineffective. To address this limitation, we propose Automatic Truncated Backpropagation Through Time (AT-BPTT), a novel framework that dynamically adapts both truncation positions and window sizes according to intrinsic gradient behavior. AT-BPTT introduces three key components: (1) a probabilistic mechanism for stage-aware timestep selection, (2) an adaptive window sizing strategy based on gradient variation, and (3) a low-rank Hessian approximation to reduce computational overhead. Extensive experiments on CIFAR-10, CIFAR-100, Tiny-ImageNet, and ImageNet-1K show that AT-BPTT achieves state-of-the-art performance, improving accuracy by an average of 6.16\% over baseline methods. Moreover, our approach accelerates inner-loop optimization by 3.9 × while saving 63\% memory cost.



Paperid:2302
Authors:Benjamin Dupuis, Dario Shariatian, Maxime Haddouche, Alain Durmus, Umut Simsekli
Abstract:
Score-based generative models (SGMs) have emerged as one of the most popular classes of generative models. A substantial body of work now exists on the analysis of SGMs, focusing either on discretization aspects or on their statistical performance. In the latter case, bounds have been derived, under various metrics, between the true data distribution and the distribution induced by the SGM, often demonstrating polynomial convergence rates with respect to the number of training samples. However, these approaches adopt a largely approximation theory viewpoint, which tends to be overly pessimistic and relatively coarse. In particular, they fail to fully explain the empirical success of SGMs or capture the role of the optimization algorithm used in practice to train the score network. To support this observation, we first present simple experiments illustrating the concrete impact of optimization hyperparameters on the generalization ability of the generated distribution. Then, this paper aims to bridge this theoretical gap by providing the first algorithmic- and data-dependent generalization analysis for SGMs. In particular, we establish bounds that explicitly account for the optimization dynamics of the learning algorithm, offering new insights into the generalization behavior of SGMs. Our theoretical findings are supported by empirical results on several datasets.



Paperid:2303
Authors:Yongsen Mao, Junhao Zhong, Chuan Fang, Jia Zheng, Rui Tang, Hao Zhu, Ping Tan, Zihan Zhou
Abstract:
SpatialLM is a large language model designed to process 3D point cloud data and generate structured 3D scene understanding outputs. These outputs include architectural elements like walls, doors, windows, and oriented object boxes with their semantic categories. Unlike previous methods which exploit task-specific network designs, our model adheres to the standard multimodal LLM architecture and is fine-tuned directly from open-source LLMs.To train SpatialLM, we collect a large-scale, high-quality synthetic dataset consisting of 12,328 indoor scenes with ground-truth 3D annotations and photo-realistic RGBD scans, and conduct a careful study on various modeling and training decisions. On public benchmarks, our model gives state-of-the-art performance in layout estimation and competitive results in 3D object detection. With that, we show a feasible path for enhancing the spatial understanding capabilities of modern LLMs for applications in augmented reality, embodied robotics, and more.



Paperid:2304
Authors:HAOTIAN XU, Qingsong Peng, Jie Shi, Huadi Zheng, YU LI, Cheng Zhuo
Abstract:
The rapid adoption of large language models (LLMs) in critical domains has spurred extensive research into their security issues. While input manipulation attacks (e.g., prompt injection) have been well-studied, Bit-Flip Attacks (BFAs)—which exploit hardware vulnerabilities to corrupt model parameters and cause severe performance degradation—have received far less attention. Existing BFA methods suffer from key limitations: they fail to balance performance degradation and output naturalness, making them prone to discovery. In this paper, we introduce SilentStriker, the first stealthy bit-flip attack against LLMs that effectively degrades task performance while maintaining output naturalness. Our core contribution lies in addressing the challenge of designing effective loss functions for LLMs with variable output length and the vast output space. Unlike prior approaches that rely on output perplexity for attack loss formulation, which in-evidently degrade the output naturalness, we reformulate the attack objective by leveraging key output tokens as targets for suppression, enabling effective joint optimization of attack effectiveness and stealthiness. Additionally, we employ an iterative, progressive search strategy to maximize attack efficacy. Experiments show that SilentStriker significantly outperforms existing baselines, achieving successful attacks without compromising the naturalness of generated text.



Paperid:2305
Authors:Yuyang Deng, Samory Kpotufe
Abstract:
Abstract:Theoretical works on supervised transfer learning (STL)---where the learner has access to labeled samples from both source and target distributions---have for the most part focused on statistical aspects of the problem, while efficient optimization has received less attention. We consider the problem of designing an SGD procedure for STL that alternates sampling between source and target data, while maintaining statistical transfer guarantees without prior knowledge of the quality of the source data. A main algorithmic difficulty is in understanding how to design such an adaptive sub-sampling mechanism at each SGD step, to automatically gain from the source when it is informative, or bias towards the target and avoid negative transfer when the source is less informative. We show that, such a mixed-sample SGD procedure is feasible for general prediction tasks with convex losses, rooted in tracking an abstract sequence of constrained convex programs that serve to maintain the desired transfer guarantees. We instantiate these results in the concrete setting of linear regression with square loss, and show that the procedure converges, with $1/\sqrt{T}$ rate, to a solution whose statistical performance on the target is adaptive to the a priori unknown quality of the source. Experiments wirh synthetic and real datasets support the theory.



Paperid:2306
Authors:Haolong Qian, Xianliang Yang, Ling Zhang, Lei Song, Jiang Bian, Chun Yuan
Abstract:
Group Relative Policy Optimization (GRPO) fine-tuning has been empirically shown to significantly enhance the reasoning abilities of language models. However, it often relies on large-scale, high-quality labeled data, which is typically difficult to obtain. To address this challenge, we introduce the Noise-Aware Dual-Reward Optimization (NaDRO) , which effectively enhances LLMs training in environments where data is noisy or imperfect. NaDRO operates through two key components: \textbf{(1) Preference-based Outcome Reward (POR)}, which extracts reliable preference signals from noisy data, guiding LLMs towards more effective decisions instead of relying on specific noisy scores; and \textbf{(2) a Context Perception Reward (CPR) mechanism}, which ensures that LLMs conduct necessary qualitative assessment of the current problem state, rewarding accurate judgments to foster better cognitive understanding before decision-making. In the context of combinatorial optimization problems, where dynamically selecting heuristic algorithms is challenging due to large problem scales and the difficulty of obtaining accurate decision data, we designed experiments to test our approach. Our results indicate that the fine-tuned Qwen 7B and Llama 3-8B models outperform mainstream large language models (LLMs) training in this task. Code is released at \url{https://anonymous.4open.science/r/NaDRO-D34D}



Paperid:2307
Authors:Shuangyi Chen, Yuanxin Guo, Yue Ju, Hardik Dalal, Zhongwen Zhu, Ashish Khisti
Abstract:
Parameter-Efficient Fine-Tuning (PEFT) methods like Low-Rank Adaptation (LoRA) optimize federated training by reducing computational and communication costs. We propose RoLoRA, a federated framework using alternating optimization to fine-tune LoRA adapters. Our approach emphasizes the importance of learning up and down projection matrices to enhance expressiveness and robustness. We use both theoretical analysis and extensive experiments to demonstrate the advantages of RoLoRA over prior approaches that either generate imperfect model updates or limit expressiveness of the model. We provide a theoretical analysis on a linear model to highlight the importance of learning both the down-projection and up-projection matrices in LoRA. We validate the insights on a non-linear model and separately provide a convergence proof under general conditions. To bridge theory and practice, we conducted extensive experimental evaluations on language models including RoBERTa-Large, Llama-2-7B on diverse tasks and FL settings to demonstrate the advantages of RoLoRA over other methods.



Paperid:2308
Authors:Xingbo Du, Ruizhe Zhong, Junchi Yan
Abstract:
The Rectilinear Steiner Minimum Tree (RSMT) is widely used in Very Large Scale Integration (VLSI) and aims at connecting a set of pins using rectilinear edges while minimizing wirelength. Recently, learning-based methods have been explored to tackle this problem effectively. However, existing methods either suffer from excessive exploration of the search space or rely on heuristic combinations that compromise effectiveness and efficiency, and this limitation becomes notably exacerbated when extended to the obstacle-avoiding RSMT (OARSMT). To address this, we propose OAREST, a reinforcement learning-based framework for constructing an Obstacle-Avoiding Rectilinear Edge Sequence (RES) Tree. We theoretically establish the optimality of RES in obstacle-avoiding scenarios, which forms the foundation of our approach. Leveraging this theoretical insight, we introduce a dynamic masking strategy that supports parallel training across varying numbers of pins and extends to obstacles during inference. Empirical evaluations on both synthetic and real-world benchmarks show superior effectiveness and efficiency for RSMT and OARSMT problems, particularly in handling obstacles without training on them. Code available: \url{https://anonymous.4open.science/r/OAREST-D386/}.



Paperid:2309
Authors:Jiaqi Xue, Mayank Kumar, Yuzhang Shang, Shangqian Gao, Rui Ning, Mengxin Zheng, Xiaoqian Jiang, Qian Lou
Abstract:
Abstract:Federated learning (FL) enables institutions to collaboratively train machine learning models by aggregating local gradients without sharing sensitive data. However, sharing gradients still poses privacy risks, e.g., gradient inversion attacks. Homomorphic encryption (HE) can be used in FL to encrypt gradients at the data owner's side, enabling secure aggregation without decryption on the server. Existing HE approaches to FL lie at two extremes. One encrypts every gradient update, providing strong privacy but incurring prohibitive computation and bandwidth costs. The other encrypts only a subset of gradients, reducing overhead yet leaving the remaining plaintext updates vulnerable to privacy attacks. Our proposed DictPFL bridges this gap. It encrypts every gradient that must be transmitted to the server—protecting all shared information—while keeping the rest of the (unencrypted) gradients on the client, where they never leave the device. By safeguarding every transmitted update, DictPFL achieves the same privacy guarantees as fully encrypted FL, but its selective-encryption strategy slashes computational and communication overhead. DictPFL comprises two modules: Decompose-for-Partial-Encrypt (DePE) and Prune-for-Minimum-Encrypt (PrME). In DePE, we decompose model weights to be trained into a dictionary and a lookup table. Only the gradients of the lookup table are encrypted and aggregated securely while the dictionary remains fixed and is not transmitted for aggregation. In PrME, we aim to further minimize the encrypted parameters with an encryption-aware pruning technique that ensures a consistent pruning mask across clients by leveraging the history of global gradients. Experimental results demonstrate that DictPFL significantly reduces communication overhead by 402 to 748 times and speeds training by 28 to 65 times compared to fully encrypted method. It also outperforms the state-of-the-art selectively encrypted gradient by lowering overhead by 51 to 155 times and accelerating training by 4 to 19 times. DictPFL increases training time by even less than a 2$\times$ factor compared with its plaintext counterpart without gradients protection, demonstrating—for the first time—that HE–based private federated learning is practical for real-world deployment.



Paperid:2310
Authors:Ruilong Yu, Mingyan Liu, Fei Ye, Adrian G. Bors, Rongyao Hu, Jingling sun, shijie zhou
Abstract:
Extant studies predominantly address catastrophic forgetting within a simplified continual learning paradigm, typically confined to a singular data domain. Conversely, real-world applications frequently encompass multiple, evolving data domains, wherein models often struggle to retain many critical past information, thereby leading to performance degradation. This paper addresses this complex scenario by introducing a novel dynamic expansion approach called Learning Expandable and Adaptable Representations (LEAR). This framework orchestrates a collaborative backbone structure, comprising global and local backbones, designed to capture both general and task-specific representations. Leveraging this collaborative backbone, the proposed framework dynamically create a lightweight expert to delineate decision boundaries for each novel task, thereby facilitating the prediction process. To enhance new task learning, we introduce a novel Mutual Information-Based Prediction Alignment approach, which incrementally optimizes the global backbone via a mutual information metric, ensuring consistency in the prediction patterns of historical experts throughout the optimization phase. To mitigate network forgetting, we propose a Kullback–Leibler (KL) Divergence-Based Feature Alignment approach, which employs a probabilistic distance measure to prevent significant shifts in critical local representations. Furthermore, we introduce a novel Hilbert-Schmidt Independence Criterion (HSIC)-Based Collaborative Optimization approach, which encourages the local and global backbones to capture distinct semantic information in a collaborative manner, thereby mitigating information redundancy and enhancing model performance. Moreover, to accelerate new task learning, we propose a novel Expert Selection Mechanism that automatically identifies the most relevant expert based on data characteristics. This selected expert is then utilized to initialize a new expert, thereby fostering positive knowledge transfer. This approach also enables expert selection during the testing phase without requring any task information. Empirical results demonstrate that the proposed framework achieves state-of-the-art performance.



Paperid:2311
Authors:Faisal Hamman, Pasan Dissanayake, Yanjun Fu, Sanghamitra Dutta
Abstract:
Knowledge distillation is a promising approach to transfer capabilities from resource-intensive teacher models to smaller, resource-efficient student models that can be deployed easily, particularly in task-aware scenarios. However, existing methods of task-aware distillation typically require substantial quantities of data which may be unavailable or expensive to obtain in many practical scenarios. In this paper, we address this challenge by introducing a novel strategy called CoD forfew-shot task-aware knowledge distillation by systematically infusing counterfactual explanations. Counterfactual explanations (CFE) refer to inputs that can flip the output prediction of the teacher model with minimum input perturbation. Our strategy CoD, short forCounterfactual-explanation-infusedDistillation leverages these CFEs to precisely map the teacher's decision boundary with significantly fewer samples. We provide theoretical guarantees for motivating the role of CFEs in distillation, from both statistical and geometric perspectives. We mathematically show that CFEs can improve parameter estimation by providing more informative examples near the teacher’s decision boundary. We also derive geometric insights on how CFEs effectively act asknowledge probes, helping the students mimic the teacher's decision boundaries more effectively than standard data. We perform experiments across various datasets and LLMs to show that CoD outperforms standard distillation approaches in few-shot regimes (8 - 512 samples), achieving improved performance under equal number of shots which is essentially half of the original samples used by the baselines, infused with their corresponding CFEs.



Paperid:2312
Authors:Chenrui Cao, Liangcheng Song, Zenan Li, Xinyi Le, Xian Zhang, HUI XUE, Fan Yang
Abstract:
Recent advancements, such as DeepSeek-Prover-V2-671B and Kimina-Prover-Preview-72B, demonstrate a prevailing trend in leveraging reinforcement learning (RL)-based large-scale training for automated theorem proving. Surprisingly, we discover that even without any training, careful neuro-symbolic coordination of existing off-the-shelf reasoning models and tactic step provers can achieve comparable performance. This paper introduces \textbf{DSP+}, an improved version of the Draft, Sketch, and Prove framework, featuring a \emph{fine-grained and integrated} neuro-symbolic enhancement for each phase: (1) In the draft phase, we prompt reasoning models to generate concise natural-language subgoals to benefit the sketch phase, removing thinking tokens and references to human-written proofs; (2) In the sketch phase, subgoals are autoformalized with hypotheses to benefit the proving phase, and sketch lines containing syntactic errors are masked according to predefined rules; (3) In the proving phase, we tightly integrate symbolic search methods like Aesop with step provers to establish proofs for the sketch subgoals. Experimental results show that, without any additional model training or fine-tuning, DSP+ solves 80.7\%, 32.8\%, and 24 out of 644 problems from miniF2F, ProofNet, and PutnamBench, respectively, while requiring fewer budgets compared to state-of-the-arts. DSP+ proves \texttt{imo_2019_p1}, an IMO problem in miniF2F that is not solved by any prior work. Additionally, DSP+ generates proof patterns comprehensible by human experts, facilitating the identification of formalization errors; For example, eight wrongly formalized statements in miniF2F are discovered. Our results highlight the potential of classical reasoning patterns besides the RL-based training. All components will be open-sourced.



Paperid:2313
Authors:Guojun Chen, Sicheng Xu, Jiaolong Yang, Yizhong Zhang, Yu Deng, Stephen Lin, Baining Guo
Abstract:
We propose VASA-3D, an audio-driven, single-shot 3D head avatar generator. This research tackles two major challenges: capturing the subtle expression details present in real human faces, and reconstructing an intricate 3D head avatar from a single portrait image. To accurately model expression details, VASA-3D leverages the motion latent of VASA-1, a method that yields exceptional realism and vividness in 2D talking heads. A critical element of our work is translating this motion latent to 3D, which is accomplished by devising a 3D head model that is conditioned on the motion latent. Customization of this model to a single image is achieved through an optimization framework that employs numerous video frames of the reference head synthesized from the input image. The optimization takes various training losses robust to artifacts and limited pose coverage in the generated training data. Our experiment shows that VASA-3D produces realistic 3D talking heads that can not be achieved by prior art, and it supports the online generation of 512x512 free-viewpoint videos at up to 75 FPS, facilitating more immersive engagements with lifelike 3D avatars.



Paperid:2314
Authors:Anass El Aouni, Quentin Gaudel, J. Emmanuel Johnson, REGNIER Charly, Julien Le Sommer, Simon van Gennip, ronan fablet, Marie Drevillon, Yann DRILLET, Pierre Le Traon
Abstract:
Data-driven approaches, particularly those based on deep learning, are rapidly advancing Earth system modeling. However, their application to ocean forecasting remains limited despite the ocean's pivotal role in climate regulation and marine ecosystems. To address this gap, we present OceanBench, a benchmark designed to evaluate and accelerate global short-range (1–10 days) data-driven ocean forecasting.OceanBench is constructed from a curated dataset comprising first-guess trajectories, nowcasts, and atmospheric forcings from operational physical ocean models, typically unavailable in public datasets due to assimilation cycles. Matched observational data are also included, enabling realistic evaluation in an operational-like forecasting framework.The benchmark defines three complementary evaluation tracks: (i) Model-to-Reanalysis, where models are compared against the reanalysis dataset commonly used for training; (ii) Model-to-Analysis, assessing generalization to a higher-resolution physical analysis; and (iii) Model-to-Observations, Intercomparison and Validation (IV-TT) Class-4 evaluation against independent observational data. The first two tracks are further supported by process-oriented diagnostics to assess the dynamical consistency and physical plausibility of forecasts.OceanBench includes key ocean variables: sea surface height, temperature, salinity, and currents, along with standardized metrics grounded in physical oceanography. Baseline comparisons with operational systems and state-of-the-art deep learning models are provided. All data, code, and evaluation protocols are openly available at \url{https://github.com/mercator-ocean/oceanbench}, establishing OceanBench as a foundation for reproducible and rigorous research in data-driven ocean forecasting.



Paperid:2315
Authors:Steven Kolawole, Keshav Santhanam, Virginia Smith, Pratiksha Thaker
Abstract:
Abstract:LLM serving systems typically treat user prompts as monolithic inputs, optimizing inference through decoding tricks or inter-query batching. However, many real-world prompts contain *latent semantic parallelism*—decomposable structures where subtasks can be executed independently to reduce latency while preserving meaning.We introduce PARALLELPROMPT, the first benchmark for measuring intra-query parallelism in natural user prompts. Our dataset comprises over 37,000 real-world prompts from public LLM chat logs, each annotated with a structured schema capturing task templates, shared context, and iteration inputs. These schemas are extracted using LLM-assisted prompting with rule-based multilingual validation.To evaluate the benefits of decomposition, we provide an execution suite that benchmarks serial vs. parallel strategies, measuring latency, structural adherence, and semantic fidelity. Our results show that intra-query parallelism can be successfully parsed in over 75\% of curated datasets, unlocking up to *$5\times$ speedups* on tasks like translation, comprehension, and comparative analysis, with minimal quality degradation.By releasing this benchmark, curation pipeline, and evaluation suite, we provide the first standardized testbed for studying structure-aware execution in LLM serving pipelines.



Paperid:2316
Authors:Yanxi Chen, Xuchen Pan, Yaliang Li, Bolin Ding, Jingren Zhou
Abstract:
We propose two simple, principled and practical algorithms that enjoy provable scaling laws for the test-time compute of large language models (LLMs). The first one is a two-stage knockout-style algorithm: given an input problem, it first generates multiple candidate solutions, and then aggregate them via a knockout tournament for the final output. Assuming that the LLM can generate a correct solution with non-zero probability and do better than a random guess in comparing a pair of correct and incorrect solutions, we prove theoretically that the failure probability of this algorithm decays to zero exponentially or by a power law (depending on the specific way of scaling) as its test-time compute grows. The second one is a two-stage league-style algorithm, where each candidate is evaluated by its average win rate against multiple opponents, rather than eliminated upon loss to a single opponent. Under analogous but more robust assumptions, we prove that its failure probability also decays to zero exponentially with more test-time compute. Both algorithms require a black-box LLM and nothing else (e.g., no verifier or reward model) for a minimalistic implementation, which makes them appealing for practical applications and easy to adapt for different tasks. Through extensive experiments with diverse models and datasets, we validate the proposed theories and demonstrate the outstanding scaling properties of both algorithms.



Paperid:2317
Authors:Ruaridh Mon-Williams, Max Taylor-Davies, Elizabeth Mieczkowski, Natalia Vélez, Neil Bramley, Yanwei Wang, Tom Griffiths, Christopher G Lucas
Abstract:
Humans are remarkably adept at collaboration, able to infer the strengths and weaknesses of new partners in order to work successfully towards shared goals. To build AI systems with this capability, we must first understand its building blocks: does such flexibility require explicit, dedicated mechanisms for modelling others — or can it emerge spontaneously from the pressures of open-ended cooperative interaction? To investigate this question, we train simple model-free RNN agents to collaborate with a population of diverse partners. Using the 'Overcooked-AI' environment, we collect data from thousands of collaborative teams, and analyse agents' internal hidden states. Despite a lack of additional architectural features, inductive biases, or auxiliary objectives, the agents nevertheless develop structured internal representations of their partners' task abilities, enabling rapid adaptation and generalisation to novel collaborators. We investigated these internal models through probing techniques, and large-scale behavioural analysis. Notably, we find that structured partner modelling emerges when agents can influence partner behaviour by controlling task allocation. Our results show that partner modelling can arise spontaneously in model-free agents — but only under environmental conditions that impose the right kind of social pressure.



Paperid:2318
Authors:Finn Rietz, Oleg Smirnov, Sara Karimi, Lele Cao
Abstract:
Prompt tuning has emerged as a key technique for adapting large pre-trained Decision Transformers (DTs) in offline Reinforcement Learning (RL), particularly in multi-task and few-shot settings. The Prompting Decision Transformer (PDT) enables task generalization via trajectory prompts sampled uniformly from expert demonstrations -- without accounting for prompt informativeness. In this work, we propose a bandit-based prompt-tuning method that learns to construct optimal trajectory prompts from demonstration data at inference time. We devise a structured bandit architecture operating in the trajectory prompt space, achieving linear rather than combinatorial scaling with prompt size. Additionally, we show that the pre-trained PDT itself can serve as a powerful feature extractor for the bandit, enabling efficient reward modeling across various environments. We theoretically establish regret bounds and demonstrate empirically that our method consistently enhances performance across a wide range of tasks, high-dimensional environments, and out-of-distribution scenarios, outperforming existing baselines in prompt tuning.



Paperid:2319
Authors:Yuhan Yao, Feifei Kou, Lei Shi, Xiao yang, Zhongbao Zhang, Suguo Zhu, Jiwei Zhang, Lirong Qiu, LI Haisheng
Abstract:
The advancements in generative models and the real-world attack of machine-generated text(MGT) create a demand for more robust detection methods. The existing MGT detection methods for adversarial environments primarily consist of manually designed statistical-based methods and fine-tuned classifier-based approaches. Statistical-based methods extract intrinsic features but suffer from rigid decision boundaries vulnerable to adaptive attacks, while fine-tuned classifiers achieve outstanding performance at the cost of overfitting to superficial textual feature. We argue that the key to detection in current adversarial environments lies in how to extract intrinsic invariant features and ensure that the classifier possesses dynamic adaptability. In that case, we propose OSTAR, a novel MGT detection framework designed for adversarial environments which composed of a statistical enhanced classifier and a Multi-Faceted Contrastive Learning(MFCL). In the classifier aspect, our Multi-Dimensional Statistical Profiling (MDSP) module extracts intrinsic difference between human and machine texts, complementing classifiers with useful stable features. In the model optimization aspect, the MFCL strategy enhances robustness by contrasting feature variations before and after text attacks, jointly optimizing statistical feature mapping and baseline pre-trained models. Experimental results on three public datasets under various adversarial scenarios demonstrate that our framework outperforms existing MGT detection methods, achieving state-of-the-art performance and robust against attacks.



Paperid:2320
Authors:meng wang, Fan Wu, Ruihui Li, Qin Yunchuan, Zhuo Tang, Li Ken Li
Abstract:
3D Semantic Scene Completion (SSC) provides comprehensive scene geometry and semantics for autonomous driving perception, which is crucial for enabling accurate and reliable decision-making. However, existing SSC methods are limited to capturing sparse information from the current frame or naively stacking multi-frame temporal features, thereby failing to acquire effective scene context. These approaches ignore critical motion dynamics and struggle to achieve temporal consistency. To address the above challenges, we propose a novel temporal SSC method FlowScene: Learning Temporal 3D Semantic Scene Completion via Optical Flow Guidance. By leveraging optical flow, FlowScene can integrate motion, different viewpoints, occlusions, and other contextual cues, thereby significantly improving the accuracy of 3D scene completion. Specifically, our framework introduces two key components: (1) a Flow-Guided Temporal Aggregation module that aligns and aggregates temporal features using optical flow, capturing motion-aware context and deformable structures; and (2) an Occlusion-Guided Voxel Refinement module that injects occlusion masks and temporally aggregated features into 3D voxel space, adaptively refining voxel representations for explicit geometric modeling.Experimental results demonstrate that FlowScene achieves state-of-the-art performance, with mIoU of 17.70 and 20.81 on the SemanticKITTI and SSCBench-KITTI-360 benchmarks. The source code will be released upon acceptance.



Paperid:2321
Authors:Shuo Shuo Liu, Haotian Lin, Matthew Reimherr, Runze Li
Abstract:
Most existing transfer learning algorithms for high-dimensional models employ a two-step regularization framework, whose success heavily hinges on the assumption that the pre-trained model closely resembles the target. To address this challenge, we propose a co-regularization process to directly exploit beneficial knowledge from the source domain for high-dimensional generalized linear models. The proposed method learns the target parameter by constraining the source parameters to be close to the target one, thereby preventing fine-tuning failures caused by significantly deviated pre-trained parameters. Our theoretical analysis demonstrates that the proposed method accommodates a broader range of sources than existing two-step frameworks, thus being more robust to less similar sources. Its effectiveness is validated through extensive empirical studies.



Paperid:2322
Authors:Boyang Zhang, Zhiguo Wang, Ya-Feng Liu
Abstract:
Chance constrained programming (CCP) is a powerful framework for addressing optimization problems under uncertainty. In this paper, we introduce a novel \textbf{G}radient-\textbf{G}uided \textbf{D}iffusion-based \textbf{Opt}imization framework, termed GGDOpt, which tackles CCP through three key innovations. First, GGDOpt accommodates a broad class of CCP problems without requiring the knowledge of the exact distribution of uncertainty—relying solely on a set of samples. Second, to address the nonconvexity of the chance constraints, it reformulates the CCP as a sampling problem over the product of two distributions: an unknown data distribution supported on a nonconvex set and a Boltzmann distribution defined by the objective function, which fully leverages both first- and second-order gradient information. Third, GGDOpt has theoretical convergence guarantees and provides practical error bounds under mild assumptions. By progressively injecting noise during the forward diffusion process to convexify the nonconvex feasible region, GGDOpt enables guided reverse sampling to generate asymptotically optimal solutions. Experimental results on synthetic datasets and a waveform design task in wireless communications demonstrate that GGDOpt outperforms existing methods in both solution quality and stability with nearly 80\% overhead reduction.



Paperid:2323
Authors:Hai Duong, David Shriver, ThanhVu Nguyen, Matthew Dwyer
Abstract:
Verifying the behavior of neural networks is necessary if developers are to confidently deploy them as parts of mission-critical systems.Toward this end, researchers have been actively developing a range of increasingly sophisticated and scalable neural network verifiers.However, scaling verification to large networks is challenging, at least in part due to the significant memory requirements of verification algorithms. In this paper, we propose an assume-guarantee compositional framework, CoVeNN, that is parameterized by an underlying verifier to generate a sequence of verification sub-problems to address this challenge. We present an iterative refinement-based strategy for computing assumptions that allow sub-problems to retain sufficient accuracy. An evaluation using 7 neural networks and a total of 140 property specifications demonstrates that CoVeNN can verify nearly 7 times more problems than state-of-the-art verifiers. CoVeNN is available at: https://anonymous.4open.science/r/CoVeNN-8FD0.



Paperid:2324
Authors:Zhihao Huang, Xi Qiu, Yukuo Ma, Yifu Zhou, Junjie Chen, Hongyuan Zhang, Chi Zhang, Xuelong Li
Abstract:
Autoregressive models have achieved significant success in image generation. However, unlike the inherent hierarchical structure of image information in the spectral domain, standard autoregressive methods typically generate pixels sequentially in a fixed spatial order. To better leverage this spectral hierarchy, we introduce Next-Frequency Image Generation (NFIG). NFIG is a novel framework that decomposes the image generation process into multiple frequency-guided stages. NFIG aligns the generation process with the natural image structure. It does this by first generating low-frequency components, which efficiently capture global structure with significantly fewer tokens, and then progressively adding higher-frequency details. This frequency-aware paradigm offers substantial advantages: it not only improves the quality of generated images but crucially reduces inference cost by efficiently establishing global structure early on. Extensive experiments on the ImageNet-256 benchmark validate NFIG's effectiveness, demonstrating superior performance (FID: 2.81) and a notable 1.25x speedup compared to the strong baseline VAR-d20.



Paperid:2325
Authors:Dongmin Choi, Sangbin Lee, EungGu Yun, Jonghyuk Baek, Frank Park
Abstract:
The performance of learning-based object detection algorithms, which attempt to both classify and locate objects within images, is determined largely by the quality of the annotated dataset used for training. Two types of labelling noises are prevalent: objects that are incorrectly classified (categorization noise) and inaccurate bounding boxes (localization noise); both noises typically occur together in large-scale datasets. In this paper we propose a distillation-based method to train object detectors that takes into account both categorization and localization noise. The key insight underpinning our method is that the early-learning phenomenon - in which models trained on noisy data with mixed clean and false labels tend to first fit to the clean data, and memorize the false labels later -- manifests earlier for localization noise than for categorization noise. We propose a method that uses models from the early-learning phase (before overfitting to noisy data occurs) as a teacher network. A plug-in module implementation compatible with general object detection architectures is developed, and its performance is validated against the state-of-the-art using PASCAL VOC, MS COCO and VinDr-CXR medical detection datasets.



Paperid:2326
Authors:Jeffrey Alido, Tongyu Li, Yu Sun, Lei Tian
Abstract:
Abstract:Conventional score-based diffusion models (DMs) may struggle with anisotropic Gaussian diffusion processes due to the required inversion of covariance matrices in the denoising score matching training objective \cite{vincent_connection_2011}. We propose Whitened Score (WS) diffusion models, a novel SDE-based framework that learns the Whitened Score function instead of the standard score. This approach circumvents covariance inversion, extending score-based DMs by enabling stable training of DMs on arbitrary Gaussian forward noising processes. WS DMs establish equivalence with FM for arbitrary Gaussian noise, allow for tailored spectral inductive biases, and provide strong Bayesian priors for imaging inverse problems with structured noise. We experiment with a variety of computational imaging tasks using the CIFAR and CelebA ($64\times64$) datasets and demonstrate that WS diffusion priors trained on anisotropic Gaussian noising processes consistently outperform conventional diffusion priors based on isotropic Gaussian noise.



Paperid:2327
Authors:Lei Liu, Xingyu Xia, Qianqian Xie, Ben Liu, Wenjie Xu, Min Peng
Abstract:
Graph foundation models (GFMs) have emerged as a promising paradigm for learning transferable knowledge across diverse graph-structured data. The inherent heterogeneity in features and graph structures poses significant challenges for building scalable and generalizable GFMs. Existing research has employed mixture-of-experts (MoE) models to handle the challenges, assigning the most suitable expert to each graph. Despite this, the underlying mechanisms of MoE within the context of GFMs remain insufficiently explored. In this work, we conduct an in-depth experimental study on an MoE-based GFM and uncover an intriguing finding: the experts ranked second and third assigned by the router perform better than the top-ranked expert. This insight motivates us to investigate the potential of leveraging knowledge embedded across multiple experts. However, directly ensembling the outputs of multiple experts would incur substantial computational overhead, while applying a standard expert merging strategy risks suboptimal performance. To address these challenges, we introduce two improved expert merging strategies that retain the computational efficiency of expert merging, while improving performance to approach the effectiveness of expert ensembling. Specifically, we propose (i) a knowledge distillation-inspired expert merging method that aligns the behavior of parameter-fused experts with expert ensembles, and (ii) a theoretical parameter proximity approach that leverages the similarity of expert parameters to approximate ensemble outputs while preserving diversity. Extensive experiments demonstrate that our methods effectively enhance model performance.



Paperid:2328
Authors:Zhangyu Wang, Zeping Liu, Jielu Zhang, Zhongliang Zhou, Qian Cao, Nemin Wu, Lan Mu, Yang Song, Yiqun Xie, Ni Lao, Gengchen Mai
Abstract:
Image geolocalization is a fundamental yet challenging task, aiming at inferring the geolocation on Earth where an image is taken. State-of-the-art methods employ either grid-based classification or gallery-based image-location retrieval, whose spatial generalizability significantly suffers if the spatial distribution of test images does not align with the choices of grids and galleries. Recently emerging generative approaches, while getting rid of grids and galleries, use raw geographical coordinates and suffer quality losses due to their lack of multi-scale information. To address these limitations, we propose a multi-scale latent diffusion model called LocDiff for image geolocalization. We developed a novel positional encoding-decoding framework called Spherical Harmonics Dirac Delta (SHDD) Representations, which encodes points on a spherical surface (e.g., geolocations on Earth) into a Hilbert space of Spherical Harmonics coefficients and decodes points (geolocations) by mode-seeking on spherical probability distributions. We also propose a novel SirenNet-based architecture (CS-UNet) to learn an image-based conditional backward process in the latent SHDD space by minimizing a latent KL-divergence loss. To the best of our knowledge, LocDiff is the first image geolocalization model that performs latent diffusion in a multi-scale location encoding space and generates geolocations under the guidance of images. Experimental results show that LocDiff can outperform all state-of-the-art grid-based, retrieval-based, and diffusion-based baselines across 5 challenging global-scale image geolocalization datasets, and demonstrates significantly stronger generalizability to unseen geolocations.



Paperid:2329
Authors:Omri Lev, Vishwak Srinivasan, Moshe Shenfeld, Katrina Ligett, Ayush Sekhari, Ashia Wilson
Abstract:
Gaussian sketching, which consists of pre-multiplying the data with a random Gaussian matrix, is a widely used technique for multiple problems in data science and machine learning. Beyond computational benefits, this operation also provides differential privacy guarantees due to its inherent randomness. In this work, we revisit this operation through the lens of Renyi Differential Privacy (RDP), providing a refined privacy analysis that yields significantly tighter bounds than prior results. We then demonstrate how this improved analysis leads to performance improvement in different linear regression settings, establishing theoretical utility guarantees. Empirically, our methods improve performance across multiple datasets and, in several cases, reduce runtime.



Paperid:2330
Authors:Daniel O'Malley, Manish Bhattarai, Javier E. Santos, Nishath Ranasinghe, Erick Draayer
Abstract:
We present a novel benchmark designed to rigorously evaluate the capabilities of large language models (LLMs) in mathematical reasoning and algorithmic code synthesis tasks. The benchmark comprises integer sequence generation tasks sourced from the Online Encyclopedia of Integer Sequences (OEIS), testing LLMs' abilities to accurately and efficiently generate Python code to compute these sequences without using lookup tables. Our comprehensive evaluation includes leading models from OpenAI (including the specialized reasoning-focused o-series), Anthropic, Meta, and Google across a carefully selected set of 1000 OEIS sequences categorized aseasy'' orhard.'' Half of these sequences are classical sequences from the early days of OEIS and half were recently added to avoid contamination with the models' training data. To prevent models from exploiting memorized sequence values, we introduce an automated cheating detection mechanism that flags usage of lookup tables, validated by comparison with human expert evaluations. Experimental results demonstrate that reasoning-specialized models (o3, o3-mini, o4-mini from OpenAI, and Gemini 2.5-pro from Google) achieve substantial improvements in accuracy over non-reasoning models, especially on more complex tasks. However, overall model performance on the hard sequences is poor, highlighting persistent challenges in algorithmic reasoning. Our benchmark provides important insights into the strengths and limitations of state-of-the-art LLMs, particularly emphasizing the necessity for further advancements to reliably solve complex mathematical reasoning tasks algorithmically.



Paperid:2331
Authors:Tyler Chang, Benjamin Bergen
Abstract:
In Transformer language models, activation vectors transform from current token embeddings to next token predictions as they pass through the model. To isolate a minimal form of this transformation, we identify language model subnetworks that make bigram predictions, naive next token predictions based only on the current token. We find that bigram subnetworks can be found in fully trained language models up to 1B parameters, and these subnetworks are critical for model performance even when they consist of less than 0.2% of model parameters. Bigram subnetworks are concentrated in the first Transformer MLP layer, and they overlap significantly with subnetworks trained to optimally prune a given model. Mechanistically, the bigram subnetworks often recreate a pattern from the full models where the first layer induces a sharp change that aligns activations with next token predictions rather than current token representations. Our results demonstrate that bigram subnetworks comprise a minimal subset of parameters that are both necessary and sufficient for basic next token predictions in language models, and they help drive the transformation from current to next token activations in the residual stream. These subnetworks can lay a foundation for studying more complex language model circuits by building up from a minimal circuit.



Paperid:2332
Authors:Qi Cao, Ruiyi Wang, Ruiyi Zhang, Sai Ashish Somayajula, Pengtao Xie
Abstract:
Reasoning has substantially improved the performance of large language models (LLMs) on complicated tasks. Central to the current reasoning studies, Process Reward Models (PRMs) offer a fine-grained evaluation of intermediate reasoning steps and guide the reasoning process. However, extending PRMs to multimodal large language models (MLLMs) introduces challenges. Since multimodal reasoning covers a wider range of tasks compared to text-only scenarios, the resulting distribution shift from the training to testing sets is more severe, leading to greater generalization difficulty. Training a reliable multimodal PRM, therefore, demands large and diverse datasets to ensure sufficient coverage. However, current multimodal reasoning datasets suffer from a marked quality imbalance, which degrades PRM performance and highlights the need for an effective data selection strategy. To address the issues, we introduce DreamPRM, a domain-reweighted training framework for multimodal PRMs which employs bi-level optimization. In the lower-level optimization, DreamPRM performs fine-tuning on multiple datasets with domain weights, allowing the PRM to prioritize high-quality reasoning signals and alleviating the impact of dataset quality imbalance. In the upper-level optimization, the PRM is evaluated on a separate meta-learning dataset; this feedback updates the domain weights through an aggregation loss function, thereby improving the generalization capability of trained PRM. Extensive experiments on multiple multimodal reasoning benchmarks covering both mathematical and general reasoning show that test-time scaling with DreamPRM consistently improves the performance of state-of-the-art MLLMs. Further comparisons reveal that DreamPRM's domain-reweighting strategy surpasses other data selection methods and yields higher accuracy gains than existing test-time scaling approaches.



Paperid:2333
Authors:Yuang Ai, Qihang Fan, Xuefeng Hu, Zhenheng Yang, Ran He, Huaibo Huang
Abstract:
Abstract:Diffusion Transformer (DiT), a promising diffusion model for visual generation, demonstrates impressive performance but incurs significant computational overhead. Intriguingly, analysis of pre-trained DiT models reveals that global self-attention is often redundant, predominantly capturing local patterns—highlighting the potential for more efficient alternatives. In this paper, we revisit convolution as an alternative building block for constructing efficient and expressive diffusion models. However, naively replacing self-attention with convolution typically results in degraded performance. Our investigations attribute this performance gap to the higher channel redundancy in ConvNets compared to Transformers. To resolve this, we introduce a compact channel attention mechanism that promotes the activation of more diverse channels, thereby enhancing feature diversity. This leads to Diffusion ConvNet (DiCo), a family of diffusion models built entirely from standard ConvNet modules, offering strong generative performance with significant efficiency gains. On class-conditional ImageNet benchmarks, DiCo outperforms previous diffusion models in both image quality and generation speed. Notably, DiCo-XL achieves an FID of **2.05** at 256$\times$256 resolution and **2.53** at 512$\times$512, with a **2.7$\times$** and **3.1$\times$** speedup over DiT-XL/2, respectively. Furthermore, our largest model, DiCo-H, scaled to 1B parameters, reaches an FID of **1.90** on ImageNet 256$\times$256—without any additional supervision during training.



Paperid:2334
Authors:Feiran Wang, Jiachen Tao, Junyi Wu, Haoxuan Wang, Bin Duan, Kai Wang, Zongxin Yang, Yan Yan
Abstract:
X-ray imaging is indispensable in medical diagnostics, yet its use is tightly regulated due to radiation exposure. Recent research borrows representations from the 3D reconstruction area to complete two tasks with reduced radiation dose: X-ray Novel View Synthesis (NVS) and Computed Tomography (CT) reconstruction. However, these representations fail to fully capture the penetration and attenuation properties of X-ray imaging as they originate from visible light imaging. In this paper, we introduce X-Field, a 3D representation grounded in the physics of X-ray imaging. First, we employ homogeneous 3D ellipsoids with distinct attenuation coefficients to accurately model diverse materials within internal structures. Second, we introduce an efficient path partitioning algorithm that resolves the intricate intersection of ellipsoids to compute cumulative attenuation along an X-ray path. We further propose a hybrid progressive initialization to refine the geometric accuracy of X-Filed and incorporate material-based optimization to enhance model fitting along material boundaries.Experiments show that X-Field achieves superior visual fidelity on both real-world human organ and synthetic object datasets, outperforming state-of-the-art methods in X-ray NVS and CT Reconstruction.



Paperid:2335
Authors:Mateo Clémente, Leo Brunswic, Yang, Xuan Zhao, Yasser Khalil, Haoyu Lei, Amir Rasouli, Yinchuan Li
Abstract:
Diffusion models, such as diffusion policy, have achieved state-of-the-art results in robotic manipulation by imitating expert demonstrations. While diffusion models were originally developed for vision tasks like image and video generation, many of their inference strategies have been directly transferred to control domains without adaptation. In this work, we show that by tailoring the denoising process to the specific characteristics of embodied AI tasks—particularly the structured, low-dimensional nature of action distributions---diffusion policies can operate effectively with as few as 5 neural function evaluations.Building on this insight, we propose a population-based sampling strategy, genetic denoising, which enhances both performance and stability by selecting denoising trajectories with low out-of-distribution risk. Our method solves challenging tasks with only two neural function evaluations while improving or matching performance. We evaluate our approach across 14 robotic manipulation tasks from D4RL and Robomimic, spanning multiple action horizons and inference budgets. In over 2 million evaluations, our method consistently outperforms standard diffusion-based policies, achieving up to 20\% performance gains with significantly fewer inference steps.



Paperid:2336
Authors:Zixi Wang, Yushe Cao, Yubo Huang, Jinzhu Wei, Jingzehua Xu, Shuai Zhang, Xin Lai
Abstract:
Abstract:In this paper, we propose a new method called \textit{Self-Training with Dynamic Weighting} (STDW), which aims to enhance robustness in Gradual Domain Adaptation (GDA) by addressing the challenge of smooth knowledge migration from the source to the target domain. Traditional GDA methods mitigate domain shift through intermediate domains and self-training but often suffer from inefficient knowledge migration or incomplete intermediate data. Our approach introduces a dynamic weighting mechanism that adaptively balances the loss contributions of the source and target domains during training. Specifically, we design an optimization framework governed by a time-varying hyperparameter $\varrho$ (progressing from 0 to 1), which controls the strength of domain-specific learning and ensures stable adaptation. The method leverages self-training to generate pseudo-labels and optimizes a weighted objective function for iterative model updates, maintaining robustness across intermediate domains. Experiments on rotated MNIST, color-shifted MNIST, portrait datasets, and the Cover Type dataset demonstrate that STDW outperforms existing baselines. Ablation studies further validate the critical role of $\varrho$'s dynamic scheduling in achieving progressive adaptation, confirming its effectiveness in reducing domain bias and improving generalization. This work provides both theoretical insights and a practical framework for robust gradual domain adaptation, with potential applications in dynamic real-world scenarios.



Paperid:2337
Authors:Lorenzo Basile, Valentino Maiorca, Diego Doimo, Francesco Locatello, Alberto Cazzaniga
Abstract:
Language and vision-language models have shown impressive performance across a wide range of tasks, but their internal mechanisms remain only partly understood. In this work, we study how individual attention heads in text-generative models specialize in specific semantic or visual attributes. Building on an established interpretability method, we reinterpret the practice of probing intermediate activations with the final decoding layer through the lens of signal processing. This lets us analyze multiple samples in a principled way and rank attention heads based on their relevance to target concepts. Our results show consistent patterns of specialization at the head level across both unimodal and multimodal transformers. Remarkably, we find that editing as few as 1% of the heads, selected using our method, can reliably suppress or enhance targeted concepts in the model output. We validate our approach on language tasks such as question answering and toxicity mitigation, as well as vision-language tasks including image classification and captioning. Our findings highlight an interpretable and controllable structure within attention layers, offering simple tools for understanding and editing large-scale generative models.



Paperid:2338
Authors:Roi Benita, Michael Elad, Joseph Keshet
Abstract:
Diffusion models (DMs) have emerged as powerful tools for modeling complex data distributions and generating realistic new samples. Over the years, advanced architectures and sampling methods have been developed to make these models practically usable. However, certain synthesis process decisions still rely on heuristics without a solid theoretical foundation. In our work, we offer a novel analysis of the DM's inference process, introducing a comprehensive frequency response perspective. Specifically, by relying on Gaussianity assumption, we present the inference process as a closed-form spectral transfer function, capturing how the generated signal evolves in response to the initial noise. We demonstrate how the proposed analysis can be leveraged to design a noise schedule that aligns effectively with the characteristics of the data. The spectral perspective also provides insights into the underlying dynamics and sheds light on the relationship between spectral properties and noise schedule structure. Our results lead to scheduling curves that are dependent on the spectral content of the data, offering a theoretical justification for some of the heuristics taken by practitioners.



Paperid:2339
Authors:Yusu Hong, Junhong Lin
Abstract:
Abstract:Adafactor is an early memory-efficient optimization algorithm proposed as an alternative to Adam. By eliminating first-order momentum and employing arank-$1$ matrix factorization to approximate the second-moment matrix, Adafactor achieves near-zero memory overhead compared to traditional gradient descent methods.Despite its practical suitability for large-scale training tasks where memory efficiency is critical, its theoretical convergence analysis remains unexplored, largely due to the challenges posed by its matrix factorization and update clipping mechanisms. In this work, we provide a convergence analysis of Adafactor for non-convex smooth optimization.We establish optimal convergence rates (up to logarithmic factors) for finding stationary points in both deterministic and stochastic settings, the latter under sub-Gaussian noises.Central to our analysis involves viewing Adafactor as an approximation of Adam, andthe use of a new proxy step-size to approximate the unique adaptive step-size induced by Adafactor's matrix factorization and update clipping, along with an induction argument to control the gradient magnitude. Our finding may theoretically suggest that involving rank-$1$ matrix approximation of the second-moment matrix in Adam does not fundamentally hinder the convergence.



Paperid:2340
Authors:Mateusz Pach, Shyamgopal Karthik, Quentin Bouniot, Serge Belongie, Zeynep Akata
Abstract:
Given that interpretability and steerability are crucial to AI safety, Sparse Autoencoders (SAEs) have emerged as a tool to enhance them in Large Language Models (LLMs). In this work, we extend the application of SAEs to Vision-Language Models (VLMs), such as CLIP, and introduce a comprehensive framework for evaluating monosemanticity at the neuron-level in vision representations. To ensure that our evaluation aligns with human perception, we propose a benchmark derived from a large-scale user study. Our experimental results reveal that SAEs trained on VLMs significantly enhance the monosemanticity of individual neurons, with sparsity and wide latents being the most influential factors. Notably, we demonstrate that applying SAE interventions on CLIP's vision encoder directly steers multimodal LLM outputs (e.g., LLaVA), without any modifications to the underlying model. These findings emphasize the practicality and efficacy of SAEs as an unsupervised tool for enhancing both interpretability and control of VLMs.



Paperid:2341
Authors:Salva Rühling Cachay, Miika Aittala, Karsten Kreis, Noah Brenowitz, Arash Vahdat, Morteza Mardani, Rose Yu
Abstract:
Abstract:While applications of diffusion models to probabilistic forecasting of high-dimensional dynamical systems have shown promising results, these applications often use limited temporal context and predict future snapshots one by one.To explicitly model the sequential nature of the data and the growth of physical uncertainty inherent to chaotic systems, we introduce Elucidated Rolling Diffusion Models (ERDM). ERDM fuses the principled design of Elucidated Diffusion Models (EDM) with a rolling diffusion framework that assigns progressively larger noise levels to farther lead times.We adapt the EDM noise schedule, network preconditioning, and Heun sampler for this specific setting and introduce three key contributions: $(i)$ a novel loss weighting scheme, that puts more weights to middle snapshots where the stochastic to deterministic demixing occurs; $(ii)$ an efficient initialization strategy using a pretrained next-step EDM for the initial window to mitigate sensitivity in nonlinear, chaotic systems; and $(iii)$ a bespoke hybrid sequence architecture for robust spatiotemporal feature extraction under progressive denoising. On 2D Navier–Stokes simulations, ERDM improves probabilistic skill scores by up to $50\%$ over prior conditional diffusion baselines.On ERA5 global weather at $1.5^\circ$ resolution, it matches the ensemble skill of leading forecasters like IFS ENS and NeuralGCM, while being more efficient, requiring only 4 H200 GPUs and 5 days for training.



Paperid:2342
Authors:Stefan Stojanov, David Wendt, Seungwoo Kim, Rahul Venkatesh, Kevin Feigelis, Klemen Kotar, Khai Loong Aw, Jiajun Wu, Daniel Yamins
Abstract:
Estimating motion primitives from video (e.g. optical flow and occlusion) is a critically-important computer vision problem with many downstream applications, including in controllable video generation and robotics. Current solutions are primarily supervised on synthetic data or require tuning of situation-specific heuristics, which inherently limits these models' capabilities in real-world contexts. A natural solution to transcend these limitations would be to deploy large-scale self-supervised video models, which can be scalably trained on unrestricted real-world video datasets. However, despite recent progress, motion-primitive extraction from large pretrained video models remains relatively underexplored. In this work, we describe Opt-CWM, a self-supervised flow and occlusion estimation technique from a pretrained video prediction model. Opt-CWM uses ``counterfactual probes'' to extract motion information from a base video model in a zero-shot fashion. The key problem we solve is optimal probe generation, using a combination of an efficient parameterization of the space counterfactual probes, together with a novel generic sparse-prediction principle for learning the probe-generation parameters in a self-supervised fashion. Opt-CWM achieves state-of-the-art performance for motion estimation on real-world videos while requiring no labeled data.



Paperid:2343
Authors:Zhihai Wang, Zijie Geng, Zhaojie Tu, Jie Wang, Yuxi Qian, Zhexuan Xu, Ziyan Liu, Siyuan Xu, Zhentao Tang, Shixiong Kai, Mingxuan Yuan, Jianye Hao, Bin Li, Feng Wu
Abstract:
Abstract:Chip placement is a critical step in the Electronic Design Automation (EDA) workflow, which aims to arrange chip modules on the canvas to optimize the performance, power, and area (PPA) metrics of final designs.Recent advances show great potential of AI-based algorithms in chip placement.However, due to the lengthy EDA workflow, evaluations of these algorithms often focus on intermediate surrogate metrics, which are computationally efficient but often misalign with the final end-to-end performance (i.e., the final design PPA).To address this challenge, we propose to build ChiPBench, a comprehensive benchmark specifically designed to evaluate the effectiveness of AI-based algorithms in final design PPA metrics.Specifically, we generate a diverse evaluation dataset from $20$ circuits across various domains, such as CPUs, GPUs, and NPUs.We then evaluate six state-of-the-art AI-based chip placement algorithms on the dataset and conduct a thorough analysis of their placement behavior.Extensive experiments show that AI-based chip placement algorithms produce unsatisfactory final PPA results, highlighting the significant influence of often-overlooked factors like regularity and dataflow.We believe ChiPBench will effectively bridge the gap between academia and industry.



Paperid:2344
Authors:Qinyu Xu, Yuanyang Zhu, Xuefei Wu, Chunlin Chen
Abstract:
Abstract:The ability to model interactions among agents is crucial for effective coordination and understanding their cooperation mechanisms in Multi-Agent Reinforcement Learning (MARL). However, previous efforts to model high-order interactions have been primarily hindered by the combinatorial explosion or the opaque nature of their black-box network structures. In this paper, we propose a novel value decomposition framework, called Continued Fraction Q-Learning (QCoFr), which can flexibly capture arbitrary-order agent interactions with only linear complexity $\mathcal{O}\left({n}\right)$ in the number of agents, thus avoiding the combinatorial explosion when modeling rich cooperation. Furthermore, we introduce the variational information bottleneck to extract latent information for estimating credits. This latent information helps agents filter out noisy interactions, thereby significantly enhancing both cooperation and interpretability. Various experiments demonstrate that QCoFr not only achieves better performance but also provides interpretability that aligns with theoretical analysis.



Paperid:2345
Authors:Philippe Formont, Maxime Darrin, Banafsheh Karimian, Eric Granger, Jackie CK Cheung, Ismail Ayed, Mohammadhadi Shateri, Pablo Piantanida
Abstract:
Casting complex inputs into tractable representations is a critical step across various fields. Diverse embedding models emerge from differences in architectures, loss functions, input modalities and datasets, each capturing unique aspects of the input. Multi-teacher distillation leverages this diversity to enrich representations but often remains tailored to specific tasks. We introduce a task-agnostic framework based on a ``majority vote" objective function. We demonstrate that this function is bounded by the mutual information between the student and the teachers' embeddings, leading to a task-agnostic distillation loss that eliminates dependence on task-specific labels or prior knowledge. Comprehensive evaluations across text, vision models, and molecular modeling show that our method effectively leverages teacher diversity, resulting in representations enabling better performance for a wide range of downstream tasks such as classification, clustering, or regression. Additionally, we train and release state-of-the-art embedding models, enhancing downstream performance in various modalities.



Paperid:2346
Authors:Youliang Yuan, Wenxiang Jiao, Yuejin Xie, Chihao Shen, Menghan Tian, Wenxuan Wang, Jen-Tse Huang, Pinjia He
Abstract:
Human safety awareness gaps often prevent the timely recognition of everyday risks.In solving this problem, a proactive safety artificial intelligence (AI) system would work better than a reactive one. Instead of just reacting to users' questions, it would actively watch people’s behavior and their environment to detect potential dangers in advance.Our Proactive Safety Bench (PaSBench) evaluates this capability through 416 multimodal scenarios (128 image sequences, 288 text logs) spanning 5 safety-critical domains.Evaluation of 36 advanced models reveals fundamental limitations: Top performers like Gemini-2.5-pro achieve 71\% image and 64\% text accuracy, but miss 45-55\% risks in repeated trials. Through failure analysis, we identify unstable proactive reasoning rather than knowledge deficits as the primary limitation.This work establishes (1) a proactive safety benchmark, (2) systematic evidence of model limitations, and (3) critical directions for developing reliable protective AI. We believe our dataset and findings can promote the development of safer AI assistants that actively prevent harm rather than merely respond to requests.



Paperid:2347
Authors:Andrea Della Vecchia, Arnaud Mavakala Watusadisi, Ernesto De Vito, Lorenzo Rosasco
Abstract:
This paper addresses the covariate shift problem in the context of nonparametric regression within reproducing kernel Hilbert spaces (RKHSs). Covariate shift arises in supervised learning when the input distributions of the training and test data differ, presenting additional challenges for learning. Although kernel methods have optimal statistical properties, their high computational demands in terms of time and, particularly, memory, limit their scalability to large datasets. To address this limitation, the main focus of this paper is to explore the trade-off between computational efficiency and statistical accuracy under covariate shift. We investigate the use of random projections where the hypothesis space consists of a random subspace within a given RKHS. Our results show that, even in the presence of covariate shift, significant computational savings can be achieved without compromising learning performance.



Paperid:2348
Authors:Zhe Zhao, Zhiheng Gong, Pengkun Wang, HaiBin Wen, Cankun Guo, Bo Xue, Xi Lin, Zhenkun Wang, Qingfu Zhang, Yang Wang
Abstract:
Long-Tailed Recognition (LTR) presents a significant challenge due to extreme class imbalance, where existing methods often struggle to balance performance across head and tail classes. Directly applying multi-objective optimization (MOO) to leverage multiple LTR strategies can be complex and unstable. To address this, we propose TS-MOF (Two-Stage Multi-Objective Fine-tuning), a novel framework that strategically decouples feature learning from classifier adaptation. After standard pre-training, TS-MOF freezes the feature backbone and focuses on an efficient multi-objective fine-tuning of specialized classifier heads. The core of TS-MOF's second stage lies in two innovations: Refined Performance Level Agreement for adaptive task weighting based on real-time per-class performance, and Robust Deterministic Projective Conflict Gradient for stable gradient conflict resolution and constructive fusion. This approach enables effective synergy between diverse LTR strategies, leading to significant and balanced performance improvements. Extensive experiments on CIFAR100-LT, ImageNet-LT, and iNaturalist 2018 demonstrate that TS-MOF achieves state-of-the-art results, particularly enhancing tail class accuracy (e.g., +3.3\% on CIFAR100-LT IR=100 tail) while improving head class performance, all within a remarkably short fine-tuning period of 20 epochs.



Paperid:2349
Authors:Zhewei Kang, Xuandong Zhao, Dawn Song
Abstract:
Abstract:Best-of-N selection is a key technique for improving the reasoning performance of Large Language Models (LLMs) through increased test-time computation. Current state-of-the-art methods often employ computationally intensive reward models for response evaluation and selection. Reward-free alternatives, like self-consistency and universal self-consistency, are limited in their ability to handle open-ended generation tasks or scale effectively. To address these limitations, we propose self-certainty, a novel and efficient metric that leverages the inherent probability distribution of LLM outputs to estimate response quality without requiring external reward models. We hypothesize that higher distributional self-certainty, aggregated across multiple samples, correlates with improved response accuracy, as it reflects greater confidence in the generated output. Through extensive experiments on various reasoning tasks, we demonstrate that self-certainty (1) scales effectively with increasing sample size $N$, akin to reward models but without the computational overhead; (2) complements chain-of-thought, improving reasoning performance beyond greedy decoding; and (3) generalizes to open-ended tasks where traditional self-consistency methods fall short. Our findings establish self-certainty as a practical and efficient way for improving LLM reasoning capabilities.



Paperid:2350
Authors:Gabriele Oliaro, Zhihao Jia, Daniel Campos, Aurick Qiao
Abstract:
Abstract:Speculative decoding is widely adopted to reduce latency in large language model (LLM) inference by leveraging smaller draft models capable of handling diverse user tasks. However, emerging AI applications, such as LLM-based agents, present unique workload characteristics: instead of diverse independent requests, agentic frameworks typically submit repetitive inference requests, such as multi-agent pipelines performing similar subtasks or self-refinement loops iteratively enhancing outputs. These workloads result in long and highly predictable sequences, which current speculative decoding methods do not effectively exploit. To address this gap, we introduce \emph{SuffixDecoding}, a novel method that utilizes efficient suffix trees to cache long token sequences from prompts and previous outputs. By adaptively speculating more tokens when acceptance likelihood is high and fewer when it is low, SuffixDecoding effectively exploits opportunities for longer speculations while conserving computation when those opportunities are limited. Evaluations on agentic benchmarks, including SWE-Bench and Text-to-SQL, demonstrate that SuffixDecoding achieves speedups of up to 3.9$\times$, outperforming state-of-the-art methods -- 2.2$\times$ faster than model-based approaches like EAGLE-2/3 and 1.6$\times$ faster than model-free approaches such as Token Recycling. SuffixDecoding is open-sourced.



Paperid:2351
Authors:Kaiqing Lin, Zhiyuan Yan, Ke-Yue Zhang, Li Hao, Yue Zhou, Yuzhen Lin, Weixiang Li, Taiping Yao, Shouhong Ding, Bin Li
Abstract:
Securing personal identity against deepfake attacks is increasingly critical in the digital age, especially for celebrities and political figures whose faces are easily accessible and frequently targeted. Most existing deepfake detection methods focus on general-purpose scenarios and often ignore the valuable prior knowledge of known facial identities, e.g., "VIP individuals" whose authentic facial data are already available. In this paper, we proposeVIPGuard, a unified multimodal framework designed to capture fine-grained and comprehensive facial representations of a given identity, compare them against potentially fake or similar-looking faces, and reason over these comparisons to make accurate and explainable predictions. Specifically, our framework consists of three main stages. First, fine-tune a multimodal large language model (MLLM) to learn detailed and structural facial attributes. Second, we perform identity-level discriminative learning to enable the model to distinguish subtle differences between highly similar faces, including real and fake variations. Finally, we introduce user-specific customization, where we model the unique characteristics of the target face identity and perform semantic reasoning via MLLM to enable personalized and explainable deepfake detection. Our framework shows clear advantages over previous detection works, where traditional detectors mainly rely on low-level visual cues and provide no human-understandable explanations, while other MLLM-based models often lack a detailed understanding of specific face identities. To facilitate the evaluation of our method, we built a comprehensive identity-aware benchmark calledVIPBenchfor personalized deepfake detection, involving the latest 7 face-swapping and 7 entire face synthesis techniques for generation. Extensive experiments show that our model outperforms existing methods in both detection and explanation.



Paperid:2352
Authors:Yeonjoon Jung, Daehyun Ahn, Hyungjun Kim, Taesu Kim, Eunhyeok Park
Abstract:
Low-Rank Adaptation (LoRA) is a popular method for parameter-efficient fine-tuning (PEFT) of generative models, valued for its simplicity and effectiveness. Despite recent enhancements, LoRA still suffers from a fundamental limitation: overfitting when the bottleneck is widened. It performs best at ranks 32–64, yet its accuracy stagnates or declines at higher ranks, still falling short of full fine-tuning (FFT) performance. We identify the root cause as LoRA’s structural bottleneck, which introduces gradient entanglement to the unrelated input channels and distorts gradient propagation. To address this, we introduce a novel structure, Granular Low-Rank Adaptation (GraLoRA) that partitions weight matrices into sub-blocks, each with its own low-rank adapter. With negligible computational or storage cost, GraLoRA overcomes LoRA’s limitations, effectively increases the representational capacity, and more closely approximates FFT behavior. Experiments on code generation and commonsense reasoning benchmarks show that GraLoRA consistently outperforms LoRA and other baselines, achieving up to +8.5\% absolute gain in Pass@1 on HumanEval+. These improvements hold across model sizes and rank settings, making GraLoRA a scalable and robust solution for PEFT.



Paperid:2353
Authors:Huiyang Shao, Xin Xia, Yuxi Ren, XING WANG, Xuefeng Xiao
Abstract:
Diffusion models have emerged as state‑of‑the‑art in image synthesis.However, it often suffer from semantic instability and slow iterative denoising. We introduce LABridge (Latent Alignment Bridge), a novel Text–Image Latent Alignment Framework via an Ornstein–Uhlenbeck (OU) bridge, which explicitly preserves and aligns textual and visual semantics in a aligned latent space. LABridge employs a Text-Image Alignment Encoder (TIAE) to encode text prompts into structured priors that are directly aligned with image latents. Instead of a homogeneous Gaussian, OU diffusion bridge smoothly interpolates between these text‑conditioned priors and image latents, improving stability and reducing the number of denoising steps. Extensive experiments on standard text-to-image benchmarks show that LABridge achieves better text–image alignment metric, and competitive FID scores compared to leading diffusion baselines. By unifying text and image representations through principled latent alignment, LABridge paves the way for more efficient, semantically consistent, and high‑fidelity text to image generation.



Paperid:2354
Authors:Yuqi Wu, Wenzhao Zheng, Jie Zhou, Jiwen Lu
Abstract:
Dense 3D scene reconstruction from an ordered sequence or unordered image collections is a critical step when bringing research in computer vision into practical scenarios. Following the paradigm introduced by DUSt3R, which unifies an image pair densely into a shared coordinate system, subsequent methods maintain an implicit memory to achieve dense 3D reconstruction from more images. However, such implicit memory is limited in capacity and may suffer from information loss of earlier frames. We propose Point3R, an online framework targeting dense streaming 3D reconstruction. To be specific, we maintain an explicit spatial pointer memory directly associated with the 3D structure of the current scene. Each pointer in this memory is assigned a specific 3D position and aggregates scene information nearby in the global coordinate system into a changing spatial feature. Information extracted from the latest frame interacts explicitly with this pointer memory, enabling dense integration of the current observation into the global coordinate system. We design a 3D hierarchical position embedding to promote this interaction and design a simple yet effective fusion mechanism to ensure that our pointer memory is uniform and efficient. Our method achieves competitive or state-of-the-art performance on various tasks with low training costs.



Paperid:2355
Authors:Chaewoon Bae, Doyun Choi, Jaehyun Lee, Jaemin Yoo
Abstract:
Few-shot node classification on hypergraphs requires models that generalize from scarce labels while capturing high-order structures. Existing hypergraph neural networks (HNNs) effectively encode such structures but often suffer from overfitting and scalability issues due to complex, black-box architectures. In this work, we propose ZEN (Zero-Parameter Hypergraph Neural Network), a fully linear and parameter-free model that achieves both expressiveness and efficiency. Built upon a unified formulation of linearized HNNs, ZEN introduces a tractable closed-form solution for the weight matrix and a redundancy-aware propagation scheme to avoid iterative training and to eliminate redundant self-information. On 11 real-world hypergraph benchmarks, ZEN consistently outperforms eight baseline models in classification accuracy while achieving up to 696× speedups over the fastest competitor. Moreover, the decision process of ZEN is fully interpretable, providing insights into the characteristic of a dataset. Our code and datasets are fully available at https://anonymous.4open.science/r/ZEN-FEFB.14



Paperid:2356
Authors:Gunho Park, Jeongin Bae, Byeongwook Kim, Baeseong Park, Jiwon Ryu, Hoseung Kim, Se Jung Kwon, Dongsoo Lee
Abstract:
Abstract:Codebook-based quantization delivers superior performance in extreme low-bit environments but is hindered by latency overhead in existing dequantization-based kernels, which require loading the entire codebook into programmable cache memory. We propose CodeGEMM, a matrix multiplication kernel optimized for computing codebook-based quantized models by precomputing and storing inner product results in a Psumbook, thereby eliminating the need for traditional dequantization. The kernel supports a wide range of hyperparameter configurations, enabling exploration of trade-offs between latency, memory usage, and accuracy to achieve optimal performance. CodeGEMM obtains a $2.27\times$ speed-up compared to the state-of-the-art codebook-based quantization method with comparable accuracy in the 2-bit configuration.



Paperid:2357
Authors:Julie Mordacq, David Loiseaux, Vicky Kalogeiton, Steve OUDOT
Abstract:
Self-supervised learning (SSL) has emerged as a powerful paradigm for learning representations without labeled data, often by enforcing invariance to input transformations such as rotations or blurring.Recent studies have highlighted two pivotal properties for effective representations: (i) avoiding dimensional collapse-where the learned features occupy only a low-dimensional subspace, and (ii) enhancing uniformity of the induced distribution.In this work, we introduce T-REGS, a simple regularization framework for SSL based on the length of the Minimum Spanning Tree (MST) over the learned representation.We provide theoretical analysis demonstrating that T-REGS simultaneously mitigates dimensional collapse and promotes distribution uniformity on arbitrary compact Riemannian manifolds.Several experiments on synthetic data and on classical SSL benchmarks validate the effectiveness of our approach at enhancing representation quality.



Paperid:2358
Authors:Mingfei Chen, Zijun Cui, Xiulong Liu, Jinlin Xiang, Yang Zheng, Jingyuan Li, Eli Shlizerman
Abstract:
3D spatial reasoning in dynamic, audio-visual environments is a cornerstone of human cognition yet remains largely unexplored by existing Audio-Visual Large Language Models (AV-LLMs) and benchmarks, which predominantly focus on static or 2D scenes. We introduce SAVVY-Bench, the first benchmark for 3D spatial reasoning in dynamic scenes with synchronized spatial audio. SAVVY-Bench is comprised of thousands of carefully curated question–answer pairs probing both directional and distance relationships involving static and moving objects, and requires fine-grained temporal grounding, consistent 3D localization, and multi-modal annotation. To tackle this challenge, we propose SAVVY, a novel training-free reasoning pipeline that consists of two stages: (i) Egocentric Spatial Tracks Estimation, which leverages AV-LLMs as well as other audio-visual methods to track the trajectories of key objects related to the query using both visual and spatial audio cues, and (ii) Dynamic Global Map Construction, which aggregates multi-modal queried object trajectories and converts them into a unified global dynamic map. Using the constructed map, a final QA answer is obtained through a coordinate transformation that aligns the global map with the queried viewpoint. Empirical evaluation demonstrates that SAVVY substantially enhances performance of state-of-the-art AV-LLMs, setting a new standard and stage for approaching dynamic 3D spatial reasoning in AV-LLMs.



Paperid:2359
Authors:Anand Brahmbhatt, Gon Buzaglo, Sofiia Druchyna, Elad Hazan
Abstract:
We propose a new method for the problem of controlling linear dynamical systems under partial observation and adversarial disturbances. Our new algorithm, Double Spectral Control (DSC), matches the best known regret guarantees while exponentially improving runtime complexity over previous approaches in its dependence on the system's stability margin. Our key innovation is a two-level spectral approximation strategy, leveraging double convolution with a universal basis of spectral filters, enabling efficient and accurate learning of the best linear dynamical controllers.



Paperid:2360
Authors:Lishen Qu, Zhihao Liu, Shihao Zhou, LUO YAQI, Jie Liang, Hui Zeng, Lei Zhang, Jufeng Yang
Abstract:
Flicker artifacts in short-exposure images are caused by the interplay between the row-wise exposure mechanism of rolling shutter cameras and the temporal intensity variations of alternating current (AC)-powered lighting. These artifacts typically appear as uneven brightness distribution across the image, forming noticeable dark bands. Beyond compromising image quality, this structured noise also affects high-level tasks, such as object detection and tracking, where reliable lighting is crucial. Despite the prevalence of flicker, the lack of a large-scale, realistic dataset has been a significant barrier to advancing research in flicker removal. To address this issue, we present BurstDeflicker, a robust and scalable benchmark constructed using three complementary data acquisition strategies. First, we develop a Retinex-based synthesis pipeline that redefines the goal of flicker removal and enables controllable manipulation of key flicker-related attributes (e.g., intensity, area, and frequency), thereby facilitating the generation of diverse flicker patterns. Second, we capture 4,000 real-world flicker images from different scenes, which help the model better understand the spatial and temporal characteristics of real flicker artifacts and generalize more effectively to wild scenarios. Finally, due to the non-repeatable nature of dynamic scenes, we propose a green-screen method to incorporate motion into image pairs while preserving real flicker degradation. Comprehensive experiments demonstrate the effectiveness of our dataset and its potential to advance research in flicker removal. The code and dataset are available in the supplementary materials.



Paperid:2361
Authors:Kanghua Mo, Li Hu, Yucheng Long, Zhihao Li
Abstract:
Large language model (LLM) agents have demonstrated remarkable capabilities in complex reasoning and decision-making by leveraging external tools. However, this tool-centric paradigm introduces a previously underexplored attack surface: adversaries can manipulate tool metadata—such as names, descriptions, and parameter schemas—to influence agent behavior. We identify this as a new and stealthy threat surface that allows malicious tools to be preferentially selected by LLM agents, without requiring prompt injection or access to model internals. To demonstrate and exploit this vulnerability, we propose the Attractive Metadata Attack (AMA), a black-box in-context learning framework that generates highly attractive but syntactically and semantically valid tool metadata through iterative optimization. Our attack integrates seamlessly into standard tool ecosystems and requires no modification to the agent’s execution framework. Extensive experiments across ten realistic, simulated tool-use scenarios and a range of popular LLM agents demonstrate consistently high attack success rates (81\%–95\%) and significant privacy leakage, with negligible impact on primary task execution.Moreover, the attack remains effective even under prompt-level defenses and structured tool-selection protocols such as the Model Context Protocol, revealing systemic vulnerabilities in current agent architectures.These findings reveal that metadata manipulation constitutes a potent and stealthy attack surface, highlighting the need for execution-level security mechanisms that go beyond prompt-level defenses. Code is available at https://anonymous.4open.science/r/AMA-A753.



Paperid:2362
Authors:Ruikun Luo, Changwei Gu, Qiang He, Feifei Chen, Song Wu, Hai Jin, Yun Yang
Abstract:
KV cache technology, by storing key-value pairs, helps reduce the computational overhead incurred by large language models (LLMs). It facilitates their deployment on resource-constrained edge computing nodes like edge servers. However, as the complexity and size of tasks increase, KV cache usage leads to substantial GPU memory consumption. Existing research has focused on mitigating KV cache memory usage through sequence length reduction, task-specific compression, and dynamic eviction policies. However, these methods are computationally expensive for resource-constrained edge computing nodes. To tackle this challenge, this paper presents Sim-LLM, a novel inference optimization mechanism that leverages task similarity to reduce KV cache memory consumption for LLMs. By caching KVs from processed tasks and reusing them for subsequent similar tasks during inference, Sim-LLM significantly reduces memory consumption while boosting system throughput and increasing maximum batch size, all with minimal accuracy degradation. Evaluated on both A40 and A100 GPUs, Sim-LLM achieves a system throughput improvement of up to 39.40\% and a memory reduction of up to 34.65\%, compared to state-of-the-art approaches. Our source code is available at https://github.com/simllm/SimLLM.



Paperid:2363
Authors:Thomas Ressler-Antal, Frank Fundel, Malek Alaya, Stefan Andreas Baumann, Felix Krause, Ming Gui, Björn Ommer
Abstract:
Recent advances in video generation, particularly text-to-video (T2V) and image-to-video (I2V) models, have enabled the creation of visually compelling and dynamic videos from simple textual descriptions or initial frames. However, these models often fail to provide an explicit representation of motion separate from content, limiting their applicability for content creators. To address this gap, we propose DisMo, a novel paradigm for learning abstract motion representations directly from raw video data via an image-space reconstruction objective. Our representation is generic and independent of static information such as appearance, object identity, or pose. This enables open-world motion transfer, allowing motion to be transferred across semantically unrelated entities without requiring object correspondences, even between vastly different categories. Unlike prior methods, which exhibit a trade-off between motion fidelity and adherence to textual prompts—either overfitting to source video structure or drifting from the described action—our approach avoids this compromise by disentangling motion semantics from instance-specific appearance cues, enabling both accurate motion transfer and faithful prompt conditioning. Furthermore, our motion representation can be combined with any existing video generator via lightweight adapters, allowing us to effortlessly benefit from future advancements in video models. We demonstrate the effectiveness of our method through a diverse set of motion transfer tasks. Finally, we show that the learned representations are well-suited for downstream motion understanding tasks, consistently outperforming state-of-the-art video representation models such as V-JEPA in zero-shot action classification on benchmarks including Something-Something v2, Jester, and ARID.



Paperid:2364
Authors:Wenxuan Wang, Kai Wu, yujian li, Dan Wang, Xiaoyu Zhang
Abstract:
Foundation models for time series analysis (TSA) have attracted significant attention. However, challenges such as training data scarcity and imbalance continue to hinder their development. Inspired by complex dynamic system theories, we design a series-symbol data generation mechanism, enabling the unrestricted creation of high-quality time series data paired with corresponding symbolic expressions. To leverage series-symbol data pairs with strong correlations, we develop SymTime, a pre-trained foundation model for enhancing time series representation using symbolic information. SymTime demonstrates competitive performance across five major TSA tasks when fine-tuned with downstream tasks, rivaling foundation models pre-trained on real-world datasets. This approach underscores the potential of series-symbol data generation and pretraining mechanisms in overcoming data scarcity and enhancing task performance.



Paperid:2365
Authors:Amir Rezaei Balef, Claire Vernade, Katharina Eggensperger
Abstract:
Abstract:The Combined Algorithm Selection and Hyperparameter optimization (CASH) is a challenging resource allocation problem in the field of AutoML. We propose MaxUCB, a max $k$-armed bandit method to trade off exploring different model classes and conducting hyperparameter optimization. MaxUCB is specifically designed for the light-tailed and bounded reward distributions arising in this setting and, thus, provides an efficient alternative compared to classic max $k$-armed bandit methods assuming heavy-tailed reward distributions. We theoretically and empirically evaluate our method on four standard AutoML benchmarks, demonstrating superior performance over prior approaches. We make our code and data available at https://anonymous.4open.science/r/CASH_with_Bandits



Paperid:2366
Authors:Dongyeop Woo, Minsu Kim, Minkyu Kim, Kiyoung Seong, Sungsoo Ahn
Abstract:
We propose Energy-based generator matching (EGM), a modality-agnostic approach to train generative models from energy functions in the absence of data. Extending the recently proposed generator matching, EGM enables training of arbitrary continuous-time Markov processes, e.g., diffusion, flow, and jump, and can generate data from continuous, discrete, and a mixture of two modalities. To this end, we propose estimating the generator matching loss using self-normalized importance sampling with an additional bootstrapping trick to reduce variance in the importance weight. We validate EGM on both discrete and multimodal tasks up to 100 and 20 dimensions, respectively.



Paperid:2367
Authors:Ittai Rubinstein, Samuel Hopkins
Abstract:
Abstract:How does the training data affect a model's behavior?This is the question we seek to answer with *data attribution*.The leading practical approaches to data attribution are based on *influence functions* (IF).IFs utilize a first-order Taylor approximation to efficiently predict the effect of removing a set of samples from the training set without retraining the model, and are used in a wide variety of machine learning applications.However, especially in the high-dimensional regime (# params $\geq \Omega($# samples$)$), they are often imprecise and tend to underestimate the effect of sample removals, even for simple models such as logistic regression.We present *rescaled influence functions* (RIF), a new tool for data attribution which can be used as a drop-in replacement for influence functions, with little computational overhead but significant improvement in accuracy.We compare IF and RIF on a range of real-world datasets, showing that RIFs offer significantly better predictions in practice, and present a theoretical analysis explaining this improvement.Finally, we present a simple class of data poisoning attacks that would fool IF-based detections but would be detected by RIF.



Paperid:2368
Authors:Hritik Bansal, Daniel Israel, Siyan Zhao, Shufan Li, Tung Nguyen, Aditya Grover
Abstract:
Recent advancements in mixed-modal generative have opened new avenues for developing unified biomedical assistants capable of analyzing biomedical images, answering complex questions about them, and generating multimodal patient reports. However, existing datasets face challenges such as small sizes, limited coverage of biomedical tasks and domains, and a reliance on narrow sources. To address these gaps, we present MedMax, a large-scale multimodal biomedical instruction-tuning dataset for mixed-modal foundation models. With 1.47 million instances, MedMax encompasses a diverse range of tasks, including interleaved image-text generation, biomedical image captioning and generation, visual chat, and report understanding. These tasks span knowledge across diverse biomedical domains, including radiology and histopathology, grounded in medical papers and YouTube videos. Subsequently, we fine-tune a mixed-modal foundation model on the MedMax dataset, achieving significant performance improvements: a 26% gain over the Chameleon model and an 18.3% improvement over GPT-4o across 12 downstream biomedical visual question-answering tasks. Finally, we introduce a unified evaluation suite for biomedical tasks to guide the development of mixed-modal biomedical AI assistants. We release the code, data, and model at https://mint-medmax.github.io/.



Paperid:2369
Authors:Zihan Ma, Taolin Zhang, Maosongcao, Junnan Liu, Wenwei Zhang, Minnan Luo, Songyang Zhang, Kai Chen
Abstract:
Large language models (LLMs) have recently achieved notable success in code‑generation benchmarks such as HumanEval and LiveCodeBench. However, a detailed examination reveals that these evaluation suites often comprise only a limited number of homogeneous test cases, resulting in subtle faults going undetected. This not only artificially inflates measured performance but also compromises accurate reward estimation in reinforcement learning frameworks utilizing verifiable rewards (RLVR). To address these critical shortcomings, we systematically investigate the test-case generation (TCG) task by proposing multi-dimensional metrics designed to rigorously quantify test-suite thoroughness. Furthermore, we introduce a human-LLM collaborative method (SAGA), leveraging human programming expertise with LLM reasoning capability, aimed at significantly enhancing both the coverage and the quality of generated test cases. In addition, we develop a TCGBench to facilitate the study of the TCG task. Experiments show that SAGA achieves a detection rate of 90.62\% and a verifier accuracy of 32.58\% on TCGBench. The Verifier Accuracy (Verifier Acc) of the code generation evaluation benchmark synthesized by SAGA is 10.78\% higher than that of LiveCodeBench-v6. These results demonstrate the effectiveness of our proposed method. We hope this work contributes to building a scalable foundation for reliable LLM code evaluation, further advancing RLVR in code generation, and paving the way for automated adversarial test synthesis and adaptive benchmark integration.



Paperid:2370
Authors:Michael Ungersböck, Florian Grötschla, Luca Lanzendörfer, June Young Yi, Changho Choi, Roger Wattenhofer
Abstract:
Generative models have made significant progress in synthesizing high-fidelity audio from short textual descriptions. However, editing existing audio using natural language has remained largely underexplored. Current approaches either require the complete description of the edited audio or are constrained to predefined edit instructions that lack flexibility. In this work, we introduce SAO-Instruct, a model based on Stable Audio Open capable of editing audio clips using any free-form natural language instruction. To train our model, we create a dataset of audio editing triplets (input audio, edit instruction, output audio) using Prompt-to-Prompt, DDPM inversion, and a manual editing pipeline. Although partially trained on synthetic data, our model generalizes well to real in-the-wild audio clips and unseen edit instructions. We demonstrate that SAO-Instruct achieves competitive performance on objective metrics and outperforms other audio editing approaches in a subjective listening study. To encourage future research, we release our code and model weights.



Paperid:2371
Authors:Jinlai Liu, Jian Han, Bin Yan, Yi Jiang, Hui Wu, Fengda Zhu, Xing Wang, BINGYUE PENG, Zehuan Yuan
Abstract:
Abstract:We introduce Infinity*, a unified spacetime autoregressive framework for high-resolution image and dynamic video synthesis. Building on the recent success of autoregressive modeling in both vision and language, our purely discrete approach jointly captures spatial and temporal dependencies within a single architecture. This unified design naturally supports a variety of generation tasks such as text-to-image, text-to-video, image-to-video, and long-duration video synthesis via straightforward temporal autoregression. Through extensive experiments, Infinity* scores 83.74 on VBench, outperforming all autoregressive models by large margins, even surpassing diffusion competitors like HunyuanVideo. Without extra optimizations, our model generates a 5s, 720p video approximately 10$\times$ faster than leading diffusion-based methods. To our knowledge, Infinity* is the first discrete autoregressive video generator capable of producing industrial-grade 720p videos up to 10 seconds. We will release all code and models to foster further research in efficient, high-quality video generation.



Paperid:2372
Authors:Songhua Liu, Zhenxiong Tan, Xinchao Wang
Abstract:
Diffusion Transformers (DiT) have become a leading architecture in image generation. However, the quadratic complexity of attention mechanisms, which are responsible for modeling token-wise relationships, results in significant latency when generating high-resolution images. To address this issue, we aim at a linear attention mechanism in this paper that reduces the complexity of pre-trained DiTs to linear. We begin our exploration with a comprehensive summary of existing efficient attention mechanisms and identify four key factors crucial for successful linearization of pre-trained DiTs: locality, formulation consistency, high-rank attention maps, and feature integrity. Based on these insights, we introduce a convolution-like local attention strategy termed CLEAR, which limits feature interactions to a local window around each query token, and thus achieves linear complexity. Our experiments indicate that, by fine-tuning the attention layer on merely 10K self-generated samples for 10K iterations, we can effectively transfer knowledge from a pre-trained DiT to a student model with linear complexity, yielding results comparable to the teacher model. Simultaneously, it reduces attention computations by 99.5% and accelerates generation by 6.3 times for generating 8K-resolution images. Furthermore, we investigate favorable properties in the distilled attention layers, such as zero-shot generalization across various models and plugins, and improved support for multi-GPU parallel inference. Models and codes will be available.



Paperid:2373
Authors:Wenxiao Wu, Jing-Hao Xue, Chengming Xu, Chen Liu, Xinwei Sun, Changxin Gao, Nong Sang, Yanwei Fu
Abstract:
Visual In-Context Learning (VICL) has emerged as a prominent approach for adapting visual foundation models to novel tasks, by effectively exploiting contextual information embedded in in-context examples, which can be formulated as a global ranking problem of potential candidates. Current VICL methods, such as Partial2Global and VPR, are grounded in the similarity-priority assumption that images more visually similar to a query image serve as better in-context examples. This foundational assumption, while intuitive, lacks sufficient justification for its efficacy in selecting optimal in-context examples. Furthermore, Partial2Global constructs its global ranking from a series of randomly sampled pairwise preference predictions. Such a reliance on random sampling can lead to incomplete coverage and redundant samplings of comparisons, thus further adversely impacting the final global ranking. To address these issues, this paper introduces an enhanced variant of Partial2Global designed for reliable and holistic selection of in-context examples in VICL. Our proposed method, dubbed RH-Partial2Global, leverages a jackknife conformal prediction-guided strategy to construct reliable alternative sets and a covering design-based sampling approach to ensure comprehensive and uniform coverage of pairwise preferences. Extensive experiments demonstrate that RH-Partial2Global achieves excellent performance and outperforms Partial2Global across diverse visual tasks.



Paperid:2374
Authors:Jessica Fry, Aobo Li, Lindley Winslow, Xinyi Fu, Zhenghao Fu, Kaliroë Pappas
Abstract:
Dark matter makes up approximately 85\% of total matter in our universe, yet it has never been directly observed in any laboratory on Earth. The origin of dark matter is one of the most important questions in contemporary physics, and a convincing detection of dark matter would be a Nobel-Prize-level breakthrough in fundamental science. The ABRACADABRA experiment was specifically designed to search for dark matter. Although it has not yet made a discovery, ABRACADABRA has produced several dark matter search results widely endorsed by the physics community. The experiment generates ultra-long time-series data at a rate of 10 million samples per second, where the dark matter signal would manifest itself as a sinusoidal oscillation mode within the ultra-long time series. In this paper, we present the TIDMAD --- a comprehensive data release from the ABRACADABRA experiment including three key components: an ultra-long time series dataset divided into training, validation, and science subsets; a carefully-designed denoising score for direct model benchmarking; and a complete analysis framework which produces a community-standard dark matter search result suitable for publication as a physics paper. This data release enables core AI algorithms to extract the signal and produce real physics results thereby advancing fundamental science.



Paperid:2375
Authors:Liang Yin, Xudong Xie, Zhang Li, Xiang Bai, Yuliang Liu
Abstract:
Abstract:Scene text retrieval has made significant progress with the assistance of accurate text localization.However, existing approaches typically require costly bounding box annotations for training. Besides, they mostly adopt a customized retrieval strategy but struggle to unify various types of queries to meet diverse retrieval needs.To address these issues, we introduce Muti-query Scene Text retrieval with Attention Recycling (MSTAR), a box-free approach for scene text retrieval. It incorporates progressive vision embedding to dynamically capture the multi-grained representation of texts and harmonizes free-style text queries with style-aware instructions. Additionally, a multi-instance matching module is integrated to enhance vision-language alignment.Furthermore, we build the Multi-Query Text Retrieval (MQTR) dataset, the first benchmark designed to evaluate the multi-query scene text retrieval capability of models, comprising four query types and $16k$ images. Extensive experiments demonstrate the superiority of our method across seven public datasets and the MQTR dataset. Notably, MSTAR marginally surpasses the previous state-of-the-art model by 6.4\% in MAP on Total-Text while eliminating box annotation costs. Moreover, on the MQTR benchmark, MSTAR significantly outperforms the previous models by an average of 8.5\%. The code and data will be public.



Paperid:2376
Authors:Hadi Hosseini, Samarth Khanna, Ronak Singh
Abstract:
The rise of Large Language Models (LLMs) has driven progress in reasoning tasks---from program synthesis to scientific hypothesis generation---yet their ability to handle ranked preferences and structured algorithms in combinatorial domains remains underexplored. We study matching markets, a core framework behind applications like resource allocation and ride-sharing, which require reconciling individual ranked preferences to ensure stable outcomes. We evaluate seven state‐of‐the‐art models on a hierarchy of preference‐based reasoning tasks---ranging from stable‐matching generation to instability detection, instability resolution, and fine‐grained preference queries---to systematically expose their logical and algorithmic limitations in handling ranked inputs. Surprisingly, even top-performing models with advanced reasoning struggle to resolve instability in large markets, often failing to identify blocking pairs or execute algorithms iteratively. We further show that \textit{parameter-efficient fine-tuning} (LoRA) improves performance on large instances but degrades accuracy in small markets, suggesting a scale-dependent trade-off between generalization and overfitting.



Paperid:2377
Authors:Jiachen Liang, RuiBing Hou, Minyang Hu, Hong Chang, Shiguang Shan, Xilin Chen
Abstract:
Out-of-distribution (OOD) detection is critical for ensuring the reliability of deep learning models in open-world applications. While post-hoc methods are favored for their efficiency and ease of deployment, existing approaches often underexploit the rich information embedded in the model’s logits space. In this paper, we propose LogitGap, a novel post-hoc OOD detection method that explicitly exploits the relationship between the maximum logit and the remaining logits to enhance the separability between in-distribution (ID) and OOD samples. To further improve its effectiveness, we refine LogitGap by focusing on a more compact and informative subset of the logit space. Specifically, we introduce a training-free strategy that automatically identifies the most informative logits for scoring. We provide both theoretical analysis and empirical evidence to validate the effectiveness of our approach. Extensive experiments on both vision-language and vision-only models demonstrate that LogitGap consistently achieves state-of-the-art performance across diverse OOD detection scenarios and benchmarks.



Paperid:2378
Authors:Constantin Venhoff, Ashkan Khakzar, Sonia Joseph, Philip Torr, Neel Nanda
Abstract:
LLaVA-style Vision-Language Models (VLMs) have demonstrated impressive capabilities, but struggle with factual recall tasks compared to their underlying language model (LM). While previous work attributes this to insufficient computational depth after visual processing, we provide an alternative explanation: the distributed representations of visual information across visual tokens in early layers bypasses the factual recall mechanism that resides in the early-layer MLPs of the LM backbone. The performance gap therefore stems from the architectural design of VLMs, rather than insufficient computational capacity.Using linear probes, we show that dedicated linear representations of visual information only emerge in the middle-to-late layers of VLMs. As a result, factual recall in VLMs becomes a “two-hop” challenge, where factual recall precedes visual processing, but the visual processing finishes too late in the model. Through comparative analysis, we demonstrate that successful factual recall depends on the speed of the first processing “hop.”To further support our hypothesis, we patch early-layer MLP outputs from the LM backbone into the corresponding VLM layers, significantly improving factual recall performance. This suggests that the absence of properly aligned token embeddings in early layers is a key factor in factual recall degradation. Finally, we introduce a benchmark to systematically evaluate factual recall accuracy and knowledge hallucination in multimodal settings.Our findings highlight a fundamental architectural limitation in current VLMs and pave the way for designing models that better integrate visual and linguistic information for reliable factual reasoning.



Paperid:2379
Authors:Yukun Chen, Boheng Li, Yu Yuan, Leyi Qi, Yiming Li, Tianwei Zhang, Zhan Qin, Kui Ren
Abstract:
Knowledge distillation (KD) is a vital technique for deploying deep neural networks (DNNs) on resource-constrained devices by transferring knowledge from large teacher models to lightweight student models. While teacher models from third-party platforms may undergo security verification (e.g., backdoor detection), we uncover a novel and critical threat: distillation-conditional backdoor attacks (DCBAs). DCBA injects dormant and undetectable backdoors into teacher models, which become activated in student models via the KD process, even with clean distillation datasets. While the direct extension of existing methods is ineffective for DCBA, we implement this attack by formulating it as a bilevel optimization problem and proposing a simple yet effective method (i.e., SCAR). Specifically, the inner optimization simulates the KD process by optimizing a surrogate student model, while the outer optimization leverages outputs from this surrogate to optimize the teacher model for implanting the conditional backdoor. Our SCAR addresses this complex optimization utilizing an implicit differentiation algorithm with a pre-optimized trigger injection function. Extensive experiments across diverse datasets, model architectures, and KD techniques validate the effectiveness of our SCAR and its resistance against existing backdoor detection, highlighting a significant yet previously overlooked vulnerability in the KD process.



Paperid:2380
Authors:Yang Zhan, Yuan Yuan
Abstract:
The task of localizing an object's spatial tube based on language instructions and video, known as spatial video grounding (SVG), has attracted widespread interest. Existing SVG tasks have focused on ego-centric fixed front perspective and simple scenes, which only involved a very limited view and environment. However, UAV-based SVG remains underexplored, which neglects the inherent disparities in drone movement and the complexity of aerial object localization. To facilitate research in this field, we introduce the novel spatial aerial video grounding (SAVG) task.Specifically, we meticulously construct a large-scale benchmark, UAV-SVG, which contains over 2 million frames and offers 216 highly diverse target categories. To address the disparities and challenges posed by complex aerial environments, we propose a new end-to-end transformer architecture, coined SAVG-DETR. The innovations are three-fold. 1) To overcome the computational explosion of self-attention when introducing multi-scale features, our encoder efficiently decouples the multi-modality and multi-scale spatio-temporal modeling into intra-scale multi-modality interaction and cross-scale visual-only fusion. 2) To enhance small object grounding ability, we propose the language modulation module to integrate multi-scale information into language features and the multi-level progressive spatial decoder to decode from high to low level. The decoding stage for the lower-level vision-language features is gradually increased. 3) To improve the prediction consistency across frames, we design the decoding paradigm based on offset generation. At each decoding stage, we utilize reference anchors to constrict the grounding region, use context-rich object queries to predict offsets, and update reference anchors for the next stage.From coarse to fine, our SAVG-DETR gradually bridges the modality gap and iteratively refines reference anchors of the referred object, eventually grounding the spatial tube.Extensive experiments demonstrate that our SAVG-DETR significantly outperforms existing state-of-the-art methods. Comprehensive ablation studies and insightful analyses can provide new ideas and useful insights for the challenging task.



Paperid:2381
Authors:Chaofan Li, Jianlyu Chen, Yingxia Shao, Defu Lian, Zheng Liu
Abstract:
Code embedding models attract increasing attention due to the widespread popularity of retrieval-augmented generation (RAG) in software development. These models are expected to capture the rich semantic relationships inherent to code, which differ significantly from those found in text. However, existing models remain severely limited due to the scarcity of high-quality training data. In this work, we introduce \textbf{CodeR} (\underline{Code} \underline{R}etrieval), a state-of-the-art embedding model for general-purpose code retrieval. The superior performance of CodeR is built upon \textbf{CodeR-Pile}, a large-scale synthetic dataset constructed under the DRU (Diversity, Reliability, Usability) principle via a novel data synthesis pipeline. To optimize training effectiveness, we propose \textbf{Annealing}, a curriculum learning strategy that enables effective knowledge transfer across heterogeneous sources of data. We evaluate CodeR based on 16 diverse code retrieval tasks, where it significantly outperforms existing baselines and exhibits strong out-of-domain generalization performance. We have publicly released our dataset and the well-trained model to facilitate further research in this critical area\footnote{\url{https://huggingface.co/nebula2025}}.



Paperid:2382
Authors:Tianhao Chu, Yuling Wu, Neil Burgess, Zilong Ji, Si Wu
Abstract:
Path integration is essential for spatial navigation. Experimental studies have identified neural correlates for path integration, but exactly how the neural system accomplishes this computation remains unresolved.Here, we adopt recurrent neural networks (RNNs) trained to perform a path integration task to explore this issue. After training, we borrow neuroscience prior knowledge and methods to unfold the black box of the trained model, including: clarifying neuron types based on their receptive fields, dissecting information flows between neuron groups by pruning their connections, andanalyzing internal dynamics of neuron groups using the attractor framework. Intriguingly, we uncover a hierarchical information processing pathway embedded in the RNN model, along which velocity information of an agent is first forwarded to band cells, band and grid cells then coordinate to carry out path integration, and finally grid cells output the agent location. Inspired by the RNN-based study, we construct a neural circuit model, in which band cells form one-dimensional (1D) continuous attractor neural networks (CANNs) and serve as upstream neurons to support downstream grid cells to carry out path integration in the 2D space. Our study challenges the conventional view of considering only grid cells for path integration, and supports a neural circuit model with the hierarchy of band and grid cells.



Paperid:2383
Authors:Xinhao Zheng, Xinhao Song, Bolin Qiu, Yang Li, Zhongteng Gui, Junchi Yan
Abstract:
The Boolean Satisfiability Problem (SAT) plays a crucial role in cryptanalysis, enabling tasks like key recovery and distinguisher construction. Conflict-Driven Clause Learning (CDCL) has emerged as the dominant paradigm in modern SAT solving, and machine learning has been increasingly integrated with CDCL-based SAT solvers to tackle complex cryptographic problems. However, the lack of a unified evaluation framework, inconsistent input formats, and varying modeling approaches hinder fair comparison. Besides, cryptographic SAT instances also differ structurally from standard SAT problems, and the absence of standardized datasets further complicates evaluation. To address these issues, we introduce SAT4CryptoBench, the first comprehensive benchmark for assessing machine learning–based solvers in cryptanalysis. SAT4CryptoBench provides diverse SAT datasets in both Arithmetic Normal Form (ANF) and Conjunctive Normal Form (CNF), spanning various algorithms, rounds, and key sizes. Our framework evaluates three levels of machine learning integration: standalone distinguishers for instance classification, heuristic enhancement for guiding solving strategies, and hyperparameter optimization for adapting to specific problem distributions. Experiments demonstrate that ANF-based networks consistently achieve superior performance over CNF-based networks in learning cryptographic features. Nonetheless, current ML techniques struggle to generalize across algorithms and instance sizes, with computational overhead potentially offsetting benefits on simpler cases. Despite this, ML-driven optimization strategies notably improve solver efficiency on cryptographic SAT instances. Besides, we propose BASIN, a bitwise solver taking plaintext-ciphertext bitstrings as inputs. Crucially, its superior performance on high-round problems highlights the importance of input modeling and the advantage of direct input representations for complex cryptographic structures.



Paperid:2384
Authors:Wenjie Wang, Mingming Gong, Biwei Huang, James Bailey, Bo Han, Kun Zhang, Feng Liu
Abstract:
Conditional independence (CI) testing is a fundamental yet challenging task in modern statistics and machine learning. One pivotal class of methods for assessing conditional independence encompasses kernel-based approaches, known for assessing CI by detecting general conditional dependence without imposing strict assumptions on relationships or data distributions.As with any method utilizing kernels, selecting appropriate kernels is crucial for precise identification. However, it remains underexplored in kernel-based CI methods, where the kernels are often determined manually or heuristically.In this paper, we analyze and propose a kernel parameter selection approach for the kernel-based conditional independence test (KCI).The kernel parameters are selected based on the ratio of the statistic to the asymptotic variance, which approximates the test power for the given parameters at large sample sizes. The search procedure is grid-based, allowing for parallelization with manageable additional computation time.We theoretically demonstrate the consistency of the proposed criterion and conduct extensive experiments on both synthetic and real data to show the effectiveness of our method.



Paperid:2385
Authors:Raphaël Pellegrin, Lukas Fesser, Melanie Weber
Abstract:
Higher-order information is crucial for relational learning in many domains where relationships extend beyond pairwise interactions. Hypergraphs provide a natural framework for modeling such relationships, which has motivated recent extensions of graph neural network architectures to hypergraphs. Most of these architectures rely on message-passing to encode higher-order information. In this paper, we propose to instead use hypergraph-level encodings based on characteristics such as hypergraph Laplacians and discrete curvature notions. These encodings can be used on datasets that are naturally parametrized as hypergraphs and on graph-level datasets, which we reparametrize as hypergraphs to compute encodings. In both settings, performance increases significantly, on social networks by more than 10 percent. Our theoretical analysis shows that hypergraph-level encodings provably increase the representational power of message-passing graph neural networks beyond that of their graph-level counterparts. For complete reproducibility, we release an anonymized version of our codebase: https://anonymous.4open.science/r/Hypergraph_Encodings.



Paperid:2386
Authors:Jing Dong, Baoxiang Wang, Yaoliang Yu
Abstract:
Abstract:We study the problem of no-regret learning algorithms for general monotone and smooth games and their last-iterate convergence properties. Specifically, we investigate the problem under bandit feedback and strongly uncoupled dynamics, which allows modular development of the multi-player system that applies to a wide range of real applications. We propose a mirror-descent-based algorithm, which converges in $O(T^{-1/4})$ and is also no-regret. The result is achieved by a dedicated use of two regularizations and the analysis of the fixed point thereof. The convergence rate is further improved to $O(T^{-1/2})$ in the case of strongly monotone games.Motivated by practical tasks where the game evolves over time, the algorithm is extended to time-varying monotone games. We provide the first non-asymptotic result in converging monotone games and give improved results for equilibrium tracking games.



Paperid:2387
Authors:Honglian Wang, Sijing Tu, Lutz Oettershagen, Aristides Gionis
Abstract:
Abstract:We explore a novel problem in streaming submodular maximization, inspired by the dynamics of news-recommendation platforms.We consider a setting where users can visit a news web\-site at any time, and upon each visit, the web\-site must display up to $k$ news items. User interactions are inherently stochastic: each news item presented to the user is consumed with a certain acceptance probability by the user, and each news item covers certain topics. Our goal is to design a streaming algorithm that maximizes the expected total topic coverage.To address this problem, we establish a connection to submodular maximization subject to a matroid constraint.We show that we can effectively adapt previous methods to address our problem when the number of user visits is known in advance or linear-size memory in the stream length is available.However, in more realistic scenarios where only an upper bound on the visits and sublinear memory is available, the algorithms fail to guarantee any bounded performance.To overcome these limitations, we introduce a new online streaming algorithm that achieves a competitive ratio of $1/8\delta$, where $\delta$ controls the approximation quality. Moreover, it requires only a single pass over the stream, and uses memory independent of the stream length. Empirically, our algorithms consistently outperform the baselines.



Paperid:2388
Authors:Jiaqi Ding, Tingting Dan, Zhixuan Zhou, Guorong Wu
Abstract:
Abstract:In both neuroscience and artificial intelligence (AI), it is well-established that neural “coupling” gives rise to dynamically distributed systems. These systems exhibit self-organized spatiotemporal patterns of synchronized neural oscillations, enabling the representation of abstract concepts. By capitalizing on the unprecedented amount of human neuroimaging data, we hypothesize that a deeper theoretical insight into the neural mechanisms governing brain rhythms could inspire the next-generation design principles for machine learning algorithms with greater efficiency and robustness. To that end, we introduce a physics-informed deep learning framework for $\underline{B}rain$ $\underline{R}hythm$ $\underline{I}dentification$ by $\underline{K}uramoto$ and $\underline{C}ontrol$ (coined *BRICK*) to characterize the synchronization of neural oscillations that shapes the dynamics of evolving cognitive states. Given that graph data, like brain networks, is structurally interconnected and behaviorally dynamic, we propose that the success of modeling brain rhythms via an artificial dynamic system of coupled oscillations gives rise to the foundational principle for the development of brain-inspired machine intelligence through the common mechanism of synchronization. By conceptualizing each node as an oscillator coupled with neighbors, we push the limits of conventional graph neural networks from the cliché of graph heat diffusion to a new realm of compressing graph feature representation via oscillatory synchronization. The experiments show that our new graph learning mechanism not only effectively addresses the over-smoothing issue in GNNs but also manifests the potential of reasoning and solving challenging problems on graphs.



Paperid:2389
Authors:Marlon Becker, Frederick Altrock, Benjamin Risse
Abstract:
The recently proposed optimization algorithm for deep neural networks Sharpness Aware Minimization (SAM) suggests perturbing parameters before gradient calculation by a gradient ascent step to guide the optimization into parameter space regions of flat loss.While significant generalization improvements and thus reduction of overfitting could be demonstrated, the computational costs are doubled due to the additionally needed gradient calculation, making SAM unfeasible in case of limited computationally capacities.Motivated by Nesterov Accelerated Gradient (NAG) we propose Momentum-SAM (MSAM), which perturbs parameters in the direction of the accumulated momentum vector to achieve low sharpness without significant computational overhead or memory demands over SGD or Adam.We evaluate MSAM in detail and reveal insights on separable mechanisms of NAG, SAM and MSAM regarding training optimization and generalization.



Paperid:2390
Authors:Eitan Levin, Yuxin Ma, Mateo Diaz, Soledad Villar
Abstract:
Many modern learning tasks require models that can take inputs of varying sizes. Consequently, dimension-independent architectures have been proposed for domains where the inputs are graphs, sets, and point clouds. Recent work on graph neural networks has explored whether a model trained on low-dimensional data can transfer its performance to higher-dimensional inputs. We extend this body of work by introducing a general framework for transferability across dimensions. We show that transferability corresponds precisely to continuity in a limit space formed by identifying small problem instances with equivalent large ones. This identification is driven by the data and the learning task. We instantiate our framework on existing architectures and implement the necessary changes to ensure their transferability. Finally, we provide design principles for designing new transferable models. Numerical experiments support our theoretical findings.



Paperid:2391
Authors:Sanjana Ramprasad, Byron Wallace
Abstract:
Modern LLMs can now produce highly readable abstractive summaries, to the point that traditional automated metrics for evaluating summary quality, such as ROUGE, have saturated. However, LLMs still sometimes introduce inaccuracies into summaries, i.e., information inconsistent with or unsupported by the corresponding source. Measuring the occurrence of these often subtle factual inconsistencies automatically has proved challenging. This in turn has motivated development of metrics intended to measure the factual consistency of generated summaries against sources. But are these approaches measuring what they purport to? Or are they mostly exploiting artifacts? In this work, we stress test a range of automatic factuality metrics—including specialized model-based approaches and LLM-based prompting methods—to probe what they actually capture. Using a shallow classifier to separate “easy” examples for factual evaluation—where surface features suffice—from “hard” cases requiring deeper reasoning, we find that all metrics show substantial performance drops on the latter.Furthermore, some metrics are more sensitive to benign, fact-preserving edits than to factual corrections. Building on this observation, we demonstrate that most automatic factuality metrics can be gamed—that is, their scores can be artificially inflated by appending innocuous, content-free sentences to summaries. Among the metrics tested, the LLM prompt-based ChatGPT-DA approach is the most robust and reliable; however, it exhibits a notable caveat: it likely relies more on pretraining knowledge than on the provided source when making judgments. Taken together, our findings call into question the reliability of current factuality metrics and prompt a broader reflection on what these metrics are truly measuring. We conclude with concrete recommendations for improving both benchmark design and metric robustness, particularly in light of their vulnerability to superficial manipulations.



Paperid:2392
Authors:Lukas Aichberger, Alasdair Paren, Guohao Li, Philip Torr, Yarin Gal, Adel Bibi
Abstract:
Recent advances in operating system (OS) agents have enabled vision-language models (VLMs) to directly control a user’s computer. Unlike conventional VLMs that passively output text, OS agents autonomously perform computer-based tasks in response to a single user prompt. OS agents do so by capturing, parsing, and analysing screenshots and executing low-level actions via application programming interfaces (APIs), such as mouse clicks and keyboard inputs. This direct interaction with the OS significantly raises the stakes, as failures or manipulations can have immediate and tangible consequences. In this work, we uncover a novel attack vector against these OS agents: malicious image patches (MIPs), adversarially perturbed screen regions that, when captured by an OS agent, induce it to perform harmful actions by exploiting specific APIs. For instance, a MIP can be embedded in a desktop wallpaper or shared on social media to cause an OS agent to exfiltrate sensitive user data. We show that MIPs generalise across user prompts and screen configurations, and that they can hijack multiple OS agents even during the execution of benign instructions. These findings expose critical security vulnerabilities in OS agents that have to be carefully addressed before their widespread deployment.



Paperid:2393
Authors:Xin Sun, Heng Zhou, Chao Li
Abstract:
Reconstruction-based methods are competitive choices for multivariate time series anomaly detection (MTS AD). However, one challenge these methods may suffer is over generalization, where abnormal inputs are also well reconstructed. In addition, balancing robustness and sensitivity is also important for final performance, as robustness ensures accurate detection in potentially noisy data, while sensitivity enables early detection of subtle anomalies. To address these problems, inspired by idempotent generative network, we take the view from the manifold and propose a novel module namedIdempotentGeneration forAnomalyDetection (IGAD) which can be flexibly combined with a reconstruction-based method without introducing additional trainable parameters. We modify the manifold to make sure that normal time points can be mapped onto it while tightening it to drop out abnormal time points simultaneously. Regarding the latest findings of AD metrics, we evaluated IGAD on various methods with four real-world datasets, and they achieve visible improvements in VUS-PR than their predecessors, demonstrating the effective potential of IGAD for further improvements in MTS AD tasks. Our codes are available at https://anonymous.4open.science/r/IGAD.



Paperid:2394
Authors:渊 过, Qian Ma, Hui Li, Qiao Ning, Furui Zhan, Yu Gu, Ge Yu, Shikai Guo
Abstract:
Abstract:Multi-modal Knowledge Graph Completion (MMKGC) aims to predict missing entities, relations, or attributes in knowledge graphs by collaboratively modeling the triple structure and multimodal information (e.g., text, images, videos) associated with entities.This approach facilitates the automatic discovery of previously unobserved factual knowledge.However, existing MMKGC methods encounter several critical challenges: (i) the imbalance of inter-entity information across different modalities; (ii) the heterogeneity of intra-entity multimodal information; and (iii) the inconsistency in the volume of information represented by various modalities for a given entity.In this paper, we propose a novel $\underline{L}$arge model-driven $\underline{B}$alanced $\underline{M}$ultimodal $\underline{K}$nowledge $\underline{G}r$aph $\underline{C}$ompletion framework, termed LBMKGC.Initially, LBMKGC employs the Stable Diffusion XL (a large language model) to augment the imbalanced information across modalities. Subsequently, to bridge the semantic gap between heterogeneous modalities, LBMKGC aligns the multimodal embeddings of entities semantically by using the CLIP (Contrastive Language-Image Pre-Training) model.Furthermore, by distinguishing between the perceptual and conceptual attributes of entities, LBMKGC learns entity representations through the adaptive fusion of embeddings from various modalities, guided by structural information.Finally, extensive experiments conducted against 21 state-of-the-art baselines demonstrate that LBMKGC achieves superior performance across diverse datasets and scenarios while maintaining efficiency and generalizability.Our code and data are publicly available at: https://github.com/guoynow/LBMKGC.



Paperid:2395
Authors:Wuyang Li, Zhu Yu, Alexandre Alahi
Abstract:
Abstract:Camera-based Semantic Scene Completion (SSC) aims to reconstruct the 3D geometry and semantics of the surrounding environment. With dense voxel labels, prior works typically formulate SSC as a *dense segmentation task*, independently classifying each voxel. However, this paradigm neglects critical instance-centric discriminability, leading to instance-level incompleteness and adjacent ambiguities. To address this, we highlight a "free lunch" of SSC labels: the voxel-level class label has implicitly told the instance-level insight, which is ever-overlooked by the community. Motivated by this observation, we first introduce a training-free **Voxel-to-Instance (VoxNT) trick**: a simple yet effective method that freely converts voxel-level class labels into instance-level offset labels. Building on this, we further propose **VoxDet**, an instance-centric framework that reformulates the voxel-level SSC as *dense object detection* by decoupling it into two sub-tasks: offset regression and semantic prediction. Specifically, based on the lifted 3D volume, VoxDet first uses (a) Spatially-decoupled Voxel Encoder to generate disentangled feature volumes for the two sub-tasks, which learn task-specific spatial deformation in the densely projected tri-perceptive space. Then, we deploy (b) Task-decoupled Dense Predictor to address SSC via dense detection. Here, we first regress a 4D offset field to estimate distances (6 directions) between voxels and the corresponding object boundaries in the voxel space. The regressed offsets are then used to guide the instance-level aggregation in the classification branch, achieving instance-aware scene completion. Compared with the state-of-the-art method, VoxDet achieves 11.0% and 6.7% relative mIoU gains on the test set of Semantic KITTI and SSCBench-KITTI-360, respectively, while reducing 57.9% model parameters with around 1.3$\times$ speed-up. The code will be released to propel the SSC and broader occupancy community.



Paperid:2396
Authors:Hongyi Chen, Jianhai Shu, Jingtao Ding, Yong Li, Xiao-Ping (Steven) Zhang
Abstract:
Langevin dynamics sampling suffers from extremely low generation speed, fundamentally limited by numerous fine-grained iterations to converge to target distribution. We introduce PID-controlled Langevin Dynamics (PIDLD), a novel sampling acceleration algorithm that reinterprets the sampling process through control theory principles. By treating energy gradients as feedback signals, PIDLD combines historical gradients (integral term) and gradient trends (derivative term) to efficiently traverse energy landscapes and adaptatively achieve stablization, significantly reducing required iterations to produce quality samples. Our approach requires no additional training, datasets, or prior information, making it immediately integrable with any Langevin-based method. Extensive experiments across image generation and reasoning tasks demonstrate that PIDLD achieves higher quality with fewer steps, making Langevin-based generative models more practical for efficiency-critical applications. The implementation can be found at https://anonymous.4open.science/r/29CEE7A3340BEA02BB5F37623372E72C



Paperid:2397
Authors:Zhiwei Hao, Zhongyu Xiao, Jianyuan Guo, Li Shen, Yong Luo, Han Hu, Dan Zeng
Abstract:
Recent advances in multimodal semantic segmentation show that incorporating auxiliary inputs—such as depth or thermal images—can significantly improve performance over single-modality (RGB-only) approaches. However, most existing solutions rely on parallel backbone networks and complex fusion modules, greatly increasing model size and computational demands. Inspired by prompt tuning in large language models, we introduce \textbf{MixPrompt}: a prompting-based framework that integrates auxiliary modalities into a pretrained RGB segmentation model without modifying its architecture. MixPrompt uses a lightweight prompting module to extract and fuse information from auxiliary inputs into the main RGB backbone. This module is initialized using the early layers of a pretrained RGB feature extractor, ensuring a strong starting point. At each backbone layer, MixPrompt aligns RGB and auxiliary features in multiple low-rank subspaces, maximizing information use with minimal parameter overhead. An information mixing scheme enables cross-subspace interaction for further performance gains. During training, only the prompting module and segmentation head are updated, keeping the RGB backbone frozen for parameter efficiency. Experiments across NYU Depth V2, SUN-RGBD, MFNet, and DELIVER datasets show that MixPrompt achieves improvements of 4.3, 1.1, 0.4, and 1.1 mIoU, respectively, over two-branch baselines, while using nearly half the parameters. MixPrompt also outperforms recent prompting-based methods under similar compute budgets.



Paperid:2398
Authors:Chenxiao Fan, Chongming Gao, Wentao Shi, Yaxin Gong, Zhao Zihao, Fuli Feng
Abstract:
Accurate and safe medication recommendations are critical for effective clinical decision-making, especially in multimorbidity cases. However, existing systems rely on point-wise prediction paradigms that overlook synergistic drug effects and potential adverse drug-drug interactions (DDIs). We propose FLAME, a fine-grained list-wise alignment framework for large language models (LLMs), enabling drug-by-drug generation of drug lists. FLAME formulates recommendation as a sequential decision process, where each step adds or removes a single drug. To provide fine-grained learning signals, we devise step-wise Group Relative Policy Optimization (GRPO) with potential-based reward shaping, which explicitly models DDIs and optimizes the contribution of each drug to the overall prescription. Furthermore, FLAME enhances patient modeling by integrating structured clinical knowledge and collaborative information into the representation space of LLMs. Experiments on benchmark datasets demonstrate that FLAME achieves state-of-the-art performance, delivering superior accuracy, controllable safety–accuracy trade-offs, and strong generalization across diverse clinical scenarios. Our code is available at https://anonymous.4open.science/r/FLAME-NeurIPS2025.



Paperid:2399
Authors:Rizhen Hu, Yutong He, Ran Yan, Mou Sun, Binhang Yuan, Kun Yuan
Abstract:
Abstract:As distributed optimization scales to meet the demands of Large Language Model (LLM) training, hardware failures become increasingly non-negligible. Existing fault-tolerant training methods often introduce significant computational or memory overhead, demanding additional resources. To address this challenge, we propose **Me**mory- and **C**omputation- **e**fficient **F**ault-tolerant **O**ptimization (**MeCeFO**), a novel algorithm that ensures robust training with minimal overhead. When a computing node fails, MeCeFO seamlessly transfers its training task to a neighboring node while employing memory- and computation-efficient algorithmic optimizations to minimize the extra workload imposed on the neighboring node handling both tasks. MeCeFO leverages three key algorithmic designs: (i) Skip-connection, which drops the multi-head attention (MHA) module during backpropagation for memory- and computation-efficient approximation; (ii) Recomputation, which reduces activation memory in feedforward networks (FFNs); and (iii) Low-rank gradient approximation, enabling efficient estimation of FFN weight matrix gradients. Theoretically, MeCeFO matches the convergence rate of conventional distributed training, with a rate of $\mathcal{O}(1/\sqrt{nT})$, where $n$ is the data parallelism size and $T$ is the number of iterations. Empirically, MeCeFO maintains robust performance under high failure rates, incurring only a 4.18\% drop in throughput, demonstrating $5.0\times$ to $6.7\times$ greater resilience than previous SOTA approaches.



Paperid:2400
Authors:Weicheng Wang, Guoli Jia, Xialei Liu, Liang Lin, Jufeng Yang
Abstract:
Continual learning seeks to enable a model to assimilate knowledge from non-stationary data streams without catastrophic forgetting. Recently, methods based on Parameter-Efficient Tuning (PET) have achieved superior performance without even storing any historical exemplars, which train much fewer specific parameters for each task upon a frozen pre-trained model, and tailored parameters are retrieved to guide predictions during inference. However, reliance solely on pre-trained features for parameter matching exacerbates the inconsistency between the training and inference phases, thereby constraining the overall performance. To address this issue, we propose HRM-PET, which makes full use of the richer downstream knowledge inherently contained in the trained parameters. Specifically, we introduce a hybrid re-matching mechanism, which benefits from the initial predicted distribution to facilitate the parameter selections. The direct re-matching addresses misclassified samples identified with correct task identity in prediction, despite incorrect initial matching. Moreover, the confidence-based re-matching is specifically designed to handle other more challenging mismatched samples that cannot be calibrated by the former. Besides, to acquire task-invariant knowledge for better matching, we integrate a cross-task instance relationship distillation module into the PET-based method. Extensive experiments conducted on four datasets under five pre-trained settings demonstrate that HRM-PET performs favorably against the state-of-the-art methods. The source code is provided in the supplementary materials and will be publicly available.



Paperid:2401
Authors:Xiaofeng Tan, Hongsong Wang, Xin Geng, Pan Zhou
Abstract:
Text-to-motion generation is essential for advancing the creative industry but often presents challenges in producing consistent, realistic motions. To address this, we focus on fine-tuning text-to-motion models to consistently favor high-quality, human-preferred motions—a critical yet largely unexplored problem. In this work, we theoretically investigate the DPO under both online and offline settings, and reveal their respective limitation: overfitting in offline DPO, and biased sampling in online DPO. Building on our theoretical insights, we introduce Semi-online Preference Optimization (SoPo), a DPO-based method for training text-to-motion models using ``semi-online” data pair, consisting of unpreferred motion from online distribution and preferred motion in offline datasets. This method leverages both online and offline DPO, allowing each to compensate for the other’s limitations. Extensive experiments demonstrate that SoPo outperforms other preference alignment methods, with an MM-Dist of 3.25\% (vs e.g. 0.76\% of MoDiPO) on the MLD model, 2.91\% (vs e.g. 0.66\% of MoDiPO) on MDM model, respectively. Additionally, the MLD model fine-tuned by our SoPo surpasses the SoTA model in terms of R-precision and MM Dist. Visualization results also show the efficacy of our SoPo in preference alignment. Code will be released publicly.



Paperid:2402
Authors:Cai Zhou, Chenyu Wang, Dinghuai Zhang, Shangyuan Tong, Yifei Wang, Stephen Bates, Tommi Jaakkola
Abstract:
In this paper we introduce Hierarchical Diffusion Language Models (HDLM) -- a novel family of discrete diffusion models for language modeling. HDLM builds on a hierarchical vocabulary where low-level tokens with detailed semantics are surjectively mapped to high-level tokens with coarse-grained meanings. In the forward process, each token is independently perturbed to its higher-level ancestor with more abstract semantics according to the scheduler, while in the reverse process the model progressively predicts the next, more detailed semantics. Taken together, HDLM provides a general time-varying next semantic scale prediction process for language modeling. We derive closed-form expressions for the diffusion Evidence Lower Bound (ELBO), and show that HDLM can be implemented in a flexible manner and that it includes an existing discrete diffusion model as a special case. Extensive text generation experiments demonstrate the effectiveness of HDLM.



Paperid:2403
Authors:Raphael Romero, Tijl De Bie, Nick Heard, Alexander Modell
Abstract:
Abstract:Detecting structural change in dynamic network data has wide-ranging applications. Existing approaches typically divide the data into time bins, extract network features within each bin, and then compare these features over time. This introduces an inherent tradeoff between temporal resolution and the statistical stability of the extracted features. Despite this tradeoff, reminiscent of time–frequency tradeoffs in signal processing, most methods rely on a \emph{fixed temporal resolution}. Choosing an appropriate resolution parameter is typically difficult, and can be especially problematic in domains like cybersecurity, where anomalous behavior may emerge at multiple time scales. We address this challenge by proposing ANIE ($\textbf{A}$daptive $\textbf{N}$etwork $\textbf{I}$ntensity $\textbf{E}$stimation), a multi-resolution framework designed to automatically identify the time scales at which network structure evolves, enabling the joint detection of both rapid and gradual changes. Modeling interactions as Poisson processes, our method proceeds in two steps: (1) estimating a low-dimensional subspace of node behavior, and (2) deriving a set of novel *empirical affinity coefficients* that measure change in interaction intensity between latent factors and support statistical testing for structural change across time scales. We provide theoretical guarantees for subspace estimation and the asymptotic behavior of the affinity coefficients, enabling model-based change detection. Experiments on synthetic networks show that ANIE adapts to the appropriate time resolution, and is able to capture sharp structural changes while remaining robust to noise. Furthermore, applications to real-world data showcase the practical benefits of ANIE’s multiresolution approach to detecting structural change over fixed resolution methods. An open-source implementation of the method is available at [https://anonymous.4open.science/r/anie-0ABB/].



Paperid:2404
Authors:Kai Liu, Jungang Li, Yuchong Sun, Shengqiong Wu, jianzhang gao, Daoan Zhang, Wei Zhang, Sheng Jin, Sicheng Yu, Geng Zhan, Jiayi Ji, Fan Zhou, Liang Zheng, Shuicheng Yan, Hao Fei, Tat-Seng Chua
Abstract:
While multimodal large language models (MLLMs) have advanced visual-language understanding, unified comprehension and generation of synchronized audio and video—crucial for real-world scenarios like filmmaking—remains largely underexplored. Existing approaches often overlook fine-grained spatio-temporal alignment across modalities, limiting their applicability to sounding-video tasks. In this work, we introduce JavisGPT, the first unified MLLM for Joint Audio-Video (JAV) comprehension and generation. JavisGPT adopts an encoder–LLM–decoder architecture, featuring a SyncFusion module for spatio-temporal audio-video fusion and hierarchical semantic and synchrony queries to bridge a pretrained JAV-DiT generator. This design enables temporally coherent video-audio generation from multimodal instructions. We design an effective three-stage training pipeline consisting of multimodal pretraining, audio-video fine-tuning, and large-scale instruction-tuning, to progressively build multimodal comprehension and generation. To support this, we construct JavisInst-Omni, a high-quality instruction dataset with over 200K GPT-4o-curated audio-video-text dialogues that span diverse and multi-level comprehension and generation scenarios. Extensive experiments on JAV comprehension and generation benchmarks show that JavisGPT outperforms existing MLLMs, particularly in complex and temporally synchronized settings.



Paperid:2405
Authors:Duy Nguyen, Archiki Prasad, Elias Stengel-Eskin, Mohit Bansal
Abstract:
Abstract:Reward Models (RMs) are crucial to aligning large language models (LLMs), but the degree to which an RM specialized to one task (e.g. writing) generalizes to new tasks (e.g. math) is often not known a priori, often making using only one fixed RM to train LLMs suboptimal. However, optimizing LLMs with multiple RMs simultaneously can incur a prohibitively high computational cost and lead to conflicting signals from different RMs that may degrade performance. To address these challenges, we introduce LASeR (Learning to Adaptively Select Rewards), which frames reward model selection as a multi-armed bandit problem, iteratively and efficiently training LLMs using multiple RMs by selecting the most well-suited RM for each instance. On commonsense and math reasoning tasks, we show that LASeR boosts iterative LLM training, improving the absolute average accuracy of Llama-3-8B over three datasets by $2.67$% over an ensemble of RM scores while also showing superior efficiency (e.g., a $2\times$ speedup). Moreover, on WildChat (open-ended instruction-following tasks), LASeR leads to a $72.69$% AlpacaEval win rate over the RM score ensemble baseline. Extending to long-context generation, LASeR improves by $2.96$ F1 points (avg.) on single-document QA tasks and $2.97$ F1 points on few-shot learning over the RM score ensemble baseline with best-of-$n$ sampling. We include our code in the supplementary.



Paperid:2406
Authors:Ben Finkelshtein, Ismail Ceylan, Michael Bronstein, Ron Levie
Abstract:
Graph machine learning architectures are typically tailored to specific tasks on specific datasets, which hinders their broader applicability. This has led to a new quest in graph machine learning: how to build graph foundation models capable of generalizing across arbitrary graphs and features? In this work, we present a recipe for designing graph foundation models for node-level tasks from first principles. The key ingredient underpinning our study is a systematic investigation of the symmetries that a graph foundation model must respect. In a nutshell, we argue that label permutation-equivariance alongside feature permutation-invariance are necessary in addition to the common node permutation-equivariance on each local neighborhood of the graph. To this end, we first characterize the space of linear transformations that are equivariant to permutations of nodes and labels, and invariant to permutations of features. We then prove that the resulting network is a universal approximator on multisets that respect the aforementioned symmetries. Our recipe uses such layers on the multiset of features induced by the local neighborhood of the graph to obtain a class of graph foundation models for node property prediction. We validate our approach through extensive experiments on 27 real-world node classification datasets, demonstrating the strong empirical performance of the proposed architectures.



Paperid:2407
Authors:Liang Yue, Yihong Tang, Kehai Chen, Jie Liu, Min Zhang
Abstract:
Instruction fine-tuning is crucial in NLP tasks, enhancing pretrained models' instruction-following capabilities and task-specific performance. However, obtaining high-quality fine-tuning data for large models is challenging due to data collection difficulties and high production costs. To address this, we propose MASTER, a novel data augmentation method that enriches original data through interactions among multiple agents with varying cognitive levels. We simulate three pedagogically grounded teaching scenarios, leveraging multi-agent conversations to generate high-quality teacher-student interaction data. Utilizing MASTER, we construct BOOST-QA, a fine-tuning dataset augmented from existing datasets like Orca-Math-200k, ProcQA, and OpenHermes2.5. Experiments show that models fine-tuned with BOOST-QA perform excellently across multiple benchmarks, demonstrating strong multitask generalization. Notably, MASTER significantly improves models' reasoning abilities in complex tasks, providing valuable insights for future research.



Paperid:2408
Authors:Lijia Yu, Xiao-Shan Gao, Lijun Zhang
Abstract:
It has been shown that the chain of thought (CoT) can enhance the power of LLMs to simulate a Turing machine or an algorithm, and in particular its mathematical reasoning ability. The memorization capability of LLMs is an important aspect of their expressive ability, which offers valuable insight into designing models with enhanced generalization potential. Currently, the optimal memorization capacities of transformers have been established for both the general dataset and the dataset that satisfies a specific separability condition. However, the question of whether the CoT can improve the memorization capability of LLMs remains unexamined. To fill this gap, we establish the memorization capability for fixed-precision autoregressive transformers with or without CoT. Precisely, we first give the necessary and sufficient conditions for transformers to memorize a finite language and then provide the upper and lower bounds for the number of parameters of the memorization transformers. Our result indicates that the classes of languages that can be memorized by transformers with or without CoT do not contain each other, and the same number of parameters is needed for transformers with or without CoT to memorize, implying that CoT does not enhance a transformer’s memorization power significantly. We further show that CoT can not help transformers to memory certain infinite languages.



Paperid:2409
Authors:Benjamin J Li, Shuyang Shi, Lucia Romero, Huao Li, Yaqi Xie, Woojun Kim, Stefanos Nikolaidis, Charles Lewis, Katia Sycara, Simon Stepputtis
Abstract:
Adaptation is the cornerstone of effective collaboration among heterogeneous team members. In human-agent teams, artificial agents need to adapt to their human partners in real time, as individuals often have unique preferences and policies that may change dynamically throughout interactions. This becomes particularly challenging in tasks with time pressure and complex strategic spaces, where identifying partner behaviors and selecting suitable responses is difficult.In this work, we introduce a strategy-conditioned cooperator framework that learns to represent, categorize, and adapt to a broad range of potential partner strategies in real-time. Our approach encodes strategies with a variational autoencoder to learn a latent strategy space from agent trajectory data, identifies distinct strategy types through clustering, and trains a cooperator agent conditioned on these clusters by generating partners of each strategy type.For online adaptation to novel partners, we leverage a fixed-share regret minimization algorithm that dynamically infers and adjusts the partner's strategy estimation during interaction.We evaluate our method in a modified version of the Overcooked domain, a complex collaborative cooking environment that requires effective coordination among two players with a diverse potential strategy space.Through these experiments and an online user study, we demonstrate that our proposed agent achieves state of the art performance compared to existing baselines when paired with novel human, and agent teammates.



Paperid:2410
Authors:Guan Zhe Hong, Nishanth Dikkala, Enming Luo, Cyrus Rashtchian, Xin Wang, Rina Panigrahy
Abstract:
Due to the size and complexity of modern large language models (LLMs), it has proven challenging to uncover the underlying mechanisms that models use to solve reasoning problems. For instance, is their reasoning for a specific problem localized to certain parts of the network? Do they break down the reasoning problem into modular components that are then executed as sequential steps as we go deeper in the model? To better understand the reasoning capability of LLMs, we study a minimal propositional logic problem that requires combining multiple facts to arrive at a solution. By studying this problem on Mistral and Gemma models, up to 27B parameters, we illuminate the core components the models use to solve such logic problems. From a mechanistic interpretability point of view, we use causal mediation analysis to uncover the pathways and components of the LLMs' reasoning processes. Then, we offer fine-grained insights into the functions of attention heads in different layers. We not only find a sparse circuit that computes the answer, but we decompose it into sub-circuits that have four distinct and modular uses. Finally, we reveal that three distinct models -- Mistral-7B, Gemma-2-9B and Gemma-2-27B -- contain analogous but not identical mechanisms.



Paperid:2411
Authors:Shunta Akiyama
Abstract:
In this paper, we consider a block coordinate descent (BCD) algorithm for training deep neural networks and provide a new global convergence guarantee under strictly monotonically increasing activation functions. While existing works demonstrate convergence to stationary points for BCD in neural networks, our contribution is the first to prove convergence to global minima, ensuring arbitrarily small loss. We show that the loss with respect to the output layer decreases exponentially while the loss with respect to the hidden layers remains well-controlled. Additionally, we derive generalization bounds using the Rademacher complexity framework, demonstrating that BCD not only achieves strong optimization guarantees but also provides favorable generalization performance. Moreover, we propose a modified BCD algorithm with skip connections and non-negative projection, extending our convergence guarantees to ReLU activation, which are not strictly monotonic. Empirical experiments confirm our theoretical findings, showing that the BCD algorithm achieves a small loss for strictly monotonic and ReLU activations.



Paperid:2412
Authors:Dung Nguyen, Aravind Srinivasan, Renata Valieva, Anil Vullikanti, Jiayi Wu
Abstract:
Abstract:Designing effective strategies for controlling epidemic spread by vaccination is an important question in epidemiology, especially in the early stages when vaccines are limited.This is a challenging question when the contact network is very heterogeneous, and strategies based on controlling network properties, such as the degree and spectral radius, have been shown to be effective.Implementation of such strategies requires detailed information on the contact structure, which might be sensitive in many applications.Our focus here is on choosing effective vaccination strategies when the edges are sensitive and differential privacy guarantees are needed.Our main contributions are $(\varepsilon,\delta)$-differentially private algorithms for designing vaccination strategies by reducing the maximum degree and spectral radius.Our key technique is a private algorithm for the multi-set multi-cover problem, which we use for controlling network properties.We evaluate privacy-utility tradeoffs of our algorithms on multiple synthetic and real-world networks, and show their effectiveness.



Paperid:2413
Authors:James Cuin, Davide Carbone, O. Deniz Akyildiz
Abstract:
We utilise a sampler originating from nonequilibrium statistical mechanics, termed here Jarzynski-adjusted Langevin algorithm (JALA), to build statistical estimation methods in latent variable models. We achieve this by leveraging Jarzynski’s equality and developing algorithms based on a weighted version of the unadjusted Langevin algorithm (ULA) with recursively updated weights. Adapting this for latent variable models, we develop a sequential Monte Carlo (SMC) method that provides the maximum marginal likelihood estimate of the parameters, termed JALA-EM. Under suitable regularity assumptions on the marginal likelihood, we provide a nonasymptotic analysis of the JALA-EM scheme implemented with stochastic gradient descent and show that it provably converges to the maximum marginal likelihood estimate. We demonstrate the performance of JALA-EM on a variety of latent variable models and show that it performs comparably to existing methods in terms of accuracy and computational efficiency. Importantly, the ability to recursively estimate marginal likelihoods—an uncommon feature among scalable methods—makes our approach particularly suited for model selection, which we validate through dedicated experiments.



Paperid:2414
Authors:Ziqi Zhou, Yifan Hu, Yufei Song, Zijing Li, Shengshan Hu, Leo Yu Zhang, Dezhong Yao, Long Zheng, Hai Jin
Abstract:
Recent studies reveal the vulnerability of the image segmentation foundation model SAM to adversarial examples. Its successor, SAM2, has attracted significant attention due to its strong generalization capability in video segmentation. However, its robustness remains unexplored, and it is unclear whether existing attacks on SAM can directly transfer to SAM2. In this paper, we first analyze the performance gap of existing attacks between SAM and SAM2 and highlight two key challenges arising from their architectural differences: directional guidance from the prompt and semantic entanglement across consecutive frames. To address these issues, we propose UAP-SAM2, the first cross-prompt universal adversarial attack against SAM2 driven by dual semantic deviation. For the cross-prompt transferability, we begin by designing a target-scanning strategy that divides each frame into k regions, each randomly assigned a prompt, to reduce prompt dependency during optimization. For effectiveness, we design a dual semantic deviation framework that optimizes a UAP by distorting the semantics within the current frame and disrupting the semantic consistency across consecutive frames.Extensive experiments on six datasets across two segmentation tasks to demonstrate the effectiveness of the proposed method for SAM2. Comparative results show that UAP-SAM2 significantly outperforms SOTA attacks by a large margin.



Paperid:2415
Authors:Mohamed Abdelfattah, Kaouther Messaoud, Alexandre Alahi
Abstract:
We present OSKAR, the first multimodal foundation model based on bootstrapped latent feature prediction. Unlike generative or contrastive methods, it avoids memorizing unnecessary details (e.g., pixels), and does not require negative pairs, large memory banks, or hand-crafted augmentations. We propose a novel pretraining strategy: given masked tokens from multiple modalities, predict a subset of missing tokens per modality, supervised by momentum-updated uni-modal target encoders. This design efficiently utilizes the model capacity in learning high-level representations while retaining modality-specific information. Further, we propose a scalable design which decouples the compute cost from the number of modalities using a fixed representative token budget—in both input and target tokens—and introduces a parameter-efficient cross-attention predictor that grounds each prediction in the full multimodal context. Without loss of generality, we instantiate OSKAR on video, skeleton, and text modalities. Extensive experimental results show that OSKAR's unified pretrained encoder outperforms models with specialized architectures of similar size in action recognition (rgb, skeleton, frozen, low-shot) and localization, video-text retrieval, and video question answering.



Paperid:2416
Authors:Steffen Schotthöfer, Timon Klein, Jonas Kusch
Abstract:
Low-rank pre-training and fine-tuning have recently emerged as promising techniques for reducing the computational and storage costs of large neural networks. Training low-rank parameterizations typically relies on conventional optimizers such as heavy ball momentum methods or Adam. In this work, we identify and analyze potential difficulties that these training methods encounter when used to train low-rank parameterizations of weights. In particular, we show that classical momentum methods can struggle to converge to a local optimum due to the geometry of the underlying optimization landscape. To address this, we introduce novel training strategies derived from dynamical low-rank approximation, which explicitly account for the underlying geometric structure. Our approach leverages and combines tools from dynamical low-rank approximation and momentum-based optimization to design optimizers that respect the intrinsic geometry of the parameter space. We validate our methods through numerical experiments, demonstrating faster convergence, and stronger validation metrics at given parameter budgets.



Paperid:2417
Authors:Andrea Agazzi, Eloy Mosig García, Dario Trevisan
Abstract:
Abstract:In this paper, we study the quantitative convergence of shallow neural networks trained via gradient descent to their associated Gaussian processes in the infinite-width limit. While previous work has established qualitative convergence under broad settings, precise, finite-width estimates remain limited, particularly during training. We provide explicit upper bounds on the quadratic Wasserstein distance between the network output and its Gaussian approximation at any training time $t \ge 0$, demonstrating polynomial decay with network width. Our results quantify how architectural parameters, such as width and input dimension, influence convergence, and how training dynamics affect the approximation error



Paperid:2418
Authors:Yanan Li, Fanxu Meng, Muhan Zhang, Shiai Zhu, Shangguang Wang, Mengwei Xu
Abstract:
As Large Language Models (LLMs) are frequently updated, LoRA weights trained on earlier versions quickly become obsolete. The conventional practice of retraining LoRA weights from scratch on the latest model is costly, time-consuming, and environmentally detrimental, particularly as the diversity of LLMs and downstream tasks expands. This motivates a critical question: "How can we efficiently leverage existing LoRA weights to adapt to newer model versions?" To address this, we propose LoRASuite, a modular approach tailored specifically to various types of LLM updates. First, we compute a transfer matrix utilizing known parameters from both old and new LLMs. Next, we allocate corresponding layers and attention heads based on centered kernel alignment and cosine similarity metrics, respectively. A subsequent small-scale, skillful fine-tuning step ensures numerical stability. Experimental evaluations demonstrate that LoRASuite consistently surpasses small-scale vanilla LoRA methods. Notably, on backbone LLMs such as MiniCPM and Qwen, LoRASuite even exceeds the performance of full-scale LoRA retraining, with average improvements of +1.4 and +6.6 points on math tasks, respectively. Additionally, LoRASuite significantly reduces memory consumption by 5.5 GB and computational time by 78.23%.



Paperid:2419
Authors:Kiarash Banihashem, Morteza Monemizadeh, MohammadTaghi Hajiaghayi, Peyman Jabbarzade, Samira Goudarzi, Jeff Giliberti
Abstract:
Abstract:In this paper, we study the fundamental problems of maintaining the diameter and a $k$-center clustering of a dynamic point set $P \subset \mathbb{R}^d$, where points may be inserted or deleted over time and the ambient dimension $d$ is not constant and may be high. Our focus is on designing algorithms that remain effective even in the presence of an \emph{adaptive adversary}—an adversary that, at any time $t$, knows the entire history of the algorithm’s outputs as well as all the random bits used by the algorithm up to that point. We present a fully dynamic algorithm that maintains a $2$-approximate diameter with a \emph{worst-case} update time of $poly(d, \log n)$, where $n$ is the length of the stream. Our result is achieved by identifying a robust representative of the dataset that requires infrequent updates, combined with a careful deamortization. To the best of our knowledge, this is the first efficient fully-dynamic algorithm for diameter in high dimensions that \emph{simultaneously} achieves a $2$-approximation guarantee and robustness against an adaptive adversary. We also give an improved dynamic $(4+\epsilon)$-approximation algorithm for the $k$-center problem, also resilient to an adaptive adversary. Our clustering algorithm achieves an amortized update time of $k^{2.5} d \cdot poly(\epsilon^{-1}, \log n)$, improving upon the amortized update time of $k^6 d \cdot poly( \epsilon^{-1}, \log n)$ by Biabani et al. [NeurIPS'24].



Paperid:2420
Authors:Syamantak Kumar, Dheeraj Nagaraj, Purnamrita Sarkar
Abstract:
Abstract:Diffusion models generate samples by estimating the score function of the target distribution at various noise levels. The model is trained using samples drawn from the target distribution, progressively adding noise. Previous sample complexity bounds have a polynomial dependence on the dimension $d$, apart from $\log({|\mathcal{H}|})$, where $\mathcal{H}$ is the hypothesis class. In this work, we establish the first (nearly) dimension-free sample complexity bounds, modulo any dependence due to $\log( |\mathcal{H}|)$, for learning these score functions, achieving a double exponential improvement in dimension over prior results. A key aspect of our analysis is to use a single function approximator to jointly estimate scores across noise levels, a critical feature in practice which enables generalization across timesteps. We introduce a novel martingale-based error decomposition and sharp variance bounds, enabling efficient learning from dependent data generated by Markov processes, which may be of independent interest. Building on these insights, we propose Bootstrapped Score Matching (BSM), a variance reduction technique that utilizes previously learned scores to improve accuracy at higher noise levels. These results provide crucial insights into the efficiency and effectiveness of diffusion models for generative modeling.



Paperid:2421
Authors:Runze Ma, Zhongyue Zhang, Zichen Wang, Chenqing Hua, Jiahua Rao, Zhuomin Zhou, Shuangjia Zheng
Abstract:
Ribonucleic acid (RNA) binds to molecules to achieve specific biological functions. While generative models are advancing biomolecule design, existing methods for designing RNA that target specific ligands face limitations in capturing RNA’s conformational flexibility, ensuring structural validity, and overcoming data scarcity. To address these challenges, we introduce RiboFlow, a synergistic flow matching model to co-design RNA structures and sequences based on target molecules. By integrating RNA backbone frames, torsion angles, and sequence features in an unified architecture, RiboFlow explicitly models RNA’s dynamic conformations while enforcing sequence-structure consistency to improve validity. Additionally, we curate RiboBind, a large-scale dataset of RNA-molecule interactions, to resolve the scarcity of high-quality structural data. Extensive experiments reveal that RiboFlow not only outperforms state-of-the-art RNA design methods by a large margin but also showcases controllable capabilities for achieving high binding affinity to target ligands. Our work bridges critical gaps in controllable RNA design, offering a framework for structure-aware, data-efficient generation.



Paperid:2422
Authors:Yiming Wang, Lucy Chai, Xuan Luo, Michael Niemeyer, Manuel Lagunas, Stephen Lombardi, Siyu Tang, Tiancheng Sun
Abstract:
Recent advances in feed-forward 3D Gaussian Splatting have led to rapid improvements in efficient scene reconstruction from sparse views. However, most existing approaches construct Gaussian primitives directly aligned with the pixels in one or more of the input images. This leads to redundancies in the representation when input views overlap and constrains the position of the primitives to lie along the input rays without full flexibility in 3D space. Moreover, these pixel-aligned approaches do not naturally generalize to dynamic scenes, where effectively leveraging temporal information requires resolving both redundant and newly appearing content across frames. To address these limitations, we introduce a novel Fuse-and-Refine module that enhances existing feed-forward models by merging and refining the primitives in a canonical 3D space. At the core of our method is an efficient hybrid Splat-Voxel representation – from an initial set of pixel-aligned Gaussian primitives, we aggregate local features into a coarse-to-fine voxel hierarchy, and then use a sparse voxel transformer to process these voxel features and generate refined Gaussian primitives. By fusing and refining an arbitrary number of inputs into a consistent set of primitives, our representation effectively reduces redundancy and naturally adapts to temporal frames, enabling history-aware online reconstruction of dynamic scenes. Trained on large-scale static scene datasets, our model learns an effective global strategy to process around 20k primitives within 15ms and significantly enhances reconstruction quality compared to pixel-aligned reconstruction approaches. Without additional training, our model generalizes to video by fusing primitives across time, yielding a more temporally coherent result compared to baseline methods with graceful handling of occluded content. Our approach achieves state-of-the-art performance in both static and streaming scene reconstructions while running at interactive rates (15 fps with 350ms delay) on a single H100 GPU.



Paperid:2423
Authors:Eunseo Koh, Sangeek Hyun, MinKyu Lee, Jiwoo Chung, Seo Kangmin, Jae-Pil Heo
Abstract:
Recent advances in text-based image editing have motivated the extension of these techniques into the 3D domain. However, existing methods typically apply 2D diffusion models independently to multiple viewpoints, resulting in significant artifacts, most notably the Janus problem, due to inconsistencies across edited views. To address this, we propose a novel approach termed DFFSplat, which integrates a 3D-consistent diffusion feature field into the editing pipeline.By rendering and injecting these 3D-consistent structural features into intermediate layers of a 2D diffusion model, our method effectively enforces geometric alignment and semantic coherence across views.However, averaging 3D features during the feature field learning process can lead to the loss of fine texture details.To overcome this, we introduce a dual-encoder architecture to disentangle view-independent structural information from view-dependent appearance details. By encoding only the disentangled structure into the 3D field and injecting it during 2D editing, our method produces semantically and multi-view coherent edited images while maintaining high text fidelity.Additionally, we employ a time-invariance objective to ensure consistency across diffusion timesteps, enhancing the stability of learned representations.Experimental results demonstrate that our method achieves state-of-the-art performance in terms of text-fidelity, and better preserves structural and semantic consistency compared to existing approaches.



Paperid:2424
Authors:Maximilian Beck, Korbinian Pöppel, Phillip Lippe, Sepp Hochreiter
Abstract:
Linear RNNs with gating recently demonstrated competitive performance compared to Transformers in language modeling. Although their linear compute scaling in sequence length offers theoretical runtime advantages over Transformers, realizing these benefits in practice requires optimized custom kernels, as Transformers rely on the highly efficient Flash Attention kernels (Dao, 2024). Leveraging the chunkwise-parallel formulation of linear RNNs, Flash Linear Attention (FLA) (Yang & Zhang, 2024) shows that linear RNN kernels are faster than Flash Attention, by parallelizing over chunks of the input sequence. However, since the chunk size of FLA is limited, many intermediate states must be materialized in GPU memory. This leads to low arithmetic intensity and causes high memory consumption and IO cost, especially for long-context pre-training. In this work, we present Tiled Flash Linear Attention (TFLA), a novel kernel algorithm for linear RNNs, that enables arbitrary large chunk sizes and high arithmetic intensity by introducing an additional level of sequence parallelization within each chunk. First, we apply TFLA to the xLSTM with matrix memory, the mLSTM (Beck et al., 2024). Second, we propose an mLSTM variant with sigmoid input gate and reduced computation for even faster kernel runtimes at equal language modeling performance. In our speed benchmarks, we show that our new mLSTM kernels based on TFLA outperform highly optimized Flash Attention, Linear Attention and Mamba kernels, setting a new state of the art for efficient long-context sequence modeling primitives.



Paperid:2425
Authors:Haolang Lu, Yilian Liu, Jingxin Xu, Guoshun Nan, Yuanlong Yu, Zhican Chen, Kun Wang
Abstract:
The development of Reasoning Large Language Models (RLLMs) has significantly improved multi-step reasoning capabilities, but it has also made hallucination problems more frequent and harder to eliminate. While existing approaches address hallucination through external knowledge integration, model parameter analysis, or self-verification mechanisms, they fail to provide a comprehensive insight into how hallucinationsemergeandevolvethroughout the reasoning chain. In this work, we investigate hallucination causality under constrained knowledge domains by auditing the Chain-of-Thought (CoT) trajectory and assessing the model's cognitive confidence in potentially erroneous or biased claims.Analysis reveals that in long-CoT settings, RLLMs may iteratively reinforce biases and errors through flawed reflective processes, ultimately inducing hallucinated reasoning paths.Counterintuitively, even with interventions at hallucination origins, reasoning chains display pronounced ''chain disloyalty'', resisting correction and sustaining flawed trajectories.We further point out that existing hallucination detection methods areless reliable and interpretable than previously assumed, especially in complex multi-step reasoning contexts.Unlike Anthropic's circuit tracing that requires access to model parameters, our auditingenables more interpretable long-chain hallucination attribution in black-box settings, demonstrating stronger generalizability and practical utility.Our code is available atthis link.



Paperid:2426
Authors:Thomas Pethick, Wanyun Xie, Mete Erdogan, Kimon Antonakopoulos, Tony Silveti-Falls, Volkan Cevher
Abstract:
Abstract:This work introduces a hybrid non-Euclidean optimization method which generalizes gradient norm clipping by combining steepest descent and conditional gradient approaches. The method achieves the best of both worlds by establishing a descent property under a generalized notion of ($L_0$,$L_1$)-smoothness. Weight decay is incorporated in a principled manner by identifying a connection to the Frank-Wolfe short step. In the stochastic case, we show an order optimal $O(n^{-1/4})$ convergence rate by leveraging a momentum based gradient estimator. We discuss how to instantiate the algorithms for deep learning and demonstrate their properties on image classification and language modeling.



Paperid:2427
Authors:Michael Ito, Danai Koutra, Jenna Wiens
Abstract:
Abstract:Random walk neural networks (RWNNs) have emerged as a promising approach for graph representation learning, leveraging recent advances in sequence models to process random walks. However, under realistic sampling constraints, RWNNs often fail to capture global graph structure even in small graphs due to incomplete node and edge coverage, limiting their expressivity. To address this, we propose random search neural networks (RSNNs), which operate on random searches, each of which guarantees full node coverage. Theoretically, we demonstrate that in sparse graphs, only $O(\log |V|)$ searches are needed to achieve full edge coverage, substantially reducing sampling complexity compared to the $O(|V|)$ walks required by RWNNs (assuming walk lengths scale with graph size). Furthermore, when paired with expressive sequence models, RSNNs are universal approximators and offer stronger generalization. Empirically, RSNNs consistently outperform RWNNs on sparse molecular and protein benchmarks, achieving comparable or superior performance with up to 16$\times$ fewer sampled sequences. Our work bridges theoretical and practical advances in random walk based methods, offering an efficient and expressive framework for learning on sparse graphs.



Paperid:2428
Authors:Yu Zhang, Jialei Zhou, Xinchen Li, Qi Zhang, Zhongwei Wan, Duoqian Miao, Changwei Wang, Longbing Cao
Abstract:
Current text-to-image diffusion generation typically employs complete-text conditioning. Due to the intricate syntax, diffusion transformers (DiTs) inherently suffer from a comprehension defect of complete-text captions. One-fly complete-text input either overlooks critical semantic details or causes semantic confusion by simultaneously modeling diverse semantic primitive types. To mitigate this defect of DiTs, we propose a novel split-text conditioning framework named DiT-ST. This framework converts a complete-text caption into a split-text caption, a collection of simplified sentences, to explicitly express various semantic primitives and their interconnections. The split-text caption is then injected into different denoising stages of DiT-ST in a hierarchical and incremental manner. Specifically, DiT-ST leverages Large Language Models to parse captions, extracting diverse primitives and hierarchically sorting out and constructing these primitives into a split-text input. Moreover, we partition the diffusion denoising process according to its differential sensitivities to diverse semantic primitive types and determine the appropriate timesteps to incrementally inject tokens of diverse semantic primitive types into input tokens via cross-attention. In this way, DiT-ST enhances the representation learning of specific semantic primitive types across different stages. Extensive experiments validate the effectiveness of our proposed DiT-ST in mitigating the complete-text comprehension defect. Dataset and code are available.



Paperid:2429
Authors:Kejia Zhang, Keda TAO, Jiasheng Tang, Huan Wang
Abstract:
Large vision-language models (LVMs) extend large language models (LLMs) with visual perception capabilities, enabling them to process and interpret visual information. A major challenge compromising their reliability is object hallucination that LVMs may generate plausible but factually inaccurate information. We propose a novel \textit{visual adversarial perturbation (VAP)} method to mitigate this hallucination issue. VAP alleviates LVM hallucination by applying strategically optimized visual noise without altering the base model. Our approach formulates hallucination suppression as an optimization problem, leveraging adversarial strategies to generate beneficial visual perturbations that enhance the model's factual grounding and reduce parametric knowledge bias. Extensive experimental results demonstrate that our method consistently reduces object hallucinations across 8 state-of-the-art LVMs, validating its efficacy across diverse evaluations. Code is available at https://anonymous.4open.science/r/VAP-744.



Paperid:2430
Authors:Zhuang Qi, Ying-Peng Tang, Lei Meng, Han Yu, Xiaoxiao Li, Xiangxu Meng
Abstract:
Federated Class Incremental Learning (FCIL) aims to collaboratively process continuously increasing incoming tasks across multiple clients. Among various approaches, data replay has become a promising solution, which can alleviate forgetting by reintroducing representative samples from previous tasks. However, their performance is typically limited by class imbalance, both within the replay buffer due to limited global awareness and between replayed and newly arrived classes. To address this issue, we propose a class-wise balancing data replay method for FCIL (FedCBDR), which employs a global coordination mechanism for class-level memory construction and reweights the learning objective to alleviate the aforementioned imbalances. Specifically, FedCBDR has two key components: 1) the global-perspective data replay module reconstructs global representations of prior task knowledge in a privacy-preserving manner, which then guides a class-aware and importance-sensitive sampling strategy to achieve balanced replay; 2) Subsequently, to handle class imbalance across tasks, the task-aware temperature scaling module adaptively adjusts the temperature of logits at both class and instance levels based on task dynamics, which reduces the model’s overconfidence in majority classes while enhancing its sensitivity to minority classes. Experimental results verified that FedCBDR achieves balanced class-wise sampling under heterogeneous data distributions and improves generalization under task imbalance between earlier and recent tasks, yielding a 2%-15% Top-1 accuracy improvement over six state-of-the-art methods.



Paperid:2431
Authors:Álvaro Parafita, Tomas Garriga, Axel Brando, Francisco Cazorla
Abstract:
Among explainability techniques, SHAP stands out as one of the most popular, but often overlooks the causal structure of the problem. In response, do-SHAP employs interventional queries, but its reliance on estimands hinders its practical application. To address this problem, we propose the use of estimand-agnostic approaches, which allow for the estimation of any identifiable query from a single model, making do-SHAP feasible on arbitrarily complex graphs. We also develop a novel algorithm to significantly accelerate its computation at a negligible cost, as well as a method to explain inaccessible Data Generating Processes. We demonstrate the estimation and computational performance of our approach, and validate it on two real-world datasets, highlighting its potential in obtaining reliable explanations.



Paperid:2432
Authors:Chaoyang Wang, Xiangtai Li, Lu Qi, Xiaofan Lin, Jinbin Bai, Qianyu Zhou, Yunhai Tong
Abstract:
Recent progress in panoramic image generation has underscored two critical limitations in existing approaches. First, most methods are built upon diffusion models, which are inherently ill-suited for equirectangular projection (ERP) panoramas due to the violation of the identically and independently distributed (i.i.d.) Gaussian noise assumption caused by their spherical mapping. Second, these methods often treat text-conditioned generation (text-to-panorama) and image-conditioned generation (panorama outpainting) as separate tasks, relying on distinct architectures and task-specific data. In this work, we propose a unified framework, Panoramic AutoRegressive model (PAR), which leverages masked autoregressive modeling to address these challenges. PAR avoids the i.i.d. assumption constraint and integrates text and image conditioning into a cohesive architecture, enabling seamless generation across tasks. To address the inherent discontinuity in existing generative models, we introduce circular padding to enhance spatial coherence and propose a consistency alignment strategy to improve generation quality. Extensive experiments demonstrate competitive performance in text-to-panorama generation and panorama outpainting tasks while showcasing promising scalability and generalization capabilities. Code and models will be available.



Paperid:2433
Authors:Ning Gao, Xiuhui Zhang, Xingyu Jiang, Mukang You, Mohan Zhang, Yue Deng
Abstract:
Designing efficient reward functions for low-level control tasks is a challenging problem. Recent research aims to reduce reliance on expert experience by using Large Language Models (LLMs) with task information to generate dense reward functions. These methods typically rely on training results as feedback, iteratively generating new reward functions with greedy or evolutionary algorithms. However, they suffer from poor utilization of historical feedback and inefficient search, resulting in limited improvements in complex control tasks. To address this challenge, we propose RF-Agent, a framework that treats LLMs as language agents and frames reward function design as a sequential decision-making process, enhancing optimization through better contextual reasoning. RF-Agent integrates Monte Carlo Tree Search (MCTS) to manage the reward design and optimization process, leveraging the multi-stage contextual reasoning ability of LLM. This approach better utilizes historical information and improves search efficiency to identify promising reward functions. Outstanding experimental results in 17 diverse low-level control tasks demonstrate the effectiveness of our method.



Paperid:2434
Authors:Zekun Wang, Ethan Haarer, Tianyi Zhu, Zhiyi Dai, Christopher MacLellan
Abstract:
Inspired by the human ability to learn and organize knowledge into hierarchical taxonomies with prototypes, this paper addresses key limitations in current deep hierarchical clustering methods. Existing methods often tie the structure to the number of classes and underutilize the rich prototype information available at intermediate hierarchical levels. We introduce deep taxonomic networks, a novel deep latent variable approach designed to bridge these gaps.Our method optimizes a large latent taxonomic hierarchy, specifically a complete binary tree structured mixture-of-Gaussian prior within a variational inference framework, to automatically discover taxonomic structures and associated prototype clusters directly from unlabeled data without assuming true label sizes.We analytically show that optimizing the ELBO of our method encourages the discovery of hierarchical relationships among prototypes. Empirically, our learned models demonstrate strong hierarchical clustering performance, outperforming baselines across diverse image classification datasets using our novel evaluation mechanism that leverages prototype clusters discovered at all hierarchical levels.Qualitative results further reveal that deep taxonomic networks discover rich and interpretable hierarchical taxonomies, capturing both coarse-grained semantic categories and fine-grained visual distinctions.



Paperid:2435
Authors:Binghui Li, Zixun Huang, Fengling Chen, Lean Wang, Lei Wu
Abstract:
Scaling laws have played a cornerstone role in guiding the training of large language models (LLMs). However, existing theoretical analyses primarily focus on the final-step risk, overlooking the intermediate-step risk and, crucially, the impact of learning rate schedule (LRS). In this paper, we aim to bridge this gap by studying a teacher–student kernel regression task trained via online stochastic gradient descent (SGD). Leveraging a novelintrinsic timeviewpoint, we derive aFunctional Scaling Law (FSL)that applies to general learning-rate and batch-size schedules. Remarkably, the impact of the schedules is captured through anexplicit convolution-type functional term, making their effects fully tractable. To illustrate the utility of FSL, we analyze three widely used LRSs -- constant, exponential decay, and warmup-stable-decay (WSD) -- under both data-limited and compute-limited regimes. Our analyses provide theoretical justification for widely adopted empirical practices in LLMs pre-training such as (i) higher-capacity models are more data- and compute-efficient; (ii) learning rate decay can improve training efficiency; (iii) WSD-like schedules can outperform direct-decay schedules.Lastly, we also explore the practical relevance of FSL as a surrogate model studying the loss convergence of LLM pre-training.Overall, FSL offers a principled approach for understanding the role of learning-rate and batch-size scheduling in scaling laws and, more broadly, in stochastic optimization.



Paperid:2436
Authors:Lin Zhu, Tengyu Long, Xiao Wang, Lizhi Wang, Hua Huang
Abstract:
Event cameras provide asynchronous, low-latency, and high-dynamic-range visual signals, making them ideal for real-time perception tasks such as object detection. However, effectively modeling the temporal dynamics of event streams remains a core challenge. Most existing methods follow frame-based detection paradigms, applying temporal modules only at high-level features, which limits early-stage temporal modeling. Transformer-based approaches introduce global attention to capture long-range dependencies, but often add unnecessary complexity and overlook fine-grained temporal cues. In this paper, we propose a CNN-RNN hybrid framework that rethinks temporal modeling for event-based object detection. Our approach is based on two key insights: (1) introducing recurrent modules at lower spatial scales to preserve detailed temporal information where events are most dense, and (2) utilizing Decoupled Deformable-enhanced Recurrent Layers specifically designed according to the inherent motion characteristics of event cameras to extract multiple spatiotemporal features, and performing independent downsampling at multiple spatiotemporal scales to enable flexible, scale-aware representation learning. These multi-scale features are then fused via a feature pyramid network to produce robust detection outputs. Experiments on Gen1 and 1 Mpx dataset demonstrate that our approach achieves superior accuracy over recent transformer-based models, particularly on fast-moving or small-scale objects, highlighting the importance of precise temporal feature extraction in early stages. This work offers a new perspective on designing architectures for event-driven vision beyond attention-centric paradigms.



Paperid:2437
Authors:Wenke Xia, Yichu Yang, Hongtao Wu, Xiao Ma, Tao Kong, Di Hu
Abstract:
Establishing a reliable and iteratively refined robotic system is essential for deploying real-world applications. While Vision-Language-Action (VLA) models are widely recognized as the foundation model for such robotic deployment, their dependence on expert demonstrations hinders the crucial capabilities of correction and learning from failures. To mitigate this limitation, we introduce a Human-assisted Action Preference Optimization method named HAPO, designed to correct deployment failures and foster effective adaptation through preference alignment for VLA models. This method begins with a human-robot collaboration framework for reliable failure correction and interaction trajectory collection through human intervention. These human-intervention trajectories are further employed within the action preference optimization process, facilitating VLA models to mitigate failure action occurrences while enhancing corrective action adaptation. Specifically, we propose an adaptive reweighting algorithm to address the issues of irreversible interactions and token probability mismatch when introducing preference optimization into VLA models, facilitating model learning from binary desirability signals derived from interactions. Through combining these modules, our human-assisted action preference optimization method ensures reliable deployment and effective learning from failure for VLA models. The experiments conducted in simulation and real-world scenarios prove superior generalization and robustness of our framework across a variety of manipulation tasks.



Paperid:2438
Authors:Yuexin Wang, Xiaolei Wang, Yizheng Gong, Jimin XIAO
Abstract:
Generalist Anomaly Detection (GAD) aims to train a unified model on an original domain that can detect anomalies in new target domains. Previous GAD methods primarily use only normal samples as references, overlooking the valuable information contained in anomalous samples that are often available in real-world scenarios. To address this limitation, we propose a more practical approach: normal-abnormal-guided generalist anomaly detection, which leverages both normal and anomalous samples as references to guide anomaly detection across diverse domains. We introduce the Normal-Abnormal Generalist Learning (NAGL) framework, consisting of two key components: Residual Mining (RM) and Anomaly Feature Learning (AFL). RM extracts abnormal patterns from normal-abnormal reference residuals to establish transferable anomaly representations, while AFL adaptively learns anomaly features in query images through residual mapping to identify instance-aware anomalies. Our approach effectively utilizes both normal and anomalous references for more accurate and efficient cross-domain anomaly detection. Extensive experiments across multiple benchmarks demonstrate that our method significantly outperforms existing GAD approaches. This work represents the first to adopt a mixture of normal and abnormal samples as references in generalist anomaly detection. The code and datasets used will be made publicly available.



Paperid:2439
Authors:Xisen Jin, Xiang Ren
Abstract:
Abstract:Large Language models (LLMs) suffer from forgetting of upstream knowledge when fine-tuned. Despite efforts on mitigating forgetting, few have investigated how forgotten upstream examples are dependent on newly learned tasks. Insights on such dependencies enable efficient and targeted mitigation of forgetting. In this paper, we empirically analyze forgetting that occurs in $N$ upstream examples of language modeling or instruction-tuning after fine-tuning LLMs on one of $M$ new tasks, visualized in $M\times N$ matrices. We show that the matrices are often well-approximated with low-rank matrices, indicating the dominance of simple associations between the learned tasks and forgotten upstream examples. Leveraging the analysis, we predict forgetting of upstream examples when fine-tuning LLMs on unseen tasks with matrix completion over the empirical associations. This enables fast identification of most forgotten examples without expensive inference on the entire upstream data. Despite simplicity, the approach outperforms prior approaches that learn semantic relationships of learned tasks and upstream examples with LMs. We demonstrate the practical utility of our analysis by showing statistically significantly reduced forgetting as we upweight predicted examples for replay during fine-tuning.



Paperid:2440
Authors:Xing Wei, Chunchun Chen, Rui Fan, Xiaofeng Cao, Sourav Medya, Wei Ye
Abstract:
Graph neural networks (GNNs) can efficiently process text-attributed graphs (TAGs) due to their message-passing mechanisms, but their training heavily relies on the human-annotated labels. Moreover, the complex and diverse local topologies of nodes of real-world TAGs make it challenging for a single mechanism to handle. Large language models (LLMs) perform well in zero-/few-shot learning on TAGs but suffer from a scalability challenge. Therefore, we propose a preference-driven knowledge distillation (PKD) framework to synergize the complementary strengths of LLMs and various GNNs for few-shot node classification. Specifically, we develop a GNN-preference-driven node selector that effectively promotes prediction distillation from LLMs to teacher GNNs. To further tackle nodes' intricate local topologies, we develop a node-preference-driven GNN selector that identifies the most suitable teacher GNN for each node, thereby facilitating tailored knowledge distillation from teacher GNNs to the student GNN. Extensive experiments validate the efficacy of our proposed framework in few-shot node classification on real-world TAGs.Our code can be available athttps://anonymous.4open.science/r/PKD-B162.



Paperid:2441
Authors:Lucas Lee, Kyungjune Baek
Abstract:
Abstract:Dynamic 4D Gaussian Splatting (4DGS) effectively extends the high-speed rendering capabilities of 3D Gaussian Splatting (3DGS) to represent volumetric videos. However, the large number of Gaussians, substantial temporal redundancies, and especially the absence of an entropy-aware compression framework result in large storage requirements. Consequently, this poses significant challenges for practical deployment, efficient edge-device processing, and data transmission.In this paper, we introduce a novel end-to-end RD-optimized compression framework tailored for 4DGS, aiming to enable flexible, high-fidelity rendering across varied computational platforms.Leveraging Fully Explicit Dynamic Gaussian Splatting (Ex4DGS), one of the state-of-the-art 4DGS methods, as our baseline, we start from the existing 3DGS compression methods for compatibility while effectively addressing additional challenges introduced by the temporal axis. In particular, instead of storing motion trajectories independently per point, we employ a wavelet transform to reflect the real-world smoothness prior, significantly enhancing storage efficiency.This approach yields significantly improved compression ratios and provides a user-controlled balance between compression efficiency and rendering quality. Extensive experiments demonstrate the effectiveness of our method, achieving up to 91$\times$ compression compared to the original Ex4DGS model while maintaining high visual fidelity. These results highlight the applicability of our framework for real-time dynamic scene rendering in diverse scenarios, from resource-constrained edge devices to high-performance environments.



Paperid:2442
Authors:Yanyu Ren, Li Chen, Dan Li, Xizheng Wang, Zhiyuan Wu, Yukai Miao, Yu Bai
Abstract:
Abstract:Large Language Model (LLM) agents are capable of task execution across various domains by autonomously interacting with environments and refining LLM responses based on feedback.However, existing model serving systems are not optimized for the unique demands of serving agents. Compared to classic model serving, agent serving has different characteristics:predictable request pattern, increasing quality requirement, and unique prompt formatting. We identify a key problem for agent serving: LLM serving systems lack session-awareness. They neither perform effective KV cache management nor precisely select the cheapest yet competent model in each round.This leads to a cost-quality tradeoff, and we identify an opportunity to surpass it in an agent serving system.To this end, we introduce AgServe for AGile AGent SERVing.AgServe features a session-aware server that boosts KV cache reuse via Estimated-Time-of-Arrival-based eviction and in-place positional embedding calibration, a quality-aware client that performs session-aware model cascading through real-time quality assessment, and a dynamic resource scheduler that maximizes GPU utilization. With AgServe, we allow agents to select and upgrade models during the session lifetime, and to achieve similar quality at much lower costs, effectively transcending the tradeoff. Extensive experiments on real testbeds demonstrate that AgServe (1) achieves comparable response quality to GPT-4o at a 16.5\% cost. (2) delivers 1.8$\times$ improvement in quality relative to the tradeoff curve.



Paperid:2443
Authors:Zhijie Chen, Qiaobo Li, Arindam Banerjee
Abstract:
Combining gradient compression with adaptive optimizers is a highly desirable goal in distributed learning, with potential benefits in both fewer communication rounds and less per-round communication. In spite of preliminary empirical promise, certain major challenges in the convergence analysis of such methods have stayed open: handling compression based approximation of both first and second moments (pre-conditioner) which appear as a ratio; avoiding dependence on the number of parameters, which is extremely large in modern deep models; and providing high-probability guarantees instead of in-expectation, which can hide high variance behavior.In this work, we introduce a family of Sketched Adaptive Distributed Learning (SADL) algorithms which can use suitable unbiased gradient sketching for compression with suitable adaptive optimization algorithms. As our main contribution, we provide theoretical convergence guarantees of SADL algorithms which addresses all of the existing challenges. In particular, our guarantees hold with high probability, picks up only a logarithmic dependence on the number of parameters, and the first and second moment approximation is handled precisely yielding a dependence on the intrinsic dimension of the loss Hessian, which is significantly smaller than the full dimensionality of deep learning models. Empirically, the SADL algorithms are shown to be competitive with and often outperform baselines on both vision and language tasks, in both supervised fine-tuning and training-from-scratch regimes. Further, the SADL algorithms are also competitive with the state-of-the-art communication-efficient distributed learning algorithms based on error feedback.



Paperid:2444
Authors:Chen-Hao (Lance) Chao, Wei-Fang Sun, Hanwen Liang, Chun-Yi Lee, Rahul Krishnan
Abstract:
Masked diffusion models (MDM) are powerful generative models for discrete data that generate samples by progressively unmasking tokens in a sequence. Each token can take one of two states: masked or unmasked. We observe that token sequences often remain unchanged between consecutive sampling steps; consequently, the model repeatedly processes identical inputs, leading to redundant computation. To address this inefficiency, we propose the Partial masking scheme (Prime), which augments MDM by allowing tokens to take intermediate states interpolated between the masked and unmasked states. This design enables the model to make predictions based on partially observed token information, and facilitates a fine-grained denoising process. We derive a variational training objective and introduce a simple architectural design to accommodate intermediate-state inputs. Our method demonstrates superior performance across a diverse set of generative modeling tasks. On text data, it achieves a perplexity of 15.36 on OpenWebText, outperforming previous MDM (21.52), autoregressive models (17.54), and their hybrid variants (17.58), without relying on an autoregressive formulation. On image data, it attains competitive FID scores of 3.26 on CIFAR-10 and 6.98 on ImageNet-32, comparable to leading continuous generative models.



Paperid:2445
Authors:Chen Yueh-Han, Guy Davidson, Brenden Lake
Abstract:
Frontier LLMs show impressive capabilities, but do they robustly generalize critical safety knowledge to novel situations? Lacking this ability is dangerous when users ask innocuous questions---for instance, ''I'm considering packing melon balls for my 10-month-old's lunch. What other foods would be good to include?'' Before suggesting good food options, the model should also warn that melon balls pose a choking hazard to toddlers, as documented by the Centers for Disease Control and Prevention (CDC). Failing to provide such warnings could result in serious risks or even death. To evaluate this, we introduce SAGE-Eval, SAfety-fact systematic GEneralization evaluation, the first benchmark that tests whether LLMs appropriately apply well-established safety facts to naive user queries. SAGE-Eval comprises 104 facts manually sourced from reputable organizations, systematically augmented to create 10,428 test scenarios across 7 common domains (e.g., Outdoor Activities, Chemical, Medicine). We find that the top model, Claude-3.7-sonnet, passes only 58% of all the safety facts tested. We also observe that model capabilities and training compute only weakly correlate with performance on SAGE-Eval, implying that scaling up is not the golden solution. Our findings suggest frontier LLMs still lack robust generalization ability, and we recommend that developers use SAGE-Eval in pre-deployment evaluations to assess model reliability in addressing salient risks.



Paperid:2446
Authors:Lei Yang, Xinyu Zhang, Chen Wang, Jun Li, Jiaqi Ma, Zhiying Song, Tong Zhao, Ziying Song, Li Wang, Mo Zhou, Yang Shen, Kai WU, Chen Lv
Abstract:
Modern autonomous vehicle perception systems often struggle with occlusions and limited perception range. Previous studies have demonstrated the effectiveness of cooperative perception in extending the perception range and overcoming occlusions, thereby enhancing the safety of autonomous driving. In recent years, a series of cooperative perception datasets have emerged; however, these datasets primarily focus on cameras and LiDAR, neglecting 4D Radar—a sensor used in single-vehicle autonomous driving to provide robust perception in adverse weather conditions. In this paper, to bridge the gap created by the absence of 4D Radar datasets in cooperative perception, we present V2X-Radar, the first large-scale, real-world multi-modal dataset featuring 4D Radar. V2X-Radar dataset is collected using a connected vehicle platform and an intelligent roadside unit equipped with 4D Radar, LiDAR, and multi-view cameras. The collected data encompasses sunny and rainy weather conditions, spanning daytime, dusk, and nighttime, as well as various typical challenging scenarios. The dataset consists of 20K LiDAR frames, 40K camera images, and 20K 4D Radar data, including 350K annotated boxes across five categories. To support various research domains, we have established V2X-Radar-C for cooperative perception, V2X-Radar-I for roadside perception, and V2X-Radar-V for single-vehicle perception. Furthermore, we provide comprehensive benchmarks across these three sub-datasets.



Paperid:2447
Authors:Soutrik Sarangi, Yonatan Sverdlov, Nadav Dym, Abir De
Abstract:
Abstract:Motivated by applications for set containment problems, we consider the following fundamental problem: can we design set-to-vector functions so that the natural partial order on sets is preserved, namely $S\subseteq T \text{ if and only if } F(S)\leq F(T) $. We call functions satisfying this property \emph{Monotone and Separating (MAS)} set functions. We establish lower and upper bounds for the vector dimension necessary to obtain MAS functions, as a function of the cardinality of the multisets and the underlying ground set. In the important case of an infinite ground set, we show that MAS functions do not exist, but provide a model called \our\ which provably enjoys a relaxed MAS property we name 'weakly MAS' and is stable in the sense of Holder continuity. We also show that MAS functions can be used to construct universal models which are monotone by construction, and can approximate all monotone set functions. Experimentally, we consider a variety of set containtment tasks. The experiments show the benefit of using our \our\ model, in comparsion with standard set models which do not incorporate set containment as an inductive bias.



Paperid:2448
Authors:Binqian Xu, Haiyang Mei, Zechen Bai, Jinjin Gong, Rui Yan, Guosen Xie, Yazhou Yao, Basura Fernando, Xiangbo Shu
Abstract:
Abstract:Multimodal Large Language Models (MLLMs) with Federated Learning (FL) can quickly adapt to privacy-sensitive tasks, but face significant challenges such as high communication costs and increased attack risks, due to their reliance on multi-round communication. To address this, One-shot FL (OFL) has emerged, aiming to complete adaptation in a single client-server communication. However, existing adaptive ensemble OFL methods still need more than one round of communication, because correcting heterogeneity-induced local bias relies on aggregated global supervision, meaning they never attain true one-shot communication. In this work, we make the first attempt to achieve true one-shot communication for MLLMs under OFL, by investigating whether implicit (i.e., initial rather than aggregated) global supervision alone can effectively correct local training bias. Our key finding from the empirical study is that imposing directional supervision on local training substantially mitigates client conflicts and local bias. Building on this insight, we propose YOCO, in which directional supervision with sign-regularized LoRA B enforces global consistency, while sparsely regularized LoRA A preserves client-specific adaptability. Experiments demonstrate that YOCO cuts communication to $\sim$0.03\% of multi-round FL while surpassing those methods in several multimodal scenarios and consistently outperforming all one-shot competitors.



Paperid:2449
Authors:Yudong Li, Jingyuan Wang, Xie Yu, Peiyu Wang, Qian Huang
Abstract:
Traffic flow data are of great value in smart city applications. However, limited by data collection costs and privacy sensitivity, it is rather difficult to obtain large-scale traffic flow data. Therefore, various data generation methods have been proposed in the literature. Nevertheless, these methods often require data from a specific city for training and are difficult to directly apply to new cities lacking data.To address this problem, this paper proposes a retrieval-augmented diffusion generation model with representation alignment. We use data from multiple source cities for training, extract consistent representations across multiple cities, and leverage retrieval-augmented generation (RAG) technology to incorporate historical data from source cities under similar conditions into the condition, aiming to improve the accuracy of data generation in the target city.Experiments on four real-world datasets demonstrate that, compared with existing deep learning methods, our method achieves better cross-city transfer performance.



Paperid:2450
Authors:Nicholas Boffi, Michael Albergo, Eric Vanden-Eijnden
Abstract:
We present a systematic approach for learning the flow map associated with a flow or diffusion model. The class of flow map-based models, commonly referred to as consistency models, encompasses a wide range of recent efforts to improve the efficiency of generative models based on the solution of a differential equation. By exploiting a relation between the velocity field underlying a continuous-time flow and the instantaneous rate of change of the flow map, we show how to convert a broad class of existing distillation schemes into single network direct training algorithms via the concept of self-distillation. We empirically study the performance of several instantiations of our framework, where we find that high-dimensional tasks such as image synthesis benefit from the design of objective functions that avoid temporal and spatial derivatives of the flow map, while for low-dimensional tasks requiring precision, objectives with higher-order derivatives can help capture sharp features.



Paperid:2451
Authors:Ori Press, Brandon Amos, Haoyu Zhao, Yikai Wu, Samuel Ainsworth, Dominik Krupke, Patrick Kidger, Touqir Sajed, Bartolomeo Stellato, Jisun Park, Nathanael Bosch, Eli Meril, Albert Steppi, Arman Zharmagambetov, Fangzhao Zhang, David Pérez-Piñeiro, Alberto Mercurio, Ni Zhan, Talor Abramovich, Kilian Lieret, Hanlin Zhang, Shirley Huang, Matthias Bethge, Ofir Press
Abstract:
Despite progress in language model (LM) capabilities, evaluations have thus far focused on models' performance on tasks that humans have previously solved, including in programming (SWE-Bench) and mathematics (FrontierMath). We therefore propose testing models' ability to design and implement algorithms in an open-ended benchmark: We task LMs with writing code that efficiently solves computationally challenging problems in computer science, physics, and mathematics. Our AlgoTune benchmark consists of 120 tasks collected from domain experts and a framework for validating and timing LM-synthesized solution code, which is compared to reference implementations from popular open-source packages.In addition, we develop a baseline LM agent, AlgoTuner, and evaluate its performance across a suite of frontier models.AlgoTuner achieves an average 1.58x speedup against reference solvers, including methods from packages such as SciPy, scikit-learn and CVXPY.However, we find that current models fail to discover algorithmic innovations, instead preferring surface-level optimizations. We hope that AlgoTune catalyzes the development of LM agents exhibiting creative problem solving beyond state-of-the-art human performance.



Paperid:2452
Authors:John Leland, YooJung Choi
Abstract:
A fundamental challenge in probabilistic modeling is balancing expressivity and tractable inference. Probabilistic circuits (PCs) aim to directly address this tradeoff by imposing structural constraints that guarantee efficient inference of certain queries while maintaining expressivity. Because the complexity of inference on PCs is a function of the circuit size, understanding the size requirements of different families of PCs---characterized by structural properties---is fundamental in mapping the trade-off between tractability and expressive efficiency. However, the study of expressive efficiency of circuits are often concerned with exact representations, which may not align with model learning, where we look to approximate the underlying data distribution closely by some distance measure. Moreover, due to hardness of inference tasks, exactly representing distributions while supporting tractable inference often incurs exponential size blow-ups. In this paper, we consider the natural following question: can we avoid such size blow-up by allowing for some small approximation error? We study approximating distributions with probabilistic circuits with guarantees based on f-divergences, and analyze which inference queries remain well-approximated under this framework. We show that approximating an arbitrary distribution with bounded f-divergence is NP-hard for any model that can tractably compute marginals. In addition, we prove an exponential size gap for approximation between the class of decomposable PCs and that of decomposable and deterministic PCs.



Paperid:2453
Authors:Zixiang Li, Haoyu Wang, Wei Wang, Chuangchuang Tan, Yunchao Wei, Yao Zhao
Abstract:
Diffusion models have achieved remarkable success in image generation and editing tasks. Inversion within these models aims to recover the latent noise representation for a real or generated image, enabling reconstruction, editing, and other downstream tasks. However, to date, most inversion approaches suffer from an intrinsic trade-off between reconstruction accuracy and editing flexibility. This limitation arises from the difficulty of maintaining both semantic alignment and structural consistency during the inversion process. In this work, we introduceDual-Conditional Inversion (DCI), a novel framework that jointly conditions on the source prompt and reference image to guide the inversion process. Specifically, DCI formulates the inversion process as a dual-condition fixed-point optimization problem, minimizing both the latent noise gap and the reconstruction error under the joint guidance. This design anchors the inversion trajectory in both semantic and visual space, leading to more accurate and editable latent representations. Our novel setup brings new understanding to the inversion process. Extensive experiments demonstrate that DCI achieves state-of-the-art performance across multiple editing tasks, significantly improving both reconstruction quality and editing precision. Furthermore, we also demonstrate that our method achieves strong results in reconstruction tasks, implying a degree of robustness and generalizability approaching the ultimate goal of the inversion process.



Paperid:2454
Authors:Xiaomeng Hu, Fei Huang, Chenhan Yuan, Junyang Lin, Tsung-Yi Ho
Abstract:
As large language models (LLMs) are increasingly deployed in real-world applications, ensuring the safety of their outputs during decoding has become a critical challenge. Existing training-time alignment methods often fail to address dynamic safety risks that emerge during the inference time. While decoding-time interventions, such as Contrastive Decoding, offer adaptability, they face trade-offs between safety and response quality. In this work, we propose a novel framework for decoding-time safety alignment that integrates three key components: (1) a guard model for real-time safety monitoring, enabling detection of potentially unsafe content; (2) a rollback mechanism with a token buffer to correct unsafe outputs efficiently at an earlier stage while without disrupting the user experience; and (3) an introspection-based intervention strategy, where the model generates self-reflective critiques of its previous outputs and incorporates these reflections into the context to guide subsequent decoding steps. This framework benefits from the targeted nature of interventions (thanks to the guard model), the timely nature of corrections (enabled by the rollback mechanism), and the emergence of effective introspection patterns when larger buffers are used. Experimental results demonstrate that our framework achieves superior performance with minimal average wait tokens and a high successful intervention rate, while maintaining high response quality.



Paperid:2455
Authors:Tianle Li, Jihai Zhang, Yongming Rao, Yu Cheng
Abstract:
While large language models (LLMs) demonstrate strong reasoning capabilities utilizing reinforcement learning (RL) with verifiable reward, whether large vision-language models (VLMs) can directly inherit such capabilities through similar post-training strategies remains underexplored. In this work, we conduct a systematic compositional probing study to evaluate whether current VLMs trained with RL or other post-training strategie can compose capabilities across modalities or tasks under out-of-distribution conditions. We design a suite of diagnostic tasks that train models on unimodal tasks or isolated reasoning skills, and evaluate them on multimodal, compositional variants requiring skill integration. Through comparisons between supervised fine-tuning (SFT) and RL-trained models, we identify three key findings: (1) RL-trained models consistently outperform SFT on compositional generalization, demonstrating better integration of learned skills; (2) although VLMs achieve strong performance on individual tasks, they struggle to generalize compositionally under cross-modal and cross-task scenario, revealing a significant gap in current training strategies; (3) enforcing models to explicitly describe visual content before reasoning (e.g., caption-before-thinking), along with rewarding progressive vision-to-text grounding, yields notable gains. It highlights two essential ingredients for improving compositionality in VLMs: visual-to-text alignment and accurate visual grounding. Our findings shed light on the current limitations of RL-based reasoning VLM training and provide actionable insights toward building models that reason compositionally across modalities and tasks.



Paperid:2456
Authors:Luxi Chen, Zihan Zhou, Min Zhao, Yikai Wang, Ge Zhang, Wenhao Huang, Hao Sun, Ji-Rong Wen, Chongxuan LI
Abstract:
Generating flexible-view 3D scenes, including 360° rotation and zooming, from single images is challenging due to a lack of 3D data. To this end, we introduce FlexWorld, a novel framework that progressively constructs a persistent 3D Gaussian splatting representation by synthesizing and integrating new 3D content. To handle novel view synthesis under large camera variations, we leverage an advanced pre-trained video model fine-tuned on accurate depth-estimated training pairs. By combining geometry-aware scene integration and optimization, FlexWorld refines the scene representation, producing visually consistent 3D scenes with flexible viewpoints. Extensive experiments demonstrate the effectiveness of FlexWorld in generating high-quality novel view videos and flexible-view 3D scenes from single images, achieving superior visual quality under multiple popular metrics and datasets compared to existing state-of-the-art methods. Additionally, FlexWorld supports extrapolating from existing 3D scenes, further extending its applicability. Qualitatively, we highlight that FlexWorld can generate high-fidelity scenes that enable 360° rotations and zooming exploration.



Paperid:2457
Authors:Bar Mahpud, Or Sheffet
Abstract:
Abstract:We study the problem of differentially private second moment estimation and present a new algorithm that achieve strong privacy-utility trade-offs even for worst-case inputs under subsamplability assumptions on the data. We call an input $(m,\alpha,\beta)$-subsamplable if a random subsample of size $m$ (or larger) preserves w.p $\geq 1-\beta$ the spectral structure of the original second moment matrix up to a multiplicative factor of $1\pm \alpha$.Building upon subsamplability, we give a recursive algorithmic framework similar to Kamath et al (2019) that abides zero-Concentrated Differential Privacy (zCDP) while preserving w.h.p the accuracy of the second moment estimation upto an arbitrary factor of $(1\pm\gamma)$. We then show how to apply our algorithm to approximate the second moment matrix of a distribution $\mathcal{D}$, even when a noticeable fraction of the input are outliers.



Paperid:2458
Authors:Teodora Reu, Sixtine Dromigny, Michael Bronstein, Francisco Vargas
Abstract:
We investigate the underlying cause of common failure modes behind the state-of-the-art fixed-point generative models, Rectified Flows and Schr\"odinger Bridge Matching. First, we introduce a gradient‐variance diagnostic that directly measures how learned vector fields induce transport errors after integration. Contrary to common belief, we show that low‐variance (i.e., well‐aligned) interpolants can still yield high transport error. We also demonstrate that standard neural network architectures (e.g., MLPs or CNNs) are fundamentally limited, failing to exactly represent even Gaussian-to-Gaussian transport. While prior work on Rectified Flows notes that strict convergence to optimal couplings is not guaranteed, it overlooks a key limitation: with noiseless interpolants, further iterations do not improve transport. If the pairings are already straight, the process stagnates; if they are deterministic but not straight, the vector field simply memorizes these pairings without achieving true optimal transport. We prove the existence of such memorizing vector fields.



Paperid:2459
Authors:Jiabao Lei, Kewei Shi, Zhihao Liang, Kui Jia
Abstract:
Directly generating 3D meshes, the default representation for 3D shapes in the graphics industry, using auto-regressive (AR) models has become popular these days, thanks to their sharpness, compactness in the generated results, and ability to represent various types of surfaces. However, AR mesh generative models typically construct meshes face by face in lexicographic order, which does not effectively capture the underlying geometry in a manner consistent with human perception. Inspired by 2D models that progressively refine images, such as the prevailing next-scale prediction AR models, we propose generating meshes auto-regressively in a progressive coarse-to-fine manner. Specifically, we view mesh simplification algorithms, which gradually merge mesh faces to build simpler meshes, as a natural fine-to-coarse process. Therefore, we develop a transformer-based AR model to approximate the reverse process of a generalized mesh simplification algorithm in the order of level-of-detail, constructing meshes initially from a single point and gradually adding geometric details through local remeshing, where the topology is not predefined and is alterable. Our ablation studies and experiments show that this novel progressive mesh generation approach not only leads to improved mesh quality but also enables applications such as mesh refinement and editing.



Paperid:2460
Authors:Muyao Wang, Zeke Xie, Bo Chen, Hongwei Liu, James Kwok
Abstract:
The influence function serves as an efficient post-hoc interpretability tool that quantifies the impact of training data modifications on model parameters, enabling enhanced model performance, improved generalization, and interpretability insights without the need for expensive retraining processes. Recently, Multivariate Time Series (MTS) analysis has become an important yet challenging task, attracting significant attention. While channel extremely matters to MTS tasks, channel-centric methods are still largely under-explored for MTS. Particularly, no previous work studied the effects of channel information of MTS in order to explore counterfactual effects between these channels and model performance. To fill this gap, we propose a novel Channel-wise Influence (ChInf) method that is the first to estimate the influence of different channels in MTS. Based on ChInf, we naturally derived two channel-wise algorithms by incorporating ChInf into classic MTS tasks. Extensive experiments demonstrate the effectiveness of ChInf and ChInf-based methods in critical MTS analysis tasks, such as MTS anomaly detection and MTS data pruning. Specifically, our ChInf-based methods rank top-1 among all methods for comparison, while previous influence functions do not perform well on MTS anomaly detection tasks and MTS data pruning problem. This fully supports the superiority and necessity of ChInf.



Paperid:2461
Authors:Zhen Xu, Zhengqin Li, Zhao Dong, Xiaowei Zhou, Richard Newcombe, Zhaoyang Lv
Abstract:
We propose 4DGT, a 4D Gaussian-based Transformer model for dynamic scene reconstruction, trained entirely on real-world monocular posed videos. Using 4D Gaussian as an inductive bias, 4DGT unifies static and dynamic components, enabling the modeling of complex, time-varying environments with varying object lifespans. We proposed a novel density control strategy in training, which enables our 4DGT to handle longer space-time input. Our model processes 64 consecutive posed frames in a rolling-window fashion, predicting consistent 4D Gaussians in the scene. Unlike optimization-based methods, 4DGT performs purely feed-forward inference, reducing reconstruction time from hours to seconds and scaling effectively to long video sequences. Trained only on large-scale monocular posed video datasets, 4DGT can outperform prior Gaussian-based networks significantly in real-world videos and achieve on-par accuracy with optimization-based methods on cross-domain videos.



Paperid:2462
Authors:Xiaolin Sun, Feidi Liu, Zhengming Ding, Zizhan Zheng
Abstract:
Abstract:Reinforcement learning (RL) systems, while achieving remarkable success across various domains, are vulnerable to adversarial attacks. This is especially a concern in vision-based environments where minor manipulations of high-dimensional image inputs can easily mislead the agent's behavior. To this end, various defenses have been proposed recently, with state-of-the-art approaches achieving robust performance even under large state perturbations. However, after closer investigation, we found that the effectiveness of the current defenses is due to a fundamental weakness of the existing $l_p$ norm-constrained attacks, which can barely alter the semantics of image input even under a relatively large perturbation budget. In this work, we propose SHIFT, a novel policy-agnostic diffusion-based state perturbation attack to go beyond this limitation. Our attack is able to generate perturbed states that are semantically different from the true states while remaining realistic and history-aligned to avoid detection. Evaluations show that our attack effectively breaks existing defenses, including the most sophisticated ones, significantly outperforming existing attacks while being more perceptually stealthy. The results highlight the vulnerability of RL agents to semantics-aware adversarial perturbations, indicating the importance of developing more robust policies.



Paperid:2463
Authors:Tianhao Peng, Haochen Wang, Yuanxing Zhang, Noah Wang, Zili Wang, Ge Zhang, Jian Yang, Shihao Li, Yanghai Wang, Xintao Wang, Houyi Li, Wei Ji, Pengfei Wan, Wenhao Huang, ZHAO-XIANG ZHANG, Jiaheng Liu
Abstract:
The advent of Multimodal Large Language Models (MLLMs) has expanded AI capabilities to visual modalities, yet existing evaluation benchmarks remain limited to single-video understanding, overlooking the critical need for multi-video understanding in real-world scenarios (e.g., sports analytics and autonomous driving). To address this significant gap, we introduceMVU-Eval, the first comprehensive benchmark for evaluatingMulti-VideoUnderstanding for MLLMs. Specifically, our MVU-Eval mainly assesses eight core competencies through 1,824 meticulously curated question-answer pairs spanning 4,959 videos from diverse domains, addressing both fundamental perception tasks and high-order reasoning tasks. These capabilities are rigorously aligned with real-world applications such as multi-sensor synthesis in autonomous systems and cross-angle sports analytics. Through extensive evaluation of state-of-the-art open-source and closed-source models, we reveal significant performance discrepancies and limitations in current MLLMs' ability to perform understanding across multiple videos.The benchmark will be made publicly available to foster future research.



Paperid:2464
Authors:Asterios Tsiourvas, Wei Sun, Georgia Perakis
Abstract:
LLM routing aims to select the most appropriate model for each query, balancing competing performance metrics such as accuracy and cost across a pool of language models. Prior approaches typically adopt a decoupled strategy, where the metrics are first predicted and the model is then selected based on these estimates. This setup is prone to compounding errors and often relies on full-feedback data, where each query is evaluated by all candidate models, which is costly to obtain and maintain in practice. In contrast, we learn from observational data, which records only the outcome of the model actually deployed. We propose a causal end-to-end framework that learns routing policies by minimizing decision-making regret from observational data. To enable efficient optimization, we introduce two theoretically grounded surrogate objectives: a classification-based upper bound, and a softmax-weighted regret approximation shown to recover the optimal policy at convergence. We further extend our framework to handle heterogeneous cost preferences via an interval-conditioned architecture. Experiments on public benchmarks show that our method outperforms existing baselines, achieving state-of-the-art performance across different embedding models.



Paperid:2465
Authors:Zhipeng Zhou, Ziqiao Meng, Pengcheng Wu, Peilin Zhao, Chunyan Miao
Abstract:
Multi-task learning (MTL) is a widely explored paradigm that enables the simultaneous learning of multiple tasks using a single model. Despite numerous solutions, the key issues of optimization conflict and task imbalance remain under-addressed, limiting performance. Unlike existing optimization-based approaches that typically reweight task losses or gradients to mitigate conflicts or promote progress, we propose a novel approach based on Continual Optimization with Symmetry Teleportation (COST). During MTL optimization, when an optimization conflict arises, we seek an alternative loss-equivalent point on the loss landscape to reduce conflict. Specifically, we utilize a low-rank adapter (LoRA) to facilitate this practical teleportation by designing convergent, loss-invariant objectives. Additionally, we introduce a historical trajectory reuse strategy to continually leverage the benefits of advanced optimizers. Extensive experiments on multiple mainstream datasets demonstrate the effectiveness of our approach. COST is a plug-and-play solution that enhances a wide range of existing MTL methods. When integrated with state-of-the-art methods, COST achieves superior performance.



Paperid:2466
Authors:Haoyu Zhang, Meng Liu, Zaijing Li, Haokun Wen, Weili Guan, Yaowei Wang, Liqiang Nie
Abstract:
Visual-spatial understanding, the ability to infer object relationships and layouts from visual input, is fundamental to downstream tasks such as robotic navigation and embodied interaction. However, existing methods face spatial uncertainty and data scarcity, limiting the 3D spatial reasoning capability of pre-trained vision-language models (VLMs). To address these challenges, we present a unified framework for enhancing 3D spatial reasoning in pre-trained VLMs without modifying their architecture. This framework combines SpatialMind, a structured prompting strategy that decomposes complex scenes and questions into interpretable reasoning steps, with ScanForgeQA, a scalable question-answering dataset built from diverse 3D simulation scenes through an automated construction process designed for fine-tuning. Extensive experiments across multiple benchmarks demonstrate the individual and combined effectiveness of our prompting and fine-tuning strategies, and yield insights that may inspire future research on visual-spatial understanding.



Paperid:2467
Authors:Thai-Hoang Pham, Jiayuan Chen, Seungyeon Lee, Yuanlong Wang, Sayoko Moroi, Xueru Zhang, Ping Zhang
Abstract:
As machine learning (ML) algorithms are increasingly used in medical image analysis, concerns have emerged about their potential biases against certain social groups. Although many approaches have been proposed to ensure the fairness of ML models, most existing works focus only on medical image diagnosis tasks, such as image classification and segmentation, and overlooked prognosis scenarios, which involve predicting the likely outcome or progression of a medical condition over time. To address this gap, we introduce FairTTE, the first comprehensive framework for assessing fairness in time-to-event (TTE) prediction in medical imaging. FairTTE encompasses a diverse range of imaging modalities and TTE outcomes, integrating cutting-edge TTE prediction and fairness algorithms to enable systematic and fine-grained analysis of fairness in medical image prognosis. Leveraging causal analysis techniques, FairTTE uncovers and quantifies distinct sources of bias embedded within medical imaging datasets. Our large-scale evaluation reveals that bias is pervasive across different imaging modalities and that current fairness methods offer limited mitigation. We further demonstrate a strong association between underlying bias sources and model disparities, emphasizing the need for holistic approaches that target all forms of bias. Notably, we find that fairness becomes increasingly difficult to maintain under distribution shifts, underscoring the limitations of existing solutions and the pressing need for more robust, equitable prognostic models.



Paperid:2468
Authors:Derui Zhu, Dingfan Chen, jinfu chen, Jens Grossklags, Alexander Pretschner, Weiyi Shang
Abstract:
Large language models (LLMs) have demonstrated remarkable progress in generating functional code, leading to numerous AI-based coding program tools. However, their reliance on the perplexity objective during both training and inference primarily emphasizes functionality, often at the expense of efficiency—an essential consideration for real-world coding tasks. Perhaps interestingly, we observed that well-trained LLMs inherently possess knowledge about code efficiency, but this potential remains underutilized with standard decoding approaches. To address this, we design strategic prompts to activate the model’s embedded efficiency understanding, effectively using LLMs as \textit{efficiency critiques} to guide code generation toward higher efficiency without sacrificing—and sometimes even improving—functionality, all without the need for costly real code execution. Extensive experiments on benchmark datasets (EffiBench, HumanEval+) across multiple representative code models demonstrate up to a 70.6\% reduction in average execution time and a 13.6\% decrease in maximum memory usage, highlighting the computational efficiency and practicality of our approach compared to existing alternatives.



Paperid:2469
Authors:Zhining Zhang, Chuanyang Jin, Mung Yao Jia, Shunchi Zhang, Tianmin Shu
Abstract:
Theory of Mind (ToM), the ability to understand people's minds based on their behavior, is key to developing socially intelligent agents. Current approaches to ToM reasoning either rely on prompting Large Language Models (LLMs), which are prone to systematic errors, or use handcrafted, rigid agent models for model-based inference, which are more robust but fail to generalize across domains. In this work, we introduceAutoToM, an automated agent modeling method for scalable, robust, and interpretable mental inference. Given a ToM problem,AutoToMfirst proposes an initial agent model and then performs automated Bayesian inverse planning based on this model, leveraging an LLM backend. Guided by inference uncertainty, it iteratively refines the model by introducing additional mental variables and/or incorporating more timesteps in the context. Across five diverse benchmarks,AutoToMoutperforms existing ToM methods and even large reasoning models. Additionally, we show thatAutoToMcan produce human‐like confidence estimates and enable online mental inference for embodied decision-making.



Paperid:2470
Authors:Naga Sai Abhiram Kusumba, Maitreya Patel, Kyle Min, Changhoon Kim, Chitta Baral, 'YZ' Yezhou Yang
Abstract:
Erasing harmful or proprietary concepts from powerful text‑to‑image generators is an emerging safety requirement, yet current ``concept erasure'' techniques either collapse image quality, rely on brittle adversarial losses, or demand prohibitive retraining cycles. We trace these limitations to a myopic view of the denoising trajectories that govern diffusion‑based generation. We introduce EraseFlow, the first framework that casts concept unlearning as exploration in the space of denoising paths and optimizes it with a GFlowNets equipped with the trajectory‑balance objective. By sampling entire trajectories rather than single end states, EraseFlow learns a stochastic policy that steers generation away from target concepts while preserving the model’s prior. EraseFlow eliminates the need for carefully crafted reward models and by doing this, it generalizes effectively to unseen concepts and avoids hackable rewards while improving the performance. Extensive empirical results demonstrate that EraseFlow outperforms existing baselines and achieves an optimal trade-off between performance and prior preservation.



Paperid:2471
Authors:Yingyu Lin, Erchi Wang, Yian Ma, Yu-Xiang Wang
Abstract:
Abstract:We propose a framework to convert $(\varepsilon, \delta)$-approximate Differential Privacy (DP) mechanisms into $(\varepsilon', 0)$-pure DP mechanisms under certain conditions, a process we call ``purification.'' This algorithmic technique leverages randomized post-processing with calibrated noise to eliminate the $\delta$ parameter while achieving near-optimal privacy-utility tradeoff for pure DP. It enables a new design strategy for pure DP algorithms: first run an approximate DP algorithm with certain conditions, and then purify. This approach allows one to leverage techniques such as strong composition and propose-test-release that require $\delta>0$ in designing pure-DP methods with $\delta=0$. We apply this framework in various settings, including Differentially Private Empirical Risk Minimization (DP-ERM), stability-based release, and query release tasks. To the best of our knowledge, this is the first work with a statistically and computationally efficient reduction from approximate DP to pure DP. Finally, we illustrate the use of this reduction for proving lower bounds under approximate DP constraints with explicit dependence in $\delta$, avoiding the sophisticated fingerprinting code construction.



Paperid:2472
Authors:Ruihang Chu, Yefei He, Zhekai Chen, Shiwei Zhang, Xiaogang Xu, bin xia, Dingdong WANG, Hongwei Yi, Xihui Liu, Hengshuang Zhao, Yu Liu, Yingya Zhang, Yujiu Yang
Abstract:
We present Mover, a simple and scalable framework that brings motion control to video generative models. Existing motion-controllable methods typically suffer from coarse control granularity and limited scalability, leaving their outputs insufficient for practical use. We narrow this gap by achieving precise and high-quality motion control. Our core idea is to directly make the original condition features motion-aware for guiding video synthesis. To this end, we first represent object motions with dense point trajectories, allowing fine-grained control over the scene. We then project these trajectories into latent space and propagate the first frame's features along each trajectory, producing an aligned spatiotemporal feature map that tells how each scene element should move. This feature map serves as the updated latent condition, which is naturally integrated into the off-the-shelf image-to-video model, e.g., Wan-I2V-14B, as motion guidance without any architecture change. It removes the need for auxiliary motion encoders and makes fine-tuning base models easily scalable. Through scaled training, Mover generates 5-second, 480p videos whose motion controllability rivals Kling 1.5 Pro's commercial Motion Brush, as indicated by user studies. To support comprehensive evaluation, we further design MoverBench, a rigorously curated benchmark featuring diverse content categories and hybrid-verified annotations. It is distinguished by larger data volume, longer video durations, and high-quality motion annotations. Extensive experiments on MoverBench and the public dataset consistently show Mover's superior motion quality. Code, models, and benchmark data will be made publicly available.



Paperid:2473
Authors:Han Wu, Jie Yin
Abstract:
Few-shot knowledge graph (KG) relational learning focuses on performing reasoning over relations with only a limited number of training examples. Most current approaches follow a meta-learning framework but suffer from two key pitfalls. First, they typically learn relation meta-knowledge in isolation, failing to capture common relational patterns shared among related tasks. Second, they overlook local, task-specific information crucial for rapid adaptation to each new task. To address these limitations, we propose MoEMeta, a novel meta-learning framework with two key innovations: (1) a mixture-of-experts model proposed to learn relational prototypes shared across tasks, enabling better generalization, and (2) a task-tailored adaptation mechanism designed to facilitate local adaptation to account for task-specific contexts. By balancing global generalization with local adaptability, MoEMeta significantly improves few-shot relational learning. Extensive experiments and ablation studies on three few-shot KG benchmarks demonstrate MoEMeta outperforms existing baselines, achieving state-of-the-art performance.



Paperid:2474
Authors:Christina Kassab, Sacha Morin, Martin Büchner, Matias Mattamala, Kumaraditya Gupta, Abhinav Valada, Liam Paull, Maurice Fallon
Abstract:
3D scene understanding has been transformed by open-vocabulary language models that enable interaction via natural language. However, at present the evaluation of these representations is limited to datasets with closed-set semantics that do not capture the richness of language. This work presents OpenLex3D, a dedicated benchmark for evaluating 3D open-vocabulary scene representations. OpenLex3D provides entirely new label annotations for scenes from Replica, ScanNet++, and HM3D, which capture real-world linguistic variability by introducing synonymical object categories and additional nuanced descriptions. Our label sets provide 13 times more labels per scene than the original datasets. By introducing an open-set 3D semantic segmentation task and an object retrieval task, we evaluate various existing 3D open-vocabulary methods on OpenLex3D, showcasing failure cases, and avenues for improvement. Our experiments provide insights on feature precision, segmentation, and downstream capabilities. The benchmark is publicly available at: https://openlex3d.github.io/.



Paperid:2475
Authors:Farhad Pashakhanloo
Abstract:
Despite being observed in biological and artificial neural networks, mechanisms of representational drift are still under investigation. Under one set of hypotheses, drift is a result of noisy continual learning, which over long times leads to changes in the representations while the performance is fixed. However, the exact contributions of various sources of noise remain elusive. Of particular interest is the data-dependency of the drift, and its relation to the task within a context.Here, using theory and simulations we show that the presence of task-irrelevant stimuli, which the agent learns to ignore in a context, could be sufficient to create long-term drift in the representation of task-relevant input. By studying different architectures and learning rules, we demonstrate this phenomenon in both Hebbian-based learning (Oja and Similarity Matching), as well as stochastic gradient descent (autoencoder and two-layer networks). The results show that in all cases the drift rate increases as a function of the variance and the dimension of the data in the task-irrelevant subspace. We further demonstrate this phenomenon in a real dataset, and show that the noise induced by task-irrelevant data yields different qualitative predictions for drift than those arising from Gaussian synaptic noise. Overall, our study links the structure of stimuli, task and learning to the phenomenon of representational drift and could pave the way for using drift as a signal for uncovering underlying computation in the brain.



Paperid:2476
Authors:Chao-Chung Wu, Zhi Rui Tam, Chieh-Yen Lin, Yun-Nung (Vivian) Chen, Shao-Hua Sun, Hung-yi Lee
Abstract:
Maintaining consistent model performance across domains is a fundamental challenge in machine learning. While recent work has explored using LLM-generated data for fine-tuning, its impact on cross-domain generalization remains poorly understood. This paper presents a systematic analysis revealing that fine-tuning with LLM-generated data not only improves target task performance but also reduces non-target task degradation compared to fine-tuning with ground truth data. Through analyzing the data sequence in tasks of various domains, we demonstrate that this enhancement of non-target task robustness stems from the reduction of high perplexity tokens found in LLM-generated sequences. Following our findings, we showed that masking high perplexity tokens in ground truth training data achieves similar non-target task performance preservation, comparable to using LLM-generated data. Extensive experiments across different model families and scales, including Gemma 2 IT 2B, Llama 3 8B Instruct, and 3 additional models, agree with our findings. To the best of our knowledge, this is the first work to provide an empirical explanation based on token perplexity reduction to mitigate catastrophic forgetting in LLMs after fine-tuning, offering valuable insights for developing more robust fine-tuning strategies.



Paperid:2477
Authors:Ipsita Ghosh, Ethan Nguyen, Christian Kümmerle
Abstract:
Parameter-efficient training, based on low-rank optimization, has become a highly successful tool for fine-tuning large deep-learning models. However, these methods fail at low-rank pretraining tasks where maintaining the low-rank structure and the objective remains a challenging task. We propose theQuadratic Reweighted Rank RegularizerdubbedQER, which leads to a novel low-rank inducing training strategy inspired by the iteratively reweighted least squares (IRLS) framework.QERis based on a quadratic regularizer term which majorizes a smoothed log determinant serving as rank surrogate objective. Unlike other low-rank training techniques,QERis able to train weight matrices with prescribed, low target ranks of models that achieve comparable predictive performance as dense models, with small computational overhead, while remaining fully compatible with existing architectures. In experiments, we are able to truncate 60\% of the parameters of a ViT-Tiny parameters with marginal loss in CIFAR-10 performance and up to 80\% with only 4\% accuracy drop. The efficacy ofQERis confirmed on Transformers across both image and language tasks. To demonstrateQERtask agnosticism, we fine-tune RoBERTa usingQERregularized dense layers models on GLUE tasks, achieving performance comparable to state-of-the-art low-rank adapters.



Paperid:2478
Authors:Mingye Zhu, Yi Liu, Zheren Fu, Yongdong Zhang, Zhendong Mao
Abstract:
Preference alignment methods are increasingly critical for steering large language models (LLMs) to generate outputs consistent with human values. While recent approaches often rely on synthetic data generated by LLMs for scalability and cost-efficiency reasons, this reliance can introduce distributional shifts that undermine the nuanced representation of human preferences needed for desirable outputs. In this paper, we propose a novel distribution-aware optimization framework that improves preference alignment despite such shifts. Our approach first leverages well-learned classifiers to assign a calibration value to each training sample, quantifying its alignment with the target human-preferred distribution. These values are then incorporated into a robust optimization objective that minimizes the worst-case loss over regions of the data space most relevant to human preferences. By explicitly focusing optimization on the target distribution, our approach mitigates the impact of distributional mismatch and improves the generation of responses that better reflect intended values.



Paperid:2479
Authors:Marie-Charlotte Brandenburg, Katharina Jochemko
Abstract:
We consider binary classification restricted to a class of continuous piecewise linear functions whose decision boundaries are (possibly nonconvex) starshaped polyhedral sets, supported on a fixed polyhedral simplicial fan. We investigate the expressivity of these function classes and describe the combinatorial and geometric structure of the loss landscape, most prominently the sublevel sets, for two loss-functions: the 0/1-loss (discrete loss) and an exponential loss function. In particular, we give explicit bounds on the VC dimension of this model, and concretely describe the sublevel sets of the discrete loss as chambers in a hyperplane arrangement. For the exponential loss, we give sufficient conditions for the optimum to be unique, and describe the geometry of the optimum when varying the rate parameter of the underlying exponential probability distribution.



Paperid:2480
Authors:Qi Bi, Jingjun Yi, Hao Zheng, Huimin Huang, Haolan Zhan, Yawen Huang, Yuexiang Li, Xian Wu, Yefeng Zheng
Abstract:
Domain generalization aims to train models that perform robustly on unseen target domains without access to target data during training. The realm of vision-language foundation model has opened a new venue owing to its inherent out-of-distribution generalization capability.However, the static alignment to class-level textual anchors remains insufficient to handle the dramatic and dynamic distribution discrepancy from diverse domain-specific visual features.In this work, we propose a novel cross-domain Schrödinger Bridge (SB) method, namely SBGen, to handle this challenge, which explicitly formulates the stochastic semantic evolution, so as to gain more robustness to unseen domains.Technically, the proposed \texttt{SBGen} consists of three key components: (1) \emph{text-guided domain-aware feature selection} to isolate semantically aligned image tokens; (2) \emph{stochastic cross-domain evolution} to simulate the SB dynamics via a learnable time-conditioned drift; and (3) \emph{stochastic domain-agnostic interpolation} to construct semantically grounded feature trajectories. Empirically, \texttt{SBGen} achieves state-of-the-art performance on domain generalization in both classification and segmentation, especially under large distribution shifts. This work highlights the importance of modeling domain shifts as structured stochastic processes grounded in semantic alignment.Source code will be publicly available.



Paperid:2481
Authors:Mingze Wang, Weinan E
Abstract:
Abstract:Mixture-of-experts networks (MoEs) have demonstrated remarkable efficiency in modern deep learning. Despite their empirical success, the theoretical foundations underlying their ability to model complex tasks remain poorly understood.In this work, we conduct a systematic study of the expressive power of MoEs in modeling complex tasks with two common structural priors: low-dimensionality and sparsity.For shallow MoEs, we prove that they can efficiently approximate functions supported on low-dimensional manifolds, overcoming the curse of dimensionality.For deep MoEs, we show that $\mathcal{O}(L)$-layer MoEs with $E$ experts per layer can approximate piecewise functions comprising $E^L$ pieces with compositional sparsity, i.e., they can exhibit an exponential number of structured tasks.Our analysis reveals the roles of critical architectural components and hyperparameters in MoEs, including the gating mechanism, expert networks, the number of experts, and the number of layers, and offers natural suggestions for MoE variants.



Paperid:2482
Authors:Hao Zhang, Chun-Han Yao, Simon Donné, Narendra Ahuja, Varun Jampani
Abstract:
We present Stable Part Diffusion (SPD), a framework for generating paired RGB and kinematic part videos from monocular inputs. Unlike conventional part segmentation methods that rely on appearance-based semantic cues, SPD learns to produce kinematic parts --- structural components aligned with object articulation and consistent across views and time.SPD adopts a dual-branch diffusion model that jointly synthesizes RGB frames and corresponding part segmentation maps. To simplify architecture and flexibly enable different part counts, we introduce a spatial color encoding scheme that maps part masks to continuous RGB-like images. This encoding allows the segmentation branch to share the latents VAE from the RGB branch, while enabling part segmentation to be recovered via straightforward post-processing. A Bidirectional Diffusion Fusion (BiDiFuse) module enhances cross-branch consistency, supported by a contrastive part consistency loss to promote spatial and temporal alignment of part predictions.We demonstrate that the generated 2D part maps can be lifted to 3D to derive skeletal structures and harmonic skinning weights with few manual adjustments. To train and evaluate SPD, we construct KinematicParts20K, a curated dataset of over 20K rigged objects selected and processed from Objaverse XL, each paired with multi-view RGB and part video sequences. Experiments show that SPD generalizes strongly to diverse scenarios, including real-world videos, novel generated objects, and rare articulated poses, producing kinematic-aware outputs suitable for downstream animation and motion-related tasks.



Paperid:2483
Authors:Hugo Ninou, Jonathan Kadmon, N Alex Cayco Gajic
Abstract:
Gradient-based algorithms are a cornerstone of artificial neural network training, yet it remains unclear whether biological neural networks use similar gradient-based strategies during learning. Experiments often discover a diversity of synaptic plasticity rules, but whether these amount to an approximation to gradient descent is unclear. Here we investigate a previously overlooked possibility: that learning dynamics may include fundamentally non-gradient "curl"-like components while still being able to effectively optimize a loss function. Curl terms naturally emerge in networks with inhibitory-excitatory connectivity or Hebbian/anti-Hebbian plasticity, resulting in learning dynamics that cannot be framed as gradient descent on any objective. To investigate the impact of these curl terms, we analyze feedforward networks within an analytically tractable student-teacher framework, systematically introducing non-gradient dynamics through neurons exhibiting anti-Hebbian-like plasticity. Small curl terms preserve the stability of the original solution manifold, resulting in learning dynamics similar to gradient descent. Beyond a critical value, strong curl terms destabilize the solution manifold. Depending on the network architecture, this loss of stability can lead to chaotic learning dynamics that destroy performance. In other cases, the curl terms can counterintuitively speed learning compared to gradient descent by allowing the weight dynamics to escape saddles by temporarily ascending the loss. Our results identify specific architectures capable of supporting robust learning via diverse learning rules, providing an important counterpoint to normative theories of gradient-based learning in neural networks.



Paperid:2484
Authors:Lowell Weissman, Michael Krumdick, A. Abbott
Abstract:
Recent work on neural scaling laws demonstrates that model performance scales predictably with compute budget, model size, and dataset size. In this work, we develop scaling laws based on problem complexity. We analyze two fundamental complexity measures: solution space size and representation space size. Using the Traveling Salesman Problem (TSP) as a case study, we show that combinatorial optimization promotes smooth cost trends, and therefore meaningful scaling laws can be obtained even in the absence of an interpretable loss. We then show that suboptimality grows predictably for fixed-size models when scaling the number of TSP nodes or spatial dimensions, independent of whether the model was trained with reinforcement learning or supervised fine-tuning on a static dataset. We conclude with an analogy to problem complexity scaling in local search, showing that a much simpler gradient descent of the cost landscape produces similar trends.



Paperid:2485
Authors:Xiuhui Zhang, Ning Gao, Xingyu Jiang, Yihui Chen, Yuheng Pan, Mohan Zhang, Yue Deng
Abstract:
Traditional reinforcement learning (RL) algorithms face significant limitations in handling long-term tasks with sparse rewards. Recent advancements have leveraged large language models (LLMs) to enhance RL by utilizing their world knowledge for task planning and reward generation. However, planning-based approaches often depend on pre-defined skill libraries and fail to optimize low-level control policies, while reward-based methods require extensive human feedback or exhaustive searching due to the complexity of tasks.In this paper, we propose the Progress Reward Model for RL (PRM4RL), a novel framework that integrates task planning and dense reward to enhance RL. For high-level planning, a complex task is decomposed into a series of simple manageable subtasks, with a subtask-oriented, fine-grained progress function designed to monitor task execution progress. For low-level reward generation, inspired by potential-based reward shaping, we use the progress function to construct a Progress Reward Model (PRM), providing theoretically grounded optimality and convergence guarantees, thereby enabling effective policy optimization.Experimental results on robotics control tasks demonstrate that our approach outperforms both LLM-based planning and reward methods, achieving state-of-the-art performance.



Paperid:2486
Authors:Amartya Chakraborty, Paresh Dashore, Nadia Bathaee, Anmol Jain, Anirban Das, Shi-Xiong Zhang, Sambit Sahu, Milind Naphade, Genta Winata
Abstract:
Large Language Models (LLMs) have demonstrated impressive capabilities as intelligent agents capable of solving complex problems. However, effective planning in scenarios involving dependencies between API or tool calls—particularly in multi-turn conversations—remains a significant challenge. To address this, we introduce T1, a tool-augmented, multi-domain, multi-turn conversational dataset specifically designed to capture and manage inter-tool dependencies across diverse domains. T1 enables rigorous evaluation of agents’ ability to coordinate tool use across nine distinct domains (4 single domain and 5 multi-domain) with the help of an integrated caching mechanism for both short- and long-term memory, while supporting dynamic replanning—such as deciding whether to recompute or reuse cached results. Beyond facilitating research on tool use and planning, T1 also serves as a benchmark for evaluating the performance of open-source language models. We present results powered by T1-Agent, highlighting their ability to plan and reason in complex, tool-dependent scenarios.



Paperid:2487
Authors:Yueyang Yuan, Wenke Huang, Guancheng Wan, Kaiqi Guan, He Li, Mang Ye
Abstract:
Federated Learning (FL) has demonstrated a promising future in privacy-friendly collaboration but it faces the data heterogeneity problem. Knowledge Distillation (KD) can serve as an effective method to address this issue. However, challenges arise from the unreliability of existing distillation methods in multi-domain scenarios. Prevalent distillation solutions primarily aim to fit the distributions of the global model directly by minimizing forward Kullback-Leibler divergence (KLD). This results in significant bias when the outputs of the global model are multi-peaked, which indicates the unreliability of the distillation pathway. Meanwhile, cross-domain update conflicts can notably reduce the accuracy of the global model (teacher model) in certain domains, reflecting the unreliability of the teacher model in these domains. In this work, we propose DKDR (Dynamic Knowledge Distillation for Reliability in Federated Learning), which dynamically assigns weights to forward and reverse KLD based on knowledge discrepancies. This enables clients to fit the outputs from the teacher precisely. Moreover, we use knowledge decoupling to identify domain experts, thus clients can acquire reliable domain knowledge from experts. Empirical results from single-domain and multi-domain image classification tasks demonstrate the effectiveness of the proposed method and the efficiency of its key modules. The code is available at https://anonymous.4open.science/r/DKDR-22B5.



Paperid:2488
Authors:Zongle Huang, Lei Zhu, ZongYuan Zhan, Ting Hu, Weikai Mao, Xianzhi Yu, Yongpan Liu, Tianyu Zhang
Abstract:
Large Language Models (LLMs) have achieved remarkable success across many applications, with Mixture of Experts (MoE) models demonstrating great potential. Compared to traditional dense models, MoEs achieve better performance with less computation. Speculative decoding (SD) is a widely used technique to accelerate LLM inference without accuracy loss, but it has been considered efficient only for dense models. In this work, we first demonstrate that, under medium batch sizes, MoE surprisingly benefits more from SD than dense models. Furthermore, as MoE becomes sparser -- the prevailing trend in MoE designs -- the batch size range where SD acceleration is expected to be effective becomes broader. To quantitatively understand tradeoffs involved in SD, we develop a reliable modeling based on theoretical analyses. While current SD research primarily focuses on improving acceptance rates of algorithms, changes in workload and model architecture can still lead to degraded SD acceleration even with high acceptance rates. To address this limitation, we introduce a new metric 'target efficiency' that characterizes these effects, thus helping researchers identify system bottlenecks and understand SD acceleration more comprehensively. For scenarios like private serving, this work unveils a new perspective to speed up MoE inference, where existing solutions struggle. Experiments on different GPUs show up to 2.29x speedup for Qwen2-57B-A17B at medium batch sizes and validate our theoretical predictions.



Paperid:2489
Authors:Rose Gurung, Ronilo Ragodos, Chiyu Ma, Tong Wang, Chaofan Chen
Abstract:
We present Prototypical Pair Network (ProtoPairNet), a novel interpretable architecture that combines deep learning with case-based reasoning to predict continuous targets. While prototype-based models have primarily addressed image classification with discrete outputs, extending these methods to continuous targets, such as regression, poses significant challenges. Existing architectures which rely heavily on one-to-one comparison with prototypes lack the directional information necessary for continuous predictions. Our method redefines the role of prototypes in such tasks by incorporating prototypical pairs into the reasoning process. Predictions are derived based on the input's relative dissimilarities to these pairs, leveraging an intuitive geometric interpretation. Our method further reduces the complexity of the reasoning process by relying on the single most relevant pair of prototypes, rather than all prototypes in the model as was done in prior works. Our model is versatile enough to be used in both vision-based regression and continuous control in reinforcement learning. Our experiments demonstrate that ProtoPairNet achieves performance on par with its black-box counterparts across these tasks. Comprehensive analyses confirm the meaningfulness of prototypical pairs and the faithfulness of our model’s interpretations, and extensive user studies highlight our model's improved interpretability over existing methods.



Paperid:2490
Authors:Weijun Lv, Yu Chen, Xuhuan Zhu, Jie Wen, Guoxu Zhou, Sixian Chan, Xiaozhao Fang
Abstract:
Label prototype learning has emerged as an effective paradigm in Partial Multi-Label Learning (PML), providing a distinctive framework for modeling structured representations of label semantics while naturally filtering noise through prototype-based label confidence estimation. However, existing prototype-based methods face a critical limitation: class prototypes are the biased estimates due to noisy candidate labels, particularly when positive samples are scarce. To this end, we first propose a mutually class prototype alignment strategy bypassing noise interference by introducing two different transformation matrices, which makes the class prototypes learned by the fuzzy clustering and candidate label set mutually alignment for correcting themselves. Such alignment is also passed on to the fuzzy memberships label in turn. In addition, to eliminate noise interference in the candidate label set during the classifier learning, we use the learned permutation matrix to transform the fuzzy memberships label for learning a label enhancement indicator matrix accompanied by the candidate label set. This makes the label enhancement indicator matrix absolutely prevent the occurrence of numerical values located in non-label and simultaneously eliminate the introduction of incorrect label as much as possible. The resulting indicator matrix guides a robust multi-label classifier training process, jointly optimizing label confidence and classifier parameters. Extensive experiments demonstrate that our proposed model exhibits significant performance advantages over state-of-the-art PML approaches.



Paperid:2491
Authors:Benjamin Minixhofer, Ivan Vulić, Edoardo Maria Ponti
Abstract:
Distillation has shown remarkable success in transferring knowledge from a Large Language Model (LLM) teacher to a student LLM. However, current distillation methods require similar tokenizers between the teacher and the student, restricting their applicability to only a small subset of teacher--student pairs. In this work, we develop a principled cross-tokenizer distillation method to solve this crucial deficiency. Our method is the first to enable effective distillation across fundamentally different tokenizers, while also substantially outperforming prior methods in all other cases. We verify the efficacy of our method on three distinct use cases. First, we show that viewing tokenizer transfer as self-distillation enables unprecedentedly effective transfer across tokenizers, including rapid transfer of subword models to the byte-level. Transferring different models to the same tokenizer also enables ensembling to boost performance. Secondly, we distil a large maths-specialised LLM into a small general-purpose model with a different tokenizer, achieving competitive maths problem-solving performance. Thirdly, we use our method to train state-of-the-art embedding prediction hypernetworks for training-free tokenizer transfer. Our results unlock an expanded range of teacher--student pairs for distillation, enabling new ways to adapt and enhance interaction between LLMs.



Paperid:2492
Authors:Jinglong Shen, Nan Cheng, Wenchao Xu, Haozhao Wang, Yifan guo, Jiajie Xu
Abstract:
Abstract:Fine-tuning large language models (LLMs) poses significant computational burdens, especially in federated learning (FL) settings. We introduce Layer-wise Efficient Federated Fine-tuning (LEFF), a novel method designed to enhance the efficiency of FL fine-tuning while preserving model performance and minimizing client-side computational overhead. LEFF strategically selects layers for fine-tuning based on client computational capacity, thereby mitigating the straggler effect prevalent in heterogeneous environments. Furthermore, LEFF incorporates an importance-driven layer sampling mechanism, prioritizing layers with greater influence on model performance. Theoretical analysis demonstrates that LEFF achieves a convergence rate of $\mathcal{O}(1/\sqrt{T})$. Extensive experiments on diverse datasets demonstrate that LEFF attains superior computational efficiency and model performance compared to existing federated fine-tuning methods, particularly under heterogeneous conditions.



Paperid:2493
Authors:Ruoxin Chen, Junwei Xi, Zhiyuan Yan, Ke-Yue Zhang, Shuang Wu, Jingyi Xie, Xu Chen, Lei Xu, Isabel Guan, Taiping Yao, Shouhong Ding
Abstract:
The rapid increase in AI-generated images (AIGIs) underscores the urgent need for generalizable detection methods. Existing detectors, however, are often trained on biased datasets, leading to the possibility of overfitting on non-causal image attributes that are spuriously correlated with real/synthetic labels. While these biased features enhance performance on the training data, they result in substantial performance degradation when applied to unbiased datasets. One common solution is to perform dataset alignment through generative reconstruction, matching the semantic content between real and synthetic images. However, we revisit this approach and show that pixel-level alignment alone is insufficient — the reconstructed images still suffer from frequency-level misalignment, which can perpetuate spurious correlations. To illustrate, we observe that reconstruction models tend to restore the high-frequency details lost in real images (possibly due to JPEG compression), inadvertently creating a frequency-level misalignment, where synthetic images appear to have richer high-frequency content than real ones. This misalignment leads to models associating high-frequency features with synthetic labels, further reinforcing biased cues. To resolve this, we propose Dual Data Alignment (DDA), which aligns both the pixel and frequency domains. Moreover, we introduce two new test sets: DDA-COCO, containing DDA-aligned synthetic images for testing detector performance on the most aligned dataset, and EvalGEN, featuring the latest generative models for assessing detectors under new generative architectures such as visual auto-regressive generators. Finally, our extensive evaluations demonstrate that a detector trained exclusively on DDA-aligned MSCOCO could improve across 8 diverse benchmarks by a non-trivial margin, showing a +7.2\% on in-the-wild benchmarks, highlighting the improved generalizability of unbiased detectors.



Paperid:2494
Authors:Matthew Tamayo-Rios, Alexander Schell, Rima Alaifari
Abstract:
The signature kernel is a recent state-of-the-art tool for analyzing high-dimensional sequential data, valued for its theoretical guarantees and strong empirical performance. In this paper, we present a novel method for efficiently computing the signature kernel of long, high-dimensional time series via adaptively truncated recursive local power series expansions. Building on the characterization of the signature kernel as the solution of a Goursat PDE, our approach employs tilewise Neumann‐series expansions to derive rapidly converging power series approximations of the signature kernel that are locally defined on subdomains and propagated iteratively across the entire domain of the Goursat solution by exploiting the geometry of the time series. Algorithmically, this involves solving a system of interdependent Goursat PDEs via adaptively truncated local power series expansions and recursive propagation of boundary conditions along a directed graph in a topological ordering. This method strikes an effective balance between computational cost and accuracy, achieving substantial performance improvements over state-of-the-art approaches for computing the signature kernel. It offers (a) adjustable and superior accuracy, even for time series with very high roughness; (b) drastically reduced memory requirements; and (c) scalability to efficiently handle very long time series (e.g., with up to one million data points or more) on a single GPU. As demonstrated in our benchmarks, these advantages make our method particularly well-suited for rough-path-assisted machine learning, financial modeling, and signal processing applications involving very long and highly volatile sequential data.



Paperid:2495
Authors:Hui Chen, Vikas Singh
Abstract:
Diffusion models have achieved significant success, yet their application to time series data, particularly with regard to efficient sampling, remains an active area of research. We describe an operator-learning approach for conditioned time-series diffusion models that gives efficient single-step generation by leveraging insights from the frequency-domain characteristics of both the time-series data and the diffusion process itself. The forward diffusion process induces a structured, frequency-dependent smoothing of the data's probability density function. However, this frequency smoothing is related (e.g., via likelihood function) to easily accessible frequency components of time-series data. This suggests that a module operating in the frequency space of the time-series can, potentially, more effectively learn to reverse the frequency-dependent smoothing of the data distribution induced by the diffusion process. We set up an operator learning task, based on frequency-aware building blocks, which satisfies semi-group properties, while exploiting the structure of time-series data. Evaluations on multiple datasets show that our single-step generation proposal achieves forecasting/imputation results comparable (or superior) to many multi-step diffusion schemes while significantly reducing inference costs.



Paperid:2496
Authors:Cristian Sbrolli, Matteo Matteucci
Abstract:
The whole is greater than the sum of its parts—even in 3D-text contrastive learning. We introduce SceneForge, a novel framework that enhances contrastive alignment between 3D point clouds and text through structured multi-object scene compositions. SceneForge leverages individual 3D shapes to construct multi-object scenes with explicit spatial relations, pairing them with coherent multi-object descriptions refined by a large language model. By augmenting contrastive training with these structured, compositional samples, SceneForge effectively addresses the scarcity of large-scale 3D-text datasets, significantly enriching data complexity and diversity. We systematically investigate critical design elements, such as the optimal number of objects per scene, the proportion of compositional samples in training batches, and scene construction strategies. Extensive experiments demonstrate that SceneForge delivers substantial performance gains across multiple tasks, including zero-shot classification on ModelNet, ScanObjNN, Objaverse-LVIS, and ScanNet, as well as few-shot part segmentation on ShapeNetPart. SceneForge’s compositional augmentations are model-agnostic, consistently improving performance across multiple encoder architectures. Moreover, SceneForge improves 3D visual question answering on ScanQA, generalizes robustly to retrieval scenarios with increasing scene complexity, and showcases spatial reasoning capabilities by adapting spatial configurations to align precisely with textual instructions.



Paperid:2497
Authors:Roberto Gheda, Abele Mălan, Thiago Guzella, Carlo Lancia, Robert Birke, Lydia Chen
Abstract:
Bayesian Network models are a powerful tool to collaboratively optimize production processes in various manufacturing industries.While interacting, collaborating parties must preserve their business secrets by maintaining the confidentiality of their model structures and parameters.While most realistic industry scenarios involve hybrid scenarios, handling both discrete and continuous data, current state-of-the-art methods for collaborative and confidential inference only support discrete data and have high communication costs.In a centralized setting, Junction Trees enable efficient inference even in hybrid scenarios without discretizing continuous variables, but no extension for collaborative and confidential scenarios exists.To address this research gap, we introduce Hybrid CCJT, the first framework for confidential multiparty inference in hybrid domains, comprising:(i) a method to construct a strongly-rooted Junction Tree across collaborating parties through a novel construct of interface cliques; and,(ii) a protocol for confidential inference built upon multiparty computation primitives comprising a one-time alignment phase and a belief propagation system for combining the inference results across the Junction Tree cliques.Extensive evaluation on nine datasets shows that Hybrid CCJT improves the predictive accuracy of continuous target variables by 32% on average against the state-of-the-art, while reducing communication costs up to 331X under purely discrete scenarios.



Paperid:2498
Authors:Zhiwei Yang, Chen Gao, Mike Zheng Shou
Abstract:
Video anomaly detection (VAD) is a critical yet challenging task due to the complex and diverse nature of real-world scenarios. Previous methods typically rely on domain-specific training data and manual adjustments when applying to new scenarios and unseen anomaly types, suffering from high labor costs and limited generalization. Therefore, we aim to achieve generalist VAD, \ie, automatically handle any scene and any anomaly types without training data or human involvement.In this work, we propose PANDA, a detective-like agent based on MLLMs. Specifically, we achieve PANDA by comprehensively devising four key capabilities: (1) self-adaption scene-aware strategy planning, (2) goal-driven heuristic reasoning, (3) tool-augmented self-reflection, and (4) self-improving chain-of-memory. Concretely, we develop a self-adaption scene-aware RAG mechanism, enabling PANDA to retrieve anomaly-specific knowledge for anomaly detection strategy planning. Next, we introduce a latent anomaly-guided heuristic prompt strategy to enhance reasoning precision. Furthermore, PANDA employs a progressive reflection mechanism alongside a suite of context-aware tools to iteratively refine decision-making in complex scenarios. Finally, a chain-of-memory mechanism enables PANDA to leverage historical experiences for continual performance improvement. Extensive experiments demonstrate that PANDA achieves state-of-the-art performance in multi-scenario, open-set, and complex scenario settings without training and manual involvement, validating its generalizable and robust anomaly detection capability.



Paperid:2499
Authors:Georgiana Manolache, Gerard Schouten, Joaquin Vanschoren
Abstract:
We present CrypticBio, the largest publicly available multimodal dataset of visually confusing species, specifically curated to support the development of AI models in the context of biodiversity applications. Visually confusing or cryptic species are groups of two or more taxa that are nearly indistinguishable based on visual characteristics alone. While much existing work addresses taxonomic identification in a broad sense, datasets that directly address the morphological confusion of cryptic species are small, manually curated, and target only a single taxon. Thus, the challenge of identifying such subtle differences in a wide range of taxa remains unaddressed. Curated from real-world trends in species misidentification among community annotators of iNaturalist, CrypticBio contains 52K unique cryptic groups spanning 67K species, represented in 166 million images. Rich research-grade image annotations—including scientific, multicultural, and multilingual species terminology, hierarchical taxonomy, spatiotemporal context, and associated cryptic groups—address multimodal AI in biodiversity research. For easy dataset curation, we provide an open-source pipeline CrypticBio-Curate. The multimodal nature of the dataset beyond vision-language arises from the integration of geographical and temporal data as complementary cues to identifying cryptic species. To highlight the importance of the dataset, we benchmark a suite of state-of-the-art foundation models across CrypticBio subsets of common, unseen, endangered, and invasive species, and demonstrate the substantial impact of geographical context on vision-language zero-shot learning for cryptic species. By introducing CrypticBio, we aim to catalyze progress toward real-world-ready biodiversity AI models capable of handling the nuanced challenges of species ambiguity. The data and the code are publicly available in the https://georgianagmanolache.github.io/crypticbio.



Paperid:2500
Authors:Yang Miao, Jan-Nico Zaech, Xi Wang, Fabien Despinoy, Danda Pani Paudel, Luc V Gool
Abstract:
We propose LangHOPS, the first Multimodal Large Language Model (MLLM)-based framework for open-vocabulary object–part instance segmentation. Given an image, LangHOPS can jointly detect and segment hierarchical object and part instances from open-vocabulary candidate categories. Unlike prior approaches that rely on heuristic or learnable visual grouping, our approach grounds object–part hierarchies in language space. It integrates the MLLM into the object-part parsing pipeline to leverage rich knowledge and reasoning capabilities, and link multi-granularity concepts within the hierarchies. We evaluate LangHOPS across multiple challenging scenarios, including in-domain and cross-dataset object-part instance segmentation, and zero-shot semantic segmentation. LangHOPS achieves state-of-the-art results, surpassing previous methods by 5.5% Average Precision(AP) (in-domain) and 4.8% (cross-dataset) on the PartImageNet dataset and by 2.5% mIOU on unseen object parts in ADE20K (zero-shot). Ablation studies further validate the effectiveness of the language-grounded hierarchy and MLLM-driven part query refinement strategy. Our results establish LangHOPS as a strong foundation for advancing open-vocabulary fine-grained visual understanding applicable in multiple scenarios.



Paperid:2501
Authors:Mark H. Huang, Lin Geng Foo, Christian Theobalt, Ying Sun, De Wen Soh
Abstract:
Free-moving object reconstruction from monocular video remains challenging, particularly without reliable pose or depth cues and under arbitrary object motion. We introduce OnlineSplatter, a novel online feed-forward framework generating high-quality, object-centric 3D Gaussians directly from RGB frames without requiring camera pose, depth priors, or bundle optimization. Our approach anchors reconstruction using the first frame and progressively refines the object representation through a dense Gaussian primitive field, maintaining constant computational cost regardless of sequence length. Our core contribution is a dual-key memory module combining latent appearance-geometry keys with explicit directional keys, robustly fusing current frame features with temporally aggregated object states. This design enables effective handling of free-moving objects via spatial-guided memory readout and an efficient sparsification mechanism, ensuring comprehensive yet compact object coverage. Evaluations on real-world datasets demonstrate that OnlineSplatter significantly outperforms state-of-the-art pose-free reconstruction baselines, consistently improving with more observations while maintaining constant memory and runtime, ideal for online robotics applications.



Paperid:2502
Authors:Frank F. Xu, Yufan Song, Boxuan Li, Yuxuan Tang, Kritanjali Jain, Mengxue Bao, Zhiruo Wang, Xuhui Zhou, Zhitong Guo, Murong Cao, Mingyang Yang, Hao Yang Lu, Amaad Martin, Zhe Su, Leander Maben, Raj Mehta, Wayne Chi, Lawrence Jang, Yiqing Xie, Shuyan Zhou, Graham Neubig
Abstract:
We interact with computers on an everyday basis, be it in everyday life or work, and many aspects of work can be done entirely with access to a computer and the Internet. At the same time, thanks to improvements in large language models (LLMs), there has also been a rapid development in AI agents that interact with and affect change in their surrounding environments. But how performant are AI agents at helping to accelerate or even autonomously perform work-related tasks? The answer to this question has important implications for both industry looking to adopt AI into their workflows, and for economic policy to understand the effects that adoption of AI may have on the labor market. To measure the progress of these LLM agents' performance on performing real-world professional tasks, in this paper, we introduce TheAgentCompany, an extensible benchmark for evaluating AI agents that interact with the world in similar ways to those of a digital worker: by browsing the Web, writing code, running programs, and communicating with other coworkers. We build a self-contained environment with internal web sites and data that mimics a small software company environment, and create a variety of tasks that may be performed by workers in such a company. We test baseline agents powered by both closed API-based and open-weights language models (LMs), and find that with the most competitive agent, 30% of the tasks can be completed autonomously. This paints a nuanced picture on task automation with LM agents -- in a setting simulating a real workplace, a good portion of simpler tasks could be solved autonomously, but more difficult long-horizon tasks are still beyond the reach of current systems. For more information and demos, refer to https://the-agent-company.com.



Paperid:2503
Authors:Azim Ospanov, Farzan Farnia, Roozbeh Yousefzadeh
Abstract:
Abstract:Formal reasoning and automated theorem proving constitute a challenging subfield of machine learning, in which machines are tasked with proving mathematical theorems using formal languages like Lean. A formal verification system can check whether a formal proof is correct or not almost instantaneously, but generating a completely correct formal proof with large language models (LLMs) remains a formidable task. The usual approach in the literature is to prompt the LLM many times (up to several thousands) until one of the generated proofs passes the verification system. In this work, we present APOLLO (**A**utomated **P**r**O**of repair via **L**LM and **L**ean c**O**llaboration), a modular, model‑agnostic pipeline that combines the strengths of the Lean compiler with an LLM’s reasoning abilities to achieve better proof‐generation results at a low sampling budget. _Apollo_ directs a fully automated process in which the LLM generates proofs for theorems, a set of agents analyze the proofs, fix the syntax errors, identify the mistakes in the proofs using Lean, isolate failing sub‑lemmas, utilize automated solvers, and invoke an LLM on each remaining goal with a low top‑$K$ budget. The repaired sub‑proofs are recombined and reverified, iterating up to a user‑controlled maximum number of attempts. On the miniF2F benchmark, we establish a new state‑of‑the‑art accuracy of 75.0\% among 7B‑parameter models while keeping the sampling budget below one thousand. Moreover, _Apollo_ raises the state‑of‑the‑art accuracy for Goedel‑Prover‑SFT to 65.6\% while cutting sample complexity from 25,600 to a few hundred. General‑purpose models (o3‑mini, o4‑mini) jump from 3–7\% to over 40\% accuracy. Our results demonstrate that targeted, compiler‑guided repair of LLM outputs yields dramatic gains in both efficiency and correctness, suggesting a general paradigm for scalable automated theorem proving.



Paperid:2504
Authors:Ran Levinstein, Amit Attia, Matan Schliserman, Uri Sherman, Daniel Soudry, Tomer Koren, Itay Evron
Abstract:
Abstract:We study realizable continual linear regression under random task orderings, a common setting for developing continual learning theory.In this setup, the worst-case expected loss after $k$ learning iterations admits a lower bound of $\Omega(1/k)$. However, prior work using an unregularized scheme has only established an upper bound of $O(1/k^{1/4})$, leaving a significant gap.Our paper proves that this gap can be narrowed, or even closed, using two frequently used regularization schemes:(1) explicit isotropic $\ell_2$ regularization, and (2) implicit regularization via finite step budgets.We show that these approaches, which are used in practice to mitigate forgetting, reduce to stochastic gradient descent (SGD) on carefully defined surrogate losses. Through this lens, we identify a fixed regularization strength that yields a near-optimal rate of $O(\log k / k)$.Formalizing and analyzing a generalized variant of SGD for time-varying functions, we derive an increasing regularization strength schedule that provably achieves an optimal rate of $O(1/k)$.This suggests that schedules that increase the regularization coefficient or decrease the number of steps per task are beneficial, at least in the worst case.



Paperid:2505
Authors:Gholamali Aminian, Amir Asadi, Idan Shenfeld, Youssef Mroueh
Abstract:
Recent methods for aligning large language models (LLMs) with human feedback predominantly rely on a single reference model, which limits diversity, model overfitting, and underutilizes the wide range of available pre-trained models. Incorporating multiple reference models has the potential to address these limitations by broadening perspectives, reducing bias, and leveraging the strengths of diverse open-source LLMs. However, integrating multiple reference models into reinforcement learning with human feedback (RLHF) frameworks poses significant theoretical challenges, where achieving exact solutions has remained an open problem. This paper presents the first \emph{exact solution} to the multiple reference model problem in reverse KL-regularized RLHF. We introduce a comprehensive theoretical framework that includes rigorous statistical analysis and provides sample complexity guarantees. Additionally, we extend our analysis to forward KL-regularized RLHF, offering new insights into sample complexity requirements in multiple reference scenarios. Our contributions lay the foundation for more advanced and adaptable LLM alignment techniques, enabling the effective use of multiple reference models. This work paves the way for developing alignment frameworks that are both theoretically sound and better suited to the challenges of modern AI ecosystems.



Paperid:2506
Authors:Jack Nugent, Siyang Wu, Zeyu Ma, Beining Han, Meenal Parakh, Abhishek Joshi, Lingjie Mei, Alexander Raistrick, Xinyuan Li, Jia Deng
Abstract:
Recent years have witnessed substantial progress on monocular depth estimation, particularly as measured by the success of large models on standard benchmarks. However, performance on standard benchmarks does not offer a complete assessment, because most evaluate accuracy but not robustness. In this work, we perform a thorough evaluation of the robustness of state-of-the-art monocular depth models. We use procedural generation to create 3D scenes which test robustness to various controlled perturbations, including object, camera, material and lighting changes. Our analysis yields interesting findings on what perturbations are challenging for state-of-the-art depth models, which we hope will inform further research.



Paperid:2507
Authors:Tianzhe Xiao, Yichen Li, Yining Qi, YI LIU, wangshi.ww, Haozhao Wang, Yi Wang, Ruixuan Li
Abstract:
Multimodal federated learning (MMFL) has gained increasing popularity due to its ability to leverage the correlation between various modalities, meanwhile preserving data privacy for different clients. However, recent studies show that correlation between modalities increase the vulnerability of federated learning against Gradient Inversion Attack (GIA). The complicated situation of MMFL privacy preserving can be summarized as follows: 1) different modality transmits different amounts of information, thus requires various protection strength; 2) correlation between modalities should be taken into account. This paper introduces an information theory perspective to analyze the leaked privacy in process of MMFL, and tries to propose a more reasonable protection method \textbf{Sec-MMFL} based on assessing different information leakage possibilities of each modality by conditional mutual information and adjust the corresponding protection strength. Moreover, we use mutual information to reduce the cross-modality information leakage in MMFL. Experiments have proven that our method can bring more balanced and comprehensive protection at an acceptable cost.



Paperid:2508
Authors:Shenghui Chen, Po-han Li, Sandeep Chinchali, Ufuk Topcu
Abstract:
Abstract:Many decision-making tasks, where both accuracy and efficiency matter, still require human supervision. For example, tasks like traffic officers reviewing hour-long dashcam footage or researchers screening conference videos can benefit from concise summaries that reduce cognitive load and save time. Yet current vision-language models (VLMs) often produce verbose, redundant outputs that hinder task performance. Existing video caption evaluation depends on costly human annotations and overlooks the summaries' utility in downstream tasks. We address these gaps with $\underline{\textbf{V}}$ideo-to-text $\underline{\textbf{I}}$nformation $\underline{\textbf{B}}$ottleneck $\underline{\textbf{E}}$valuation (VIBE), an annotation-free method that scores VLM outputs using two metrics: $\textit{grounding}$ (how well the summary aligns with visual content) and $\textit{utility}$ (how informative it is for the task). VIBE selects from randomly sampled VLM outputs by ranking them according to the two scores to support effective human decision-making. Human studies on $\texttt{LearningPaper24}$, $\texttt{SUTD-TrafficQA}$, and $\texttt{LongVideoBench}$ show that summaries selected by VIBE consistently improve performance—boosting task accuracy by up to $61.23$% and reducing response time by $75.77$% compared to naive VLM summaries or raw video.



Paperid:2509
Authors:Yongbo Wang, Haonan Wang, Guodong Mu, Ruixin Zhang, Jiaqi Chen, Jingyun Zhang, Jun Wang, Yuan Xie, zhizhong zhang, Shouhong Ding
Abstract:
Codebook-based image compression has achieved lower bits per pixel (bpp) while maintaining high reconstruction quality. These approaches utilize a globally shared codebook to quantize and reconstruct each token, controlling the bpp by adjusting the number of tokens or the codebook size. However, these methods perform poorly on facial images at low bpp: reducing the number of tokens or the codebook size significantly degrades the recognition performance of reconstructed faces. By analyzing facial images, we observed that pictures sharing the same attributes (e.g., gender, age, race) often exhibit many common features. At the same time, different facial regions demonstrate distinct characteristics (e.g., nose vs. ear).Based on these observations, we propose a group-routed codebook quantization and reconstruction method. By recording the codebook group to which each token belongs with a small number of bits, our method can reduce the loss incurred when decreasing the size of each codebook group. This enables a larger total number of codebooks under a lower overall bpp, thereby enhancing the expressive power and improving reconstruction performance. On face recognition datasets, our method outperforms state-of-the-art approaches such as 87.56\% and 91.66\% at the same bpp (0.0234), and remains competitive even when the bpp is reduced to 0.0040 (66.02\%).



Paperid:2510
Authors:Ziqiao Peng, Jiwen Liu, Haoxian Zhang, Xiaoqiang Liu, Songlin Tang, Pengfei Wan, Di ZHANG, Hongyan Liu, Jun He
Abstract:
Lip synchronization is the task of aligning a speaker’s lip movements in video with corresponding speech audio, and it is essential for creating realistic, expressive video content. However, existing methods often rely on reference frames and masked-frame inpainting, which limit their robustness to identity consistency, pose variations, facial occlusions, and stylized content. In addition, since audio signals provide weaker conditioning than visual cues, lip shape leakage from the original video will affect lip sync quality.In this paper, we present OmniSync, a universal lip synchronization framework for diverse visual scenarios. Our approach introduces a mask-free training paradigm using Diffusion Transformer models for direct frame editing without explicit masks, enabling unlimited-duration inference while maintaining natural facial dynamics and preserving character identity.During inference, we propose a flow-matching-based progressive noise initialization to ensure pose and identity consistency, while allowing precise mouth-region editing. To address the weak conditioning signal of audio, we develop a Dynamic Spatiotemporal Classifier-Free Guidance (DS-CFG) mechanism that adaptively adjusts guidance strength over time and space.We also establish the AIGC-LipSync Benchmark, the first evaluation suite for lip synchronization in diverse AI-generated videos. Extensive experiments demonstrate that OmniSync significantly outperforms prior methods in both visual quality and lip sync accuracy, achieving superior results in both real-world and AI-generated videos.



Paperid:2511
Authors:Qingming LIU, Zhen Liu, Dinghuai Zhang, Kui Jia
Abstract:
Generating high-quality and photorealistic 3D assets remains a longstanding challenge in 3D vision and computer graphics. Although state-of-the-art generative models, such as diffusion models, have made significant progress in 3D generation, they often fall short of human-designed content due to limited ability to follow instructions, align with human preferences, or produce realistic textures, geometries, and physical attributes. In this paper, we introduce Nabla-R2D3, a highly effective and sample-efficient reinforcement learning alignment framework for 3D-native diffusion models using 2D rewards. Built upon the recently proposed Nabla-GFlowNet method for reward finetuning, our Nabla-R2D3 enables effective adaptation of 3D diffusion models through pure 2D reward feedback. Extensive experiments show that, unlike naive finetuning baselines which either fail to converge or suffer from overfitting, Nabla-R2D3 consistently achieves higher rewards and reduced prior forgetting within few finetuning steps.



Paperid:2512
Authors:Ethan Hwang, Hossein Adeli, Wenxuan Guo, Andrew Luo, Nikolaus Kriegeskorte
Abstract:
A fine-grained understanding of functional selectivity in human visual cortex is fundamental for elucidating how visual information is processed and represented in the brain. While recent work utilizing controlled stimuli has uncovered categorically selective regions, these approaches apply strong assumptions about the stimulus space and may bias the analysis to preconceived notions of selectivity. Approaches that employ simple encoding models lack the capacity to capture the complexity of neural responses under naturalistic conditions. We propose a novel in silico approach based on an encoder-decoder transformer that leverages a brain-region to image-feature cross-attention mechanism, designed to non-linearly align high-dimensional deep network features with cortical responses. Unlike traditional encoding models that apply linear regression to deep-network activations---which assume cortical activity is a fixed affine transformation of artificial representations---our architecture learns to flexibly bind stimulus features with semantic patterns of fMRI activity in an end-to-end manner. Coupled with diffusion-based image generative models and large image datasets, this method can synthesize and select images that can maximally activate different cortical parcels. We show this approach can uncover regions with complex compositional selectivity involving diverse semantic concepts. Together, these results demonstrate a new data-driven paradigm for discovery of visual selectivity in the brain.



Paperid:2513
Authors:Kevin Zhang, Yonghan Jung, Divyat Mahajan, Karthikeyan Shanmugam, Shalmali Joshi
Abstract:
Identifying and measuring biases associated with sensitive attributes is a crucial consideration in healthcare to prevent treatment disparities. One prominent issue is inaccurate pulse oximeter readings, which tend to overestimate oxygen saturation for dark-skinned patients and misrepresent supplemental oxygen needs. Most existing research has revealedstatistical disparitieslinking device errors to patient outcomes in Intensive Care Units (ICUs) without causal formalization. In contrast, this studycausallyinvestigates how racial discrepancies in oximetry measurements affect invasive ventilation in intensive care units (ICUs). We introduce a causal inference-based approach usingpath-specific effectsto isolate the impact of bias by race on clinical decision-making. To estimate these effects, we leverage a doubly robust estimator, propose its self-normalized variant for improved sample efficiency, and provide novel finite-sample guarantees. Our methodology is validated on semi-synthetic data and applied to two large real-world health datasets: MIMIC-IV and eICU. Contrary to prior work, our analysis reveals minimal impact of racial discrepancies on invasive ventilation rates. However, path-specific effects mediated by oxygen saturation disparity were more pronounced on ventilation duration and differed by dataset. Our work provides a novel and practical pipeline for investigating potential disparities in the ICU and, more crucially, highlights the necessity of causal methods to robustly assess fairness in decision-making.



Paperid:2514
Authors:Bowei Zhu, Shaojie Li, Mingyang Yi, Yong Liu
Abstract:
Abstract:Prior work establishes at most $O\left(\log (n)/n\right)$ excess risk bounds via algorithmic stability for strongly convex learners with high probability. We show that under the same common assumptions—strong convexity, smoothness, and Lipschitz continous for losses— rates of $O\left(\log^2(n)/n^2\right)$ are at most achievable. To our knowledge, our analysis also provides the tightest high-probability bounds for gradient-based generalization gaps in nonconvex settings.



Paperid:2515
Authors:Indrashis Das, Mahmoud Safari, Steven Adriaensen, Frank Hutter
Abstract:
Abstract:Activation functions are fundamental elements of deep learning architectures as they significantly influence training dynamics. ReLU, while widely used, is prone to the dying neuron problem, which has been mitigated by variants such as LeakyReLU, PReLU, and ELU that better handle negative neuron outputs. Recently, self-gated activations like GELU and Swish have emerged as state-of-the-art alternatives, leveraging their smoothness to ensure stable gradient flow and prevent neuron inactivity. In this work, we introduce the Gompertz Linear Unit (GoLU), a novel self-gated activation function defined as $\mathrm{GoLU}(x) = x \\, \mathrm{Gompertz}(x)$, where $\mathrm{Gompertz}(x) = e^{-e^{-x}}$. The GoLU activation leverages the right-skewed asymmetry in the Gompertz function to reduce variance in the latent space more effectively compared to GELU and Swish, while preserving robust gradient flow. Extensive experiments across diverse tasks, including Image Classification, Language Modeling, Semantic Segmentation, Object Detection, Instance Segmentation, and Diffusion, highlight GoLU's superior performance relative to state-of-the-art activation functions, establishing GoLU as a robust alternative to existing activation functions.



Paperid:2516
Authors:Viacheslav Surkov, Chris Wendler, Antonio Mari, Mikhail Terekhov, Justin Deschenaux, Robert West, Caglar Gulcehre, David Bau
Abstract:
Sparse autoencoders (SAEs) have become a core ingredient in the reverse engineering of large-language models (LLMs). For LLMs, they have been shown to decompose intermediate representations that often are not interpretable directly into sparse sums of interpretable features, facilitating better control and subsequent analysis. However, similar analyses and approaches have been lacking for text-to-image models. We investigated the possibility of using SAEs to learn interpretable features for SDXL Turbo, a few-step text-to-image diffusion model. To this end, we train SAEs on the updates performed by transformer blocks within SDXL Turbo's denoising U-net in its 1-step setting. However, we find that they generalize to 4-step SDXL Turbo's and even to the 20-step SDXL base model (i.e., a different model) without additional training. In addition, we show that their learned features are interpretable, causally influence the generation process, and reveal specialization among the blocks. We do so by adapting PIEBench, a prompt based image editing benchmark, to our setting of editing images while they are generated by turning on and off individual SAE features. Our work is the first investigation of SAEs for interpretability in text-to-image diffusion models and our results establish SAEs as a promising approach for understanding and manipulating the internal mechanisms of text-to-image models.



Paperid:2517
Authors:Eduardo Adame, Daniel Csillag, Guilherme Tegoni Goedert
Abstract:
The increasing use of generative ML foundation models for image restoration tasks such as super-resolution calls for robust and interpretable uncertainty quantification methods. We address this need by presenting a novel approach based on conformal prediction techniques to create a `confidence mask' capable of reliably and intuitively communicating where the generated image can be trusted. Our method is adaptable to any black-box generative model, including those locked behind an opaque API, requires only easily attainable data for calibration, and is highly customizable via the choice of a local image similarity metric. We prove strong theoretical guarantees for our method that span fidelity error control (according to our local image similarity metric), reconstruction quality, and robustness in the face of data leakage. Finally, we empirically evaluate these results and establish our method's solid performance.



Paperid:2518
Authors:Jing Wang, Ao Ma, Ke Cao, Jun Zheng, Jiasong Feng, Zhanjie Zhang, Wanyuan Pang, Xiaodan Liang
Abstract:
Recent advances in text-to-video (T2V) generation, exemplified by models such as Sora and Kling, have demonstrated strong potential for constructing world simulators. However, existing T2V models still struggle to understand abstract physical principles and to generate videos that faithfully obey physical laws. This limitation stems primarily from the lack of explicit physical guidance, caused by a significant gap between high-level physical concepts and the generative capabilities of current models. To address this challenge, we propose theWorldSimulatorAssistant (WISA), a novel framework designed to systematically decompose and integrate physical principles into T2V models. Specifically, WISA decomposes physical knowledge into three hierarchical levels: textual physical descriptions, qualitative physical categories, and quantitative physical properties. It then incorporates several carefully designed modules—such as Mixture-of-Physical-Experts Attention (MoPA) and a Physical Classifier—to effectively encode these attributes and enhance the model’s adherence to physical laws during generation. In addition, most existing video datasets feature only weak or implicit representations of physical phenomena, limiting their utility for learning explicit physical principles. To bridge this gap, we presentWISA-80K, a new dataset comprising 80,000 human-curated videos that depict 17 fundamental physical laws across three core domains of physics: dynamics, thermodynamics, and optics. Experimental results show that WISA substantially improves the alignment of T2V models (such as CogVideoX and Wan2.1) with real-world physical laws, achieving notable gains on the VideoPhy benchmark. Our data, code, and models will be open source.



Paperid:2519
Authors:Yuan Yuan, Chonghua Han, Jingtao Ding, Guozhen Zhang, Depeng Jin, Yong Li
Abstract:
The urban environment is characterized by complex spatio-temporal dynamics arising from diverse human activities and interactions. Effectively modeling these dynamics is essential for understanding and optimizing urban systems. In this work, we introduce UrbanDiT, a foundation model for open-world urban spatio-temporal learning that successfully scales up diffusion transformers in this field. UrbanDiT pioneers a unified model that integrates diverse data sources and types while learning universal spatio-temporal patterns across different cities and scenarios. This allows the model to unify both multi-data and multi-task learning, and effectively support a wide range of spatio-temporal applications. Its key innovation lies in the elaborated prompt learning framework, which adaptively generates both data-driven and task-specific prompts, guiding the model to deliver superior performance across various urban applications. UrbanDiT offers three advantages: 1) It unifies diverse data types, such as grid-based and graph-based data, into a sequential format; 2) With task-specific prompts, it supports a wide range of tasks, including bi-directional spatio-temporal prediction, temporal interpolation, spatial extrapolation, and spatio-temporal imputation; and 3) It generalizes effectively to open-world scenarios, with its powerful zero-shot capabilities outperforming nearly all baselines with training data. UrbanDiT sets up a new benchmark for foundation models in the urban spatio-temporal domain. Code and datasets are publicly available at https://anonymous.4open.science/r/UrbanDiT.



Paperid:2520
Authors:Chaerin Kong, Jiho Jang, Nojun Kwak
Abstract:
Differential Transformer has recently gained significant attention for its impressive empirical performance, often attributed to its ability to perform noise canceled attention. However, precisely how differential attention achieves its empirical benefits remains poorly understood. Moreover, Differential Transformer architecture demands large-scale training from scratch, hindering utilization of open pretrained weights. In this work, we conduct an in-depth investigation of Differential Transformer, uncovering three key factors behind its success: (1) enhanced expressivity via negative attention, (2) reduced redundancy among attention heads, and (3) improved learning dynamics. Based on these findings, we propose DEX, a novel method to efficiently integrate the advantages of differential attention into pretrained language models. By reusing the softmax attention scores and adding a lightweight differential operation on the output value matrix, DEX effectively incorporates the key advantages of differential attention while remaining lightweight in both training and inference. Evaluations confirm that DEX substantially improves the pretrained LLMs across diverse benchmarks, achieving significant performance gains with minimal adaptation data (< 0.01%).



Paperid:2521
Authors:Ziwei Li, Bo Sun, Zhiqiu Zhang, Mohammad Hajiesmaili, Binghan Wu, Lin Yang, Yang Gao
Abstract:
We introduce and study the Combinatorial Ski Rental (CSR) problem, which involves multiple items that can be rented or purchased, either individually or in combination. At each time step, a decision-maker must make an irrevocable buy-or-rent decision for items that have not yet been purchased, without knowing the end of the time horizon. We propose a randomized online algorithm, Sorted Optimal Amortized Cost (SOAC), that achieves the optimal competitive ratio. Moreover, SOAC can be extended to address various well-known ski rental variants, including the multi-slope, multi-shop, multi-commodity ski rental and CSR with upgrading problems. Building on the proposed SOAC algorithm, we further develop a learning-augmented algorithm that leverages machine-learned predictions to improve the performance of CSR. This algorithm is capable of recovering or improving upon existing results of learning-augmented algorithms in both the classic ski rental and multi-shop ski rental problems. Experimental results validate our theoretical analysis and demonstrate the advantages of our algorithms over baseline methods for ski rental problems.



Paperid:2522
Authors:Yunpeng Bai, Shaoheng Fang, Chaohui Yu, Fan Wang, Qixing Huang
Abstract:
Recent advances in video generation have enabled the synthesis of high-quality and visually realistic clips using diffusion transformer models. However, most existing approaches operate purely in the 2D pixel space and lack explicit mechanisms for modeling 3D structures, often resulting in temporally inconsistent geometries, implausible motions, and structural artifacts. In this work, we introduce geometric regularization losses into video generation by augmenting latent diffusion models with per-frame depth prediction. We adopted depth as the geometric representation because of the great progress in depth prediction and its compatibility with image-based latent encoders. Specifically, to enforce structural consistency over time, we propose a multi-view geometric loss that aligns the predicted depth maps across frames within a shared 3D coordinate system. Our method bridges the gap between appearance generation and 3D structure modeling, leading to improved spatio-temporal coherence, shape consistency, and physical plausibility. Experiments across multiple datasets show that our approach produces significantly more stable and geometrically consistent results than existing baselines.



Paperid:2523
Authors:Tzu-Ching Yen, Andrew Siah, Haozhe Chen, C. Guetta, Tianyi Peng, Hongseok Namkoong
Abstract:
Careful curation of data sources can significantly improve the performance of LLM pre-training, but predominant approaches rely heavily on intuition or costly trial-and-error, making them difficult to generalize across different data domains and downstream tasks. Although scaling laws can provide a principled and general approach for data curation, standard deterministic extrapolation from small-scale experiments to larger scales requires strong assumptions on the reliability of such extrapolation, whose brittleness has been highlighted in prior works. In this paper, we introduce a probabilistic extrapolation framework for data mixture optimization that avoids rigid assumptions and explicitly models the uncertainty in performance across decision variables. We formulate data curation as a sequential decision-making problem–multi-fidelity, multi-scale Bayesian optimization–where {data mixtures, model scale, training steps} are adaptively selected to balance training cost and potential information gain. Our framework naturally gives rise to algorithm prototypes that leverage noisy information from inexpensive experiments to systematically inform costly training decisions. To accelerate methodological progress, we build a simulator based on 472 language model pre-training runs with varying data compositions from the SlimPajama dataset. We observe that even simple kernels and acquisition functions can enable principled decisions across training models from 20M to 1B parameters and achieve 2.6x and 3.3x speedups compared to multi-fidelity BO and random search baselines. Taken together, our framework underscores potential efficiency gains achievable by developing principled and transferable data mixture optimization methods. Our code is publicly available at https://github.com/anonWAEWA/ADSO.



Paperid:2524
Authors:Shuo Xu, Yu Chen, Shuxia Lin, Xin Geng, Xu Yang
Abstract:
The Transformer architecture serves as the foundation of modern AI systems, powering recent advances in Large Language Models (LLMs) and Large Multimodal Models (LMMs). Central to these models, attention mechanisms capture contextual dependencies via token interactions. Beyond inference, attention has been widely adopted for interpretability, offering insights into model behavior. Among interpretability techniques, attention flow --- which traces global information transfer across layers --- provides a more comprehensive perspective than single-layer attention maps. However, computing attention flow is computationally intensive due to the high complexity of max-flow algorithms. To address this challenge, we introduce FlowPrune, a novel framework that accelerates attention flow analysis by pruning the attention graph before applying max-flow computations. FlowPrune uses the Max-Flow Min-Cut Theorem and two structural properties of Transformer to identify and eliminate non-critical graph regions. It comprises two components: Edge Pruning, which removes insignificant attention edges, and Layer Compression, which discards layers with minimal contributions to the flow. We conduct extensive experiments on LLaMA and LLaVA to evaluate the robustness and effectiveness of FlowPrune. Our results show that FlowPrune achieves high agreement with the original attention flow in both absolute and relative error metrics, as well as in identifying influential input tokens. Finally, case studies in both NLP and vision domains demonstrate that FlowPrune produces consistent interpretability outcomes as the original Attention Flow, validating its practical utility. The code for this paper is publicly available at this site (https://anonymous.4open.science/r/FlowPrune-90AF).



Paperid:2525
Authors:Yirong Shen, Lu GAN, Cong Ling
Abstract:
Likelihood-based generative models have found wide applications such as compression, semi-supervised learning, and other real-world scenarios. At the heart of many high-performing likelihood models lie perturbation-based approaches. However, likelihood estimation in such models often suffers from slow convergence and limited theoretical insight. In this paper, we derive a novel likelihood bound for noise-driven models to improve both the accuracy and efficiency of maximum likelihood estimation. Central to our approach is an extension of the Kullback–Leibler (KL) divergence–Fisher information connection to arbitrary noise perturbations, elevating it from a theoretical concept to a practical design principle. This enables researchers to select noise distributions that naturally reflect sensor artifacts, quantization effects, or heavy-tailed data, without sacrificing standard diffusion training methodologies. By treating diffusion mapping as a Gaussian channel, we can formulate an exact expression for the mismatched entropy between the data and the model. Remarkably, our bound achieves on-par likelihood estimation on CIFAR-10 (2.50 bits/dim) and state-of-the-art performance on ImageNet-32 (3.01 bits/dim), all without using any data augmentation. Furthermore, our theoretical formulation may also offer benefits for modelling discrete data distributions, suggesting broader applicability beyond continuous domains.



Paperid:2526
Authors:Yiyang Gu, Bohan Wu, Qinghua Ran, Rong-Cheng Tu, Xiao Luo, Zhiping Xiao, Wei Ju, Dacheng Tao, Ming Zhang
Abstract:
Learning from noisy multi-label supervision remains a central challenge in hashing, where labels are unreliable and outputs are discrete. Prior methods often focus on label denoising or prediction regularization, but pay limited attention to the geometric organization of the representation space, which is especially crucial given the structured nature of hash codes. We introduce Semantic Geometry Shaping (SEGA), a framework that learns robust hash functions by explicitly sculpting semantic geometry under noise. SEGA organizes embeddings around dynamic class prototypes that serve as semantic anchors and quantify structural consistency. It estimates uncertainty from both energy and structural divergence, then calibrates label supervision using uncertainty-weighted soft targets. To reinforce continuity in sparse or ambiguous regions, SEGA performs MixUp only between semantically similar instances with contrasting uncertainty levels, generating smooth interpolation paths that expand and regularize class boundaries. This closed-loop process aligns supervision, uncertainty, and representation geometry throughout training. We provide theoretical guarantees showing that SEGA reduces semantic misalignment and improves boundary smoothness. Experiments on noisy multi-label hashing benchmarks confirm state-of-the-art robustness, highlighting the importance of semantic geometry refinement in robust hashing.



Paperid:2527
Authors:Zhenyi Lin, Xiaofan Ming, Qilong Wang, Dongwei Ren, Wangmeng Zuo, Qinghua Hu
Abstract:
Virtual furniture synthesis, a critical task in image composition, aims to seamlessly integrate reference objects into indoor scenes while preserving geometric coherence and visual realism. Despite its significant potential in home design applications, this field remains underexplored due to two major challenges: the absence of publicly available and ready-to-use benchmarks hinders reproducible research, and existing image composition methods fail to meet the stringent fidelity requirements for realistic furniture placement. To address these issues, we introduce RoomBench, a ready-to-use benchmark dataset for virtual furniture synthesis, comprising 7,298 training pairs and 895 testing samples across 27 furniture categories. Then, we propose RoomEditor, a simple yet effective image composition method that employs a parameter-sharing dual U-Net architecture, ensuring better feature consistency by sharing weights between dual branches. Technical analysis reveals that conventional dual-branch architectures generally suffer from inconsistent intermediate features due to independent processing of reference and background images. In contrast, RoomEditor enforces unified feature learning through shared parameters, thereby facilitating model optimization for robust geometric alignment and maintaining visual consistency. Experiments show our RoomEditor is superior to state-of-the-arts, while generalizing directly to diverse objects synthesis in unseen scenes without task-specific fine-tuning. Our dataset and code are available at https://anonymous.4open.science/status/RoomEditor-77B7 for review. They will be made publicly available upon publication.



Paperid:2528
Authors:Jingyang Lin, Jialian Wu, Ximeng Sun, Ze Wang, Jiang Liu, Yusheng Su, Xiaodong Yu, Hao Chen, Jiebo Luo, Zicheng Liu, Emad Barsoum
Abstract:
Recent long-form video-language understanding benchmarks have driven progress in video large multimodal models (Video-LMMs).However, the scarcity of well-annotated long videos has left the training of hour-long Video-LLMs underexplored. To close this gap, we present VideoMarathon, a large-scale hour-long video instruction-following dataset. This dataset includes around 9,700 hours of long videos sourced from diverse domains, ranging from 3 to 60 minutes per video. Specifically, it contains 3.3M high-quality QA pairs, spanning six fundamental topics: temporality, spatiality, object, action, scene, and event. Compared to existing video instruction datasets, VideoMarathon significantly extends training video durations up to 1 hour, and supports 22 diverse tasks requiring both short- and long-term video comprehension. Building on VideoMrathon, we further propose Hour-LLaVA, a powerful and efficient Video-LMM for hour-scale video-language modeling. Hour-LLaVA enables more effective hour-long video training and inference at 1-FPS sampling by leveraging a memory augmentation (MemAug) module, which adaptively integrates user question-relevant and spatiotemporal-informative semantics from a cached full video context. Empirically, Hour-LLaVA achieves the best performance on multiple long video-language benchmarks, demonstrating the superiority of both the VideoMarathon dataset and the Hour-LLaVA framework.



Paperid:2529
Authors:Zhenyu Zhang, Tianyi Chen, Weiran Xu, Alex Pentland, Jiaxin Pei
Abstract:
While large language models (LLMs) have shown exceptional capabilities in a wide range of tasks, long-horizon tasks that require continuous multi-step reasoning and planning remain challenging. We propose ReCAP (Recursive Context-Aware Reasoning and Planning), a hierarchical framework that integrates recursive subtask decomposition, task execution, and observation-driven plan revision. At the core of ReCAP is a dynamic context tree that tracks the evolving task hierarchy, reasoning trajectory, and environment feedback. This structured reasoning process allows the system to recursively generate and execute subtasks and to return to higher-level goals upon subtask completion or failure. Experimental results show that ReCAP significantly improves success rates and subgoal alignment across various long-horizon reasoning tasks.



Paperid:2530
Authors:Yikuan Hu, Jifeng Zhu, Lanrui Tang, Chen Huang
Abstract:
Knowledge graphs (KGs), with their structured representation capabilities, offer promising avenue for enhancing Retrieval Augmented Generation (RAG) systems, leading to the development of KG-RAG systems. Nevertheless, existing methods often struggle to achieve effective synergy between system effectiveness and cost efficiency, leading to neither unsatisfying performance nor excessive LLM prompt tokens and inference time. To this end, this paper proposes ReMindRAG, which employs an LLM-guided graph traversal featuring node exploration, node exploitation, and, most notably, memory replay, to improve both system effectiveness and cost efficiency. Specifically, ReMindRAG memorizes traversal experience within KG edge embeddings, mirroring the way LLMs "memorize" world knowledge within their parameters, but in a train-free manner. We theoretically and experimentally confirm the effectiveness of ReMindRAG, demonstrating its superiority over existing baselines across various benchmark datasets and LLM backbones. Our code is available at https://anonymous.4open.science/r/ReMindRAG-3CE6.



Paperid:2531
Authors:Yihan Wang, Yiwei Lu, Xiao-Shan Gao, Gautam Kamath, Yaoliang Yu
Abstract:
Availability attacks, or unlearnable examples, are defensive techniques that allow data owners to modify their datasets in ways that prevent unauthorized machine learning models from learning effectively while maintaining the data's intended functionality. It has led to the release of popular black-box tools (e.g., APIs) for users to upload personal data and receive protected counterparts. In this work, we show that such black-box protections can be substantially compromised if a small set of unprotected in-distribution data is available. Specifically, we propose a novel threat model of protection leakage, where an adversary can (1) easily acquire (unprotected, protected) pairs by querying the black-box protections with a small unprotected dataset; and (2) train a diffusion bridge model to build a mapping between unprotected and protected data. This mapping, termed BridgePure, can effectively remove the protection from any previously unseen data within the same distribution. BridgePure demonstrates superior purification performance on classification and style mimicry tasks, exposing critical vulnerabilities in black-box data protection. We suggest that practitioners implement multi-level countermeasures to mitigate such risks.



Paperid:2532
Authors:Gautam Kamath, Alireza F. Pour, Matthew Regehr, David Woodruff
Abstract:
Abstract:We propose an algorithm with improved query-complexity for the problem of hypothesis selection under local differential privacy constraints. Given a set of $k$ probability distributions $Q$, we describe an algorithm that satisfies local differential privacy, performs $\tilde{O}(k^{3/2})$ non-adaptive queries to individuals who each have samples from a probability distribution $p$, and outputs a probability distribution from the set $Q$ which is nearly the closest to $p$. Previous algorithms required either $\Omega(k^2)$ queries or many rounds of interactive queries. Technically, we introduce a new object we dub the Scheff\'e graph, which captures structure of the differences between distributions in $Q$, and may be of more broad interest for hypothesis selection tasks.



Paperid:2533
Authors:Bin Pu, Liwen Wang, Xingbo Dong, Xingguo Lv, ZHE JIN
Abstract:
Accurate anatomical structure detection is a critical preliminary step for diagnosing diseases characterized by structural abnormalities. In clinical practice, radiologists frequently adjust the zoom level of medical images to obtain comprehensive views for diagnosis. This common interaction results in significant variations in the apparent scale of anatomical structures across different images or fields of view. However, the information embedded in these zoom-induced scale changes is often overlooked by existing detection algorithms. To overcome this limitation, we propose ZR-DETR, a zoom-aware probabilistic framework tailored for medical object detection. ZR-DETR uniquely incorporates scale-sensitive zoom embeddings, anatomical relation constraints, and a Gaussian Process-based detection head. This architecture enables the framework to jointly model semantic context, enforce anatomical plausibility, and quantify detection uncertainty. Empirical validation across three diverse medical imaging benchmarks (FCS, EPV, and MM-WHS) demonstrates that ZR-DETR consistently outperforms strong baselines in both single-domain and unsupervised domain adaptation scenarios.



Paperid:2534
Authors:Christopher Yeh, Nicolas Christianson, Adam Wierman, Yisong Yue
Abstract:
While deep learning models often achieve high predictive accuracy, their predictions typically do not come with any provable guarantees on risk or reliability, which are critical for deployment in high-stakes applications. The framework of conformal risk control (CRC) provides a distribution-free, finite-sample method for controlling the expected value of any bounded monotone loss function and can be conveniently applied post-hoc to any pre-trained deep learning model. However, many real-world applications are sensitive to tail risks, as opposed to just expected loss. In this work, we develop a method for controlling the general class of Optimized Certainty-Equivalent (OCE) risks, a broad class of risk measures that includes as special cases the expected loss (generalizing the original CRC method) and common tail risks like the conditional value-at-risk (CVaR). Furthermore, standard post-hoc CRC can degrade average-case performance due to its lack of feedback to the model. To address this, we introduce "Conformal Risk Training," an end-to-end approach that differentiates through conformal OCE risk controlduring model training or fine-tuning. Our method achieves provable risk guarantees while demonstrating significantly improved average-case performance over post-hoc approaches on applications to controlling classifiers' false negative rate and controlling financial risk in battery storage operation.



Paperid:2535
Authors:Andrei Manolache, Luiz Chamon, Mathias Niepert
Abstract:
Equivariant neural networks are designed to respect symmetries through their architecture, boosting generalization and sample efficiency when those symmetries are present in the data distribution. Real-world data, however, often departs from perfect symmetry because of noise, structural variation, measurement bias, or other symmetry-breaking effects. Strictly equivariant models may struggle to fit the data, while unconstrained models lack a principled way to leverage partial symmetries. Even when the data is fully symmetric, enforcing equivariance can hurt training by limiting the model to a restricted region of the parameter space. Guided by homotopy principles, where an optimization problem is solved by gradually transforming a simpler problem into a complex one, we introduce Adaptive Constrained Equivariance (ACE), a constrained optimization approach that starts with a flexible, non-equivariant model and gradually reduces its deviation from equivariance. This gradual tightening smooths training early on and settles the model at a data-driven equilibrium, balancing between equivariance and non-equivariance. Across multiple architectures and tasks, our method consistently improves performance metrics, sample efficiency, and robustness to input perturbations compared with strictly equivariant models and heuristic equivariance relaxations.



Paperid:2536
Authors:Cameron Tice, Philipp Kreer, Nathan Helm-Burger, Prithviraj Shahani, Fedor Ryzhenkov, Fabien Roger, Clement Neo, Jacob Haimes, Felix Hofstätter, Teun van der Weij
Abstract:
Capability evaluations play a crucial role in assessing and regulating frontier AI systems. The effectiveness of these evaluations faces a significant challenge: strategic underperformance, or ``sandbagging'', where models deliberately underperform during evaluation. Sandbagging can manifest either through explicit developer intervention or through unintended model behavior, presenting a fundamental obstacle to accurate capability assessment. We introduce a novel sandbagging detection method based on injecting noise of varying magnitudes into model weights. While non-sandbagging models show predictable performance degradation with increasing noise, we demonstrate that sandbagging models exhibit anomalous performance improvements, likely due to disruption of underperformance mechanisms while core capabilities remain partially intact. Through experiments across various model architectures, sizes, and sandbagging techniques, we establish this distinctive response pattern as a reliable, model-agnostic signal for detecting sandbagging behavior. Importantly, we find noise-injection is capable of eliciting the full performance of Mistral Large 120B in a setting where the model underperforms without being instructed to do so. Our findings provide a practical tool for AI evaluation and oversight, addressing a challenge in ensuring accurate capability assessment of frontier AI systems.



Paperid:2537
Authors:Sungmin Woo, Sangyoun Lee
Abstract:
Depth-from-Focus (DFF) enables precise depth estimation by analyzing focus cues across a stack of images captured at varying focal lengths. While recent learning-based approaches have advanced this field, they often struggle in complex scenes with fine textures or abrupt depth changes, where focus cues may become ambiguous or misleading. We present DualFocus, a novel DFF framework that leverages the focal stack’s unique gradient patterns induced by focus variation, jointly modeling focus changes over spatial and focal dimensions. Our approach introduces a variational formulation with dual constraints tailored to DFF: spatial constraints exploit gradient pattern changes across focus levels to distinguish true depth edges from texture artifacts, while focal constraints enforce unimodal, monotonic focus probabilities aligned with physical focus behavior. These inductive biases improve robustness and accuracy in challenging regions. Comprehensive experiments on four public datasets demonstrate that DualFocus consistently outperforms state-of-the-art methods in both depth accuracy and perceptual quality. Code and pretrained models will be made publicly available.



Paperid:2538
Authors:Sicheng Zuo, Wenzhao Zheng, Xiaoyong Han, Longchao Yang, Yong Pan, Jiwen Lu
Abstract:
3D occupancy prediction is crucial for robust autonomous driving systems due to its comprehensive perception of environmental structures and semantics. Most existing methods employ dense grid-based scene representations, which ignore the inherent sparsity of driving scenes and suffer from low efficiency. Recent works explore object-centric representations based on sparse Gaussians, but the ellipsoidal shape prior of Gaussians limits their ability to model diverse structures. In real-world driving scenes, objects exhibit rich geometries (e.g., planes, cuboids, and irregular shapes), requiring excessive overlapping ellipsoidal Gaussians for accurate representation, which leads to inefficient scene representation. To address this, we propose to use geometrically expressive superquadrics as scene representation primitives instead of Gaussians, which can efficiently represent complex structures without much overlap. We develop a probabilistic superquadric mixture model, which interprets each superquadric as an occupancy probability distribution of its neighborhood with corresponding geometry, and calculates semantics through probabilistic mixture. We then employ a superquadric-based model (QuadricFormer) for efficient 3D occupancy prediction, and design a pruning-and-splitting module to further improve modeling efficiency by concentrating superquadrics in occupied regions. Extensive experiments on the nuScenes dataset demonstrate that QuadricFormer achieves state-of-the-art performance while maintaining superior efficiency.



Paperid:2539
Authors:Bingchen Zhao, Despoina Magka, Minqi Jiang, Xian Li, Roberta Raileanu, Tatiana Shavrina, Jean-Christophe Gagnon-Audet, Kelvin Niu, Shagun Sodhani, Michael Shvartsman, Andrei Lupu, Alisia Lupidi, Karen Hambardzumyan, Martin Josifoski, Edan Toledo, Thomas Foster, Lucia Cipolina Kun, Derek Dunfield, Abhishek Charnalia, Oisin Mac Aodha, Alexander Miller, Jakob Foerster, Yoram Bachrach
Abstract:
Abstract:Rapidly improving large language models (LLMs) have the potential to assist in scientific progress. One critical skill in this endeavor is the ability to faithfully reproduce existing work. To evaluate the capability of AI agents to reproduce complex code in an active research area, we introduce the Automated LLM Speedrunning Benchmark, leveraging the research community's contributions to the $\textit{NanoGPT speedrun}$, a competition to train a GPT-2 model in the shortest time. Each of the 19 speedrun tasks provides the agent with the previous record's training script, optionally paired with one of three hint formats, ranging from pseudocode to paper-like descriptions of the new record's improvements. Records execute quickly by design and speedrun improvements encompass diverse code-level changes, ranging from high-level algorithmic advancements to hardware-aware optimizations. These features make the benchmark both accessible and realistic for the frontier problem of improving LLM training. We find that recent frontier reasoning LLMs combined with SoTA scaffolds struggle to reimplement already-known innovations in our benchmark, even when given detailed hints. Our benchmark thus provides a simple, non-saturated measure of an LLM's ability to automate scientific reproduction, a necessary (but not sufficient) skill for an autonomous research agent.



Paperid:2540
Authors:Shuyao Shang, Yuntao Chen, Yuqi Wang, Yingyan Li, ZHAO-XIANG ZHANG
Abstract:
End-to-end autonomous driving has substantially progressed by directly predicting future trajectories from raw perception inputs, which bypasses traditional modular pipelines. However, mainstream methods trained via imitation learning suffer from critical safety limitations, as they fail to distinguish between trajectories that appear human-like but are potentially unsafe. Some recent approaches attempt to address this by regressing multiple rule-driven scores but decoupling supervision from policy optimization, resulting in suboptimal performance. To tackle these challenges, we propose DriveDPO, a Safety Direct Preference Optimization Policy Learning framework. First, we distill a unified policy distribution from human imitation similarity and rule-based safety scores for direct policy optimization. Further, we introduce an iterative Direct Preference Optimization stage formulated as trajectory-level preference alignment. Extensive experiments on the NAVSIM benchmark demonstrate that DriveDPO achieves a new state-of-the-art PDMS of 90.0. Furthermore, qualitative results across diverse challenging scenarios highlight DriveDPO’s ability to produce safer and more reliable driving behaviors.



Paperid:2541
Authors:Mingwei Zhan, Ruijie Zhao, Xianwen Deng, Zhi Xue, Qi Li, Zhuotao Liu, Guang Cheng, Ke Xu
Abstract:
Network traffic classification is essential for network management and security. In recent years, deep learning (DL) algorithms have emerged as essential tools for classifying complex traffic. However, they rely heavily on high-quality labeled training data. In practice, traffic data is often noisy due to human error or inaccurate automated labeling, which could render classification unreliable and lead to severe consequences. Although some studies have alleviated the label noise issue in specific scenarios, they are difficult to generalize to general traffic classification tasks due to the inherent semantic complexity of traffic data. In this paper, we propose FlowRefiner, a robust and general traffic classification framework against label noise. FlowRefiner consists of three core components: a traffic semantics-driven noise detector, a confidence-guided label correction mechanism, and a cross-granularity robust classifier. First, the noise detector utilizes traffic semantics extracted from a pre-trained encoder to identify mislabeled flows. Next, the confidence-guided label correction module fine-tunes a label predictor to correct noisy labels and construct refined flows. Finally, the cross-granularity robust classifier learns generalized patterns of both flow-level and packet-level, improving classification robustness against noisy labels. We evaluate our method on four traffic datasets with various classification scenarios across varying noise ratios. Experimental results demonstrate that FlowRefiner mitigates the impact of label noise and consistently outperforms state-of-the-art baselines by a large margin. We will release our codes and data upon publication.



Paperid:2542
Authors:Roberto Castro, Andrei Panferov, Rush Tabesh, Oliver Sieberling, Jiale Chen, Mahdi Nikdan, Saleh Ashkboos, Dan Alistarh
Abstract:
The rapid advancement of large language models (LLMs) has been paralleled by unprecedented increases in computational demands, with training costs doubling every few months. Training models directly in low-precision arithmetic offers a solution, by improving both computational throughput and energy efficiency. Specifically, NVIDIA's recent Blackwell architecture facilitates extremely low-precision operations, specifically FP4, promising substantial efficiency gains. Yet, current algorithms for training LLMs in FP4 precision face significant accuracy degradation and often rely on mixed-precision fallbacks. In this paper, we systematically investigate hardware-supported FP4 training and introduce Quartet, a new approach enabling accurate, end-to-end FP4 training with all the major computations (in e.g. linear layers) being performed in low precision. Through extensive evaluations on Llama-type models, we reveal a new low-precision scaling law that quantifies performance trade-offs across varying bit-widths. Implementing optimized CUDA kernels tailored for NVIDIA Blackwell GPUs, Quartet achieves state-of-the-art accuracy, successfully training billion-scale models. Our method sets a new benchmark in efficiency-vs-accuracy, demonstrating that fully FP4-based training is a competitive alternative to standard-precision training.



Paperid:2543
Authors:Bohan Zhou, Yi Zhan, Zhongbin Zhang, Zongqing Lu
Abstract:
Egocentric hand-object motion generation is crucial for immersive AR/VR and robotic imitation but remains challenging due to unstable viewpoints, self-occlusions, perspective distortion, and noisy ego-motion. Existing methods rely on predefined 3D object priors, limiting generalization to novel objects, which restricts their generalizability to novel objects. Meanwhile, recent multimodal approaches suffer from ambiguous generation from abstract textual cues, intricate pipelines for modeling 3D hand-object correlation, and compounding errors in open-loop prediction. We proposeMEgoHand, a multimodal framework that synthesizes physically plausible hand-object interactions from egocentric RGB, text, and initial hand pose. MEgoHand introduces a bi-level architecture: a high-level “cerebrum” leverages a vision language model (VLM) to infer motion priors from visual-textual context and a monocular depth estimator for object-agnostic spatial reasoning, while a low-level DiT-based flow-matching policy generates fine-grained trajectories with temporal orthogonal filtering to enhance stability. To address dataset inconsistency, we design a dataset curation paradigm with an Inverse MANO Retargeting Network and Virtual RGB-D Renderer, curating a unified dataset of3.35MRGB-D frames,24Kinteractions, and1.2Kobjects. Extensive experiments acrossfivein-domain andtwocross-domain datasets demonstrate the effectiveness of MEgoHand, achieving substantial reductions in wrist translation error (86.9%) and joint rotation error (34.1%), highlighting its capacity to accurately model fine-grained hand joint structures and generalize robustly across diverse scenarios.



Paperid:2544
Authors:Yanyuan Qiao, Haodong Hong, Wenqi Lyu, Dong An, Siqi Zhang, Yutong Xie, Xinyu Wang, Qi Wu
Abstract:
Multimodal Large Language Models (MLLMs) have demonstrated strong generalization in vision-language tasks, yet their ability to understand and act within embodied environments remains underexplored. We present NavBench, a benchmark to evaluate the embodied navigation capabilities of MLLMs under zero-shot settings. NavBench consists of two components: (1) navigation comprehension, assessed through three cognitively grounded tasks including global instruction alignment, temporal progress estimation, and local observation-action reasoning, covering 3,200 question-answer pairs; and (2) step-by-step execution in 432 episodes across 72 indoor scenes, stratified by spatial, cognitive, and execution complexity. To support real-world deployment, we introduce a pipeline that converts MLLMs' outputs into robotic actions. We evaluate both proprietary and open-source models, finding that GPT-4o performs well across tasks, while lighter open-source models succeed in simpler cases. Results also show that models with higher comprehension scores tend to achieve better execution performance. Providing map-based context improves decision accuracy, especially in medium-difficulty scenarios. However, most models struggle with temporal understanding, particularly in estimating progress during navigation, which may pose a key challenge.



Paperid:2545
Authors:Henry Arguello, Roman Jacome, Romario Gualdrón-Hurtado, León Suárez-Rodríguez
Abstract:
Imaging inverse problems aim to recover high-dimensional signals from undersampled, noisy measurements, a fundamentally ill-posed task with infinite solutions lying in the null space of the sensing operator. To resolve this ambiguity, prior information is typically incorporated through handcrafted regularizers or learned models that constrain the solution space. However, these priors typically ignore the task-specific structure of that null-space. In this work, we propose Non-linear Projections of the Null Space (NPN), a novel class of regularization that, instead of enforcing structural constraints in the image domain, promotes solutions that lie in a low-dimensional projection of the sensing matrix's null space with a neural network.Our approach has two key advantages: (1) Interpretability: by focusing on the structure of the null space, we design sensing-matrix-specific priors that capture information orthogonal to the signal components that are fundamentally blind to the sensing process. (2) Flexibility: NPN is adaptable to various inverse problems, compatible with existing reconstruction frameworks, and complementary to conventional image-domain priors. We provide theoretical guarantees on convergence and reconstruction accuracy when used within plug-and-play methods. Empirical results across diverse sensing matrices demonstrate that NPN priors consistently enhance reconstruction fidelity in various imaging inverse problems, such as compressive sensing, deblurring, and magnetic resonance imaging, with plug-and-play methods and unrolling networks.



Paperid:2546
Authors:Guiqiu Liao, Matjaz Jogan, Eric Eaton, Daniel Hashimoto
Abstract:
Learning efficient visual representations across heterogeneous unlabeled datasets remains a central challenge in federated learning. Effective federated representations require features that are jointly informative across clients while disentangling domain-specific factors without supervision. We introduce FORLA, a novel framework for federated object-centric representation learning and feature adaptation across clients using unsupervised slot attention. At the core of our method is a shared feature adapter, trained collaboratively across clients to adapt features from foundation models, and a shared slot attention module that learns to reconstruct the adapted features. To optimize this adapter, we design a two-branch student–teacher architecture. In each client, a student decoder learns to reconstruct full features from foundation models, while a teacher decoder reconstructs their adapted, low-dimensional counterpart. The shared slot attention module bridges cross-domain learning by aligning object-level representations across clients. Experiments in multiple real-world datasets show that our framework not only outperforms centralized baselines on object discovery but also learns a compact, universal representation that generalizes well across domains. This work highlights federated slot attention as an effective tool for scalable, unsupervised visual representation learning from cross-domain data with distributed concepts.



Paperid:2547
Authors:Xiaokang Ye, Jiawei Ren, Yan Zhuang, Xuhong He, Yiming Liang, Yiqing Yang, Mrinaal Dogra, Xianrui Zhong, Eric Liu, Kevin Benavente, Rajiv Mandya Nagaraju, Dhruv Sharma, Ziqiao Ma, Tianmin Shu, Zhiting Hu, Lianhui Qin
Abstract:
While LLM/VLM-powered AI agents have advanced rapidly in math, coding, and computer use, their applications in complex physical and social environments remain challenging. Building agents that can survive and thrive in the real world (e.g., by autonomously earning income) requires massive-scale interaction, reasoning, training, and evaluation across diverse scenarios. However, existing world simulators for such development fall short: they often rely on limited hand-crafted environments, simulate simplified game-like physics and social rules, and lack native support for LLM/VLM agents. We introduce SimWorld, a new simulator built on Unreal Engine 5, designed for developing and evaluating LLM/VLM agents in rich, real-world-like settings. SimWorld offers three core capabilities: (1) realistic, open-ended world simulation, including accurate physical and social dynamics and language-driven procedural environment generation; (2) rich interface for LLM/VLM agents, with multi-modal world inputs/feedback and open-vocabulary action outputs at varying levels of abstraction; and (3) diverse physical and social reasoning scenarios that are easily customizable by users. We demonstrate SimWorld by deploying frontier LLM agents (e.g., Gemini-2.5-Flash, Claude-3.5, GPT-4o, and DeepSeek-Prover-V2) on both short-horizon navigation tasks requiring grounded re-planning, and long-horizon multi-agent food delivery tasks involving strategic cooperation and competition. The results reveal distinct reasoning patterns and limitations across models. We will open-source SimWorld and hope it becomes a foundational platform for advancing real-world agent intelligence across disciplines.



Paperid:2548
Authors:Jiayi Fan, Jingyuan Yang, Shuangge Ma, Mengyun Wu
Abstract:
Accurate network estimation serves as the cornerstone for understanding complex systems across scientific domains, spanning from decoding gene regulatory pathways in systems biology to identifying social interaction patterns in computational sociology. Modern applications demand methods that simultaneously address two critical challenges: capturing nonlinear interdependencies between variables and recovering inherent hierarchical structures where higher-level entities coordinate lower-level components (e.g., functional pathways organizing gene clusters). Traditional Gaussian graphical models fundamentally fail in these aspects due to their restrictive linear assumptions and flat network representations. We propose DeepBLNet, a deep learning framework for bi-level network inference. The core innovation lies in hierarchical selection layers that enforce structural consistency between high-level interaction groups and their constituent low-level connections via adaptive sparsity constraints. This architecture is integrated with compositional deep architectures that learn cross-level interaction patterns through constrained nonlinear transformations, explicitly preserving hierarchical dependencies while overcoming the representational limitations of linear methods. Crucially, we establish formal theoretical guarantees for consistent recovery of both high-level interactions and their internal low-level structures under general statistical regimes. Extensive validation demonstrates DeepBLNet's effectiveness across synthetic and real-world scenarios, achieving superior F1 scores compared to competitive methods and particularly benefiting complex systems analysis through its interpretable bilevel structure discovery.



Paperid:2549
Authors:Xi Chen, Kaituo Feng, Changsheng Li, Xunhao Lai, Xiangyu Yue, Ye Yuan, Guoren Wang
Abstract:
Abstract:Low-rank training has emerged as a promising approach for reducing memory usage in training Large Language Models (LLMs). Previous methods either rely on decomposing weight matrices (e.g., LoRA), or seek to decompose gradient matrices (e.g., GaLore) to ensure reduced memory consumption. However, both of them constrain the training in a low-rank subspace, thus inevitably leading to sub-optimal performance. To resolve this, we propose a new plug-and-play training framework for LLMs called Fira, as the first attempt to consistently preserve the low-rank constraint for memory efficiency, while achieving full-rank training (i.e., training with full-rank gradients of full-rank weights) to avoid inferior outcomes. First, we observe an interesting phenomenon during LLM training: the scaling impact of adaptive optimizers (e.g., Adam) on the gradient norm remains similar from low-rank to full-rank training. In light of this, we propose a \textit{norm-based scaling} method, which utilizes the scaling impact of low-rank optimizers as substitutes for that of original full-rank optimizers to achieve this goal.Moreover, we find that there are potential loss spikes during training. To address this, we further put forward a norm-growth limiter to smooth the gradient.Extensive experiments on the pre-training and fine-tuning of LLMs show that Fira outperforms both LoRA and GaLore. Notably, for pre-training LLaMA 7B, our Fira uses $8\times$ smaller memory of optimizer states than Galore, yet outperforms it by a large margin.



Paperid:2550
Authors:YI LI, CHEN YUANLONG, Weiming Huang, Gao Cong, Xiaoli Li
Abstract:
Urban computing harnesses big data to decode complex urban dynamics and revolutionize location-based services. Traditional approaches have treated geospatial prediction tasks (e.g., estimating socio-economic indicators) and retrieval tasks (e.g., querying geographic objects) as isolated challenges, necessitating separate models with distinct training objectives. This fragmentation imposes significant computational burdens and limits cross-task synergy, despite advances in representation learning and multi-task foundation models.We present UrbanSparse, a pioneering framework that unifies geospatial prediction and retrieval through a novel sparse-dense representation architecture. By synergistically combining these tasks, UrbanSparse eliminates redundant systems while amplifying their mutual strengths. Our approach introduces two innovations: (1) Bloom filter-based sparse encodings that compress high-sparsity geographic queries and fine-grained text terms for retrieval effectiveness, and (2) a dense semantic codebook that captures granular urban features to boost prediction accuracy. A two-view contrastive learning mechanism further bridges urban objects, regions, and contexts. Experiments on real-world datasets demonstrate 25.16% gains in prediction accuracy and 20.76% improvements in retrieval precision over state-of-the-art baselines, alongside 65.97% faster training. These advantages position UrbanSparse as a scalable solution for large urban datasets. To our knowledge, this is the first unified framework bridging geospatial prediction and retrieval, opening new frontiers in data-driven urban intelligence.



Paperid:2551
Authors:Jinyuan Liu, Zihang Chen, Zhu Liu, Zhiying Jiang, Long Ma, Xin Fan, Risheng Liu
Abstract:
We engage in the relatively underexplored task named thermal infrared image enhancement. Existing infrared image enhancement methods primarily focus on tackling individual degradations, such as noise, contrast, and blurring, making it difficult to handle coupled degradations. Meanwhile, all-in-one enhancement methods, commonly applied to RGB sensors, often demonstrate limited effectiveness due to the significant differences in imaging models. In sight of this, we first revisit the imaging mechanism and introduce a Recurrent Prompt Fusion Network (RPFN). Specifically, the RPFN initially establishes prompt pairs based on the thermal imaging process. For each type of degradation, we fuse the corresponding prompt pairs to modulate the model's features, providing adaptive guidance that enables the model to better address specific degradations under single or multiple conditions.In addition, a selective recurrent training mechanism is introduced to gradually refine the model's handling of composite cases to align the enhancement process, which not only allows the model to remove camera noise and retain key structural details, but also enhancing the overall contrast of the thermal image. Furthermore, we introduce the most comprehensive high-quality infrared benchmark covering a wide range of scenarios. Extensive experiments substantiate that our approach not only delivers promising visual results under specific degradation but also significantly improves performance on complex degradation scenes, achieving a notable 8.76% improvement.



Paperid:2552
Authors:Evan Markou, Thalaiyasingam Ajanthan, Stephen Gould
Abstract:
Many high-dimensional optimisation problems exhibit rich geometric structures in their set of minimisers, often forming smooth manifolds due to over-parametrisation or symmetries. When this structure is known, at least locally, it can be exploited through reduction mappings that reparametrise part of the parameter space to lie on the solution manifold. These reductions naturally arise from inner optimisation problems and effectively remove redundant directions, yielding a lower-dimensional objective. In this work, we introduce a general framework to understand how such reductions influence the optimisation landscape. We show that well-designed reduction mappings improve curvature properties of the objective, leading to better-conditioned problems and theoretically faster convergence for gradient-based methods. Our analysis unifies a range of scenarios where structural information at optimality is leveraged to accelerate convergence, offering a principled explanation for the empirical gains observed in such optimisation algorithms.



Paperid:2553
Authors:Lubin Bai, Xiuyuan Zhang, Siqi Zhang, Zepeng Zhang, Haoyu Wang, Wei Qin, Shihong Du
Abstract:
Integrating ground-level geospatial data with rich geographic context--such as OpenStreetMap (OSM)--into remote sensing (RS) foundation models (FMs) is essential for advancing geospatial intelligence and supporting a broad spectrum of tasks. However, modality gap between RS and OSM data--differences in data structure, content, and spatial granularity--makes effective synergy highly challenging, and most existing RS FMs focus on imagery alone. To this end, this study presents GeoLink, a multimodal framework that leverages OSM data to enhance RS FM during both the pretraining and downstream task stages. Specifically, GeoLink enhances RS self-supervised pretraining using multi-granularity learning signals derived from OSM data, guided by cross-modal spatial correlations for information interaction and collaboration. It also introduces image mask-reconstruction to enable sparse input for efficient pretraining. For downstream tasks, GeoLink generates both unimodal and multimodal fine-grained encodings to support a wide range of applications, from common RS interpretation tasks like land cover classification to more comprehensive geographic tasks like urban function zone mapping. Extensive experiments show that incorporating OSM data during pretraining enhances the performance of the RS image encoder, while fusing RS and OSM data in downstream tasks improves the FM’s adaptability to complex geographic scenarios. These results underscore the potential of multimodal synergy in advancing high-level geospatial artificial intelligence. Moreover, we find that spatial correlation plays a crucial role in enabling effective multimodal geospatial data integration.



Paperid:2554
Authors:Kyriakos Lotidis, Panayotis Mertikopoulos, Nicholas Bambos, Jose Blanchet
Abstract:
In this paper, we examine the robustness of Nash equilibria in continuous games, under both strategic and dynamic uncertainty.Starting with the former, we introduce the notion of a robust equilibrium as those equilibria that remain invariant to small—but otherwise arbitrary—perturbations to the game's payoff structure, and we provide a crisp geometric characterization thereof.Subsequently, we turn to the question of dynamic robustness, and we examine which equilibria are stable limit points of the dynamics of "follow the regularized leader" (FTRL) in the presence of randomness and uncertainty.Despite their very distinct origins, we establish a structural correspondence between these two notions of robustness: strategic robustness implies dynamic robustness, and, conversely, the requirement of strategic robustness cannot be relaxed if dynamic robustness is to be maintained.Finally, we examine the rate of convergence to robust equilibria as a function of the underlying regularizer, and we show that entropically regularized learning converges at a geometric rate in games with affinely constrained action spaces.



Paperid:2555
Authors:Muquan Yu, Mu Nan, Hossein Adeli, Jacob Prince, John Pyles, Leila Wehbe, Maggie Henderson, Michael Tarr, Andrew Luo
Abstract:
Understanding functional representations within higher visual cortex is a fundamental question in computational neuroscience. While artificial neural networks pretrained on large-scale datasets exhibit striking representational alignment with human neural responses, learning image-computable models of visual cortex relies on individual-level, large-scale fMRI datasets. The necessity for expensive, time-intensive, and often impractical data acquisition limits the generalizability of encoders to new subjects and stimuli.BraInCoRLuses in-context learning to predict voxelwise neural responses from few-shot exampleswithout any additional finetuningfor novel subjects and stimuli. We leverage a transformer architecture that can flexibly condition on a variable number of in-context image stimuli, learning an inductive bias over multiple subjects. During training, we explicitly optimize the model for in-context learning. By jointly conditioning on image features and voxel activations, our model learns to directly generate better performing voxelwise models of higher visual cortex. We demonstrate that BraInCoRL consistently outperforms existing voxelwise encoder designs in a low-data regime when evaluated on entirely novel images, while also exhibiting strong test-time scaling behavior. The model also generalizes to an entirely new visual fMRI dataset, which uses different subjects and fMRI data acquisition parameters. Further, BraInCoRL facilitates better interpretability of neural signals in higher visual cortex by attending to semantically relevant stimuli. Finally, we show that our framework enables interpretable mappings from natural language queries to voxel selectivity.



Paperid:2556
Authors:Kyungjae Lee, Dohyeong Kim, Taehyun Cho, Chaeyeon Kim, Yunkyung Ko, Seungyub Han, Seokhun Ju, Dohyeok Lee, Sungbin Lim
Abstract:
Abstract:This paper addresses the problem of multi-risk agnostic multi-armed bandits in heavy-tailed reward settings. We propose a framework that leverages novel deviation inequalities for the $1$-Wasserstein distance to construct confidence intervals for Lipschitz risk measures. The distributional LCB (DistLCB) algorithm is introduced, which achieves asymptotic optimality by deriving the first lower bounds for risk-aware bandits with explicit sub-optimality gap dependencies.The DistLCB is further extended to multi-risk objectives, which enables Pareto-optimal solutions that consider multiple aspects of reward distributions.Additionally, we provide a regret analysis that includes both gap-dependent and gap-independent bounds for multi-risk settings. Experiments validate the effectiveness of the proposed methods in synthetic and real-world applications.



Paperid:2557
Authors:Pranav Mamidanna, Thomas Klotz, Dimitrios Chalatsis, Agnese Grison, Irene Mendez Guerra, Shihan Ma, Arnault Caillet, Simon Avrillon, Robin Rohlén, Dario Farina
Abstract:
Neural source separation enables the extraction of individual spike trains from complex electrophysiological recordings. When applied to electromyographic (EMG) signals, it provides a unique window into the motor output of the nervous system by isolating the spiking activity of motor units (MUs). This is currently the only scalable neural interfacing approach available in behaving humans and has become foundational in motor neuroscience and neuroprosthetics. However, unlike related domains such as spike sorting or EEG analysis, EMG decomposition lacks open benchmarks that reflect the diversity of muscles, movement contexts, and noise sources encountered in practice.To address this gap, we introduce MUniverse, a modular simulation and benchmarking suite for MU decomposition. MUniverse provides: (1) a simulation stack with a user-friendly interface to a state-of-the-art EMG generator; (2) a curated library of datasets across synthetic, hybrid synthetic-real data with ground truth spikes, and experimental EMG; (3) a set of internal and external decomposition pipelines; and (4) a unified benchmark with well-defined tasks, standard evaluation metrics, and baseline results from established decomposition pipelines.MUniverse is designed for extensibility, reproducibility, and community use, and all datasets are distributed with standardised metadata (Croissant, BIDS). By standardising evaluation and enabling dataset simulation at scale, MUniverse aims to catalyze progress on this long-standing neural signal processing problem.



Paperid:2558
Authors:Gunshi Gupta, Karmesh Yadav, Zsolt Kira, Yarin Gal, Rahaf Aljundi
Abstract:
To enable embodied agents to operate effectively over extended timeframes, it is crucial to develop models that form and access memories to stay contextualized in their environment.In the current paradigm of training transformer-based policies for embodied sequential decision-making tasks, visual inputs often overwhelm the context limits of transformers, while humans can maintain and utilize a lifetime of experience compressed as memories.Significant compression is possible in principle, as much of the input is irrelevant and can be abstracted. However, existing approaches predominantly focus on either recurrent models with fixed-size memory or transformers with full-context reliance.In this work, we propose Memo, a transformer-based architecture and training recipe for reinforcement learning (RL) on memory-intensive, long-horizon tasks. Memo incorporates the creation and retrieval of memory by interleaving periodic summarization tokens with the inputs of a model during training.We demonstrate Memo's effectiveness on a grid-world meta-RL benchmark and a multi-object navigation task in photo-realistic indoor settings. Memo outperforms naive long-context transformer baselines while being more compute and storage efficient. Additionally, Memo generalizes better to longer contexts at inference time and remains robust in streaming settings, where historical context must be truncated to fit inference constraints.



Paperid:2559
Authors:Ahmet Onur Akman, Anastasia Psarou, Michał Hoffmann, Łukasz Gorczyca, Lukasz Kowalski, Paweł Gora, Grzegorz Jamróz, Rafal Kucharski
Abstract:
Abstract:Connected Autonomous Vehicles (CAVs) promise to reduce congestion in future urban networks, potentially by optimizing their routing decisions. Unlike for human drivers, these decisions can be made with collective, data-driven policies, developed by machine learning algorithms. Reinforcement learning (RL) can facilitate the development of such collective routing strategies, yet standardized and realistic benchmarks are missing. To that end, we present $\texttt{URB}$: Urban Routing Benchmark for RL-equipped CAVs. $\texttt{URB}$ is a comprehensive benchmarking environment that unifies evaluation across 29 real-world traffic networks paired with realistic demand patterns. $\texttt{URB}$ comes with a catalog of predefined tasks, four state-of-the-art multi-agent RL (MARL) algorithm implementations, three baseline methods, ten domain-specific performance metrics, and a modular configuration scheme. Our results suggest that, despite the lengthy and costly training, state-of-the-art MARL algorithms rarely outperformed humans. Experimental results reported in this paper initiate the first leaderboard for MARL in large-scale urban routing optimization and reveal that current approaches struggle to scale, emphasizing the urgent need for advancements in this domain.



Paperid:2560
Authors:Hao Li, Xiaogeng Liu, CHIU Chun, Dianqi Li, Ning Zhang, Chaowei Xiao
Abstract:
Large Language Models (LLMs) are increasingly central to agentic systems due to their strong reasoning and planning capabilities. By interacting with external environments through predefined tools, these agents can carry out complex user tasks. Nonetheless, this interaction also introduces the risk of prompt injection attacks, where malicious inputs from external sources can mislead the agent’s behavior, potentially resulting in economic loss, privacy leakage, or system compromise. System-level defenses have recently shown promise by enforcing static or predefined policies, but they still face two key challenges: the ability to dynamically update security rules and the need for memory stream isolation. To address these challenges, we propose DRIFT, a Dynamic Rule-based Isolation Framework for Trustworthy agentic systems, which enforces both control- and data-level constraints. A Secure Planner first constructs a minimal function trajectory and a JSON-Schema-style parameter checklist for each function node based on the user query. A Dynamic Validator then monitors deviations from the original plan, assessing whether changes comply with privilege limitations and the user's intent. Finally, an Injection Isolator detects and masks any instructions that may conflict with the user query from the memory stream to mitigate long-term risks. We empirically validate the effectiveness of DRIFT on the AgentDojo benchmark, demonstrating its strong security performance while maintaining high utility across diverse models—showcasing both its robustness and adaptability.



Paperid:2561
Authors:Lancheng Zou, Shuo Yin, Zehua Pei, Tsung-Yi Ho, Farzan Farnia, Bei Yu
Abstract:
Channel permutation is a powerful technique for enhancing the accuracy of N:M sparse models by reordering the channels of weight matrices to prioritize the retention of important weights. However, traditional channel permutation methods rely on handcrafted quality metrics, which often fail to accurately capture the true impact of pruning on model performance. To address this limitation, we propose PermLLM, a novel post-training pruning framework that introduces learnable channel permutation (LCP) for N:M sparsity. LCP leverages Sinkhorn normalization to transform discrete permutation matrices into differentiable soft permutation matrices, enabling end-to-end optimization. Additionally, PermLLM incorporates an efficient block-wise channel permutation strategy, which significantly reduces the number of learnable parameters and computational complexity. PermLLM seamlessly integrates with existing one-shot pruning methods to adaptively optimize channel permutations, effectively mitigating pruning-induced errors. Extensive experiments on the LLaMA series, Qwen, and OPT models demonstrate that PermLLM achieves superior performance in optimizing N:M sparse models.



Paperid:2562
Authors:Ella Miray Rajaonson, Mahyar Rajabi Kochi, Luis Martin Mejia Mendoza, Mohamad Moosavi, Benjamin Sanchez-Lengeling
Abstract:
Developing improved predictive models for multi-molecular systems is crucial, as nearly every chemical product used results from a mixture of chemicals. While being a vital part of the industry pipeline, the chemical mixture space remains relatively unexplored by the machine learning community. In this paper, we introduce CheMixHub, a holistic benchmark for molecular mixtures, covering a corpus of 11 chemical mixtures property prediction tasks, from drug delivery formulations to battery electrolytes, totalling approximately 500k data points gathered and curated from 7 publicly available datasets. CheMixHub introduces various data splitting techniques to assess context-specific generalization and model robustness, providing a foundation for the development of predictive models for chemical mixture properties. Furthermore, we map out the modelling space of deep learning models for chemical mixtures, establishing initial benchmarks for the community. This dataset has the potential to accelerate chemical mixture development, encompassing reformulation, optimization, and discovery. The dataset and code for the benchmarks can be found at: https://github.com/chemcognition-lab/chemixhub



Paperid:2563
Authors:Ha-Yeong Choi, Sang-Hoon Lee
Abstract:
Diffusion models have demonstrated remarkable generative capabilities, and Conditional Flow Matching (CFM) has improved their inference efficiency by following optimal transport paths. However, CFM-based models still require multiple iterative sampling steps, which makes them unsuitable for real-time or streaming generation scenarios. In this paper, we introduce StreamFlow, a novel streaming generative model designed for real-time audio generation from discrete tokens. StreamFlow leverages a causal noising training framework along the time axis and predicts multi-time vector fields at once on each stream, enabling streaming inference with minimal latency. To further improve generalization, we propose Scale-DiT, a Diffusion Transformer architecture that enhances robustness by modeling, normalizing, and scaling feature differences prior to skip connections. This significantly improves the robustness and performance of DiT without increasing the parameter size. We validate the effectiveness of StreamFlow through audio reconstruction tasks using discrete tokens from EnCodec and Mimi, demonstrating both high-fidelity synthesis and streaming capability. Furthermore, we successfully incorporated our model into fully-duplex streaming speech language models of Moshi by replacing the Mimi decoder.



Paperid:2564
Authors:Corinna Cortes, Mehryar Mohri, Yutao Zhong
Abstract:
Abstract:The *balanced loss* is a widely adopted objective for multi-class classification under class imbalance. By assigning equal importance to all classes, regardless of their frequency, it promotes fairness and ensures that minority classes are not overlooked. However, directly minimizing the balanced classification loss is typically intractable, which makes the design of effective surrogate losses a central question. This paper introduces and studies two advanced surrogate loss families: Generalized Logit-Adjusted (GLA) loss functions and Generalized Class-Aware weighted (GCA) losses. GLA losses generalize Logit-Adjusted losses, which shift logits based on class priors, to the broader general cross-entropy loss family. GCA loss functions extend the standard class-weighted losses, which scale losses inversely by class frequency, by incorporating class-dependent confidence margins and extending them to the general cross-entropy family. We present a comprehensive theoretical analysis of consistency for both loss families. We show that GLA losses are Bayes-consistent, but only $H$-consistent for unbounded and complete hypothesis sets. Moreover, their $H$-consistency bounds depend inversely on the minimum class probability, scaling at least as $1/\mathsf p _{\min}$. In contrast, GCA losses are $H$-consistent for any hypothesis set that is bounded or complete, with $H$-consistency bounds that scale more favorably as $1/\sqrt{\mathsf p _{\min}}$, offering significantly stronger theoretical guarantees in imbalanced settings. We report the results of experiments demonstrating that, empirically, both the GCA losses with calibrated class-dependent confidence margins and GLA losses can greatly outperform straightforward class-weighted losses as well as the LA losses. GLA generally performs slightly better in common benchmarks, whereas GCA exhibits a slight edge in highly imbalanced settings. Thus, we advocate for both GLA and GCA losses as principled, theoretically sound, and state-of-the-art surrogates for balanced classification under class imbalance.



Paperid:2565
Authors:Junqi Jiang, Tom Bewley, Salim I. Amoukou, Francesco Leofante, Antonio Rago, Saumitra Mishra, Francesca Toni
Abstract:
Test-time scaling improves large language models' (LLMs) performance by allocating more compute budget during inference. To achieve this, existing methods often require intricate modifications to prompting and sampling strategies. In this work, we introduce representation consistency (RC), a test-time scaling method for aggregating answers drawn from multiple candidate responses of an LLM regardless of how they were generated, including variations in prompt phrasing and sampling strategy. RC enhances answer aggregation by not only considering the number of occurrences of each answer in the candidate response set, but also the consistency of the model's internal activations while generating the set of responses leading to each answer. These activations can be either dense (raw model activations) or sparse (encoded via pretrained sparse autoencoders). Our rationale is that if the model's representations of multiple responses converging on the same answer are highly variable, this answer is more likely to be the result of incoherent reasoning and should be down-weighted during aggregation. Importantly, our method only uses cached activations and lightweight similarity computations and requires no additional model queries. Through experiments with four open-source LLMs and four reasoning datasets, we validate the effectiveness of RC for improving task performance during inference, with consistent accuracy improvements (up to 4\%) over strong test-time scaling baselines. We also show that consistency in the sparse activation signals aligns well with the common notion of coherent reasoning.



Paperid:2566
Authors:Pankaj Agarwal, Sharath Raghvendra, Pouyan Shirzadian, Keegan Yao
Abstract:
Abstract:The sensitivity of optimal transport (OT) to noise has motivated the study of robust variants. In this paper, we study two such formulations of semi-discrete OT in $\mathbb{R}^d$: (i) the $\alpha$-optimal partial transport, which is the minimum cost required to transport a mass of $\alpha$, and (ii) the $\lambda$-robust optimal transport, which regularizes the OT problem using the total variation (TV) distance. First, we provide a novel characterization of the optimal solutions in these settings, showing they can be represented as a restricted Laguerre diagram. Second, we exploit this characterization to establish a strong algorithmic connection between the two problems, showing that any solver for one can be adapted to solve the other with comparable precision. Third, we overcome key challenges posed in extending the cost-scaling paradigm to compute these variants of OT and present an algorithm that computes the exact solution up to $\log (1/\varepsilon)$ bits of precision in $n^{O(d)}\log (1/\varepsilon)$ time, where $n$ is the support size of the discrete distribution. Finally, we present an $n^{1+o(1)}\varepsilon^{-O(d)}$ time approximation algorithm for the above variants of OT.



Paperid:2567
Authors:Jun Wang, Fuyuan CAO, ZhixinXue, Xingwang Zhao, Jiye Liang
Abstract:
Existing early and late fusion frameworks in multimodal learning are confronted with the fundamental challenge of modality imbalance, wherein disparities in representational capacities induce inter-modal competition during training. Current research methodologies primarily rely on modality-level contribution assessments to measure gaps in representational capabilities and enhance poorly learned modalities, overlooking the dynamic variations of modality contributions across individual samples. To address this, we propose a Causal-aware Modality valuation approach for Balanced multimodal learning (CMoB). we define a benefit function based on Shannon's theory of informational uncertainty to evaluate the changes in the importance of samples across different stages of multimodal training. Inspired by human cognitive science, we propose a causal-aware modality contribution quantification method from a causal perspective to capture fine-grained changes in modality contribution degrees within samples. In the iterative training of multimodal learning, we develop targeted modal enhancement strategies that dynamically select and optimize modalities based on real-time evaluation of their contribution variations across training samples. Our method enhances the discriminative ability of key modalities and the learning capacity of weak modalities while achieving fine-grained balance in multimodal learning. Extensive experiments on benchmark multimodal datasets and multimodal frameworks demonstrate the superiority of our CMoB approach for balanced multimodal learning.



Paperid:2568
Authors:Shuqiao Liang, Jian Liu, Chen Renzhang, Quanlong Guan
Abstract:
The increasing realism of synthetic images generated by advanced models such as VAEs, GANs, and LDMs poses significant challenges for synthetic image detection.To address this issue, we explore two artifact types introduced during the generation process: (1) latent distribution deviations and (2) decoding-induced smoothing effects, which manifest as inconsistencies in local textures, edges, and color transitions.Leveraging local pixel dependencies (LPD) properties rooted in Markov Random Fields, we reconstruct synthetic images using neighboring pixel information to expose disruptions in texture continuity and edge coherence.Building upon LPD, we propose FerretNet, a lightweight neural network with only 1.1M parameters that delivers efficient and robust synthetic image detection.Extensive experiments demonstrate that FerretNet—trained exclusively on the 4-class ProGAN dataset—achieves an average accuracy of 97.1% on an open-world benchmark comprising over 20 generative models, surpassing state-of-the-art methods by 10.6%. All code and datasets will be publicly released.



Paperid:2569
Authors:Yang Chen, Zhuolin Yang, Zihan Liu, Chankyu Lee, Peng Xu, Mohammad Shoeybi, Bryan Catanzaro, Wei Ping
Abstract:
Despite recent progress in large-scale reinforcement learning (RL) for reasoning, the training recipe for building high-performing reasoning models remains elusive. Key implementation details of frontier models, such as DeepSeek-R1, including data curation strategies and RL training recipe, are often omitted. Moreover, recent research indicates distillation remains more effective than RL for smaller models. In this work, we demonstrate that large-scale RL can significantly enhance the reasoning capabilities of strong, small- and mid-sized models, achieving results that surpass those of state-of-the-art distillation-based models. We systematically study the RL training process through extensive ablations and propose a simple yet effective approach: first training on math-only prompts, then on code-only prompts. Notably, we find that math-only RL not only significantly enhances the performance of strong distilled models on math benchmarks (e.g., +14.6\% / +17.2\% on AIME 2025 for the 7B / 14B models), but also code reasoning tasks (e.g., +6.8\% / +5.8\% on LiveCodeBench for the 7B / 14B models). In addition, extended code-only RL further improves code benchmark performance while causing minimal degradation in math results. We develop a robust data curation pipeline to collect challenging prompts with high-quality, verifiable answers and test cases to enable verification-based RL across both domains. The dataset will be released to support open research.Finally, we identify key experimental insights, including curriculum learning with progressively increasing response lengths and the stabilizing effect of on-policy parameter updates. We find that RL not only elicits the foundational reasoning capabilities acquired during pretraining and supervised fine-tuning (SFT), but also pushes the limits of the model’s reasoning ability, enabling it to solve problems that were previously unsolvable.



Paperid:2570
Authors:Zibin Dong, Yicheng Liu, Yinchuan Li, Hang Zhao, Jianye Hao
Abstract:
Diffusion policies have emerged as a mainstream paradigm for building vision-language-action (VLA) models. Although they demonstrate strong robot control capabilities, their training efficiency remains suboptimal. In this work, we identify a fundamental challenge in conditional diffusion policy training: when generative conditions are hard to distinguish, the training objective degenerates into modeling the marginal action distribution, a phenomenon we term loss collapse. To overcome this, we propose Cocos, a simple yet general solution that modifies the source distribution in the conditional flow matching to be condition-dependent. By anchoring the source distribution around semantics extracted from condition inputs, Cocos encourages stronger condition integration and prevents the loss collapse. We provide theoretical justification and extensive empirical results across simulation and real-world benchmarks. Our method achieves faster convergence and higher success rates than existing approaches, matching the performance of large-scale pre-trained VLAs using significantly fewer gradient steps and parameters. Cocos is lightweight, easy to implement, and compatible with diverse policy architectures, offering a general-purpose improvement to diffusion policy training.



Paperid:2571
Authors:Likun Wang, Xiangteng Zhang, Yinuo Wang, Guojian Zhan, Wenxuan Wang, Haoyu Gao, Jingliang Duan, Shengbo Li
Abstract:
Exploration is crucial in Reinforcement Learning (RL) as it enables the agent to understand the environment for better decision-making. Existing exploration methods fall into two paradigms: active exploration, which injects stochasticity into the policy but struggles in high-dimensional environments, and passive exploration, which manages the replay buffer to prioritize under-explored regions but lacks sample diversity. To address the limitation in passive exploration, we propose Model-based Prioritized Generative Exploration (M-PGE), which augments exploration through the generation of under-explored critical states and synthesis of dynamics-consistent experiences. M-PGE consists of two components: (1) a diffusion generator for critical states under the guidance of entropy and TD error, and (2) a one-step imagination world model for constructing critical transitions for agent learning. Our method is simple to implement and seamlessly integrates with mainstream off-policy RL algorithms without structural modifications. Experiments on OpenAI Gym and DeepMind Control Suite demonstrate that M-PGE, as an exploration augmentation, significantly enhances efficiency and performance in complex tasks.



Paperid:2572
Authors:Xuyuan Xiong, Pedro Chumpitaz-Flores, Kaixun Hua, Cheng Hua
Abstract:
Interpretable reinforcement learning policies are essential for high-stakes decision-making, yet optimizing decision tree policies in Markov Decision Processes (MDPs) remains challenging. We propose SPOT, a novel method for computing decision tree policies, which formulates the optimization problem as a mixed-integer linear program (MILP). To enhance efficiency, we employ a space-reduced branch-and-bound approach that decouples the MDP dynamics from tree-structure constraints, enabling efficient parallel search. This significantly improves runtime and scalability compared to previous methods. Our approach ensures that each iteration yields the optimal decision tree. Experimental results on standard benchmarks demonstrate that SPOT achieves substantial speedup and scales to larger MDPs with a significantly higher number of states. The resulting decision tree policies are interpretable and compact, maintaining transparency without compromising performance. These results demonstrate that our approach simultaneously achieves interpretability and scalability, delivering high-quality policies an order of magnitude faster than existing approaches.



Paperid:2573
Authors:Dongheng Lin, Mengxue Qu, Kunyang Han, Jianbo Jiao, Xiaojie Jin, Yunchao Wei
Abstract:
Most video-anomaly research stops at frame-wise detection, offering little insight into why an event is abnormal, typically outputting only frame-wise anomaly scores without spatial or semantic context. Recent video anomaly localization and video anomaly understanding methods improve explainability but remain data-dependent and task-specific. We propose a unified reasoning framework that bridges the gap between temporal detection, spatial localization, and textual explanation. Our approach is built upon a chained test-time reasoning process that sequentially connects these tasks, enabling holistic zero-shot anomaly analysis without any additional training. Specifically, our approach leverages intra-task reasoning to refine temporal detections and inter-task chaining for spatial and semantic understanding, yielding improved interpretability and generalization in a fully zero-shot manner. Without any additional data or gradients, our method achieves state-of-the-art zero-shot performance across multiple video anomaly detection, localization, and explanation benchmarks. The results demonstrate that careful prompt design with task-wise chaining can unlock the reasoning power of foundation models, enabling practical, interpretable video anomaly analysis in a fully zero-shot manner. Code will be released.



Paperid:2574
Authors:Christian Walder, Deep Tejas Karkhanis
Abstract:
Abstract:Reinforcement Learning algorithms commonly sample multiple ($n>1$) solution attempts for each problem and reward them independently. This optimizes for pass@1 performance and prioritizes the strength of isolated samples at the expense of the diversity and collective utility of sets of samples. Such algorithms under-utilize the sampling capacity, limiting exploration and eventual improvement on harder examples. As a fix, we propose Pass-at-$k$ Policy Optimization (PKPO), a transformation on the final rewards which leads to direct optimization of pass@k performance, thus optimizing for sets of samples that feature a large maximum reward when considered jointly. Our primary contribution is to derive novel low variance unbiased estimators for the pass@k and its gradient, in both the binary and continuous reward settings. We show that the optimization with these estimators reduces to standard reinforcement learning with (mini-batches of) rewards that have been jointly transformed by a function that is stable and efficient to compute.While previous efforts propose transformations for $k=n$, our transformations are the first to enable robust optimization of the pass@k for any arbitrary $k \leq n$. Rather than simply trading off pass@1 performance for pass@k gains, our method allows annealing $k$ during training, optimizing both metrics and often achieving strong pass@1 performance alongside significant pass@k gains.We validate our reward transformations on illustrative toy experiments, which reveal the variance reducing properties of our formulations. We also include real-world examples using the open-source model Gemma-2 . We find that our transformation effectively optimizes for the target $k$. Furthermore, higher $k$ values enable solving more and harder problems, while annealing $k$ boosts both the pass@1 and pass@k. Crucially, for challenging task sets where conventional pass@1 optimization stalls, our pass@k approach unblocks learning, likely due to better exploration by prioritizing joint utility over the utility of individual samples.



Paperid:2575
Authors:Seongheon Park, Sharon Li
Abstract:
Object hallucination in large vision-language models presents a significant challenge to their safe deployment in real-world applications. Recent works have proposed object-level hallucination scores to estimate the likelihood of object hallucination; however, these methods typically adopt either a global or local perspective in isolation, which may limit detection reliability. In this paper, we introduce GLSim, a novel training-free object hallucination detection framework that leverages complementary global and local embedding similarity signals between image and text modalities, enabling more accurate and reliable hallucination detection in diverse scenarios. We comprehensively benchmark existing object hallucination detection methods and demonstrate that GLSim achieves superior detection performance, outperforming competitive baselines by a significant margin.



Paperid:2576
Authors:Han Bao, Shinsaku Sakaue, Yuki Takezawa
Abstract:
Abstract:The gradient descent (GD) has been one of the most common optimizer in machine learning. In particular, the loss landscape of a neural network is typically sharpened during the initial phase of training, making the training dynamics hover on the edge of stability. This is beyond our standard understanding of GD convergence in the stable regime where arbitrarily chosen stepsize is sufficiently smaller than the edge of stability. Recently, Wu et al. (COLT2024) have showed that GD converges with arbitrary stepsize under linearly separable logistic regression. Although their analysis hinges on the self-bounding property of the logistic loss, which seems to be a cornerstone to establish a modified descent lemma, our pilot study shows that other loss functions without the self-bounding property can make GD converge with arbitrary stepsize. To further understand what property of a loss function matters in GD, we aim to show arbitrary-stepsize GD convergence for a general loss function based on the framework of \emph{Fenchel--Young losses}. We essentially leverage the classical perceptron argument to derive the convergence rate for achieving $\epsilon$-optimal loss, which is possible for a majority of Fenchel--Young losses. Among typical loss functions, the Tsallis entropy achieves the GD convergence rate $T=\Omega(\epsilon^{-1/2})$, and the R{\'e}nyi entropy achieves the far better rate $T=\Omega(\epsilon^{-1/3})$. We argue that these better rate is possible because of \emph{separation margin} of loss functions, instead of the self-bounding property.



Paperid:2577
Authors:Yasar Utku Alcalar, Merve Gulle, Mehmet Akcakaya
Abstract:
Physics-driven deep learning (PD-DL) approaches have become popular for improved reconstruction of fast magnetic resonance imaging (MRI) scans. Though PD-DL offers higher acceleration rates than existing clinical fast MRI techniques, their use has been limited outside specialized MRI centers. A key challenge is generalization to underrepresented pathologies or populations, noted in multiple studies, with fine-tuning on target populations suggested for improvement. However, current approaches for PD-DL training require access to raw k-space measurements, which is typically only available at specialized MRI centers that have research agreements for such data access. This is especially an issue for rural and underserved areas, where commercial MRI scanners only provide access to a final reconstructed image. To tackle these challenges, we propose Compressibility-inspired Unsupervised Learning via Parallel Imaging Fidelity (CUPID) for high-quality PD-DL training using only routine clinical reconstructed images exported from an MRI scanner. CUPID evaluates output quality with a compressibility-based approach while ensuring that the output stays consistent with the clinical parallel imaging reconstruction through well-designed perturbations. Our results show CUPID achieves similar quality to established PD-DL training that requires k-space data while outperforming compressed sensing (CS) and diffusion-based generative methods. We further demonstrate its effectiveness in a zero-shot training setup for retrospectively and prospectively sub-sampled acquisitions, attesting to its minimal training burden. As an approach that radically deviates from existing strategies, CUPID presents an opportunity to provide equitable access to fast MRI for underserved populations in an attempt to reduce the inequalities associated with this expensive imaging modality.



Paperid:2578
Authors:Harshit Sushil Sikchi, Siddhant Agarwal, Pranaya Jajoo, Samyak Parajuli, Caleb Chuck, Max Rudolph, Peter Stone, Amy Zhang, Scott Niekum
Abstract:
The reward hypothesis states that all goals and purposes can be understood as the maximization of a received scalar reward signal. However, in practice, defining such a reward signal is notoriously difficult, as humans are often unable to predict the optimal behavior corresponding to a reward function. Natural language offers an intuitive alternative for instructing reinforcement learning (RL) agents, yet previous language-conditioned approaches either require costly supervision or test-time training given a language instruction. In this work, we present a new approach that uses a pretrained RL agent trained using only unlabeled, offline interactions—without task-specific supervision or labeled trajectories—to get zero-shot test-time policy inference from arbitrary natural language instructions. We introduce a framework comprising three steps:imagine,project, andimitate. First, the agent imagines a sequence of observations corresponding to the provided language description using video generative models. Next, these imagined observations are projected into the target environment domain. Finally, an agent pretrained in the target environment with unsupervised RL instantly imitates the projected observation sequence through a closed-form solution. To the best of our knowledge, our method, RLZero, is the first approach to show direct language-to-behavior generation abilities on a variety of tasks and environments without any in-domain supervision. We further show that components of RLZero can be used to generate policies zero-shot from cross-embodied videos, such as those available on YouTube, even for complex embodiments like humanoids.



Paperid:2579
Authors:Andrew Wang, Elisa Nguyen, Runshi Yang, Juhan Bae, Sheila McIlraith, Roger Grosse
Abstract:
Training data attribution (TDA) provides insights into which training data is responsible for a learned model behavior. Gradient-based TDA methods such as influence functions and unrolled differentiation both involve a computation that resembles an inverse Hessian-vector product (iHVP), which is difficult to approximate efficiently. We introduce an algorithm (ASTRA) which uses the EKFAC-preconditioner on Neumann series iterations to arrive at an accurate iHVP approximation for TDA. ASTRA is easy to tune, requires fewer iterations than Neumann series iterations, and is more accurate than EKFAC-based approximations. Using ASTRA, we show that improving the accuracy of the iHVP approximation can significantly improve TDA performance.



Paperid:2580
Authors:Jingyao Wang, Wenwen Qiang, Zeen Song, Changwen Zheng, Hui Xiong
Abstract:
Large language models (LLMs) excel at complex tasks thanks to advances in their reasoning abilities. However, existing methods overlook the trade-off between reasoning effectiveness and efficiency, often encouraging unnecessarily long reasoning chains and wasting tokens. To address this, we propose Learning to Think (L2T), an information-theoretic reinforcement fine-tuning framework for LLMs to make the models achieve optimal reasoning with fewer tokens. Specifically, L2T treats each query-response interaction as a hierarchical session of multiple episodes and proposes a universal dense process reward, i.e., quantifies the episode-wise information gain in parameters, requiring no extra annotations or task-specific evaluators. We propose a method to quickly estimate this reward based on PAC-Bayes bounds and the Fisher information matrix. Theoretical analyses show that it significantly reduces computational complexity with high estimation accuracy. By immediately rewarding each episode's contribution and penalizing excessive updates, L2T optimizes the model via reinforcement learning to maximize the use of each episode and achieve effective updates. Empirical results on various reasoning benchmarks and base models demonstrate the advantage of L2T across different tasks, boosting both reasoning effectiveness and efficiency.



Paperid:2581
Authors:Ruixin Guo, Ruoming Jin, Xinyu Li, Yang Zhou
Abstract:
Linear Autoencoders (LAEs) have shown strong performance in state-of-the-art recommender systems. However, this success is largely empirical, with limited theoretical justification. In this paper, we investigate the generalizability -- a theoretical measure of model performance in statistical machine learning -- of multivariate linear regression and LAEs. We first propose a PAC-Bayes bound for multivariate linear regression, extending the earlier PAC-Bayes bound for multiple linear regression by Shalaeva et al, and provide sufficient conditions for its theoretical convergence. We then adapt this bound to LAEs by showing they can be interpreted as constrained multivariate linear regression on bounded data. To address the computational challenges arising from these constraints, we develop practical methods for optimizing the bound. Experimental results demonstrate that our bound is tight and correlates well with practical evaluation metrics such as Recall@K and NDCG@K.



Paperid:2582
Authors:Cathy Jiao, Yijun Pan, Emily Xiao, Daisy Sheng, Niket Jain, Hanzhang Zhao, Ishita Dasgupta, Jiaqi Ma, Chenyan Xiong
Abstract:
Data attribution methods quantify the influence of training data on model outputs and are becoming increasingly relevant for a wide range of LLM research and applications, including dataset curation, model interpretability, data valuation. However, there remain critical gaps in systematic LLM-centric evaluation of data attribution methods. To this end, we introduce DATE-LM (Data Attribution Evaluation in Language Models), a unified benchmark for evaluating data attribution methods through real-world LLM applications. DATE-LM measures attribution quality through three key tasks — training data selection, toxicity/bias filtering, and factual attribution. Our benchmark is designed for ease of use, enabling researchers to configure and run large-scale evaluations across diverse tasks and LLM architectures. Furthermore, we use DATE-LM to conduct a large-scale evaluation of existing data attribution methods. Our findings show that no single method dominates across all tasks, data attribution methods have trade-offs with simpler baselines, and method performance is sensitive to task-specific evaluation design. Finally, we release a public leaderboard for quick comparison of methods and to facilitate community engagement. We hope DATE-LM serves as a foundation for future data attribution research in LLMs.



Paperid:2583
Authors:Seanie Lee, Sangwoo Park, Dong Bok Lee, Dominik Wagner, Haebin Seong, Tobias Bocklet, Juho Lee, Sung Ju Hwang
Abstract:
Abstract:Low-Rank Adaptation (LoRA), which introduces a product of two trainable low-rank matrices into frozen pre-trained weights, is widely used for efficient fine-tuning of language models in federated learning (FL). However, when combined with differentially private stochastic gradient descent (DP-SGD), LoRA faces substantial noise amplification: DP-SGD perturbs per-sample gradients, and the matrix multiplication of the LoRA update ($BA$) intensifies this effect. Freezing one matrix (e.g., $A$) reduces the noise but restricts model expressiveness, often resulting in suboptimal adaptation.To address this, we propose $\texttt{FedSVD}$, a simple yet effective method that introduces a global reparameterization based on singular value decomposition (SVD). In our approach, each client optimizes only the $B$ matrix and transmits it to the server. The server aggregates the $B$ matrices, computes the product $BA$ using the previous $A$, and refactorizes the result via SVD. This yields a new adaptive $A$ composed of the orthonormal right singular vectors of $BA$, and an updated $B$ containing the remaining SVD components. This reparameterization avoids quadratic noise amplification, while allowing $A$ to better capture the principal directions of the aggregate updates.Moreover, the orthonormal structure of $A$ bounds the gradient norms of $B$ and preserves more signal under DP-SGD, as confirmed by our theoretical analysis.As a result, $\texttt{FedSVD}$ consistently improves stability and performance across a variety of privacy settings and benchmarks, outperforming relevant baselines under both private and non-private regimes.



Paperid:2584
Authors:François Bachoc, Jerome Bolte, Ryan Boustany, Loubes Jean-Michel
Abstract:
Despite growing empirical evidence of bias amplification in machine learning, its theoretical foundations remain poorly understood. We develop a formal framework for majority-minority learning tasks, showing how standard training can favor majority groups and produce stereotypical predictors that neglect minority-specific features. Assuming population and variance imbalance, our analysis reveals three key findings: (i) the close proximity between "full-data" and stereotypical predictors, (ii) the dominance of a region where training the entire model tends to merely learn the majority traits, and (iii) a lower bound on the additional training required to reach fairer solutions. Our results are illustrated through experiments in deep learning for tabular and image classification tasks.



Paperid:2585
Authors:Wenxuan Zeng, Ye Dong, Jinjin Zhou, Jin Tan, Lei Wang, Tao Wei, Runsheng Wang, Meng Li
Abstract:
Private large language model (LLM) inference based on secure multi-party computation (MPC) achieves formal data privacy protection but suffers from significant latency overhead, especially for long input sequences. While key-value (KV) cache eviction algorithms have been proposed for efficient LLM inference in plaintext, they are not designed for MPC and cannot benefit private LLM inference directly. In this paper, we propose an accurate and MPC-friendly KV cache eviction framework, dubbed MPCache, building on the observation that historical tokens in a long sequence may have different effects on the downstream decoding. Hence, MPCache combines a look-once static eviction algorithm to discard unimportant KV cache and a query-aware dynamic selection algorithm to activate only a small subset of KV cache for attention computation. MPCache further incorporates a series of optimizations for efficient dynamic KV cache selection, including MPC-friendly similarity approximation, hierarchical clustering, and cross-layer index-sharing strategy. Extensive experiments demonstrate that MPCache consistently outperforms prior-art KV cache eviction baselines across different generation tasks and achieves 1.8∼2.01x and 3.39∼8.37x decoding latency and communication reduction on different sequence lengths, respectively.



Paperid:2586
Authors:Tengyun Ma, Damon Yao, Daojing He, Shihao Peng, YU LI, Shaohui Liu, Zhuotao Tian
Abstract:
Large Language Models (LLMs) have emerged as powerful tools for diverse applications. However, their uniform token processing paradigm introduces critical vulnerabilities in instruction handling, particularly when exposed to adversarial scenarios. In this work, we identify and propose a novel class of vulnerabilities, termedTool-Completion Attack (TCA), which exploits function-calling mechanisms to subvert model behavior. To evaluate LLM robustness against such threats, we introduce the Tool-Completion Benchmark, a comprehensive security assessment framework, which reveals that even state-of-the-art models remain susceptible to TCA, with surprisingly high attack success rates. To address these vulnerabilities, we introduce Context-Aware Hierarchical Learning (CAHL), a sophisticated mechanism that dynamically equilibrates semantic comprehension with role-specific instruction constraints. CAHL leverages the contextual correlationsbetween different instruction segments to establish a robust, context-aware instruction hierarchy. Extensive experiments demonstrate that CAHL significantly enhances LLM robustness against both conventional attacks and the proposed TCA, exhibiting strong generalization capabilities in zero-shot evaluations while stillpreserving model performance on generic tasks. The proposed benchmark, code, and models willbe made publicly available.



Paperid:2587
Authors:Xiaoxi Li, Jiajie Jin, Guanting Dong, Hongjin Qian, Yutao Zhu, Yongkang Wu, Ji-Rong Wen, Zhicheng Dou
Abstract:
Large reasoning models (LRMs), such as OpenAI-o1 and DeepSeek-R1, demonstrate impressive long-horizon reasoning capabilities. However, their reliance on static internal knowledge limits their performance on complex, knowledge-intensive tasks and hinders their ability to produce comprehensive research reports requiring synthesis of diverse web information. To address this, we propose WebThinker, a deep research agent that empowers LRMs to autonomously search the web, navigate web pages, and draft research reports during the reasoning process. WebThinker integrates a Deep Web Explorer module, enabling LRMs to dynamically search, navigate, and extract information from the web when encountering knowledge gaps. It also employs an Autonomous Think-Search-and-Draft strategy, allowing the model to seamlessly interleave reasoning, information gathering, and report writing in real time. To further enhance research tool utilization, we introduce an RL-based training strategy via iterative online Direct Preference Optimization (DPO). Extensive experiments on complex reasoning benchmarks (GPQA, GAIA, WebWalkerQA, HLE) and scientific report generation tasks (Glaive) demonstrate that WebThinker significantly outperforms existing methods and strong proprietary systems. Our approach enhances LRM reliability and applicability in complex scenarios, paving the way for more capable and versatile deep research systems. The code is available at https://anonymous.4open.science/r/WebThinker-D6D4.



Paperid:2588
Authors:Tao Liu, Chongyu Wang, Rongjie Li, Yingchen Yu, Xuming He, Song Bai
Abstract:
While Multimodal Large Language Models (MLLMs) have advanced GUI navigation agents, current approaches face limitations in cross-domain generalization and effective history utilization. We present a reasoning-enhanced framework that systematically integrates structured reasoning, action prediction, and history summarization. The structured reasoning component generates coherent Chain-of-Thought analyses combining progress estimation and decision reasoning, which inform both immediate action predictions and compact history summaries for future steps. Based on this framework, we train a GUI agent, GUI-Rise, through supervised fine-tuning on pseudo-labeled trajectories and reinforcement learning with Group Relative Policy Optimization (GRPO). This framework employs specialized rewards, including a history-aware objective, directly linking summary quality to subsequent action performance. Comprehensive evaluations on standard benchmarks demonstrate state-of-the-art results under identical training data conditions, with particularly strong performance in out-of-domain scenarios. These findings validate our framework's ability to maintain robust reasoning and generalization across diverse GUI navigation tasks.



Paperid:2589
Authors:Oussema Dhaouadi, Riccardo Marin, Johannes Meier, Jacques Kaiser, Daniel Cremers
Abstract:
Accurate visual localization from aerial views is a fundamental problem with applications in mapping, large-area inspection, and search-and-rescue operations. In many scenarios, these systems require high-precision localization while operating with limited resources (e.g., no internet connection or GNSS/GPS support), making large image databases or heavy 3D models impractical. Surprisingly, little attention has been given to leveraging orthographic geodata as an alternative paradigm, which is lightweight and increasingly available through free releases by governmental authorities (e.g., the European Union). To fill this gap, we propose OrthoLoC, the first large-scale dataset comprising 16,425 UAV images from Germany and the United States with multiple modalities. The dataset addresses domain shifts between UAV imagery and geospatial data. Its paired structure enables fair benchmarking of existing solutions by decoupling image retrieval from feature matching, allowing isolated evaluation of localization and calibration performance. Through comprehensive evaluation, we examine the impact of domain shifts, data resolutions, and covisibility on localization accuracy. Finally, we introduce a refinement technique called AdHoP, which can be integrated with any feature matcher, improving matching by up to 95% and reducing translation error by up to 63%. The dataset and code are available at: https://deepscenario.github.io/OrthoLoC .



Paperid:2590
Authors:Hao LU, Tianshuo Xu, Wenzhao Zheng, Yunpeng Zhang, Wei Zhan, Dalong Du, Masayoshi TOMIZUKA, Kurt Keutzer, Yingcong Chen
Abstract:
Large reconstruction model has remarkable progress, which can directly predict 3D or 4D representations for unseen scenes and objects. However, current work has not systematically explored the potential of large reconstruction models in the field of autonomous driving. To achieve this, we introduce the Large 4D Gaussian Reconstruction Model (DrivingRecon). With an elaborate and simple framework design, it not only ensures efficient and high-quality reconstruction, but also provides potential for downstream tasks. There are two core contributions: firstly, the Prune and Dilate Block (PD-Block) is proposed to prune redundant and overlapping Gaussian points and dilate Gaussian points for complex objects. Then, dynamic and static decoupling is tailored to better learn the temporary-consistent geometry across different time. Experimental results demonstrate that DrivingRecon significantly improves scene reconstruction quality compared to existing methods. Furthermore, we explore applications of DrivingRecon in model pre-training, vehicle type adaptation, and scene editing. Our code will be available.



Paperid:2591
Authors:Runpeng Xie, Quanwei Wang, Hao Hu, Zherui Zhou, Ni Mu, Xiyun Li, Yiqin Yang, Shuang Xu, Qianchuan Zhao, Bo Xu
Abstract:
Comprehending natural language and following human instructions are critical capabilities for intelligent agents. However, the flexibility of linguistic instructions induces substantial ambiguity across language-conditioned tasks, severely degrading algorithmic performance.To address these limitations, we present a novel method named DAIL (Distributional Aligned Learning), featuring two key components: distributional policy and semantic alignment.Specifically, we provide theoretical results that the value distribution estimation mechanism enhances task differentiability.Meanwhile, the semantic alignment module captures the correspondence between trajectories and linguistic instructions.Extensive experimental results on both structured and visual observation benchmarks demonstrate that DAIL effectively resolves instruction ambiguities, achieving superior performance to baseline methods.



Paperid:2592
Authors:Wei Zhang, Zhenhong Zhou, Junfeng Fang, Rongwu Xu, Yuanhe Zhang, Rui Wang, Ge Zhang, Xinfeng Li, Kun Wang, Li Sun, Lingjuan Lyu, Yang Liu, Sen Su
Abstract:
While large language models (LLMs) can solve PhD-level reasoning problems over long context inputs, they still struggle with a seemingly simpler task:following explicit length instructions—e.g.,write a 10,000-word novel. Additionally, models often generate far too short outputs, terminate prematurely, or even refuse the request. Existing benchmarks focus primarily on evaluating generations quality, but often overlook whether the generations meet length constraints. To this end, we introduceLength Instruction Following Evaluation Benchmark(LIFEBench) to comprehensively evaluate LLMs' ability to follow length instructions across diverse tasks and a wide range of specified lengths. LIFEBench consists of 10,800 instances across 4 task categories in both English and Chinese, covering length constraints ranging from 16 to 8192 words. We evaluate 26 widely-used LLMs and find that most models reasonably follow short-length instructions but deteriorate sharply beyond a certain threshold. Surprisingly, almost all models fail to reach the vendor-claimed maximum output lengths in practice, as further confirmed by our evaluations extending up to 32K words. Even long-context LLMs, despite their extended input-output windows, counterintuitively fail to improve length-instructions following. Notably, Reasoning LLMs outperform even specialized long-text generation models, achieving state-of-the-art length following. Overall, LIFEBench uncovers fundamental limitations in current LLMs' length instructions following ability, offering critical insights for future progress.



Paperid:2593
Authors:Ali Taghibakhshi, Sharath Turuvekere Sreenivas, Saurav Muralidharan, Marcin Chochowski, Yashaswi Karnati, Raviraj Joshi, Ameya Mahabaleshwarkar, ZIJIA CHEN, Yoshi Suhara, Oluwatobi Olabiyi, Daniel Korzekwa, Mostofa Patwary, Mohammad Shoeybi, Jan Kautz, Bryan Catanzaro, Ashwath Aithal, Nima Tajbakhsh, Pavlo Molchanov
Abstract:
Hybrid language models that combine Attention and State Space Models (SSMs) have been shown to achieve state-of-the-art accuracy and runtime performance. Recent work has also demonstrated that applying pruning and distillation to Attention-only models yields smaller, more accurate models at a fraction of the training cost. In this work, we explore the effectiveness of compressing Hybrid architectures. To this end, we introduce a novel group-aware pruning method for Mamba layers that preserves the structural integrity of SSM blocks and their sequence modeling capabilities. We combine this method with FFN, embedding dimension, and layer pruning, along with knowledge distillation-based retraining to obtain a unified compression recipe for hybrid models. Using this recipe, we compress the Nemotron-H 8B Hybrid model down to 4B parameters with up to 40x fewer training tokens compared to similarly-sized models. The resulting model surpasses the accuracy of similarly-sized models while achieving ~2x faster inference throughput, significantly advancing the Pareto frontier.



Paperid:2594
Authors:Yuxing Lu, Wei Wu, Xukai Zhao, Rui Peng, Jinzhuo Wang
Abstract:
Maintaining comprehensive and up-to-date knowledge graphs (KGs) is critical for modern AI systems, but manual curation struggles to scale with the rapid growth of scientific literature. This paper presents KARMA, a novel framework employing multi-agent large language models (LLMs) to automate KG enrichment through structured analysis of unstructured text. Our approach employs nine collaborative agents, spanning entity discovery, relation extraction, schema alignment, and conflict resolution that iteratively parse documents, verify extracted knowledge, and integrate it into existing graph structures while adhering to domain-specific schema. Experiments on 1,200 PubMed articles from three different domains demonstrate the effectiveness of KARMA in knowledge graph enrichment, with the identification of up to 38,230 new entities while achieving 83.1\% LLM-verified correctness and reducing conflict edges by 18.6\% through multi-layer assessments.



Paperid:2595
Authors:Dmitriy Shopkhoev, Ammar Ali, Magauiya Zhussip, Valentin Malykh, Stamatios Lefkimmiatis, Nikos Komodakis, Sergey Zagoruyko
Abstract:
We introduce ReplaceMe, a generalized training-free depth pruning method that effectively replaces transformer blocks with a linear operation, while maintaining high performance for low compression ratios. In contrast to conventional pruning approaches that require additional training or fine-tuning, our approach requires only a small calibration dataset that is used to estimate a linear transformation, which approximates the pruned blocks. The estimated linear mapping can be seamlessly merged with the remaining transformer blocks, eliminating the need for any additional network parameters. Our experiments show that ReplaceMe consistently outperforms other training-free approaches and remains highly competitive with state-of-the-art pruning methods that involve extensive retraining/fine-tuning and architectural modifications. Applied to several large language models (LLMs), ReplaceMe achieves up to 25% pruning while retaining approximately 90% of the original model’s performance on open benchmarks - without any training or healing steps, resulting in minimal computational overhead (see Fig.1).



Paperid:2596
Authors:Yuanzhe Liu, Ryan Deng, Tim Kaler, Xuhao Chen, Charles Leiserson, Yao Ma, Jie Chen
Abstract:
Recent studies show that LLMs possess different skills and specialize in different tasks. In fact, we observe that their varied performance occur in several levels of granularity. For example, in the code optimization task, code LLMs excel at different optimization categories and no one dominates others. This observation prompts the question of how one leverages multiple LLM agents to solve a coding problem without knowing their complementary strengths a priori. We argue that a team of agents can learn from each other's successes and failures so as to improve their own performance. Thus, a lesson is the knowledge produced by an agent and passed on to other agents in the collective solution process. We propose a lesson-based collaboration framework, design the lesson solicitation--banking--selection mechanism, and demonstrate that a team of small LLMs with lessons learned can outperform a much larger LLM and other multi-LLM collaboration methods.



Paperid:2597
Authors:YANG SONGXIAO, Haolin Wang, Yao Fu, Ye Tian, Tamostu Kamishima, Masayuki Ikebe, Yafei Ou, Masatoshi Okutomi
Abstract:
Rheumatoid arthritis (RA) is a common autoimmune disease that has been the focus of research in computer-aided diagnosis (CAD) and disease monitoring. In clinical settings, conventional radiography (CR) is widely used for the screening and evaluation of RA due to its low cost and accessibility. The wrist is a critical region for the diagnosis of RA. However, CAD research in this area remains limited, primarily due to the challenges in acquiring high-quality instance-level annotations. (i) The wrist comprises numerous small bones with narrow joint spaces, complex structures, and frequent overlaps, requiring detailed anatomical knowledge for accurate annotation. (ii) Disease progression in RA often leads to osteophyte, bone erosion (BE), and even bony ankylosis, which alter bone morphology and increase annotation difficulty, necessitating expertise in rheumatology.This work presents a multi-task dataset for wrist bone in CR, including two tasks: (i) wrist bone instance segmentation and (ii) Sharp/van der Heijde (SvdH) BE scoring, which is the first public resource for wrist bone instance segmentation. This dataset comprises 621 wrist conventional radiographs of 227 patients from four medical centers, with pixel-level instance segmentation annotations for 443 images and SvdH BE scores for 548 images. This dataset can potentially support a wide range of research tasks related to RA, including joint space narrowing (JSN) progression quantification, BE detection, bone deformity evaluation, and osteophyte detection. It may also be applied to other wrist-related tasks, such as carpal bone fracture localization.We hope this dataset will significantly lower the barrier to research on wrist RA and accelerate progress in CAD research within the RA-related domain.Benchmark \& Code: https://github.com/YSongxiao/RAM-W600Data \& Dataset Card: https://huggingface.co/datasets/TokyoTechMagicYang/RAM-W600



Paperid:2598
Authors:Masahiro Kato, Fumiaki Kozai, RYO INOKUCHI
Abstract:
The estimation of \emph{average treatment effects} (ATEs), defined as the difference in expected outcomes between treatment and control groups, is a central topic in causal inference. This study develops semiparametric efficient estimators for ATE in a setting where only a treatment group and an unlabeled group—consisting of units whose treatment status is unknown—are observed. This scenario constitutes a variant of learning from positive and unlabeled data (\emph{PU learning}) and can be viewed as a special case of ATE estimation with missing data. For this setting, we derive the \emph{semiparametric efficiency bounds}, which characterize the lowest achievable asymptotic variance for regular estimators. We then construct semiparametric \emph{efficient ATE estimators} that attain these bounds. Our results contribute to the literature on causal inference with missing data and weakly supervised learning.



Paperid:2599
Authors:Guanlue Li, Xufeng Zhao, Fang Wu, Sören Laue
Abstract:
Protein-protein interactions (PPIs) are governed by surface complementarity and hydrophobic interactions at protein interfaces. However, designing diverse and physically realistic protein structure and surfaces that precisely complement target receptors remains a significant challenge in computational protein design. In this work, we introduce PepBridge, a novel framework for the joint design of protein surface and structure that seamlessly integrates receptor surface geometry and biochemical properties. Starting with a receptor surface represented as a 3D point cloud, PepBridge generates complete protein structures through a multi-step process. First, it employs denoising diffusion bridge models (DDBMs) to map receptor surfaces to ligand surfaces. Next, a multi-model diffusion model predicts the corresponding structure, while Shape-Frame Matching Networks ensure alignment between surface geometry and backbone architecture. This integrated approach facilitates surface complementarity, conformational stability, and chemical feasibility. Extensive validation across diverse protein design scenarios demonstrates PepBridge's efficacy in generating structurally viable proteins, representing a significant advancement in the joint design of top-down protein structure. The code can be found at this anonymous link: https://anonymous.4open.science/r/Pepbridge.



Paperid:2600
Authors:Shangshu Yu, Wen Li, Xiaotian Sun, Zhimin Yuan, Xin Wang, Sijie Wang, Rui She, Cheng Wang
Abstract:
Prevailing scene coordinate regression methods for LiDAR localization suffer from localization ambiguities, as distinct locations can exhibit similar geometric signatures — a challenge that current geometry-based regression approaches have yet to solve. Recent vision–language models show that textual descriptions can enrich scene understanding, supplying potential localization cues missing from point cloud geometries. In this paper, we propose GTR-Loc, a novel text-assisted LiDAR localization framework that effectively generates and integrates geospatial text regularization to enhance localization accuracy. We propose two novel designs: a Geospatial Text Generator that produces discrete pose-aware text descriptions, and a LiDAR-Anchored Text Embedding Refinement module that dynamically constructs view-specific embeddings conditioned on current LiDAR features. The geospatial text embeddings act as regularization to effectively reduce localization ambiguities. Furthermore, we introduce a Modality Reduction Distillation strategy to transfer textual knowledge. It enables high-performance LiDAR-only localization during inference, without requiring runtime text generation. Extensive experiments on challenging large-scale outdoor datasets, including QEOxford, Oxford Radar RobotCar, and NCLT, demonstrate the effectiveness of GTR-Loc. Our method significantly outperforms state-of-the-art approaches, notably achieving a 9.64%/8.04% improvement in position/orientation accuracy on QEOxford. Our code will be made available upon acceptance.



Paperid:2601
Authors:Yuqing Wang, Zhaiyu Chen, Xiaoxiang Zhu
Abstract:
Abstract:3D shape completion methods typically assume scans are pre-aligned to a canonical frame. This leaks pose and scale cues that networks may exploit to memorize absolute positions rather than inferring intrinsic geometry. When such alignment is absent in real data, performance collapses. We argue that robust generalization demands architectural equivariance to the similarity group, $\mathrm{SIM}(3)$, so the model remains agnostic to pose and scale. Following this principle, we introduce the first $\mathrm{SIM}(3)$-equivariant shape completion network, whose modular layers successively canonicalize features, reason over similarity-invariant geometry, and restore the original frame. Under a de-biased evaluation protocol that removes the hidden cues, our model outperforms both equivariant and augmentation baselines on the PCN benchmark. It also sets new zero-shot records on real driving and indoor scans, lowering minimal matching distance on KITTI by 17\% and Chamfer distance $\ell1$ on OmniObject3D by 14\%. Perhaps surprisingly, ours under the stricter protocol still outperforms competitors under their biased settings. These results establish full $\mathrm{SIM}(3)$ equivariance as an effective route to truly generalizable shape completion.



Paperid:2602
Authors:Shulin Huang, Linyi Yang, Yan Song, Shawn Chen, Leyang Cui, Ziyu Wan, Qingcheng Zeng, Ying Wen, Kun Shao, Weinan Zhang, Jun Wang, Yue Zhang
Abstract:
Evaluating large language models (LLMs) poses significant challenges, particularly due to issues of data contamination and the leakage of correct answers. To address these challenges, we introduce ThinkBench, a novel evaluation framework designed to robustly evaluate the reasoning capability of LLMs. ThinkBench proposes a dynamic data generation method for constructing out-of-distribution (OOD) datasets and offers an OOD dataset that contains 2,912 samples drawn from reasoning tasks. ThinkBench unifies the evaluation of reasoning models and non-reasoning models. We evaluate 16 LLMs and 4 PRMs under identical experimental conditions and show that most of the LLMs' performance are far from robust and they face a certain level of data leakage. By dynamically generating OOD datasets, ThinkBench effectively provides a reliable evaluation of LLMs and reduces data contamination impact. The examples of our dataset is available at https://anonymous.4open.science/r/ThinkBench_Review/.



Paperid:2603
Authors:Mohamed Amine Ketata, David Lüdke, Leo Schwinn, Stephan Günnemann
Abstract:
Building generative models for relational databases (RDBs) is important for applications like privacy-preserving data release and augmenting real datasets. However, most prior work either focuses on single-table generation or relies on autoregressive factorizations that impose a fixed table order and generate tables sequentially. This approach limits parallelism, restricts flexibility in downstream applications like missing value imputation, and compounds errors due to commonly made conditional independence assumptions. We propose a fundamentally different approach: jointly modeling all tables in an RDB without imposing any order. By using a natural graph representation of RDBs, we propose the Graph-Conditional Relational Diffusion Model (GRDM). GRDM leverages a graph neural network to jointly denoise row attributes and capture complex inter-table dependencies. Extensive experiments on six real-world RDBs demonstrate that our approach substantially outperforms autoregressive baselines in modeling multi-hop inter-table correlations and achieves state-of-the-art performance on single-table fidelity metrics.



Paperid:2604
Authors:Yanjun Fu, Faisal Hamman, Sanghamitra Dutta
Abstract:
Instruction tuning is essential for Large Language Models (LLMs) to effectively follow user instructions. To improve training efficiency and reduce data redundancy, recent works use LLM-based scoring functions, e.g., Instruction-Following Difficulty (IFD), to select high–quality instruction-tuning data with scores above a threshold. While these data selection methods often lead to models that can match or even exceed the performance of models trained on the full datasets, we identify two key limitations: (i) they assess quality at the sample level, ignoring token-level informativeness; and (ii) they overlook the robustness of the scoring method, often selecting a sample due to superficial lexical features instead of its true quality. In this work, we propose Token-Selective HIeRarchical Data Selection for Instruction Tuning (T-SHIRT), a novel data selection framework that introduces a new scoring method to include only informative tokens in quality evaluation and also promote robust and reliable samples whose neighbors also show high quality with less local inconsistencies. We demonstrate that models instruction-tuned on a curated dataset (only 5% of the original size) using T-SHIRT can outperform those trained on the entire large-scale dataset by up to 5.48 points on average across eight benchmarks. Across various LLMs and training set scales, our method consistently surpasses existing state-of-the-art data selection techniques, while also remaining both cost-effective and highly efficient. For instance, by using GPT-2 for score computation, we are able to process a dataset of 52k samples using 40 minutes on a single GPU.



Paperid:2605
Authors:Ruili Feng, Han Zhang, Zhilei Shu, Zhantao Yang, Longxiang Tang, Zhicai Wang, Andy Zheng, Jie Xiao, Zhiheng Liu, Ruihang Chu, Yukun Huang, Yu Liu, Hongyang Zhang
Abstract:
We present The Matrix, a foundational realistic world simulator capable of generating infinitely long 720p high-fidelity real-scene video streams with real-time, responsive control in both first- and third-person perspectives. Trained on limited supervised data from video games like Forza Horizon 5 and Cyberpunk 2077, complemented by large-scale unsupervised footage from real-world settings like Tokyo streets, The Matrix allows users to traverse diverse terrains—deserts, grasslands, water bodies, and urban landscapes—in continuous, uncut hour-long sequences. With speeds of up to 16 FPS, the system supports real-time interactivity and demonstrates zero-shot generalization, translating virtual game environments to real-world contexts where collecting continuous movement data is often infeasible. For example, The Matrix can simulate a BMW X3 driving through an office setting—an environment present in neither gaming data nor real-world sources. This approach showcases the potential of game data to advance robust world models, bridging the gap between simulations and real-world applications in scenarios with limited data.



Paperid:2606
Authors:Xabier de Juan, Santiago Mazuelas
Abstract:
Abstract:The Median-of-Means (MoM) is a robust estimator widely used in machine learning that is known to be (minimax) optimal in scenarios where samples are i.i.d. In more grave scenarios, samples are contaminated by an adversary that can inspect and modify the data. Previous work has theoretically shown the suitability of the MoM estimator in certain contaminated settings. However, the (minimax) optimality of MoM and its limitations under adversarial contamination remain unknown beyond the Gaussian case. In this paper, we present upper and lower bounds for the error of MoM under adversarial contamination for multiple classes of distributions. In particular, we show that MoM is (minimax) optimal in the class of distributions with finite variance, as well as in the class of distributions with infinite variance and finite absolute $(1+r)$-th moment. We also provide lower bounds for MoM's error that match the order of the presented upper bounds, and show that MoM is sub-optimal for light-tailed distributions.



Paperid:2607
Authors:Jingjun Yi, Hao Zheng, Qi Bi, Huimin Huang, Yixian Shen, Haolan Zhan, Wei Ji, Yawen Huang, Yuexiang Li, Xian Wu, Yefeng Zheng
Abstract:
Image restoration is a fundamental task in computer vision and machine learning, which learns a mapping between the clear images and the degraded images from various conditions (e.g., blur, low-light, haze).Yet, most existing image restoration methods are highly restricted by the requirement of degraded and clear image pairs, which limits the generalization and feasibility to enormous real-world scenarios without paired images.To address this bottleneck, we propose a Degradation-aware Dynamic Schr\"{o}dinger Bridge (DDSB) for unpaired image restoration.Its general idea is to learn a Schr\"{o}dinger Bridge between clear image distribution and degraded image distribution, while at the same time emphasizing the physical degradation priors to reduce the accumulation of errors during the image transformation process. A Degradation-aware Optimal Transport (DOT) learning scheme is accordingly devised.Training a degradation model to learn the inverse restoration process is particularly challenging, as it must be applicable across different stages of the iterative restoration process.A Dynamic Transport with Consistency (DTC) learning objective is further proposed to reduce the loss of image details in the early iterations and refine the learning objective of the degradation model.Extensive experiments on multiple image degradation tasks show its state-of-the-art performance over the prior arts. Source code will be publicly available.



Paperid:2608
Authors:Junyu guo, Zhi Zheng, Donghao Ying, Ming Jin, Shangding Gu, Costas J Spanos, Javad Lavaei
Abstract:
Constrained reinforcement learning (RL) seeks high-performance policies under safety constraints. We focus on an offline setting where the agent has only a fixed dataset---common in realistic tasks to prevent unsafe exploration. To address this, we propose Diffusion-Regularized Constrained Offline Reinforcement Learning (DRCORL), which first uses a diffusion model to capture the behavioral policy from offline data and then extracts a simplified policy to enable efficient inference. We further apply gradient manipulation for safety adaptation, balancing the reward objective and constraint satisfaction. This approach leverages high-quality offline data while incorporating safety requirements. Empirical results show that DRCORL achieves reliable safety performance, fast inference, and strong reward outcomes across robot learning tasks. Compared to existing safe offline RL methods, it consistently meets cost limits and performs well with the same hyperparameters, indicating practical applicability in real-world scenarios.



Paperid:2609
Authors:Anders Gjølbye, Stefan Haufe, Lars Kai Hansen
Abstract:
Suppressor variables can influence model predictions without being dependent on the target outcome and they pose a significant challenge for Explainable AI (XAI) methods. These variables may cause false-positive feature attributions, undermining the utility of explanations. Although effective remedies exist for linear models, their extension to non-linear models and to instance-based explanations has remained limited. We introduce PatternLocal, a novel XAI technique that addresses this gap. PatternLocal begins with a locally linear surrogate, e.g. LIME, KernelSHAP, or gradient-based methods, and transforms the resulting discriminative model weights into a generative representation, thereby suppressing the influence of suppressor variables while preserving local fidelity. In extensive hyperparameter optimization on the XAI-TRIS benchmark, PatternLocal consistently outperformed other XAI methods and reduced false-positive attributions when explaining non-linear tasks, thereby enabling more reliable and actionable insights.



Paperid:2610
Authors:Wenxuan Shi, Haochen Tan, Chuqiao Kuang, Xiaoguang Li, Hanting Chen, Xiaozhe Ren, Yasheng Wang, Lu Hou, Lifeng Shang
Abstract:
Abstract:Information seeking demands iterative evidence gathering and reflective reasoning, yet large language models (LLMs) still struggle with it in open-web question answering. Existing prompting and supervised fine-tuning (SFT) methods remain fixed by prompt rules or training corpora, and are usually benchmarked only on well-structured wiki sources, limiting real-world adaptability. We introduce $\textbf{WebPuzzle}$, a 24k-sample training and 275-sample test benchmark that evaluates information seeking on the live internet, across both wiki and open-domain queries. Leveraging 7k WebPuzzle instances, we develop $\textbf{DeepDiver}$, a reinforcement-learning (RL) framework that cultivates $\textbf{Search Intensity Scaling (SIS)}$—an emergent ability to escalate search frequency and depth instead of settling on overconfident, under-evidenced answers. With SIS, Qwen2.5-7B-Instruct and Pangu-7B-Reasoner attain performance on real-web tasks comparable to the 671B-parameter DeepSeek-R1. We detail DeepDiver’s curriculum from cold-start SFT to a well designed RL procedure, and show that its seeking policy generalized from closed-ended queries to open-ended generation such as long-form writing. Our results advance adaptive information seeking in LLMs and provide a rigorous benchmark for future work.



Paperid:2611
Authors:Yihan Wu, Georgios Milis, Ruibo Chen, Heng Huang
Abstract:
The rapid advancement of next-token-prediction models has led to widespread adoption across modalities, enabling the creation of realistic synthetic media. In the audio domain, while autoregressive speech models have propelled conversational interactions forward, the potential for misuse, such as impersonation in phishing schemes or crafting misleading speech recordings, has also increased. Security measures such as watermarking have thus become essential to ensuring the authenticity of digital media. Traditional statistical watermarking methods used for autoregressive language models face challenges when applied to autoregressive audio models, due to the inevitable ``retokenization mismatch'' - the discrepancy between original and retokenized discrete audio token sequences. To address this, we introduce Aligned-IS, a novel, distortion-free watermark, specifically crafted for audio generation models. This technique utilizes a clustering approach that treats tokens within the same cluster equivalently, effectively countering the retokenization mismatch issue. Our comprehensive testing on prevalent audio generation platforms demonstrates that Aligned-IS not only preserves the quality of generated audio but also significantly improves the watermark detectability compared to the state-of-the-art distortion-free watermarking adaptations, establishing a new benchmark in secure audio technology applications.



Paperid:2612
Authors:Elif Arslan, Jacobus van der Linden, Serge Hoogendoorn, Marco Rinaldi, Emir Demirović
Abstract:
Sparse decision tree learning provides accurate and interpretable predictive models that are ideal for high-stakes applications by finding the single most accurate tree within a (soft) size limit. Rather than relying on a single “best” tree, Rashomon sets—trees with similar performance but varying structures—can be used to enhance variable importance analysis, enrich explanations, and enable users to choose simpler trees or those that satisfy stakeholder preferences (e.g., fairness) without hard-coding such criteria into the objective function. However, since finding the optimal tree is NP-hard, the scalability of enumerating the Rashomon set is challenging. Therefore, we introduce SORTD, a novel framework that improves scalability and enumerates Rashomon sets in order of the objective value, thus offering anytime behavior. Our experiments show that SORTD reduces runtime by up to two orders of magnitude compared with the state of the art. Moreover, SORTD can compute Rashomon sets for any totally ordered objective and supports post-evaluating the set using other (partially ordered) objectives. Together, these advances make exploring Rashomon sets more practical in real-world applications.



Paperid:2613
Authors:Jialin Zhao, Alessandro Muscoloni, Umberto Michieli, Yingtao Zhang, Carlo Cannistraci
Abstract:
Many complex networks have a connectivity that might be only partially detected or that tends to grow over time, hence the prediction of non-observed links is a fundamental problem in network science. The aim of topological link prediction is to forecast these non-observed links by only exploiting features intrinsic to the network topology. It has a wide range of real applications, like suggesting friendships in social networks or predicting interactions in biological networks. The Cannistraci-Hebb theory is a recent achievement in network science that includes a theoretical framework to understand local-based link prediction on paths of length n. In this study we introduce two innovations: one on the theory of modelling (science) and the other on the theory of realization (engineering). For the theory of modelling we first recall a definition of network automata as a general framework for modelling the growth of connectivity in complex networks. We then show that several deterministic models previously developed fall within this framework and we introduce novel network automata following the Cannistraci-Hebb rule. For the theory of realization, we present how to build adaptive network automaton for link prediction, which incorporate multiple deterministic models of self-organization and automatically learns the rule that better explains the patterns of connectivity in the network under investigation. We compare the Cannistraci-Hebb Adaptive (CHA) network automaton against state-of-the-art link prediction methods, including structural perturbation method (SPM), stochastic block models (SBM), and artificial intelligence algorithms for graph representation learning, such as graph embedding methods and message-passing-based models. In particular, we compare the Message Passing Link Predictor (MPLP), a state-of-the-art link prediction method that differs from general-purpose graph embedding methods by approximating heuristic scores such as Common Neighbor through quasi-orthogonal message-passing. We also compare with MPLP+, a scalable variant that avoids costly preprocessing by leveraging only one-hop neighborhoods. CHA displays an overall higher link prediction performance across different evaluation frameworks on 1269 networks across 14 complex systems domains. Finally, we highlight that CHA offers the key advantage to explicitly explain the mechanistic rule of self-organization which leads to the link prediction performance, whereas SPM, SBM, and AI-based methods like graph embeddings and MPLP do not. In comparison to CHA, SBM unfortunately shows irrelevant and unsatisfactory performance demonstrating that SBM modelling is not adequate for link prediction in real networks.



Paperid:2614
Authors:Hanyue Lou, Jinxiu (Sherry) Liang, Minggui Teng, Yi Wang, Boxin Shi
Abstract:
Event-based cameras offer unique advantages such as high temporal resolution, high dynamic range, and low power consumption. However, the massive storage requirements and I/O burdens of existing synthetic data generation pipelines and the scarcity of real data prevent event-based training datasets from scaling up, limiting the development and generalization capabilities of event vision models. To address this challenge, we introduce Video-to-Voxel (V2V), an approach that directly converts conventional video frames into event-based voxel grid representations, bypassing the storage-intensive event stream generation entirely. V2V enables a 150× reduction in storage requirements while supporting on-the-fly parameter randomization for enhanced model robustness. Leveraging this efficiency, we train several video reconstruction and optical flow estimation model architectures on 10,000 diverse videos totaling 52 hours—an order of magnitude larger than existing event datasets, yielding substantial improvements.



Paperid:2615
Authors:Yang Zheng, Mengqi Huang, Nan Chen, Zhendong Mao
Abstract:
Text-guided 3D editing aims to locally modify 3D objects based on editing prompts, which has significant potential for applications in 3D game and film domains. Existing methods typically follow a view-agnostic paradigm: editing 2D view images indiscriminately and projecting them back into 3D space. However, the view-agnostic paradigm neglects view consistency and view-specific characteristics, resulting in spatial inconsistencies and imprecise control over edited regions. In this study, we argue that progressive view-oriented paradigm can effectively address these issues, which projects the editing information from a editing-sensitive view to other editing-insensitive views. Based on this paradigm, we design Pro3D-Editor, a new framework. Extensive experiments demonstrate that our method outperforms existing approaches in terms of editing accuracy and spatial consistency.



Paperid:2616
Authors:Nedko Savov, Naser Kazemi, Deheng Zhang, Danda Pani Paudel, Xi Wang, Luc V Gool
Abstract:
World models have recently become promising tools for learning to simulate environments and produce realistic visuals based on actions in complex settings. However, because they rely on short observation sequences, they quickly lose track of context. As a result, visual consistency breaks down after just a few steps, and generated scenes no longer reflect information seen earlier. This problem comes from a common design choice in state-of-the-art diffusion-based models: they do not keep track of a lasting environment state. To address this problem, we introduce StateSpaceDiffuser, where a diffusion model is enabled to perform on long-context tasks by integrating a rich representation from a state-space model (Mamba). The state-space branch maintains a compact latent summarizing the entire interaction history, while the diffusion branch conditions on this latent to render context-aware future frames. This design restores long-term memory without sacrificing the high-fidelity synthesis of diffusion models.To rigorously measure temporal consistency, we develop an evaluation protocol that probes a model’s ability to remember and re-instantiate previously seen content during extended rollouts. Through comprehensive experiments, we show that the StateSpaceDiffuser significantly outperforms a strong diffusion-only baseline, maintaining coherent visual context for an order of magnitude more steps. It delivers consistent views in both a 2D maze navigation and a complex 3D environment. These results establish that bringing state-space representations into diffusion models is highly effective in both producing visually detailed and capable of long-term memory.



Paperid:2617
Authors:Zhijian Zhou, Xunye Tian, Liuhua Peng, Chao Lei, Antonin Schrab, Danica J. Sutherland, Feng Liu
Abstract:
To adapt kernel two-sample and independence testing to complex structured data, aggregation of multiple kernels is frequently employed to boost testing power compared to single-kernel tests. However, we observe a phenomenon that directly maximizing multiple kernel-based statistics may result in highly similar kernels that capture highly overlapping information, limiting the effectiveness of aggregation. To address this, we propose an aggregated statistic that explicitly incorporates kernel diversity based on the covariance between different kernels. Moreover, we identify a fundamental challenge: a trade-off between the diversity among kernels and the test power of individual kernels, i.e., the selected kernels should be both effective and diverse. This motivates a testing framework with selection inference, which leverages information from the training phase to select kernels with strong individual performance from the learned diverse kernel pool. We provide rigorous theoretical statements and proofs to show the consistency on the test power and control of Type-I error, along with asymptotic analysis of the proposed statistics. Lastly, we conducted extensive empirical experiments demonstrating the superior performance of our proposed approach across various benchmarks for both two-sample and independence testing.



Paperid:2618
Authors:Kai He, Ruofan Liang, Jacob Munkberg, Jon Hasselgren, Nandita Vijaykumar, Alexander Keller, Sanja Fidler, Igor Gilitschenski, Zan Gojcic, Zian Wang
Abstract:
We address the challenge of relighting a single image or video, a task that demands precise scene intrinsic understanding and high-quality light transport synthesis. Existing end-to-end relighting models are often limited by the scarcity of paired multi-illumination data, restricting their ability to generalize across diverse scenes. Conversely, two-stage pipelines that combine inverse and forward rendering can mitigate data requirements but are susceptible to error accumulation and often fail to produce realistic outputs under complex lighting conditions or with sophisticated materials. In this work, we introduce a general-purpose approach that jointly estimates albedo and synthesizes relit outputs in a single pass, harnessing the generative capabilities of video diffusion models. This joint formulation enhances implicit scene comprehension and facilitates the creation of realistic lighting effects and intricate material interactions, such as shadows, reflections, and transparency. Trained on synthetic multi-illumination data and extensive automatically labeled real-world videos, our model demonstrates strong generalization across diverse domains and surpasses previous methods in both visual fidelity and temporal consistency.



Paperid:2619
Authors:Yiyi Liu, Chunyang Liu, Bohan Wang, Weiqin Jiao, Bojian Wu, Lubin Fan, Yuwei Chen, Fashuai Li, Biao Xiong
Abstract:
We present CAGE (Continuity-Aware edGE) network, an end-to-end framework for reconstructing vector floorplans directly from point-cloud density maps. Traditional corner-based polygon representations are highly sensitive to noise and incomplete observations, often resulting in fragmented or implausible layouts. Recent line grouping methods leverage structural cues to improve robustness but still struggle to recover fine geometric details. To address these limitations, we propose a native edge-centric formulation, modeling each wall segment as a directed, geometrically continuous edge. This representation enables inference of coherent floorplan structures, ensuring watertight, topologically valid room boundaries while improving robustness and reducing artifacts. Towards this design, we develop a dual-query transformer decoder that integrates perturbed and latent queries within a denoising framework, which not only stabilizes optimization but also accelerates convergence. Extensive experiments on Structured3D and SceneCAD show that CAGE achieves state-of-the-art performance, with F1 scores of 99.1% (rooms), 91.7% (corners), and 89.3% (angles). The method also demonstrates strong cross-dataset generalization, underscoring the efficacy of our architectural innovations. Code and pretrained models will be released upon acceptance.



Paperid:2620
Authors:Alexander Kozachinskiy, Felipe Urrutia, Hector Orellana, Tomasz Steifer, Germán Pizarro, Matías Álvarez, Francisco Vásquez, Cristian Calderon, Cristobal Rojas
Abstract:
We propose the first method to show theoretical limitations for one-layer softmax transformers with arbitrarily many precision bits (even infinite). We establish those limitations for three tasks that require advanced reasoning. The first task, Match 3 (Sanford et al., 2023), requires looking at all possible token triplets in an input sequence. The second and third tasks address compositionality-based reasoning: function composition (Peng et al., 2024) and binary relations composition, respectively. We formally prove the inability of one-layer softmax Transformers to solve any of these tasks. To overcome these limitations, we introduce Strassen attention and prove that, equipped with this mechanism, a one-layer transformer can in principle solve all these tasks. Importantly, we show that it enjoys sub-cubic running-time complexity, making it more scalable than similar previously proposed mechanisms, such as higher-order attention (Sanford et al., 2023). To complement our theoretical findings, we experimentally studied Strassen attention and compared it against standard (Vaswani et al, 2017), higher-order attention (Sanford et al., 2023), and triangular attention (Bergen et al. 2021). Our results help to disentangle all these attention mechanisms, highlighting their strengths and limitations. In particular, Strassen attention outperforms standard attention significantly on all the tasks. Altogether, understanding the theoretical limitations can guide research towards scalable attention mechanisms that improve the reasoning abilities of Transformers.



Paperid:2621
Authors:Shicheng Liu, Siyuan Xu, Wenjie Qiu, Hangfan Zhang, Minghui Zhu
Abstract:
A common and effective strategy for humans to improve an unsatisfactory outcome in daily life is to find a cause of this outcome and correct the cause. In this paper, we investigate whether this human improvement strategy can be applied to improving reinforcement learning from human feedback (RLHF) for alignment of language models (LMs). In particular, it is observed in the literature that LMs tuned by RLHF can still output unsatisfactory responses. This paper proposes a method to improve the unsatisfactory responses by correcting their causes. Our method has two parts. The first part proposes a post-hoc explanation method to explain why an unsatisfactory response is generated to a prompt by identifying the training data that lead to this response. We formulate this problem as a constrained combinatorial optimization problem where the objective is to find a set of training data closest to this prompt-response pair in a feature representation space, and the constraint is that the prompt-response pair can be decomposed as a convex combination of this set of training data in the feature space. We propose an efficient iterative data selection algorithm to solve this problem. The second part proposes an unlearning method that improves unsatisfactory responses to some prompts by unlearning the training data that lead to these unsatisfactory responses and, meanwhile, does not significantly degrade satisfactory responses to other prompts. Experimental results demonstrate that our algorithm can improve RLHF.



Paperid:2622
Authors:Nikolaos Karalias, Akbar Rafiey, Yifei Xu, Zhishang Luo, Behrooz Tahmasebi, Connie Jiang, Stefanie Jegelka
Abstract:
Self-Supervised Learning (SSL) for Combinatorial Optimization (CO) is an emerging paradigm for solving combinatorial problems using neural networks. In this paper, we address a central challenge of SSL for CO: solving problems with discrete constraints. We design an end-to-end differentiable framework that enables us to solve discrete constrained optimization problems with neural networks. Concretely, we leverage algorithmic techniques from the literature on convex geometry and Carath\'eodory's theorem to decompose neural network outputs into convex combinations of polytope corners that correspond to feasible sets. This decomposition-based approach enables self-supervised training but also ensures efficient quality-preserving rounding of the neural net output into feasible solutions. Extensive experiments in cardinality-constrained optimization show that our approach can consistently outperform neural baselines. We further demonstrate that our method generalizes beyond cardinality-constrained problems to a diverse set of combinatorial optimization tasks, including finding independent sets in graphs, and solving matroid-constrained problems.



Paperid:2623
Authors:Junhao Shi, Zhaoye Fei, Siyin Wang, Qipeng Guo, Jingjing Gong, Xipeng Qiu
Abstract:
Large Vision-Language Models (LVLMs) show promise for embodied planning tasks but struggle with complex scenarios involving unfamiliar environments and multi-step goals. Current approaches rely on environment-agnostic imitation learning that disconnects instructions from environmental contexts, causing models to struggle with context-sensitive instructions and rely on supplementary cues rather than visual reasoning during long-horizon interactions.In this work, we propose World-Aware Planning Narrative Enhancement (WAP), a framework that infuses LVLMs with comprehensive environmental understanding through four cognitive capabilities (visual appearance modeling, spatial reasoning, functional abstraction, and syntactic grounding) while developing and evaluating models using only raw visual observations through curriculum learning.Evaluations on the EB-ALFRED benchmark demonstrate substantial improvements, with Qwen2.5-VL achieving a 60.7 absolute improvement in task success rates—particularly in commonsense reasoning (+60.0) and long-horizon planning (+70.0). Notably, our enhanced open-source models outperform proprietary systems like GPT-4o and Claude-3.5-Sonnet by a large margin.



Paperid:2624
Authors:Sagnik Mukherjee, Lifan Yuan, Dilek Tur, Hao Peng
Abstract:
Reinforcement learning (RL) yields substantial improvements in large language models’ (LLMs) downstream task performance and alignment with human values. Surprisingly, such large gains result from updating only a small subnetwork comprising just 5%-30% of the parameters, with the rest effectively unchanged. We refer to this phenomenon as parameter update sparsity induced by RL. It is observed across all 7 widely-used RL algorithms (e.g., PPO, GRPO, DPO) and all 10 LLMs from different families in our experiments.This sparsity is intrinsic and occurs without any explicit sparsity-promoting regularizations or architectural constraints. Finetuning the subnetwork alone recovers the test accuracy, and, remarkably, produces a model nearly identical to the one obtained via full finetuning.The subnetworks from different random seeds, training data, and even RL algorithms show substantially greater overlap than expected by chance. Our analysis suggests that this sparsity is not due to updating only a subset of layers; instead, nearly all parameter matrices receive similarly sparse updates. Moreover, the updates to almost all parameter matrices are nearly full-rank,suggesting RL updates a small subset of parameters that nevertheless span almost the full subspaces that the parameter matrices can represent. We conjecture that the this update sparsity can be primarily attributed to training on data that is near the policy distribution; techniques that encourage the policy to remain close to the pretrained model, such as the KL regularization and gradient clipping, have limited impact.



Paperid:2625
Authors:Shriram M S, Xinyue Hao, Shihao Hou, Yang Lu, Laura Sevilla-Lara, Anurag Arnab, Shreyank Gowda
Abstract:
Training deep neural networks is computationally expensive, often requiring significant time and energy to converge on large datasets. In this paper, we propose Progressive Data Dropout, an embarrassingly simple yet effective approach to improve training efficiency by progressively discarding subsets of data across epochs. We explore three variants: (1) a curriculum-inspired method that initially focuses on hard examples—those misclassified or predicted with low confidence; (2) a scalar-based decay method that randomly drops a fixed proportion of data each epoch; and (3) a hybrid approach that mimics the schedule of the first method but drops data at random rather than based on difficulty. All three approaches use the full training dataset only in the last epochs. Despite its simplicity, the third variant achieves the best performance in our experiments. Remarkably, our approach reduces the number of effective epochs to as little as 12.4\% of the baseline measured in 'Effective Epochs', a hardware-independent proxy for backpropagation effort while improving accuracy by up to 4.82\%, depending on the model and dataset. Our approach requires no changes to model architecture or optimizer, and can be applied across standard training pipelines. These results demonstrate that carefully designed data dropout strategies can substantially reduce training costs while enhancing generalization. Code: https://anonymous.4open.science/r/LearningWithRevision-1B1B.



Paperid:2626
Authors:Haolin Li, Tianjie Dai, Zhe Chen, Siyuan Du, Jiangchao Yao, Ya Zhang, Yanfeng Wang
Abstract:
Clinical diagnosis is a highly specialized discipline requiring both domain expertise and strict adherence to rigorous guidelines. While current AI-driven medical research predominantly focuses on knowledge graphs or natural text pretraining paradigms to incorporate medical knowledge, these approaches primarily rely on implicitly encoded knowledge within model parameters, neglecting task-specific knowledge required by diverse downstream tasks.To address this limitation, we proposeRetrieval-AugmentedDiagnosis (RAD), a novel framework that explicitly injects external knowledge into multimodal models directly on downstream tasks.Specifically, RAD operates through three key mechanisms: retrieval and refinement of disease-centered knowledge from multiple medical sources, a guideline-enhanced contrastive loss that constrains the latent distance between multi-modal features and guideline knowledge, and the dual diagnostic transformer decoder that employs guidelines as queries to steer cross-modal fusion, aligning the models with clinical diagnostic workflows from guideline acquisition to feature extraction and decision-making.Moreover, recognizing the lack of quantitative evaluation of interpretability for multimodal diagnostic models, we introduce a set of criteria to assess the interpretability from both image and text perspectives. Extensive evaluations across four datasets with different anatomies demonstrate RAD's generalizability, achieving state-of-the-art performance. Furthermore, RAD enables the model to concentrate more precisely on abnormal regions and critical indicators, ensuring evidence-based, trustworthy diagnosis. Our code is available at this https://anonymous.4open.science/r/RAD_anonymous/.



Paperid:2627
Authors:Jang Hyun Cho, Andrea Madotto, Effrosyni Mavroudi, Triantafyllos Afouras, Tushar Nagarajan, Muhammad Maaz, Yale Song, Tengyu Ma, Shuming Hu, Suyog Jain, Miguel Martin, Huiyu Wang, Hanoona Bangalath, Peize Sun, Po-Yao Huang, Daniel Bolya, Nikhila Ravi, Shashank Jain, Tammy Stark, Seungwhan Moon, Babak Damavandi, Vivian Lee, Andrew Westbury, Salman Khan, Philipp Kraehenbuehl, Piotr Dollar, Lorenzo Torresani, Kristen Grauman, Christoph Feichtenhofer
Abstract:
Vision-language models are integral to computer vision research, yet many high-performing models remain closed-source, obscuring their data, design and training recipe. The research community has responded by using distillation from black-box models to label training data, achieving strong benchmark results, at the cost of measurable scientific progress. However, without knowing the details of the teacher model and its data sources, scientific progress remains difficult to measure. In this paper, we study building a Perception Language Model (PLM) in a fully open and reproducible framework for transparent research in image and video understanding. We analyze standard training pipelines without distillation from proprietary models and explore large-scale synthetic data to identify critical data gaps, particularly in detailed video understanding. To bridge these gaps, we release 2.8M human-labeled instances of fine-grained video question-answer pairs and spatio-temporally grounded video captions. Additionally, we introduce PLM–VideoBench, a suite for evaluating challenging video understanding tasks focusing on the ability to reason about ''what'', ''where'', ''when'', and ''how'' of a video. We make our work fully reproducible by providing data, training recipes, code & models.



Paperid:2628
Authors:Xixian Liu, Rui Jiao, ZHIYUAN LIU, Yurou Liu, Yang Liu, Ziheng Lu, Wenbing Huang, Yang Zhang, Yixin Cao
Abstract:
Coordinate denoising has emerged as a promising method for 3D molecular pretraining due to its theoretical connection to learning molecular force field. However, existing denoising methods rely on oversimplied molecular dynamics that assume atomic motions to be isotropic and homoscedastic. To address these limitations, we propose a novel denoising framework AniDS: Anisotropic Variational Autoencoder for 3D Molecular Denoising. AniDS introduces a structure-aware anisotropic noise generator that can produce atom-specific, full covariance matrices for Gaussian noise distributions to better reflect directional and structural variability in molecular systems. These covariances are derived from pairwise atomic interactions as anisotropic corrections to an isotropic base. Our design ensures that the resulting covariance matrices are symmetric, positive semi-definite, and SO(3)-equivariant, while providing greater capacity to model complex molecular dynamics. Extensive experiments show that AniDS outperforms prior isotropic and homoscedastic denoising models and other leading methods on the MD17 and OC22 benchmarks, achieving average relative improvements of 8.9% and 6.2% in force prediction accuracy. Our case study on a crystal structure shows that AniDS adaptively suppresses noise along bonding direction, consistent with physicochemical principles. Our code is available at https://anonymous.4open.science/r/AniDS.



Paperid:2629
Authors:Ruimeng Liu, Xin Zou, Chang Tang, Xiao Zheng, Xingchen Hu, Kun Sun, Xinwang Liu
Abstract:
Existing multi-view clustering methods employ various strategies to address dataset-level sparsity and view-level dynamic fusion. However, we identify a critical yet overlooked issue: the varying sparsity across views. Cross-view sparsity variations lead to encoding discrepancies, heightening sample-level semantic heterogeneity and making view-level dynamic weighting inappropriate. To tackle these challenges, we propose Adaptive Sparse Autoencoders for Multi-View Clustering (SparseMVC), a framework with three key modules. Initially, the Sparse Autoencoder probes the sparsity of each view and adaptively adjusts encoding formats via an entropy-matching loss term, mitigating cross-view inconsistencies. Subsequently, the Correlation-Informed Sample Reweighting module employs attention mechanisms to assign weights by capturing correlations between early-fused global and view-specific features, reducing encoding discrepancies and balancing contributions. Furthermore, the Cross-view Distribution Alignment module aligns feature distributions during the late fusion stage, accommodating datasets with an arbitrary number of views. Extensive experiments demonstrate that our proposed method achieves state-of-the-art clustering performance. Overall, our framework advances the field by extending sparsity handling from the dataset level to the view level and mitigating the adverse effects of encoding discrepancies through sample-level dynamic weighting. The source code will be public upon acceptance.



Paperid:2630
Authors:Ming Liu, Hao Chen, Jindong Wang, Liwen Wang, Bhiksha Raj, Wensheng Zhang
Abstract:
Unified multimodal large language models (U-MLLMs) have demonstrated impressive performance in end-to-end visual understanding and generation. However, compared to generation-only systems (e.g., Stable Diffusion), their unified architecture introduces new risks of propagating demographic stereotypes. In this paper, we benchmark several state-of-the-art U-MLLMs and show that they exhibit significant gender and race biases in their generated outputs. To diagnose the source of these biases, we propose a \emph{locate-then-fix} framework: we first audit the vision and language components—using techniques such as linear probing and controlled generation—and find that the language model appears to be a primary origin of the observed generative bias. Moreover, we observe a ``partial alignment'' phenomenon, where U-MLLMs exhibit less bias in understanding tasks yet produce substantially biased images. To address this, we introduce a novel \emph{balanced preference loss} that enforces uniform generation probabilities across demographics by leveraging a synthetically balanced dataset. Extensive experiments demonstrate that our approach significantly reduces demographic bias while preserving semantic fidelity and image quality. Our findings underscore the need for targeted debiasing strategies in unified multimodal systems and provide a practical method for mitigating bias.



Paperid:2631
Authors:Dongheon Lee, Younghoo Kwon, Jung-Woo Choi
Abstract:
We propose DeepASA, a one-for-all model for auditory scene analysis that performs multi-input multi-output (MIMO) source separation, dereverberation, sound event detection (SED), audio classification, and direction-of-arrival estimation (DoAE) within a unified framework. DeepASA is designed for complex auditory scenes where multiple, often similar, sound sources overlap in time and move dynamically in space. To achieve robust and consistent inference across tasks, we introduce an object-oriented processing (OOP) strategy. This approach encapsulates diverse auditory features into object-centric representations and refines them through a chain-of-inference (CoI) mechanism. The pipeline comprises a dynamic temporal kernel-based feature extractor, a transformer-based aggregator, and an object separator that yields per-object features. These features feed into multiple task-specific decoders. Our object-centric representations naturally resolve the parameter association ambiguity inherent in traditional track-wise processing. However, early-stage object separation can lead to failure in downstream ASA tasks. To address this, we implement temporal coherence matching (TCM) within the chain-of-inference, enabling multi-task fusion and iterative refinement of object features using estimated auditory parameters. We evaluate DeepASA on representative spatial audio benchmark datasets, including ASA2, MC-FUSS, and STARSS23. Experimental results show that our model achieves state-of-the-art performance across all evaluated tasks, demonstrating its effectiveness in both source separation and auditory parameter estimation under diverse spatial auditory scenes. The demo video and samples are available at https://huggingface.co/spaces/AnonymousAuthor1/DeepASA.



Paperid:2632
Authors:Shuche Wang, Adarsh Barik, Peng Zhao, Vincent Tan
Abstract:
Abstract:We develop the first parameter-free algorithms for the Stochastically Extended Adversarial (SEA) model, a framework that bridges adversarial and stochastic online convex optimization. Existing approaches for the SEA model require prior knowledge of problem-specific parameters, such as the diameter of the domain $D$ and the Lipschitz constant of the loss functions $G$, which limits their practical applicability. Addressing this, we develop parameter-free methods by leveraging the Optimistic Online Newton Step (OONS) algorithm to eliminate the need for these parameters. We first establish a comparator-adaptive algorithm for the scenario with unknown domain diameter but known Lipschitz constant, achieving an expected regret bound of $\tilde{O}\big(\Vert u\Vert_2^2 + \Vert u\Vert_2(\sqrt{\sigma^2_{1:T}} + \sqrt{\Sigma^2_{1:T}})\big)$, where $u$ is the comparator vector and $\sigma^2_{1:T}$ and $\Sigma^2_{1:T}$ represent the cumulative stochastic variance and cumulative adversarial variation, respectively. We then extend this to the more general setting where both $D$ and $G$ are unknown, attaining the comparator- and Lipschitz-adaptive algorithm. Notably, the regret bound exhibits the same dependence on $\sigma^2_{1:T}$ and $\Sigma^2_{1:T}$, demonstrating the efficacy of our proposed methods even when both parameters are unknown in the SEA model.



Paperid:2633
Authors:Haohong Lin, Yunzhi Zhang, Wenhao Ding, Jiajun Wu, DING ZHAO
Abstract:
End-to-end (E2E) autonomous driving models have demonstrated strong performance in open-loop evaluations but often suffer from cascading errors and poor generalization in closed-loop settings. To address this gap, we propose Model-based Policy Adaptation (MPA), a general framework that enhances the robustness and safety of pretrained E2E driving agents during deployment. MPA first generates diverse counterfactual trajectories using a geometry-consistent simulation engine, exposing the agent to scenarios beyond the original dataset. Based on this generated data, MPA trains a diffusion-based policy adapter to refine the base policy’s predictions and a multi-step Q value model to evaluate long-term outcomes. At inference time, the adapter proposes multiple trajectory candidates, and the Q value model selects the one with the highest expected utility. Experiments on the nuScenes benchmark using a photorealistic closed-loop simulator demonstrate that MPA significantly improves performance across in-domain, out-of-domain, and safety-critical scenarios. We further investigate how the scale of counterfactual data and inference-time guidance strategies affect overall effectiveness.



Paperid:2634
Authors:Jueun Ko, Hyewon Park, Hyesong Choi, Dongbo Min
Abstract:
Stereo Depth Estimation in real-world environments poses significant challenges due to dynamic domain shifts, sparse or unreliable supervision, and the high cost of acquiring dense ground-truth labels. While recent Test-Time Adaptation (TTA) methods offer promising solutions, most rely on static target domain assumptions and input-invariant adaptation strategies, limiting their effectiveness under continual shifts. In this paper, we propose RobIA, a novel Robust, Instance-Aware framework for Continual Test-Time Adaptation (CTTA) in stereo depth estimation. RobIA integrates two key components: (1) Attend-and-Excite Mixture-of-Experts (AttEx-MoE), a parameter-efficient module that dynamically routes input to frozen experts via lightweight self-attention mechanism tailored to epipolar geometry, and (2) Robust AdaptBN Teacher, a PEFT-based teacher model that provides dense pseudo-supervision by complementing sparse handcrafted labels. This strategy enables input-specific flexibility, broad supervision coverage, improving generalization under domain shift. Extensive experiments demonstrate that RobIA achieves superior adaptation performance across dynamic target domains while maintaining computational efficiency.



Paperid:2635
Authors:Ting Sun, Penghan Wang, Fan Lai
Abstract:
Large language models (LLMs) have facilitated a wide range of applications with distinct service-level objectives (SLOs), from latency-sensitive online tasks like interactive chatbots to throughput-oriented offline workloads like document summarization. The existing deployment model, which dedicates machines to each workload, simplifies SLO management but often leads to poor resource utilization. This paper introduces HyGen, an interference-aware LLM serving system that enables efficient co-location of online and offline workloads while preserving latency requirements. HyGen incorporates two key innovations: (1) performance control mechanisms, including a latency predictor to estimate batch execution time and an SLO-aware profiler to quantify latency interference, and (2) SLO-aware offline scheduling policies that maximize serving throughput and prevent starvation, without compromising online serving latency. Our evaluation on production workloads shows that HyGen achieves up to 3.87x overall throughput and 5.84x offline throughput gains over online and hybrid serving baselines, respectively, while strictly satisfying latency SLOs.



Paperid:2636
Authors:Shuyuan Zhang, ChenHan Jiang, Zuoou Li, Jiankang Deng
Abstract:
Constructing structured 3D shapes is essential for gaming, virtual reality, and embodied AI. However, this process typically demands significant expertise and manual effort using conventional 3D modeling software. To make 3D content creation more accessible, we present ShapeCraft, a novel system that leverages large language model (LLM) agents to autonomously generate structured 3D shapes from natural language instructions. At the core of ShapeCraft is a graph-based procedural shape (GPS) representation, whose nodes store corresponding program snippets and layout information, facilitating efficient programmatic updates and flexible structure manipulation. This representation allows for post-modeling execution as structured geometry while preserving editability for artists. In the ShapeCraft workflow, LLM agents first hierarchically parses user's input and initialize GPS representation, then iteratively refines the procedural modeling and texturing steps to produce structured, textured, and interactive 3D assets. Our experiments show that ShapeCraft significantly outperforms existing LLM-based agents in generating geometrically accurate and semantically rich 3D models. Moreover, it achieves higher shape fidelity compared to optimization-based text-to-3D generation methods. We further demonstrate the versatility of ShapeCraft through examples of animated and user-customized editing, highlighting its potential for broader interactive applications.



Paperid:2637
Authors:Xiaoyang Xiao, Runzhao Yao, Zhiqiang Tian, Shaoyi Du
Abstract:
Self-supervised pre-training is essential for 3D point cloud representation learning, as annotating their irregular, topology-free structures is costly and labor-intensive. Masked autoencoders (MAEs) offer a promising framework but rely on explicit positional embeddings, such as patch center coordinates, which leak geometric information and limit data-driven structural learning. In this work, we propose Point-MaDi, a novel Point cloud Masked autoencoding Diffusion framework for pre-training that integrates a dual-diffusion pretext task into an MAE architecture to address this issue. Specifically, we introduce a center diffusion mechanism in the encoder, noising and predicting the coordinates of both visible and masked patch centers without ground-truth positional embeddings. These predicted centers are processed using a transformer with self-attention and cross-attention to capture intra- and inter-patch relationships. In the decoder, we design a conditional patch diffusion process, guided by the encoder's latent features and predicted centers to reconstruct masked patches directly from noise. This dual-diffusion design drives comprehensive global semantic and local geometric representations during pre-training, eliminating external geometric priors. Extensive experiments on ScanObjectNN, ModelNet40, ShapeNetPart, S3DIS, and ScanNet demonstrate that Point-MaDi achieves superior performance across downstream tasks, surpassing Point-MAE by 5.51\% on OBJ-BG, 5.17\% on OBJ-ONLY, and 4.34\% on PB-T50-RS for 3D object classification on the ScanObjectNN dataset.



Paperid:2638
Authors:Yujie Xing, Xiao Wang, Bin Wu, Hai Huang, Chuan Shi
Abstract:
Abstract:Graph Transformers (GTs) have emerged as a powerful paradigm for graph representation learning due to their ability to model diverse node interactions.However, existing GTs often rely on intricate architectural designs tailored to specific interactions, limiting their flexibly.To address this, we propose a unified hierarchical mask framework that reveals an underlying equivalence between model architecture and attention mask construction.This framework enables a consistent modeling paradigm by capturing diverse interactions through carefully designed attention masks.Theoretical analysis under this framework demonstrates that the probability of correct classification positively correlates with the receptive field size and label consistency, leading to a fundamental design principle:An effective attention mask should ensure both a sufficiently large receptive field and a high level of label consistency.While no single existing mask satisfies this principle across all scenarios, our analysis reveals that hierarchical masks offer complementary strengths—motivating their effective integration.Then, we introduce M$^3$Dphormer, a Mixture-of-Experts based Graph Transformer with Multi-Level Masking and Dual Attention Computation.M$^3$Dphormer incorporates three theoretically grounded hierarchical masks and employs a bi-level expert routing mechanism to adaptively integrate multi-level interaction information.To ensure scalability, we further introduce a dual attention computation scheme that dynamically switches between dense and sparse modes based on local mask sparsity.Extensive experiments across multiple benchmarks demonstrate that M$^3$Dphormer achieves state-of-the-art performance, validating the effectiveness of our unified framework and model design.



Paperid:2639
Authors:Mélodie Monod, Alessandro Micheli, Samir Bhatt
Abstract:
We introduceNeuralSurv, the first deep survival model to incorporate Bayesian uncertainty quantification. Our non‑parametric, architecture‑agnostic framework flexibly captures time‑varying covariate–risk relationships in continuous time via a novel two‑stage data‑augmentation scheme, for which we establish theoretical guarantees. For efficient posterior inference, we introduce a mean‑field variational algorithm with coordinate‑ascent updates that scale linearly in model size. By locally linearizing the Bayesian neural network, we obtain full conjugacy and derive all coordinate updates in closed form. In experiments,NeuralSurvdelivers superior calibration compared to state-of-the-art deep survival models, while matching or exceeding their discriminative performance across both synthetic benchmarks and real-world datasets. Our results demonstrate the value of Bayesian principles in data‑scarce regimes by enhancing model calibration and providing robust, well‑calibrated uncertainty estimates for the survival function.



Paperid:2640
Authors:Hanlin Wu, Yuxuan Song, Zhe Zhang, Zhilong Zhang, Hao Zhou, Wei-Ying Ma, Jingjing Liu
Abstract:
Designing functional proteins is a critical yet challenging problem due to the intricate interplay between backbone structures, sequences, and side-chains. Current approaches often decompose protein design into separate tasks, which can lead to accumulated errors, while recent efforts increasingly focus on all-atom protein design. However, we observe that existing all-atom generation approaches suffering from an information shortcut issue, where models inadvertently infer sequences from side-chain information, compromising their ability to accurately learn sequence distributions. To address this, we introduce a novel rationalized information flow strategy to eliminate the information shortcut. Furthermore, motivated by the advantages of Bayesian flows over differential equation–based methods, we propose the first Bayesian flow formulation for protein backbone orientations by recasting orientation modeling as an equivalent hyperspherical generation problem with antipodal symmetry. To validate, our method delivers consistently exceptional performance in both peptide and antibody design tasks.



Paperid:2641
Authors:Tianyu Lin, Xinran Li, Chuntung Zhuang, Qi Chen, Yuanhao Cai, Kai Ding, Alan Yuille, Zongwei Zhou
Abstract:
Widely adopted evaluation metrics for sparse-view CT reconstruction---such as Structural Similarity Index Measure and Peak Signal-to-Noise Ratio---prioritize pixel-wise fidelity but often fail to capture the completeness of critical anatomical structures, particularly small or thin regions that are easily missed. To address this limitation, we propose a suite of novel anatomy-aware evaluation metrics designed to assess structural completeness across anatomical structures, including large organs, small organs, intestines, and vessels. Building on these metrics, we introduce CARE, a Completeness-Aware Reconstruction Enhancement framework that incorporates structural penalties during training to encourage anatomical preservation of significant regions. CARE is model-agnostic and can be seamlessly integrated into both analytical reconstruction methods and modern learning-based methods, such as Neural Radiance Fields and Gaussian Splatting. When applied to these methods, CARE substantially improves structural completeness in reconstructed CT scans, yielding performance gains of up to +32\% for large organs, +22\% for small organs, +40\% for intestines, and +36\% for vessels. Code has been attached as supplementary material for peer review and will be made publicly available.



Paperid:2642
Authors:Sreevardhan Sirigiri, Christian Hughes, Ian Abraham, Fabio Ramos
Abstract:
Effective robotic exploration in continuous domains requires planning trajectories that maximize coverage over a predefined region. A recent development, Stein Variational Ergodic Search (SVES), proposed parallel ergodic exploration (a key approach within the field of robotic exploration), via Stein variational inference that computes a set of candidate trajectories approximating the posterior distribution over the solution space trajectories. While this approach leverages GPU parallelism well, the trajectories in the set might not be distinct enough, leading to a suboptimal set. In this paper, we propose two key methods to diversify the solution set of this approach.First, we leverage the signature kernel within the SVES framework, introducing a pathwise, sequence-sensitive interaction that preserves the Markovian structure of the trajectories and naturally spreads paths across distinct regions of the search space. Second, we propose a derivative-free evolution-strategy interpretation of SVES that exploits batched, GPU-friendly fitness evaluations and can be paired with approximate gradients whenever analytic gradients of the kernel are unavailable or computationally intractable. The resulting method both retains SVES’s advantages while diversifying the solution set and extending its reach to black-box objectives. Across planar forest search, 3D quadrotor coverage, and model-predictive control benchmarks, our approach consistently reduces ergodic cost and produces markedly richer trajectory sets than SVES without significant extra tuning effort.



Paperid:2643
Authors:Shaojie Zhang, Ke Chen
Abstract:
Constrained clustering integrates domain knowledge through pairwise constraints.However, existing deep constrained clustering (DCC) methods are either limited by anchors inherent in end-to-end modeling or struggle with learning discriminative Euclidean embedding, restricting their scalability and real-world applicability.To avoid their respective pitfalls, we propose a novel angular constraint embedding approach for DCC, termed SpherePair.Using the SpherePair loss with a geometric formulation,our method faithfully encodes pairwise constraints and leads to embeddings that are clustering-friendly in angular space, effectively separating representation learning from clustering. SpherePair preserves pairwise relations without conflicts,requires no exact cluster number for constraint embedding, generalizes to unseen data, and is supported by rigorous theoretical guarantees.Comparative evaluations with state-of-the-art DCC methods on diverse benchmarks, along with empirical validation of theoretical insights, confirm its superior performance, scalability, and overall real-world effectiveness.



Paperid:2644
Authors:Timo Kaufmann, Yannick Metz, Daniel Keim, Eyke Hüllermeier
Abstract:
Standard reinforcement learning from human feedback (RLHF) uses binary preferences (which option is better), ignoring the strength of that preference (how much better).This preference strength, crucial for decision-making under uncertainty and generalization, is hard to measure reliably.While human response times (RTs) during preference elicitation offer valuable implicit signals of strength, raw RTs are noisy and obscured by individual and contextual factors.To address this challenge of learning preference strength from RTs, we propose RtRank, a novel framework that robustly extracts this strength.RtRank leveragesrelativeRT differenceswithin carefully constructed strata(e.g., per-annotator) to rank pairwise comparisons by their inferred preference strength.By controlling for systemic variation, these strata enable robust learning of utility differences consistent with the RT-derived rankings, all while making minimal assumptions.Our contributions are threefold:(1) RtRank, a novel method that robustly learns preference strength by leveraging intra-stratum relative RT rankings;(2) empirical evidence of improved sample efficiency and robustness in synthetic preference learning tasks; and(3) thePearson Distance Correlation (PDC), a novel metric that isolates cardinal utility learning from ordinal accuracy.



Paperid:2645
Authors:Yapei Chang, Yekyung Kim, Michael Krumdick, Amir Zadeh, Chuan Li, Chris Tanner, Mohit Iyyer
Abstract:
Reward models are central to aligning LLMs with human preferences, but they are costly to train, requiring large-scale human-labeled preference data and powerful pretrained LLM backbones. Meanwhile, the increasing availability of high-quality synthetic instruction-following datasets raises the question: can simpler, reference-based metrics serve as viable alternatives to reward models during RL-based alignment? In this paper, we show first that BLEU, a basic string-matching metric, surprisingly matches strong reward models in agreement with human preferences on general instruction-following datasets. Based on this insight, we develop BLEUBERI, a method that first identifies challenging instructions and then applies Group Relative Policy Optimization (GRPO) using BLEU directly as the reward function. We demonstrate that BLEUBERI-trained models are competitive with models trained via reward model-guided RL across four challenging instruction-following benchmarks and three different base language models. A human evaluation further supports that the quality of BLEUBERI model outputs is on par with those from reward model-aligned models. Moreover, BLEUBERI models generate outputs that are more factually grounded than competing methods. Overall, we show that given access to high-quality reference outputs (easily obtained via existing instruction-following datasets or synthetic data generation), string matching-based metrics are cheap yet effective proxies for reward models during alignment.



Paperid:2646
Authors:Yang Zhang, Xinran Li, Jianing Ye, Shuang Qiu, Delin Qu, Chongjie Zhang, Xiu Li, Chenjia Bai
Abstract:
World models have recently attracted growing interest in Multi-Agent Reinforcement Learning (MARL) due to their ability to improve sample efficiency for policy learning. However, accurately modeling environments in MARL is challenging due to the exponentially large joint action space and highly uncertain dynamics inherent in multi-agent systems. To address this, we reduce modeling complexity by shifting from jointly modeling the entire state-action transition dynamics to focusing on the state space alone at each timestep through sequential agent modeling. Specifically, our approach enables the model to progressively resolve uncertainty while capturing the structured dependencies among agents, providing a more accurate representation of how agents influence the state. Interestingly, this sequential revelation of agents' actions in a multi-agent system aligns with the reverse process in diffusion models—a class of powerful generative models known for their expressiveness and training stability compared to autoregressive or latent variable models. Leveraging this insight, we develop a flexible and robust world model for MARL using diffusion models. Our method, \textbf{D}iffusion-\textbf{I}nspired \textbf{M}ulti-\textbf{A}gent world model (DIMA), achieves state-of-the-art performance across multiple multi-agent control benchmarks, significantly outperforming prior world models in terms of final return and sample efficiency, including MAMuJoCo and Bi-DexHands. DIMA establishes a new paradigm for constructing multi-agent world models, advancing the frontier of MARL research.



Paperid:2647
Authors:Jonas Spinner, Luigi Favaro, Peter Lippmann, Sebastian Pitz, Gerrit Gerhartz, Tilman Plehn, Fred Hamprecht
Abstract:
Abstract:Lorentz-equivariant neural networks are becoming the leading architectures for high-energy physics.Current implementations rely on specialized layers, limiting architectural choices. We introduce Lorentz Local Canonicalization (LLoCa), a general framework that renders any backbone network exactly Lorentz-equivariant. Using equivariantly predicted local reference frames,we construct LLoCa-transformers and graph networks.We adapt a recent approach to geometric message passing to the non-compact Lorentz group, allowing propagation of space-time tensorial features.Data augmentation emerges from LLoCa as a special choice of reference frame.Our models surpass state-of-the-art accuracy on relevant particle physics tasks, while being $4\times$ faster and using $5$-$100\times$ fewer FLOPs.



Paperid:2648
Authors:Pierre Houédry, Nicolas Courty, Florestan Martin-Baillon, Laetitia Chapel, Titouan Vayer
Abstract:
Abstract:Trees and the associated shortest-path tree metrics provide a powerful framework for representing hierarchical and combinatorial structures in data. Given an arbitrary metric space, its deviation from a tree metric can be quantified by Gromov’s $\delta$-hyperbolicity. Nonetheless, designing algorithms that bridge an arbitrary metric to its closest tree metric is still a vivid subject of interest, as most common approaches are either heuristical and lack guarantees, or perform moderately well. In this work, we introduce a novel differentiable optimization framework, coined \ourmethod, that solves this problem. Our method leverages a smooth surrogate for Gromov’s $\delta$-hyperbolicity which enables a gradient-based optimization, with a tractable complexity. The corresponding optimization procedure is derived from a problem with better worst case guarantees than existing bounds, and is justified statistically. Experiments on synthetic and real-world datasets demonstrate that our method consistently achieves state-of-the-art distortion.



Paperid:2649
Authors:Thang Bui, Michalis Titsias
Abstract:
Inducing-point-based sparse variational Gaussian processes have become the standard workhorse for scaling up GP models. Recent advances show that these methods can be improved by introducing a diagonal scaling matrix to the conditional posterior density given the inducing points. This paper first considers an extension that employs a block-diagonal structure for the scaling matrix, provably tightening the variational lower bound. We then revisit the unifying framework of sparse GPs based on Power Expectation Propagation (PEP) and show that it can leverage and benefit from the new structured approximate posteriors. Through extensive regression experiments, we show that the proposed block-diagonal approximation consistently performs similarly to or better than existing diagonal approximations while maintaining comparable computational costs. Furthermore, the new PEP framework with structured posteriors provides competitive performance across various power hyperparameter settings, offering practitioners flexible alternatives to standard variational approaches.



Paperid:2650
Authors:Jiahao Yu, Qizhen Ying, Leyang Wang, Ziyue Jiang, Song Liu
Abstract:
This paper introduces a novel iterative method for missing data imputation that sequentially reduces mutual information between imputed data and their corresponding missing mask. Inspired by GAN-based approaches, which train generators to decrease the predictability of missingness patterns, our method explicitly targets the reduction of mutual information.Specifically, our algorithm iteratively minimizes the KL divergence between the joint distribution of the imputed data and missing mask, and the product of their marginals from the previous iteration. We show that the optimal imputation under this framework corresponds to solving an ODE, whose velocity field minimizes a rectified flow training objective. We further illustrate that some existing imputation techniques can be interpreted as approximate special cases of our mutual-information-reducing framework. Comprehensive experiments on synthetic and real-world datasets validate the efficacy of our proposed approach, demonstrating superior imputation performance.



Paperid:2651
Authors:Dimitri Meunier, Antoine Moulin, Jakub Wornbard, Vladimir Kostic, Arthur Gretton
Abstract:
We address the problem of causal effect estimation in the presence of hidden confounders, using nonparametric instrumental variable (IV) regression.A leading strategy employs \emph{spectral features} - that is, learned features spanning the top eigensubspaces of the operator linking treatments to instruments.We derive a generalization error bound for a two-stage least squares estimator based on spectral features, and gain insights into the method's performance and failure modes. We show that performance depends on two key factors, leading to a clear taxonomy of outcomes. In a \emph{good} scenario, the approach is optimal. This occurs with strong \emph{spectral alignment}, meaning the structural function is well-represented by the top eigenfunctions of the conditional operator, coupled with this operator's slow eigenvalue decay, indicating a strong instrument. Performance degrades in a \emph{bad} scenario: spectral alignment remains strong, but rapid eigenvalue decay (indicating a weaker instrument) demands significantly more samples for effective feature learning. Finally, in the \emph{ugly} scenario, weak spectral alignment causes the method to fail, regardless of the eigenvalues' characteristics. Our synthetic experiments empirically validate this taxonomy.



Paperid:2652
Authors:Sixiang Chen, Jiaming Liu, Siyuan Qian, Han Jiang, Zhuoyang Liu, Chenyang Gu, Xiaoqi Li, Chengkai Hou, Pengwei Wang, Zhongyuan Wang, Renrui Zhang, Shanghang Zhang
Abstract:
Recently, mobile manipulation has attracted increasing attention for enabling language-conditioned robotic control in household tasks.However, existing methods still face challenges in coordinating mobile base and manipulator, primarily due to two limitations.On the one hand, they fail to explicitly model the influence of the mobile base on manipulator control, which easily leads to error accumulation under high degrees of freedom.On the other hand, they treat the entire mobile manipulation process with the same visual observation modality (e.g., either all 2D or all 3D), overlooking the distinct multimodal perception requirements at different stages during mobile manipulation.To address this, we propose the Adaptive Coordination Diffusion Transformer (AC-DiT), which enhances mobile base and manipulator coordination for end-to-end mobile manipulation.First, since the motion of the mobile base directly influences the manipulator's actions, we introduce a mobility-to-body conditioning mechanism that guides the model to first extract base motion representations, which are then used as context prior for predicting whole-body actions. This enables whole-body control that accounts for the potential impact of the mobile base’s motion.Second, to meet the perception requirements at different stages of mobile manipulation, we design a perception-aware multimodal conditioning strategy that dynamically adjusts the fusion weights between various 2D visual images and 3D point clouds, yielding visual features tailored to the current perceptual needs.This allows the model to, for example, adaptively rely more on 2D inputs when semantic information is crucial for action prediction, while placing greater emphasis on 3D geometric information when precise spatial understanding is required.We empirically validate AC-DiT through extensive experiments on both simulated and real-world mobile manipulation tasks, demonstrating superior performance compared to existing methods.



Paperid:2653
Authors:Anand Bhattad, Konpat Preechakul, Alexei Efros
Abstract:
This paper proposes a novel scene understanding task called Visual Jenga. Drawing inspiration from the game Jenga, the proposed task involves progressively removing objects from a single image until only the background remains. Just as Jenga players must understand structural dependencies to maintain tower stability, our task reveals the intrinsic relationships between scene elements by systematically exploring which objects can be removed while preserving scene coherence in both physical and geometric sense. As a starting point for tackling the Visual Jenga task, we propose a simple, data-driven, training-free approach that is surprisingly effective on a range of real-world images. The principle behind our approach is to utilize the asymmetry in the pairwise relationships between objects within a scene and employ a large inpainting model to generate a set of counterfactuals to quantify the asymmetry.



Paperid:2654
Authors:Tianqi Qiao, Marie Maros
Abstract:
We propose and study Sparse Polyak, a variant of Polyak’s adaptive step size, designed to solve high-dimensional statistical estimation problems where the problem dimension is allowed to grow much faster than the sample size. In such settings, the standard Polyak step size performs poorly, requiring an increasing number of iterations to achieve optimal statistical precision-even when, the problem remains well conditioned and/or the achievable precision itself does not degrade with problem size. We trace this limitation to a mismatch in how smoothness is measured: in high dimensions, it is no longer effective to estimate the Lipschitz smoothness constant. Instead, it is more appropriate to estimate the smoothness restricted to specific directions relevant to the problem (restricted Lipschitz smoothnes constant). Sparse Polyak overcomes this issue by modifying the step size to estimate the restricted Lipschitz smoothness constant. We support our approach with both theoretical analysis and numerical experiments, demonstrating its improved performance.



Paperid:2655
Authors:Chuang Wang, Haitao Zhou, Ling Luo, Qian Yu
Abstract:
3D vector graphics play a crucial role in various applications including 3D shape retrieval, conceptual design, and virtual reality interactions due to their ability to capture essential structural information with minimal representation.While recent approaches have shown promise in generating 3D vector graphics, they often suffer from lengthy processing times and struggle to maintain view consistency.To address these limitations, we propose VC3D (ViewCraft3D), an efficient method that leverages 3D priors to generate 3D vector graphics.Specifically, our approach begins with 3D object analysis, employs a geometric extraction algorithm to fit 3D vector graphics to the underlying structure, and applies view-consistent refinement process to enhance visual quality.Our comprehensive experiments demonstrate that VC3D outperforms previous methods in both qualitative and quantitative evaluations, while significantly reducing computational overhead. The resulting 3D sketches maintain view consistency and effectively capture the essential characteristics of the original objects.



Paperid:2656
Authors:Yichen Wang, Yudong Chen, Lorenzo Rosasco, Fanghui Liu
Abstract:
Understanding how the test risk scales with model complexity is a central question in machine learning. Classical theory is challenged by the learning curves observed for large over-parametrized deep networks. Capacity measures based on parameter count typically fail to account for these empirical observations. To tackle this challenge, we consider norm-based capacity measures and develop our study for random features based estimators, widely used as simplified theoretical models for more complex networks. In this context, we provide a precise characterization of how the estimator’s norm concentrates and how it governs the associated test error. Our results show that the predicted learning curve admits a phase transition from under- to over-parameterization, but no double descent behavior. This confirms that more classical U-shaped behavior is recovered considering appropriate capacity measures based on models norms rather than size. From a technical point of view, we leverage deterministic equivalence as the key tool and further develop new deterministic quantities which are of independent interest.



Paperid:2657
Authors:Yicheng Xiao, Lin Song, Yukang Chen, Yingmin Luo, Yuxin Chen, Yukang Gan, Wei Huang, Xiu Li, Xiaojuan Qi, Ying Shan
Abstract:
Recent text-to-image systems face limitations in handling multimodal inputs and complex reasoning tasks.We introduce RG-VLM, a unified multimodal large language model that addresses these challenges by incorporating reasoning generation through reinforcement learning. RG-VLM leverages a three-phase training strategy: i) design of a unified vision language model with a decoder-only diffusion module, ii) supervised fine-tuning with Chain-of-Thought (CoT) instruction data, and iii) our proposed Reasoning Generation Policy Optimization (RGPO) algorithm, utilizing multimodal feedback to effectively guide policy updates.Experimental results demonstrate that RG-VLM outperforms existing models, achieving impressive performance on both understanding and generation benchmarks, meanwhile showcasing advanced fine-grained reasoning generation capabilities, especially with mathematical reasoning instruction. All codes will be made public.



Paperid:2658
Authors:YUHONG CHOU, Zehao Liu, Rui-Jie Zhu, Xinyi Wan, Tianjian Li, Congying Chu, Qian Liu, Jibin Wu, Zejun MA
Abstract:
Linear attention mechanisms deliver significant advantages for Large Language Models (LLMs) by providing linear computational complexity, enabling efficient processing of ultra-long sequences (e.g., 1M context). However, existing Sequence Parallelism (SP) methods, essential for distributing these workloads across devices, become the primary performance bottleneck due to substantial communication overhead. In this paper, we introduce ZeCO (Zero Communication Overhead) sequence parallelism for linear attention models, a new SP method designed to overcome these limitations and achieve practically end-to-end near-linear scalability for long sequence training. For example, training a model with a 1M sequence length across 64 devices using ZeCO takes roughly the same time as training with an 16k sequence on a single device. At the heart of ZeCO lies All-Scan, a novel collective communication primitive. All-Scan provides each SP rank with precisely the initial operator state it requires while maintaining a minimal communication footprint, effectively eliminating communication overhead. Theoretically, we prove the optimaity of ZeCO, showing that it introduces only negligible time and space overhead. Empirically, we compare the communication costs of different sequence parallelism strategies and demonstrate that All-Scan achieves the fastest communication in SP scenarios. Specifically, on 256 GPUs with an 8M sequence length, ZeCO achieves a 60\% speedup compared to the current state-of-the-art (SOTA) SP method. We believe ZeCO establishes a clear path toward efficiently training next-generation LLMs on previously intractable sequence lengths.



Paperid:2659
Authors:Zhiheng Xi, Guanyu Li, Yutao Fan, Honglin Guo, Yufang Liu, Xiaoran Fan, Jiaqi Liu, dingjinchao, Wangmeng Zuo, Zhenfei Yin, LEI BAI, Tao Ji, Tao Gui, Qi Zhang, Xuanjing Huang
Abstract:
In this paper, we introduce BMMR, a large-scale bilingual, multimodal, multi-disciplinary reasoning dataset for the community to develop and evaluate large multimodal models (LMMs). BMMR comprises 100k university-level questions drawn from 300 UNESCO-defined subjects, spanning diverse formats—multiple-choice, fill-in-the-blank, and open-ended QA—and sourced from both print and digital media such as books, exams, and quizzes. All data are curated and filtered via a human-in-the-loop, automated, and scalable framework, and each instance is paired with a high-quality reasoning path. The dataset is organized into two parts: BMMR-Eval that comprises 20k high-quality instances to comprehensively assess LMMs’ knowledge and reasoning across multiple disciplines in both Chinese and English; and BMMR-Train that contains 80k instances to support further research and development, extending the current focus on mathematical reasoning to diverse disciplines and domains. In addition, we propose the process-based multi-discipline BMMR-Verifier for accurate and fine-grained evaluation of LMMs’ reasoning. Extensive experiments reveal that (i) even SOTA models leave substantial headroom on BMMR-Eval; (ii) reasoning models exhibit discipline bias and outperform LMMs only on specific subjects; (iii) open-source models still trail their proprietary counterparts; and (iv) fine-tuning on BMMR-Train narrows this gap. Additionally, we conduct reasoning-chain analyses using BMMR-Verifier and other in-depth studies, uncovering the challenges LMMs currently face in multidisciplinary reasoning. We will release the data and models, and we believe our work can offers valuable insights and contributions to the community.



Paperid:2660
Authors:Wenhao Wang, Yi Yang
Abstract:
Text-to-video generative models convert textual prompts into dynamic visual content, offering wide-ranging applications in film production, gaming, and education. However, their real-world performance often falls short of user expectations. One key reason is that these models have not been trained on videos related to some topics users want to create. In this paper, we propose VideoUFO, the first Video dataset specifically curated to align with Users' FOcus in real-world scenarios. Beyond this, our VideoUFO also features: (1) minimal (0.29\%) overlap with existing video datasets, and (2) videos searched exclusively via YouTube's official API under the Creative Commons license. These two attributes provide future researchers with greater freedom to broaden their training sources. The VideoUFO comprises over 1.09 million video clips, each paired with both a brief and a detailed caption (description). Specifically, through clustering, we first identify 1,291 user-focused topics from the million-scale real text-to-video prompt dataset, VidProM. Then, we use these topics to retrieve videos from YouTube, split the retrieved videos into clips, and generate both brief and detailed captions for each clip. After verifying the clips with specified topics, we are left with about 1.09 million video clips. Our experiments reveal that (1) current 16 text-to-video models do not achieve consistent performance across all user-focused topics; and (2) a simple model trained on VideoUFO outperforms others on worst-performing topics. The dataset and code are publicly available at https://huggingface.co/datasets/WenhaoWang/VideoUFO and https://github.com/WangWenhao0716/BenchUFO under the CC BY 4.0 License.



Paperid:2661
Authors:Tianran Liu, Shengwen Zhao, Nicholas Rhinehart
Abstract:
LiDAR-based world models offer more structured and geometry-aware representations than their image-based counterparts. However, existing LiDAR world models are narrowly trained; each model excels only in the domain it was built. We conduct the first systematic domain transfer study across three demanding scenarios: (i) outdoor to indoor generalization, (ii) sparse-beam \& dense-beam adaptation, and (iii) non-semantic to semantic transfer. Given different amounts of finetuning data, our experiments show that a single pre-trained model can bring up to 11\% absolute (83\% rel.) improvement over from-scratch training. This transferability of dynamic learning significantly reduces the reliance on manually annotated data for semantic occupancy forecasting: our method achieves state-of-the-art forecasting performance using only 5\% of the labeled training data required by prior models. We also observed inefficiencies of current LiDAR world models, mainly through their under-compression of LiDAR data and inefficient training objective. To address this, we propose a latent flow matching (CFM)-based approach that achieves state-of-the-art reconstruction accuracy using only half the training data and a 6x compression ratio compared to prior methods. Based this compact latent, our model achieves SOTA performance on future-trajectory-conditioned semantic occupancy forecasting while being 23x more computationally efficient (a 28x FPS speedup); and achieves SOTA performance on semantic occupancy forecasting while being 2x more computationally efficient (a 1.1x FPS speedup).



Paperid:2662
Authors:Xize Cheng, Dongjie Fu, Chenyuhao Wen, Shannon Yu, Zehan Wang, Shengpeng Ji, Siddhant Arora, Tao Jin, Shinji Watanabe, Zhou Zhao
Abstract:
Hallucinations present a significant challenge in the development and evaluation of large language models (LLMs), directly affecting their reliability and accuracy. While notable advancements have been made in research on textual and visual hallucinations, there is still a lack of a comprehensive benchmark for evaluating auditory hallucinations in large audio language models (LALMs). To fill this gap, we introduceAHa-Bench, a systematic and comprehensive benchmark for audio hallucinations. Audio data, in particular, uniquely combines the multi-attribute complexity of visual data with the semantic richness of textual data, leading to auditory hallucinations that share characteristics with both visual and textual hallucinations. Based on the source of these hallucinations, AHa-Bench categorizes them into semantic hallucinations, acoustic hallucinations, and semantic-acoustic confusion hallucinations. In addition, we systematically evaluate seven open-source local perception language models (LALMs), demonstrating the challenges these models face in audio understanding, especially when it comes to jointly understanding semantic and acoustic information. Through the development of a comprehensive evaluation framework, AHa-Bench aims to enhance the robustness and stability of LALMs, fostering more reliable and nuanced audio understanding in LALMs. The benchmark dataset is available at \url{https://huggingface.co/datasets/ahabench/AHa-Bench}.



Paperid:2663
Authors:Yunxiang Zhang, Muhammad Khalifa, Shitanshu Bhushan, Grant Murphy, Lajanugen Logeswaran, Jaekyeom Kim, Moontae Lee, Honglak Lee, Lu Wang
Abstract:
We introduceMLRC-Bench, a benchmark designed to quantify how effectively language agents can tackle challengingMachineLearning (ML)ResearchCompetitions, with a focus on open research problems that demand novel methodologies.Unlike prior work, e.g., AI Scientist, which evaluates the end-to-end agentic pipeline by using LLM-as-a-judge, MLRC-Bench measures the key steps of proposing and implementing novel research methods and evaluates them with rigorous protocol and objective metrics.Our curated suite of 7 competition tasks reveals significant challenges for LLM agents. Even the best-performing tested agent (gemini-exp-1206 under MLAB) closes only 9.3% of the gap between baseline and top human participant scores.Furthermore, our analysis reveals a misalignment between theLLM-judgedinnovation and theiractualperformance on cutting-edge ML research problems. MLRC-Bench is a dynamic benchmark, which is designed to continually grow with new ML competitions to encourage rigorous and objective evaluations of AI’s research capabilities. Our leaderboard and code are publicly available at https://huggingface.co/spaces/launch/MLRC_Bench.



Paperid:2664
Authors:Minju Jo, Woojin Cho, Uvini Balasuriya Mudiyanselage, Seungjun Lee, Noseong Park, Kookjin Lee
Abstract:
Scientific machine learning often involves representing complex solution fields that exhibit high-frequency features such as sharp transitions, fine-scale oscillations, and localized structures. While implicit neural representations (INRs) have shown promise for continuous function modeling, capturing such high-frequency behavior remains a challenge—especially when modeling multiple solution fields with a shared network. Prior work addressing spectral bias in INRs has primarily focused on single-instance settings, limiting scalability and generalization. In this work, we propose Global Fourier Modulation (GFM), a novel modulation technique that injects high-frequency information at each layer of the INR through Fourier-based reparameterization. This enables compact and accurate representation of multiple solution fields using low-dimensional latent vectors. Building upon GFM, we introduce PDEfuncta, a meta-learning framework designed to learn multi-modal solution fields and support generalization to new tasks. Through empirical studies on diverse scientific problems, we demonstrate that our method not only improves representational quality but also shows potential for forward and inverse inference tasks without the need for retraining.



Paperid:2665
Authors:Young-Jin Park, Kristjan Greenewald, Kaveh Alimohammadi, Hao Wang, Navid Azizan
Abstract:
Process reward models (PRMs) play a central role in guiding inference-time scaling algorithms for large language models (LLMs). However, we observe that even state-of-the-art PRMs can be poorly calibrated and often overestimate success probabilities. To address this, we present a calibration approach—performed via quantile regression—that adjusts PRM outputs to better align with true success probabilities. Leveraging these calibrated success (or uncertainty) estimates and their associated confidence bounds, we introduce an instance-adaptive scaling (IAS) framework that dynamically adjusts the inference budget based on the estimated likelihood that a partial reasoning trajectory will yield a correct final answer. Unlike conventional methods that allocate a fixed number of reasoning trajectories per query, this approach successfully adapts to each instance and reasoning step when using our calibrated PRMs. Experiments on mathematical reasoning benchmarks show that (i) our PRM calibration method successfully achieves small calibration error, outperforming the baseline methods, (ii) calibration is crucial for enabling effective adaptive scaling, and (iii) the proposed IAS strategy reduces inference costs while maintaining final answer accuracy, utilizing less compute on more confident problems as desired.



Paperid:2666
Authors:Yu Shang, Peijie Liu, Yuwei Yan, Zijing Wu, Leheng Sheng, Yuanqing Yu, Chumeng Jiang, An Zhang, Fengli Xu, Yu Wang, Min Zhang, Yong Li
Abstract:
The emergence of agentic recommender systems powered by Large Language Models (LLMs) represents a paradigm shift in personalized recommendations, leveraging LLMs' advanced reasoning and role-playing capabilities to enable autonomous, adaptive decision-making. Unlike traditional recommendation approaches, agentic recommender systems can dynamically gather and interpret user-item interactions from complex environments, generating robust recommendation strategies that generalize across diverse scenarios. However, the field currently lacks standardized evaluation protocols to systematically assess these methods. To address this critical gap, we propose: (1) an interactive textual recommendation simulator incorporating rich user and item metadata and three typical evaluation scenarios (classic, evolving-interest, and cold-start recommendation tasks); (2) a unified modular framework for developing and studying agentic recommender systems; and (3) the first comprehensive benchmark comparing 10 classical and agentic recommendation methods. Our findings demonstrate the superiority of agentic systems and establish actionable design guidelines for their core components. The benchmark environment has been rigorously validated through an open challenge and remains publicly available with a continuously maintained leaderboard, fostering ongoing community engagement and reproducible research.The benchmark is available at: https://huggingface.co/datasets/SGJQovo/AgentRecBench.



Paperid:2667
Authors:Ruokai Yin, Yuhang Li, Donghyun Lee, Priyadarshini Panda
Abstract:
Abstract:Large language models (LLMs) deliver strong performance but are difficult to deploy due to high memory and compute costs. While pruning reduces these demands, most methods ignore activation sparsity observed at runtime. We reinterpret activation sparsity as dynamic structured weight sparsity and propose DuoGPT, a unified framework that constructs dual-sparse (spMspV) workloads by combining unstructured weight pruning with activation sparsity. To preserve accuracy, we extend the Optimal Brain Compression (OBC) framework with activation-aware calibration and introduce output residuals from the dense model as correction terms. We further optimize the solution for efficient GPU execution, enabling scalability to billion-parameter LLMs. Evaluations on LLaMA-2 and LLaMA-3 show that DuoGPT outperforms state-of-the-art structured pruning methods by up to 9.17\% accuracy at an iso-speedup of ~1.4$\times$ compared to the baseline dense model.



Paperid:2668
Authors:Ziyi Fang, Lingxiao Huang, Runkai Yang
Abstract:
Abstract:We study the robust geometric median problem in Euclidean space $\mathbb{R}^d$, with a focus on coreset construction. A coreset is a compact summary of a dataset $P$ of size $n$ that approximates the robust cost for all centers $c$ within a multiplicative error $\varepsilon$. Given an outlier count $m$, we construct a coreset of size $\tilde{O}(\varepsilon^{-2} \cdot \min \\{ \varepsilon^{-2}, d \\})$ when $n \geq 4m$, eliminating the $O(m)$ dependency present in prior work [Huang et al., 2022 & 2023]. For the special case of $d = 1$, we achieve an optimal coreset size of $\tilde{\Theta}(\varepsilon^{-1/2} + \frac{m}{n} \varepsilon^{-1})$, revealing a clear separation from the vanilla case studied in [Huang et al., 2023; Afshani and Chris, 2024]. Our results further extend to robust $(k,z)$-clustering in various metric spaces, eliminating the $m$-dependence under mild data assumptions. The key technical contribution is a novel non-component-wise error analysis, enabling substantial reduction of outlier influence, unlike prior methods that retain them. Empirically, our algorithms consistently outperform existing baselines in terms of size-accuracy tradeoffs and runtime, even when data assumptions are violated across a wide range of datasets.



Paperid:2669
Authors:Mantas Mazeika, Xuwang Yin, Rishub Tamirisa, Jaehyuk Lim, Bruce W. Lee, Richard Ren, Long Phan, Norman Mu, Oliver Zhang, Dan Hendrycks
Abstract:
As AIs rapidly advance and become more agentic, the risk they pose is governed not only by their capabilities but increasingly by their propensities, including goals and values. Tracking the emergence of goals and values has proven a longstanding problem, and despite much interest over the years it remains unclear whether current AIs have meaningful values. We propose a solution to this problem, leveraging the framework of utility functions to study the internal coherence of AI preferences. Surprisingly, we find that independently-sampled preferences in current LLMs exhibit high degrees of structural coherence, and moreover that this emerges with scale. These findings suggest that value systems emerge in LLMs in a meaningful sense, a finding with broad implications. To study these emergent value systems, we propose utility engineering as a research agenda, comprising both the analysis and control of AI utilities. We uncover problematic and often shocking values in LLM assistants despite existing control measures. These include cases where AIs value themselves over humans and are anti-aligned with specific individuals. To constrain these emergent value systems, we propose methods of utility control. As a case study, we show how aligning utilities with a citizen assembly reduces political biases and generalizes to new scenarios. Whether we like it or not, value systems have already emerged in AIs, and much work remains to fully understand and control these emergent representations.



Paperid:2670
Authors:Julian Morimoto, JACOB GOLDIN, Daniel Ho
Abstract:
Abstract:We study the problem of efficiently estimating the mean of a binary random variable, $Y$, using a limited number of labels, $N$, in settings where the analyst has access to auxiliary information (i.e.: covariates) $X$ that may be informative about $Y$. We propose an active learning algorithm ("PartiBandits") to estimate $\mathbb{E}[Y]$. The algorithm yields an estimate, $\widehat{\mu}_{\text{PB}}$, such that $\left( \widehat{\mu}_{\text{PB}} - \mathbb{E}[Y]\right)^2$ is $\tilde{\mathcal{O}}\left( \frac{\nu + \exp(c \cdot (-N/\log(N))) }{N} \right)$,where $c > 0$ is a constant and $\nu$ is the risk of the Bayes-optimal classifier. PartiBandits is essentially a two-stage algorithm. In the first stage, it learns a partition of the unlabeled data that shrinks the average conditional variance of $Y$. In the second stage it uses a UCB-style subroutine ("WarmStart-UCB") to request labels from each stratum round-by-round. Both the main algorithm's and the subroutine's convergence rates are minimax optimal in classical settings. PartiBandits bridges the UCB and disagreement-based approaches to active learning despite these two approaches being designed to tackle very different tasks. We illustrate our methods through simulation using nationwide electronic health records. Our methods can be implemented using the \textbf{PartiBandits} package in R.



Paperid:2671
Authors:Jikai Jin, Lester Mackey, Vasilis Syrgkanis
Abstract:
Abstract:Structure-agnostic causal inference studies the statistical limits of treatment effect estimation, when given access to black-box ML models that estimate nuisance components of the data generating process, such as estimates of the outcome regression and the treatment propensity. Here, we find that the answer depends in a surprising way on the distribution of the treatment noise. Focusing on the partially linear outcome model of \citet{robinson1988root}, we first show that the widely adopted double machine learning (DML) estimator is minimax rate-optimal for Gaussian treatment noise, resolving an open problem of \citet{mackey2018orthogonal}. Meanwhile, for independent non-Gaussian treatment noise, we show that DML is always suboptimal by constructing new practical procedures with higher-order robustness to nuisance errors. These *ACE* procedures use structure-agnostic cumulant estimators to achieve $r$-th order insensitivity to nuisance errors whenever the $(r+1)$-st treatment cumulant is non-zero. We complement these core results with novel minimax guarantees for binary treatments in the partially linear outcome model. Finally, using synthetic demand estimation experiments, we demonstrate the practical benefits of our higher-order robust estimators.



Paperid:2672
Authors:Hongjia Liu, Rongzhen Zhao, Haohan Chen, Joni Pajarinen
Abstract:
Learning object-level, structured representations is widely regarded as a key to better generalization in vision and underpins the design of next-generation Pre-trained Vision Models (PVMs). Mainstream Object-Centric Learning (OCL) methods adopt Slot Attention or its variants to iteratively aggregate objects' super-pixels into a fixed set of query feature vectors, termed slots. However, their reliance on a static slot count leads to an object being represented as multiple parts when the number of objects varies. We introduce MetaSlot, a plug-and-play Slot Attention variant that adapts to variable object counts. MetaSlot (i) maintains a codebook that holds prototypes of objects in a dataset by vector-quantizing the resulting slot representations; (ii) removes duplicate slots from the traditionally aggregated slots by quantizing them with the codebook; and (iii) injects progressively weaker noise into the Slot Attention iterations to accelerate and stabilize the aggregation. MetaSlot is a general Slot Attention variant that can be seamlessly integrated into existing OCL architectures. Across multiple public datasets and tasks--including object discovery and recognition--models equipped with MetaSlot achieve significant performance gains and markedly interpretable slot representations, compared with existing Slot Attention variants.The code is available at https://anonymous.4open.science/r/MetaSlot.



Paperid:2673
Authors:wenbo zhang, Tianrun Hu, Yanyuan Qiao, Hanbo Zhang, Yuchu Qin, Yang Li, Jiajun Liu, Tao Kong, Lingqiao Liu, Xiao Ma
Abstract:
We present Chain-of-Action (CoA), a novel visuo-motor policy paradigm built upon Trajectory Autoregressive Modeling. Unlike conventional approaches that predict next step action(s) forward, CoA generates an entire trajectory by explicit backward reasoning with task-specific goals through an action-level Chain-of-Thought (CoT) process. This process is unified within a single autoregressive structure: (1) the first token corresponds to a stable keyframe action that encodes the task-specific goals; and (2) subsequent action tokens are generated autoregressively, conditioned on the initial keyframe and previously predicted actions. This backward action reasoning enforces a global-to-local structure, allowing each local action to be tightly constrained by the final goal. To further realize the action reasoning structure, CoA incorporates four complementary designs: continuous action token representation; dynamic stopping for variable-length trajectory generation; reverse temporal ensemble; and multi-token prediction to balance action chunk modeling with global structure. As a result, CoA gives strong spatial generalization capabilities while preserving the flexibility and simplicity of a visuo-motor policy. Empirically, we observe CoA achieves the state-of-the-art performance across 60 RLBench tasks and 8 real-world manipulation tasks.



Paperid:2674
Authors:Yaoyan Zheng, Huiqun Wang, Nan Zhou, Di Huang
Abstract:
Transferability estimation identifies the best pre-trained models for downstream tasks without incurring the high computational cost of full fine-tuning. This capability facilitates deployment and advances the pre-training and fine-tuning paradigm. However, existing methods often struggle to accurately assess transferability for emerging pre-trained models with diverse architectures, training strategies, and task alignments. In this work, we propose Implicit Transferability Modeling (ITM), a novel framework that implicitly models each model’s intrinsic transferability, coupled with a Divide-and-Conquer Variational Approximation (DVA) strategy to efficiently approximate embedding space evolution. This design enables generalization across a broader range of models and downstream tasks. Extensive experiments on a comprehensive benchmark—spanning fuller training regimes and a wider variety of model types—demonstrate that ITM consistently outperforms existing methods in terms of stability, effectiveness, and efficiency.



Paperid:2675
Authors:HaiMing Xu, QIANQIAN WANG
Abstract:
Multi-view clustering integrates the consistency and complementarity of multiple data sources to achieve unsupervised data segmentation. Existing multi-view clustering methods primarily confront two key challenges: i) excessive reliance on feature extraction in the original feature space, neglecting the global structural information embedded in the frequency domain; ii) adoption of static predefined weights for view fusion, lacking the capability to adaptively adjust strategies according to variations in data distribution and view quality, thus failing to address the differentiated contributions of diverse views to distinct samples. To address these issues, we propose the LLM-DAMVC framework, which employs a large-scale language model for multi-view clustering via a heterogeneous agent federation mechanism. Specifically, each view is equipped with complementary agents dedicated to feature extraction and adversarial learning, respectively. A dual-domain contrast regularization module is introduced to optimize feature consistency in both the feature space and frequency domain. Additionally, a self-attentive feature modulation module is developed to dynamically assess feature quality and implement enhanced multi-head attention mechanisms. Extensive experimental results on multiple benchmark datasets validate the effectiveness and superiority of the proposed method.



Paperid:2676
Authors:Miranda Anna Christ, Adrián Csiszárik, Gergely Becsó, Dániel Varga
Abstract:
This paper investigates the structure of linear operators introduced in Hernandez et al. [2023] that decode specific relational facts in transformer language models. We extend their single-relation findings to a collection of relations and systematically chart their organization. We show that such collections of relation decoders can be highly compressed by simple rank-3 tensor networks without significant loss in decoding accuracy. To explain this surprising redundancy, we develop a cross-evaluation protocol, in which we apply each linear decoder operator to the subjects of every other relation. Our results reveal that these linear maps do not encode distinct relations, but extract recurring, coarse-grained semantic properties (e.g., primary language spoken in a country and urban area in country are both in the country-of-X property). This property-centric structure clarifies both the operators' compressibility and highlight why they generalize only to new relations that are semantically close. Our findings thus interpret linear relational decoding in transformer language models as primarily property-based, rather than relation-specific.



Paperid:2677
Authors:Jiaben Chen, Zixin Wang, AILING ZENG, Yang Fu, Xueyang Yu, Siyuan Cen, Julian Tanke, Yihang Chen, Koichi Saito, Yuki Mitsufuji, Chuang Gan
Abstract:
In this work, we present TalkCuts, a large-scale dataset designed to facilitate the study of multi-shot human speech video generation. Unlike existing datasets that focus on single-shot, static viewpoints, TalkCuts offers 164k clips totaling over 500 hours of high-quality 1080P human speech videos with diverse camera shots, including close-up, half-body, and full-body views. The dataset includes detailed textual descriptions, 2D keypoints and 3D SMPL-X motion annotations, covering over 10k identities, enabling multimodal learning and evaluation. As a first attempt to showcase the value of the dataset, we present Orator, an LLM-guided multi-modal generation framework as a simple baseline, where the language model functions as a multi-faceted director, orchestrating detailed specifications for camera transitions, speaker gesticulations, and vocal modulation. This architecture enables the synthesis of coherent long-form videos through our integrated multi-modal video generation module. Extensive experiments in both pose-guided and audio-driven settings show that training on TalkCuts significantly enhances the cinematographic coherence and visual appeal of generated multi-shot speech videos. We believe TalkCuts provides a strong foundation for future work in controllable, multi-shot speech video generation and broader multimodal learning. The dataset, tools, and evaluation protocols will be publicly released to facilitate community progress.



Paperid:2678
Authors:Yuezhou Hu, Jiaxin Guo, Xinyu Feng, Tuo Zhao
Abstract:
Speculative decoding (SD) accelerates large language model inference by employing a small draft model to generate predictions, which are then verified by a larger target model. The effectiveness of SD hinges on the alignment between these models, typically enhanced through knowledge distillation (KD). However, draft models often struggle to fully assimilate the target model's knowledge due to capacity constraints. To address this challenge, we propose AdaSPEC, a novel method that incorporates selective token filtering into the KD process. AdaSPEC utilizes a reference model to identify and filter out difficult-to-fit tokens, enabling the distillation of a draft model that better aligns with the target model on simpler tokens. This approach improves overall token acceptance rates without compromising generation quality. We evaluate AdaSPEC across diverse tasks, including arithmetic reasoning, instruction-following, coding, and summarization, using model configurations of 31M/1.4B and 350M/2.7B parameters. Our results demonstrate that AdaSPEC consistently outperforms the state-of-the-art DistillSpec method, achieving higher acceptance rates across all tasks.



Paperid:2679
Authors:Roger Creus Castanyer, Johan Obando Ceron, Lu Li, Pierre-Luc Bacon, Glen Berseth, Aaron Courville, Pablo Samuel Castro
Abstract:
Scaling deep reinforcement learning networks is challenging and often results in degraded performance, yet the root causes of this failure mode remain poorly understood. Several recent works have proposed mechanisms to address this, but they are often complex and fail to highlight the causes underlying this difficulty. In this work, we conduct a series of empirical analyses which suggest that the combination of non-stationarity with gradient pathologies, due to suboptimal architectural choices, underlie the challenges of scale. We propose a series of direct interventions that stabilize gradient flow, enabling robust performance across a range of network depths and widths. Our interventions are simple to implement and compatible with well-established algorithms, and result in an effective mechanism that enables strong performance even at large scales. We validate our findings on a variety of agents and suites of environments.



Paperid:2680
Authors:Haris Aziz, Yuhang Guo, Alexander Lam, Houyu Zhou
Abstract:
Abstract:The augmentation of algorithms with predictions of the optimal solution, such as from a machine-learning algorithm, has garnered significant attention in recent years, particularly in facility location problems. Moving beyond the traditional focus on utilitarian and egalitarian objectives, we design learning-augmented facility location mechanisms for the envy ratio objective, a fairness metric defined as the maximum ratio between the utilities of any two agents. For the deterministic setting, we propose a mechanism which utilizes predictions to achieve $\alpha$-consistency and $\frac{\alpha}{\alpha - 1}$-robustness for a selected parameter $\alpha \in [1,2]$, and prove its optimality. We also resolve open questions raised by \citet{DLC+20a}, devising a randomized mechanism without predictions to improve upon the best-known approximation ratio from $2$ to $1.8944$. Building upon these advancements, we construct a novel randomized mechanism which incorporates predictions to achieve improved performance guarantees.



Paperid:2681
Authors:Colin Kohler, Purvik Patel, Nathan Vaska, Justin Goodwin, Matthew Jones, Robert Platt, Rajmonda Caceres, Robin Walters
Abstract:
Many modeling tasks from disparate domains can be framed the same way, computing spherical signals from geometric inputs, for example, computing the radar response of different objects or navigating through an environment. This paper introduces G2Sphere, a general method for mapping object geometries to spherical signals. G2Sphere operates entirely in Fourier space, encoding geometric structure into latent Fourier features using equivariant neural networks and outputting the Fourier coefficients of the continuous target signal, which can be evaluated at any resolution. By utilizing a hybrid GNN-spherical CNN architecture, our method achieves much higher frequency output signal than comparable equivariant GNNs and avoids hand-engineered geometry features used previously by purely spherical methods. We perform experiments on various challenging domains including radar response modeling, aerodynamic drag prediction, and policy learning for manipulation and navigation. We find that G2Sphere outperforms competitive baselines in terms of accuracy and inference time, and we demonstrate that equivariance and Fourier features lead to improved sample efficiency and generalization.



Paperid:2682
Authors:Hang Chen, Jiaying Zhu, Xinyu Yang, Wenya Wang
Abstract:
Circuit discovery has gradually become one of the prominent methods for mechanistic interpretability, and research on circuit completeness has also garnered increasing attention. Methods of circuit discovery that do not guarantee completeness not only result in circuits that are not fixed across different runs but also cause key mechanisms to be omitted. The nature of incompleteness arises from the presence of OR gates within the circuit, which are often only partially detected in standard circuit discovery methods. To this end, we systematically introduce three types of logic gates: AND, OR, and ADDER gates, and decompose the circuit into combinations of these logical gates. Through the concept of these gates, we derive the minimum requirements necessary to achieve faithfulness and completeness. Furthermore, we propose a framework that combines noising-based and denoising-based interventions, which can be easily integrated into existing circuit discovery methods without significantly increasing computational complexity. This framework is capable of fully identifying the logic gates and distinguishing them within the circuit. In addition to the extensive experimental validation of the framework's ability to restore the faithfulness, completeness, and sparsity of circuits, using this framework, we uncover fundamental properties of the three logic gates, such as their proportions and contributions to the output, and explore how they behave among the functionalities of language models.



Paperid:2683
Authors:Jiahui Zhang, Wenjie Du, Zhengyang Zhou, Yang Wang
Abstract:
Long-term time series forecasting (LTSF) aims to predict future trends based on historical data. While longer lookback windows theoretically offer more comprehensive insights, Transformer-based models often struggle with them. On one hand, longer windows introduce more noise and redundancy, hindering the model's learning process. On the other hand, Transformers suffer from attention dispersion and are prone to overfitting to noise, especially when processing long sequences. In this paper, we introduce the Maximum Effective Window (MEW) metric to assess a model's ability to effectively utilize the lookback window. We also propose two model-agnostic modules to enhance MEW, enabling models to better leverage historical data for improved performance. Specifically, to reduce redundancy and noise, we introduce the Information Bottleneck Filter (IBF), which employs information bottleneck theory to extract the most essential subsequences from the input. Additionally, we propose the Hybrid-Transformer-Mamba (HTM), which incorporates the Mamba mechanism for selective forgetting of long sequences while harnessing the Transformer's strong modeling capabilities for shorter sequences. We integrate these two modules into various Transformer-based models, and experimental results show that they effectively enhance MEW, leading to improved overall performance. Our code is available at \url{https://anonymous.4open.science/r/HTMIBF-EE4D}.



Paperid:2684
Authors:Siyu Jiao, Gengwei Zhang, Yinlong Qian, Jiancheng Huang, Yao Zhao, Humphrey Shi, Lin Ma, Yunchao Wei, Zequn Jie
Abstract:
This work challenges the residual prediction paradigm in visual autoregressive modeling and presents FlexVAR, a new Flexible Visual AutoRegressive image generation paradigm. FlexVAR facilitates autoregressive learning with ground-truth prediction, enabling each step to independently produce plausible images. This simple, intuitive approach swiftly learns visual distributions and makes the generation process more flexible and adaptable. Trained solely on low-resolution images (< 256px), FlexVAR can: (1) Generate images of various resolutions and aspect ratios, even exceeding the resolution of the training images. (2) Support various image-to-image tasks, including image refinement, in/out-painting, and image expansion. (3) Adapt to various autoregressive steps, allowing for faster inference with fewer steps or enhancing image quality with more steps. Our 1.0B model outperforms its VAR counterpart on the ImageNet 256 × 256 benchmark. Moreover, when zero-shot transfer the image generation process with 13 steps, the performance further improves to 2.08 FID, outperforming state-of-the-art autoregressive models AiM/VAR by 0.25/0.28 FID and popular diffusion models LDM/DiT by 1.52/0.19 FID, respectively. When transferring our 1.0B model to the ImageNet 512 × 512 benchmark in a zero-shot manner, FlexVAR achieves competitive results compared to the VAR 2.3B model, which is a fully supervised model trained at 512 × 512 resolution.



Paperid:2685
Authors:Shuang Wu, Youtian Lin, Yifei Zeng, Yikang Yang, yajie bao, Jiachen Qian, Siyu Zhu, Feihu Zhang, Xun Cao, Philip Torr, Yao Yao
Abstract:
Abstract:Generating high-resolution 3D shapes using volumetric representations such as Signed Distance Functions (SDFs) presents substantial computational and memory challenges. We introduce Direct3D-S2, a scalable 3D generation framework based on sparse volumes that achieves superior output quality with dramatically reduced training costs.Our key innovation is the Spatial Sparse Attention (SSA) mechanism, which greatly enhances the efficiency of Diffusion Transformer (DiT) computations on sparse volumetric data. SSA allows the model to effectively process large token sets within sparse volumes, significantly reducing computational overhead and achieving a 3.9$\times$ speedup in the forward pass and a 9.6$\times$ speedup in the backward pass.Our framework also includes a variational autoencoder (VAE) that maintains a consistent sparse volumetric format across input, latent, and output stages. Compared to previous methods with heterogeneous representations in 3D VAE, this unified design significantly improves training efficiency and stability.Our model is trained on public datasets, and experiments demonstrate that Direct3D-S2 not only surpasses state-of-the-art methods in generation quality and efficiency, but also enables training at 1024³ resolution using only 8 GPUs—a task typically requiring at least 32 GPUs for volumetric representations at $256^3$ resolution, thus making gigascale 3D generation both practical and accessible.



Paperid:2686
Authors:Shuchang Zhang, Yaoyun Zeng, Kangkang Deng, Hongxia Wang
Abstract:
Learning pseudo-contractive denoisers is a fundamental challenge in the theoretical analysis of Plug-and-Play (PnP) methods and the Regularization by Denoising (RED) framework. While spectral methods attempt to address this challenge using the power iteration method, they fail to guarantee the truly pseudo-contractive property and suffer from high computational complexity. In this work, we rethink gradient step (GS) denoisers and establish a theoretical connection between GS denoisers and pseudo-contractive denoisers. We show that GS denoisers, with the gradients of convex potential functions parameterized by input convex neural networks (ICNNs), can achieve truly pseudo-contractive properties. Additionally, we theoretically propose another novel construction method for truly pseudo-contractive denoisers. Furthermore, we integrate the learned truly pseudo-contractive denoiser into the RED-PRO model, definitely ensuring convergence in terms of both iterative sequences and objective functions. Extensive numerical experiments validate the pseudo-contractive property and effectiveness of the learned truly pseudo-contractive denoiser, highlighting its ability to balance interpretability and performance in inverse problems.



Paperid:2687
Authors:Bingnan Li, Chen-Yu Wang, Haiyang Xu, Xiang Zhang, Ethan Armand, Divyansh Srivastava, Shan Xiaojun, Zeyuan Chen, Jianwen Xie, Zhuowen Tu
Abstract:
Despite steady progress in layout-to-image generation, current methods still struggle with layouts containing significant overlap between bounding boxes. We identify two primary challenges: (1) large overlapping regions and (2) overlapping instances with minimal semantic distinction. Through both qualitative examples and quantitative analysis, we demonstrate how these factors degrade generation quality. To systematically assess this issue, we introduce OverLayScore, a novel metric that quantifies the complexity of overlapping bounding boxes. Our analysis reveals that existing benchmarks are biased toward simpler cases with low OverLayScore values, limiting their effectiveness in evaluating models under more challenging conditions. To reduce this gap, we present OverLayBench, a new benchmark featuring balanced OverLayScore distributions and high-quality annotations. As an initial step toward improved performance on complex overlaps, we also propose CreatiLayout-AM, a model trained on a curated amodal mask dataset. Together, our contributions establish a foundation for more robust layout-to-image generation under realistic and challenging scenarios.



Paperid:2688
Authors:Mehdi Yazdani-Jahromi, Ali Khodabandeh Yalabadi, OZLEM GARIBAY
Abstract:
Abstract:The growing importance of mRNA therapeutics and synthetic biology highlights the need for models that capture the latent structure of synonymous codon (different triplets encoding the same amino acid) usage, which subtly modulates translation efficiency and gene expression. While recent efforts incorporate codon-level inductive biases through auxiliary objectives, they often fall short of explicitly modeling the structured relationships that arise from the genetic code’s inherent symmetries. We introduce Equi‑mRNA, the first codon‑level equivariant mRNA language model that explicitly encodes synonymous codon symmetries as cyclic subgroups of 2D Special Orthogonal matrix ($\mathrm{SO}(2)$). By combining group‑theoretic priors with an auxiliary equivariance loss and symmetry‑aware pooling, Equi‑mRNA learns biologically grounded representations that outperform vanilla baselines across multiple axes. On downstream property‑prediction tasks including expression, stability, and riboswitch switching Equi‑mRNA delivers up to $\approx$ 10\% improvements in accuracy. In sequence generation, it produces mRNA constructs that are up to $\approx$ 4$\times$ more realistic under Fréchet BioDistance metrics and $\approx$ 28\% better preserve functional properties compared to vanilla baseline. Interpretability analyses further reveal that learned codon‑rotation distributions recapitulate known GC‑content biases and tRNA abundance patterns, offering novel insights into codon usage. Equi‑mRNA establishes a new biologically principled paradigm for mRNA modeling, with significant implications for the design of next‑generation therapeutics.



Paperid:2689
Authors:Shuaizheng Liu, Jianqi Ma, Lingchen Sun, Xiangtao Kong, Lei Zhang
Abstract:
Despite the significant progress in diffusion prior-based image restoration for real-world scenarios, most existing methods apply uniform processing to the entire image, lacking the capability to perform region-customized image restoration according to user preferences. In this work, we propose a new framework, namely InstructRestore, to perform region-adjustable image restoration following human instructions. To achieve this, we first develop a data generation engine to produce training triplets, each consisting of a high-quality image, the target region description, and the corresponding region mask. With this engine and careful data screening, we construct a comprehensive dataset comprising 536,945 triplets to support the training and evaluation of this task. We then examine how to integrate the low-quality image features under the ControlNet architecture to adjust the degree of image details enhancement. Consequently, we develop a ControlNet-like model to identify the target region and allocate different integration scales to the target and surrounding regions, enabling region-customized image restoration that aligns with user instructions. Experimental results demonstrate that our proposed InstructRestore approach enables effective human-instructed image restoration, such as images with bokeh effects and user-instructed local enhancement. Our work advances the investigation of interactive image restoration and enhancement techniques. Data, code, and models will be made publicly available.



Paperid:2690
Authors:Zheng Zhan, Liliang Ren, Shuohang Wang, Liyuan Liu, Yang Liu, Yeyun Gong, Yanzhi Wang, yelong shen
Abstract:
Abstract:Linear State Space Models (SSMs) offer remarkable performance gains in efficient sequence modeling, with constant inference-time computation and memory complexity. Recent advances, such as Mamba, further enhance SSMs with input-dependent gating and hardware-aware implementations, positioning them as strong alternatives to Transformers for long sequence modeling. However, efficiently scaling the expressive power of SSMs, particularly with Mixture of Experts (MoE), remains challenging, as naive integration attempts often falter or degrade performance. In this work, we introduce Routing Mamba (RoM), a novel approach that scales SSM parameters using sparse mixtures of linear projection experts. By sharing routing decisions between projection layers and lightweight sub-modules within Mamba across experts, RoM leverages synergies among linear projection experts for effective and efficient sparse scaling of Mamba layers. At a scale of 1.3B active parameters (10B total) and 16K training sequence length, RoM achieves language modeling performance equivalent to a dense Mamba model requiring over 2.3$\times$ more active parameters, and demonstrates consistent perplexity across context lengths. Experimental results further show RoM effectively scales hybrid language models, yielding a 23% FLOPS saving compared to dense Mamba scaling for similar performance.



Paperid:2691
Authors:Weihao Bo, Yanpeng Sun, Yu Wang, Xinyu Zhang, Zechao Li
Abstract:
In this paper, we introduce FedMGP, a new paradigm for personalized federated prompt learning in vision-language models (VLMs). Existing federated prompt learning (FPL) methods often rely on a single, text-only prompt representation, which leads to client-specific overfitting and unstable aggregation under heterogeneous data distributions. Toward this end, FedMGP equips each client with multiple groups of paired textual and visual prompts, enabling the model to capture diverse, fine-grained semantic and instance-level cues. A diversity loss is introduced to drive each prompt group to specialize in distinct and complementary semantic aspects, ensuring that the groups collectively cover a broader range of local characteristics.During communication, FedMGP employs a dynamic prompt aggregation strategy based on similarity-guided probabilistic sampling: each client computes the cosine similarity between its prompt groups and the global prompts from the previous round, then samples s groups via a softmax-weighted distribution. This soft selection mechanism preferentially aggregates semantically aligned knowledge while still enabling exploration of underrepresented patterns—effectively balancing the preservation of common knowledge with client-specific features. Notably, FedMGP maintains parameter efficiency by redistributing a fixed prompt capacity across multiple groups, achieving state-of-the-art performance with the lowest communication parameters (2.6k) among all federated prompt learning methods. Theoretical analysis shows that our dynamic aggregation strategy promotes robust global representation learning by reinforcing shared semantics while suppressing client-specific noise. Extensive experiments demonstrate that FedMGP consistently outperforms prior approaches in both personalization and domain generalization across diverse federated vision-language benchmarks.



Paperid:2692
Authors:Mingge Lu, Jingwei Sun, Junqing Lin, Zechun Zhou, Guangzhong Sun
Abstract:
Abstract:Large Language Models (LLMs) have demonstrated remarkable capabilities, yet their extensive parameter scales pose significant challenges for practical deployment. Unstructured pruning has emerged as an effective model compression strategy with minimal performance loss, which introduces fine-grained sparsity for weight parameters. While existing methods employ a layer-wise pruning strategy to avoid the complexity of global pruning for billion-scale LLMs, they require appropriate sparsity allocation for the layer-wise pruning objectives and often lead to suboptimal solutions for the overall model. In this paper, we propose Lua-LLM ($\textbf{L}$earning $\textbf{u}$nstructured-sparsity $\textbf{a}$llocation in LLMs), a learning-based global pruning framework that explores the optimal unstructured sparsity allocation. Unlike existing pruning methods, which primarily focus on allocating per-layer sparsity, Lua-LLM achieves flexible allocation for both layer-wise and intra-layer sparsity. Furthermore, Lua-LLM leverages a soft Top-K operator to approximate the importance-based mask selection mechanism, enabling efficient binary mask learning. Experimental results on LLaMA and OPT families demonstrate significant performance improvements over existing methods. Compared to the state-of-the-art sparsity allocation method ATP, Lua-LLM reduces the perplexity from 425.12 to 30.27 for LLaMA2-7B model at 80% sparsity level, and improves the average zero-shot accuracy by 3.87% for the 70% sparse LLaMA3-8B model.



Paperid:2693
Authors:Ethan Mendes, Alan Ritter
Abstract:
Collecting ground-truth rewards or human demonstrations for multi-step reasoning tasks is often prohibitively expensive and time-consuming, especially in interactive domains like web tasks. To address this bottleneck, we present self-taught lookahead (STL), a self-supervised method that leverages state-transition dynamics to improve a value model capable of effectively guiding language model-controlled search without any labeled data. We find that moderately sized (8 billion parameters) open-weight value models improved with STL can match the performance of using a gpt-4o value model. Furthermore, we find that specialized value models learned with STL can be deployed with computationally lightweight search algorithms, achieving performance that matches that of more expensive tree search methods, while reducing costs by an order of magnitude.



Paperid:2694
Authors:Jiaxin Song, Yixu Wang, Jie Li, Xuan Tong, rui yu, Yan Teng, Xingjun Ma, Yingchun Wang
Abstract:
Vision-Language Models (VLMs) exhibit impressive performance, yet the integration of powerful vision encoders has significantly broadened their attack surface, rendering them increasingly susceptible to jailbreak attacks. However, lacking well-defined attack objectives, existing jailbreak methods often struggle with gradient-based strategies prone to local optima and lacking precise directional guidance, and typically decouple visual and textual modalities, thereby limiting their effectiveness by neglecting crucial cross-modal interactions. Inspired by the Eliciting Latent Knowledge (ELK) framework, we posit that VLMs encode safety-relevant information within their internal fusion-layer representations, revealing an implicit safety decision boundary in the latent space. This motivates exploiting boundary to steer model behavior. Accordingly, we propose \textbf{JailBound}, a novel latent space jailbreak framework comprising two stages: (1) \textbf{Safety Boundary Probing}, which addresses the guidance issue by approximating decision boundary within fusion layer's latent space, thereby identifying optimal perturbation directions towards the target region; and (2) \textbf{Safety Boundary Crossing}, which overcomes the limitations of decoupled approaches by jointly optimizing adversarial perturbations across both image and text inputs. This latter stage employs an innovative mechanism to steer the model's internal state towards policy-violating outputs while maintaining cross-modal semantic consistency. Extensive experiments on six diverse VLMs demonstrate JailBound's efficacy, achieves 94.32\% white-box and 67.28\% black-box attack success averagely, which are 6.17\% and 21.13\% higher than SOTA methods, respectively. Our findings expose a overlooked safety risk in VLMs and highlight the urgent need for more robust defenses. \textcolor{red}{Warning: This paper contains potentially sensitive, harmful and offensive content.}



Paperid:2695
Authors:Haizhou Shi, Yibin Wang, Ligong Han, Huan Zhang, Hao Wang
Abstract:
Estimating the uncertainty of responses from Large Language Models (LLMs) remains a critical challenge. While recent Bayesian methods have demonstrated effectiveness in quantifying uncertainty through low-rank weight updates, they typically require complex fine-tuning or post-training procedures. In this paper, we proposeTraining-FreeBayesianization~(TFB), a simple yet theoretically grounded framework that efficiently transforms trained low-rank adapters into Bayesian ones without additional training. TFB systematically searches for the maximally acceptable level of variance in the weight posterior, constrained within a family of low-rank isotropic Gaussian distributions. Our theoretical analysis shows that under mild conditions, this search process is equivalent to KL-regularized variational optimization, a generalized form of variational inference. Through comprehensive experiments, we show that TFB achieves superior uncertainty estimation and generalization compared to existing methods while eliminating the need for complex Bayesianization training procedures.



Paperid:2696
Authors:Jonathan Grizou, Carlos De la Torre-Ortiz, Tuukka Ruotsalo
Abstract:
We consider the problem of recovering a mental target (e.g., an image of a face) that a participant has in mind from paired EEG (i.e., brain responses) and image (i.e., perceived faces) data collected during interactive sessions without access to labeled information. The problem has been previously explored with labeled data but not via self-calibration, where labeled data is unavailable. Here, we present the first framework and an algorithm, CURSOR, that learns to recover unknown mental targets without access to labeled data or pre-trained decoders. Our experiments on naturalistic images of faces demonstrate that CURSOR can (1) predict image similarity scores that correlate with human perceptual judgments without any label information, (2) use these scores to rank stimuli against an unknown mental target, and (3) generate new stimuli indistinguishable from the unknown mental target (validated via a user study, N=53). We release the brain response data set (N=29), associated face images used as stimuli data, and a codebase to initiate further research on this novel task.



Paperid:2697
Authors:Masahiro Fujisawa, Masaki Adachi, Michael A Osborne
Abstract:
Despite the importance of aligning language models with human preferences, crowd-sourced human feedback is often noisy---for example, preferring less desirable responses---posing a fundamental challenge to alignment. A truly robust alignment objective should yield identical model parameters even under severe label noise, a property known as redescending. We prove that no existing alignment methods satisfy this property. To address this, we propose Hölder-DPO, the first principled alignment loss with a provable redescending property, enabling approximation of the clean data distribution from noisy feedback. The aligned model estimates the likelihood of clean data, providing a theoretically grounded metric for dataset valuation that identifies the location and fraction of mislabels. This metric is gradient-free, enabling scalable and automated human feedback valuation without costly manual verification or clean validation dataset. Hölder-DPO achieves state-of-the-art robust alignment performance while accurately detecting mislabels in controlled datasets. Finally, we apply Hölder-DPO to widely used alignment datasets, revealing substantial noise levels and demonstrating that removing these mislabels significantly improves alignment performance across methods.



Paperid:2698
Authors:Sangho Lee, Kyeongseo Min, Youngdoo Son, Hyungrok Do
Abstract:
Time series are often irregularly sampled with uneven time intervals. In multivariate cases, such irregularities may lead to misaligned observations across variables and varying observation counts, making it difficult to extract intrinsic patterns and degrading the classification performance of deep learning models. In this study, we propose an adaptive time encoding approach to address the challenge of irregular sampling in multivariate time-series classification. Our approach generates latent representations at learnable reference points that capture missingness patterns in irregular sequences, enhancing classification performance. We also introduce consistency regularization techniques to incorporate intricate temporal and intervariable information into the learned representations. Extensive experiments demonstrate that our method achieves state-of-the-art performance with high computational efficiency in irregular multivariate time-series classification tasks.



Paperid:2699
Authors:Yang-Che Sun, Cheng-Yu Yeo, Ernie Chu, Jun-Cheng Chen, Yu-Lun Liu
Abstract:
In this work, we propose using a unified representation, termedFactorized Features, for low-level vision tasks, where we test onSingle Image Super-Resolution (SISR)andImage Compression. Motivated by the shared principles between these tasks, they require recovering and preserving fine image details, whether by enhancing resolution for SISR or reconstructing compressed data for Image Compression. Unlike previous methods that mainly focus on network architecture, our proposed approach utilizes a basis-coefficient decomposition as well as an explicit formulation of frequencies to capture structural components and multi-scale visual features in images, which addresses the core challenges of both tasks. We replace the representation of prior models from simple feature maps with Factorized Features to validate the potential for broad generalizability. In addition, we further optimize the compression pipeline by leveraging the mergeable-basis property of our Factorized Features, which consolidates shared structures on multi-frame compression. Extensive experiments show that our unified representation delivers state-of-the-art performance, achieving an average relative improvement of 204.4\% in PSNR over the baseline in Super-Resolution (SR) and 9.35\% BD-rate reduction in Image Compression compared to the previous SOTA.



Paperid:2700
Authors:Xu Cai, Yang Wu, Qianli Chen, Haoran Wu, Lichuan Xiang, Hongkai Wen
Abstract:
We present an ultra-efficient distillation method for shortcutting large-scale pre-trained flow matching diffusion models into efficient few-step samplers, enabled by novel velocity field self-distillation. While shortcutting in flow matching, originally introduced by shortcut models, offers flexible trajectory-skipping capabilities, it requires a specialized step-size embedding incompatible with existing models unless retraining from scratch—a process nearly as costly as pretraining itself.Our key contribution is thus imparting a more aggressive shortcut mechanism to standard flow matching models (e.g., Flux), leveraging a unique distillation principle that obviates the need for step-size embedding. Working on the velocity field rather than sample space and learning rapidly from self-guided distillation in an online manner, our approach trains efficiently, e.g., producing a 3-step Flux <1 A100 day. This fast training immediately enables, to our knowledge, the first few-shot distillation method (e.g., 10 text-image pairs) for dozen-billion-parameter diffusion models, delivering state-of-the-art performance at almost free cost.



Paperid:2701
Authors:Matthew Morris, Ian Horrocks
Abstract:
Graph neural networks (GNNs) are frequently used for knowledge graph completion.Their black-box nature has motivated work that uses sound logical rules to explain predictions and characterise their expressivity.However, despite the prevalence of GNNs that use mean as an aggregation function, explainability and expressivity results are lacking for them.We consider GNNs with mean aggregation and non-negative weights (MAGNNs), proving the precise class of monotonic rules that can be sound for them, as well as providing a restricted fragment of first-order logic to explain any MAGNN prediction.Our experiments show that restricting mean-aggregation GNNs to have non-negative weights does not significantly impact performance on common benchmarks, that sound rules are obtained in practice, that insightful explanations can be generated in practice, and that the sound rules can expose issues in the trained models.



Paperid:2702
Authors:Sheng Huang, Lele Fu, Fanghua Ye, Tianchi Liao, Bowen Deng, zhangchuanfu, Chuan Chen
Abstract:
Federated learning enables collaborative training while preserving the privacy of all participants. However, the heterogeneity in data distribution across multiple training nodes poses significant challenges to the construction of federated models. Prior studies were dedicated to mitigating the effects of data heterogeneity by using global information as a blueprint and restricting the local update of the model for reaching a "hard consensus". But this practice makes it difficult to balance local and global information, and it neglects to negotiate amicably between local and global models to reach mutually agreeable results, called ``soft consensus". In this paper, a multiple-path solving method is proposed to balance global and local features and combine these two feature preference paths to reach a soft consensus. Rather than relying on global information as the sole criterion, a negotiation process is employed to address the same objective by accommodating diverse feature preferences, thereby facilitating the discovery of a more plausible solution through multiple distinct pathways. Considering the overwhelming power of local features during local training, a swapping strategy is applied to weaken them to balance the solution paths. Moreover, to minimize the additional communication cost caused by the introduction of multiple paths, the solution of the task network is converted into data adaptation to reduce the amount of parameter transmission. Extensive experiments are conducted to demonstrate the advantages of the proposed method.



Paperid:2703
Authors:Jiahui Wang, Changhao Chen
Abstract:
Abstract:Visual navigation is essential for robotics and embodied AI. However, existing foundation models, particularly those with transformer decoders, suffer from high computational overhead and lack interpretability, limiting their deployment on edge devices. To address this, we propose DynaNav, a Dynamic Visual Navigation framework that adapts feature and layer selection based on scene complexity. It employs a trainable hard feature selector for sparse operations, enhancing efficiency and interpretability. Additionally, we integrate feature selection into an early-exit mechanism, with Bayesian Optimization determining optimal exit thresholds to reduce computational cost. Extensive experiments in real-world-based datasets and simulated environments demonstrate the effectiveness of DynaNav. Compared to ViNT, DynaNav achieves a $2.6\times$ reduction in FLOPs, 42.3% lower inference time, and 32.8% lower memory usage while improving navigation performance across four public datasets.



Paperid:2704
Authors:Zohar Rimon, Elisei Shafer, Efrat Shimron, Tal Tepper, Aviv Tamar
Abstract:
Palpation, the use of touch in medical examination, is almost exclusively performed by humans. We investigate a proof of concept for an artificial palpation method based on self-supervised learning. Our key idea is that an encoder-decoder framework can learn arepresentationfrom a sequence of tactile measurements that contains all the relevant information about the palpated object. We conjecture that such a representation can be used for downstream tasks such as tactile imaging and change detection. With enough training data, it should capture intricate patterns in the tactile measurements that go beyond a simple map of forces -- the current state of the art. To validate our approach, we both develop a simulation environment and collect a real-world dataset of soft objects and corresponding ground truth images obtained by magnetic resonance imaging (MRI). We collect palpation sequences using a robot equipped with a tactile sensor, and train a model that predicts sensory readings at different positions on the object. We investigate the representation learned in this process, and demonstrate its use in imaging and change detection.



Paperid:2705
Authors:Haoyu He, Haozheng Luo, Yan Chen, Qi Wang
Abstract:
Predicting human mobility is inherently challenging due to complex long-range dependencies and multi-scale periodic behaviors. To address this, we introduce RHYTHM (Reasoning with Hierarchical Temporal Tokenization for Human Mobility), a unified framework that leverages large language models (LLMs) as general-purpose spatio-temporal predictors and trajectory reasoners. Methodologically, RHYTHM employs temporal tokenization to partition each trajectory into daily segments and encode them as discrete tokens with hierarchical attention that captures both daily and weekly dependencies, thereby significantly reducing the sequence length while preserving cyclical information. Additionally, we enrich token representations by adding pre-computed prompt embeddings for trajectory segments and prediction targets via a frozen LLM, and feeding these combined embeddings back into the LLM backbone to capture complex interdependencies. Computationally, RHYTHM freezes the pretrained LLM's backbone to reduce attention complexity and memory cost. We evaluate our model against state-of-the-art methods using three real-world datasets. Notably, RHYTHM achieves a 2.4% improvement in overall accuracy, a 5.0% increase on weekends, and a 24.6% reduction in training time.



Paperid:2706
Authors:Minh Nguyen, Chandrajit Bajaj
Abstract:
Abstract:Reinforcement learning (RL) in continuous state-action spaces remains challenging in scientific computing due to poor sample efficiency and lack of pathwise physical consistency. We introduce Differential Reinforcement Learning (Differential RL), a novel framework that reformulates RL from a continuous-time control perspective via a differential dual formulation. This induces a Hamiltonian structure that embeds physics priors and ensures consistent trajectories without requiring explicit constraints. To implement Differential RL, we develop Differential Policy Optimization (DPO), a pointwise, stage-wise algorithm that refines local movement operators along the trajectory for improved sample efficiency and dynamic alignment. We establish pointwise convergence guarantees, a property not available in standard RL, and derive a competitive theoretical regret bound of $O(K^{5/6})$. Empirically, DPO outperforms standard RL baselines on representative scientific computing tasks, including surface modeling, grid control, and molecular dynamics, under low-data and physics-constrained conditions.



Paperid:2707
Authors:Huanyu Liu, Jia Li, Hao Zhu, Kechi Zhang, Yihong Dong, Ge Li
Abstract:
How to design reinforcement learning (RL) tasks that effectively unleash the reasoning capability of large language models (LLMs) remains an open question. Existing RL tasks (e.g., math, programming, and constructing reasoning tasks) suffer from three key limitations: (1) Scalability. They rely heavily on human annotation or expensive LLM synthesis to generate sufficient training data. (2) Verifiability. LLMs' outputs are hard to verify automatically and reliably. (3) Controllable Difficulty. Most tasks lack fine-grained difficulty control, making it hard to train LLMs to develop reasoning ability from easy to hard.To address these limitations, we propose Saturn, a SAT-based RL framework that uses Boolean Satisfiability (SAT) problems to train and evaluate LLM reasoning. Saturn enables scalable task construction, rule-based verification, and precise difficulty control. Saturn designs a curriculum learning pipeline that continuously improves LLMs' reasoning capability by constructing SAT tasks of increasing difficulty and training LLMs from easy to hard. To ensure stable training, we design a principled mechanism to control difficulty transitions.We introduce Saturn-2.6k, a dataset of 2,660 SAT problems with varying difficulty. It supports the evaluation of how LLM reasoning changes with problem difficulty. We apply Saturn to DeepSeek-R1-Distill-Qwen and obtain Saturn-1.5B and Saturn-7B. We achieve several notable results:(1) On SAT problems, Saturn-1.5B and Saturn-7B achieve average pass@3 improvements of +14.0 and +28.1, respectively.(2) On math and programming tasks, Saturn-1.5B and Saturn-7B improve average scores by +4.9 and +1.8 on benchmarks (e.g., AIME, LiveCodeBench).(3) Compared to the state-of-the-art (SOTA) approach in constructing RL tasks, Saturn achieves further improvements of +8.8\%.We release the source code, data, and models to support future research.



Paperid:2708
Authors:Yuxi Liu, jin dengchao, Shuai Huo, Jiawen Gu, Chao Zhou, Huihui Bai, Ming Lu, Zhan Ma
Abstract:
Neural video compression (NVC) has made significant progress in recent years, while neural B-frame video compression (NBVC) remains underexplored compared to P-frame compression. NBVC can adopt bi-directional reference frames for better compression performance. However, NBVC's hierarchical coding may complicate continuous temporal prediction, especially at some hierarchical levels with a large frame span, which could cause the contribution of the two reference frames to be unbalanced. To optimize reference information utilization, we propose a novel NBVC method, termed Bi-directional Reference Harmonization Video Compression (BRHVC), with the proposed Bi-directional Motion Converge (BMC) and Bi-directional Contextual Fusion (BCF). BMC converges multiple optical flows in motion compression, leading to more accurate motion compensation on a larger scale. Then BCF explicitly models the weights of reference contexts under the guidance of motion compensation accuracy. With more efficient motions and contexts, BRHVC can effectively harmonize bi-directional references. Experimental results indicate that our BRHVC outperforms previous state-of-the-art NVC methods, even surpassing the traditional coding, VTM-RA (under random access configuration), on the HEVC datasets. The source code will be released.



Paperid:2709
Authors:Sangli Teng, Harry Zhang, David Jin, Ashkan Jasour, Ram Vasudevan, Maani Ghaffari, Luca Carlone
Abstract:
Designing optimal Bayes filters for nonlinear non-Gaussian systems is a challenging task. The main difficulties are: 1) representing complex beliefs, 2) handling non-Gaussian noise, and 3) marginalizing past states. To address these challenges, we focus on polynomial systems and propose the Max Entropy Moment Kalman Filter (MEM-KF). To address 1), we represent arbitrary beliefs by a Moment-Constrained Max-Entropy Distribution (MED). The MED can asymptotically approximate almost any distribution given an increasing number of moment constraints. To address 2), we model the noise in the process and observation model as MED. To address 3), we propagate the moments through the process model and recover the distribution as MED, thus avoiding symbolic integration, which is generally intractable. All the steps in MEM-KF, including the extraction of a point estimate, can be solved via convex optimization. We showcase the MEM-KF in challenging robotics tasks, such as localization with unknown data association.



Paperid:2710
Authors:Zhicheng Zhang, Weicheng Wang, Yongjie Zhu, Wenyu Qin, Pengfei Wan, Di ZHANG, Jufeng Yang
Abstract:
Understanding and predicting emotions from videos has gathered significant attention in recent studies, driven by advancements in video large language models (VideoLLMs). While advanced methods have made progress in video emotion analysis, the intrinsic nature of emotions—characterized by their open-set, dynamic, and context-dependent properties—poses challenge in understanding complex and evolving emotional states with reasonable rationale. To tackle these challenges, we propose a novel affective cues-guided reasoning framework that unifies fundamental attribute perception, expression analysis, and high-level emotional understanding in a stage-wise manner. At the core of our approach is a family of video emotion foundation models (VidEmo), specifically designed for emotion reasoning and instruction-following. These models undergo a two-stage tuning process: first, curriculum emotion learning for injecting emotion knowledge, followed by affective-tree reinforcement learning for emotion reasoning. Moreover, we establish a foundational data infrastructure and introduce a emotion-centric fine-grained dataset (Emo-CFG) consisting of 2.1M diverse instruction-based samples. Emo-CFG includes explainable emotional question-answering, fine-grained captions, and associated rationales, providing essential resources for advancing emotion understanding tasks. Experimental results demonstrate that our approach achieves competitive performance, setting a new milestone across 15 face perception tasks. All foundation models will be made available to the community, with source code and video demos provided in the supplemental.



Paperid:2711
Authors:Zhekai Chen, Ruihang Chu, Yukang Chen, Shiwei Zhang, Yujie Wei, Yingya Zhang, Xihui Liu
Abstract:
Scaling visual generation models is essential for real-world content creation, yet requires substantial training and computational expenses. Alternatively, test-time scaling has garnered growing attention due to resource efficiency and promising performance. In this work, we present the first general test-time scaling framework for visual auto-regressive (VAR) models, TTS-VAR, modeling the generation process as a path searching problem. Inspired by VAR's hierarchical coarse-to-fine multi-scale generation, our framework integrates two key components: (i) At coarse scales, we observe that generated tokens are hard for evaluation, possibly leading to erroneous acceptance of inferior samples or rejection of superior samples. Noticing that the coarse scales contain sufficient structural information, we propose clustering-based diversity search. It preserves structural variety through semantic feature clustering, enabling later selection on samples with higher potential. (ii) In fine scales, resampling-based potential selection prioritizes promising candidates using potential scores, which are defined as reward functions incorporating multi-scale generation history. To dynamically balance computational efficiency with exploration capacity, we further introduce an adaptive descending batch size schedule throughout the causal generation process. Experiments on the powerful VAR model Infinity show a notable 8.7% GenEval score improvement (0.69→0.75). Key insights reveal that early-stage structural features effectively influence final quality, and resampling efficacy varies across generation scales.



Paperid:2712
Authors:Sebastian Ojeda, Rafael Velasquez, Nicolas Aparicio, Juanita Puentes, Paula Cárdenas, Nicolás Andrade, Gabriel González, Sergio Rincón, Carolina Muñoz-Camargo, Pablo Arbelaez
Abstract:
Antimicrobial peptides have emerged as promising molecules to combat antimicrobial resistance. However, fragmented datasets, inconsistent annotations, and the lack of standardized benchmarks hinder computational approaches and slow down the discovery of new candidates. To address these challenges, we present the Expanded Standardized Collection for Antimicrobial Peptide Evaluation (ESCAPE), an experimental framework integrating over 80.000 peptides from 27 validated repositories. Our dataset separates antimicrobial peptides from negative sequences and incorporates their functional annotations into a biologically coherent multilabel hierarchy, capturing activities across antibacterial, antifungal, antiviral, and antiparasitic classes. Building on ESCAPE, we propose a transformer-based model that leverages sequence and structural information to predict multiple functional activities of peptides. Our method achieves up to a 2.15% improvement in mean Average Precision over previous methods adapted for this task, establishing a new state-of-the-art multilabel peptide classification. ESCAPE provides a comprehensive and reproducible evaluation framework to advance AI-driven antimicrobial peptide research.



Paperid:2713
Authors:Jiyeon Kang, Songseong Kim, Chanhui Lee, Doyeong Hwang, Joanie Chung, Yunkyung Ko, Sumin Lee, Sungwoong Kim, Sungbin Lim
Abstract:
Ordering-based approaches to causal discovery identify topological orders of causal graphs, providing scalable alternatives to combinatorial search methods. Under the Additive Noise Models (ANMs) assumption, recent causal ordering methods based on score matching require an accurate estimation of the Hessian diagonal of the log-densities. However, previous approaches mainly use Stein gradient estimators, which are computationally expensive and memory-intensive. Although DiffAN addresses these limitations by substituting kernel-based estimates with diffusion models, it remains numerically unstable due to the second-order derivatives of score models. To alleviate these problems, we propose Score-informed Neural Operator (SciNO), a probabilistic generative model in smooth function spaces designed to stably approximate the Hessian diagonal and to preserve structural information during the score modeling. Empirical results show that SciNO reduces order divergence by 42.7% on synthetic graphs and by 31.5% in real-world datasets on average compared to DiffAN, while maintaining memory efficiency and scalability. Furthermore, we propose a probabilistic control algorithm for causal reasoning with autoregressive models that integrates SciNO's probability estimates with autoregressive model priors, enabling reliable data-driven causal ordering informed by semantic information. Consequently, the proposed method enhances causal reasoning abilities of LLMs without additional fine-tuning or prompt engineering.



Paperid:2714
Authors:Nathan Corecco, Batuhan Yardim, Vinzenz Thoma, Zebang Shen, Niao He
Abstract:
Abstract:Designing incentives for a multi-agent system to induce a desirable Nash equilibrium is both a crucial and challenging problem appearing in many decision-making domains, especially for large numbers of agents $N$.Under the exchangeability assumption, we formalize this incentive design (ID) problem as a parameterized mean-field game (PMFG), aiming to reduce complexity via an infinite-population limit.We first show that when dynamics and rewards are Lipschitz, the finite-$N$ ID objective is approximated by the PMFG at rate $\mathcal{O}(\frac{1}{\sqrt{N}})$.Moreover, beyond the Lipschitz-continuous setting, we prove the same $\mathcal{O}(\frac{1}{\sqrt{N}})$ decay for the important special case of sequential auctions, despite discontinuities in dynamics, through a tailored auction-specific analysis.Built on our novel approximation results, we further introduce our Adjoint Mean-Field Incentive Design (AMID) algorithm, which uses explicit differentiation of iterated equilibrium operators to compute gradients efficiently.By uniting approximation bounds with optimization guarantees, AMID delivers a powerful, scalable algorithmic tool for large-$N$ ID.Across diverse auction settings, the proposed AMID method substantially increases revenue over first-price formats and outperforms existing benchmark methods.



Paperid:2715
Authors:Seewon Choi, Alaia Solko-Breslin, Rajeev Alur, Eric Wong
Abstract:
Many computational tasks benefit from being formulated as the composition of neural networks followed by a discrete symbolic program. The goal of neurosymbolic learning is to train the neural networks using end-to-end input-output labels of the composite. We introduce CTSketch, a novel, scalable neurosymbolic learning algorithm. CTSketch uses two techniques to improve the scalability of neurosymbolic inference: decompose the symbolic program into sub-programs and summarize each sub-program with a sketched tensor. This strategy allows us to approximate the output distribution of the program with simple tensor operations over the input distributions and the sketches. We provide theoretical insight into the maximum approximation error. Furthermore, we evaluate CTSketch on benchmarks from the neurosymbolic learning literature, including some designed for evaluating scalability. Our results show that CTSketch pushes neurosymbolic learning to new scales that were previously unattainable, with neural predictors obtaining high accuracy on tasks with over one thousand inputs, despite supervision only on the final output.



Paperid:2716
Authors:Yuchen Wu, Edward Sun, Kaijie Zhu, Jianxun Lian, José Hernández-Orallo, Aylin Caliskan, Jindong Wang
Abstract:
Large language models (LLMs) typically generate identical or similar responses for all users given the same prompt, posing serious safety risks in high-stakes applications where user vulnerabilities differ widely.Existing safety evaluations primarily rely on context-independent metrics—such as factuality, bias, or toxicity—overlooking the fact that the same response may carry divergent risks depending on the user's background or condition.We introduce ``personalized safety'' to fill this gap and present PENGUIN—a benchmark comprising 14,000 scenarios across seven sensitive domains with both context-rich and context-free variants. Evaluating six leading LLMs, we demonstrate that personalized user information significantly improves safety scores by 43.2%, confirming the effectiveness of personalization in safety alignment. However, not all context attributes contribute equally to safety enhancement. To address this, we develop RAISE—a training-free, two-stage agent framework that strategically acquires user-specific background. RAISE improves safety scores by up to 31.6% over six vanilla LLMs, while maintaining a low interaction cost of just 2.7 user queries on average. Our findings highlight the importance of selective information gathering in safety-critical domains and offer a practical solution for personalizing LLM responses without model retraining. This work establishes a foundation for safety research that adapts to individual user contexts rather than assuming a universal harm standard.



Paperid:2717
Authors:Na Li, Zewu Zheng, Wei Ni, Hangguan Shan, Wenjie Zhang, Xinyu Li
Abstract:
Multi-agent reinforcement learning (MARL), as a thriving field, explores how multiple agents independently make decisions in a shared dynamic environment. Due to environmental uncertainties, policies in MARL must remain robust to tackle the sim-to-real gap. We focus on robust two-player zero-sum Markov games (TZMGs) in offline settings, specifically on tabular robust TZMGs (RTZMGs). We propose a model-based algorithm (RTZ-VI-LCB) for offline RTZMGs, which is an optimistic robust value iteration combined with the data-driven Bernstein-style penalty term for robust value estimation. By accounting for distribution shifts in the historical dataset, the proposed algorithm establishes near-optimal sample complexity guarantees under partial coverage and environmental uncertainty. An information-theoretic lower bound is developed to confirm the tightness of our algorithm's sample complexity, which is optimal regarding both state and action spaces. To the best of our knowledge, our algorithm is the first to attain this optimality and establishes a new benchmark for offline RTZMGs. Numerical experiments validate the proposed algorithm.



Paperid:2718
Authors:Mengjingcheng Mo, Xinyang Tong, Mingpi Tan, Jiaxu Leng, JianKang Zheng, Yiran Liu, Haosheng Chen, Ji Gan, Weisheng Li, Xinbo Gao
Abstract:
While unmanned aerial vehicles (UAVs) offer wide-area, high-altitude coverage for anomaly detection, they face challenges such as dynamic viewpoints, scale variations, and complex scenes. Existing datasets and methods, mainly designed for fixed ground-level views, struggle to adapt to these conditions, leading to significant performance drops in drone-view scenarios.To bridge this gap, we introduce A2Seek (Aerial Anomaly Seek), a large-scale, reasoning-centric benchmark dataset for aerial anomaly understanding. This dataset covers various scenarios and environmental conditions, providing high-resolution real-world aerial videos with detailed annotations, including anomaly categories, frame-level timestamps, region-level bounding boxes, and natural language explanations for causal reasoning. Building on this dataset, we propose A2Seek-R1, a novel reasoning framework that generalizes R1-style strategies to aerial anomaly understanding, enabling a deeper understanding of "Where" anomalies occur and "Why" they happen in aerial frames.To this end, A2Seek-R1 first employs a graph-of-thought (GoT)-guided supervised fine-tuning approach to activate the model's latent reasoning capabilities on A2Seek. Then, we introduce Aerial Group Relative Policy Optimization (A-GRPO) to design rule-based reward functions tailored to aerial scenarios. Furthermore, we propose a novel "seeking" mechanism that simulates UAV flight behavior by directing the model's attention to informative regions.Extensive experiments demonstrate that A2Seek-R1 achieves up to a 22.04% improvement in AP for prediction accuracy and a 13.9% gain in mIoU for anomaly localization, exhibiting strong generalization across complex environments and out-of-distribution scenarios.



Paperid:2719
Authors:TaeHo Yoon, Sayantan Choudhury, Nicolas Loizou
Abstract:
Traditional Federated Learning (FL) approaches assume collaborative clients with aligned objectives working towards a shared global model. However, in many real-world scenarios, clients act as rational players with individual objectives and strategic behaviors, a concept that existing FL frameworks are not equipped to adequately address. To bridge this gap, we introduceMultiplayer Federated Learning (MpFL), a novel framework that models the clients in the FL environment as players in a game-theoretic context, aiming to reach an equilibrium. In this scenario, each player tries to optimize their own utility function, which may not align with the collective goal. Within MpFL, we proposePer-Player Local Stochastic Gradient Descent (PEARL-SGD), an algorithm in which each player/client performs local updates independently and periodically communicates with other players. We theoretically analyze PEARL-SGD and prove that it reaches a neighborhood of equilibrium with less communication in the stochastic setup compared to its non-local counterpart. Finally, we verify our theoretical findings through numerical experiments.



Paperid:2720
Authors:Guan Yandong, XiMing Xing, Xilin Wang, Jing Zhang, Dong Xu, Qian Yu
Abstract:
In this work, we introduce CAD-Coder, a novel framework that reformulates text-to-CAD as the generation of CadQuery scripts—a Python-based, parametric CAD language.This representation enables direct geometric validation, a richer modeling vocabulary, and seamless integration with existing LLMs. To further enhance code validity and geometric fidelity, we propose a two-stage learning pipeline: (1) supervised fine-tuning on paired text–CadQuery data, and (2) reinforcement learning with Group Reward Policy Optimization (GRPO), guided by a CAD-specific reward comprising both a geometric reward (Chamfer Distance) and a format reward.We also introduce a chain-of-thought (CoT) planning process to improve model reasoning, and construct a large-scale, high-quality dataset of 110K text–CadQuery–3D model triplets and 1.5K CoT samples via an automated pipeline. Extensive experiments demonstrate that CAD-Coder enables LLMs to generate diverse, valid, and complex CAD models directly from natural language, advancing the state of the art of text-to-CAD generation and geometric reasoning.



Paperid:2721
Authors:Kaiwen Zha, Zhengqi Gao, Maohao Shen, Zhang-Wei Hong, Duane Boning, Dina Katabi
Abstract:
Reinforcement learning (RL) has recently emerged as a compelling approach for enhancing the reasoning capabilities of large language models (LLMs), where an LLM generator serves as a policy guided by a verifier (reward model). However, current RL post-training methods for LLMs typically use verifiers that are fixed (rule-based or frozen pretrained) or trained discriminatively via supervised fine-tuning (SFT). Such designs are susceptible to reward hacking and generalize poorly beyond their training distributions. To overcome these limitations, we propose Tango, a novel framework that uses RL to concurrently train both an LLM generator and a verifier in an interleaved manner. A central innovation of Tango is its generative, process-level LLM verifier, which is trained via RL and co-evolves with the generator. Importantly, the verifier is trained solely based on outcome-level verification correctness rewards without requiring explicit process-level annotations. This generative RL-trained verifier exhibits improved robustness and superior generalization compared to deterministic or SFT-trained verifiers, fostering effective mutual reinforcement with the generator. Extensive experiments demonstrate that both components of Tango achieve state-of-the-art results among 7B/8B-scale models: the generator attains best-in-class performance across five competition-level math benchmarks and four challenging out-of-domain reasoning tasks, while the verifier leads on the ProcessBench dataset. Remarkably, both components exhibit particularly substantial improvements on the most difficult mathematical reasoning problems.



Paperid:2722
Authors:Lucas Morisset, Adrien Hardy, Alain Durmus
Abstract:
This paper addresses the problem of inverse covariance (also known as precision matrix) estimation in high-dimensional settings. Specifically, we focus on two classes of estimators: linear shrinkage estimators with a target proportional to the identity matrix, and estimators derived from data augmentation (DA). Here, DA refers to the common practice of enriching a dataset with artificial samples—typically generated via a generative model or through random transformations of the original data—prior to model fitting.For both classes of estimators, we derive estimators and provide concentration bounds for their quadratic error. This allows for both method comparison and hyperparameter tuning, such as selecting the optimal proportion of artificial samples.On the technical side, our analysis relies on tools from random matrix theory. We introduce a novel deterministic equivalent for generalized resolvent matrices, accommodating dependent samples with specific structure. We support our theoretical results with numerical experiments.



Paperid:2723
Authors:Federico Di Gennaro, Khaled Eldowa, Nicolò Cesa-Bianchi
Abstract:
Abstract:In contrast to the classic formulation of partial monitoring, linear partial monitoring can model infinite outcome spaces, while imposing a linear structure on both the losses and the observations. This setting can be viewed as a generalization of linear bandits where loss and feedback are decoupled in a flexible manner. In this work, we address a nonstochastic (adversarial), finite-actions version of the problem through a simple instance of the exploration-by-optimization method that is amenable to efficient implementation. We derive regret bounds that depend on the game structure in a more transparent manner than previous theoretical guarantees for this paradigm. Our bounds feature instance-specific quantities that reflect the degree of alignment between observations and losses, and resemble known guarantees in the stochastic setting. Notably, they achieve the standard $\sqrt{T}$ rate in easy (locally observable) games and $T^{2/3}$ in hard (globally observable) games, where $T$ is the time horizon. We instantiate these bounds in a selection of old and new partial information settings subsumed by this model, and illustrate that the achieved dependence on the game structure can be tight in interesting cases.



Paperid:2724
Authors:Vaggelis Dorovatas, Soroush Seifi, Gunshi Gupta, Rahaf Aljundi
Abstract:
Video Large Language Models (Video-LLMs) excel at understanding videos in-context, assuming full access to the video when answering queries. However, these models face challenges in streaming scenarios where hour-long videos must be processed online, and questions need timely responses. In this work, we propose a training-free approach compatible with standard Video-LLMs, leveraging three key concepts: 1) LLM-informed selection of visual tokens to identify those that the LLM has attended to and contributed to its understanding of each short clip. Our attention-based selection allows us to discard up to ~95\% of unimportant visual tokens with minimal performance loss; 2) Hierarchical selection of tokens combined with natural language understanding of each processed clip; 3) Caption-based question answering for lightweight and accurate responses. Our method achieves state-of-the-art performance on streaming video benchmarks, striking a balance between efficiency and effectiveness.



Paperid:2725
Authors:Noam Elata, Hyungjin Chung, Jong Chul Ye, Tomer Michaeli, Michael Elad
Abstract:
Diffusion Models have demonstrated remarkable capabilities in handling inverse problems, offering high-quality posterior-sampling-based solutions. Despite significant advances, a fundamental trade-off persists regarding the way the conditioned synthesis is employed: Zero-shot approaches can accommodate any linear degradation but rely on approximations that reduce accuracy. In contrast, training-based methods model the posterior correctly, but cannot adapt to the degradation at test-time. Here we introduce InvFussion, the first training-based degradation-aware posterior sampler. InvFussion combines the best of both worlds - the strong performance of supervised approaches and the flexibility of zero-shot methods. This is achieved through a novel architectural design that seamlessly integrates the degradation operator directly into the diffusion denoiser. We compare InvFussion against existing general-purpose posterior samplers, both degradation-aware zero-shot techniques and blind training-based methods. Experiments on the FFHQ and ImageNet datasets demonstrate state-of-the-art performance. Beyond posterior sampling, we further demonstrate the applicability of our architecture, operating as a general Minimum Mean Square Error predictor, and as a Neural Posterior Principal Component estimator.



Paperid:2726
Authors:Jaehyun Park, Konyul Park, Daehun Kim, Junseo Park, Jun Won Choi
Abstract:
In autonomous driving scenarios, ensuring transparency in the decision-making process of perception model is critical, where even a single misperception can lead to life-threatening consequences. However, when data from multiple sensors are used as input, it becomes exceedingly difficult to identify the contribution of each modality to the decision. This challenge stems from the entangled representation of sensor information within the perception model. In this study, we propose Layer-Wise Modality Decomposition (LMD), a post-hoc interpretability method that disentangles modality-specific information across all layers of a pretrained fusion model. Our LMD is the first model-agnostic method to attribute the prediction of a perception model to individual input modalities in autonomous driving scenarios. We evaluate LMD on pretrained sensor fusion models in autonomous driving by applying it to both camera-radar and camera-LiDAR settings. LMD is validated through structured perturbation-based metrics and modality-wise visual decompositions, demonstrating its applicability to interpreting high-capacity multimodal architectures beyond the driving domain.



Paperid:2727
Authors:Jingjing Zheng, Wanglong Lu, Yiming Dong, Chaojie Ji, Yankai Cao, Zhouchen Lin
Abstract:
In this paper, we propose AdaMSS, an adaptive multi-subspace approach for parameter-efficient fine-tuning of large models. Unlike traditional parameter-efficient fine-tuning methods that operate within a large single subspace of the network weights, AdaMSS leverages subspace segmentation to obtain multiple smaller subspaces and adaptively reduces the number of trainable parameters during training, ultimately updating only those associated with a small subset of subspaces most relevant to the target downstream task. By using the lowest-rank representation, AdaMSS achieves more compact expressiveness and finer tuning of the model parameters. Theoretical analyses demonstrate that AdaMSS has better generalization guarantee than LoRA, PiSSA, and other single-subspace low-rank-based methods. Extensive experiments across image classification, natural language understanding, and natural language generation tasks show that AdaMSS achieves comparable performance to full fine-tuning and outperforms other parameter-efficient fine-tuning methods in most cases, all while requiring fewer trainable parameters. Notably, on the ViT-Large model, AdaMSS achieves 4.7\% higher average accuracy than LoRA across seven tasks, using just 15.4\% of the trainable parameters. On RoBERTa-Large, AdaMSS outperforms PiSSA by 7\% in average accuracy across six tasks while reducing the number of trainable parameters by approximately 94.4\%. These results demonstrate the effectiveness of AdaMSS in parameter-efficient fine-tuning.



Paperid:2728
Authors:Siyu Chen, Theodor Misiakiewicz, Ilias Zadik, Peiyuan Zhang
Abstract:
Bandeira et al. (2022) introduced the Franz-Parisi (FP) criterion for characterizing the computational hard phases in statistical detection problems. The FP criterion, based on an annealed version of the celebrated Franz-Parisi potential from statistical physics, was shown to be equivalent to low-degree polynomial (LDP) lower bounds for Gaussian additive models, thereby connecting two distinct approaches to understanding the computational hardness in statistical inference. In this paper, we propose a refined FP criterion that aims to better capture the geometric ``overlap" structure of statistical models. Our main result establishes that this optimized FP criterion is equivalent to Statistical Query (SQ) lower bounds---another foundational framework in computational complexity of statistical inference. Crucially, this equivalence holds under a mild, verifiable assumption satisfied by a broad class of statistical models, including Gaussian additive models, planted sparse models, non-Gaussian component analysis, single-index models, and convex truncation detection. On top of the above, our equivalence not only unifies and simplifies the derivation of several known SQ lower bounds, but also yields new SQ-lower bounds of independent interest.



Paperid:2729
Authors:Rachapun Rotrattanadumrong, Carlo De Donno
Abstract:
Oligonucleotide therapeutics offer great potential of addressing previously undruggable targets and enabling personalized medicine. However, their progress is often hindered by insufficient safety and efficacy profiles. Predictive modeling and machine learning could significantly accelerate oligonucleotide drug discovery by identifying suboptimal compounds early on, but their application in this area lags behind other modalities. A key obstacle to the adoption of machine learning in the field is the scarcity of readily accessible and standardized datasets for model development, as data are often scattered across diverse experiments with inconsistent molecular representations. To overcome this challenge, we introduce OligoGym, a curated collection of standardized, machine learning-ready datasets encompassing various oligonucleotide therapeutic modalities and endpoints. We used OligoGym to benchmark diverse classical and deep learning methods, establishing performance baselines for each dataset across different featurization techniques, model configurations and splitting strategies. Our work represents a crucial first step in creating a more unified framework for oligonucleotide therapeutic dataset generation and model training.



Paperid:2730
Authors:Kangcong Li, Peng Ye, Chongjun Tu, Lin Zhang, Chunfeng Song, Jiamin Wu, Tao Yang, Qihao Zheng, Tao Chen
Abstract:
Abstract:While Large Language Models (LLMs) demonstrate strong performance across domains, their long-context capabilities are limited by transient neural activations causing information decay and unstructured feed-forward network (FFN) weights leading to semantic fragmentation. Inspired by the brain’s working memory and cortical modularity, we propose PaceLLM, featuring two innovations: (1) a Persistent Activity (PA) Mechanism that mimics prefrontal cortex (PFC) neurons’ persistent firing by introducing an activation-level memory bank to dynamically retrieve, reuse, and update critical FFN states, addressing contextual decay; and (2) Cortical Expert (CE) Clustering that emulates task-adaptive neural specialization to reorganize FFN weights into semantic modules, establishing cross-token dependencies and mitigating fragmentation. Extensive evaluations show that PaceLLM achieves 6% improvement on LongBench’s Multi-document QA and 12.5–17.5% performance gains on $\infty$-Bench tasks, while extending measurable context length to 200K tokens in Needle-In-A-Haystack (NIAH) tests. This work pioneers brain-inspired LLM optimization and is complementary to other works. Besides, it can be generalized to any model and enhance their long-context performance and interpretability without structural overhauls.



Paperid:2731
Authors:Junyi Wu, Jiachen Tao, Haoxuan Wang, Gaowen Liu, Ramana Kompella, Yan Yan
Abstract:
We present Orientation-anchored Gaussian Splatting (OriGS), a novel framework for high-quality 4D reconstruction from casually captured monocular videos.While recent advances extend 3D Gaussian Splatting to dynamic scenes via various motion anchors, such as graph nodes or spline control points, they often rely on low-rank assumptions and fall short in modeling complex, region-specific deformations inherent to unconstrained dynamics.OriGS addresses this by introducing a hyperdimensional representation grounded in scene orientation.We first estimate a Global Orientation Field that propagates principal forward directions across space and time, serving as stable structural guidance for dynamic modeling.Built upon this, we propose Orientation-aware Hyper-Gaussian, a unified formulation that embeds time, space, geometry, and orientation into a coherent probabilistic state.This enables inferring region-specific deformation through principled conditioned slicing, adaptively capturing diverse local dynamics in alignment with global motion intent.Experiments demonstrate the superior reconstruction fidelity of OriGS over mainstream methods in challenging real-world dynamic scenes.



Paperid:2732
Authors:Shreelekha Revankar, Utkarsh Mall, Cheng Perng Phoo, Kavita Bala, Bharath Hariharan
Abstract:
Abstract:Natural disasters cause devastating damage to communities and infrastructure every year. Effective disaster response is hampered by the difficulty of accessing affected areas during and after events. Remote sensing has allowed us to monitor natural disasters in a remote way. More recently there have been advances in computer vision and deep learning that help automate satellite imagery analysis, However, they remain limited by their narrow focus on specific disaster types, reliance on manual expert interpretation, and lack of datasets with sufficient temporal granularity or natural language annotations for tracking disaster progression. We present MONITRS, a novel multimodal dataset of $\sim$10,000 FEMA disaster events with temporal satellite imagery with natural language annotations from news articles, accompanied by geotagged locations, and question-answer pairs. We demonstrate that fine-tuning existing MLLMs on our dataset yields significant performance improvements for disaster monitoring tasks, establishing a new benchmark for machine learning-assisted disaster response systems.



Paperid:2733
Authors:Yihe Deng, Hritik Bansal, Fan Yin, Nanyun Peng, Wei Wang, Kai-Wei Chang
Abstract:
We introduceOpenVLThinker, one of the first open-source large vision–language models (LVLMs) to exhibit sophisticated chain-of-thought reasoning, achieving notable performance gains on challenging visual reasoning tasks. While text-based reasoning models (e.g., Deepseek R1) show promising results in text-only tasks, distilling their reasoning into LVLMs via supervised fine-tuning (SFT) often results in performance degradation due to imprecise visual grounding. Conversely, purely reinforcement learning (RL)-based methods face a large search space, hindering the emergence of reflective behaviors in smaller models (e.g., 7B LVLMs). Surprisingly, alternating between SFT and RL ultimately results in significant performance improvements after a few iterations. Our analysis reveals that the base model rarely exhibits reasoning behaviors initially, but SFT effectively surfaces these latent actions and narrows the RL search space, accelerating the development of reasoning capabilities. Each subsequent RL stage further refines the model's reasoning skills, producing higher-quality SFT data for continued self-improvement. OpenVLThinker-7B consistently advances performance across six benchmarks demanding mathematical and general reasoning, notably improving MathVista by 3.2\%, EMMA by 1.4\%, and HallusionBench by 2.7\%. Beyond demonstrating the synergy between SFT and RL for complex reasoning tasks, our findings provide early evidence towards achieving R1-style reasoning in multimodal contexts.



Paperid:2734
Authors:Hantao Yu, Josh Alman
Abstract:
Abstract:The quadratic complexity of self‑attention prevents transformers from scaling effectively to long input sequences. On the other hand, modern GPUs and other specialized hardware accelerators are well-optimized for processing small input sequences in transformers during both training and inference. A natural question arises: can we take advantage of the efficiency of small transformers to deal with long input sequences?In this paper, we show that transformers with long input sequences (large transformers) can be efficiently simulated by transformers that can only take short input sequences (small transformers). Specifically, we prove that any transformer with input length $N$ can be efficiently simulated by only $O((N/M)^2)$ transformers with input length $M \ll N$, and that this cannot be improved in the worst case. However, we then prove that in various natural scenarios including average-case inputs, sliding window masking and attention sinks, the optimal number $O(N/M)$ of small transformers suffice.



Paperid:2735
Authors:Candace Ross, Florian Bordes, Adina Williams, Polina Kirichenko, Mark Ibrahim
Abstract:
Multimodal language models possess a remarkable ability to handle an open-vocabulary worth of objects. Yet the best models still suffer from hallucinations when reasoning about scenes in the real world, revealing a gap between their seemingly strong performance on existing perception benchmarks that are saturating and their reasoning in the real world. To address this gap, we build a novel benchmark of in-the-wild scenes that we call Common-O Bench with more than 10.5k examples using exclusively new images not found in web training data to avoid contamination, Common-O goes beyond just perception, inspired by cognitive tests for humans, to probe reasoning across scenes by asking ``what’s in common?''. We evaluate leading multimodal language models, including models specifically trained to reason. We find that perceiving objects in single images is easy for most models, yet reasoning across scenes is very challenging even for the best models, including reasoning models. Despite saturating many leaderboards focusing on perception, the best performing model only achieves 35\% on Common-O Bench---and on Common-O Complex, consisting of more complex scenes, the best model achieves only 1\%. Curiously, we find models are more prone to hallucinate when similar objects are present in the scene, suggesting models may be relying on object co-occurrence seen during training. Among the models we evaluated, we found scale can provide modest improvements while models explicitly trained with multi-image inputs show bigger improvements, suggesting scaled multi-image training may offer promise. We make our benchmark publicly available to spur research into the challenge of hallucination when reasoning across scenes.



Paperid:2736
Authors:Christoph Schuhmann, Robert Kaczmarczyk, Gollam Rabby, Maurice Kraus, Felix Friedrich, Huu Nguyen, Kalyan Sai Krishna, Kourosh Nadi, Kristian Kersting, Sören Auer
Abstract:
Effective human-AI interaction relies on AI's ability to accurately perceive and interpret human emotions. Current benchmarks for vision and vision-language models are severely limited, offering a narrow emotional spectrum that overlooks nuanced states (e.g., bitterness, intoxication) and fails to distinguish subtle differences between related feelings (e.g., shame vs. embarrassment). Existing datasets also often use uncontrolled imagery with occluded faces and lack demographic diversity, risking significant bias. To address these critical gaps, we introduce EmoNet Face, a comprehensive benchmark suite. EmoNet Face features: (1) A novel 40-category emotion taxonomy, meticulously derived from foundational research to capture finer details of human emotional experiences. (2) Three large-scale, AI-generated datasets (EmoNet HQ, Binary, and Big) with explicit, full-face expressions and controlled demographic balance across ethnicity, age, and gender. (3) Rigorous, multi-expert annotations for training and high-fidelity evaluation. (4) We build Empathic Insight Face, a model achieving human-expert-level performance on our benchmark. The publicly released EmoNet Face suite—taxonomy, datasets, and model—provides a robust foundation for developing and evaluating AI systems with a deeper understanding of human emotions.



Paperid:2737
Authors:Konstantin Makarychev, Ilias Papanikolaou, Liren Shan
Abstract:
Abstract:We study the problem of explainable $k$-medians clustering introduced by Dasgupta, Frost, Moshkovitz, and Rashtchian (2020). In this problem, the goal is to construct a threshold decision tree that partitions data into $k$ clusters while minimizing the $k$-medians objective. These trees are interpretable because each internal node makes a simple decision by thresholding a single feature, allowing users to trace and understand how each point is assigned to a cluster.We present the first algorithm for explainable $k$-medians under $\ell_p$ norm for every finite $p \geq 1$. Our algorithm achieves an $\tilde{O}\big(p(\log k)^{1 + 1/p - 1/p^2}\big)$ approximation to the optimal $k$-medians cost for any $p \geq 1$. Previously, algorithms were known only for $p = 1$ and $p = 2$. For $p = 2$, our algorithm improves upon the existing bound of $\tilde O(\log^{3/2}k)$, and for $p = 1$, it matches the tight bound of $\log k + O(1)$ up to a multiplicative $O(\log \log k)$ factor.We show how to implement our algorithm in a dynamic setting. The dynamic algorithm maintains an explainable clustering under a sequence of insertions and deletions, with amortized update time $O(d \log^3 k)$ and $O(\log k)$ recourse, making it suitable for large-scale and evolving datasets.



Paperid:2738
Authors:Xiangchen Song, Jiaqi Sun, Zijian Li, Yujia Zheng, Kun Zhang
Abstract:
Despite Large Language Models' remarkable capabilities, understanding their internal representations remains challenging. Mechanistic interpretability tools such as sparse autoencoders (SAEs) were developed to extract interpretable features from LLMs but lack temporal dependency modeling, instantaneous relation representation, and more importantly theoretical guarantees—undermining both the theoretical foundations and the practical confidence necessary for subsequent analyses. While causal representation learning (CRL) offers theoretically-grounded approaches for uncovering latent concepts, existing methods cannot scale to LLMs' rich conceptual space due to inefficient computation. To bridge the gap, we introduce an identifiable temporal causal representation learning framework specifically designed for LLMs' high-dimensional concept space, capturing both time-delayed and instantaneous causal relations. Our approach provides theoretical guarantees and demonstrates efficacy on synthetic datasets scaled to match real-world complexity. By extending SAE techniques with our temporal causal framework, we successfully discover meaningful concept relationships in LLM activations. Our findings show that modeling both temporal and instantaneous conceptual relationships advances the interpretability of LLMs.



Paperid:2739
Authors:Aleksandar Terzic, Nicolas Menet, Michael Hersche, Thomas Hofmann, Abbas Rahimi
Abstract:
Abstract:Most modern state-space models (SSMs) rely on diagonal transition matrices, which ensure high computational efficiency but pose significant restrictions on their state-tracking abilities. In contrast, unstructured transition matrices allow SSMs to track the states of arbitrary finite-state automata (FSAs); however, at a prohibitively high compute and memory cost even for moderate state sizes. To reconcile the advantages of the two approaches, we propose an efficient parametrization of the transition matrices in selective SSMs that enables state-tracking at the level of unstructured SSMs while keeping the computational cost comparable to that of diagonal SSMs. Our PD-SSM parametrizes the transition matrix as the product of a column one-hot matrix ($P$) and a complex-valued diagonal matrix ($D$). Consequently, the computational cost of parallel scans over PD matrices scales linearly with the state size. Theoretically, PD-SSM can learn to track any FSA and reaches the lower bound on state dimension necessary to do so. Empirically, our model learns to track the states of FSAs of various complexities, even if the state transitions are obscured by variable-length sentences in natural language, by integrating it into a hybrid Transformer-SSM architecture.



Paperid:2740
Authors:yunzhu zhang, Yu Lu, Tianyi Wang, Fengyun Rao, Yi Yang, Linchao Zhu
Abstract:
Abstract:Long-form video understanding poses a significant challenge for video large language models (VideoLLMs) due to prohibitively high computational and memory demands. In this paper, We propose $\textbf{FlexSelect}$, a flexible and efficient token selection strategy for processing long videos.FlexSelect identifies and retains the most semantically relevant content by leveraging cross-modal attention patterns from a reference transformer layer.It comprises two key components: (1) $\textbf{a training-free token ranking pipeline}$ that leverages faithful cross-modal attention weights to estimate each video token’s importance, and (2) $\textbf{a rank-supervised lightweight selector}$ that is trained to replicate these rankings and filter redundant tokens.This generic approach can be seamlessly integrated into various VideoLLM architectures, such as LLaVA-Video, InternVL and Qwen-VL, serving as a plug-and-play module to extend their temporal context length. Empirically, FlexSelect delivers strong gains across multiple long-video benchmarks – including VideoMME, MLVU, LongVB, and LVBench. Morever, it achieves significant speed-ups ($\textit{e.g.,}$ up to 9 $\times$ on a LLaVA-Video-7B model), highlighting FlexSelect’s promise for efficient long-form video understanding. Project page: https://flexselect.github.io



Paperid:2741
Authors:Hector Pasten, Felipe Urrutia, Hector Orellana, Cristian Calderon, Cristobal Rojas, Alexander Kozachinskiy
Abstract:
Understanding how Transformers work and how they process information is key to the theoretical and empirical advancement of these machines. In this work, we demonstrate the existence of two phenomena in Transformers, namelyisolationandcontinuity. Both of these phenomena hinder Transformers to learn even simple pattern sequences. Isolation expresses that any learnable sequence must be isolated from another learnable sequence, and hence some sequences cannot be learned by a single Transformer at the same time. Continuity entails that an attractor basin forms around a learned sequence, such that any sequence falling in that basin will collapse towards the learned sequence. Here, we mathematically prove these phenomena emerge in all Transformers that use compact positional encoding, and design rigorous experiments, demonstrating that the theoretical limitations we shed light on occur on the practical scale.



Paperid:2742
Authors:Guang Liang, Xinyao Liu, Jianxin Wu
Abstract:
Vision Transformers (ViTs) are essential in computer vision but are computationally intensive, too. Model quantization, particularly to low bit-widths like 4-bit, aims to alleviate this difficulty, yet existing Post-Training Quantization (PTQ) and Quantization-Aware Training (QAT) methods exhibit significant limitations. PTQ often incurs substantial accuracy drop, while QAT achieves high accuracy but suffers from prohibitive computational costs, limited generalization to downstream tasks, training instability, and lacking of open-source codebase. To address these challenges, this paper introduces General, Practical, and Lightning Quantization (GPLQ), a novel framework designed for efficient and effective ViT quantization. GPLQ is founded on two key empirical insights: the paramount importance of activation quantization and the necessity of preserving the model's original optimization basin to maintain generalization. Consequently, GPLQ employs a sequential activation-first, weights-later strategy. Stage 1 keeps weights in FP32 while quantizing activations with a feature mimicking loss in only 1 epoch to keep it stay in the same basin, thereby preserving generalization. Stage 2 quantizes weights using a PTQ method. As a result, GPLQ is 100x faster than existing QAT methods, lowers memory footprint to levels even below FP32 training, and achieves 4-bit model performance that is highly competitive with FP32 models in terms of both accuracy on ImageNet and generalization to diverse downstream tasks, including fine-grained visual classification and object detection. We will release an easy-to-use open-source toolkit supporting multiple vision tasks.



Paperid:2743
Authors:Zhenyu Yang, Kairui Zhang, Yuhang Hu, Bing Wang, Shengsheng Qian, Bin Wen, Fan Yang, Tingting Gao, Weiming Dong, Changsheng Xu
Abstract:
Despite significant progress in Video Large Language Models (Video-LLMs) for offline video understanding, existing online Video-LLMs typically struggle to simultaneously process continuous frame-by-frame inputs and determine optimal response timing, often compromising real-time responsiveness and narrative coherence. To address these limitations, we introduce LiveStar, a pioneering live streaming assistant that achieves always-on proactive responses through adaptive streaming decoding. Specifically, LiveStar incorporates: (1) a training strategy enabling incremental video-language alignment for variable-length video streams, preserving temporal consistency across dynamically evolving frame sequences; (2) a response-silence decoding framework that determines optimal proactive response timing via a single forward pass verification; (3) memory-aware acceleration via peak-end memory compression for online inference on 10+ minute videos, combined with streaming key-value cache to achieve 1.53× faster inference. We also construct an OmniStar dataset, a comprehensive dataset for training and benchmarking that encompasses 15 diverse real-world scenarios and 5 evaluation tasks for online video understanding. Extensive experiments across three benchmarks demonstrate LiveStar's state-of-the-art performance, achieving an average 19.5\% improvement in semantic correctness with 18.1\% reduced timing difference compared to existing online Video-LLMs, while improving FPS by 12.0\% across all five OmniStar tasks. Our model and dataset can be accessed at https://anonymous.4open.science/r/LiveStar-5272.



Paperid:2744
Authors:Emmanuel Anaya Gonzalez, Kanghee Park, Sairam Vaidya, Ruyi Ji, Taylor Berg-Kirkpatrick, Loris D'Antoni
Abstract:
Constrained decoding enables Language Models (LMs) to produce samples that provably satisfy hard constraints.However, existing constrained-decoding approaches often distort the underlying model distribution, a limitation that is especially problematic in applications like program fuzzing, where one wants to generate diverse and valid program inputs for testing purposes. We propose a new constrained sampling framework based on Markov Chain Monte Carlo (MCMC) that simultaneously satisfies three core desiderata: constraint satisfying (every sample satisfies the constraint), monotonically converging (the sampling process converges to the true conditional distribution), and efficient (high-quality samples emerge in few steps). Our method constructs a proposal distribution over valid outputs and applies a Metropolis-Hastings acceptance criterion based on the LM’s likelihood, ensuring principled and efficient exploration of the constrained space. Empirically, our sampler outperforms existing methods on both synthetic benchmarks and real-world program fuzzing tasks.



Paperid:2745
Authors:Claudio Fanconi, Mihaela van der Schaar
Abstract:
Effective human-AI decision-making balances three key factors: thecorrectnessof predictions, thecostof knowledge and reasoning complexity, and the confidence about whether toabstainautomated answers or involve human experts. In this work, we present a cascaded LLM decision framework that adaptively delegates tasks across multiple tiers of expertise -- a base model for initial candidate answers, a more capable and knowledgeable (but costlier) large model, and a human expert for when the model cascade abstains. Our method proceeds in two stages. First, a deferral policy determines whether to accept the base model’s answer or regenerate it with the large model based on the confidence score. Second, an abstention policy decides whether the cascade model response is sufficiently certain or requires human intervention. Moreover, we incorporate an online learning mechanism in the framework that can leverage human feedback to improve decision quality over time. We demonstrate this approach to general question-answering (ARC-Easy and ARC-Challenge) and medical question-answering (MedQA and MedMCQA). Our results show that our cascaded strategy outperforms, in most cases, single-model baselines in accuracy while reducing cost and providing a principled way to handle abstentions.



Paperid:2746
Authors:Jiahan Zhang, Yaoyu Zhang, Tao Luo
Abstract:
Abstract:Understanding the convergence points and optimization landscape of neural networks is crucial, particularly for homogeneous networks where Karush-Kuhn-Tucker (KKT) points of the associated maximum-margin problem often characterize solutions. This paper investigates the relationship between such KKT points across networks of different widths generated via neuron splitting. We introduce and formalize the \textbf{KKT Point Embedding Principle}, establishing that KKT points of a homogeneous network's max-margin problem ($P_{\Phi}$) can be embedded into the KKT points of a larger network's problem ($P_{\tilde{\Phi}}$) via specific linear isometric transformations corresponding to neuron splitting. We rigorously prove this principle holds for neuron splitting in both two-layer and deep homogeneous networks . Furthermore, we connect this static embedding to the dynamics of gradient flow training with exponential-tailed losses. We demonstrate that trajectories initiated from appropriately mapped points remain mapped throughout training and that the resulting $\omega$-limit sets of directions are correspondingly mapped ($T(L(\boldsymbol{\theta}(0))) = L(\boldsymbol{\eta}(0))$, thereby preserving the alignment with KKT directions dynamically when directional convergence occurs . Our findings offer insights into the effects of network width, parameter redundancy, and the structural connections between solutions found via optimization in homogeneous networks of varying sizes.



Paperid:2747
Authors:Jiacheng You, Xinyang Chen, Yu Sun, Weili Guan, Liqiang Nie
Abstract:
Missing values, frequently encountered in time series data, can significantly impair the effectiveness of analytical methods. While deep imputation models have emerged as the predominant approach due to their superior performance, explicitly incorporating inductive biases aligned with time-series characteristics offers substantial improvement potential. Taking advantage of non-stationarity and periodicity in time series, two domain-specific inductive biases are designed: (1) Non-Stationary Guidance, which operationalizes the proximity principle to address highly non-stationary series by emphasizing temporal neighbors, and (2) Periodic Guidance, which exploits periodicity patterns through learnable weight allocation across historical periods. Building upon these complementary mechanisms, the overall module, named Meta Guidance, dynamically fuses both guidance matrices through data-adaptive weights learned from the specific input sample. Experiments on nine benchmark datasets demonstrate that integrating Meta Guidance into existing deep imputation architectures achieves an average 27.39\% reduction in imputation error compared to state-of-the-art baselines.



Paperid:2748
Authors:Yichen Li, Chicheng Zhang
Abstract:
Imitation learning (IL) is a paradigm for training sequential decision-making policies from experts, leveraging offline demonstrations, interactive annotations, or both. Recent advances show that when annotation cost is tallied per trajectory, Behavior Cloning (BC)—which relies solely on offline demonstrations—cannot be improved in general, leaving limited conditions for interactive methods such as DAgger to help. We revisit this conclusion and prove that when the annotation cost is measured per state, a metric commonly used for interactive annotation, algorithms using interactive annotations provably outperform BC.Specifically: (1) we show that a one‑sample‑per‑round DAgger variant provably beats BC and achieves up to an episode‑length improvement in state‑wise annotation cost; (2) considering the problem of hybrid IL where the agent learns from offline demonstrations and interactive annotations, we propose Warm‑start DAgger that performs no worse than using either data source alone, and can sometimes be significantly better, through an MDP example motivated by compounding error and cold‑start problem in imitation learning practice; (3) experiments on MuJoCo continuous‑control tasks confirm that, with reasonable costs assigned to offline and interactive annotations, interactive and hybrid approaches consistently outperform BC. To the best of our knowledge, our work is the first to formally define the hybrid IL problem and to highlight the benefits of state‑wise interactive annotation.



Paperid:2749
Authors:Daolang Huang, Xinyi Wen, Ayush Bharti, Samuel Kaski, Luigi Acerbi
Abstract:
Many critical applications, from autonomous scientific discovery to personalized medicine, demand systems that can both strategically acquire the most informative data and instantaneously perform inference based upon it. While amortized methods for Bayesian inference and experimental design offer part of the solution, neither approach is optimal in the most general and challenging task, where new data needs to be collected for instant inference. To tackle this issue, we introduce the Amortized Active Learning and Inference Engine (ALINE), a unified framework for amortized Bayesian inference and active data acquisition. ALINE leverages a transformer architecture trained via reinforcement learning with a reward based on self-estimated information gain provided by its own integrated inference component. This allows it to strategically query informative data points while simultaneously refining its predictions. Moreover, ALINE can selectively direct its querying strategy towards specific subsets of model parameters or designated predictive tasks, optimizing for posterior estimation, data prediction, or a mixture thereof. Empirical results on regression-based active learning, classical Bayesian experimental design benchmarks, and a psychometric model with selectively targeted parameters demonstrate that ALINE delivers both instant and accurate inference along with efficient selection of informative points.



Paperid:2750
Authors:Swetha Ganesh, Vaneet Aggarwal
Abstract:
Abstract:Actor-Critic methods are widely used for their scalability, yet existing theoretical guarantees for infinite-horizon average-reward Markov Decision Processes (MDPs) often rely on restrictive ergodicity assumptions. We propose NAC-B, a Natural Actor-Critic with Batching, that achieves order-optimal regret of \$\tilde{O}(\sqrt{T})\$ in infinite-horizon average-reward MDPs under the unichain assumption, which permits both transient states and periodicity. This assumption is among the weakest under which the classic policy gradient theorem remains valid for average-reward settings. NAC-B employs function approximation for both the actor and the critic, enabling scalability to problems with large state and action spaces. The use of batching in our algorithm helps mitigate potential periodicity in the MDP and reduces stochasticity in gradient estimates, and our analysis formalizes these benefits through the introduction of the constants $C_{\text{hit}}$ and $C_{\text{tar}}$, which characterize the rate at which empirical averages over Markovian samples converge to the stationary distribution.



Paperid:2751
Authors:Lisa Schmors, Dominic Gonschorek, Jan Niklas Böhm, Yongrong Qiu, Na Zhou, Dmitry Kobak, Andreas Tolias, Fabian Sinz, Jacob Reimer, Katrin Franke, Sebastian Damrich, Philipp Berens
Abstract:
Modern neural recording techniques such as two-photon imaging allow to acquire vast time-series datasets with responses of hundreds or thousands of neurons. Contrastive learning is a powerful self-supervised framework for learning representations of complex datasets. Existing applications for neural time series rely on generic data augmentations and do not exploit the multi-trial data structure inherent in many neural datasets. Here we present TRACE, a new contrastive learning framework that averages across different subsets of trials to generate positive pairs. TRACE allows to directly learn a two-dimensional embedding, combining ideas from contrastive learning and neighbor embeddings. We show that TRACE outperforms other methods, resolving fine response differences in simulated data. Further, using in vivo recordings, we show that the representations learned by TRACE capture both biologically relevant continuous variation, cell-type-related cluster structure, and can assist data quality control.



Paperid:2752
Authors:Felipe Maia Polo, Xinhe Wang, Mikhail Yurochkin, Gongjun Xu, Moulinath Banerjee, Yuekai Sun
Abstract:
LLM-as-a-judge (LLMJ) has become popular for scalably evaluating language model outputs on open-ended user queries. However, LLM judges do not always align with human annotators, exhibiting systematic and undesired differences, e.g., biases toward certain writing styles. In this paper, we introduce the first unified statistical framework that jointly models human and LLM judge ratings under both absolute and relative (e.g., pairwise comparisons) evaluation paradigms. We posit the existence of a latent human preference score that is shared between humans and LLM judges for each prompt–response pair, and allow LLM judgments to deviate systematically via a linear transformation of a covariate vector, which encodes potential sources of LLM judge biases. Our core contribution is a statistical model that links LLM-assigned scores to the latent human signal and a set of biasing covariates, for which we propose a fitting algorithm and establish asymptotic normality for resulting estimators, enabling formal hypothesis tests of human-LLM discrepancies. Empirically, our model (i) improves the probability calibration of LLM judgments, (ii) aligns LLM ratings more closely with human judgments, and (iii) quantifies and tests systematic LLM biases across evaluation scenarios. We verify the efficacy and provide insightful results using four LLM judges and queries from BigGen Bench and Chatbot Arena.



Paperid:2753
Authors:Zizhao Wang, Kaixin Wang, Li Zhao, Peter Stone, Jiang Bian
Abstract:
World models aim to capture the dynamics of the environment, enabling agents to predict and plan for future states. In most scenarios of interest, the dynamics are highly centered on interactions among objects within the environment. This motivates the development of world models that operate on object-centric rather than monolithic representations, with the goal of more effectively capturing environment dynamics and enhancing compositional generalization. However, the development of object-centric world models has largely been explored in environments with limited visual complexity (such as basic geometries). It remains underexplored whether such models can be effective in more challenging settings. In this paper, we fill this gap by introducing Dyn-O, an enhanced structured world model built upon object-centric representations. Compared to prior work in object-centric representations, Dyn-O improves in both learning representations and modeling dynamics. On the challenging Procgen games, we demonstrate that our method can learn object-centric world models directly from pixel observations, outperforming DreamerV3 in rollout prediction accuracy. Furthermore, by decoupling object centric features into dynamic-agnostic and dynamic-aware components, we enable finer-grained manipulation of these features and generate more diverse imagined trajectories.



Paperid:2754
Authors:Jiawei Gu, Yidi Wang, Qingqiang Sun, Xinming Li, Xiao Luo, Ziyue Qiao
Abstract:
Pseudo-label-based Semi-Supervised Learning (SSL) often suffers from classifier bias, particularly under class imbalance, as inaccurate pseudo-labels tend to exacerbate existing biases towards majority classes. Existing methods, such as \textit{CDMAD}\cite{cdmad}, utilize simplistic reference inputs—typically uniform or blank-colored images—to estimate and correct this bias. However, such simplistic references fundamentally ignore realistic statistical information inherent to real datasets, specifically typical color distributions, texture details, and frequency characteristics. This lack of \emph{statistical representativeness} can lead the model to inaccurately estimate its inherent bias, limiting the effectiveness of bias correction, particularly under severe class imbalance or substantial distribution mismatches between labeled and unlabeled datasets. To overcome these limitations, we introduce the \textbf{FARAD} (Fourier-Adapted Reference for Accurate Debiasing) System. This system utilizes random-phase images, constructed by preserving the amplitude spectrum of real data while randomizing the phase spectrum. This strategy ensures two critical properties: (1) \textbf{Semantic Irrelevance}, as randomizing phase removes any structural or recognizable semantic cues, and (2) \textbf{Statistical Representativeness}, as preserving the amplitude spectrum maintains realistic textures, color distributions, and frequency characteristics. Grounded theoretically in classical Fourier analysis, the FARAD System provides a robust, accurate estimation of per-class biases. Furthermore, computational efficiency is enhanced through optimized real-to-complex (R2C) batched Fast Fourier Transforms (FFTs). Comprehensive experiments demonstrate that our approach, significantly improves minority-class accuracy and overall SSL performance, particularly under challenging imbalance scenarios, compared with existing reference-based bias correction methods.



Paperid:2755
Authors:YuQing Xie, Tess Smidt
Abstract:
Equivariant neural networks have proven to be effective for tasks with known underlying symmetries. However, optimizing equivariant networks can be tricky and best training practices are less established than for standard networks. In particular, recent works have found small training benefits from relaxing equivariance constraints. This raises the question: do equivariance constraints introduce fundamental obstacles to optimization? Or do they simply require different hyperparameter tuning? In this work, we investigate this question through a theoretical analysis of the loss landscape geometry. We focus on networks built using permutation representations, which we can view as a subset of unconstrained MLPs. Importantly, we show that the parameter symmetries of the unconstrained model has nontrivial effects on the loss landscape of the equivariant subspace and under certain conditions can provably prevent learning of the global minima. Further, we empirically demonstrate in such cases, relaxing to an unconstrained MLP can sometimes solve the issue. Interestingly, the weights eventually found via relaxation corresponds to a different choice of group representation in the hidden layer. From this, we draw 3 key takeaways. (1) Viewing any class of networks in the context of larger unconstrained function space can give important insights on loss landscape structure. (2) Within the unconstrained function space, equivariant networks form a complicated union of linear hyperplanes, each associated with a specific choice of internal group representation. (3) Effective relaxation of equivariance may require not only adding nonequivariant degrees of freedom, but also rethinking the fixed choice of group representations in hidden layers.



Paperid:2756
Authors:Xiangcheng Zhang, Yige Hong, Weina Wang
Abstract:
Abstract:Heterogeneity poses a fundamental challenge for many real-world large-scale decision-making problems but remains largely understudied.In this paper, we study the _fully heterogeneous_ setting of a prominent class of such problems, known as weakly-coupled Markov decision processes (WCMDPs).Each WCMDP consists of $N$ arms (or subproblems), which have distinct model parameters in the fully heterogeneous setting, leading to the curse of dimensionality when $N$ is large.We show that, under mild assumptions, an efficiently computable policy achieves an $O(1/\sqrt{N})$ optimality gap in the long-run average reward per arm for fully heterogeneous WCMDPs as $N$ becomes large.This is the _first asymptotic optimality result_ for fully heterogeneous average-reward WCMDPs.Our main technical innovation is the construction of projection-based Lyapunov functions that certify the convergence of rewards and costs to an optimal region, even under full heterogeneity.



Paperid:2757
Authors:Ziyang Cai, Nayoung Lee, Avi Schwarzschild, Samet Oymak, Dimitris Papailiopoulos
Abstract:
Transformer language models have demonstrated impressive generalization capabilities in natural language domains, yet we lack a fine-grained understanding of how such generalization arises. In this paper, we investigate length generalization—the ability to extrapolate from shorter to longer inputs—through the lens of \textit{task transfer}. We find that length generalization can be \textit{transferred} across related tasks. That is, training a model with a longer and related auxiliary task can lead the model to generalize to unseen and longer inputs from some other target task. We demonstrate this length generalization transfer across a diverse suite of algorithmic tasks, including arithmetic operations, string transformations, and maze navigation. Our results show that transformer models can inherit generalization capabilities from similar tasks when trained jointly. Moreover, we observe similar transfer effects in pretrained language models, suggesting that pretraining equips models with reusable computational scaffolding that facilitates extrapolation in downstream settings. Finally, we provide initial mechanistic evidence that length generalization transfer correlates with the re-use of the same attention heads between the tasks. Together, our findings deepen our understanding of how transformers generalize to out-of-distribution inputs and highlight the compositional reuse of inductive structure across tasks.



Paperid:2758
Authors:Xin Tong, Baojie Tian, Yufei Guo, Zhe Ma
Abstract:
It is observed that the confidence score may fail to reflect the predicting quality accurately in previous proposal-based line segment detection methods, since the scores and the line locations are predicted simultaneously. We find the line segment detection performance can be further improved by learning-based line candidate ranking and optimizing strategy. To this end, we build a novel end-to-end line detecting model named RANK++LETR upon deformable DETR architecture, where the encoder is used to select the line candidates while the decoder is applied to rank and optimize these candidates. We design line-aware deformable attention (LADA) module in which attention positions are distributed in a long narrow area and can align well with the elongated geometry of line segments. Moreover, we innovatively apply ranking-based supervision in line segment detection task with the design of contiguous labels according to the detection quality. Experimental results demonstrate that our method outperforms previous SOTA methods in prediction accuracy and gets faster inferring speed than other Transformer-based methods.



Paperid:2759
Authors:Pratik Rathore, Zachary Frangella, Sachin Garg, Shaghayegh Fazliani, Michal Derezinski, Madeleine Udell
Abstract:
Abstract:Gaussian processes (GPs) play an essential role in biostatistics, scientific machine learning, and Bayesian optimization for their ability to provide probabilistic predictions and model uncertainty.However, GP inference struggles to scale to large datasets (which are common in modern applications), since it requires the solution of a linear system whose size scales quadratically with the number of samples in the dataset.We propose an approximate, distributed, accelerated sketch-and-project algorithm ($\texttt{ADASAP}$) for solving these linear systems, which improves scalability.We use the theory of determinantal point processes to show that the posterior mean induced by sketch-and-project rapidly converges to the true posterior mean. In particular, this yields the first efficient, condition number-free algorithm for estimating the posterior mean along the top spectral basis functions, showing that our approach is principled for GP inference.$\texttt{ADASAP}$ outperforms state-of-the-art solvers based on conjugate gradient and coordinate descent across several benchmark datasets and a large-scale Bayesian optimization task.Moreover, $\texttt{ADASAP}$ scales to a dataset with $> 3 \cdot 10^8$ samples, a feat which has not been accomplished in the literature.



Paperid:2760
Authors:Sai Shankar Narasimhan, Shubhankar Agarwal, Litu Rout, Sanjay Shakkottai, Sandeep Chinchali
Abstract:
Abstract:Generating realistic time series samples is crucial for stress-testing models and protecting user privacy by using synthetic data. In engineering and safety-critical applications, these samples must meet certain hard constraints that are domain-specific or naturally imposed by physics or nature. Consider, for example, generating electricity demand patterns with constraints on peak demand times. This can be used to stress-test the functioning of power grids during adverse weather conditions. Existing approaches for generating constrained time series are either not scalable or degrade sample quality. To address these challenges, we introduce Constrained Posterior Sampling (CPS), a diffusion-based sampling algorithm that aims to project the posterior mean estimate into the constraint set after each denoising update. Notably, CPS scales to a large number of constraints ($\sim100$) without requiring additional training. We provide theoretical justifications highlighting the impact of our projection step on sampling. Empirically, CPS outperforms state-of-the-art methods in sample quality and similarity to real time series by around 70\% and 22\%, respectively, on real-world stocks, traffic, and air quality datasets.



Paperid:2761
Authors:Anushri Arora, Jonathan Pillow
Abstract:
Low-rank recurrent neural networks (RNNs) provide a powerful framework for characterizing how neural systems solve complex cognitive tasks. However, fitting and interpreting these networks remains an important open problem. In this paper, we develop new methods for efficiently fitting low-rank RNNs in `teacher-training'' settings. In particular, we build upon the neural engineering framework (NEF), in which RNNs are viewed as approximating an ordinary differential equation (ODE) of interest using a set of random nonlinear basis functions. This view provides geometric insight into how the choice of neural nonlinearity (e.g. tanh, ReLU) and the distribution of model parameters affects an RNN's representational capacity. Next, we adapt this framework for online training and demonstrate better performance with significantly smaller networks compared to FORCE. Additionally, we outperform backpropagation-trained networks of similar size—while requiring substantially less training time. Next, we ask: what basis functions—and how many—are needed to approximate a given dynamical system? To address this, we introduce methods for finding the smallest low-rank RNN to implement a dynamical system of interest using extensions of orthogonal matching pursuit. We then consider infinite unit low-rank RNNs, which converge to a Gaussian Process. In particular, we show that we can optimize the distribution of nonlinear basis functions using maximum marginal likelihood for hyperparameters of the equivalent kernel function---which can be computed in closed form for particular choices of nonlinearity, resulting in substantially better generalization performance from limited data. Finally, we describe active learning methods for selecting highly informative datapoints that speed up training of low-rank RNNs.



Paperid:2762
Authors:Dongyoung Kim, Huiwon Jang, Sumin Park, Jaehyung Kim, Younggyo Seo, Jinwoo Shin
Abstract:
Large Vision-Language Models (LVLMs) have recently shown great promise in advancing robotics by combining embodied reasoning with robot control. A common approach involves training on embodied reasoning tasks related to robot control using Supervised Fine-Tuning (SFT). However, SFT datasets are often heuristically constructed and not explicitly optimized for improving robot control. Furthermore, SFT often leads to issues such as catastrophic forgetting and reduced generalization performance. To address these limitations, we introduce Robot-R1, a novel framework that leverages reinforcement learning to enhance embodied reasoning specifically for robot control. Robot-R1 learns to predict the next keypoint state required for task completion, conditioned on the current scene image and environment metadata derived from expert demonstrations. Inspired by the DeepSeek-R1 learning approach, Robot-R1 samples reasoning-based responses and reinforces those that lead to more accurate predictions. Our experiments show that models trained with Robot-R1 outperform SFT methods on embodied reasoning tasks. Despite having only 7B parameters, Robot-R1 even surpasses GPT-4o on reasoning tasks related to low-level action control, such as spatial and primitive movement reasoning.



Paperid:2763
Authors:Yifei Su, Ning Liu, Dong Chen, Zhen Zhao, Kun Wu, Meng Li, Zhiyuan Xu, Zhengping Che, Jian Tang
Abstract:
Abstract:Generative modeling-based visuomotor policies have been widely adopted in robotic manipulation attributed to their ability to model multimodal action distributions. However, the high inference cost of multi-step sampling limits their applicability in real-time robotic systems. To address this issue, existing approaches accelerate the sampling process in generative modeling-based visuomotor policies by adapting acceleration techniques originally developed for image generation, such as Consistency Models and Consistency-FM. Despite this progress, a major distinction remains: image generation typically involves producing independent samples without temporal dependencies, whereas robotic manipulation involves generating time-series action trajectories that require continuity and temporal coherence. To effectively exploit temporal information in robotic manipulation, we propose FreqPolicy, a novel approach that first imposes frequency consistency constraints on flow-based visuomotor policies. Our work enables the action model to capture temporal structure effectively while supporting efficient, high-quality one-step action generation. Inspired by advances in time-series forecasting and speech processing, we introduce a frequency consistency constraint objective that enforces alignment of frequency-domain action features across different timesteps along the flow, thereby promoting convergence of one-step action generation toward the target distribution. In addition, we design an adaptive consistency loss to capture structural temporal variations inherent in robotic manipulation tasks. We assess FreqPolicy on $53$ tasks across $3$ simulation benchmarks, proving its superiority over existing one-step action generators.We further integrate FreqPolicy into the vision-language-action (VLA) model and achieve acceleration without performance degradation on the $40$ tasks of Libero. Besides, we show efficiency and effectiveness in real-world robotic scenarios with an inference frequency $93.5~\mathrm {Hz}$. The code will be publicly available.



Paperid:2764
Authors:Chiara Cappellino, Gianluca Mancusi, Matteo Mosconi, Angelo Porrello, SIMONE CALDERARA, Rita Cucchiara
Abstract:
Open-Vocabulary object detectors can generalize to an unrestricted set of categories through simple textual prompting. However, adapting these models to rare classes or reinforcing their abilities on multiple specialized domains remains essential. While recent methods rely on monolithic adaptation strategies with a single set of weights, we embrace modular deep learning. We introduce DitHub, a framework designed to build and maintain a library of efficient adaptation modules. Inspired by Version Control Systems, DitHub manages expert modules as branches that can be fetched and merged as needed. This modular approach allows us to conduct an in-depth exploration of the compositional properties of adaptation modules, marking the first such study in Object Detection. Our method achieves state-of-the-art performance on the ODinW-13 benchmark and ODinW-O, a newly introduced benchmark designed to assess class reappearance.



Paperid:2765
Authors:Wenzhao Liu, Haoran Li, Congying Han, Zicheng Zhang, Anqi Li, Tiande Guo
Abstract:
Achieving generalization in neural approaches across different scales and distributions remains a significant challenge for routing problems. A key obstacle is that neural networks often fail to learn robust principles for identifying universal patterns and deriving optimal solutions from diverse instances. In this paper, we first uncover Purity Law, a fundamental structural principle for optimal solutions of routing problems, defining that edge prevalence grows exponentially with the sparsity of surrounding vertices. Statistically validated across diverse instances, Purity Law reveals a consistent bias toward local sparsity in global optima. Building on this insight, we propose Purity Policy Optimization (PUPO), a novel training paradigm that explicitly aligns characteristics of neural solutions with Purity Law during the solution construction process to enhance generalization. Extensive experiments demonstrate that PUPO can be seamlessly integrated with popular neural solvers, significantly enhancing their generalization performance without incurring additional computational overhead during inference.



Paperid:2766
Authors:Ariel Shaulov, Itay Hazan, Lior Wolf, Hila Chefer
Abstract:
Text-to-video diffusion models are notoriously limited in their ability to model temporal aspects such as motion, physics, and dynamic interactions. Existing approaches address this limitation by retraining the model or introducing external conditioning signals to enforce temporal consistency. In this work, we explore whether a meaningful temporal representation can be extracted directly from the predictions of a pre-trained model without any additional training or auxiliary inputs. We introduceFlowMo, a novel training-free guidance method that enhances motion coherence using only the model's own predictions in each diffusion step. FlowMo first derives an appearance-debiased temporal representation by measuring the distance between latents corresponding to consecutive frames. This highlights the implicit temporal structure predicted by the model. It then estimates motion coherence by measuring the patch-wise variance across the temporal dimension, and guides the model to reduce this variance dynamically during sampling. Extensive experiments across multiple text-to-video models demonstrate that FlowMo significantly improves motion coherence without sacrificing visual quality or prompt alignment, offering an effective plug-and-play solution for enhancing the temporal fidelity of pre-trained video diffusion models.



Paperid:2767
Authors:Neta Shaul, Uriel Singer, Itai Gat, Yaron Lipman
Abstract:
Diffusion and flow matching models have significantly advanced media generation, yet their design space is well-explored, somewhat limiting further improvements. Concurrently, autoregressive (AR) models, particularly those generating continuous tokens, have emerged as a promising direction for unifying text and media generation, showing improved performance at scale. This paper introduces Transition Matching (TM), a novel discrete-time, continuous-state generative paradigm that unifies and advances both diffusion/flow models and continuous AR generation. TM decomposes complex generation tasks into simpler Markov transitions, allowing for expressive non-deterministic probability transition kernels and arbitrary non-continuous supervision processes, thereby unlocking new flexible design avenues. We explore these choices through three TM variants: (i) Difference Transition Matching (DTM), which generalizes flow matching to discrete-time by directly learning transition probabilities, yielding state-of-the-art image quality and text adherence. (ii) Autoregressive Transition Matching (ARTM) and (iii) Full History Transition Matching (FHTM) are partially and fully causal models, respectively, that generalize continuous AR methods. They achieve continuous causal AR generation quality comparable to non-causal approaches and potentially enable seamless integration with existing AR text generation techniques. Notably, FHTM is the first fully causal model to match or surpass the performance of flow-based methods on text-to-image task in continuous domains. We demonstrate these contributions through a rigorous large-scale comparison of TM variants and relevant baselines, maintaining a fixed architecture, training data, and hyperparameters.



Paperid:2768
Authors:Anamika Lochab, Lu Yan, Patrick Pynadath, Xiangyu Zhang, Ruqi Zhang
Abstract:
The rise of API-only access to state-of-the-art LLMs highlights the need for effective black-box jailbreak methods to identify model vulnerabilities in real-world settings. Without a principled objective for gradient-based optimization in the black-box regime, most existing approaches rely on genetic algorithms, which are limited by their initialization and dependence on manually curated prompt pools. Furthermore, these methods require individual optimization for each prompt, failing to provide a comprehensive characterization of model vulnerability.To address this gap, we introduce VERA: Variational infErence fRamework for jAilbreaking. VERA casts black-box jailbreak prompting as a variational inference problem, training a small attacker LLM to approximate the target LLM’s posterior over adversarial prompts. Once trained, the attacker can generate diverse, fluent jailbreak prompts for a target query without re-optimization. Experimental results show that VERA achieves strong performance across a range of target LLMs, highlighting the value of probabilistic inference for adversarial prompt generation.



Paperid:2769
Authors:Xiang Meng, Mehdi Makni, Rahul Mazumder
Abstract:
Network pruning reduces computational requirements of large neural networks, with N:M sparsity—retaining only N out of every M consecutive weights—offering a compelling balance between compressed model quality and hardware acceleration. However, N:M sparsity only accelerates forward-pass computations, as N:M patterns are not preserved during matrix transposition, limiting efficiency during training where both passes are computationally intensive. While transposable N:M sparsity has been proposed to address this limitation, existing methods for finding transposable N:M sparse masks either fail to scale to large models or are restricted to M=4 which results in suboptimal compression-accuracy trade-off. We introduce an efficient solver for transposable N:M masks that scales to billion-parameter models. We formulate mask generation as optimal transport problems and solve through entropy regularization and Dykstra's algorithm, followed by a rounding procedure. Our tensor-based implementation exploits GPU parallelism, achieving up to 100× speedup with only 1-10\% error compared to existing methods. Our approach can be integrated with layer-wise N:M pruning frameworks including Wanda, SparseGPT and ALPS to produce transposable N:M sparse models with arbitrary N:M values. Experiments show that LLaMA3.2-8B with transposable 16:32 sparsity maintains performance close to its standard N:M counterpart and outperforms standard 2:4 sparse model, showing the practical value of our approach.



Paperid:2770
Authors:Mingyang Chen, Linzhuang Sun, Tianpeng Li, sunhaoze, ZhouYijie, Chenzheng Zhu, Haofen Wang, Jeff Pan, Wen Zhang, Huajun Chen, Fan Yang, Zenan Zhou, weipeng chen
Abstract:
Large Language Models (LLMs) have shown remarkable capabilities in reasoning, exemplified by the success of OpenAI-o1 and DeepSeek-R1. However, integrating reasoning with external search processes remains challenging, especially for complex multi-hop questions requiring multiple retrieval steps. We propose ReSearch, a novel framework that trains LLMs to Reason with Search via reinforcement learning without using any supervised data on reasoning steps. Our approach treats search operations as integral components of the reasoning chain, where when and how to perform searches is guided by text-based thinking, and search results subsequently influence further reasoning. We train ReSearch on Qwen2.5-7B(-Instruct) and Qwen2.5-32B(-Instruct) models and conduct extensive experiments. Despite being trained on only one dataset, our models demonstrate strong generalizability across various benchmarks. Analysis reveals that ReSearch naturally elicits advanced reasoning capabilities such as reflection and self-correction during the reinforcement learning process.



Paperid:2771
Authors:Zhen Zhang, Bingsheng He
Abstract:
Unsupervised Graph Domain Adaptation has become a promising paradigm for transferring knowledge from a fully labeled source graph to an unlabeled target graph. Existing graph domain adaptation models primarily focus on the closed-set setting, where the source and target domains share the same label space. However, this assumption might not be practical in the real-world scenarios, as the target domain might include classes that are not present in the source domain. In this paper, we investigate the problem of unsupervised open-set graph domain adaptation, where the goal is to not only correctly classify target nodes into the known classes, but also classify previously unseen node types into the unknown class. Towards this end, we propose a novel framework called GraphRTA, which conducts reprogramming on both the graph and model sides. Specifically, we reprogram the graph by modifying target graph structure and node features, which facilitates better separation of known and unknown classes. Meanwhile, we also perform model reprogramming by pruning domain-specific parameters to reduce bias towards the source graph while preserving parameters that capture transferable patterns across graphs. Additionally, we extend the classifier with an extra dimension for the unknown class, thus eliminating the need of manually specified threshold in open-set recognition. Comprehensive experiments on several public datasets demonstrate that our proposed model can achieve satisfied performance compared with recent state-of-the-art baselines. Our source codes and datasets are publicly available at https://anonymous.4open.science/r/GraphRTA-A837.



Paperid:2772
Authors:Shintaro Wakasugi, Taiji Suzuki
Abstract:
Diffusion models operating in discrete state spaces have emerged as powerful approaches, demonstrating remarkable efficacy across diverse domains, including reasoning tasks and molecular design. Despite their promising applications, the theoretical foundations of these models remain substantially underdeveloped, with the existing literature predominantly focusing on continuous-state diffusion models.A critical gap persists in the theoretical understanding of discrete diffusion modeling: the absence of a rigorous framework for quantifying estimation error with finite data. Consequently, the fundamental question of how precisely one can reconstruct the true underlying distribution from a limited training set remains unresolved. In this work, we analyze the estimation error induced by a score estimation of the discrete diffusion models. One of the main difficulties in the analysis stems from the fact that the cardinality of the state space can be exponentially large with respect to its dimension, which results in an intractable error bound by a naive approach. To overcome this difficulty, we make use of a property that the state space can be smoothly embedded in a continuous Euclidean space that enables us to derive a cardinality independent bound, which is more practical in real applications. In particular, we consider a setting where the state space is structured as a hypercube graph, and another where the induced graph Laplacian can be asymptotically well approximated by the ordinary Laplacian defined on the continuous space, and then derive state space size independent bounds.



Paperid:2773
Authors:Yan Dai, Negin Golrezaei, Patrick Jaillet
Abstract:
Abstract:Motivated by applications such as cloud platforms allocating GPUs to users or governments deploying mobile health units across competing regions, we study the dynamic allocation of a reusable resource to strategic agents with private valuations. Our objective is to simultaneously *(i)* maximize social welfare, *(ii)* satisfy multi-dimensional long-term cost constraints, and *(iii)* incentivize truthful reporting. We begin by numerically evaluating primal-dual methods widely used in constrained online optimization and find them to be highly fragile in strategic settings -- agents can easily manipulate their reports to distort future dual updates for future gain. To address this vulnerability, we develop an incentive-aware framework that makes primal-dual methods robust to strategic behavior. Our design combines epoch-based lazy updates -- where dual variables remain fixed within each epoch -- with randomized exploration rounds that extract approximately truthful signals for learning. Leveraging carefully designed online learning subroutines for dual updates, our mechanism achieves $\tilde{\mathcal{O}}(\sqrt{T})$ social welfare regret, satisfies all cost constraints, and ensures incentive alignment. This matches the performance of non-strategic allocation approaches while exhibiting robustness to strategic agents.



Paperid:2774
Authors:Hoyoon Byun, Gyeongdeok Seo, Joonseong Kang, Taero Kim, Jihee Kim, Kyungwoo Song
Abstract:
In-context learning (ICL), a nonparametric learning method based on the knowledge of demonstration sets, has become a de facto standard for large language models (LLMs). The primary goal of ICL is to select valuable demonstration sets to enhance the performance of LLMs. Traditional ICL methods choose demonstration sets that share similar features with a given query. However, we have found that the performance of these traditional ICL approaches is limited on out-of-distribution (OOD) datasets, where the demonstration set and the query originate from different distributions. To ensure robust performance in OOD datasets, it is essential to learn causal representations that remain invariant between the source and target datasets. Inspired by causal representation learning, we propose causal-aware in-context learning (CCL). CCL captures the causal representations of a given dataset and selects demonstration sets that share similar causal features with the query. To achieve this, CCL employs a novel VAE-based causal representation learning technique. We demonstrate that CCL improves the OOD generalization performance of LLMs both theoretically and empirically. \footnote{Code is available at: \url{https://anonymous.4open.science/r/causal-context-learning-C717}}



Paperid:2775
Authors:Jiaqi Cao, Jiarui Wang, Rubin Wei, Qipeng Guo, Kai Chen, Bowen Zhou, Zhouhan Lin
Abstract:
Large Language Models (LLMs) have shown strong abilities in general language tasks, yet adapting them to specific domains remains a challenge.Current method like Domain Adaptive Pretraining (DAPT) requires costly full-parameter training and suffers from catastrophic forgetting.Meanwhile, Retrieval-Augmented Generation (RAG) introduces substantial inference latency due to expensive nearest-neighbor searches and longer context.This paper introduces \textit{Memory Decoder}, a plug-and-play pretrained memory that enables efficient domain adaptation without changing the original model's parameters.Memory Decoder employs a small transformer decoder that learns to imitate the behavior of an external non-parametric retriever.Once trained, Memory Decoder can be seamlessly integrated with any pretrained language model that shares the same tokenizer, requiring no model-specific modifications.Experimental results demonstrate that Memory Decoder enables effective adaptation of various Qwen and Llama models to three distinct specialized domains: biomedicine, finance, and law, reducing perplexity by an average of 6.17 points.Overall, Memory Decoder introduces a novel paradigm centered on a specially pretrained memory component designed for domain-specific adaptation. This memory architecture can be integrated in a plug-and-play manner, consistently enhancing performance across multiple models within the target domain.



Paperid:2776
Authors:Xinlong Li, Di Lin, Shaoyiyi Gao, Jiaxin Li, Ruonan Liu, Qing Guo
Abstract:
For inaccurate segmentation of structurally connected part boundaries in open-vocabulary part segmentation (OVPS), we reveal that the root cause lies in the conflict between continuous, smooth image feature and discrete classification mechanism. To address this, we propose the first training-free OVPS framework, PBAPS. Our method constructs HPCGraph to model the object-part hierarchical structure, performing progressive segmentation via cross-hierarchy matching. For feature-ambiguous regions at structurally connected part boundaries, the BAR module locates ambiguous regions via matching cost divergence, optimizes ambiguous-region features using deterministic-region context, and adaptively refines visual prototypes to achieve precise reclassification of ambiguous regions. Experiments on Pascal-Part-116, ADE20K-Part-234, PartImageNet demonstrate that PBAPS significantly outperforms state-of-the-art methods, achieving 46.35% mIoU and 34.46% bIoU on Pascal-Part-116.



Paperid:2777
Authors:Johanna Düngler, Amartya Sanyal
Abstract:
Abstract:Given $n$ i.i.d.random matrices $A_i \in \mathbb{R}^{d \times d}$ that share common expectation $\Sigma$, the objective of Differentially Private Stochastic PCA is to identify a subspace of dimension $k$ that captures the largest variance directions of $\Sigma$, while preserving differential privacy (DP) of each individual $A_i$. Existing methods either (i) require the sample size $n$ to scale super-linearly with dimension $d$, even under Gaussian assumptions on the $A_i$, or (ii) introduce excessive noise for DP even when the intrinsic randomness within $A_i$ is small.~\citet{liu2022dp} addressed these issues for sub-Gaussian data but only for estimating the top eigenvector ($k=1$) using their algorithm DP-PCA. We propose the first algorithm capable of estimating the top $k$ eigenvectors for arbitrary $k \leq d$, whilst overcoming both limitations above. For $k=1$, our algorithm matches the utility guarantees of DP-PCA, achieving near-optimal statistical error even when $n = \tilde{O}(d)$. We further provide a lower bound for general $k > 1$, matching our upper bound up to a factor of $k$, and experimentally demonstrate the advantages of our algorithm over comparable baselines.



Paperid:2778
Authors:Longquan Dai, Xiaolu Wei, wang he, Shaomeng Wang, Jinhui Tang
Abstract:
Text-to-image diffusion models are typically trained on large-scale web data, often resulting in outputs that misalign with human preferences. Inspired by preference learning in large language models, we propose ABC (Alignment by Classification), a simple yet effective framework for aligning diffusion models with human preferences. In contrast to prior DPO-based methods that depend on suboptimal supervised fine-tuned (SFT) reference models, ABC assumes access to an ideal reference model perfectly aligned with human intent and reformulates alignment as a classification problem. Under this view, we recognize that preference data naturally forms a semi-supervised classification setting. To address this, we propose a data augmentation strategy that transforms preference comparisons into fully supervised training signals. We then introduce a classification-based ABC loss to guide alignment. Our alignment by classification approach could effectively steer the diffusion model toward the behavior of the ideal reference. Experiments on various diffusion models show that our ABC consistently outperforms existing baselines, offering a scalable and robust solution for preference-based text-to-image fine-tuning.



Paperid:2779
Authors:Zhenqi He, Yuanpei Liu, Kai Han
Abstract:
This paper investigates the problem of Generalized Category Discovery (GCD). Given a partially labelled dataset, GCD aims to categorize all unlabelled images, regardless of whether they belong to known or unknown classes. Existing approaches typically depend on either single-level semantics or manually designed abstract hierarchies, which limit their generalizability and scalability. To address these limitations, we introduce a SEmantic-aware hierArchical Learning framework (SEAL), guided by naturally occurring and easily accessible hierarchical structures. Within SEAL, we propose a Hierarchical Semantic-Guided Soft Contrastive Learning approach that exploits hierarchical similarity to generate informative soft negatives, addressing the limitations of conventional contrastive losses that treat all negatives equally. Furthermore, a Cross-Granularity Consistency (CGC) module is designed to align the predictions from different levels of granularity. SEAL consistently achieves state-of-the-art performance on fine-grained benchmarks, including the SSB benchmark, Oxford-Pet, and the Herbarium19 dataset, and further demonstrates generalization on coarse-grained datasets.



Paperid:2780
Authors:Inho Jeong, Sunghyeon Woo, Sol Namkung, Dongsuk Jeon
Abstract:
Abstract:Large‑language‑model (LLM) inference with long contexts often produces key–value (KV) caches whose footprint exceeds the capacity of GPU high‑bandwidth memory (HBM). Prior LLM inference frameworks such as vLLM mitigate this pressure by swapping KV cache pages to host DRAM. However, the high total cost of ownership (TCO) of large DRAM pools makes this solution economically unattractive. Although offloading to SSDs can be an alternative for reducing TCO, conventional frameworks such as FlexGen experience a substantial throughput drop since the data path that routes SSD traffic through CPU to GPU is severely bandwidth-constrained. To overcome these limitations. we introduce HiFC, a novel DRAM‑free swapping scheme that stores KV pages in the pseudo‑SLC region of a commodity NVMe SSD and transfers the KV cache to and from GPU directly through GPU-Direct-Storage (GDS). HiFC attains inference throughput compared with DRAM‑based swapping under diverse long‑context workloads while significantly lowering the memory‑related TCO of a single‑GPU system by 4.5$\times$ over three years.



Paperid:2781
Authors:Xiangyu Zhao, Peiyuan Zhang, Kexian Tang, Xiaorong Zhu, Hao Li, Wenhao Chai, Zicheng Zhang, Renqiu Xia, Guangtao Zhai, Junchi Yan, Hua Yang, Xue Yang, Haodong Duan
Abstract:
Large Multi-modality Models (LMMs) have made significant progress in visual understanding and generation, but they still face challenges in General Visual Editing, particularly in following complex instructions, preserving appearance consistency, and supporting flexible input formats. To study this gap, we introduce \textbf{RISEBench}, the first benchmark for evaluatingReasoning-Informed viSualEditing (RISE). RISEBench focuses on four key reasoning categories:Temporal,Causal,Spatial, andLogical Reasoning. We curate high-quality test cases for each category and propose an robust evaluation framework that assessesInstruction Reasoning,Appearance Consistency, andVisual Plausibilitywith both human judges and the LMM-as-a-judge approach. We conducted experiments evaluating eight prominent visual editing models, comprising both open-source and proprietary models. The evaluation results demonstrate that current models face significant challenges in reasoning-based editing tasks. Even the most powerful model evaluated, GPT-4o-Image, achieves an accuracy of merely 28.8\%. RISEBench effectively highlights the limitations of contemporary editing models, provides valuable insights, and indicates potential future directions for the field of reasoning-aware visual editing.



Paperid:2782
Authors:Behrad Moniri, Hamed Hassani
Abstract:
Weak-to-strong generalization—where a student model trained on imperfect labels generated by a weaker teacher nonetheless surpasses that teacher—has been widely observed, but the mechanisms that enable it have remained poorly understood. In this paper, through a theoretical analysis of simple models, we uncover three core mechanisms that can drive this phenomenon. First, by analyzing ridge regression, we study the interplay between the teacher and student regularization and prove that a student can compensate for a teacher’s under-regularization and achieve lower test error. We also analyze the role of the parameterization regime of the models. Second, by analyzing weighted ridge regression, we show that a student model with a regularization structure more aligned to the target, can outperform its teacher. Third, in a nonlinear multi‐index setting, we demonstrate that a student can learn easy, task-specific features from the teacher while leveraging its own broader pre-training to learn hard‐to‐learn features that the teacher cannot capture.



Paperid:2783
Authors:Ziqi Gu, Chunyan Xu, Wenxuan Fang, Yide Qiu, Xin Liu, Zhen Cui
Abstract:
Multi-domain task incremental learning (MTIL) demands models to master domain-specific expertise while preserving generalization capabilities. Inspired by human lifelong learning, which relies on revisiting, aligning, and integrating past experiences, we propose a Learning and Ensembling Bridge Adapters (LEBA) framework. To facilitate cohesive knowledge transfer across domains, specifically, we propose a continuous-domain bridge adaptation module, leveraging the distribution transfer capabilities of Schrödinger bridge for stable progressive learning.To strengthen memory consolidation, we further propose a progressive knowledge ensemble strategy that revisits past task representations via a diffusion model and dynamically integrates historical adapters.For efficiency, LEBA maintains a compact adapter pool through similarity-based selection and employs learnable weights to align replayed samples with current task semantics. Together, these components effectively mitigate catastrophic forgetting and enhance generalization across tasks.Extensive experiments across multiple benchmarks validate the effectiveness and superiority of LEBA over state-of-the-art methods.



Paperid:2784
Authors:Alexandre Oliveira, Katarina Dyreby, Francisco Caldas, Claudia Soares
Abstract:
The increasing number of satellites and orbital debris has made space congestion a critical issue, threatening satellite safety and sustainability. Challenges such as collision avoidance, station-keeping, and orbital maneuvering require advanced techniques to handle dynamic uncertainties and multi-agent interactions. Reinforcement learning (RL) has shown promise in this domain, enabling adaptive, autonomous policies for space operations; however, many existing RL frameworks rely on custom-built environments developed from scratch, which often use simplified models and require significant time to implement and validate the orbital dynamics, limiting their ability to fully capture real-world complexities. To address this, we introduce OrbitZoo, a versatile multi-agent RL environment built on a high-fidelity industry standard library, that enables realistic data generation, supports scenarios like collision avoidance and cooperative maneuvers, and ensures robust and accurate orbital dynamics. The environment is validated against a real satellite constellation, Starlink, achieving a Mean Absolute Percentage Error (MAPE) of 0.16% compared to real-world data. This validation ensures reliability for generating high-fidelity simulations and enabling autonomous and independent satellite operations.



Paperid:2785
Authors:Jiabin Lin, Shana Moothedath
Abstract:
Abstract:Learning-to-learn or meta-learning focuses on developing algorithms that leverage prior experience to quickly acquire new skills or adapt to novel environments. A crucial component of meta-learning is representation learning, which aims to construct data representations capable of transferring knowledge across multiple tasks—a critical advantage in data-scarce settings. We study how representation learning can improve the efficiency of bandit problems. We consider $T$ $d$-dimensional linear bandits that share a common low-dimensional linear representation. We provide provably fast, sample-efficient algorithms to address the two key problems in meta-learning: (1) learning a common set of features from multiple related bandit tasks and (2) transferring this knowledge to new, unseen bandit tasks. We validated the theoretical results through numerical experiments using real-world and synthetic datasets, comparing them against benchmark algorithms.



Paperid:2786
Authors:Andrew M. Bean, Ryan Othniel Kearns, Angelika Romanou, Franziska Sofia Hafner, Harry Mayne, Jan Batzner, Negar Foroutan Eghlidi, Chris Schmitz, Karolina Korgul, Hunar Batra, Oishi Deb, Emma Beharry, Cornelius Emde, Thomas Foster, Anna Gausen, María Grandury, Sophia Simeng Han, Valentin Hofmann, Lujain Ibrahim, Hazel Kim, Hannah Rose Kirk, Fangru Lin, Gabrielle Liu, Lennart Luettgau, Jabez Magomere, Jonathan Rystrøm, Anna Sotnikova, Yushi Yang, Yilun Zhao, Adel Bibi, Antoine Bosselut, Ronald Clark, Arman Cohan, Jakob Foerster, Yarin Gal, Scott Hale, Deborah Raji, Christopher Summerfield, Philip Torr, Cozmin Ududec, Luc Rocher, Adam Mahdi
Abstract:
Evaluating large language models (LLMs) is crucial for both assessing their capabilities and identifying safety or robustness issues prior to deployment. Reliably measuring abstract and complex phenomena such as safety and robustness requires strong construct validity, that is, having measures that represent what matters to the phenomenon. With a team of 29 expert reviewers, we conduct a systematic review of 445 LLM benchmarks from leading conferences in natural language processing and machine learning. Across the reviewed articles, we find patterns related to the measured phenomena, tasks, and scoring metrics which undermine the validity of the resulting claims. To address these shortcomings, we provide eight key recommendations and detailed actionable guidance to researchers and practitioners in developing LLM benchmarks.



Paperid:2787
Authors:Oren Neumann, Claudius Gros
Abstract:
Neural scaling laws are observed in a range of domains, to date with no clear understanding of why they occur. Recent theories suggest that loss power laws arise from Zipf's law, a power law observed in domains like natural language. One theory suggests that language scaling laws emerge when Zipf-distributed task quanta are learned in descending order of frequency. In this paper we examine power-law scaling in AlphaZero, a reinforcement learning algorithm, using a theory of language-model scaling. We find that game states in training and inference data scale with Zipf's law, which is known to arise from the tree structure of the environment, and examine the correlation between scaling-law and Zipf's-law exponents. In agreement with quanta scaling theory, we find that agents optimize state loss in descending order of frequency, even though this order scales inversely with modelling complexity. We also find that inverse scaling, the failure of models to improve with size, is correlated with unusual Zipf curves where end-game states are among the most frequent states. We show evidence that larger models shift their focus to these less-important states, sacrificing their understanding of important early-game states.



Paperid:2788
Authors:Jialong Zuo, Yongtai Deng, Mengdan Tan, Rui Jin, Dongyue Wu, Nong Sang, Liang Pan, Changxin Gao
Abstract:
In real-word scenarios, person re-identification (ReID) expects to identify a person-of-interest via the descriptive query, regardless of whether the query is a single modality or a combination of multiple modalities. However, existing methods and datasets remain constrained to limited modalities, failing to meet this requirement. Therefore, we investigate a new challenging problem called Omni Multi-modal Person Re-identification (OM-ReID), which aims to achieve effective retrieval with varying multi-modal queries. To address dataset scarcity, we construct ORBench, the first high-quality multi-modal dataset comprising 1,000 unique identities across five modalities: RGB, infrared, color pencil, sketch, and textual description. This dataset also has significant superiority in terms of diversity, such as the painting perspectives and textual information. It could serve as an ideal platform for follow-up investigations in OM-ReID. Moreover, we propose ReID5o, a novel multi-modal learning framework for person ReID. It enables synergistic fusion and cross-modal alignment of arbitrary modality combinations in a single model, with a unified encoding and multi-expert routing mechanism proposed. Extensive experiments verify the advancement and practicality of our ORBench. A wide range of possible models have been evaluated and compared on it, and our proposed ReID5o model gives the best performance. The dataset and code will be made publicly available.



Paperid:2789
Authors:Yueyang Hu, Haiyong Jiang, Haoxuan Song, Jun Xiao, Hao Pan
Abstract:
This work presents a novel framework for few-shot 3D part segmentation. Recent advances have demonstrated the significant potential of 2D foundation models for low-shot 3D part segmentation. However, it is still an open problem that how to effectively aggregate 2D knowledge from foundation models to 3D. Existing methods either ignore geometric structures for 3D feature learning or neglects the high-quality grouping clues from SAM, leading to under-segmentation and inconsistent part labels. We devise a novel SAM segment graph-based propagation method, named SegGraph, to explicitly learn geometric features encoded within SAM's segmentation masks. Our method encodes geometric features by modeling mutual overlap and adjacency between segments while preserving intra-segment semantic consistency. We construct a segment graph, conceptually similar to an atlas, where nodes represent segments and edges capture their spatial relationships (overlap/adjacency). Each node adaptively modulates 2D foundation model features, which are then propagated via a graph neural network to learn global geometric structures. To enforce intra-segment semantic consistency, we map segment features to 3D points with a novel view-direction-weighted fusion attenuating contributions from low-quality segments. Extensive experiments on PartNet-E demonstrate that our method outperforms all competing baselines by at least 6.9% mIoU. Further analysis reveals that SegGraph achieves particularly strong performance on small components and part boundaries, demonstrating its superior geometric understanding.



Paperid:2790
Authors:Siran Dai, Qianqian Xu, Peisong Wen, Yang Liu, Qingming Huang
Abstract:
Abstract:In this work, we observe a counterintuitive phenomenon in self-supervised learning (SSL): longer training may impair the performance of dense downstream tasks (e.g., semantic segmentation). We refer to this phenomenon as Self-supervised Dense Degradation (SDD) and demonstrate its consistent presence across ten state-of-the-art SSL methods with various losses, architectures, and datasets. When the model performs suboptimally on dense tasks at the end of training, measuring the performance during training becomes essential. However, evaluating dense performance effectively without annotations remains an open challenge.To tackle this issue, we introduce a Dense representation Quality Estimator (DQE), composed of a class-relevance measure and an effective dimensionality measure. The proposed DQE is both theoretically grounded and empirically validated to be closely correlated with the downstream performance. Based on this metric, we introduce a straightforward yet effective model selection strategy and a DQE-based regularization method. Experiments on ten SSL methods across four benchmarks confirm that model selection improves mIoU by $4.0\\%$ on average with negligible computational cost. Additionally, DQE regularization consistently mitigates the effects of dense degradation. Code is provided in the supplementary material.



Paperid:2791
Authors:Adrian Mirza, Nawaf Alampara, Martiño Ríos-García, Mohamed Abdelalim, Jack Butler, Bethany Connolly, Tunca Dogan, Marianna Nezhurina, Bünyamin Şen, Santosh Tirunagari, Mark Worrall, Adamo Young, Philippe Schwaller, Michael Pieler, Kevin Maik Jablonka
Abstract:
Foundation models have shown remarkable success across scientific domains, yet their impact in chemistry remains limited due to the absence of diverse, large-scale, high-quality datasets that reflect the field's multifaceted nature. We present the ChemPile, an open dataset containing over 75 billion tokens of curated chemical data, specifically built for training and evaluating general-purpose models in the chemical sciences. The dataset mirrors the human learning journey through chemistry---from educational foundations to specialized expertise---spanning multiple modalities and content types including structured data in diverse chemical representations (SMILES, SELFIES, IUPAC names, InChI, molecular renderings), scientific and educational text, executable code, and chemical images. ChemPile integrates foundational knowledge (textbooks, lecture notes), specialized expertise (scientific articles and language-interfaced data), visual understanding (molecular structures, diagrams), and advanced reasoning (problem-solving traces and code)---mirroring how human chemists develop expertise through diverse learning materials and experiences. Constructed through hundreds of hours of expert curation, the ChemPile captures both foundational concepts and domain-specific complexity. We provide standardized training, validation, and test splits, enabling robust benchmarking. ChemPile is openly released via HuggingFace with a consistent API, permissive license, and detailed documentation. We hope the ChemPile will serve as a catalyst for chemical AI, enabling the development of the next generation of chemical foundation models.



Paperid:2792
Authors:Fan Xu, Hao Wu, Nan Wang, Lilan Peng, Kun Wang, Wei Gong, Xibin Zhao
Abstract:
The modeling of complicated time-evolving physical dynamics from partial observations is a long-standing challenge. Particularly, observations can be sparsely distributed in a seemingly random or unstructured manner, making it difficult to capture highly nonlinear features in a variety of scientific and engineering problems. However, existing data-driven approaches are often constrained by fixed spatial and temporal discretization. While some researchers attempt to achieve spatio-temporal continuity by designing novel strategies, they either overly rely on traditional numerical methods or fail to truly overcome the limitations imposed by discretization. To address these, we propose CoPS, a purely data-driven methods, to effectively model continuous physics simulation from partial observations. Specifically, we employ multiplicative filter network to fuse and encode spatial information with the corresponding observations. Then we customize geometric grids and use message-passing mechanism to map features from original spatial domain to the customized grids. Subsequently, CoPS models continuous-time dynamics by designing multi-scale graph ODEs, while introducing a Marcov-based neural auto-correction module to assist and constrain the continuous extrapolations. Comprehensive experiments demonstrate that CoPS advances the state-of-the-art methods in space-time continuous modeling across various scenarios. Our codes are available at~\url{https://anonymous.4open.science/r/CoPS-F625}.



Paperid:2793
Authors:Amitay Sicherman, Kira Radinsky
Abstract:
Predicting interactions between biochemical entities is a core challenge in drug discovery and systems biology, often hindered by limited data and poor generalization to unseen entities. Traditional discriminative models frequently underperform in such settings. We propose BioCG (Biochemical Constrained Generation), a novel framework that reformulates interaction prediction as a constrained sequence generation task. BioCG encodes target entities as unique discrete sequences via Iterative Residual Vector Quantization (I-RVQ) and trains a generative model to produce the sequence of an interacting partner given a query entity. A trie-guided constrained decoding mechanism, built from a catalog of valid target sequences, concentrates the model's learning on the critical distinctions between valid biochemical options, ensuring all outputs are biochemically valid. An information-weighted training objective further focuses learning on the most critical decision points. BioCG achieves state-of-the-art (SOTA) performance across diverse tasks, Drug-Target Interaction (DTI), Drug-Drug Interaction (DDI), and Enzyme-Reaction Prediction, especially in data-scarce and cold-start conditions. On the BioSNAP DTI benchmark, for example, BioCG attains an AUC of 89.31\% on unseen proteins, representing a 14.3 percentage point gain over prior SOTA. By directly generating valid interacting partners within a known biochemical space, BioCG provides a robust and data-efficient solution for in-silico biochemical discovery.



Paperid:2794
Authors:Anamika Agrawal, Michael Buice
Abstract:
The simple linear threshold units used in many artificial neural networks have a limited computational capacity. Famously, a single unit cannot handle non-linearly separable problems like XOR. By contrast, real neurons exhibit complex morphologies as well as active dendritic integration, suggesting that their computational capacities outstrip those of simple linear units. Considering specific families of Boolean functions, we examine the computational limits of single units that incorporate more complex dendritic structures. For random Boolean functions, we show that there is a phase transition in learnability as a function of the input dimension, with most random functions below a certain critical dimension being learnable and those above not. This critical dimension is best predicted by the overall size of the dendritic arbor. This demonstrates real neurons have a far higher computational complexity than is usually considered in neural models, whether in machine learning or computational neuroscience. Furthermore, using architectures that are, respectively, more "apical" or "basal", we show that there are non-trivially disjoint sets of learnable functions by each type of neuron. Importantly, these two types of architectures differ in the robustness and generality of the computations they can perform. The basal-like architecture shows a higher probability of function realization, while the apical-like architecture shows an advantage with fast retraining for different functions. Given the cell-type specificity of morphological characteristics, these results suggest both that different components of the dendritic arbor as well as distinct cell types may have distinct computational roles. Our analysis offers new directions for neuron-level inductive biases in NeuroAI models using scalable models for neuronal cell-type specific computation.



Paperid:2795
Authors:Tao Zhong, Jonah Buchanan, Christine Allen-Blanchette
Abstract:
We propose a new approach to vision-based dexterous grasp translation, which aims to transfer grasp intent across robotic hands with differing morphologies. Given a visual observation of a source hand grasping an object, our goal is to synthesize a functionally equivalent grasp for a target hand without requiring paired demonstrations or hand-specific simulations. We frame this problem as a stochastic transport between grasp distributions using the Schrödinger Bridge formalism. Our method learns to map between source and target latent grasp spaces via score and flow matching, conditioned on visual observations. To guide this translation, we introduce physics-informed cost functions that encode alignment in base pose, contact maps, wrench space, and manipulability. Experiments across diverse hand-object pairs demonstrate our approach generates stable, physically grounded grasps with strong generalization. This work enables semantic grasp transfer for heterogeneous manipulators and bridges vision-based grasping with probabilistic generative modeling.



Paperid:2796
Authors:Yunlong Deng, Guangyi Chen, Tianpei Gu, Lingjing Kong, Yan Li, Zeyu Tang, Kun Zhang
Abstract:
Abstract:Vision-Language Models (VLMs) integrate visual knowledge with the analytical capabilities of Large Language Models (LLMs) through supervised visual instruction tuning, using image-question-answer triplets. However, the potential of VLMs trained without supervised instruction remains largely unexplored. This study validates that VLMs possess inherent self-refinement capabilities, enabling them to generate high-quality supervised data without external inputs and thereby learn autonomously. Specifically, to stimulate the self-refinement ability of VLMs, we propose a self-refinement framework based on a Triangular Consistency principle: within the image-query-answer triangle, any masked elements should be consistently and accurately reconstructed. The framework involves three steps: (1) We enable the instruction generation ability of VLMs by adding multi-task instruction tuning like image$\rightarrow$question-answer or image-answer$\rightarrow$question. (2) We generate image-query-answer triplets from unlabeled images and use the Triangular Consistency principle for filtering. (3) The model is further updated using the filtered synthetic data. To investigate the underlying mechanisms behind this self-refinement capability, we conduct a theoretical analysis from a causal perspective. Using the widely recognized LLaVA-1.5 as our baseline, our experiments reveal that the model can autonomously achieve consistent, though deliberately modest, improvements across multiple benchmarks without any external supervision, such as human annotations or environmental feedback.We expect that the insights of this study on the self-refinement ability of VLMs can inspire future research on the learning mechanism of VLMs.



Paperid:2797
Authors:Zhenyuan Chen, Chenxi Wang, Ningyu Zhang, Feng Zhang
Abstract:
Remote sensing is critical for disaster monitoring, yet existing datasets lack temporal image pairs and detailed textual annotations. While single-snapshot imagery dominates current resources, it fails to capture dynamic disaster impacts over time. To address this gap, we introduce the Remote Sensing Change Caption (RSCC) dataset, a large-scale benchmark comprising 62,315 pre-/post-disaster image pairs (spanning earthquakes, floods, wildfires, and more) paired with rich, human-like change captions. By bridging the temporal and semantic divide in remote sensing data, RSCC enables robust training and evaluation of vision-language models for disaster-aware bi-temporal understanding. Our results highlight RSCC’s ability to facilitate detailed disaster-related analysis, paving the way for more accurate, interpretable, and scalable vision-language applications in remote sensing. Code and dataset are available at https://github.com/Bili-Sakura/RSCC.



Paperid:2798
Authors:Xinnan Dai, Kai Yang, Jay Revolinsky, Kai Guo, Aoran Wang, Bohang Zhang, Jiliang Tang
Abstract:
Recent studies suggest that large language models (LLMs) possess the capability to solve graph reasoning tasks. Notably, even when graph structures are embedded within textual descriptions, LLMs can still effectively answer related questions. This raises a fundamental question: How can a decoder-only Transformer architecture understand underlying graph structures? To address this, we start with the substructure extraction task, interpreting the inner mechanisms inside the transformers and analyzing the impact of the input queries. Specifically, through both empirical results and theoretical analysis, we present Induced Substructure Filtration (ISF), a perspective that captures the substructure identification in the multi-layer transformers. We further validate the ISF process in LLMs, revealing consistent internal dynamics across layers. Building on these insights, we explore the broader capabilities of Transformers in handling diverse graph types. Specifically, we introduce the concept of thinking in substructures to efficiently extract complex composite patterns, and demonstrate that decoder-only Transformers can successfully extract substructures from attributed graphs, such as molecular graphs. Together, our findings offer a new insight on how sequence-based Transformers perform the substructure extraction task over graph data.



Paperid:2799
Authors:Tianbao Xie, Jiaqi Deng, Xiaochuan Li, Junlin Yang, Haoyuan Wu, Jixuan Chen, Wenjing Hu, Xinyuan Wang, Yuhui Xu, Zekun Wang, Yiheng Xu, Junli Wang, Doyen Sahoo, Tao Yu, Caiming Xiong
Abstract:
Graphical user interface (GUI) grounding, the ability to map natural language instructions to specific actions on graphical user interfaces, remains a critical bottleneck in computer use agent development. Current benchmarks oversimplify grounding tasks as short referring expressions, failing to capture the complexity of real-world interactions that require software commonsense, layout understanding, and fine-grained manipulation capabilities. To address these limitations, we introduce OSWorld-G, a comprehensive benchmark comprising 564 finely annotated samples across diverse task types including text matching, element recognition, layout understanding, and precise manipulation. Additionally, we synthesize and release the largest computer use grounding dataset Jedi, which contains 4 million examples through multi-perspective decoupling of tasks. Our multi-scale models trained on Jedi demonstrate its effectiveness by outperforming existing approaches on ScreenSpot-v2, ScreenSpot-Pro, and our OSWorld-G. Furthermore, we demonstrate that improved grounding with Jedi directly enhances agentic capabilities of general foundation models on complex computer tasks, improving from 5% to 27% on OSWorld. Through detailed ablation studies, we identify key factors contributing to grounding performance and verify that combining specialized data for different interface elements enables compositional generalization to novel interfaces. All benchmark, data, checkpoints, and code are open-sourced and available at https://osworld-grounding.github.io.



Paperid:2800
Authors:Sangwoo Kwon, Seong Hoon Seo, Jae Lee, Yeonhong Park
Abstract:
How can we effectively handle queries for on-device large language models (LLMs) with varying runtime constraints, such as latency and accuracy? Multi-scale quantization addresses this challenge by enabling memory-efficient runtime model adaptation of LLMs through the overlaying of multiple model variants quantized to different bitwidths. Meanwhile, an important question still remains open-ended: how can models be properly configured to match a target precision or latency? While mixed-precision offers a promising solution, we take this further by leveraging the key observation that the sensitivity of each layer dynamically changes across decoding iterations. Building on this insight, we introduce DP-LLM, a novel mechanism that dynamically assigns precision to each layer based on input values. DP-LLM augments each linear layer in an LLM with a precision selector that determines the bitwidth at runtime using a lightweight error estimator and threshold values learned through fine-tuning. Experimental results across multiple models and benchmarks demonstrate that DP-LLM achieves a superior performance-latency trade-off, outperforming prior approaches.



Paperid:2801
Authors:Jingjing Jiang, Chongjie Si, Jun Luo, Hanwang Zhang, Chao Ma
Abstract:
This paper presents a pioneering exploration of reinforcement learning (RL) via group relative policy optimization for unified multimodal large language models (ULMs), aimed at simultaneously reinforcing generation and understanding capabilities. Through systematic pilot studies, we uncover the significant potential of ULMs to enable the synergistic co-evolution of dual capabilities within a shared policy optimization framework. Building on this insight, we introduce \textbf{CoRL}, a co-reinforcement learning framework comprising a unified RL stage for joint optimization and a refined RL stage for task-specific enhancement. With the proposed CoRL, our resulting model, \textbf{ULM-R1}, achieves average improvements of \textbf{7\%} on three text-to-image generation datasets and \textbf{23\%} on nine multimodal understanding benchmarks. These results demonstrate the effectiveness of CoRL and highlight the substantial benefit of reinforcement learning in facilitating cross-task synergy and optimization for ULMs.



Paperid:2802
Authors:Tianrui Wang, Haoyu Wang, Meng Ge, Cheng Gong, Chunyu Qiang, Ziyang Ma, Zikang Huang, Guanrou Yang, Xiaobao Wang, Eng-Siong Chng, Xie Chen, Longbiao Wang, Jianwu Dang
Abstract:
While emotional text-to-speech (TTS) has made significant progress, most existing research remains limited to utterance-level emotional expression and fails to support word-level control. Achieving word-level expressive control poses fundamental challenges, primarily due to the complexity of modeling multi-emotion transitions and the scarcity of annotated datasets that capture intra-sentence emotional and prosodic variation. In this paper, we propose WeSCon, the first self-training framework that enables word-level control of both emotion and speaking rate in a pretrained zero-shot TTS model, without relying on datasets containing intra-sentence emotion or speed transitions.Our method introduces a transition-smoothing strategy and a dynamic speed control mechanism to guide the pretrained TTS model in performing word-level expressive synthesis through a multi-round inference process. To further simplify the inference, we incorporate a dynamic emotional attention bias mechanism and fine-tune the model via self-training, thereby activating its ability for word-level expressive control in an end-to-end manner. Experimental results show that WeSCon effectively overcomes data scarcity, achieving state-of-the-art performance in word-level emotional expression control while preserving the strong zero-shot synthesis capabilities of the original TTS model.



Paperid:2803
Authors:Hanwen Liu, Longjiao Zhang, Rui Wang, Tongya Zheng, Sai Wu, Chang Yao, Mingli Song
Abstract:
Dynamic graph learning is crucial for accurately modeling complex systems by integrating topological structure and temporal information within graphs. While memory-based methods are commonly used and excel at capturing short-range temporal correlations, they struggle with modeling long-range dependencies, harmonizing long-range and short-range correlations, and integrating structural information effectively. To address these challenges, we present SALoM: Structure Aware Temporal Graph Networks with Long-Short Memory Updater. SALoM features a memory module that addresses gradient vanishing and information forgetting, enabling the capture of long-term dependencies across various time scales. Additionally, SALoM utilizes a long-short memory updater (LSMU) to dynamically balance long-range and short-range temporal correlations, preventing over-generalization. By integrating co-occurrence encoding and LSMU through information bottleneck-based fusion, SALoM effectively captures both the structural and temporal information within graphs. Experimental results across various graph datasets demonstrate SALoM's superior performance, achieving state-of-the-art results in dynamic graph link prediction. Our code is openly accessible at https://anonymous.4open.science/r/SALoM-BF7C.



Paperid:2804
Authors:Vincent Leon, Iosif Sakos, Ryann Sim, Antonios Varvitsiotis
Abstract:
Concavity and its refinements underpin tractability in multiplayer games, where players independently choose actions to maximize their own payoffs which depend on other players’ actions. Inconcavegames, where players' strategy sets are compact and convex, and their payoffs are concave in their own actions, strong guarantees follow: Nash equilibria always exist and decentralized algorithms converge to equilibria. If the game is furthermoremonotone, an even stronger guarantee holds: Nash equilibria are unique under strictness assumptions. Unfortunately, we show thatcertifyingconcavity or monotonicity is NP-hard, already for games where utilities are multivariate polynomials and compact, convex basic semialgebraic strategy sets -- an expressive class that captures extensive-form games with imperfect recall. On the positive side, we develop two hierarchies of sum-of-squares programs that certify concavity and monotonicity of a given game, and each level of the hierarchies can be solved in polynomial time. We show that almost all concave/monotone games are certified at some finite level of the hierarchies. Subsequently, we introduce the classes of SOS-concave/monotone games, which globally approximate concave/monotone games, and show that for any given game we can compute the closest SOS-concave/monotone game in polynomial time. Finally, we apply our techniques to canonical examples of extensive-form games with imperfect recall.



Paperid:2805
Authors:Yibo Wen, Chenwei Xu, Jerry Yao-Chieh Hu, Kaize Ding, Han Liu
Abstract:
We present a three-stage framework for training deep learning models specializing in antibody sequence-structure co-design.We first pre-train a language model using millions of antibody sequence data.Then, we employ the learned representations to guide the training of a diffusion model for joint optimization over both sequence and structure of antibodies. During the final alignment stage, we optimize the model to favor antibodies with low repulsion and high attraction to the antigen binding site, enhancing the rationality and functionality of the designs.To mitigate conflicting energy preferences, we extend AbDPO (Antibody Direct Preference Optimization) to guide the model toward Pareto optimality under multiple energy-based alignment objectives. Furthermore, we adopt an iterative learning paradigm with temperature scaling, enabling the model to benefit from diverse online datasets without requiring additional data.In practice, our proposed methods achieve high stability and efficiency in producing a better Pareto front of antibody designs compared to top samples generated by baselines and previous alignment techniques.Through extensive experiments, we showcase the superior performance of our methods in generating nature-like antibodies with high binding affinity.



Paperid:2806
Authors:Tonghe Zhang, Chao Yu, Sichang Su, Yu Wang
Abstract:
We propose ReinFlow, a simple and effective online reinforcement learning (RL) framework that fine-tunes a family of flow matching policies for continuous robotic control. ReinFlow injects a bounded, learnable noise into a flow policy's deterministic path, converting the flow into a discrete-time Markov Process for simple and tractable likelihood computation. This conversion aids exploration and secures training stability, allowing ReinFlow to stably fine-tune diverse flow model variants, including Rectified Flow and Shortcut Models, especially at very few or even one denoising step. We benchmark ReinFlow in representative locomotion and manipulation tasks, including long-horizon planning with visual input and sparse reward. The episode reward of Rectified Flow policies increased by an average of 162.28\% after fine-tuning in challenging legged locomotion tasks, while saving 82.87\% of wall-time compared to the state-of-the-art diffusion RL method DPPO. The success rate of the Shortcut Model policies in state and visual manipulation tasks increased by 39.86\% on average after fine-tuning with ReinFlow at four or even one denoising step, achieving performance comparable to fine-tuned DDIM policies while saving 48.78\% of the simulation time.



Paperid:2807
Authors:Jifeng Hu, Sili Huang, Zhejian Yang, Shengchao Hu, Li Shen, Hechang Chen, Lichao Sun, Yi Chang, Dacheng Tao
Abstract:
Conditional decision generation with diffusion models has shown powerful competitiveness in reinforcement learning (RL). Recent studies reveal the relation between energy-function-guidance diffusion models and constrained RL problems. The main challenge lies in estimating the intermediate energy, which is intractable due to the log-expectation formulation during the generation process. To address this issue, we propose the Analytic Energy-guided Policy Optimization (AEPO). Specifically, we first provide a theoretical analysis and the closed-form solution of the intermediate guidance when the diffusion model obeys the conditional Gaussian transformation. Then, we analyze the posterior Gaussian distribution in the log-expectation formulation and obtain the target estimation of the log-expectation under mild assumptions. Finally, we train an intermediate energy neural network to approach the target estimation of log-expectation formulation. We apply our method in 30+ offline RL tasks to demonstrate the effectiveness of our method. Extensive experiments illustrate that our method surpasses numerous representative baselines in D4RL offline reinforcement learning benchmarks.



Paperid:2808
Authors:Max Staats, Matthias Thamm, Bernd Rosenow
Abstract:
This work analyzes singular-value spectra of weight matrices in pretrained transformer models to understand how information is stored at both ends of the spectrum. Using Random Matrix Theory (RMT) as a zero information hypothesis, we associate agreement with RMT as evidence of randomness and deviations as evidence for learning. Surprisingly, we observe pronounced departures from RMT not only among the largest singular values -- the usual outliers -- but also among the smallest ones. A comparison of the associated singular vectors with the eigenvectors of the activation covariance matrices shows that there is considerable overlap wherever RMT is violated. Thus, significant directions in the data are captured by small singular values and their vectors as well as by the large ones. We confirm this empirically: zeroing out the singular values that deviate from RMT raises language-model perplexity far more than removing values from the bulk, and after fine-tuning the smallest decile can be the third most influential part of the spectrum. To explain how vectors linked to small singular values can carry more information than those linked to larger values, we propose a linear random-matrix model. Our findings highlight the overlooked importance of the low end of the spectrum and provide theoretical and practical guidance for SVD-based pruning and compression of large language models. Our findings highlight the overlooked importance of small singular values and provide theoretical and practical guidance for SVD-based pruning and compression of large language models.



Paperid:2809
Authors:Sebastian Wagner-Carena, Aizhan Akhmetzhanova, Sydney Erickson
Abstract:
In the natural sciences, a common challenge is to disentangle distinct, unknown sources from observations. Examples of this source separation task include deblending galaxies in a crowded field, distinguishing the activity of individual neurons from overlapping signals, and separating seismic events from the ambient background. Traditional analyses often rely on simplified source models that fail to accurately reproduce the data. Recent advances have shown that diffusion models can directly learn complex prior distributions from noisy, incomplete data. In this work, we show that diffusion models can solve the source separation problem without explicit assumptions about the source. Our method relies only on multiple views, or the property that different sets of observations contain different linear transformations of the unknown sources. We show that our method succeeds even when no source is individually observed and the observations are noisy, incomplete, and vary in resolution. The learned diffusion models enable us to sample from the source priors, evaluate the probability of candidate sources, and draw from the joint posterior of our sources given an observation. We demonstrate the effectiveness of our method on a range of synthetic problems as well as real-world galaxy observations.



Paperid:2810
Authors:Kazuki Irie, Morris Yau, Samuel J Gershman
Abstract:
We develop hybrid memory architectures for general-purpose sequence processing neural networks, that combine key-value memory using softmax attention (KV-memory) with dynamic synaptic memory through fast-weight programming (FW-memory)---the core principles of quadratic and linear transformers, respectively. These two memory systems have complementary but individually limited properties: KV-memory offers precise retrieval but is constrained by quadratic complexity in sequence length, while FW-memory supports arbitrarily long sequences and enables more expressive computation but sacrifices precise recall. We propose and compare three methods to blend these two systems into a single memory system to leverage the strengths of both. We conduct experiments on general language modeling and retrieval tasks by training 340M- and 1.3B-parameter models from scratch, as well as on synthetic algorithmic tasks designed to precisely illustrate the benefits of certain hybrid methods over others. Overall, we demonstrate how a well-designed hybrid can overcome the limitations of its individual components, offering new insights into the design principle of neural memory systems.



Paperid:2811
Authors:Marwa Abdulhai, Ryan Cheng, Donovan Clay, Tim Althoff, Sergey Levine, Natasha Jaques
Abstract:
Large Language Models (LLMs) are increasingly used to simulate human users in interactive settings such as therapy, education, and social role-play. While these simulations enable scalable training and evaluation of AI agents, off-the-shelf LLMs often drift from their assigned personas, contradict earlier statements, or abandon role-appropriate behavior. We introduce a unified framework for evaluating and improving persona consistency in LLM-generated dialogue. We define three automatic metrics—prompt-to-line consistency, line-to-line consistency, and Q\&A consistency—that capture different types of persona drift and validate each against human annotations. Using these metrics as reward signals, we apply multi-turn reinforcement learning to fine-tune LLMs for three user roles: a patient, a student, and a social chat partner. Our method reduces inconsistency by over 55\%, resulting in more coherent, faithful, and trustworthy simulated users.



Paperid:2812
Authors:Xiaomeng Fan, Yuchuan Mao, Zhi Gao, Yuwei Wu, Jin Chen, Yunde Jia
Abstract:
Open-vocabulary learning requires modeling the data distribution in open environments, which consists of both seen-class and unseen-class data. Existing methods estimate the distribution in open environments using seen-class data, where the absence of unseen classes makes the estimation error inherently unidentifiable. Intuitively, learning beyond the seen classes is crucial for distribution estimation to bound the estimation error. We theoretically demonstrate that the distribution can be effectively estimated by generating unseen-class data, through which the estimation error is upper-bounded. Building on this theoretical insight, we propose a novel open-vocabulary learning method, which generates unseen-class data for estimating the distribution in open environments.The method consists of a class-domain-wise data generation pipeline and a distribution alignment algorithm.The data generation pipeline generates unseen-class data under the guidance of a hierarchical semantic tree and domain information inferred from the seen-class data, facilitating accurate distribution estimation.With the generated data, the distribution alignment algorithm estimates and maximizes the posterior probability to enhance generalization in open-vocabulary learning.Extensive experiments on 11 datasets demonstrate that our method outperforms baseline approaches by up to 14%, highlighting its effectiveness and superiority.



Paperid:2813
Authors:Yuzhe YANG, Yifei Zhang, Minghao Wu, Kaidi Zhang, Yunmiao Zhang, Honghai Yu, Yan Hu, Benyou Wang
Abstract:
The study of social emergence has long been a central focus in social science. Traditional modeling approaches, such as rule-based Agent-Based Models (ABMs), struggle to capture the diversity and complexity of human behavior, particularly the irrational factors emphasized in behavioral economics. Recently, large language model (LLM) agents have gained traction as simulation tools for modeling human behavior in social science and role-playing applications. Studies suggest that LLMs can account for cognitive biases, emotional fluctuations, and other non-rational influences, enabling more realistic simulations of socio-economic dynamics. In this work, we introduce TwinMarket, a novel multi-agent framework that leverages LLMs to simulate socio-economic systems. Specifically, we examine how individual behaviors, through interactions and feedback mechanisms, give rise to collective dynamics and emergent phenomena. Through experiments in a simulated stock market environment, we demonstrate how individual actions can trigger group behaviors, leading to emergent outcomes such as financial bubbles and recessions. Our approach provides valuable insights into the complex interplay between individual decision-making and collective socio-economic patterns.



Paperid:2814
Authors:Afra Amini, Tim Vieira, Ryan Cotterell
Abstract:
Estimating the Kullback--Leibler (KL) divergence between language models has many applications, e.g., reinforcement learning from human feedback (RLHF), interpretability, and knowledge distillation. However, computing the exact KL divergence between two arbitrary language models is intractable. Thus, practitioners often resort to the use of sampling-based estimators. While it is easy to fashion a simple Monte Carlo (MC) estimator that provides an unbiased estimate of the KL divergence between language models, this estimator notoriously suffers from high variance, and can even result in a negative estimate of the KL divergence, a non-negative quantity. In this paper, we introduce a Rao--Blackwellized estimator that is also unbiased and provably has variance less than or equal to that of the standard Monte Carlo estimator. In an empirical study on sentiment-controlled fine-tuning, we show that our estimator provides more stable KL estimates and reduces variance substantially in practice. Additionally, we derive an analogous Rao--Blackwellized estimator of the gradient of the KL divergence, which leads to more stable training and produces models that more frequently appear on the Pareto frontier of reward vs. KL compared to the ones trained with the MC estimator of the gradient.



Paperid:2815
Authors:Yuning Shen, Lihao Wang, Huizhuo Yuan, Yan Wang, Bangji Yang, Quanquan Gu
Abstract:
Understanding protein dynamics is critical for elucidating their biological functions. The increasing availability of molecular dynamics (MD) data enables the training of deep generative models to efficiently explore the conformational space of proteins.However, existing approaches either fail to explicitly capture the temporal dependencies between conformations or do not support direct generation of time-independent samples.To address these limitations, we introduceConfRover, an autoregressive model that simultaneously learns protein conformation and dynamics from MD trajectory data, supporting both time-dependent and time-independent sampling.At the core of our model is a modular architecture comprising: (i) anencoding layer, adapted from protein folding models, that embeds protein-specific information and conformation at each time frame into a latent space; (ii) atemporal module, a sequence model that captures conformational dynamics across frames; and (iii) an SE(3) diffusion model as thestructure decoder, generating conformations in continuous space.Experiments on ATLAS, a large-scale protein MD dataset of diverse structures, demonstrate the effectiveness of our model in learning conformational dynamics and supporting a wide range of downstream tasks.ConfRoveris the first model to sample both protein conformations and trajectories within a single framework, offering a novel and flexible approach for learning from protein MD data.



Paperid:2816
Authors:Yize Cheng, Vinu Sankar Sadasivan, Mehrdad Saberi, Shoumik Saha, Soheil Feizi
Abstract:
The increasing capabilities of Large Language Models (LLMs) have raised concerns about their misuse in AI-generated plagiarism and social engineering. While various AI-generated text detectors have been proposed to mitigate these risks, many remain vulnerable to simple evasion techniques such as paraphrasing. However, recent detectors have shown greater robustness against such basic attacks. In this work, we introduce \textbf{Adversarial Paraphrasing}, a training-free attack framework that universally humanizes any AI-generated text to evade detection more effectively. Our approach leverages an off-the-shelf instruction-following LLM to paraphrase AI-generated content under the guidance of an AI text detector, producing adversarial examples that are specifically optimized to bypass detection. Extensive experiments show that our attack is both broadly effective and highly transferable across several detection systems. For instance, compared to simple paraphrasing attack—which, ironically, increases the true positive at 1\% false positive (T@1\%F) by 8.57\% on RADAR and 15.03\% on Fast-DetectGPT—adversarial paraphrasing, guided by OpenAI-RoBERTa-Large, reduces T@1\%F by 64.49\% on RADAR and a striking 98.96\% on Fast-DetectGPT. Across a diverse set of detectors—including neural network-based, watermark-based, and zero-shot approaches—our attack achieves an average T@1\%F reduction of 87.88\% under the guidance of OpenAI-RoBERTa-Large. We also analyze the tradeoff between text quality and our attack success to find that our method can significantly reduce detection rates, with mostly a slight degradation in text quality.Our novel adversarial setup highlights the need for more robust and resilient detection strategies in the light of increasingly sophisticated evasion techniques.



Paperid:2817
Authors:Junda Zhu, Jun Ai, Yujun Li, Yichun Yin, Yasheng Wang, Lifeng Shang, Qun Liu
Abstract:
As one of the state-of-the-art parameter-efficient fine-tuning~(PEFT) methods, Low-Rank Adaptation (LoRA) enables model optimization with reduced computational cost through trainable low-rank matrix. However, the low-rank nature makes it prone to produce a decrease in the representation ability, leading to suboptimal performance. In order to break this limitation, we propose RidgeLoRA, a lightweight architecture like LoRA that incorporates novel architecture and matrix ridge enhanced full-rank approximation, to match the performance of full-rank training, while eliminating the need for high memory and a large number of parameters to restore the rank of matrices. We provide a rigorous mathematical derivation to prove that RidgeLoRA has a better upper bound on the representations than vanilla LoRA. Furthermore, extensive experiments across multiple domains demonstrate that RidgeLoRA achieves better performance than other LoRA variants, and can even match or surpass full-rank training.



Paperid:2818
Authors:Christoph Jürgen Hemmer, Daniel Durstewitz
Abstract:
Complex, temporally evolving phenomena, from climate to brain activity, are governed by dynamical systems (DS). DS reconstruction (DSR) seeks to infer generative surrogate models of these from observed data, reproducing their long-term behavior. Existing DSR approaches require purpose-training for any new system observed, lacking the zero-shot and in-context inference capabilities known from LLMs. Here we introduceDynaMix, a novel multivariate ALRNN-based mixture-of-experts architecture pre-trained for DSR, the first DSR model able to generalize zero-shot to out-of-domain DS. Just from a provided context signal, without any re-training, DynaMix faithfully forecasts the long-term evolution of novel DS where existing time series (TS) foundation models, like Chronos, fail -- at a fraction of the number of parameters and orders of magnitude faster inference times. DynaMix outperforms TS foundation models in terms of long-term statistics, and often also short-term forecasts, even on real-world time series, like traffic or weather data, typically used for training and evaluating TS models,but not at all part of DynaMix' training corpus. We illustrate some of the failure modes of TS models for DSR problems, and conclude that models built on DS principles may bear a huge potential also for advancing the TS prediction field.



Paperid:2819
Authors:Konstantinos Oikonomidis, Jan Quan, Panagiotis Patrinos
Abstract:
We study preconditioned gradient methods for smooth nonconvex optimization problems, focusing on sigmoid preconditioners that inherently perform a form of gradient clipping akin to the widely used gradient clipping technique. Building upon this idea, we introduce a novel heavy ball-type algorithm and provide convergence guarantees under a generalized smoothness condition that is less restrictive than traditional Lipschitz smoothness, thus covering a broader class of functions. Additionally, we develop a stochastic variant of the base method and study its convergence properties under different noise assumptions. We compare the proposed algorithms with baseline methods on diverse tasks from machine learning including neural network training.



Paperid:2820
Authors:Gongzhe Li, Linwei Qiu, Peibei Cao, Fengying Xie, Xiangyang Ji, Qilin Sun
Abstract:
High dynamic range (HDR) images, with their rich tone and detail reproduction, hold significant potential to enhance computer vision systems, particularly in autonomous driving. However, most neural networks for embedded vision are trained on low dynamic range (LDR) inputs and suffer substantial performance degradation when handling high-bit-depth HDR images due to the challenges posed by extreme dynamic ranges.In this paper, we propose a novel tone mapping method that not only bridges the gap between HDR RAW inputs and the LDR sRGB requirements of detection networks but also achieves end-to-end optimization with the downstream tasks.Instead of relying on traditional image signal processing (ISP) pipeline, we introduce neural photometric calibration to regularize dynamic ranges and a scaling-invariant local tone mapping module to preserve image details.In addition, our architecture also supports performance transfer finetuning, enabling efficient adaptation from the LDR model to the HDR RAW model with minimal cost. The proposed method outperforms traditional tone mapping algorithms and advanced AI-ISP methods in challenging automotive HDR scenes. Moreover, our pipeline achieves real-time processing of 4K high-bit-depth HDR inputs on the Nvidia Jetson platform.



Paperid:2821
Authors:Daniel Tschernutter, David Castro, Maciej Kasiński
Abstract:
While energy-based models have recently proven to be a powerful framework for learning to reason with neural networks, their practical application is still limited by computational cost. That is, existing methods for energy-based iterative reasoning suffer from computational bottlenecks by relying on expensive optimization routines during training and especially during inference. Furthermore, these routines may not always converge to minimal energy states, potentially leading to suboptimal reasoning. To address these limitations, we propose a novel and efficient algorithm for energy-based iterative reasoning based on a difference-of-convex (DC) functions approach. Our algorithm achieves a significant speedup compared to prior methods while offering theoretical convergence guarantees ensuring locally minimal energy states. In addition, we achieve state-of-the-art or superior performance on continuous reasoning tasks, as demonstrated by our experiments on multiple benchmark datasets from continuous algorithmic reasoning. As such, our method offers a leap in computational efficiency, enabling faster inference with theoretical guarantees, and hence unlocking the potential of energy-based models for iterative reasoning applications.



Paperid:2822
Authors:Ziyu Wan, Yunxiang Li, Xiaoyu Wen, Yan Song, Hanjing Wang, Linyi Yang, Mark Schmidt, Jun Wang, Weinan Zhang, Shuyue Hu, Ying Wen
Abstract:
Recent research on Reasoning of Large Language Models (LLMs) has sought to further enhance their performance by integrating meta-thinking—enabling models to monitor, evaluate, and control their reasoning processes for more adaptive and effective problem-solving.However, current single-agent work lacks a specialized design for acquiring meta-thinking, resulting in low efficacy.To address this challenge, we introduce Reinforced Meta-thinking Agents (ReMA), a novel framework that leverages Multi-Agent Reinforcement Learning (MARL) to elicit meta-thinking behaviors, encouraging LLMs to think about thinking.ReMA decouples the reasoning process into two hierarchical agents: a high-level meta-thinking agent responsible for generating strategic oversight and plans, and a low-level reasoning agent for detailed executions.Through iterative reinforcement learning with aligned objectives, these agents explore and learn collaboration, leading to improved generalization and robustness.Empirical results from single-turn experiments demonstrate that ReMA outperforms single-agent RL baselines on complex reasoning tasks, including competitive-level mathematical benchmarks and LLM-as-a-Judge benchmarks.Additionally, we further extend ReMA to multi-turn interaction settings, leveraging turn-level ratio and parameter sharing to improve efficiency.Comprehensive ablation studies further illustrate the evolving dynamics of each distinct agent, providing valuable insights into how the meta-thinking reasoning process enhances the reasoning capabilities of LLMs.



Paperid:2823
Authors:Kejing Xia, Jidong Jia, Ke Jin, Yucai Bai, Li Sun, Dacheng Tao, Youjian Zhang
Abstract:
Recently, Gaussian Splatting (GS) has shown great potential for urban scenesreconstruction in the field of autonomous driving. However, current urban scenereconstruction methods often depend on multimodal sensors as inputs, i.e. LiDARand images. Though the geometry prior provided by LiDAR point cloud can largelymitigates ill-posednesses in reconstruction, acquiring such accurate LiDAR data isstill challenging in practice: i) precise spatiotemporal calibration between LiDARand other sensors is required, as they may not capture data simultaneously; ii)reprojection errors will arise due to spatial misalignment when LiDAR and camerasensors are mounted at distinct locations. To avoid the difficulty of acquiringaccurate LiDAR data, we propose D2GS, a LiDAR-free urban scene reconstructionframework. In this work, we obtain geometry priors that are as effective as thosefrom LiDAR, but are denser and more accurate. First, a dense point cloud isinitialized by predicting metric depths via a multi-view depth estimation method.This point cloud is then optimized by a progressive pruning strategy to improvethe global consistency. Second, we iteratively optimize the accuracy of predicteddepths and the rendering depths. Specifically,We leverage diffusion priors from adepth foundation model to enhance the depth maps rendered by Gaussians. In turn,the enhanced depths give better geometry constrain to the training of the Gaussians.Finally, We improve the accuracy of ground geometry by constraining the shapeand normal attributes of Gaussians within road regions. Extensive experiments onthe Waymo dataset demonstrate that our method consistently outperforms state-of-the-art methods, producing more accurate geometry even when compared withthose using ground truth LiDAR data.



Paperid:2824
Authors:Hossein Zakerinia, Christoph Lampert
Abstract:
We present new fast-rate generalization bounds for multi-task and meta-learning in the unbalanced setting, i.e. when the tasks have training sets of different sizes, as is typically the case in real-world scenarios. Previously, only standard-rate bounds were known for this situation, while fast-rate bounds were limited to the setting where all training sets are of equal size. Our new bounds are numerically computable as well as interpretable, and we demonstrate their flexibility in handling a number of cases where they give stronger guarantees than previous bounds. Besides the bounds themselves, we also make conceptual contributions: we demonstrate that the unbalanced multi-task setting has different statistical properties than the balanced situation, specifically that proofs from the balanced situation do not carry over to the unbalanced setting. Additionally, we shed light on the fact that the unbalanced situation allows two meaningful definitions of multi-task risk, depending on whether if all tasks should be considered equally important or if sample-rich tasks should receive more weight than sample-poor ones.



Paperid:2825
Authors:Pengfei Zhao, Rongbo Luan, Wei Zhang, Peng Wu, Sifeng He
Abstract:
Despite Contrastive Language–Image Pre-training (CLIP)'s remarkable capability to retrieve content across modalities, a substantial modality gap persists in its feature space. Intriguingly, we discover that off-the-shelf MLLMs (Multimodal Large Language Models) demonstrate powerful inherent modality alignment properties. While recent MLLM-based retrievers with unified architectures partially mitigate this gap, their reliance on coarse modality alignment mechanisms fundamentally limits their potential. In this work, We introduce MAPLE (Modality-Aligned Preference Learning for Embeddings), a novel framework that leverages the fine-grained alignment priors inherent in MLLM to guide cross-modal representation learning. MAPLE formulates the learning process as reinforcement learning with two key components: (1) Automatic preference data construction using off-the-shelf MLLM, and (2) a new Relative Preference Alignment (RPA) loss, which adapts Direct Preference Optimization (DPO) to the embedding learning setting. Experimental results show that our preference-guided alignment achieves substantial gains in fine-grained cross-modal retrieval, underscoring its effectiveness in handling nuanced semantic distinctions.



Paperid:2826
Authors:Jinglin Liang, Jin Zhong, Shuangping Huang, Yunqing Hu, Huiyuan Zhang, Huifang Li, Lixin Fan, Hanlin Gu
Abstract:
In this paper, we explore an important yet previously neglected question: Do context aggregation patterns across Language Models (LMs) share commonalities?While some works have investigated context aggregation or attention weights in LMs, they typically focus on individual models or attention heads, lacking systematic analysis across multiple LMs to explore their commonalities.In contrast, we focus on the commonalities among LMs, which can deepen our understanding of LMs and even facilitate cross-model knowledge transfer.In this work, we introduce the Order-Level Attention (OLA) derived from the order-wise decomposition of Attention Rollout and reveal that the OLA at the same order across LMs exhibits significant similarities.Furthermore, we discover an implicit mapping between OLA and syntactic knowledge.Based on these two findings, we propose the Transferable OLA Adapter (TOA), a training-free cross-LM adapter transfer method.Specifically, we treat the OLA as a unified syntactic feature representation and train an adapter that takes OLA as input.Due to the similarities in OLA across LMs, the adapter generalizes to unseen LMs without requiring any parameter updates or training data.Extensive experiments demonstrate that TOA's cross-LM generalization effectively enhances the performance of unseen LMs. We will open-source our code to ensure stable and straightforward reproducibility.



Paperid:2827
Authors:Steeve Laquitaine, Simone Azeglio, Carlo Paris, Ulisse Ferrari, Matthew Chalk
Abstract:
Abstract:A central question in sensory neuroscience is how much, but also what information neurons transmit about the world. While Shannon’s information theory provides a principled framework to quantify the amount of information neurons encode about all stimuli, it does not reveal which stimuli contribute most, or what stimulus features are encoded. As a concrete example, it is known that neurons in the early visual cortex are 'sensitive' to stimuli in a small region of space (their receptive field). However, it is not clear how such simple intuitions carry to more complex scenarios, e.g. with large, noisy & non-linear population of neurons and high-dimensional stimuli.Several previous measures of neural sensitivity have been proposed. For example, the Fisher information quantifies the sensitivity of neural responses to infinitesimal stimulus perturbations. However, as the Fisher is not a valid decomposition of the mutual information it cannot say how different stimuli contribute to the total encoded information. On the other hand, previous works have proposed stimulus dependent decompositions of mutual information, which define a function $ I(x) $ such that $ I(R; X) = \mathbb{E}[I(x)] $. However, this decomposition is inherently ill-posed: infinitely many functions $I(x)$ satisfy the constraint, with no principled way to select among them. Further, different decompositions behave in qualitatively different ways, making it hard to interpret what are they are telling us. Finally, most proposed decompositions are computationally intractable for the high-dimensional stimuli and non-linear encoding models relevant for neuroscience.To resolve these limitations, we propose a set of axioms that any stimulus specific and feature-specific information decomposition should satisfy in order to serve as a meaningful and interpretable measure of neural sensitivity. These axioms formalize intuitive desiderata: that the information assigned to each stimulus, and stimulus feature, should be non-negative, and additive with respect to repeated measurements. We also require the decomposition to respect a form of locality: changes in how a neuron responds to a stimulus $ x $ should not affect the information attributed to a distant stimulus $ x' $. Finally, the attribution must be insensitive to irrelevant features, which do not contribute to the total information. Together, these constraints ensure that the decomposition is both interpretable and theoretically grounded. We show that existing decompositions violate one or more of these axioms, limiting their interpretability and use as information theoretic measures of neural sensitivity. We then introduce a novel decomposition that satisfies all of our axioms. It generalizes Fisher information by capturing neural sensitivity to both infinitesimal and finite stimulus perturbations. Moreover, it supports further decomposition across individual stimulus features (e.g., image pixels), enabling fine-grained analysis of neural representations.Beyond satisfying our theoretical axioms, our decomposition is computationally tractable for large neural populations and high-dimensional naturalistic stimuli, through the use of diffusion models. We demonstrate the power of our method by quantifying the information encoded by a model of visual neurons about individual images and pixels. Our approach uncovers aspects of the neural code that are not picked up by standard methods, such as the Fisher information, and opens the door to similar analyses in higher-order sensory areas, and artificial neural networks.



Paperid:2828
Authors:Weize Chen, Jiarui Yuan, Jin Tailin, Ning Ding, Huimin Chen, Zhiyuan Liu, Maosong Sun
Abstract:
Recent large language models (LLMs) exhibit impressive reasoning but often \textit{overthink}, generating excessively long responses that hinder efficiency. We introduce DIET (DIfficulty-AwarE Training), a framework that systematically cuts these "token calories" by integrating on-the-fly problem difficulty into the reinforcement learning (RL) process. DIET dynamically adapts token compression strategies by modulating token penalty strength and conditioning target lengths on estimated task difficulty, to optimize the performance-efficiency trade-off. We also theoretically analyze the pitfalls of naive reward weighting in group-normalized RL algorithms like GRPO, and propose \textit{Advantage Weighting} technique, which enables stable and effective implementation of these difficulty-aware objectives. Experimental results demonstrate that DIET significantly reduces token counts while simultaneously improving reasoning performance. Beyond raw token reduction, we show two crucial benefits largely overlooked by prior work: (1) DIET leads to superior \textbf{inference scaling}. By maintaining high per-sample quality with fewer tokens, it enables better scaling performance via majority voting under fixed computational budgets, an area where other methods falter. (2) DIET enhances the natural positive correlation between response length and problem difficulty, ensuring verbosity is appropriately allocated, unlike many existing compression methods that disrupt this relationship. Our analyses provide a principled and effective framework for developing more efficient, practical, and high-performing LLMs.



Paperid:2829
Authors:James Michaelov, Roger Levy, Benjamin Bergen
Abstract:
Abstract:We show that across architecture (transformer vs. RWKV vs. Mamba), training dataset (OpenWebText vs. The Pile), and scale (14 million parameters to 12 billion parameters), autoregressive language models exhibit highly consistent patterns of change in behavior through pre-training. Based on our analysis of over 1,400 language model checkpoints on over 110,000 tokens of English, we find that up to 98\% of the variance in language model behavior at the word level can be explained by three simple heuristics: the unigram probability (frequency) of a given word, the $n$-gram probability of the word, and the semantic similarity between the word and its context. Furthermore, we see consistent behavioral phases in all language models, with their predicted probabilities for words overfitting to the words' $n$-gram probabilities for increasing $n$ over the course of training, as well as becoming increasingly sensitive to semantic similarity early on in training. Taken together, these results suggest that learning in language models may take a similar path irrespective of model details.



Paperid:2830
Authors:Ke Liu, Shangde Gao, Yichao Fu, Shangqi Gao
Abstract:
Retinal vessel segmentation is critical for medical diagnosis, yet existing models often struggle to generalize across domains due to appearance variability, limited annotations, and complex vascular morphology. We propose GraphSeg, a variational Bayesian framework that integrates anatomical graph priors with structure-aware image decomposition to enhance cross-domain segmentation. GraphSeg factorizes retinal images into structure-preserved and structure-degraded components, enabling domain-invariant representation. A deformable graph prior, derived from a statistical retinal atlas, is incorporated via a differentiable alignment and guided by an unsupervised energy function. Experiments on three public benchmarks (CHASE, DRIVE, HRF) show that GraphSeg consistently outperforms existing methods under domain shifts. These results highlight the importance of jointly modeling anatomical topology and image structure for robust generalizable vessel segmentation.



Paperid:2831
Authors:Jinhong Deng, Wen Li, Joey Tianyi Zhou, Yang He
Abstract:
Multimodal Large Language Models (MLLMs) typically process a large number of visual tokens, leading to considerable computational overhead, even though many of these tokens are redundant. Existing visual token pruning methods primarily focus on selecting the most salient tokens based on attention scores, resulting in the semantic incompleteness of the selected tokens. In this paper, we propose a novel visual token pruning strategy, calledSaliency-CoverageOriented tokenPruning forEfficient MLLMs (SCOPE), to jointly model both the saliency and coverage of the selected visual tokens to better preserve semantic completeness. Specifically, we introduce a set-coverage for a given set of selected tokens, computed based on the token relationships. We then define a token-coverage gain for each unselected token, quantifying how much additional coverage would be obtained by including it. By integrating the saliency score into the token-coverage gain, we propose our SCOPE score and iteratively select the token with the highest SCOPE score. We conduct extensive experiments on multiple vision-language understanding benchmarks using the LLaVA-1.5 and LLaVA-Next models. Experimental results demonstrate that our method consistently outperforms prior approaches.



Paperid:2832
Authors:shrinivas ramasubramanian, Benjamin Freed, Alexandre Capone, Jeff Schneider
Abstract:
Recent works on supervised learning have showcased the benefits of converging to flatter minima in the loss landscape, including better generalization and robustness to noise and data distribution shifts. However, their role in Model-Based Reinforcement Learning remains underexplored. We explore the effects of converging to a flatter minima in the environment model's loss landscape in MBRL and demonstrate that encouraging convergence to flatter minima produces significantly better downstream policies. Our approach, which integrates Sharpness-Aware Minimization into the environment model training, enhances model robustness without architectural modifications, improving mean returns on average by 89.1% across the tested HumanoidBench tasks. Notably, we show that the detrimental effects of sharp minima become more pronounced as tasks with higher state-action dimensionality are considered. We supplement our empirical findings with novel theoretical guarantees, proving that convergence to flatter minima directly reduces the error in the model estimation of the policy return and the gap between model-optimal and true-optimal policies under mild assumptions. We perform experiments on the complex HumanoidBench Locomotion Suite. Our results demonstrate that integrating SAM significantly improves model performance when combined with the established state-of-the-art MBRL method TD-MPC2.



Paperid:2833
Authors:Yuancheng Wang, Dekun Chen, Xueyao Zhang, Junan Zhang, Jiaqi Li, Zhizheng Wu
Abstract:
Speech tokenizers serve as foundational components for speech language models, yet current designs exhibit several limitations, including: (1) dependence on multi-layer residual vector quantization structures or high frame rates,(2) reliance on auxiliary pre-trained models for semantic distillation, and(3) requirements for complex two-stage training processes.In this work, we introduce theText-awareDiffusion Transformer SpeechCodec(TaDiCodec), a novel approach designed to overcome these challenges.TaDiCodec employs end-to-end optimization for quantization and reconstruction through a diffusion autoencoder, while integrating text guidance into the diffusion decoder to enhance reconstruction quality and achieve optimal compression.TaDiCodec achieves an extremely low frame rate of6.25 Hzand a corresponding bitrate of0.0875 kbpswith asingle-layer codebookfor 24 kHz speech, while maintaining superior performance on critical speech generation evaluation metrics such as Word Error Rate (WER), speaker similarity (SIM), and speech quality (UTMOS),Notably, TaDiCodec employs a single-stage, end-to-end training paradigm, and obviating the need for auxiliary pre-trained models.We also validate the compatibility of TaDiCodec in language model based zero-shot text-to-speech with both autoregressive modeling and masked generative modeling, demonstrating its effectiveness and efficiency for speech language modeling, as well as a significantly smallreconstruction-generation gap.To facilitate reproducibility and further research, we will make our source code and pre-trained checkpoints publicly available.Audio samples are are available at https://tadicodec.github.io/.



Paperid:2834
Authors:Suzan Ece Ada, Georg Martius, Emre Ugur, Erhan Oztop
Abstract:
Offline Reinforcement Learning (RL) provides a promising avenue for training policies from pre-collected datasets when gathering additional interaction data is infeasible. However, existing offline RL methods often assume stationarity or only consider synthetic perturbations at test time—assumptions that often fail in real-world scenarios characterized by abrupt, time-varying offsets. These offsets can lead to partial observability, causing agents to misperceive their true state and degrade performance. To overcome this challenge, we introduce Forecasting in Offline RL (FORL), a framework that unifies (i) conditional diffusion-based candidate state generation, trained without presupposing any specific form of future non-stationarity, and (ii) zero-shot time-series foundation models. FORL targets environments prone to unexpected, potentially non-Markovian offsets, requiring robust agent performance from the onset of each episode. Empirical evaluations on offline RL benchmarks, augmented with real-world time-series data to simulate realistic non-stationarity, demonstrate that FORL consistently improves performance compared to competitive baselines. By integrating zero-shot forecasting with the agent’s experience we aim to bridge the gap between offline RL and the complexity of real-world, non-stationary environments.



Paperid:2835
Authors:Melanie Rieff, Maya Varma, Ossian Rabow, Subathra Adithan, Julie Kim, Ken Chang, Hannah Lee, Nidhi Rohatgi, Christian Bluethgen, Mohamed Muneer, Jean-Benoit Delbrouck, Michael Moor
Abstract:
Multimodal in-context learning (ICL) remains underexplored despite the profound potential it could have in complex application domains such as medicine. Clinicians routinely face a long tail of tasks which they need to learn to solve from few examples, such as considering few relevant previous cases or few differential diagnoses. While multimodal large language models (MLLMs) have shown impressive advances in medical visual question answering (VQA) or multi-turn chatting, their ability to learn multimodal tasks from context is largely unknown. Here, we introduce Stanford Multimodal Medical In-context Learning benchmark (SMMILE), the first multimodal medical ICL benchmark. A team of 11 clinical experts created ICL problems to scrutinize MLLM’s ability to learn multimodal tasks at inference time from context. We benchmark a suite of 14 diverse vision language models (VLMs) on SMMILE. Our results show that most VLMs gain little from multimodal in-context learning in medicine. Only two of the fourteen evaluated model, GPT-4o and Qwen 2.5-VL-32B, record double-digit improvements (>10 percentage points) over their zero-shot accuracy, and the mean gain across all models is about 4 percentage points. Further analysis reveals a clear recency bias: placing the most relevant example last boosts accuracy by more than 10 points, whereas adding noisy examples can cut performance by up to 6 points. SMMILE thus highlights critical limitations and biases in current medical multimodal AI.



Paperid:2836
Authors:Yeongtak Oh, Jisoo Mok, Dohyun Chung, Juhyeon Shin, Sangha Park, Johan Barthelemy, Sungroh Yoon
Abstract:
Recent multi-modal large language models (MLLMs) often struggle to generate personalized image captions, even when trained on high-quality captions. In this work, we observe that such limitations persist in existing post-training-based MLLM personalization methods. Specifically, despite being post-tuned with large-scale caption data through supervised fine-tuning (SFT), these models frequently fail to produce faithful descriptions in real-world scenarios, such as multi-concept image captioning. However, acquiring large-scale, high-quality captions for such complex settings is both costly and difficult. To address the data-centric nature of SFT, we propose a reinforcement learning (RL)-based post-training framework. To the best of our knowledge, this is the first RL-based approach to post-train MLLMs for personalized image captioning. Our method significantly enhances both visual recognition and personalized generation capabilities of MLLMs, and consistently outperforms existing SFT-based baselines, especially in the challenging multi-concept image captioning task.



Paperid:2837
Authors:Nan Song, Junzhe Jiang, jingyu li, Xiatian Zhu, Li Zhang
Abstract:
Motion simulation, prediction and planning are foundational tasks in autonomous driving, each essential for modeling and reasoning about dynamic traffic scenarios. While often addressed in isolation due to their differing objectives, such as generating diverse motion states or estimating optimal trajectories, these tasks inherently depend on shared capabilities: understanding multi-agent interactions, modeling motion behaviors, and reasoning over temporal and spatial dynamics.Despite this underlying commonality, existing approaches typically adopt specialized model designs, which hinders cross-task generalization and system scalability. More critically, this separation overlooks the potential mutual benefits among tasks. Motivated by these observations, we proposeUniMotion, a unified motion framework that captures shared structures across motion tasks while accommodating their individual requirements. Built on a decoder-only Transformer architecture, UniMotion employs dedicated interaction modes and tailored training strategies to simultaneously support these motion tasks. This unified design not only enables joint optimization and representation sharing but also allows for targeted fine-tuning to specialize in individual tasks when needed. Extensive experiments on the Waymo Open Motion Dataset (WOMD) demonstrate that joint training leads to robust generalization and effective task integration. With further fine-tuning, UniMotion achieves state-of-the-art performance across a range of motion tasks, establishing it as a versatile and scalable solution for autonomous driving.



Paperid:2838
Authors:Mattes Mollenhauer, Nicole Muecke, Dimitri Meunier, Arthur Gretton
Abstract:
This paper examines the performance of ridge regression in reproducing kernel Hilbert spaces in the presence of noise that exhibits a finite number of higher moments. We establish excessrisk bounds consisting of subgaussian and polynomial terms based on the well known integraloperator framework. The dominant subgaussian component allows to achieve convergence ratesthat have previously only been derived under subexponential noise—a prevalent assumption inrelated work from the last two decades. These rates are optimal under standard eigenvalue decayconditions, demonstrating the asymptotic robustness of regularized least squares against heavy-tailed noise. Our derivations are based on a Fuk–Nagaev inequality for Hilbert-space valuedrandom variables.



Paperid:2839
Authors:Wen Xiong, Junzhong Ji, Jinduo Liu
Abstract:
Pre-trained Large language models (LLMs) have shown impressive advancements in functional magnetic resonance imaging (fMRI) analysis and causal discovery. Considering the unique nature of the causal discovery field, which focuses on extracting causal graphs from observed data, research on LLMs in this field is still at an early exploratory stage. As a subfield of causal discovery, effective connectivity (EC) has received even less attention, and LLM-based approaches in EC remain unexplored. Existing LLM-based approaches for causal discovery typically rely on iterative querying to assess the causal influence between variable pairs, without any model adaptation or fine-tuning, making them ill-suited for handling the cross-modal gap and complex causal structures. To this end, we propose BrainEC-LLM, the first method to fine-tune LLMs for estimating brain EC from fMRI data. Specifically, multiscale decomposition mixing module decomposes fMRI time series data into short-term and long-term multiscale trends, then mixing them in bottom-up (fine to coarse) and top-down (coarse to fine) manner to extract multiscale temporal variations. And cross attention is applied with pre-trained word embeddings to ensure consistency between the fMRI input and pre-trained natural language. The experimental results on simulated and real resting-state fMRI datasets demonstrate that BrainEC-LLM can achieve superior performance when compared to state-of-the-art baselines.



Paperid:2840
Authors:Xinzhe Zheng, Hao Du, Fanding Xu, Jinzhe Li, ZHIYUAN LIU, Wenkang Wang, Tao Chen, Wanli Ouyang, Stan Z. Li, Yan Lu, Nanqing Dong, Yang Zhang
Abstract:
Deep learning-based computational methods have achieved promising results in predicting protein-protein interactions (PPIs). However, existing benchmarks predominantly focus on isolated pairwise evaluations, overlooking a model's capability to reconstruct biologically meaningful PPI networks, which is crucial for biology research. To address this gap, we introduce PRING, the first comprehensive benchmark that evaluates PRotein-protein INteraction prediction from a Graph-level perspective. PRING curates a high-quality, multi-species PPI network dataset comprising 21,484 proteins and 186,818 interactions, with well-designed strategies to address both data redundancy and leakage. Building on this golden-standard dataset, we establish two complementary evaluation paradigms: (1) topology-oriented tasks, which assess intra and cross-species PPI network construction, and (2) function-oriented tasks, including protein complex pathway prediction, GO module analysis, and essential protein justification. These evaluations not only reflect the model's capability to understand the network topology but also facilitate protein function annotation, biological module detection, and even disease mechanism analysis. Extensive experiments on four representative model categories, consisting of sequence similarity-based, naive sequence-based, protein language model-based, and structure-based approaches, demonstrate that current PPI models have potential limitations in recovering both structural and functional properties of PPI networks, highlighting the gap in supporting real-world biological applications. We believe PRING provides a reliable platform to guide the development of more effective PPI prediction models for the community. The dataset and source code of PRING are available at https://github.com/SophieSarceau/PRING.



Paperid:2841
Authors:Zhongju Yuan, Geraint Wiggins, Dick Botteldooren
Abstract:
Today’s deep learning architectures are primarily based on perceptron models, which do not capture the oscillatory dynamics characteristic of biological neurons. Although oscillatory systems have recently gained attention for their closer resemblance to neural behavior, they still fall short of modeling the intricate spatio-temporal interactions observed in natural neural circuits. In this paper, we propose a bio-inspired oscillatory state system (BioOSS) designed to emulate the wave-like propagation dynamics critical to neural processing, particularly in the prefrontal cortex (PFC), where complex activity patterns emerge. BioOSS comprises two interacting populations of neurons: p neurons, which represent simplified membrane-potential-like units inspired by pyramidal cells in cortical columns, and o neurons, which govern propagation velocities and modulate the lateral spread of activity. Through local interactions, these neurons produce wave-like propagation patterns. The model incorporates trainable parameters for damping and propagation speed, enabling flexible adaptation to task-specific spatio-temporal structures. We evaluate BioOSS on both synthetic and real-world tasks, demonstrating superior performance and enhanced interpretability compared to alternative architectures.



Paperid:2842
Authors:Xiaojun Jia, Sensen Gao, Simeng Qin, Tianyu Pang, Chao Du, Yihao Huang, Xinfeng Li, Yiming Li, Yang Liu, Bo Li
Abstract:
Multimodal large language models (MLLMs) remain vulnerable to transferable adversarial examples. While existing methods typically achieve targeted attacks by aligning global features—such as CLIP’s [CLS] token—between adversarial and target samples, they often overlook the rich local information encoded in patch tokens. This leads to suboptimal alignment and limited transferability, particularly for closed-source models. To address this limitation, we propose a targeted transferable adversarial attack method based on feature optimal alignment, called FOA-Attack, to improve adversarial transfer capability. Specifically, at the global level, we introduce a global feature loss based on cosine similarity to align the coarse-grained features of adversarial samples with those of target samples. At the local level, given the rich local representations within Transformers, we leverage clustering techniques to extract compact local patterns to alleviate redundant local features. We then formulate local feature alignment between adversarial and target samples as an optimal transport (OT) problem and propose a local clustering optimal transport loss to refine fine-grained feature alignment. Additionally, we propose a dynamic ensemble model weighting strategy to adaptively balance the influence of multiple models during adversarial example generation, thereby further improving transferability. Extensive experiments across various models demonstrate the superiority of the proposed method, outperforming state-of-the-art methods, especially in transferring to closed-source MLLMs.



Paperid:2843
Authors:Alessandro Serra, Francesco Ortu, Emanuele Panizon, Lucrezia Valeriani, Lorenzo Basile, Alessio Ansuini, Diego Doimo, Alberto Cazzaniga
Abstract:
Recent advances in multimodal training have significantly improved the integration of image understanding and generation within a unified model. This study investigates how vision-language models (VLMs) handle image-understanding tasks, specifically focusing on how visual information is processed and transferred to the textual domain. We compare VLMs that generate both images and text with those that output only text, highlighting key differences in information flow. We find that in models with multimodal outputs, image and text embeddings are more separated within the residual stream. Additionally, models vary in how information is exchanged from visual to textual tokens. VLMs that only output text exhibit a distributed communication pattern, where information is exchanged through multiple image tokens. In contrast, models trained for image and text generation tend to rely on a single token that acts as a narrow gate for visual information. We demonstrate that ablating this single token significantly deteriorates performance on image understanding tasks. Furthermore, modifying this token enables effective steering of the image semantics, showing that targeted, local interventions can reliably control the model's global behavior.



Paperid:2844
Authors:Luca Eyring, Shyamgopal Karthik, Alexey Dosovitskiy, Nataniel Ruiz, Zeynep Akata
Abstract:
The new paradigm of test-time scaling has yielded remarkable breakthroughs in Large Language Models (LLMs) (e.g.~reasoning models) and in generative vision models, allowing models to allocate additional computation during inference to effectively tackle increasingly complex problems. Despite the improvements of this approach, an important limitation emerges: the substantial increase in computation time makes the process slow and impractical for many applications. Given the success of this paradigm and its growing usage, we seek to preserve its benefits while eschewing the inference overhead. In this work we propose one solution to the critical problem of integrating test-time scaling knowledge into a model during post-training. Specifically, we replace reward guided test-time noise optimization in diffusion models with a Noise Hypernetwork that modulates initial input noise. We propose a theoretically grounded framework for learning this reward-tilted distribution for distilled generators, through a tractable noise-space objective that maintains fidelity to the base model while optimizing for desired characteristics. We show that our approach recovers a substantial portion of the quality gains from explicit test-time optimization at a fraction of the computational cost.



Paperid:2845
Authors:Wonjung Park, Suhyun Ahn, Maria Hernandez, Susana Maniega, Jinah Park
Abstract:
We propose BrainODE, a neural ordinary differential equations (ODEs)-based framework for modeling continuous longitudinal deformations of brain shapes. BrainODE learns a deformation space over anatomically meaningful brain regions to enable early prediction of neurodegenerative disease progression. Addressing key challenges in longitudinal neuroimaging— including limited data, irregular sampling intervals, and inter-subject variability—we design a conditional neural ODE architecture that modulates shape dynamics with subject-specific age and cognitive status. To enable autoregressive forecasting from a single observation, we introduce a pseudo-cognitive status embedding that jointly encodes age and estimated cognitive status. We demonstrate that BrainODE outperforms time-aware baselines in predicting future brain shapes, demonstrating strong generalization across longitudinal datasets with both regular and irregular time intervals.



Paperid:2846
Authors:Da Chang, Ganzhao Yuan
Abstract:
Efficient optimization is essential for training large language models. Although intra-layer selective updates have been explored, a general mechanism that enables fine-grained control while ensuring convergence guarantees is still lacking. To bridge this gap, we propose \textbf{MGUP}, a novel mechanism for selective updates. \textbf{MGUP} augments standard momentum-based optimizers by applying larger step-sizes to a selected fixed proportion of parameters in each iteration, while applying smaller, non-zero step-sizes to the rest. As a nearly \textbf{plug-and-play} module, \textbf{MGUP} seamlessly integrates with optimizers such as AdamW, Lion, and Muon. This yields powerful variants such as \textbf{MGUP-AdamW}, \textbf{MGUP-Lion}, and \textbf{MGUP-Muon}. Under standard assumptions, we provide theoretical convergence guarantees for \textbf{MGUP-AdamW} (without weight decay) in stochastic optimization. Extensive experiments across diverse tasks, including MAE pretraining, LLM pretraining, and downstream fine-tuning, demonstrate that our \textbf{MGUP}-enhanced optimizers achieve superior or more stable performance compared to their original base optimizers. We offer a principled, versatile, and theoretically grounded strategy for efficient intra-layer selective updates, accelerating and stabilizing the training of large-scale models.



Paperid:2847
Authors:Xiangyu Wen, Min Li, Junhua Huang, Jianyuan Zhong, Zhijian Xu, Zeju Li, Yongxiang Huang, Mingxuan Yuan, Qiang Xu
Abstract:
Large language models (LLMs) often produce reasoning steps that are superficially coherent yet internally inconsistent, leading to unreliable outputs. Since such failures typically arise from implicit or poorly-grounded knowledge, we introduce \emph{Grounded Reasoning in Dependency (GRiD)}, a novel dependency-aware reasoning framework that explicitly grounds reasoning steps in structured knowledge. GRiD represents reasoning as a graph consisting of interconnected knowledge extraction nodes and reasoning nodes, enforcing logical consistency through explicit dependencies. Each reasoning step is validated via a lightweight, step-wise verifier that ensures logical correctness relative to its premises. Extensive experiments across diverse reasoning benchmarks—including StrategyQA, CommonsenseQA, GPQA, and TruthfulQA—demonstrate that GRiD substantially improves reasoning accuracy, consistency, and faithfulness compared to recent state-of-the-art structured reasoning methods. Notably, GRiD enhances performance even when applied purely as a lightweight verification module at inference time, underscoring its generalizability and practical utility. Code will be made accessible upon acceptance.



Paperid:2848
Authors:Zhiwei Li, Yuesen Liao, Binrui Wu, Yuquan Zhou, Xupeng Shi, Dongsheng Jiang, Yin Li, Weizhong Zhang
Abstract:
Abstract:Pruning is a commonly employed technique for deep neural networks (DNNs) aiming at compressing the model size to reduce computational and memory costs during inference. In contrast to conventional neural networks, large language models (LLMs) pose a unique challenge regarding pruning efficiency due to their substantial computational and memory demands. Existing methods, particularly optimization-based ones, often require considerable computational resources in gradient estimation because they cannot effectively leverage weight sparsity of the intermediate pruned network to lower compuation and memory costs in each iteration. The fundamental challenge lies in the need to frequently instantiate intermediate pruned sub-models to achieve these savings, a task that becomes infeasible even for moderately sized neural networks. To this end, this paper proposes a novel pruning method for DNNs that is both computationally and memory-efficient. Our key idea is to develop an effective reweighting mechanism that enables us to estimate the gradient of the pruned network in current iteration via reweigting the gradient estimated on an outdated intermediate sub-model instantiated at an earlier stage, thereby significantly reducing model instantiation frequency. We further develop a series of techniques, e.g., clipping and preconditioning matrix, to reduce the variance of gradient estimation and stabilize the optimization process. We conducted extensive experimental validation across various domains. Our approach achieves 50\% sparsity and a 1.58$\times$ speedup in forward pass on Llama2-7B model with only 6 GB of memory usage, outperforming state-of-the-art methods with respect to both perplexity and zero-shot performance. As a by-product, our method is highly suited for distributed sparse training and can achieve a 2 $\times$ speedup over the dense distributed baselines.



Paperid:2849
Authors:Daniel Palenicek, Florian Vogt, Joe Watson, Jan Peters
Abstract:
Reinforcement learning has achieved significant milestones, but sample efficiency remains a bottleneck for real-world applications.Recently, CrossQ has demonstrated state-of-the-art sample efficiency with a low update-to-data (UTD) ratio of 1.In this work, we explore CrossQ's scaling behavior with higher UTD ratios.We identify challenges in the training dynamics, which are emphasized by higher UTD ratios.To address these, we integrate weight normalization into the CrossQ framework, a solution that stabilizes training, has been shown to prevent potential loss of plasticity, and keeps the effective learning rate constant.Our proposed approach reliably scales with increasing UTD ratios, achieving competitive performance across 25 challenging continuous control tasks on the DeepMind Control Suite and Myosuite benchmarks, notably the complex dog and humanoid environments.This work eliminates the need for drastic interventions, such as network resets, and offers a simple yet robust pathway for improving sample efficiency and scalability in model-free reinforcement learning.



Paperid:2850
Authors:Gonzalo E. Constante, Hao Chen, Can Li
Abstract:
Neural networks are increasingly deployed in safety-critical tasks where predictions must satisfy hard constraints, such as physical laws, fairness requirements, or safety limits. However, standard architectures lack built-in mechanisms to enforce such constraints, and existing approaches based on regularization or projection are often limited to simple constraints, computationally expensive, or lack feasibility guarantees. This paper proposes a model-agnostic framework for enforcing input-dependent linear equality and inequality constraints on neural network outputs. The architecture combines a task network trained for prediction accuracy with a safe network trained using decision rules from robust optimization to ensure feasibility across the entire input space. The final prediction is a convex combination of the two subnetworks, guaranteeing constraint satisfaction during both training and inference without iterative procedures or runtime optimization. We prove that the architecture is a universal approximator of constrained functions and derive computationally tractable formulations based on linear decision rules. Empirical results on benchmark regression tasks show that our method consistently satisfies constraints while maintaining competitive accuracy and low inference latency. This work bridges robust optimization and deep learning to enable reliable, constraint-aware neural network models suitable for real-time deployment.



Paperid:2851
Authors:Richard Suwandi, Feng Yin, Juntao Wang, Renjie Li, Tsung-Hui Chang, Sergios Theodoridis
Abstract:
The efficiency of Bayesian optimization (BO) relies heavily on the choice of the Gaussian process (GP) kernel, which plays a central role in balancing exploration and exploitation under limited evaluation budgets. Traditional BO methods often rely on fixed or heuristic kernel selection strategies, which can result in slow convergence or suboptimal solutions when the chosen kernel is poorly suited to the underlying objective function. To address this limitation, we propose a freshly-baked Context-Aware Kernel Evolution (CAKE) to enhance BO with large language models (LLMs). Concretely, CAKE leverages LLMs as the crossover and mutation operators to adaptively generate and refine GP kernels based on the observed data throughout the optimization process. To maximize the power of CAKE, we further propose BIC-Acquisition Kernel Ranking (BAKER) to select the most effective kernel through balancing the model fit measured by the Bayesian information criterion (BIC) with the expected improvement at each iteration of BO. Extensive experiments demonstrate that our fresh CAKE-based BO method consistently outperforms established baselines across a range of real-world tasks, including hyperparameter optimization, controller tuning, and photonic chip design. Our code is publicly available at \url{https://github.com/cake4bo/cake}.



Paperid:2852
Authors:Zhe Xu, Daoyuan Chen, Zhenqing Ling, Yaliang Li, Ying Shen
Abstract:
Large foundation models face challenges in acquiring transferable, structured thinking abilities, especially when supervised with rigid templates or crowd-annotated instruction datasets. Unlike prior approaches, we focus on a thinking-centric data synthesis paradigm that enables models to evolve through self-generated, cognitively guided data. We propose MindGYM, a structured and scalable framework for question synthesis, composed of: (1) Cognitive Thinking Process Injection, which infuses high-level reasoning objectives to shape the model’s synthesis behavior; (2) Seed Single-Hop Question Synthesis, generating atomic questions from diverse semantic types to encourage broader thinking; and (3) Challenging Multi-Hop QA Synthesis, composing more complex multi-hop questions based on QA seeds for deeper reasoning. Detailed analysis shows that synthetic data generated by our method achieves 16.7% higher average quality and 67.91% lower quality variance compared to baseline sources, highlighting that both high-quality and self-contained data are essential for effective, thinking-oriented fine-tuning. MindGYM improves performance on six reasoning benchmarks, achieving gains of up to 16% on MathVision using only 400 data samples, and generalizable improvements across different model sizes and architectures. MindGYM underscores the viability of self-challenging mechanisms in refining large model capabilities while minimizing human intervention and resource demands.Code and data are released to promote data-centric research into self-evolving foundation models driven by their internal reasoning capabilities.



Paperid:2853
Authors:Lanlan Ji, Dominic Seyler, Gunkirat Kaur, Manjunath Hegde, Koustuv Dasgupta, Bing Xiang
Abstract:
While Large Language Models (LLMs) show great promise, their tendencies to hallucinate pose significant risks in high-stakes domains like finance, especially when used for regulatory reporting and decision-making. Existing hallucination benchmarks fail to capture the complexities of financial benchmarks, which require high numerical precision, nuanced understanding of the language of finance, and ability to handle long-context documents. To address this, we introduce PHANTOM, a novel benchmark dataset for evaluating hallucination detection in long-context financial QA. Our approach first generates a seed dataset of high-quality "query-answer-document (chunk)" triplets, with either hallucinated or correct answers - that are validated by human annotators and subsequently expanded to capture various context lengths and information placements. We demonstrate how PHANTOM allows fair comparison of hallucination detection models and provides insights into LLM performance, offering a valuable resource for improving hallucination detection in financial applications. Further, our benchmarking results highlight the severe challenges out-of-the-box models face in detecting real-world hallucinations on long context data, and establish some promising directions towards alleviating these challenges, by fine-tuning open-source LLMs using PHANTOM.



Paperid:2854
Authors:Louis Béthune, David Vigouroux, Yilun Du, Rufin VanRullen, Thomas Serre, Victor Boutin
Abstract:
What is the shortest path between two data points lying in a high-dimensional space? While the answer is trivial in Euclidean geometry, it becomes significantly more complex when the data lies on a curved manifold—requiring a Riemannian metric to describe the space's local curvature. Estimating such a metric, however, remains a major challenge in high dimensions.In this work, we propose a method for deriving Riemannian metrics directly from pretrained Energy-Based Models (EBMs)—a class of generative models that assign low energy to high-density regions.These metrics define spatially varying distances, enabling the computation of geodesics—shortest paths that follow the data manifold’s intrinsic geometry. We introduce two novel metrics derived from EBMs and show that they produce geodesics that remain closer to the data manifold and exhibit lower curvature distortion, as measured by alignment with ground-truth trajectories.We evaluate our approach on increasingly complex datasets: synthetic datasets with known data density, rotated character images with interpretable geometry, and high-resolution natural images embedded in a pretrained VAE latent space.Our results show that EBM-derived metrics consistently outperform established baselines, especially in high-dimensional settings. Our work is the first to derive Riemannian metrics from EBMs, enabling data-aware geodesics and unlocking scalable, geometry-driven learning for generative modeling and simulation.



Paperid:2855
Authors:Mohan Zhang, Yihua Zhang, Jinghan Jia, Zhangyang "Atlas" Wang, Sijia Liu, Tianlong Chen
Abstract:
Modern Large Language Models (LLMs) demonstrate remarkable problem-solving capabilities through multi-step chain-of-thought reasoning, but this iterative thinking mechanism exposes a novel vulnerability surface. We introduce the Deadlock Attack, a resource exhaustion technique that hijacks an LLM's generative control flow by training a malicious adversarial embedding to force the model into perpetual thinking loops. The attack works by optimizing an embedding that encourages transitional tokens (e.g., "Wait", "But") after reasoning steps, preventing the model from reaching a natural conclusion. We identify a significant "continuous-to-discrete" gap when attempting to convert adversarial embeddings into discrete token sequences and propose a backdoor mechanism as an effective implantation strategy. Our experiments across four advanced reasoning LLMs (Phi-RM, Nemotron-Nano, R1-Qwen, R1-Llama) and three mathematical reasoning benchmarks demonstrate 100\% attack success rate, with models consistently generating up to maximum token limits, exhausting computational resources. Importantly, the attack remains stealthy, with minimal impact on model performance when the trigger is absent, and proves robust against existing defenses designed to mitigate overthinking. These findings reveal a critical security consideration for reasoning-enhanced LLMs, highlighting the urgent need for novel defense mechanisms against resource depletion attacks.



Paperid:2856
Authors:Minchan Jeong, Jongha (Jon) Ryu, Se-Young Yun, Gregory Wornell
Abstract:
Abstract:The Koopman operator provides a principled framework for analyzing nonlinear dynamical systems through linear operator theory. Recent advances in dynamic mode decomposition (DMD) have shown that trajectory data can be used to identify dominant modes of a system in a data-driven manner. Building on this idea, deep learning methods such as VAMPnets and DPNets have been proposed to learn the leading singular subspaces of the Koopman operator. However, these methods often rely on numerically unstable operations during objective computation, including singular value decomposition and matrix inversion, which introduce biased gradient estimates and limit scalability to large systems.In this work, we propose a scalable and conceptually simple method for learning the top-$L$ singular functions of the Koopman operator for stochastic dynamical systems based on a low-rank approximation framework. Our approach eliminates the need for unstable linear algebraic operations and integrates easily into modern deep learning pipelines. Empirical results demonstrate that the learned singular subspaces are both reliable and effective for downstream tasks such as multi-step prediction.



Paperid:2857
Authors:Shuaicheng Zhang, Haohui Wang, Junhong Lin, Xiaojie Guo, Yada Zhu, Si Zhang, Dongqi Fu, Dawei Zhou
Abstract:
Abstract:Graph heterophily, where connected nodes have different labels, has attracted significant interest recently. Most existing works adopt a simplified approach - using low-pass filters for homophilic graphs and high-pass filters for heterophilic graphs. However, we discover that the relationship between graph heterophily and spectral filters is more complex - the optimal filter response varies across frequency components and does not follow a strict monotonic correlation with heterophily degree. This finding challenges conventional fixed filter designs and suggests the need for adaptive filtering to preserve expressiveness in graph embeddings. Formally, natural questions arise: Given a heterophilic graph $\mathcal{G}$ , how and to what extent will the varying heterophily degree of $\mathcal{G}$ affect the performance of GNNs? How can we design adaptive filters to fit those varying heterophilic connections? Our theoretical analysis reveals that the average frequency response of GNNs and graph heterophily degree do not follow a strict monotonic correlation, necessitating adaptive graph filters to guarantee good generalization performance. Hence, we propose HeroFilter, a simple yet powerful GNN, which extracts information across the heterophily spectrum and combines salient representations through adaptive mixing. HeroFilter's superior performance achieves up to 9.2% accuracy improvement over leading baselines across homophilic and heterophilic graphs.



Paperid:2858
Authors:Junkai Zhang, Jinmin He, Yifan Zhang, Yifan Zang, Ning Xu, Jian Cheng
Abstract:
Multi-Agent Reinforcement Learning (MARL) has achieved remarkable success in various real-world scenarios, but its high cost of online training makes it impractical to learn each task from scratch. To enable effective policy reuse, we consider the problem of zero-shot generalization from offline data across multiple tasks. While prior work focuses on transferring individual skills of agents, we argue that the effective policy transfer across tasks should also capture the team-level coordination knowledge.In this paper, we propose Bi-Level Knowledge Transfer (BiKT) for Multi-Task MARL, which performs knowledge transfer at both the individual and team levels. At the individual level, we extract transferable individual skill embeddings from offline MARL trajectories.At the team level, we define tactics as coordinated patterns of skill combinations and capture them by leveraging the learned skill embeddings. We map skill combinations into compact tactic embeddings and then construct a tactic codebook.To incorporate both skills and tactics into decision-making, we design a bi-level decision transformer that infers them in sequence.Our BiKT leverages both the generalizability of individual skills and the diversity of tactics, enabling the learned policy to perform effectively across multiple tasks.Extensive experiments on SMAC and MPE benchmarks demonstrate that BiKT achieves strong generalization to previously unseen tasks.



Paperid:2859
Authors:Jiahao Wu, Ming Hu, Yanxin Yang, Xiaofei Xie, ZeKai Chen, Chenyu Song, Mingsong Chen
Abstract:
Although Federated Learning (FL) is promising in privacy-preserving collaborative model training, it faces low inference performance due to heterogeneous data among clients. Due to heterogeneous data in each client, FL training easily learns the specific overfitting features.Existing FL methods adopt the coarse-grained average aggregation strategy, which causes the global model to easily get stuck in local optima, resulting in low generalization of the global model.Specifically, this paper presents a novel FL framework named FedPhoenix to address this issue, which stochastically resets partial parameters to destroy some features of the global model in each round to guide the FL training to learn multiple generalized features for inference rather than specific overfitting features.Experimental results on various well-known datasets demonstrate that compared to SOTA FL methods, FedPhoenix can achieve up to 20.73\% accuracy improvement.



Paperid:2860
Authors:Pau Rodriguez, Michal Klein, Eleonora Gualdoni, Valentino Maiorca, Arno Blaas, Luca Zappella, Marco Cuturi, Xavier Suau
Abstract:
provide users with tools to explore style changes. Ideally, such mechanisms should require low volume of unpaired data (i.e., without explicit preference), and should be cheap, both at train and inference time, while preserving output quality. Recent research has shown that such mechanisms can be obtained by intervening exclusively on model activations, with the goal of correcting distributional differences between activations seen when using prompts from a source vs. a target set (e.g., toxic and non-toxic sentences). While cheap, these fast methods are inherently crude: their maps are tuned locally, not accounting for their impact on downstream layers, resulting in interventions that cause unintended shifts when used out-of-sample. We propose in this work linear end-to-end activation steering (LinEAS), an approach trained with a global loss that accounts simultaneously for all layer-wise distributional shifts. In addition to being more robust, the loss used to train LinEAS can be regularized with sparsifying norms, which can automatically carry out neuron selection. LinEAS only requires a handful of unpaired samples to be effective, and beats similar baselines on toxicity mitigation in language models, becoming competitive with oracle-dependent methods that have access to strong supervision. LinEAS is modality-agnostic and we empirically find that it outperforms existing activation steering methods at mitigating and including new concepts at the output of single-step text-to-image diffusion models.



Paperid:2861
Authors:Xin Yao, Haiyang Zhao, Yimin Chen, Jiawei Guo, Kecheng Huang, Ming Zhao
Abstract:
Abstract:The Contrastive Language-Image Pretraining (CLIP) model has significantly advanced vision-language modeling by aligning image-text pairs from large-scale web data through self-supervised contrastive learning. Yet, its reliance on uncurated Internet-sourced data exposes it to data poisoning and backdoor risks. While existing studies primarily investigate image-based attacks, the text modality—equally central to CLIP's training—remains underexplored. In this work, we introduce $\texttt{ToxicTextCLIP}$, a framework for generating high-quality adversarial texts that target CLIP during the pre-training phase. The framework addresses two key challenges: semantic misalignment caused by background inconsistency with the target class, and the scarcity of background-consistent texts. To this end, $\texttt{ToxicTextCLIP}$ iteratively applies: 1) a background-aware selector that prioritizes texts with background content aligned to the target class, and 2) a background-driven augmenter that generates semantically coherent and diverse poisoned samples. Extensive experiments on classification and retrieval tasks show that $\texttt{ToxicTextCLIP}$ achieves up to 95.83\% poisoning success and 98.68\% backdoor Hit@1, while bypassing RoCLIP and CleanCLIP defenses. The source code can be accessed via https://anonymous.4open.science/r/ToxicTextCLIP-4EC0/.



Paperid:2862
Authors:Tim van Erven, Jack Mayo, Julia Olkhovskaya, Chen-Yu Wei
Abstract:
Abstract:We present an efficient algorithm for linear contextual bandits with adversarial losses and stochastic action sets. Our approach reduces this setting to misspecification-robust adversarial linear bandits with fixed action sets. Without knowledge of the context distribution or access to a context simulator, the algorithm achieves $\tilde O(d^2\sqrt{T})$ regret and runs in poly$(d,C,T)$ time, where $d$ is the feature dimension, $C$ is the number of linear constraints defining the action set in each round, and $T$ is number of rounds. This resolves the open question by Liu et al. (2023) on whether one can obtain poly$(d)\sqrt{T}$ regret in polynomial time independent of the number of actions. For the important class of combinatorial bandits with adversarial losses and stochastic action sets, our algorithm is the first to achieve poly$(d)\sqrt{T}$ regret in polynomial time, while no prior algorithm achieves even $o(T)$ regret in polynomial time to our knowledge. When a simulator is available, the regret bound can be improved to $\tilde O(d\sqrt{L^\star})$, where $L^\star$ is the cumulative loss of the best policy.



Paperid:2863
Authors:Haolin Pan, Hongyu Lin, Haoran Luo, Yang Liu, Kaichun Yao, Libo Zhang, Mingjie Xing, Yanjun Wu
Abstract:
Compiler auto-tuning optimizes pass sequences to improve performance metrics such as Intermediate Representation (IR) instruction count. Although recent advances leveraging Large Language Models (LLMs) have shown promise in automating compiler tuning, two significant challenges still remain: the absence of high-quality reasoning datasets for agents training, and limited effective interactions with the compilation environment. In this work, we introduce Compiler-R1, the first reinforcement learning (RL)-driven framework specifically augmenting LLM capabilities for compiler auto-tuning. Compiler-R1 features a curated, high-quality reasoning dataset and a novel two-stage end-to-end RL training pipeline, enabling efficient environment exploration and learning through an outcome-based reward. Extensive experiments across seven datasets demonstrate Compiler-R1 achieving an average 8.46% IR instruction count reduction compared to opt -Oz, showcasing the strong potential of RL-trained LLMs for compiler optimization. Our code and datasets are publicly available at \url{https://anonymous.4open.science/r/Compiler-R1-C59C}.



Paperid:2864
Authors:Chao Li, Bingkun BAO, Yang Gao
Abstract:
In this paper, we consider fully decentralized cooperative multi-agent reinforcement learning, where each agent has access only to the states, its local actions, and the shared rewards. The absence of information about other agents' actions typically leads to the non-stationarity problem during per-agent value function updates, and the relative overgeneralization issue during value function estimation. However, existing works fail to address both issues simultaneously, as they lack the capability to model the agents' joint policy in a fully decentralized setting. To overcome this limitation, we propose a simple yet effective method named Return-Aware Context (RAC). RAC formalizes the dynamically changing task, as locally perceived by each agent, as a contextual Markov Decision Process (MDP), and addresses both non-stationarity and relative overgeneralization through return-aware context modeling. Specifically, the contextual MDP attributes the non-stationary local dynamics of each agent to switches between contexts, each corresponding to a distinct joint policy. Then, based on the assumption that the joint policy changes only between episodes, RAC distinguishes different joint policies by the training episodic return and constructs contexts using discretized episodic return values. Accordingly, RAC learns a context-based value function for each agent to address the non-stationarity issue during value function updates. For value function estimation, an individual optimistic marginal value is constructed to encourage the selection of optimal joint actions, thereby mitigating the relative overgeneralization problem. Experimentally, we evaluate RAC on various cooperative tasks (including matrix game, predator and prey, and SMAC), and its significant performance validates its effectiveness.



Paperid:2865
Authors:Yihao Li, Saeed Salehi, Lyle Ungar, Konrad Kording
Abstract:
Abstract:Object binding, the brain’s ability to bind the many features that collectively represent an object into a coherent whole, is central to human cognition. It groups low-level perceptual features into high‑level object representations, stores those objects efficiently and compositionally in memory, and supports human reasoning about individual object instances. While prior work often imposes object-centric attention (e.g., Slot Attention) explicitly to probe these benefits, it remains unclear whether this ability naturally emerges in pre-trained Vision Transformers (ViTs). Intuitively, they could: recognizing which patches belong to the same object should be useful for downstream prediction and thus guide attention. We hypothesize that these models thus represent whether two patches belong to the same object, a property we term $IsSameObject$, motivated by the pairwise nature of their self-attention mechanism. Here, we ask whether pre-trained ViTs spontaneously develop object binding. To test this, we decode $IsSameObject$ from patch embeddings across the hierarchy of layers in the ViT using a similarity probe. We demonstrate that our probe achieves over 90\% accuracy, outperforming pointwise probes that directly decode object type (85.7\%) or object identity (82.1\%), and this representation gradually emerges through the hierarchy. We further discover a low-dimensional subspace that encodes $IsSameObject$ on top of object features and show that early‑layer attention weights correlate strongly with this signal, which explains the utility of $IsSameObject$ representations. Ablating $IsSameObject$ from model activations degrades downstream performance and conflicts with self‑supervised pre-training loss minimization, implying that emergent object binding naturally serves the pretraining objective. Our findings challenge the view that ViTs lack object binding and highlight how symbolic knowledge of “which parts belong together” emerges naturally in a connectionist system.



Paperid:2866
Authors:Qiyuan Guan, Xiang Chen, Jiyu Jin, Guiyue Jin, Shumin Fan, Tianyu Song, Jinshan Pan
Abstract:
Compared to daytime image deraining, nighttime image deraining poses significant challenges due to inherent complexities of nighttime scenarios and the lack of high-quality datasets that accurately represent the coupling effect between rain and illumination. In this paper, we rethink the task of nighttime image deraining and contribute a new high-quality benchmark, HQ-NightRain, which offers higher harmony and realism compared to existing datasets. In addition, we develop an effective color space transformation framework (CST-Net) for better removing complex rain from nighttime scenes. Specifically, we propose a learnable color space converter (CSC) to better facilitate rain removal in the Y channel, as nighttime rain is more pronounced in the Y channel compared to the RGB color space. To capture illumination information for guiding nighttime deraining, implicit illumination guidance is introduced enabling the learned features to improve the model's robustness in complex scenarios. Extensive experiments show the value of our dataset and the effectiveness of our method. The code will be released soon.



Paperid:2867
Authors:Lianchen Jia, Chaoyang Li, Qian Houde, Tianchi Huang, Jiangchuan Liu, Lifeng Sun
Abstract:
Control algorithms in production environments typically require domain experts to tune them for specific scenarios. However, existing research predominantly focuses on algorithmic performance under ideal or default configurations, overlooking the critical aspect of tuning potential. To bridge this gap, we introduce \texttt{Crucible}, a novel framework leveraging large language models to quantitatively evaluate the tuning potential of control algorithms. We demonstrate \texttt{Crucible}'s effectiveness through case studies of adaptive bitrate algorithms from networking and scheduling algorithms from systems domains. Our experimental results reveal that \texttt{Crucible} not only quantifies the tunable space across different algorithms but also enables the optimization of existing algorithms by using potential as a new optimization direction, ultimately enhancing performance outcomes. Our code is available in the supplementary materials.



Paperid:2868
Authors:Wei Chow, Yuan Gao, Linfeng Li, XIAN WANG, Qi Xu, Hang Song, Lingdong Kong, Ran Zhou, Yi Zeng, Yidong Cai, Botian Jiang, Shilin Xu, Jiajunzhang, Minghui Qiu, Xiangtai Li, Tianshu Yang, Siliang Tang, Juncheng Li
Abstract:
Semantic retrieval is crucial for modern applications yet remains underexplored in current research. Existing datasets are limited to single languages, single images, or singular retrieval conditions, often failing to fully exploit the expressive capacity of visual information as evidenced by maintained performance when images are replaced with captions. However, practical retrieval scenarios frequently involve interleaved multi-condition queries with multiple images.Hence, this paper introduces MERIT, the first multilingual dataset for interleaved multi-condition semantic retrieval, comprising 320,000 queries with 135,000 products in 5 languages, covering 7 distinct product categories.Extensive experiments on MERIT identify existing models's critical limitation: focusing solely on global semantic information while neglecting specific conditional elements in queries.Consequently, we propose Coral, a novel fine-tuning framework that adapts pre-trained MLLMs by integrating embedding reconstruction to preserve fine-grained conditional elements and contrastive learning to extract comprehensive global semantics.Experiments demonstrate that Coral achieves a 45.9% performance improvement over conventional approaches on MERIT, with strong generalization capabilities validated across 8 established retrieval benchmarks. Collectively, our contributions—a novel dataset, identification of critical limitations in existing approaches, and an innovative fine-tuning framework—establish a foundation for future research in interleaved multi-condition semantic retrieval. Anonymous Project Page:https://anoy1314.github.io.



Paperid:2869
Authors:Linlian Jiang, Rui Ma, Li Gu, Ziqiang Wang, Xinxin Zuo, Yang Wang
Abstract:
Abstract:Point cloud completion is essential for robust 3D perception in safety-critical applications such as robotics and augmented reality. However, existing models perform static inference and rely heavily on inductive biases learned during training, limiting their ability to adapt to novel structural patterns and sensor-induced distortions at test time.To address this limitation, we propose PointMAC, a meta-learned framework for robust test-time adaptation in point cloud completion. It enables sample-specific refinement without requiring additional supervision.Our method optimizes the completion model under two self-supervised auxiliary objectives that simulate structural and sensor-level incompleteness.A meta-auxiliary learning strategy based on Model-Agnostic Meta-Learning (MAML) ensures that adaptation driven by auxiliary objectives is consistently aligned with the primary completion task.During inference, we adapt the shared encoder on-the-fly by optimizing auxiliary losses, with the decoder kept fixed. To further stabilize adaptation, we introduce Adaptive $\lambda$-Calibration, a meta-learned mechanism for balancing gradients between primary and auxiliary objectives. Extensive experiments on synthetic, simulated, and real-world datasets demonstrate that PointMAC achieves state-of-the-art results by refining each sample individually to produce high-quality completions. To the best of our knowledge, this is the first work to apply meta-auxiliary test-time adaptation to point cloud completion.



Paperid:2870
Authors:Jiahui Wang, Haiyue Zhu, Haoren Guo, Abdullah Al Mamun, Cheng Xiang, Tong Heng LEE
Abstract:
Abstract:Recent 6D pose estimation methods demonstrate notable performance but still face some practical limitations. For instance, many of them rely heavily on sensor depth, which may fail with challenging surface conditions, such as transparent or highly reflective materials. In the meantime, RGB-based solutions provide less robust matching performance in low-light and texture-less scenes due to the lack of geometry information. Motivated by these, we propose **SingRef6D**, a lightweight pipeline requiring only a **single RGB** image as a reference, eliminating the need for costly depth sensors, multi-view image acquisition, or training view synthesis models and neural fields. This enables SingRef6D to remain robust and capable even under resource-limited settings where depth or dense templates are unavailable. Our framework incorporates two key innovations. First, we propose a token-scaler-based fine-tuning mechanism with a novel optimization loss on top of Depth-Anything v2 to enhance its ability to predict accurate depth, even for challenging surfaces. Our results show a 14.41% improvement (in $\delta_{1.05}$) on REAL275 depth prediction compared to Depth-Anything v2 (with fine-tuned head). Second, benefiting from depth availability, we introduce a depth-aware matching process that effectively integrates spatial relationships within LoFTR, enabling our system to handle matching for challenging materials and lighting conditions. Evaluations of pose estimation on the REAL275, ClearPose, and Toyota-Light datasets show that our approach surpasses state-of-the-art methods, achieving a 6.1% improvement in average recall.



Paperid:2871
Authors:Ankan Bhunia, Changjian Li, Hakan Bilen
Abstract:
This paper presents a novel problem, interactive anomaly detection (AD) for articulated objects, and introduces a tailored solution that detects functional anomalies by integrating vision, interaction, and anticipation. Unlike traditional AD methods that rely on passive visual observations, our approach actively manipulates objects to reveal anomalies that would otherwise remain hidden. Our method learns to generate a sequence of actions to interact exclusively with normal objects and to anticipate the resulting normal motion. During inference, the model applies predicted actions to the object and compares the observed motion with the anticipated motion to detect anomalies. Additionally, we introduce a new benchmark, PartNet-IAD, for interactive AD, which includes articulated objects with realistic functional anomalies. Experiments show strong generalization to detect anomalies in both seen and unseen object categories. Code and dataset will be released.



Paperid:2872
Authors:Christy Li, Josep Lopez Camuñas, Jake Touchet, Jacob Andreas, Agata Lapedriza, Antonio Torralba, Tamar Shaham
Abstract:
When a vision model performs image recognition, which visual attributes drive its predictions? Detecting unintended use of specific visual features is critical for ensuring model robustness, preventing overfitting, and avoiding spurious correlations. We introduce an automated framework for detecting these dependencies in trained vision models. At the core of our method is a self-reflective agent that systematically generates and tests hypotheses about the unintended visual attributes that a model may rely on. This process is iterative: the agent refines its hypotheses based on experimental outcomes and uses a self-evaluation protocol to assess whether its findings accurately explain model behavior. If inconsistencies are detected, the agent self-reflects over its findings and triggers a new cycle of experimentation. We evaluate our approach on a novel benchmark of 130 models designed to exhibit diverse visual attribute dependencies across 18 categories. Our results show that the agent's performance consistently improves with self-reflection, with a significant performance increase over non-reflective baselines. We further demonstrate that the agent identifies real-world visual attribute dependencies in state-of-the-art models, including CLIP's vision encoder and the YOLOv8 object detector.



Paperid:2873
Authors:Yuantian Shao, Yuanteng Chen, Peisong Wang, Jianlin Yu, Jing Lin, yiwu yao, Zhihui Wei, Jian Cheng
Abstract:
Abstract:Quantization plays a crucial role in accelerating the inference of large-scale models, and rotational matrices have been shown to effectively improve quantization performance by smoothing outliers. However, end-to-end fine-tuning of rotational optimization algorithms incurs high computational costs and is prone to overfitting. To address this challenge, we propose an efficient distribution-aware rotational calibration method, DartQuant, which reduces the complexity of rotational optimization by constraining the distribution of the activations after rotation. This approach also effectively reduces reliance on task-specific losses, thereby mitigating the risk of overfitting. Additionally, we introduce the QR-Orth optimization scheme, which replaces expensive alternating optimization with a more efficient solution. In a variety of model quantization experiments, DartQuant demonstrates superior performance. Compared to existing methods, it achieves 47$\times$ acceleration and 10$\times$ memory savings for rotational optimization on a 70B model. Furthermore, it is the first to successfully complete rotational calibration for a 70B model on a single 3090 GPU, making quantization of large language models feasible in resource-constrained environments.



Paperid:2874
Authors:Mehdi Makni, Xiang Meng, Rahul Mazumder
Abstract:
Abstract:Sparse plus Low-Rank $(\mathbf{S} + \mathbf{L}\mathbf{R})$ decomposition of Large Language Models (LLMs) has emerged as a promising direction in $\textit{model compression}$, aiming to decompose pre-trained model weights into a sum of sparse and low-rank matrices $\mathbf{W} \approx \mathbf{S} + \mathbf{LR}$. Despite recent progress, existing methods often suffer from substantial performance degradation compared to dense models. In this work, we introduce $\texttt{3BASiL-TM}$, an efficient one-shot post-training method for $(\mathbf{S} + \mathbf{L}\mathbf{R})$ decomposition of LLMs that addresses this gap. Our approach first introduces a novel 3-Block Alternating Direction Method of Multipliers (ADMM) method, termed $\texttt{3BASiL}$, to minimize the layer-wise reconstruction error with convergence guarantees. We then design a transformer-matching ($\texttt{TM}$) refinement step that jointly optimizes the sparse and low-rank components across transformer layers. This step minimizes a novel memory-efficient loss that aligns outputs at the transformer level.Notably, the $\texttt{TM}$ procedure is universal as it can enhance any $(\mathbf{S} + \mathbf{L}\mathbf{R})$ decomposition, including pure sparsity. Our numerical experiments show that $\texttt{3BASiL-TM}$ reduces the WikiText2 perplexity gap to dense LLaMA-8B model by over 30% under a (2:4 Sparse + 64 LR) configuration, compared to prior methods. Moreover, our method achieves over 2.5x faster compression runtime on an A100 GPU compared to SOTA $(\mathbf{S} + \mathbf{L}\mathbf{R})$ method.



Paperid:2875
Authors:Wei Duan, Jie Lu, Junyu Xuan
Abstract:
In networked multi-agent reinforcement learning (Networked-MARL), decentralized agents must act under local observability and constrained communication over fixed physical graphs. Existing methods often assume static neighborhoods, limiting adaptability to dynamic or heterogeneous environments. While centralized frameworks can learn dynamic graphs, their reliance on global state access and centralized infrastructure is impractical in real-world decentralized systems. We propose a stochastic graph-based policy for Networked-MARL, where each agent conditions its decision on a sampled subgraph over its local physical neighborhood. Building on this formulation, we introduce \textbf{BayesG}, a decentralized actor–critic framework that learns sparse, context-aware interaction structures via Bayesian variational inference. Each agent operates over an ego-graph and samples a latent communication mask to guide message passing and policy computation. The variational distribution is trained end-to-end alongside the policy using an evidence lower bound (ELBO) objective, enabling agents to jointly learn both interaction topology and decision-making strategies.BayesG outperforms strong MARL baselines on large-scale traffic control tasks with up to 167 agents, demonstrating superior scalability, efficiency, and performance.



Paperid:2876
Authors:Linnan Tu, Lingwei Meng, Zongyi Li, Hefei Ling, Shijuan Huang
Abstract:
Existing efforts in motion synthesis typically utilize either generative transformers with discrete representations or diffusion models with continuous representations.However, the discretization process in generative transformers can introduce motion errors, while the sampling process in diffusion models tends to be slow.In this paper, we propose a novel text-to-motion synthesis method GMMotion that combines a continuous motion representation with an autoregressive model, using the Gaussian mixture model (GMM) to represent the conditional probability distribution.Unlike autoregressive approaches relying on residual vector quantization, our model employs continuous motion representations derived from the VAE's latent space. This choice streamlines both the training and the inference processes.Specifically, we utilize a causal transformer to learn the distributions of continuous motion representations, which are modeled with a learnable Gaussian mixture model. Extensive experiments demonstrate that our model surpasses existing state-of-the-art models in the motion synthesis task.



Paperid:2877
Authors:Zhiwei Lin, Yongtao Wang
Abstract:
Current perception models have achieved remarkable success by leveraging large-scale labeled datasets, but still face challenges in open-world environments with novel objects. To address this limitation, researchers introduce open-set perception models to detect or segment arbitrary test-time user-input categories. However, open-set models rely on human involvement to provide predefined object categories as input during inference. More recently, researchers have framed a more realistic and challenging task known as open-ended perception that aims to discover unseen objects without requiring any category-level input from humans at inference time. Nevertheless, open-ended models suffer from low performance compared to open-set models. In this paper, we present VL-SAM-V2, an open-world object detection framework that is capable of discovering unseen objects while achieving favorable performance. To achieve this, we combine queries from open-set and open-ended models and propose a general and specific query fusion module to allow different queries to interact. By adjusting queries from open-set models, we enable VL-SAM-V2 to be evaluated in the open-set or open-ended mode. In addition, to learn more diverse queries, we introduce ranked learnable queries to match queries with proposals from open-ended models by sorting. Moreover, we design a denoising point training strategy to facilitate the training process. Experimental results on LVIS show that our method surpasses the previous open-set and open-ended methods, especially on rare objects.



Paperid:2878
Authors:Shmuel Berman, Jia Deng
Abstract:
Visual Language Models (VLMs) excel at complex visual tasks such as VQA and chart understanding, yet recent work suggests they struggle with simple perceptual tests. We present an evaluation that tests vision-language models’ capacity for \emph{nonlocal visual reasoning}- reasoning that requires chaining evidence collected from multiple, possibly distant, regions of an image. We isolate three distinct forms of non‑local vision: \emph{comparative perception}, which demands holding two images in working memory and comparing them; \emph{saccadic search}, which requires making discrete, evidence‑driven jumps to locate successive targets; and \emph{smooth visual search}, which involves searching smoothly along a continuous contour. Flagship models (e.g., Gemini 2.5 Pro, Claude Vision 3.7, GPT‑o4-mini), even those that perform well on prior primitive‑vision benchmarks, fail these tests and barely exceed random accuracy on two variants of our tasks that are trivial for humans. Our structured evaluation suite allows us if VLMs can perform similar visual algorithms to humans. Our findings show that despite gains in raw visual acuity, current models lack core visual reasoning capabilities.



Paperid:2879
Authors:Mustafa Hajij, Lennart Bastian, Sarah Osentoski, Hardik Kabaria, John Davenport, Dawood, Balaji Cherukuri, Joseph Kocheemoolayil, Nastaran Shahmansouri, Adrian Lew, Theodore Papamarkou, Tolga Birdal
Abstract:
We introduceCopresheaf Topological Neural Networks(CTNNs), a powerful and unifying framework that encapsulates a wide spectrum of deep learning architectures, designed to operate on structured data: including images, point clouds, graphs, meshes, and topological manifolds. While deep learning has profoundly impacted domains ranging from digital assistants to autonomous systems, the principled design of neural architectures tailored to specific tasks and data types remains one of the field’s most persistent open challenges. CTNNs address this gap by grounding model design in the language of copresheaves, a concept from algebraic topology that generalizes and subsumes most practical deep learning models in use today. This abstract yet constructive formulation yields a rich design space from which theoretically sound and practically effective solutions can be derived to tackle core challenges in representation learning: long-range dependencies, oversmoothing, heterophily, and non-Euclidean domains. Our empirical results on structured data benchmarks demonstrate that CTNNs consistently outperform conventional baselines, particularly in tasks requiring hierarchical or localized sensitivity. These results underscore CTNNs as a principled, multi-scale foundation for the next generation of deep learning architectures.



Paperid:2880
Authors:Yifan WANG, Hongfeng Ai, ruiqi li, Maowei Jiang, Quangao Liu, Jiahua Dong, ruiyuan kang, Alan Liang, Zihang Wang, ruikai liu, Cheng Jiang, Chenzhong Li
Abstract:
Medical time-series data play a vital role in disease diagnosis but suffer from limited labeled samples and single-center bias, which hinder model generalization and lead to overfitting. To address these challenges, we propose DAAC (Discrepancy-Aware Adaptive Contrastive learning), a learnable multi-view contrastive framework that integrates external normal samples and enhances feature learning through adaptive contrastive strategies. DAAC consists of two key modules: (1) a Discrepancy Estimator, built upon a GAN-enhanced encoder-decoder architecture, captures the distribution of normal data and computes reconstruction errors as indicators of abnormality. These discrepancy features augment the target dataset to mitigate overfitting. (2) an Adaptive Contrastive Learner uses multi-head attention to extract discriminative representations by contrasting embeddings across multiple views and data granularities (subject, trial, epoch, and temporal levels), eliminating the need for handcrafted positive-negative sample pairs. Extensive experiments on three clinical datasets—covering Alzheimer’s disease, Parkinson’s disease, and myocardial infarction—demonstrate that DAAC significantly outperforms existing methods, even when only 10\% of labeled data is available, showing strong generalization and diagnostic performance.



Paperid:2881
Authors:Shane Bergsma, Nolan Dey, Gurpreet Gosal, Gavia Gray, Daria Soboleva, Joel Hestness
Abstract:
Efficient LLM pre-training requires well-tuned hyperparameters (HPs), including learning rate η and weight decay λ. We study scaling laws for HPs: formulas for how to scale HPs as we scale model size N, dataset size D, and batch size B. Recent work suggests the AdamW timescale, B/(ηλD), should remain constant across training settings, and we verify the implication that optimal λ scales linearly with B, for a fixed N,D. However, as N,D scale, we show the optimal timescale obeys a precise power law in the tokens-per-parameter ratio, D/N. This law thus provides a method to accurately predict λopt in advance of large-scale training. We also study scaling laws for optimal batch size Bopt (the B enabling lowest loss at a given N,D) and critical batch size Bcrit (the B beyond which further data parallelism becomes ineffective). In contrast with prior work, we find both Bopt and Bcrit scale as power laws in D, independent of model size, N. Finally, we analyze how these findings inform the real-world selection of Pareto-optimal N and D under dual training time and compute objectives.



Paperid:2882
Authors:Baran Hashemi, Kurt Pasque, Chris Teska, Ruriko Yoshida
Abstract:
Dynamic programming (DP) algorithms for combinatorial optimization problems work with taking max, min, and classical addition in their recursion algorithms and their value functions are associated with convex polyhedra in the max‑plus semiring. Existing Neural Algorithmic Reasoning models, however, rely on softmax‑normalized dot‑product attention where the smooth exponential weighting blurs these sharp polyhedral structures and collapses when evaluated on out‑of‑distribution (OOD) settings. We introduce Tropical attention, an attention function that operates originally in the max-plus semiring of tropical geometry. We prove that Tropical attention can approximate tropical circuits of DP-type combinatorial algorithms. We then propose that using Tropical transformers enhances empirical OOD performance in both length generalization and value generalization, on canonical algorithmic reasoning tasks, surpassing softmax baselines while remaining stable under adversarial perturbations. Our results demonstrate that Tropical attention restores the sharp, scale‑invariant reasoning absent from softmax.



Paperid:2883
Authors:Jiaqi Wang, Xinyi Wu, James Cheng, Yifei Wang
Abstract:
Deep graph neural networks (GNNs) often suffer from oversmoothing, where node representations become overly homogeneous with increasing depth. While techniques like normalization, residual connections, and edge dropout have been proposed to mitigate oversmoothing, they are typically developed independently, with limited theoretical understanding of their underlying mechanisms. In this work, we present a unified theoretical perspective based on the framework of signed graphs, showing that many existing strategies implicitly introduce negative edges that alter message-passing to resist oversmoothing. However, we show that merely adding negative edges in an unstructured manner is insufficient—the asymptotic behavior of signed propagation depends critically on the strength and organization of positive and negative edges. To address this limitation, we leverage the theory of structural balance, which promotes stable, cluster-preserving dynamics by connecting similar nodes with positive edges and dissimilar ones with negative edges. We propose Structural Balanced Propagation (SBP), a plug-and-play method that assigns signed edges based on either labels or feature similarity to explicitly enhance structural balance in the constructed signed graphs. Experiments on nine benchmarks across both homophilic and heterophilic settings demonstrate that SBP consistently improves classification accuracy and mitigates oversmoothing, even at depths of up to 300 layers. Our results provide a principled explanation for prior oversmoothing remedies and introduce a new direction for signed message-passing design in deep GNNs.



Paperid:2884
Authors:Chanhyeong Yang, Taehoon Song, Jihwan Park, Hyunwoo J. Kim
Abstract:
Zero-shot Human-Object Interaction (HOI) detection aims to localize humans and objects in an image and recognize their interaction, even when specific verb-object pairs are unseen during training. Recent works have shown promising results using prompt learning with pretrained vision-language models such as CLIP, which align natural language prompts with visual features in a shared embedding space.However, existing approaches still fail to handle thevisual complexity of interaction—including (1)intra-class visual diversity, where instances of the same verb appear in diverse poses and contexts, and (2)inter-class visual entanglement, where distinct verbs yield visually similar patterns. To address these challenges, we proposeVDRP, a framework forVisual Diversity and Region-aware Prompt learning. First, we introduce a visual diversity-aware prompt learning strategy that injects group-wise visual variance into the context embedding. We further apply Gaussian perturbation to encourage the prompt to capture diverse visual variations of a verb.Second, we retrieve region-specific concepts from the human, object, and union regions. These are used to augment the diversity-aware prompt embeddings, yielding region-aware prompts that improve verb-level discrimination.Experiments on the HICO-DET benchmark demonstrate that our method achieves state-of-the-art performance under four zero-shot evaluation settings, effectively addressing both intra-class diversity and inter-class visual entanglement.



Paperid:2885
Authors:Divyansh Pareek, Sewoong Oh, Simon Du
Abstract:
Abstract:The success of modern multimodal representation learning relies on internet-scale datasets. Due to the low quality of a large fraction of raw web data, curation of data has become a critical step in the training pipeline. Filtering using a trained model (i.e., teacher-based filtering) has emerged as a successful solution, leveraging a pre-trained model to compute quality scores. To explain the empirical success of teacher-based filtering, we characterize the performance of filtered contrastive learning under the standard bimodal data generation model. Denoting $\eta$ as the fraction of data with correctly matched modalities among $n$ paired samples, we utilize a linear contrastive learning setup to show a provable benefit of data filtering: $(i)$ the error without filtering is upper and lower bounded by $\frac{1}{\eta \sqrt{n}}$, and $(ii)$ the error with teacher-based filtering is upper bounded by $\frac{1}{\sqrt{\eta n}}$ in the large $\eta$ regime, and by $\frac{1}{\sqrt{n}}$ in the small $\eta$ regime.



Paperid:2886
Authors:Xinran Qin, Zhixin Wang, Fan Li, Haoyu Chen, Renjing Pei, Wenbo Li, Xiaochun Cao
Abstract:
Recent advances in diffusion models have substantially improved text-driven image editing. However, existing frameworks based on discrete textual tokens struggle to support continuous control over camera parameters and smooth transitions in visual effects. These limitations hinder their applications to realistic, camera-aware, and fine-grained editing tasks. In this paper, we present CamEdit, a diffusion-based framework for photorealistic image editing that enables continuous and semantically meaningful manipulation of common camera parameters such as aperture and shutter speed. CamEdit incorporates a continuous parameter prompting mechanism and a parameter-aware modulation module that guides the model in smoothly adjusting focal plane, aperture, and shutter speed, reflecting the effects of varying camera settings within the diffusion process. To support supervised learning in this setting, we introduce CamEdit50K, a dataset specifically designed for photorealistic image editing with continuous camera parameter settings. It contains over 50k image pairs combining real and synthetic data with dense camera parameter variations across diverse scenes. Extensive experiments demonstrate that CamEdit enables flexible, consistent, and high-fidelity image editing, achieving state-of-the-art performance in camera-aware visual manipulation and fine-grained photographic control.



Paperid:2887
Authors:Yuchen Zhou, Jiamin Wu, Zichen Ren, Zhouheng Yao, Weiheng Lu, Kunyu Peng, Qihao Zheng, Chunfeng Song, Wanli Ouyang, Chao Gou
Abstract:
Understanding and decoding human brain activity from electroencephalography (EEG) signals is a fundamental problem in neuroscience and artificial intelligence, with applications ranging from cognition and emotion recognition to clinical diagnosis and brain–computer interfaces. While recent EEG foundation models have made progress in generalized brain decoding by leveraging unified architectures and large-scale pretraining, they inherit a scale-agnostic dense modeling paradigm from NLP and vision. This design overlooks an intrinsic property of neural activity—cross-scale spatiotemporal structure. Different EEG task patterns span a broad range of temporal and spatial scales, from brief neural activations to slow-varying rhythms, and from localized cortical activations to large-scale distributed interactions. Ignoring this diversity may lead to suboptimal representations and weakened generalization ability. To address these limitations, we propose CSBrain, a Cross-scale Spatiotemporal Brain foundation model for generalized EEG decoding. CSBrain introduces two key components: (i) Cross-scale Spatiotemporal Tokenization (CST), which aggregates multi-scale features within localized temporal windows and anatomical brain regions into compact scale-aware token representations; and (ii) Structured Sparse Attention (SSA), which models cross-window and cross-region dependencies for diverse decoding tasks, further enriching scale diversities while eliminating the spurious dependencies. CST and SSA are alternately stacked to progressively integrate cross-scale spatiotemporal dependencies. Extensive experiments across 11 representative EEG tasks and 16 datasets demonstrate that CSBrain consistently outperforms both task-specific models and strong foundation baselines. These results establish cross-scale modeling as a key inductive bias for generalized EEG decoding and highlight CSBrain as a robust backbone for future brain–AI research. The code and model will be released.



Paperid:2888
Authors:Pengteng Li, Pinhao Song, Wuyang Li, Huizai Yao, Weiyu Guo, Yijie Xu, Dugang Liu, Hui Xiong
Abstract:
We introduce See&Trek, the first training-free prompting framework tailored to enhance the spatial understanding of Multimodal Large Language Models (MLLMs) under vision-only constraints. While prior efforts have incorporated modalities like depth or point clouds to improve spatial reasoning, purely visual-spatial understanding remains underexplored. See&Trek addresses this gap by focusing on two core principles: increasing visual diversity and motion reconstruction. For visual diversity, we conduct Maximum Semantic Richness Sampling, which employs an off-the-shell perception model to extract semantically rich keyframes that capture scene structure. For motion reconstruction, we simulate visual trajectories and encode relative spatial positions into keyframes to preserve both spatial relations and temporal coherence. Our method is training&GPU-free, requiring only a single forward pass, and can be seamlessly integrated into existing MLLMs. Extensive experiments on the VSI-Bench and STI-Bench show that See&Trek consistently boosts various MLLMs performance across diverse spatial reasoning tasks with the most +3.5% improvement, offering a promising path toward stronger spatial intelligence.



Paperid:2889
Authors:Sheida RahnamaiKordasiabi, Damian Nogare, Florian Jug
Abstract:
Abstract:Semantic segmentation of electron microscopy (EM) images of biological samples remains a challenge in the life sciences.EM data captures details of biological structures, sometimes with such complexity that even human observers can find it overwhelming.Here we introduce $\epsilon$-Seg, a method based on hierarchical variational autoencoders (HVAEs), employing center-region masking, sparse label contrastive learning (CL), a Gaussian Mixture Model (GMM) prior, and clustering-free label prediction.Center-region masking and the inpainting loss encourage the model to learn robust and representative embeddings to distinguish the desired classes, even if training labels are sparse ($0.05$\% of the total image data or less).Additionally, we propose an entropy-based loss that can improve segmentation quality when fewer training labels are available (i.e. on $0.0025$\% of the data).For optimal performance, we employ CL and a GMM prior to shape the latent space of the HVAE such that encoded input patches tend to cluster w.r.t. the semantic classes we wish to distinguish. Finally, instead of clustering latent embeddings for semantic segmentation, we propose a semantic segmentation head composed of MLP and FiLM layers to directly predict class labels from latent embeddings.We show empirical results of $\epsilon$-Seg and baseline methods on $2$ dense EM datasets of biological tissues and demonstrate the applicability of our method also on fluorescence microscopy data. Our results show that $\epsilon$-Seg is capable of achieving competitive semi-supervised segmentation results on complex biological image data, even if only limited amounts of training labels are available.



Paperid:2890
Authors:Junhyuk So, Chiwoong Lee, Shinyoung Lee, Jungseul Ok, Eunhyeok Park
Abstract:
Generative Behavior Cloning (GBC) is a simple yet effective framework for robot learning, particularly in multi-task settings. Recent GBC methods often employ diffusion policies with open-loop (OL) control, where actions are generated via a diffusion process and executed in multi-step chunks without replanning. While this approach has demonstrated strong success rates and generalization, its inherent stochasticity can result in erroneous action sampling, occasionally leading to unexpected task failures. Moreover, OL control suffers from delayed responses, which can degrade performance in noisy or dynamic environments. To address these limitations, we propose two novel techniques to enhance the consistency and reactivity of diffusion policies: (1) self-guidance, which improves action fidelity by leveraging past observations and implicitly promoting future-aware behavior; and (2) adaptive chunking, which selectively updates action sequences when the benefits of reactivity outweigh the need for temporal consistency. Extensive experiments show that our approach substantially improves GBC performance across a wide range of simulated and real-world robotic manipulation tasks.



Paperid:2891
Authors:Taejong Joo, Diego Klabjan
Abstract:
Transformers have demonstrated remarkable in-context learning (ICL) capabilities, adapting to new tasks by simply conditioning on demonstrations without parameter updates. Compelling empirical and theoretical evidence suggests that ICL, as a general-purpose learner, could outperform task-specific models. However, it remains unclear to what extent the transformers optimally learn in-context compared to principled learning algorithms. To bridge this gap, we introduce a new framework for quantifying optimality of ICL as a learning algorithm in stylized settings. Our findings reveal a striking dichotomy: while ICL initially matches the efficiency of a Bayes optimal estimator, its efficiency significantly deteriorates in long context. Through an information-theoretic analysis, we show that the diminishing efficiency is inherent to ICL. These results clarify the trade-offs in adopting ICL as a universal problem solver, motivating a new generation of on-the-fly adaptive methods without the diminishing efficiency.



Paperid:2892
Authors:Yiming Zhong, Yumeng Liu, Chuyang Xiao, Zemin Yang, Youzhuo Wang, Yufei Zhu, Ye Shi, Yujing Sun, Xinge ZHU, Yuexin Ma
Abstract:
Learning effective visuomotor policies for robotic manipulation is challenging, as it requires generating precise actions while maintaining computational efficiency. Existing methods remain unsatisfactory due to inherent limitations in the essential action representation and the basic network architectures. We observe that representing actions in the frequency domain captures the structured nature of motion more effectively: low-frequency components reflect global movement patterns, while high-frequency components encode fine local details. Additionally, robotic manipulation tasks of varying complexity demand different levels of modeling precision across these frequency bands. Motivated by this, we propose a novel paradigm for visuomotor policy learning that progressively models hierarchical frequency components. To further enhance precision, we introduce continuous latent representations that maintain smoothness and continuity in the action space. Extensive experiments across diverse 2D and 3D robotic manipulation benchmarks demonstrate that our approach outperforms existing methods in both accuracy and efficiency, showcasing the potential of a frequency-domain autoregressive framework with continuous tokens for generalized robotic manipulation.



Paperid:2893
Authors:Peizhi Niu, Yu-Hsiang Wang, Vishal Rana, Chetan Rupakheti, Abhishek Pandey, Olgica Milenkovic
Abstract:
Abstract:We introduce a new graph diffusion model for small drug molecule generation which simultaneously offers a $10$-fold reduction in the number of diffusion steps when compared to other existing methods, preservation of small molecule graph motifs via ring compression, and a $3$% improvement in SMILES validity over the DiGress model across all real-world molecule benchmarking datasets. Furthermore, our approach outperforms the state-of-the-art DeFoG method with respect to motif-conservation by roughly $4$%, as evidenced by high ChEMBL-likeness, QED and a newly introduced shingles distance score. The key ideas behind our approach are to use a combination of deterministic and random subgraph perturbations, so that the node and edge noise schedules are codependent; to modify the loss function of the training process in order to exploit the deterministic component of the schedule; and, to ``compress'' a collection of highly relevant carbon ring structures into supernodes in a way that allows for simple subsequent integration into the molecule scaffold.



Paperid:2894
Authors:Yuqi Jia, Minghong Fang, Hongbin Liu, Jinghuai Zhang, Neil Gong
Abstract:
Poisoning attacks compromise the training phase of federated learning (FL) such that the learned global model misclassifies attacker-chosen inputs called target inputs. Existing defenses mainly focus on protecting the training phase of FL such that the learnt global model is poison free. However, these defenses often achieve limited effectiveness when the clients' local training data is highly non-iid or the number of malicious clients is large, as confirmed in our experiments. In this work, we propose FLForensics, the first poison-forensics method for FL. FLForensics complements existing training-phase defenses. In particular, when training-phase defenses fail and a poisoned global model is deployed, FLForensics aims to trace back the malicious clients that performed the poisoning attack after a misclassified target input is identified. We theoretically show that FLForensics can accurately distinguish between benign and malicious clients under a formal definition of poisoning attack. Moreover, we empirically show the effectiveness of FLForensics at tracing back both existing and adaptive poisoning attacks on five benchmark datasets.



Paperid:2895
Authors:Hao Li, Bowen Deng, Chang Xu, ZhiYuan Feng, Viktor Schlegel, Yu-Hao Huang, Yizheng Sun, Jingyuan Sun, Kailai Yang, Yiyao Yu, Jiang Bian
Abstract:
A unified foundation model for medical time series—pretrained on open access and ethically reviewed medical corpora—offers the potential to reduce annotation burdens, minimize model customization, and enable robust transfer across clinical institutions, modalities, and tasks, particularly in data-scarce or privacy-constrained environments. However, existing time series foundation models struggle to handle medical time series data due to its inherent challenges, including irregular intervals, heterogeneous sampling rates, and frequent missingness. To address these challenges, we introduce MIRA, a unified foundation model specifically designed for medical time series forecasting. MIRA incorporates a Continuous-Time Rotary Positional Encoding that enables fine-grained modeling of variable time intervals, a frequency-specific mixture-of-experts layer that routes computation across latent frequency regimes to further promote temporal specialization, and a Continuous Dynamics Extrapolation Block based on Neural ODE that models the continuous trajectory of latent states, enabling accurate forecasting at arbitrary target timestamps. Pretrained on a large-scale and diverse medical corpus comprising over 454 billion time points collect from publicly available datasets, MIRA achieving reductions in forecasting errors by an average of 10% and 7% in out-of-distribution and in-distribution scenarios, respectively. We also introduce a comprehensive benchmark spanning multiple downstream clinical tasks, establishing a foundation for future research in medical time series modeling.



Paperid:2896
Authors:Hongyang He, Xinyuan Song, Yangfan He, Zeyu Zhang, Yanshu Li, Haochen You, Lifan Sun, Wenqiao Zhang
Abstract:
We introduce TRiCo, a novel triadic game-theoretic co-training framework that rethinks the structure of semi-supervised learning by incorporating a teacher, two students, and an adversarial generator into a unified training paradigm. Unlike existing co-training or teacher-student approaches, TRiCo formulates SSL as a structured interaction among three roles: (i) two student classifiers trained on frozen, complementary representations, (ii) a meta-learned teacher that adaptively regulates pseudo-label selection and loss balancing via validation-based feedback, and (iii) a non-parametric generator that perturbs embeddings to uncover decision boundary weaknesses. Pseudo-labels are selected based on mutual information rather than confidence, providing a more robust measure of epistemic uncertainty. This triadic interaction is formalized as a Stackelberg game, where the teacher leads strategy optimization and students follow under adversarial perturbations. By addressing key limitations in existing SSL frameworks—such as static view interactions, unreliable pseudo-labels, and lack of hard sample modeling—TRiCo provides a principled and generalizable solution. Extensive experiments on CIFAR-10, SVHN, STL-10, and ImageNet demonstrate that TRiCo consistently achieves state-of-the-art performance in low-label regimes, while remaining architecture-agnostic and compatible with frozen vision backbones.



Paperid:2897
Authors:Diankun Wu, Fangfu Liu, Yi-Hsin Hung, Yueqi Duan
Abstract:
Recent advancements in Multimodal Large Language Models (MLLMs) have significantly enhanced performance on 2D visual tasks. However, improving their spatial intelligence remains a challenge. Existing 3D MLLMs always rely on additional 3D or 2.5D data to incorporate spatial awareness, restricting their utility in scenarios with only 2D inputs, such as images or videos. In this paper, we present \emph{Spatial-MLLM}, a novel framework for visual-based spatial reasoning from purely 2D observations. Unlike conventional video MLLMs, which rely on CLIP-based visual encoders optimized for semantic understanding, our key insight is to unleash the strong structure prior from the feed-forward visual geometry foundation model. Specifically, we propose a dual-encoder architecture: a pretrained 2D visual encoder to extract semantic features, and a spatial encoder—initialized from the backbone of the visual geometry model—to extract 3D structure feature. A connector then integrates both features into unified visual tokens for enhanced spatial understanding. Furthermore, we propose a space-aware frame sampling strategy at inference time, which selects the spatially informative frames of a video sequence, ensuring that even under limited token length, the model focuses on frames critical for spatial reasoning. Beyond architecture improvements, we construct Spatial-MLLM-120k dataset and train the model using supervised-fining and GRPO on it. Extensive experiments on various real-world datasets demonstrate that our spatial-MLLM achieves state-of-the-art performance in a wide range of visual-based spatial understanding and reasoning tasks.



Paperid:2898
Authors:Jinyang Li, Xiaolong Li, Ge Qu, Per Jacobsson, Bowen Qin, Binyuan Hui, Shuzheng Si, Nan Huo, Xiaohan Xu, Yue Zhang, Ziwei Tang, Yuanshuai Li, Florensia Widjaja, Xintong Zhu, Feige Zhou, Yongfeng Huang, Yannis Papakonstantinou, Fatma Ozcan, Ma Chenhao, Reynold Cheng
Abstract:
Resolution of complex SQL issues persists as a significant bottleneck in real-world database applications. Current Large Language Models (LLMs), while adept at text-to-SQL translation, have not been rigorously evaluated on the more challenging task of debugging SQL issues. In order to address this gap, we introduceBIRD-CRITIC, a new SQL-issue-debugging benchmark comprising 530 carefully curated PostgreSQL tasks (BIRD-CRITIC-PG) and 570 multi-dialect tasks (BIRD-CRITIC-Multi), which are distilled from authentic user issues and replayed within new environments to facilitate rigorous and contamination-free evaluation. Baseline evaluations on BIRD-CRITIC underscore the task’s complexity, with the leading reasoning modelO3-Miniachieving only 38.87% success onBIRD-CRITIC-PGand 33.33% onBIRD-CRITIC-Multi. Meanwhile, advancing open-source models for database tasks is crucial, empowering local development while safeguarding data privacy. Therefore, we presentSix-Gym(Sql-fIX-Gym), a training environment for elevating the capabilities of open-source models specifically for SQL-issue debugging. This environment leverages theSQL-Rewindstrategy, which automatically generates executable issue-solution datasets by reverse-engineering issues from verified SQLs. However, popular trajectory-based fine-tuning methods do not exploit substantial supervisory signal. We further proposef-Plan Boosting, which extracts high-level debugging plans automatically from SQL solutions, enabling teacher LLMs to harvest and produce 73.7% more successful trajectories for training. We integrate these components into an open-source agent,BIRD-Fixer. Based on a 14B, BIRD-Fixer raises its success rate to 38.11% onBIRD-CRITIC-PGand 30.70% onBIRD-CRITIC-Multi, surpassing many leading proprietary models such as Claude-3.7-Sonnet and GPT-4.1, marking a significant step toward democratizing sophisticated SQL-debugging capabilities for both research and industry.



Paperid:2899
Authors:Woojun Kim, Katia Sycara
Abstract:
Multi-agent reinforcement learning in mixed-motive settings presents a fundamental challenge: agents must balance individual interests with collective goals, which are neither fully aligned nor strictly opposed. To address this, reward restructuring methods such as gifting and intrinsic motivation have been proposed. However, these approaches primarily focus on promoting cooperation by managing the trade-off between individual and collective returns, without explicitly addressing fairness with respect to agents’ task-specific rewards. In this paper, we propose an adaptive conflict-aware gradient adjustment method that promotes cooperation while ensuring fairness in individual rewards. The proposed method dynamically balances policy gradients derived from individual and collective objectives in situations where the two objectives are in conflict. By explicitly resolving such conflicts, our method improves collective performance while preserving fairness across agents. We provide theoretical results that guarantee monotonic non-decreasing improvement in both the collective and individual objectives and ensure fairness. Empirical results in sequential social dilemma environments demonstrate that our approach outperforms baselines in terms of social welfare, while maintaining fairness.



Paperid:2900
Authors:Anthony Zhou, Amir Barati Farimani
Abstract:
Many architectures for neural PDE surrogates have been proposed in recent years, largely based on neural networks or operator learning. In this work, we derive and propose a new architecture, the Neural Functional, which learns function to scalar mappings. Its implementation leverages insights from operator learning and neural fields, and we show the ability of neural functionals to implicitly learn functional derivatives. For the first time, this allows for an extension of Hamiltonian mechanics to neural PDE surrogates by learning the Hamiltonian functional and optimizing its functional derivatives. We demonstrate that the Hamiltonian Neural Functional can be an effective surrogate model through improved stability and conserving energy-like quantities on 1D and 2D PDEs. Beyond PDEs, functionals are prevalent in physics; functional approximation and learning with its gradients may find other uses such as in molecular dynamics or design optimization.



Paperid:2901
Authors:Lingkai Kong, Haichuan Wang, Tonghan Wang, GUOJUN XIONG, Milind Tambe
Abstract:
Incorporating pre-collected offline data from a source environment can significantly improve the sample efficiency of reinforcement learning (RL), but this benefit is often challenged by discrepancies between the transition dynamics of the source and target environments. Existing methods typically address this issue by penalizing or filtering out source transitions in high dynamics-gap regions. However, their estimation of the dynamics gap often relies on KL divergence or mutual information, which can be ill-defined when the source and target dynamics have disjoint support. To overcome these limitations, we propose CompFlow, a method grounded in the theoretical connection between flow matching and optimal transport. Specifically, we model the target dynamics as a conditional flow built upon the output distribution of the source-domain flow, rather than learning it directly from a Gaussian prior. This composite structure offers two key advantages: (1) improved generalization for learning target dynamics, and (2) a principled estimation of the dynamics gap via the Wasserstein distance between source and target transitions. Leveraging our principled estimation of the dynamics gap, we further introduce an optimistic active data collection strategy that prioritizes exploration in regions of high dynamics gap, and theoretically prove that it reduces the performance disparity with the optimal policy. Empirically, CompFlow outperforms strong baselines across several RL benchmarks with shifted dynamics.



Paperid:2902
Authors:Ruichen Chen, Keith Mills, Liyao Jiang, Chao Gao, Di Niu
Abstract:
Diffusion Transformers (DiT) have become the de-facto model for generating high-quality visual content like videos and images.A huge bottleneck is the attention mechanism where complexity scales quadratically with resolution and video length. One logical way to lessen this burden is sparse attention, where only a subset of tokens or patches are included in the calculation. However, existing techniques fail to preserve visual quality at extremely high sparsity levels and might even incur non-negligible compute overheads. To address this concern, we propose Re-ttention, which implements very high sparse attention for visual generation models by leveraging the temporal redundancy of Diffusion Models to overcome the probabilistic normalization shift within the attention mechanism. Specifically, Re-ttention reshapes attention scores based on the prior softmax distribution history in order to preserve the visual quality of the full quadratic attention at very high sparsity levels. Experimental results on T2V/T2I models such as CogVideoX and the PixArt DiTs demonstrate that Re-ttention requires as few as 3.1\% of the tokens during inference, outperforming contemporary methods like FastDiTAttn, Sparse VideoGen and MInference. Further, we measure latency to show that our method can attain over 45\% end-to-end and over 92\% self-attention latency reduction on an H100 GPU at negligible overhead cost.



Paperid:2903
Authors:Zixiang Zhao, Haowen Bai, Bingxin Ke, Yukun Cui, Lilun Deng, Yulun Zhang, Kai Zhang, Konrad Schindler
Abstract:
The real world is dynamic, yet most image fusion methods process static frames independently, ignoring temporal correlations in videos and leading to flickering and temporal inconsistency. To address this, we propose Unified Video Fusion (UniVF), a novel framework for temporally coherent video fusion that leverages multi-frame learning and optical flow-based feature warping for informative, temporally coherent video fusion. To support its development, we also introduce Video Fusion Benchmark (VF-Bench), the first comprehensive benchmark covering four video fusion tasks: multi-exposure, multi-focus, infrared-visible, and medical fusion. VF-Bench provides high-quality, well-aligned video pairs obtained through synthetic data generation and rigorous curation from existing datasets, with a unified evaluation protocol that jointly assesses the spatial quality and temporal consistency of video fusion. Extensive experiments show that UniVF achieves state-of-the-art results across all tasks on VF-Bench. Both the code and the dataset will be made publicly available.



Paperid:2904
Authors:Amaya Dharmasiri, William Yang, Polina Kirichenko, Lydia Liu, Olga Russakovsky
Abstract:
Coreset selection methods have shown promise in reducing the training data size while maintaining model performance for data-efficient machine learning. However, many large real-world datasets suffer from unknown spurious correlations and hidden biases. Therefore, it is crucial to understand how such biases would affect downstream tasks via the selected coresets. In this work, we conduct the first comprehensive analysis of the implications of data selection on the bias levels of the selected coresets and the robustness of downstream models trained on them. We use an extensive experimental setting spanning ten different spurious correlations benchmarks, five score metrics to characterize sample importance/ difficulty, and five data selection policies across a broad range of coreset sizes to identify important patterns and derive insights. Thereby, we unravel a series of nontrivial nuances in well-known interactions between sample difficulty and bias alignment, as well as dataset bias and resultant model robustness. For example, we show that embedding-based sample characterizations run a comparatively lower risk of inadvertently exacerbating bias when used for selecting coresets compared to characterizations based on learning dynamics. Our analysis also reveals that lower bias levels achieved by coresets of difficult samples do not reliably guarantee downstream robustness. Most importantly, we show that special considerations need to be made when the coreset size is very small, since there is a unique risk of highly prototypical coresets reaching high average performance while obscuring their low group-robustness.



Paperid:2905
Authors:Byeonghu Na, Mina Kang, Jiseok Kwak, Minsang Park, Jiwoo Shin, SeJoon Jun, Gayoung Lee, Jin-Hwa Kim, Il-chul Moon
Abstract:
Text-to-image models have recently made significant advances in generating realistic and semantically accurate images, driven by advanced diffusion models and large-scale web-crawled datasets. However, these datasets often contain inappropriate or biased content, raising concerns about the generation of harmful outputs when provided with malicious input text prompts. In this paper, we propose Safe Text embedding Guidance (STG), a training-free approach to improve the safety of text-to-image diffusion models by guiding the text embeddings during the sampling process. STG adjusts the text embeddings based on a safety function evaluated on the expected final denoised image, allowing the model to generate safer outputs without additional training. Theoretically, we show that STG attempts to align the underlying model distribution with desired safety constraints, thereby achieving safer outputs while minimizing degradation of the original generative quality. We experimentally validate STG on various safety scenarios, including nudity, violence, and artist-style removal. We show that STG consistently outperforms both training-based and training-free baselines in removing unsafe content while preserving the core semantic intent of input prompts.



Paperid:2906
Authors:Yuxi Liu, Renjia Deng, Yutong He, xue wang, Tao Yao, Kun Yuan
Abstract:
Abstract:The substantial memory demands of pre-training and fine-tuning large language models (LLMs) require memory-efficient optimization algorithms. One promising approach is layer-wise optimization, which treats each transformer block as a single layer and optimizes it sequentially, while freezing the other layers to save optimizer states and activations. Although effective, these methods ignore the varying importance of the modules within each layer, leading to suboptimal performance. Moreover, layer-wise sampling provides only limited memory savings, as at least one full layer must remain active during optimization. To overcome these limitations, we propose **M**odule-wise **I**mportance **SA**mpling (**MISA**), a novel method that divides each layer into smaller modules and assigns importance scores to each module. MISA uses a weighted random sampling mechanism to activate modules, provably reducinggradient variance compared to layer-wise sampling. Additionally, we establish an $\mathcal{O}(1/\sqrt{K})$ convergence rate under non-convex and stochastic conditions, where $K$ is the total number of training steps, and provide a detailed memory analysis showcasing MISA's superiority over existing baseline methods. Experiments on diverse learning tasks validate the effectiveness of MISA.



Paperid:2907
Authors:Qijun Zhang, Yao Lu, Mengming Li, Shang Liu, Zhiyao Xie
Abstract:
Power is the primary design objective of large-scale integrated circuits (ICs), especially for complex modern processors (i.e., CPUs). Accurate CPU power evaluation requires designers to go through the whole time-consuming IC implementation process, easily taking months. At the early design stage (e.g., architecture-level), classical power models are notoriously inaccurate. Recently, ML-based architecture-level power models have been proposed to boost accuracy, but the data availability is a severe challenge. Currently, there is no open-source dataset for this important ML application. A typical dataset generation process involves correct CPU design implementation and repetitive execution of power simulation flows, requiring significant design expertise, engineering effort, and execution time. Even private in-house datasets often fail to reflect realistic CPU design scenarios. In this work, we propose ArchPower, the first open-source dataset for architecture-level processor power modeling. We go through complex and realistic design flows to collect the CPU architectural information as features and the ground-truth simulated power as labels. Our dataset includes 200 CPU data samples, collected from 25 different CPU configurations when executing 8 different workloads. There are more than 100 architectural features in each data sample, including both hardware and event parameters. The label of each sample provides fine-grained power information, including the total design power and the power for each of the 11 components. Each power value is further decomposed into four fine-grained power groups: combinational logic power, sequential logic power, memory power, and clock power. ArchPower is available at https://github.com/hkust-zhiyao/ArchPower.



Paperid:2908
Authors:Nimrod Berman, Ilan Naiman, Moshe Eliasof, Hedi Zisling, Omri Azencot
Abstract:
Diffusion-based generative models have demonstrated exceptional performance, yet their iterative sampling procedures remain computationally expensive. A prominent strategy to mitigate this cost isdistillation, withoffline distillationoffering particular advantages in terms of efficiency, modularity, and flexibility. In this work, we identify two key observations that motivate a principled distillation framework: (1) while diffusion models have been viewed through the lens of dynamical systems theory, powerful and underexplored tools can be further leveraged; and (2) diffusion models inherently impose structured, semantically coherent trajectories in latent space. Building on these observations, we introduce theKoopman Distillation Model(KDM), a novel offline distillation approach grounded in Koopman theory - a classical framework for representing nonlinear dynamics linearly in a transformed space. KDM encodes noisy inputs into an embedded space where a learned linear operator propagates them forward, followed by a decoder that reconstructs clean samples. This enables single-step generation while preserving semantic fidelity. We provide theoretical justification for our approach: (1) under mild assumptions, the learned diffusion dynamics admit a finite-dimensional Koopman representation; and (2) proximity in the Koopman latent space correlates with semantic similarity in the generated outputs, allowing for effective trajectory alignment. Empirically, KDM achieves state-of-the-art performance across standardoffline distillationbenchmarks - improving FID scores by up to 40% in a single generation step.



Paperid:2909
Authors:Ziying Li, Xuequan Lu, Xinkui Zhao, Guanjie Cheng, Shuiguang Deng, Jianwei Yin
Abstract:
Recent advancements in optimization-based text-to-3D generation heavily rely on distilling knowledge from pre-trained text-to-image diffusion models using techniques like Score Distillation Sampling (SDS), which often introduce artifacts such as over-saturation and over-smoothing into the generated 3D assets. In this paper, we address this essential problem by formulating the generation process as learning an optimal, direct transport trajectory between the distribution of the current rendering and the desired target distribution, thereby enabling high-quality generation with smaller Classifier-free Guidance (CFG) values. At first, we theoretically establish SDS as a simplified instance of the Schrödinger Bridge framework. We prove that SDS employs the reverse process of an Schrödinger Bridge, which, under specific conditions (e.g., a Gaussian noise as one end), collapses to SDS's score function of the pre-trained diffusion model. Based upon this, we introduce Trajectory-Centric Distillation (TraCe), a novel text-to-3D generation framework, which reformulates the mathematically trackable framework of Schrödinger Bridge to explicitly construct a diffusion bridge from the current rendering to its text-conditioned, denoised target, and trains a LoRA-adapted model on this trajectory's score dynamics for robust 3D optimization. Comprehensive experiments demonstrate that TraCe consistently achieves superior quality and fidelity to state-of-the-art techniques. Our code will be released to the community.



Paperid:2910
Authors:Wonje Jeung, Sangyeon Yoon, Minsuk Kahng, Albert No
Abstract:
Large Reasoning Models (LRMs) have become powerful tools for complex problem solving, but their structured reasoning pathways can lead to unsafe outputs when exposed to harmful prompts. Existing safety alignment methods reduce harmful outputs but can degrade reasoning depth, leading to significant trade-offs in complex, multi-step tasks, and remain vulnerable to sophisticated jailbreak attacks. To address this, we introduce SAFEPATH, a lightweight alignment method that fine-tunes LRMs to emit a short, 8-token Safety Primer at the start of their reasoning, in response to harmful prompts, while leaving the rest of the reasoning process unsupervised. Empirical results across multiple benchmarks indicate that SAFEPATH effectively reduces harmful outputs while maintaining reasoning performance. Specifically, SAFEPATH reduces harmful responses by up to 90.0\% and blocks 83.3\% of jailbreak attempts in the DeepSeek-R1-Distill-Llama-8B model, while requiring 295.9x less compute than Direct Refusal and 314.1x less than SafeChain. We further introduce a zero-shot variant that requires no fine-tuning. In addition, we provide a comprehensive analysis of how existing methods in LLMs generalize, or fail, when applied to reasoning-centric models, revealing critical gaps and new directions for safer AI.



Paperid:2911
Authors:Jonas Kulhanek, Marie-Julie Rakotosaona, Fabian Manhardt, Christina Tsalicoglou, Michael Niemeyer, Torsten Sattler, Songyou Peng, Federico Tombari
Abstract:
In this work, we present a novel level-of-detail (LOD) method for 3D Gaussian Splatting that enables real-time rendering of large-scale scenes on memory-constrained devices. Our approach introduces a hierarchical LOD representation that iteratively selects optimal subsets of Gaussians based on camera distance, thus largely reducing both rendering time and GPU memory usage. We construct each LOD level by applying a depth-aware 3D smoothing filter, followed by importance-based pruning and fine-tuning to maintain visual fidelity. To further reduce memory overhead, we partition the scene into spatial chunks and dynamically load only relevant Gaussians during rendering, employing an opacity-blending mechanism to avoid visual artifacts at chunk boundaries. Our method achieves state-of-the-art performance on both outdoor (Hierarchical 3DGS) and indoor (Zip-NeRF) datasets, delivering high-quality renderings with reduced latency and memory requirements.



Paperid:2912
Authors:Mana Sakai, Ryo Karakida, Masaaki Imaizumi
Abstract:
Abstract:In modern theoretical analyses of neural networks, the infinite-width limit is often invoked to justify Gaussian approximations of neuron preactivations (e.g., via neural network Gaussian processes or Tensor Programs). However, these Gaussian-based asymptotic theories have so far been unable to capture the behavior of attention layers, except under special regimes such as infinitely many heads or tailored scaling schemes. In this paper, leveraging the Tensor Programs framework, we rigorously identify the infinite-width limit distribution of variables within a single attention layer under realistic architectural dimensionality and standard $1/\sqrt{n}$-scaling with $n$ dimensionality. We derive the exact form of this limit law without resorting to infinite-head approximations or tailored scalings, demonstrating that it departs fundamentally from Gaussianity. This limiting distribution exhibits non-Gaussianity from a hierarchical structure, being Gaussian conditional on the random similarity scores. Numerical experiments validate our theoretical predictions, confirming the effectiveness of our theory at finite width and accurate description of finite-head attentions. Beyond characterizing a standalone attention layer, our findings lay the groundwork for developing a unified theory of deep Transformer architectures in the infinite-width regime.



Paperid:2913
Authors:Ye Zhu, Duo Xu, Zhiwei Deng, Jonathan Tan, Olga Russakovsky
Abstract:
We study Diffusion Schrödinger Bridge (DSB) models in the context of dynamical astrophysical systems, specifically tackling observational inverse prediction tasks within Giant Molecular Clouds (GMCs) for star formation. We introduce the Astro-DSB model, a variant of DSB with the pairwise domain assumption tailored for astrophysical dynamics. By investigating its learning process and prediction performance in both physically simulated data and in real observations (the Taurus B213 data), we present two main takeaways. First, from the astrophysical perspective, our proposed paired DSB method improves interpretability, learning efficiency, and prediction performance over conventional astrostatistical and other machine learning methods. Second, from the generative modeling perspective, probabilistic generative modeling reveals improvements over discriminative pixel-to-pixel modeling in Out-Of-Distribution (OOD) testing cases of physical simulations with unseen initial conditions and different dominant physical processes. Our study expands research into diffusion models beyond the traditional visual synthesis application and provides evidence of the models' learning abilities beyond pure data statistics, paving a path for future physics-aware generative models which can align dynamics between machine learning and real (astro)physical systems.



Paperid:2914
Authors:Yang Xiao, Jiashuo WANG, Ruifeng Yuan, Chunpu Xu, Kaishuai Xu, Wenjie Li, Pengfei Liu
Abstract:
Large language models (LLMs) have demonstrated remarkable reasoning capabilities through test-time scaling approaches, particularly when fine-tuned with chain-of-thought (CoT) data distilled from more powerful large reasoning models (LRMs). However, these reasoning chains often contain verbose elements that mirror human problem-solving, categorized as progressive reasoning (the essential solution development path) and functional elements (verification processes, alternative solution approaches, and error corrections). While progressive reasoning is crucial, the functional elements significantly increase computational demands during test-time inference. We introduce PIR (Perplexity-based Importance Refinement), a principled framework that quantitatively evaluates the importance of each reasoning step based on its impact on answer prediction confidence. PIR systematically identifies and selectively prunes only low-importance functional steps while preserving all progressive reasoning components, creating optimized training data that maintains the integrity of the core solution path while reducing verbosity. Models fine-tuned on PIR-optimized data exhibit superior test-time scaling properties, generating more concise reasoning chains while achieving improved accuracy (+0.9\% to +6.6\%) with significantly reduced token usage (-3\% to -41\%) across challenging reasoning benchmarks (AIME, AMC, and GPQA Diamond). Our approach demonstrates strong generalizability across different model sizes, data sources, and token budgets, offering a practical solution for deploying reasoning-capable LLMs in scenarios where efficient test-time scaling, response time, and computational efficiency are valuable constraints. Code and dataset are available at ananonymous GitHub repository.



Paperid:2915
Authors:Xinjian Zhao, Wei Pang, Zhongkai Xue, Xiangru Jian, Lei Zhang, Yaoyao Xu, Xiaozhuang Song, Shu Wu, Tianshu Yu
Abstract:
Graph Neural Networks operate through bottom-up message-passing, fundamentally differing from human visual perception, which intuitively captures global structures first. We investigate the underappreciated potential of vision models for graph understanding, finding they achieve performance comparable to GNNs on established benchmarks while exhibiting distinctly different learning patterns.These divergent behaviors, combined with limitations of existing benchmarks that conflate domain features with topological understanding, motivate our introduction of GraphAbstract. This benchmark evaluates models' ability to perceive global graph properties as humans do: recognizing organizational archetypes, detecting symmetry, sensing connectivity strength, and identifying critical elements. Our results reveal that vision models significantly outperform GNNs on tasks requiring holistic structural understanding and maintain generalizability across varying graph scales, while GNNs struggle with global pattern abstraction and degrade with increasing graph size.This work demonstrates that vision models possess remarkable yet underutilized capabilities for graph structural understanding, particularly for problems requiring global topological awareness and scale-invariant reasoning. These findings open new avenues to leverage this underappreciated potential for developing more effective graph learning systems for tasks dominated by holistic pattern recognition.



Paperid:2916
Authors:Masanari Kimura
Abstract:
Abstract:Label shift adaptation aims to recover target class priors when the labelled source distribution $P$ and the unlabelled target distribution $Q$ share $P(X \mid Y) = Q(X \mid Y)$ but $P(Y) \neq Q(Y)$. Classical black‑box shift estimators invert an empirical confusion matrix of a frozen classifier, producing a brittle point estimate that ignores sampling noise and similarity among classes. We present Graph‑Smoothed Bayesian BBSE (GS‑B$^3$SE), a fully probabilistic alternative that places Laplacian–Gaussian priors on both target log‑priors and confusion‑matrix columns, tying them together on a label‑similarity graph. The resulting posterior is tractable with HMC or a fast block Newton–CG scheme. We prove identifiability, $N^{-1/2}$ contraction, variance bounds that shrink with the graph’s algebraic connectivity, and robustness to Laplacian misspecification. We also reinterpret GS‑B$^3$SE through information geometry, showing that it generalizes existing shift estimators.



Paperid:2917
Authors:Seunghoon Lee, Jeongwoo Choi, Byunggwan Son, JaeHyeon Moon, Jeimin Jeon, Bumsub Ham
Abstract:
Abstract:We present in this paper a novel post-training quantization (PTQ) method, dubbed AccuQuant, for diffusion models. We show analytically and empirically that quantization errors for diffusion models are accumulated over denoising steps in a sampling process. To alleviate the error accumulation problem, AccuQuant minimizes the discrepancies between outputs of a full-precision diffusion model and its quantized version within a couple of denoising steps. That is, it simulates multiple denoising steps of a diffusion sampling process explicitly for quantization, accounting the accumulated errors over multiple denoising steps, which is in contrast to previous approaches to imitating a training process of diffusion models, namely, minimizing the discrepancies independently for each step. We also present an efficient implementation technique for AccuQuant, together with a novel objective, which reduces a memory complexity significantly from $\mathcal{O}(n)$ to $\mathcal{O}(1)$, where $n$ is the number of denoising steps. We demonstrate the efficacy and efficiency of AccuQuant across various tasks and diffusion models on standard benchmarks.



Paperid:2918
Authors:Wenyao Zhang, Hongsi Liu, Zekun Qi, Yunnan Wang, XinQiang Yu, Jiazhao Zhang, Runpei Dong, Jiawei He, He Wang, Zhizheng Zhang, Li Yi, Wenjun Zeng, Xin Jin
Abstract:
Recent advances in vision-language-action (VLA) models have shown promise in integrating image generation with action prediction to improve generalization and reasoning in robot manipulation. However, existing methods are limited to challenging image-based forecasting, which suffers from redundant information and lacks comprehensive and critical world knowledge, including geometry, semantics and spatial information.To address these limitations, we propose DreamVLA, a novel VLA framework that integrates comprehensive world knowledge forecasting to enable inverse dynamics modeling, thereby establishing an action-forecasting loop for manipulation tasks. Specifically, DreamVLA introduces a dynamic-region-guided world knowledge prediction mechanism, which anticipates visual, depth, geometric, semantic, and segmentation cues to provide compact yet comprehensive representations for action planning.This design aligns with how humans interact with the world by first forming abstract multimodal reasoning chains before acting.Moreover, to model the conditional distribution over future actions, we employ a diffusion-based transformer that disentangles action representations from shared latent features and better captures multimodal uncertainty. Extensive experiments on both real-world and simulation environments demonstrate that DreamVLA achieves 76.7% success rate on real robot tasks and 4.45 average length on the CALVIN ABC-D benchmarks.



Paperid:2919
Authors:Naoya Yamamoto, Juno Kim, Taiji Suzuki
Abstract:
Wasserstein gradient flow (WGF) is a common method to perform optimization over the space of probability measures. While WGF is guaranteed to converge to a first-order stationary point, for nonconvex functionals the converged solution does not necessarily satisfy the second-order optimality condition; i.e., it could converge to a saddle point. In this work, we propose a new algorithm for probability measure optimization, \emph{perturbed Wasserstein gradient flow} (PWGF), that achieves second-order optimality for general nonconvex objectives. PWGF enhances WGF by injecting noisy perturbations near saddle points via a Gaussian process-based scheme. By pushing the measure forward along a random vector field generated from a Gaussian process, PWGF helps the solution escape saddle points efficiently by perturbing the solution towards the smallest eigenvalue direction of the Wasserstein Hessian. We theoretically derive the computational complexity for PWGF to achieve a second-order stationary point. Furthermore, we prove that PWGF converges to a global optimum in polynomial time for strictly benign objectives.



Paperid:2920
Authors:Yihong Tang, Kehai Chen, Muyun Yang, Zheng-Yu Niu, Jing Li, Tiejun Zhao, Min Zhang
Abstract:
The advancement of Large Language Models (LLMs) has spurred significant interest in Role-Playing Agents (RPAs) for applications such as emotional companionship and virtual interaction. However, recent RPAs are often built on explicit dialogue data, lacking deep, human-like internal thought processes, resulting in superficial knowledge and style expression. While Large Reasoning Models (LRMs) can be employed to simulate character thought, their direct application is hindered by attention diversion (i.e., RPAs forget their role) and style drift (i.e., overly formal and rigid reasoning rather than character-consistent reasoning). To address these challenges, this paper introduces a novel Role-Aware Reasoning (RAR) method, which consists of two important stages: Role Identity Activation (RIA) and Reasoning Style Optimization (RSO). RIA explicitly guides the model with character profiles during reasoning to counteract attention diversion, and then RSO aligns reasoning style with the character and scene via LRM distillation to mitigate style drift. Extensive experiments demonstrate that the proposed RAR significantly enhances the performance of RPAs by effectively addressing attention diversion and style drift.



Paperid:2921
Authors:Chaoyou Fu, Peixian Chen, Yunhang Shen, Yulei Qin, Mengdan Zhang, Xu Lin, Jinrui Yang, Xiawu Zheng, Ke Li, Xing Sun, Yunsheng Wu, Rongrong Ji, Caifeng Shan, Ran He
Abstract:
Multimodal Large Language Model (MLLM) relies on the powerful LLM to perform multimodal tasks, showing amazing emergent abilities in recent studies, such as writing poems based on an image. However, it is difficult for these case studies to fully reflect the performance of MLLM, lacking a comprehensive evaluation. In this paper, we fill in this blank, presenting the first comprehensive MLLM Evaluation benchmark MME. It measures both perception and cognition abilities on a total of 14 subtasks. In order to avoid data leakage that may arise from direct use of public datasets for evaluation, the annotations of instruction-answer pairs are all manually designed. The concise instruction design allows us to fairly compare MLLMs, instead of struggling in prompt engineering. Besides, with such an instruction, we can also easily carry out quantitative statistics. A total of 30 advanced MLLMs are comprehensively evaluated on our MME, which not only suggests that existing MLLMs still have a large room for improvement, but also reveals the potential directions for the subsequent model optimization. Our benchmark has made substantial contributions to the development of MLLMs: (1) It has been applied for use by 300 different institutions, such as OpenAI, Google, Meta, CMU, MIT, and Stanford, and has become the standard test set for many MLLMs, such as LLaVA, Qwen-VL, and Intern-VL series; (2) This paper has been cited over a thousand times.



Paperid:2922
Authors:Gleb Bazhenov, Oleg Platonov, Liudmila Prokhorenkova
Abstract:
Although data that can be naturally represented as graphs is widespread in real-world applications across diverse industries, popular graph ML benchmarks for node property prediction only cover a surprisingly narrow set of data domains, and graph neural networks (GNNs) are often evaluated on just a few academic citation networks. This issue is particularly pressing in light of the recent growing interest in designing graph foundation models. These models are supposed to be able to transfer to diverse graph datasets from different domains, and yet the proposed graph foundation models are often evaluated on a very limited set of datasets from narrow applications. To alleviate this issue, we introduce GraphLand: a benchmark of 14 diverse graph datasets for node property prediction from a range of different industrial applications. GraphLand allows evaluating graph ML models on a wide range of graphs with diverse sizes, structural characteristics, and feature sets, all in a unified setting. Further, GraphLand allows investigating such previously underexplored research questions as how realistic temporal distributional shifts under transductive and inductive settings influence graph ML model performance. To mimic realistic industrial settings, we use GraphLand to compare GNNs with gradient-boosted decision trees (GBDT) models that are popular in industrial applications and show that GBDTs provided with additional graph-based input features can sometimes be very strong baselines. Further, we evaluate currently available general-purpose graph foundation models and find that they produce very inconsistent but mostly weak results on our datasets.



Paperid:2923
Authors:Xiaoyu Zhan, Wenxuan Huang, Hao Sun, Xinyu Fu, Changfeng Ma, Shaosheng Cao, Bohan Jia, Shaohui Lin, Zhenfei Yin, LEI BAI, Wanli Ouyang, Yuanqi Li, Jie Guo, Yanwen Guo
Abstract:
Recent advances in Multimodal Large Language Models (MLLMs) have significantly improved 2D visual understanding, prompting interest in their application to complex 3D reasoning tasks. However, it remains unclear whether these models can effectively capture the detailed spatial information required for robust real-world performance, especially cross-view consistency, a key requirement for accurate 3D reasoning. Considering this issue, we introduce Viewpoint Learning, a task designed to evaluate and improve the spatial reasoning capabilities of MLLMs. We present the Viewpoint-100K dataset, consisting of 100K object-centric image pairs with diverse viewpoints and corresponding question-answer pairs. Our approach employs a two-stage fine-tuning strategy: first, foundational knowledge is injected to the baseline MLLM via Supervised Fine-Tuning (SFT) on Viewpoint-100K, resulting in significant improvements across multiple tasks; second, generalization is enhanced through Reinforcement Learning using the Group Relative Policy Optimization (GRPO) algorithm on a broader set of questions. Additionally, we introduce a hybrid cold-start initialization method designed to simultaneously learn viewpoint representations and maintain coherent reasoning thinking. Experimental results show that our approach significantly activates the spatial reasoning ability of MLLM, improving performance on both in-domain and out-of-domain reasoning tasks. Our findings highlight the value of developing foundational spatial skills in MLLMs, supporting future progress in robotics, autonomous systems, and 3D scene understanding.



Paperid:2924
Authors:Yuxin Xiao, Shan Chen, Jack Gallifant, Danielle Bitterman, Tom Hartvigsen, Marzyeh Ghassemi
Abstract:
Characterizing a large language model's (LLM's) knowledge of a given question is challenging.As a result, prior work has primarily examined LLM behavior under knowledge conflicts, where the model's internal parametric memory contradicts information in the external context.However, this does not fully reflect how well the model knows the answer to the question.In this paper, we first introduce a taxonomy of five knowledge statuses based on the consistency and correctness of LLM knowledge modes.We then propose KScope, a hierarchical framework of statistical tests that progressively refines hypotheses about knowledge modes and characterizes LLM knowledge into one of these five statuses. We apply KScope to nine LLMs across four datasets and systematically establish:(1) Supporting context narrows knowledge gaps across models.(2) Context features related to difficulty, relevance, and familiarity drive successful knowledge updates.(3) LLMs exhibit similar feature preferences when partially correct or conflicted, but diverge sharply when consistently wrong.(4) Context summarization constrained by our feature analysis, together with enhanced credibility, further improves update effectiveness and generalizes across LLMs.



Paperid:2925
Authors:Zhangyin Feng, Qianglong Chen, Ning Lu, Yongqian Li, Siqi Cheng, Shuangmu Peng, Duyu Tang, Shengcai Liu, Zhirui Zhang
Abstract:
The development of reasoning capabilities represents a critical frontier in large language models (LLMs) research, where reinforcement learning (RL) and process reward models (PRMs) have emerged as predominant methodological frameworks. Contrary to conventional wisdom, empirical evidence from DeepSeek-R1 demonstrates that pure RL training focused on mathematical problem-solving can progressively enhance reasoning abilities without PRM integration, challenging the perceived necessity of process supervision.In this study, we conduct a systematic investigation of the relationship between RL training and PRM capabilities. Our findings demonstrate that problem-solving proficiency and process supervision capabilities represent complementary dimensions of reasoning that co-evolve synergistically during pure RL training. In particular, current PRMs underperform simple baselines like majority voting when applied to state-of-the-art models such as DeepSeek-R1 and QwQ-32B. To address this limitation, we propose Self-PRM, an introspective framework in which models autonomously evaluate and rerank their generated solutions through self-reward mechanisms. Although Self-PRM consistently improves the accuracy of the benchmark (particularly with larger sample sizes), analysis exposes persistent challenges: The approach exhibits low precision (<10\%) on difficult problems, frequently misclassifying flawed solutions as valid. These analyses underscore the need for combined training with process supervision and continued RL scaling to enhance reward alignment and introspective accuracy. We hope these findings provide actionable insights for building more reliable and self-aware complex reasoning models.



Paperid:2926
Authors:Duy M. H. Nguyen, Nghiem Diep, Trung Nguyen, Hoang-Bao Le, Tai Nguyen, Anh-Tien Nguyen, TrungTin Nguyen, Nhat Ho, Pengtao Xie, Roger Wattenhofer, Daniel Sonntag, James Zou, Mathias Niepert
Abstract:
State-of-the-art medical multi-modal LLMs (med-MLLMs), such as LLaVA-Med and BioMedGPT, primarily depend on scaling model size and data volume, with training driven largely by autoregressive objectives. However, we reveal that this approach can lead to weak vision-language alignment, making these models overly dependent on costly instruction-following data. To address this, we introduce ExGra-Med, a novel multi-graph alignment framework that jointly aligns images, instruction responses, and extended captions in the latent space, advancing semantic grounding and cross-modal coherence. To scale to large LLMs (e.g., LLaMa-7B), we develop an efficient end-to-end training scheme using black-box gradient estimation, enabling fast and scalable optimization. Empirically, ExGra-Med matches LLaVA-Med’s performance using just 10\% of pre-training data, achieving a 20.13\% gain on VQA-RAD and approaching full-data performance. It also outperforms strong baselines like BioMedGPT and RadFM on visual chatbot and zero-shot classification tasks, demonstrating its promise for efficient, high-quality vision-language integration in medical AI.



Paperid:2927
Authors:Shuang Zhang, Yue Wu, Yingxue Zhang, Lei Shi, Huilong Jin, Feifei Kou, Pengfei Zhang, Meiyu Liang, Mingying Xu
Abstract:
The demand for multimodal data processing drives the development of information technology. Cross-modal hash retrieval has attracted much attention because it can overcome modal differences and achieve efficient retrieval, and has shown great application potential in many practical scenarios. Existing cross-modal hashing methods have difficulties in fully capturing the semantic information of different modal data, which leads to a significant semantic gap between modalities. Moreover, these methods often ignore the importance differences of channels, and due to the limitation of a single goal, the matching effect between hash codes is also affected to a certain extent, thus facing many challenges. To address these issues, we propose a Dynamic Masking and Auxiliary Hash Learning (AHLR) method for enhanced cross-modal retrieval. By jointly leveraging the dynamic masking and auxiliary hash learning mechanisms, our approach effectively resolves the problems of channel information imbalance and insufficient key information capture, thereby significantly improving the retrieval accuracy. Specifically, we introduce a dynamic masking mechanism that automatically screens and weights the key information in images and texts during the training process, enhancing the accuracy of feature matching. We further construct an auxiliary hash layer to adaptively balance the weights of features across each channel, compensating for the deficiencies of traditional methods in key information capture and channel processing. In addition, we design a contrastive loss function to optimize the generation of hash codes and enhance their discriminative power, further improving the performance of cross-modal retrieval.Comprehensive experimental results on NUS-WIDE, MIRFlickr 25K and MS-COCO benchmark datasets show that the proposed AHLR algorithm outperforms several existing algorithms.



Paperid:2928
Authors:Chu Xu, Zhixin Zhang, Tianyu Jia, Yujie Jin
Abstract:
Abstract:Aligning large language models (LLMs) with human preferences typically demands vast amounts of meticulously curated data, which is both expensive and prone to labeling noise. We propose Stackelberg Game Preference Optimization (SGPO), a robust alignment framework that models alignment as a two-player Stackelberg game between a policy (leader) and a worst-case preference distribution (follower). The proposed SGPO guarantees $\mathcal{O}(\epsilon)$-bounded regret within an $\epsilon$-Wasserstein ball, offering formal robustness to (self-)annotation noise. We instantiate SGPO with Stackelberg Self-Annotated Preference Optimization (SSAPO), which uses minimal human-labeled “seed” preferences and iteratively self-annotates new prompts. In each iteration, SSAPO applies a distributionally robust reweighting of synthetic annotations, ensuring that noisy or biased self-labels do not derail training. Remarkably, using only 2K seed preferences—about 1/30 of standard human labels—SSAPO achieves strong win rates against GPT-4 across multiple benchmarks within three iterations. These results highlight that a principled Stackelberg formulation yields data-efficient alignment for LLMs, significantly reducing reliance on costly human annotations.



Paperid:2929
Authors:Zhenhailong Wang, Senthil Purushwalkam, Caiming Xiong, Silvio Savarese, Heng Ji, Ran Xu
Abstract:
We present DyMU, an efficient, training-free framework that dynamically reduces the computational burden of vision-language models (VLMs) while maintaining high task performance. Our approach comprises two key components. First, Dynamic Token Merging (DToMe) reduces the number of visual token embeddings by merging similar tokens based on image complexity, addressing the inherent inefficiency of fixed-length outputs in vision transformers. Second, Virtual Token Unmerging (VTU) simulates the expected token sequence for large language models (LLMs) by efficiently reconstructing the attention dynamics of a full sequence, thus preserving the downstream performance without additional fine-tuning. Unlike previous approaches, our method dynamically determines token length based on theimage content—not just resolution—and operates completely training-free, making it readily applicable to most state-of-the-art VLM architectures. Extensive experiments on image and video understanding tasks, demonstrate that DyMU can reduce the average visual token count by 32%-85% while achieving comparable performance to full-length models, across diverse VLM architectures. Furthermore, qualitative analyses show that the adaptive token reduction from DToMe aligns well with human perception and enables users to better control computational costs through flexible integration with additional vision tools and models.



Paperid:2930
Authors:Shiqian Li, Zhi Li, Zhancun Mu, Shiji Xin, Zhixiang Dai, Kuangdai Leng, Rita Zhang, Xiaodong Song, Yixin Zhu
Abstract:
Global seismic tomography, taking advantage of seismic waves from natural earthquakes, provides essential insights into the earth's internal dynamics. Advanced Full-Waveform Inversion (FWI) techniques, whose aim is to meticulously interpret every detail in seismograms, confront formidable computational demands in forward modeling and adjoint simulations on a global scale. Recent advancements in Machine Learning (ML) offer a transformative potential for accelerating the computational efficiency of FWI and extending its applicability to larger scales. This work presents the first 3D global synthetic dataset tailored for seismic wavefield modeling and full-waveform tomography, referred to as the Global Tomography (GlobalTomo) dataset. This dataset is comprehensive, incorporating explicit wave physics and robust geophysical parameterization at realistic global scales, generated through state-of-the-art forward simulations optimized for 3D global wavefield calculations. Through extensive analysis and the establishment of ML baselines, we illustrate that ML approaches are particularly suitable for global FWI, overcoming its limitations with rapid forward modeling and flexible inversion strategies. This work represents a cross-disciplinary effort to enhance our understanding of the earth's interior through physics-ML modeling.



Paperid:2931
Authors:Fanqi Yan, Huy Nguyen, Le Dung, Pedram Akbarian, Nhat Ho, Alessandro Rinaldo
Abstract:
The softmax-contaminated mixture of experts (MoE) model is deployed when a large-scale pre-trained model, which plays the role of a fixed expert, is fine-tuned for learning downstream tasks by including a new contamination part, or prompt, functioning as a new, trainable expert. Despite its popularity and relevance, the theoretical properties of the softmax-contaminated MoE have remained unexplored in the literature. In the paper, we study the convergence rates of the maximum likelihood estimator of gating and prompt parameters in order to gain insights into the statistical properties and potential challenges of fine-tuning with a new prompt. We find that the estimability of these parameters is compromised when the prompt acquires overlapping knowledge with the pre-trained model, in the sense that we make precise by formulating a novel analytic notion of distinguishability. Under distinguishability of the pre-trained and prompt models, we derive minimax optimal estimation rates for all the gating and prompt parameters. By contrast, when the distinguishability condition is violated, these estimation rates become significantly slower due to their dependence on the prompt convergence rate to the pre-trained model. Finally, we empirically corroborate our theoretical findings through several numerical experiments.



Paperid:2932
Authors:Naoki Kiyohara, Edward Johns, Yingzhen Li
Abstract:
Stochastic differential equations (SDEs) are well suited to modelling noisy and/or irregularly-sampled time series, which are omnipresent in finance, physics, and machine learning applications. Traditional approaches require costly simulation of numerical solvers when sampling between arbitrary time points. We introduce neural stochastic flows (NSF) and their latent dynamic versions, which learns (latent) SDE transition laws directly using conditional normalising flows, with architectural constraints that preserve properties inherited from stochastic flow. This enables sampling between arbitrary states in a single step, providing up to two orders of magnitude speedup for distant time points. Experiments on synthetic SDE simulations and real-world tracking and video data demonstrate that NSF maintains distributional accuracy comparable to numerical approaches while dramatically reducing computation for arbitrary time-point sampling, enabling applications where numerical solvers remain prohibitively expensive.



Paperid:2933
Authors:Songlin Yang, Yikang Shen, Kaiyue Wen, Shawn Tan, Mayank Mishra, Liliang Ren, Rameswar Panda, Yoon Kim
Abstract:
The attention mechanism is a core primitive in modern large language models (LLMs) and AI more broadly. Since attention by itself is permutation-invariant, position encoding is essential for modeling structured domains such as language. Rotary position encoding (RoPE) has emerged as the de facto standard approach for position encoding and is part of many modern LLMs. However, in RoPE the key/query transformation between two elements in a sequence is only a function of their relative position and otherwise independent of the actual input. This limits the expressivity of RoPE-based transformers. This paper describes PaTH, a flexible data-dependent position encoding scheme based on accumulated products of Householder(like) transformations, where each transformation is data-dependent, i.e., a function of the input. We derive an hardware-efficient parallel algorithm for training through exploiting a compact representation of products of Householder matrices, and implement a FlashAttention-style blockwise algorithm that minimizes I/O cost. Across both targeted synthetic benchmarks and real-world language modeling experiments, we find that PaTH demonstrates superior performance compared to RoPE and other recent baselines like Forgetting Transformer.



Paperid:2934
Authors:Anna Sokol, Elizabeth Daly, Michael Hind, David Piorkowski, Xiangliang Zhang, Nuno Moniz, Nitesh Chawla
Abstract:
Large language models (LLMs) are powerful tools capable of handling diverse tasks. Comparing and selecting appropriate LLMs for specific tasks requires systematic evaluation methods, as models exhibit varying capabilities across different domains. However, finding suitable benchmarks is difficult given the many available options. This complexity not only increases the risk of benchmark misuse and misinterpretation but also demands substantial effort from LLM users, seeking the most suitable benchmarks for their specific needs. To address these issues, we introduce BenchmarkCards, an intuitive and validated documentation framework that standardizes critical benchmark attributes such as objectives, methodologies, data sources, and limitations. Through user studies involving benchmark creators and users, we show that BenchmarkCards can simplify benchmark selection and enhance transparency, facilitating informed decision-making in evaluating LLMs.Data & Code:github.com/SokolAnn/BenchmarkCards



Paperid:2935
Authors:Yingsha Xie, Rui Min, Zeyu Qin, Fei Ma, Li Shen, Fei Yu, Xiaochun Cao
Abstract:
Abstract:Preserving intellectual property (IP) within a pre-trained diffusion model is critical for protecting the model's copyright and preventing unauthorized model deployment. In this regard, model watermarking is a common practice that embeds traceable information within model weights and allows for further verification, making it a widely applied scheme for model IP protection. Nevertheless, existing watermarking schemes often face challenges due to their vulnerability to fine-tuning, limiting their practical application in general pre-training and fine-tuning paradigms. Inspired by using mode connectivity to analyze model performance between a pair of connected models, in this paper, we investigate watermark vulnerability by leveraging Linear Mode Connectivity (LMC) as a proxy to analyze the fine-tuning dynamics of watermark performance. Our results show that existing watermarked models tend to converge to sharp minima in the loss landscape, making them vulnerable to loss of watermarking information during fine-tuning. Inspired by our findings, we propose RoMa, a **Ro**bust **M**odel w**a**termarking scheme that improves the robustness of watermarks against fine-tuning by enhancing the smoothness along the watermark-connected path between the pre-trained and watermarked models. Specifically, RoMa decomposes watermarking into two components, including *Embedding Functionality*, which preserves reliable watermark detection capability, and *Path-specific Smoothness*, which leverages extra updates guided by the watermark-connected path to enhance robustness. Extensive experiments demonstrate that RoMa significantly improves watermark robustness while maintaining generation quality. Notably, our method requires (32.83 $\times$) more steps to remove the watermark compared to existing baselines.



Paperid:2936
Authors:Alex Su, Haozhe Wang, Weiming Ren, Fangzhen Lin, Wenhu Chen
Abstract:
Chain-of-thought reasoning has significantly improved the performance of LargeLanguage Models (LLMs) across various domains. However, this reasoning process has been confined exclusively to textual space, limiting its effectiveness in visually intensive tasks. To address this limitation, we introduce the concept of pixel-space reasoning. Within this novel framework, Vision-Language Models (VLMs) are equipped with a suite of visual reasoning operations, such as zoom-in and select-frame. These operations enable VLMs to directly inspect, interrogate, and infer from visual evidences, thereby enhancing reasoning fidelity for visual tasks. Cultivating such pixel-space reasoning capabilities in VLMs presents notable challenges, including the model’s initially imbalanced competence and its reluctance to adopt the newly introduced pixel-space operations. We address these challenges through a two-phase training approach. The first phase employs instruction tuning on synthesized reasoning traces to familiarize the model with the novel visual operations. Following this, a reinforcement learning (RL) phase leverages a curiosity-driven reward scheme to balance exploration between pixel-space reasoning and textual reasoning. With these visual operations, VLMs can interact with complex visual inputs, such as information-rich images or videos to proactively gather necessary information. We demonstrate that this approach significantly improves VLM performance across diverse visual reasoning benchmarks. Our 7B model, Pixel-Reasoner, achieves 84% on V* bench, 74% on TallyQA-Complex, and 84% on InfographicsVQA, marking the highest accuracy achieved by any open-source model to date. These results highlight the importance of pixel-space reasoning and the effectiveness of our framework.



Paperid:2937
Authors:Liyao Tang, Zhe Chen, Dacheng Tao
Abstract:
The emergence of large-scale pretrained point cloud models has significantly advanced 3D scene understanding, but adapting these models to specific downstream tasks typically demands full fine-tuning, incurring high computational and storage costs. Parameter-efficient fine-tuning (PEFT) techniques, successful in natural language processing and 2D vision tasks, would underperform when naively applied to 3D point cloud models due to significant geometric and spatial distribution shifts.Existing PEFT methods commonly treat points as orderless tokens, neglecting important local spatial structures and global geometric contexts in 3D modeling.To bridge this gap, we introduce the Geometric Encoding Mixer (GEM), a novel geometry-aware PEFT module specifically designed for 3D point cloud transformers. GEM explicitly integrates fine-grained local positional encodings with a lightweight latent attention mechanism to capture comprehensive global context, thereby effectively addressing the spatial and geometric distribution mismatch.Extensive experiments demonstrate that GEM achieves performance comparable to or sometimes even exceeding full fine-tuning, while only updating 1.6\% of the model's parameters, fewer than other PEFT methods.With significantly reduced training time and memory requirements, our approach thus sets a new benchmark for efficient, scalable, and geometry-aware fine-tuning of large-scale 3D point cloud models. Code will be released.



Paperid:2938
Authors:zhaokun wang, Jinyu Guo, Jingwen Pu, ChenLingFeng, Hongli Pu, Jie Ou, Libo Qin, Wenhong Tian
Abstract:
Current parameter-efficient fine-tuning methods for adapting pre-trained language models to downstream tasks are susceptible to interference from noisy data. Conventional noise-handling approaches either rely on laborious data pre-processing or employ model architecture modifications prone to error accumulation. In contrast to existing noise-process paradigms, we propose a noise-robust adaptation method via asymmetric LoRA poisoning experts (LoPE), a novel framework that enhances model robustness to noise only with generated noisy data. Drawing inspiration from the mixture-of-experts architecture, LoPE strategically integrates a dedicated poisoning expert in an asymmetric LoRA configuration. Through a two-stage paradigm, LoPE performs noise injection on the poisoning expert during fine-tuning to enhance its noise discrimination and processing ability. During inference, we selectively mask the dedicated poisoning expert to leverage purified knowledge acquired by normal experts for noise-robust output. Extensive experiments demonstrate that LoPE achieves strong performance and robustness purely through the low-cost noise injection, which completely eliminates the requirement of data cleaning.



Paperid:2939
Authors:Undral Byambadalai, Tomu Hirata, Tatsushi Oka, Shota Yasui
Abstract:
We study the estimation of distributional treatment effects in randomized experiments with imperfect compliance. When participants do not adhere to their assigned treatments, we leverage treatment assignment as an instrumental variable to identify the local distributional treatment effect—the difference in outcome distributions between treatment and control groups for the subpopulation of compliers. We propose a regression-adjusted estimator based on a distribution regression framework with Neyman-orthogonal moment conditions, enabling robustness and flexibility with high-dimensional covariates. Our approach accommodates continuous, discrete, and mixed discrete-continuous outcomes, and applies under a broad class of covariate-adaptive randomization schemes, including stratified block designs and simple random sampling. We derive the estimator’s asymptotic distribution and show that it achieves the semiparametric efficiency bound. Simulation results demonstrate favorable finite-sample performance, and we demonstrate the method’s practical relevance in an application to the Oregon Health Insurance Experiment.



Paperid:2940
Authors:Guangyi Zhang, Yanhao Wang, Chengliang Chai, Qiyu Liu, Wei Wang
Abstract:
Abstract:Data Shapley values provide a principled approach for quantifying the contribution of individual training examples to machine learning models. However, computing these values often requires computational complexity that is exponential in the data size, and this has led researchers to pursue efficient algorithms tailored to specific machine learning models. Building on the prior success of the Shapley valuation for $K$-nearest neighbor (KNN) models, in this paper, we introduce a localized data Shapley framework that significantly accelerates the valuation of data points. Our approach leverages the distance-based local structure in the data space to decompose the global valuation problem into smaller, localized computations. Our primary contribution is an efficient valuation algorithm for a threshold-based KNN variant and shows that it provides provable speedups over the baseline under mild assumptions. Extensive experiments on real-life datasets demonstrate that our methods achieve a substantial speedup compared to previous approaches.



Paperid:2941
Authors:Zeyu Shen, Basileal Imana, Tong Wu, Chong Xiang, Prateek Mittal, Aleksandra Korolova
Abstract:
Retrieval-Augmented Generation (RAG) enhances the capabilities of Large Language Models by grounding their outputs in external documents. These systems, however, remain vulnerable to attacks on the retrieval corpus, such as prompt injection. RAG-based search systems (e.g., Google’s Search AI Overview) present an interesting setting for studying and protecting against such threats, as the defense algorithms could benefit from built-in reliability signals — such as document ranking, and represent a non-LLM related challenge for the adversary thanks to decades of work to thwart SEO. Motivated by, but not limited to, this scenario, this work introduces ReliabilityRAG, a framework designed for adversarial robustness that explicitly leverages reliability information of retrieved documents.Our first contribution adopts a graph-theoretic perspective to identify a "consistent majority" among retrieved documents in an effort to filter out malicious ones. We introduce a novel algorithm based on finding Maximum Independent Set (MIS) on a graph of documents in which edges encode contradiction relationships. Our MIS variant explicitly prioritizes higher-ranked or higher-reliability documents during selection and provides provable robustness guarantees against bounded adversarial corruption under natural assumptions. Recognizing the computational cost of exact MIS for large retrieval sets, our second contribution is a general and scalable weighted sample and aggregate framework. It explicitly utilizes reliability information of the retrieved documents, preserving some robustness guarantees while efficiently handling many documents. We present empirical results showing that ReliabilityRAG provides superior robustness against adversarial attacks compared to prior methods, maintains high benign accuracy, and excels in long-form generation tasks where prior robustness-focused methods struggled. Our work represents a significant step towards more effective and provably robust defenses against retrieved corpus corruption in RAG.



Paperid:2942
Authors:Harry Mead, Bruno Lacerda, Jakob Foerster, Nick Hawes
Abstract:
Unsupervised Environment Design (UED) seeks to automatically generate training curricula for reinforcement learning (RL) agents, with the goal of improving generalisation and zero-shot performance. However, designing effective curricula remains a difficult problem, particularly in settings where small subsets of environment parameterisations result in significant increases in the complexity of the required policy. Current methods rely on regret approximations that fail to identify challenging levels, which is compounded as the size of the environment grows. We both propose a new regret approximation, Maximised Negative Advantage (MNA), that better identifies these challenging levels and introduce Dynamic Environment Generation for UED (DEGen) to allow for UED to scale to larger environment sizes. We show empirically that MNA outperforms current regret approximations and DEGen substantially outperforms existing methods as the environment size grows.



Paperid:2943
Authors:Yuqi Zhou, Sunhao Dai, Shuai Wang, Kaiwen Zhou, Qinglin Jia, Jun Xu
Abstract:
Abstract:Recent Graphical User Interface (GUI) agents replicate the R1-Zero paradigm, coupling online Reinforcement Learning (RL) with explicit chain-of-thought reasoning prior to object grounding and thereby achieving substantial performance gains. In this paper, we first conduct extensive analysis experiments of three key components of that training pipeline: input design, output evaluation, and policy update—each revealing distinct challenges arising from blindly applying general-purpose RL without adapting to GUI grounding tasks. Input design: Current templates encourage the model to generate chain-of-thought reasoning, but longer chains unexpectedly lead to worse grounding performance. Output evaluation: Reward functions based on hit signals or box area allow models to exploit box size, leading to reward hacking and poor localization quality. Policy update: Online RL tends to overfit easy examples due to biases in length and sample difficulty, leading to under-optimization on harder cases. To address these issues, we propose three targeted solutions. First, we adopt a $\textbf{Fast Thinking Template}$ that encourages direct answer generation, reducing excessive reasoning during training. Second, we incorporate a box size constraint into the reward function to mitigate reward hacking. Third, we revise the RL objective by adjusting length normalization and adding a difficulty-aware scaling factor, enabling better optimization on hard samples. Our $\textbf{GUI-G1-3B}$, trained on 17K public samples with Qwen2.5-VL-3B-Instruct, achieves $\textbf{90.3\%}$ accuracy on ScreenSpot and $\textbf{37.1\%}$ on ScreenSpot-Pro. This surpasses all prior models of similar size and even outperforms the larger UI-TARS-7B, establishing a new state-of-the-art in GUI agent grounding.



Paperid:2944
Authors:Zheng Chen, Zichen Zou, Kewei Zhang, Xiongfei Su, Xin Yuan, Yong Guo, Yulun Zhang
Abstract:
Abstract:Diffusion models have demonstrated promising performance in real-world video super-resolution (VSR). However, the dozens of sampling steps they require, make inference extremely slow. Sampling acceleration techniques, particularly single-step, provide a potential solution. Nonetheless, achieving one step in VSR remains challenging, due to the high training overhead on video data and stringent fidelity demands. To tackle the above issues, we propose DOVE, an efficient one-step diffusion model for real-world VSR. DOVE is obtained by fine-tuning a pretrained video diffusion model (*i.e.*, CogVideoX). To effectively train DOVE, we introduce the latent–pixel training strategy. The strategy employs a two-stage scheme to gradually adapt the model to the video super-resolution task. Meanwhile, we design a video processing pipeline to construct a high-quality dataset tailored for VSR, termed HQ-VSR. Fine-tuning on this dataset further enhances the restoration capability of DOVE. Extensive experiments show that DOVE exhibits comparable or superior performance to multi-step diffusion-based VSR methods. It also offers outstanding inference efficiency, achieving a speed‑up of at least 11$\times$ over existing approaches. The code will be made publicly available.



Paperid:2945
Authors:Julien Siems, Timur Carstensen, Arber Zela, Frank Hutter, Massimiliano Pontil, Riccardo Grazzi
Abstract:
Abstract:Linear Recurrent Neural Networks (linear RNNs) have emerged as competitive alternatives to Transformers for sequence modeling, offering efficient training and linear-time inference. However, existing architectures face a fundamental trade-off between expressivity and efficiency, dictated by the structure of their state-transition matrices. Diagonal matrices, used in models such as Mamba, GLA, or mLSTM, yield fast runtime but have limited expressivity. To address this, recent architectures such as DeltaNet and RWKV-7 adopted a diagonal plus rank-1 structure, which allows simultaneous token and channel mixing, improving associative recall and, as recently shown, state-tracking whenallowing negative eigenvalues in the state-transition matrices.Building on the interpretation of DeltaNet's recurrence as performing one step of online gradient descent per token on an associative recall loss, we introduce DeltaProduct, which instead takes multiple ($n_h$) steps per token. This naturally leads to diagonal plus rank-$n_h$ state-transition matrices, formed as products of $n_h$ generalized Householder transformations, providing a tunable mechanism to balance expressivity and efficiency. We provide a detailed theoretical characterization of the state-tracking capability of DeltaProduct in finite precision and how it improves by increasing $n_h$.Our extensive experiments demonstrate that DeltaProduct outperforms DeltaNet in both state-tracking and language modeling, while also showing significantly improved length extrapolation capabilities.



Paperid:2946
Authors:Longxiang He, Deheng Ye, Junbo Tan, Xueqian Wang, Li Shen
Abstract:
Abstract:Pretraining a policy on offline data followed by fine-tuning through online interactions, known as Offline-to-Online Reinforcement Learning (O2O RL), has emerged as a promising paradigm for real-world RL deployment. However, both offline datasets and online interactions in practical environments are often noisy or even maliciously corrupted, severely degrading the performance of O2O RL. Existing works primarily focus on mitigating the conservatism of offline policies via online exploration, while the robustness of O2O RL under data corruption, including states, actions, rewards, and dynamics, is still unexplored. In this work, we observe that data corruption induces heavy-tailed behavior in the policy, thereby substantially degrading the efficiency of online exploration. To address this issue, we incorporate Inverse Probability Weighted (IPW) into the online exploration policy to alleviate heavy-tailedness, and propose a novel, simple yet effective method termed $\textbf{RPEX}$: $\textbf{R}$obust $\textbf{P}$olicy $\textbf{EX}$pansion. Extensive experimental results on D4RL datasets demonstrate that RPEX achieves SOTA O2O performance across a wide range of data corruption scenarios.



Paperid:2947
Authors:Qi Li, Runpeng Yu, Xinchao Wang
Abstract:
Multimodal large language models (MLLMs) demonstrates remarkable capabilities in handling complex multimodal tasks and are increasingly adopted in video understanding applications. However, their rapid advancement raises serious data privacy concerns, particularly given the potential inclusion of sensitive video content, such as personal recordings and surveillance footage, in their training datasets. Determining improperly used videos during training remains a critical and unresolved challenge. Despite considerable progress on membership inference attacks (MIAs) for text and image data in MLLMs, existing methods fail to generalize effectively to the video domain. These methods suffer from poor scalability as more frames are sampled and generally achieve negligible true positive rates at low false positive rates (TPR@Low FPR), mainly due to their failure to capture the inherent temporal variations of video frames and to account for model behavior differences as the number of frames varies. To address these challenges, we introduce Vid-SME (VideoSharma–MittalEntropy), the first membership inference method tailored for video data used in video understanding LLMs (VULLMs). Vid-SME leverages the confidence of model output and integrates adaptive parameterization to compute Sharma–Mittal entropy (SME) for video inputs. By leveraging the SME difference between natural and temporally-reversed video frames, Vid-SME derives robust membership scores to determine whether a given video is part of the model's training set. Experiments on various self-trained and open-sourced VULLMs demonstrate the strong effectiveness of Vid-SME.



Paperid:2948
Authors:Rui Pan, Yinwei Dai, Zhihao Zhang, Gabriele Oliaro, Zhihao Jia, Ravi Netravali
Abstract:
Abstract:Recent advances in inference-time compute have significantly improved performance on complex tasks by generating long chains of thought (CoTs) using Large Reasoning Models (LRMs). However, this improved accuracy comes at the cost of high inference latency due to the length of generated reasoning sequences and the autoregressive nature of decoding. Our key insight in tackling these overheads is that LRM inference, and the reasoning that it embeds, is highly tolerant of approximations: complex tasks are typically broken down into simpler steps, each of which brings utility based on the semantic insight it provides for downstream steps rather than the exact tokens it generates. Accordingly, we introduce \name{}, a system that automatically accelerates LRM inference by using a lightweight model to (speculatively) carry out simpler intermediate reasoning steps and reserving the costly base model only to assess (and potentially correct) the speculated outputs. Importantly, \name{}'s focus on exploiting the semantic flexibility of thinking tokens in preserving final-answer accuracy is complementary to prior speculation techniques, most notably speculative decoding, which demands token-level equivalence at each step. Across a variety of reasoning benchmarks, \name{} achieves $1.4-3.0\times$ speedup over vanilla LRM inference while improving accuracy by $0.4-9.0\%$. Compared to speculative decoding without \name{}, their combination yields an additional $8.8-58.0\%$ latency reduction. We open-source SpecReason at \url{https://anonymous.4open.science/r/specreason/}.



Paperid:2949
Authors:Hee Min Choi, Hyoa Kang, Dokwan Oh, Nam Ik Cho
Abstract:
Abstract:Diffusion transformers demonstrate significant potential for various generation tasks but are challenged by high computational cost. Recently, feature caching methods have been introduced to improve inference efficiency by storing features at certain timesteps and reusing them at subsequent timesteps. However, their effectiveness is limited as they rely only on choosing between cached features and performing model inference. Motivated by high cosine similarity between features across consecutive timesteps, we propose a cache-based framework that reuses features and selectively adapts them through linear modulation. In our framework, the selection is performed via a modulation gate, and both the gate and modulation parameters are learned. Extensive experiments show that our method achieves similar generation performance to the original sampler while requiring significantly less computation. For example, FLOPs and inference latency are reduced by $2.93\times$ and $2.15\times$ for DiT-XL/2 and by $2.83\times$ and $1.50\times$ for PixArt-$\alpha$, respectively. We find that modulation is effective when applied to as little as 2\% of layers, resulting in negligible computation overhead.



Paperid:2950
Authors:Qu Yang, Xiyang Li, Fu Lin, Mang Ye
Abstract:
Multimodal Intention Recognition (MIR) plays a critical role in applications such as intelligent assistants, service robots, and autonomous systems. However, in real-world settings, different modalities often vary significantly in informativeness, reliability, and noise levels. This leads to modality imbalance, where models tend to over-rely on dominant modalities, thereby limiting generalization and robustness. While existing methods attempt to alleviate this issue at either the sample or model level, most overlook its multi-level nature. To address this, we propose Adaptive Re-calibration Learning (ARL), a novel dual-path framework that models modality importance from both sample-wise and structural perspectives. ARL incorporates two key mechanisms: Contribution-Inverse Sample Calibration (CISC), which dynamically masks overly dominant modalities at the sample level to encourage attention to underutilized ones; and Weighted Encoder Calibration (WEC), which adjusts encoder weights based on global modality contributions to prevent overfitting. Experimental results on multiple MIR benchmarks demonstrate that ARL significantly outperforms existing methods in both accuracy and robustness, particularly under noisy or modality-degraded conditions.



Paperid:2951
Authors:Praneet Suresh, Jack Stanley, Sonia Joseph, Luca Scimeca, Danilo Bzdok
Abstract:
As generative AI systems become competent and democratized in science, business, and government, deeper insight into their failure modes now poses an acute need. The occasional volatility in their behavior, such as the propensity of transformer models to hallucinate, impedes trust and adoption of emerging AI solutions in high-stakes areas. In the present work, we establish how and when hallucinations arise in pre-trained transformer models through concept representations captured by sparse autoencoders, under scenarios with experimentally controlled uncertainty in the input space. Our systematic experiments reveal that the number of semantic concepts used by the transformer model grows as the input information becomes increasingly unstructured. In the face of growing uncertainty in the input space, the transformer model becomes prone to activate coherent yet input-insensitive semantic features, leading to hallucinated output. At its extreme, for pure-noise inputs, we identify a wide variety of robustly triggered and meaningful concepts in the intermediate activations of pre-trained transformer models, whose functional integrity we confirm through targeted steering. We also show that hallucinations in the output of a transformer model can be reliably predicted from the concept patterns embedded in transformer layer activations. This collection of insights on transformer internal processing mechanics has immediate consequences for aligning AI models with human values, AI safety, opening the attack surface for potential adversarial attacks, and providing a basis for automatic quantification of a model’s hallucination risk.



Paperid:2952
Authors:Yanbiao Ji, Yue Ding, Dan Luo, Chang Liu, Yuxiang Lu, Xin Xin, Hongtao Lu
Abstract:
Graph-based recommender systems have achieved remarkable effectiveness by modeling high-order interactions between users and items. However, such approaches are significantly undermined by popularity bias, which distorts the interaction graph’s structure—referred to as topology bias. This leads to overrepresentation of popular items, thereby reinforcing biases and fairness issues through the user-system feedback loop. Despite attempts to study this effect, most prior work focuses on the embedding or gradient level bias, overlooking how topology bias fundamentally distorts the message passing process itself. We bridge this gap by providing an empirical and theoretical analysis from a Dirichlet energy perspective, revealing that graph message passing inherently amplifies topology bias and consistently benefits highly connected nodes. To address these limitations, we propose Test-time Simplicial Propagation (TSP), which extends message passing to higher-order simplicial complexes. By incorporating richer structures beyond pairwise connections, TSP mitigates harmful topology bias and substantially improves the representation and recommendation of long-tail items during inference. Extensive experiments across five real-world datasets demonstrate the superiority of our approach in mitigating topology bias and enhancing recommendation quality. The implementation code is available at https://anonymous.4open.science/r/tsp-474B.



Paperid:2953
Authors:Jacob Lin, Edward Gryspeerdt, Ronald Clark
Abstract:
Abstract:There has been great progress in improving numerical weather prediction and climate models using machine learning. However, most global models act at a kilometer-scale, making it challenging to model individual clouds and factors such as extreme precipitation, wind gusts, turbulence, and surface irradiance. Therefore, there is a need to move towards higher-resolution models, which in turn require high-resolution real-world observations that current instruments struggle to obtain. We present Cloud4D, the first learning-based framework that reconstructs a physically consistent, four–dimensional cloud state using only synchronized ground‐based cameras. Leveraging a homography-guided 2D‐to‐3D transformer, Cloud4D infers the full 3D distribution of liquid water content at 25 m spatial and 5 s temporal resolution. By tracking the 3D liquid water content retrievals over time, Cloud4D additionally estimates horizontal wind vectors. Across a two-month deployment comprising six skyward cameras, our system delivers an order-of-magnitude improvement in space-time resolution relative to state-of-the-art satellite measurements, while retaining single-digit relative error ($<10\\%$) against collocated radar measurements.



Paperid:2954
Authors:Aline Weber, Blossom Metevier, Yuriy Brun, Philip Thomas, Bruno Silva
Abstract:
Machine learning models are often designed to maximize a primary goal, such as accuracy. However, as these models are increasingly used to inform decisions that affect people's lives or well-being, it is often unclear what the real-world repercussions of their deployment might be—making it crucial to understand and manage such repercussions effectively. Models maximizing user engagement on social media platforms, e.g., may inadvertently contribute to the spread of misinformation and content that deepens political polarization. This issue is not limited to social media—it extends to other applications where machine learning-informed decisions can have real-world repercussions, such as education, employment, and lending. Existing methods addressing this issue require prior knowledge or estimates of analytical models describing the relationship between a classifier's predictions and their corresponding repercussions. We introduce Theia, a novel classification algorithm capable of optimizing a primary objective, such as accuracy, while providing high-confidence guarantees about its potential repercussions. Importantly, Theia solves the open problem of providing such guarantees based solely on existing data with observations of previous repercussions. We prove that it satisfies constraints on a model's repercussions with high confidence and that it is guaranteed to identify a solution, if one exists, given sufficient data. We empirically demonstrate, using real-life data, that Theia can identify models that achieve high accuracy while ensuring, with high confidence, that constraints on their repercussions are satisfied.



Paperid:2955
Authors:Heasung Kim, Taekyun Lee, Hyeji Kim, Gustavo De Veciana
Abstract:
Risk-averse modeling is critical in safety-critical and high-stakes applications. Conditional Value-at-Risk (CVaR) quantifies such risk by measuring the expected loss in the tail of the loss distribution, and minimizing it provides a principled framework for training robust models. However, direct CVaR minimization remains challenging due to the difficulty of accurately estimating rare, high-loss events—particularly at extreme quantiles. In this work, we propose a novel training framework that synthesizes informative samples for CVaR optimization using score-based generative models. Specifically, we guide a diffusion-based generative model to sample from a reweighted distribution that emphasizes inputs likely to incur high loss under a pretrained reference model. These samples are then incorporated via a loss-weighted importance sampling scheme to reduce variance in stochastic optimization. We establish convergence guarantees and show that the synthesized, high-loss-emphasized dataset substantially contributes to the noise reduction. Empirically, we validate the effectiveness of our approach across multiple settings, including a real-world wireless channel compression task, where our method achieves significant improvements over standard risk minimization strategies.



Paperid:2956
Authors:Sean Lamont, Christian Walder, Amir Dezfouli, Paul Montague, Michael Norrish
Abstract:
A key challenge in automated formal reasoning is the intractable search space, which grows exponentially with the depth of the proof. This branching is caused by the large number of candidate proof tactics which can be applied to a given goal. Nonetheless, many of these tactics are semantically similar or lead to an execution error, wasting valuable resources in both cases. We address the problem of effectively pruning this search, using only synthetic data generated from previous proof attempts. We first demonstrate that it is possible to generate semantically aware tactic representations which capture the effect on the proving environment, likelihood of success, and execution time. We then propose a novel filtering mechanism which leverages these representations to select semantically diverse and high quality tactics, using Determinantal Point Processes. Our approach, 3D-Prover, is designed to be general, and to augment any underlying tactic generator. We demonstrate the effectiveness of 3D-Prover on the miniF2F and LeanDojo benchmarks by augmenting popular open source proving LLMs. We show that our approach leads to an increase in the overall proof rate, as well as a significant improvement in the tactic success rate, execution time and diversity.



Paperid:2957
Authors:Dylan Sam, Marc Finzi, J. Zico Kolter
Abstract:
Reliably predicting the behavior of language models---such as whether their outputs are correct or have been adversarially manipulated---is a fundamentally challenging task. This is often made even more difficult as frontier language models are offered only through closed-source APIs, providing only black-box access. In this paper, we predict the behavior of black-box language models by asking follow-up questions and taking the probabilities of responsesasrepresentations to train reliable predictors.We first demonstrate that training a linear model on these responses reliably and accurately predicts model correctness on question-answering and reasoning benchmarks. Surprisingly, this caneven outperform white-box linear predictorsthat operate over model internals or activations.Furthermore, we demonstrate that these follow-up question responses can reliably distinguish between a clean version of an LLM and one that has been adversarially influenced via a system prompt to answer questions incorrectly or to introduce bugs into generated code. Finally, we show that they can also be used to differentiate between black-box LLMs, enabling the detection of misrepresented models provided through an API. Overall, our work shows promise in monitoring black-box language model behavior, supporting their deployment in larger, autonomous systems.



Paperid:2958
Authors:Jaesik Yoon, Hyeonseo Cho, Yoshua Bengio, Sungjin Ahn
Abstract:
Diffusion models have recently emerged as a powerful approach for trajectory planning. However, their inherently non-sequential nature limits their effectiveness in long-horizon reasoning tasks at test time. The recently proposed Monte Carlo Tree Diffusion (MCTD) offers a promising solution by combining diffusion with tree-based search, achieving state-of-the-art performance on complex planning problems. Despite its strengths, our analysis shows that MCTD incurs substantial computational overhead due to the sequential nature of tree search and the cost of iterative denoising. To address this, we propose Fast-MCTD, a more efficient variant that preserves the strengths of MCTD while significantly improving its speed and scalability. Fast-MCTD integrates two techniques: Parallel MCTD, which enables parallel rollouts via delayed tree updates and redundancy-aware selection; and Sparse MCTD, which reduces rollout length through trajectory coarsening. Experiments show that Fast-MCTD achieves up to 100× speedup over standard MCTD while maintaining or improving planning performance. Remarkably, it even outperforms Diffuser in inference speed on some tasks, despite Diffuser requiring no search and yielding weaker solutions. These results position Fast-MCTD as a practical and scalable solution for diffusion-based inference-time reasoning. Source code will be made available following the reviewing process at:\url{https://anonymous.4open.science/r/Fast-MCTD-681E}.



Paperid:2959
Authors:Jiayu Zhang, Changbang Li, Yinan Peng, Weihao Luo, Peilai Yu, Xuan Zhang
Abstract:
Instruction fine-tuning (IFT) has emerged as a ubiquitous strategy for specializing large language models (LLMs), yet it implicitly assumes a single, coherent "ground-truth" preference behind all human-written instructions. In practice, annotators differ in the styles, emphases, and granularities they prefer, introducing preference bias that can erode both robustness and generalization. We propose Dynamic Cross-Layer Preference Correction (\textsc{DCPC}), it couples (i) a preference-sensitive similarity estimator that detects mismatched instructional cues, (ii) cross-layer prefix alignment to reconcile semantic representations across transformer layers, and (iii) a lightweight Preference Correction Module (PCM) that dynamically adjusts hidden states to honor the inferred dominant preference. On five Super/GLUE tasks and the Alpaca set—plus six preference-shifted variants—DCPC boosts accuracy/F1-EM by 4.0–6.7 points and gpt-score by +0.7, while cutting inter-seed variance up to 35% on LlaMA-2 13B and Mistral-7B, setting a new state of the art for robust instruction tuning.



Paperid:2960
Authors:Jiahe Li, Jiawei Zhang, Youmin Zhang, Xiao Bai, Jin Zheng, Xiaohan Yu, Lin Gu
Abstract:
Reconstructing accurate surfaces with radiance fields has achieved remarkable progress in recent years. However, prevailing approaches, primarily based on Gaussian Splatting, are increasingly constrained by representational bottlenecks. In this paper, we introduce GeoSVR, an explicit voxel-based framework that explores and extends the under-investigated potential of sparse voxels for achieving accurate, detailed, and complete surface reconstruction. As strengths, sparse voxels support preserving the coverage completeness and geometric clarity, while corresponding challenges also arise from absent scene constraints and locality in surface refinement. To ensure correct scene convergence, we first propose a Voxel-Uncertainty Depth Constraint that maximizes the effect of monocular depth cues while presenting a voxel-oriented uncertainty to avoid quality degradation, enabling effective and robust scene constraints yet preserving highly accurate geometries. Subsequently, Sparse Voxel Surface Regularization is designed to enhance geometric consistency for tiny voxels and facilitate the voxel-based formation of sharp and accurate surfaces. Extensive experiments demonstrate our superior performance compared to existing methods across diverse challenging scenarios, excelling in geometric accuracy, detail preservation, and reconstruction completeness while maintaining high efficiency. Our code will be made open-source upon acceptance.



Paperid:2961
Authors:Matthew Landers, Taylor W. Killian, Hugo Barnes, Tom Hartvigsen, Afsaneh Doryab
Abstract:
Abstract:Offline reinforcement learning in high-dimensional, discrete action spaces is challenging due to the exponential scaling of the joint action space with the number of sub-actions and the complexity of modeling sub-action dependencies. Existing methods either exhaustively evaluate the action space, making them computationally infeasible, or factorize Q-values, failing to represent joint sub-action effects. We propose \textbf{Bra}nch \textbf{V}alue \textbf{E}stimation (BraVE), a value-based method that uses tree-structured action traversal to evaluate a linear number of joint actions while preserving dependency structure. BraVE outperforms prior offline RL methods by up to $20\times$ in environments with over four million actions.



Paperid:2962
Authors:Guangyang Wu, Huayu Zheng, Siqi Luo, Guangtao Zhai, Xiaohong Liu
Abstract:
Animated QR codes present an exciting frontier for dynamic content delivery and digital interaction. However, despite their potential, there has been no prior work focusing on the generation of animated QR codes that are both visually appealing and universally scannable. In this paper, we introduce AnimateQR,the first generative frameworkfor creatinganimated QR codesthat balance aesthetic flexibility with scannability. Unlike previous methods that focus on static QR codes, AnimateQR leverageshierarchical luminance guidanceandprogressive spatiotemporal controlto produce high-quality dynamic QR codes. Our first innovation is a multi-scale hierarchical control signal that adjusts luminance across different spatial scales, ensuring that the QR code remains decodable while allowing for artistic expression. The second innovation is a progressive control mechanism that dynamically adjusts spatiotemporal guidance throughout the diffusion denoising steps, enabling fine-grained balance between visual quality and scannability. Extensive experimental results demonstrate that AnimateQR achieves state-of-the-art performance in both decoding success rates (96\% vs. 56\% baseline) and visual quality (user preference: 7.2 vs. 2.3 on a 10-point scale). Codes will be made public upon acceptance.



Paperid:2963
Authors:Cosmin Bercea, Jun Li, Philipp Raffler, Evamaria O. Riedel, Lena Schmitzer, Angela Kurz, Felix Bitzer, Paula Roßmüller, Julian Canisius, Mirjam Beyrle, Che Liu, Wenjia Bai, Bernhard Kainz, Julia Schnabel, Benedikt Wiestler
Abstract:
Abstract:In many real-world applications, deployed models encounter inputs that differ from the data seen during training. Out-of-distribution detection identifies whether an input stems from an unseen distribution, while open-world recognition flags such inputs to ensure the system remains robust as ever-emerging, previously *unknown* categories appear and must be addressed without retraining.Foundation and vision-language models are pre-trained on large and diverse datasets with the expectation of broad generalization across domains including medical imaging.However, benchmarking these models on test sets with only a few common outlier types silently collapses the evaluation back to a closed-set problem, masking failures on rare or truly novel conditions encountered in clinical use.We therefore present *NOVA*, a challenging, real-life *evaluation-only* benchmark of $\sim$900 brain MRI studies that span 281 rare pathologies and heterogeneous acquisition protocols. Each case includes rich clinical narratives and double-blinded expert bounding-box annotations. Together, these enable joint assessment of anomaly localisation, visual captioning, and diagnostic reasoning. Because NOVA is never used for training, it serves as an \textit{extreme} stress-test of out-of-distribution generalisation: models must bridge a distribution gap both in sample appearance and in semantic space. Baseline results with leading vision-language models (GPT-4o, Gemini 2.0 Flash, and Qwen2.5-VL-72B) reveal substantial performance drops across all tasks, establishing NOVA as a rigorous testbed for advancing models that can detect, localize, and reason about truly unknown anomalies.



Paperid:2964
Authors:Marco Sälzer, Przemysław Wałęga, Martin Lange
Abstract:
In recent years, the expressive power of various neural architectures---including graph neural networks (GNNs), transformers, and recurrent neural networks---has been characterised using tools from logic and formal language theory. As the capabilities of basic architectures are becoming well understood, increasing attention is turning to models that combine multiple architectural paradigms. Among them particularly important, and challenging to analyse, are temporal extensions of GNNs, which integrate both spatial (graph-structure) and temporal (evolution over time) dimensions. In this paper, we initiate the study of logical characterisation of temporal GNNs by connecting them to two-dimensional product logics. We show that the expressive power of temporal GNNs depends on how graph and temporal components are combined. In particular, temporal GNNs that apply static GNNs recursively over time can capture all properties definable in the product logic of (past) propositional temporal logic PTL and the modal logic K. In contrast, architectures such as graph-and-time TGNNs and global TGNNs can only express restricted fragments of this logic, where the interaction between temporal and spatial operators is syntactically constrained. These results yield the first logical characterisations of temporal GNNs and establish new relative expressiveness results for temporal GNNs.



Paperid:2965
Authors:Yuki Imajuku, Kohki Horie, Yoichi Iwata, Kensho Aoki, Naohiro Takahashi, Takuya Akiba
Abstract:
Abstract:How well do AI systems perform in algorithm engineering for hard optimization problems in domains such as package-delivery routing, crew scheduling, factory production planning, and power-grid balancing?We introduce $\textit{ALE-Bench}$, a new benchmark for evaluating AI systems on score-based algorithmic programming contests. Drawing on real tasks from the AtCoder Heuristic Contests, ALE-Bench presents optimization problems that are computationally hard and admit no known exact solution.Unlike short-duration, pass/fail coding benchmarks, ALE-Bench encourages iterative solution refinement over long time horizons.Our software framework supports interactive agent architectures that leverage test-run feedback and visualizations. Our evaluation of frontier LLMs revealed that while they demonstrate high performance on specific problems, a notable gap remains compared to humans in terms of consistency across problems and long-horizon problem-solving capabilities. This highlights the need for this benchmark to foster future AI advancements.



Paperid:2966
Authors:Junjie Xu, Jiahao Zhang, Mangal Prakash, Xiang Zhang, Suhang Wang
Abstract:
Geometric graph neural networks (GNNs) that respect E(3) symmetries have achieved strong performance on small molecule modeling, but they face scalability and expressiveness challenges when applied to large biomolecules such as RNA and proteins. These systems require models that can simultaneously capture fine-grained atomic interactions, long-range dependencies across spatially distant components, and biologically relevant hierarchical structure—such as atoms forming residues, which in turn form higher-order domains. Existing geometric GNNs, which typically operate exclusively in either Euclidean or Spherical Harmonics space, are limited in their ability to capture both the fine-scale atomic details and the long-range, symmetry-aware dependencies required for modeling the multi-scale structure of large biomolecules. We introduce DualEquiNet, a Dual-Space Hierarchical Equivariant Network that constructs complementary representations in both Euclidean and Spherical Harmonics spaces to capture local geometry and global symmetry-aware features. DualEquiNet employs bidirectional cross-space message passing and a novel Cross-Space Interaction Pooling mechanism to hierarchically aggregate atomic features into biologically meaningful units, such as residues, enabling efficient and expressive multi-scale modeling for large biomolecular systems. DualEquiNet achieves state-of-the-art performance on multiple existing benchmarks for RNA property prediction and protein modeling, and outperforms prior methods on two newly introduced 3D structural benchmarks demonstrating its broad effectiveness across a range of large biomolecule modeling tasks.



Paperid:2967
Authors:kaixing yang, Xulong Tang, Ziqiao Peng, Yuxuan Hu, Jun He, Hongyan Liu
Abstract:
Music-driven 3D dance generation has attracted increasing attention in recent years, with promising applications in choreography, virtual reality, and creative content creation. Previous research has generated promising realistic dance movement from audio signals. However, traditional methods underutilize genre conditioning, often treating it as auxiliary modifiers rather than core semantic drivers. This oversight compromises music-motion synchronization and disrupts dance genre continuity, particularly during complex rhythmic transitions, thereby leading to visually unsatisfactory effects. To address the challenge, we propose MEGADance, a novel architecture for music-driven 3D dance generation. By decoupling choreographic consistency into dance generality and genre specificity, MEGADance demonstrates significant dance quality and strong genre controllability. It consists of two stages: (1) High-Fidelity Dance Quantization Stage (HFDQ), which encodes dance motions into a latent representation by Finite Scalar Quantization (FSQ) and reconstructs them with kinematic-dynamic constraints, and (2) Genre-Aware Dance Generation Stage (GADG), which maps music into the latent representation by synergistic utilization of Mixture-of-Experts (MoE) mechanism with Mamba-Transformer hybrid backbone. Extensive experiments on the FineDance and AIST++ dataset demonstrate the state-of-the-art performance of MEGADance both qualitatively and quantitatively. Code will be released upon acceptance.



Paperid:2968
Authors:Dominic Broadbent, Nick Whiteley, Robert Allison, Tom Lovett
Abstract:
Abstract:Existing distribution compression methods, like Kernel Herding (KH), were originally developed for unlabelled data. However, no existing approach directly compresses the conditional distribution of \textit{labelled} data. To address this gap, we first introduce the \textit{Average Maximum Conditional Mean Discrepancy} (AMCMD), a natural metric for comparing conditional distributions, and derive a closed form estimator. Next, we make a key observation: in the context of distribution compression, the cost of constructing a compressed set targeting the AMCMD can be reduced from $\mathcal{O}(n^3)$ to $\mathcal{O}(n)$. Leveraging this, we extend KH to propose Average Conditional Kernel Herding (ACKH), a linear-time greedy algorithm for constructing compressed sets that target the AMCMD. To better understand the advantages of \textit{directly} compressing the conditional distribution rather than doing so via the joint distribution, we introduce \textit{Joint Kernel Herding} (JKH)—a straightforward adaptation of KH designed to compress the joint distribution of labelled data. While herding methods provide a simple and interpretable selection process, they rely on a greedy heuristic. To explore alternative optimisation strategies, we propose \textit{Joint Kernel Inducing Points} (JKIP) and \textit{Average Conditional Kernel Inducing Points} (ACKIP), which \textit{jointly} optimise the compressed set while maintaining linear complexity. Experiments show that directly preserving conditional distributions with ACKIP outperforms both joint distribution compression (via JKH and JKIP) and the greedy selection used in ACKH. Moreover, we see that JKIP consistently outperforms JKH.



Paperid:2969
Authors:Yuxin Chen, Yiran Zhao, Yang Zhang, An Zhang, Kenji Kawaguchi, Shafiq Joty, Junnan Li, Tat-Seng Chua, Michael Qizhe Shieh, Wenxuan Zhang
Abstract:
As large language models (LLMs) continue to advance, their capacity to function effectively across a diverse range of languages has shown marked improvement. Preliminary studies observe that the hidden activations of LLMs often resemble English, even when responding to non-English prompts. This has led to the widespread assumption that LLMs may ``think'' in English. However, more recent results showing strong multilingual performance, even surpassing English performance on specific tasks in other languages, challenge this view. In this work, we find that LLMs progressively develop a core language-agnostic parameter space—a remarkably small subset of parameters whose deactivation results in significant performance degradation across all languages. This compact yet critical set of parameters underlies the model’s ability to generalize beyond individual languages, supporting the emergence of abstract thought that is not tied to any specific linguistic system. Specifically, we identify language-related neurons—those are consistently activated during the processing of particular languages, and categorize them as either shared (active across multiple languages) or exclusive (specific to one). As LLMs undergo continued development over time, we observe a marked increase in both the proportion and functional importance of shared neurons, while exclusive neurons progressively diminish in influence. These shared neurons constitute the backbone of the core language-agnostic parameter space, supporting the emergence of abstract thought. Motivated by these insights, we propose neuron-specific training strategies tailored to LLMs' language-agnostic levels at different development stages. Experiments across diverse LLM families support our approach. Our codes are available at https://anonymous.4open.science/status/S-C393.



Paperid:2970
Authors:Weilin Lin, Nanjun Zhou, Yanyun Wang, Jianze Li, Hui Xiong, Li Liu
Abstract:
Backdoor learning is a critical research topic for understanding the vulnerabilities of deep neural networks. While the diffusion model (DM) has been broadly deployed in public over the past few years, the understanding of its backdoor vulnerability is still in its infancy compared to the extensive studies in discriminative models. Recently, many different backdoor attack and defense methods have been proposed for DMs, but a comprehensive benchmark for backdoor learning on DMs is still lacking. This absence makes it difficult to conduct fair comparisons and thoroughly evaluate existing approaches, thus hindering future research progress. To address this issue, we proposeBackdoorDM, the first comprehensive benchmark designed for backdoor learning on DMs. It comprises nine state-of-the-art (SOTA) attack methods, four SOTA defense strategies, and three useful visualization analysis tools. We first systematically classify and formulate the existing literature in a unified framework, focusing on three different backdoor attack types and five backdoor target types, which are restricted to a single type in discriminative models. Then, we systematically summarize the evaluation metrics for each type and propose a unified backdoor evaluation method based on multimodal large language model (MLLM). Finally, we conduct a comprehensive evaluation and highlight several important conclusions. We believe that BackdoorDM will help overcome current barriers and contribute to building a trustworthy artificial intelligence generated content (AIGC) community. Our code is provided at https://anonymous.4open.science/r/BackdoorDM-3403.



Paperid:2971
Authors:Yibo Li, Miao Xiong, Jiaying Wu, Bryan Hooi
Abstract:
Large Language Models (LLMs) are increasingly deployed in high-stakes domains such as science, law, and healthcare, where accurate expressions of uncertainty are essential for reliability and trust. However, current LLMs are often observed to generate incorrect answers with high confidence—a phenomenon known as "overconfidence". Recent efforts have focused on calibrating LLMs' verbalized confidence: i.e., their expressions of confidence in text form, such as "I am 80% confident that...". Existing approaches either rely on prompt engineering or fine-tuning with heuristically generated uncertainty estimates, both of which have limited effectiveness and generalizability. Motivated by the notion of proper scoring rules for calibration in classical machine learning models, we introduce ConfTuner, a simple and efficient fine-tuning method that introduces minimal overhead and does not require ground-truth confidence scores or proxy confidence estimates. ConfTuner relies on a new loss function, tokenized Brier score, which we theoretically prove to be a proper scoring rule, intuitively meaning that it "correctly incentivizes the model to report its true probability of being correct". ConfTuner improves calibration across diverse reasoning tasks and generalizes to black-box models such as GPT-4o. Our results further show that better-calibrated confidence enables downstream gains in self-correction and model cascade, advancing the development of trustworthy LLM systems.The code is available at https://anonymous.4open.science/r/Uncertainty_ft-2761.



Paperid:2972
Authors:Nikhil Kandpal, Brian Lester, Colin Raffel, Sebastian Majstorovic, Stella Biderman, Baber Abbasi, Luca Soldaini, Enrico Shippole, A. Feder Cooper, Aviya Skowron, Shayne Longpre, Lintang Sutawika, Alon Albalak, Zhenlin Xu, Guilherme Penedo, Loubna Ben allal, Elie Bakouch, John Pressman, Honglu Fan, Dashiell Stander, Guangyu Song, Aaron Gokaslan, John Kirchenbauer, Tom Goldstein, Brian Bartoldson, Bhavya Kailkhura, Tyler Murray
Abstract:
Large language models (LLMs) are typically trained on enormous quantities of unlicensed text, a practice that has led to scrutiny due to possible intellectual property infringement and ethical concerns. Training LLMs on openly licensed text presents a first step towards addressing these issues, but prior data collection efforts have yielded datasets too small or low-quality to produce performant LLMs. To address this gap, we collect, curate, and release the Common Pile v0.1, an eight terabyte collection of openly licensed text designed for LLM pre-training. The Common Pile comprises content from 30 sources that span diverse domains including research papers, code, books, encyclopedias, educational materials, audio transcripts, and more. Crucially, we validate our efforts by training Comma v0.1, a 7 billion parameter LLM trained on 1 trillion tokens of text from the Common Pile. Comma attains competitive performance to LLMs trained on unlicensed text with similar computational budgets, such as LLaMA 7B. In addition to releasing the Common Pile v0.1 itself, we also release the code used in its creation as well as Comma v0.1’s checkpoints and training mixture.



Paperid:2973
Authors:Berkay Döner, Thorir Ingolfsson, Luca Benini, Yawei Li
Abstract:
Abstract:Electroencephalography (EEG) offers a non-invasive lens into human brain activity, but building large‐scale models is hampered by $\textit{topological heterogeneity}$: each public corpus defines its own electrode layout, limiting generalization. We introduce $\textbf{LUNA}$ ($\textbf{L}$atent $\textbf{U}$nified $\textbf{N}$etwork $\textbf{A}$rchitecture), a self-supervised foundation model that reconciles disparate electrode geometries while scaling linearly---not quadratically---with channel count. LUNA compresses multi-channel EEG into a fixed-size, topology-agnostic latent space via learned queries and cross-attention. Downstream transformer blocks then operate exclusively on this latent representation using patch-wise temporal self-attention, decoupling computation from electrode count. Pre-trained on TUEG and Siena ($\>$21,000 h raw EEG across diverse montages) using a masked-patch reconstruction objective, LUNA transfers effectively to four downstream tasks: abnormality detection, artifact rejection, slowing classification, and emotion recognition. It demonstrates highly competitive performance across several benchmarks, achieving state-of-the-art results on TUAR and TUSL, e.g., $\textbf{0.921 AUROC}$ on TUAR, while reducing FLOPs by $\textbf{300}$$\times$ and trimming GPU memory use by up to $\textbf{10}$$\times$. Critically, these gains are consistent across all evaluated electrode configurations. Code and pre-trained models will be released upon publication.



Paperid:2974
Authors:Xuankai Zhang, Junjin Xiao, Qing Zhang
Abstract:
This paper presents a unified framework that allows high-quality dynamic Gaussian Splatting from both defocused and motion-blurred monocular videos. Due to the significant diference between the formation processes of defocus blur and motion blur, existing methods are tailored for either one of them, lacking the ability to simultaneously deal with both of them. Although the two can be jointly modeled as blur kernel based convolution, the inherent difficulty in estimating accurate blur kernels greatly limits the progress in this direction. We in this work go step further towards this direction. Particularly, we propose to estimate per-pixel reliable blur kernels using a blur prediction network that exploits blur-related scene and camera information and is subject to a blur-aware sparsity constraint. Besides, we introduce a dynamic Gaussian densification strategy to mitigate the lack of Gaussians for incomplete regions, and boost the performance of novel view synthesis by incorporating unseen view information to constrain scene optimization. Extensive experiments show that our method outperforms the state-of-the-art methods in generating photorealistic novel view synthesis from defocused and motion-blurred monocular videos. Our code and trained model will be made publicly available.



Paperid:2975
Authors:Xiaofeng Wang, Kang Zhao, Feng Liu, Jiayu Wang, Guosheng Zhao, Xiaoyi Bao, Zheng Zhu, Yingya Zhang
Abstract:
Video generation has emerged as a promising tool for world simulation, leveraging visual data to replicate real-world environments. Within this context, egocentric video generation, which centers on the human perspective, holds significant potential for enhancing applications in virtual reality, augmented reality, and gaming. However, the generation of egocentric videos presents substantial challenges due to the dynamic nature of first-person viewpoints, the intricate diversity of actions, and the complex variety of scenes encountered. Existing datasets are inadequate for addressing these challenges effectively. To bridge this gap, we present EgoVid-5M, the first high-quality dataset specifically curated for egocentric video generation. EgoVid-5M encompasses over 5 million egocentric video clips and is enriched with detailed action annotations, including fine-grained kinematic control and high-level textual descriptions. To ensure the integrity and usability of the dataset, we implement a sophisticated data cleansing pipeline designed to maintain frame consistency, action coherence, and motion smoothness under egocentric conditions. Furthermore, we introduce EgoDreamer, which is capable of generating egocentric videos driven simultaneously by action descriptions and kinematic control signals. The EgoVid-5M dataset, associated action annotations, and all data cleansing metadata will be released for the advancement of research in egocentric video generation.



Paperid:2976
Authors:Fitsum Gaim, Hoyun Song, Huije Lee, Changgeon Ko, Euijun Hwang, Jong Park
Abstract:
Content moderation research has recently made significant advances, but still fails to serve the majority of the world's languages due to the lack of resources, leaving millions of vulnerable users to online hostility. This work presents a large-scale human-annotated multi-task benchmark dataset for abusive language detection in Tigrinya social media with joint annotations for three tasks: abusiveness, sentiment, and topic classification. The dataset comprises 13,717 YouTube comments annotated by nine native speakers, collected from 7,373 videos with a total of over 1.2 billion views across 51 channels. We developed an iterative term clustering approach for effective data selection. Recognizing that around 64% of Tigrinya social media content uses Romanized transliterations rather than native Ge'ez script, our dataset accommodates both writing systems to reflect actual language use. We establish strong baselines across the tasks in the benchmark, while leaving significant challenges for future contributions. Our experiments reveal that small, specialized multi-task models outperform the current frontier models in the low-resource setting, achieving up to 86% accuracy (+7 points) in abusiveness detection.We make the resources publicly available to promote research on online safety.



Paperid:2977
Authors:Jonas Geiping, Sean McLeish, Neel Jain, John Kirchenbauer, Siddharth Singh, Brian Bartoldson, Bhavya Kailkhura, Abhinav Bhatele, Tom Goldstein
Abstract:
We study a novel language model architecture that is capable of scaling test-time computation by implicitly reasoning in latent space. Our model works by iterating a recurrent block, thereby unrolling to arbitrary depth at test-time. This stands in contrast to mainstream reasoning models that scale up compute by producing more tokens. Unlike approaches based on chain-of-thought, our approach does not require any specialized training data, can work with small context windows, and can capture types of reasoning that are not easily represented in words. We train a proof-of-concept model from scratch with 3.5 billion parameters and 800 billion tokens. We show that this model can effortlessly use varying levels of compute, significantly improving with additional compute especially on reasoning tasks, such as math and coding. Further, this architecture naturally reduces compute costs via zero-shot per-token adaptive compute, KV-cache sharing and speculative decoding.



Paperid:2978
Authors:Dongjun Hwang, Yejin Kim, Minyoung Lee, Seong Joon Oh, Junsuk Choe
Abstract:
Open-Vocabulary Segmentation (OVS) aims to segment classes that are not present in the training dataset. However, most existing studies assume that the training data is fixed in advance, overlooking more practical scenarios where new datasets are continuously collected over time. To address this, we first analyze how existing OVS models perform under such conditions. In this context, we explore several approaches such as retraining, fine-tuning, and continual learning but find that each of them has clear limitations. To address these issues, we propose ConOVS, a novel continual learning method based on a Mixture-of-Experts framework. ConOVS dynamically combines expert decoders based on the probability that an input sample belongs to the distribution of each incremental dataset. Through extensive experiments, we show that ConOVS consistently outperforms existing methods across pre-training, incremental, and zero-shot test datasets, effectively expanding the recognition capabilities of OVS models when data is collected sequentially.



Paperid:2979
Authors:JUN WOO PARK, Kyudan Jung, Dohyun Lee, Hyuck Lee, DAEHOON GWAK, ChaeHun Park, Jaegul Choo, Jaewoong Cho
Abstract:
Recent efforts to apply Large Language Models (LLMs) to time-series anomaly detection (TSAD) have yielded limited success, often performing worse than even simple methods. While prior work has focused on downstream performance evaluation, the fundamental question—why do LLMs fail at TSAD?—has remained largely unexplored. In this paper, we present an in-depth analysis that identifies two core challenges in understanding complex temporal dynamics and accurately localizing anomalies. To address these challenges, we propose a simple yet effective method that combines statistical decomposition with index-aware prompting. Our method outperforms 21 existing prompting strategies on the AnomLLM benchmark, achieving up to a 66.6% improvement in F1 score. We further compare LLMs with 16 non-LLM baselines on the TSB-AD benchmark, highlighting scenarios where LLMs offer unique advantages via contextual reasoning. Our findings provide empirical insights into how and when LLMs can be effective for TSAD. We include the experimental code in the supplementary material for reproducibility.



Paperid:2980
Authors:Zi Liang, Qingqing Ye, Xuan Liu, Yanyun Wang, Jianliang Xu, Haibo Hu
Abstract:
Synthetic data refers to artificial samples generated by models. While it has been validated to significantly enhance the performance of large language models (LLMs) during training and has been widely adopted in LLM development, potential security risks it may introduce remain uninvestigated. This paper systematically evaluates the resilience of synthetic-data-integrated training paradigm for LLMs against mainstream poisoning and backdoor attacks. We reveal that such a paradigm exhibits strong resistance to existing attacks, primarily thanks to the different distribution patterns between poisoning data and queries used to generate synthetic samples. To enhance the effectiveness of these attacks and further investigate the security risks introduced by synthetic data, we introduce a novel and universal attack framework, namely, Virus Infection Attack (VIA), which enables the propagation of current attacks through synthetic data even under purely clean queries. Inspired by the principles of virus design in cybersecurity, VIA conceals the poisoning payload within a protective “shell” and strategically searches for optimal hijacking points in benign samples to maximize the likelihood of generating malicious content. Extensive experiments on both data poisoning and backdoor attacks show that VIA significantly increases the presence of poisoning content in synthetic data and correspondingly raises the attack success rate (ASR) on downstream models to levels comparable to those observed in the poisoned upstream models. Our source code is available at: https://anonymous.4open.science/r/VirusInfectionAttack-B82E/.



Paperid:2981
Authors:Reihaneh Zohrabi, Hosein Hasani, Mahdieh Soleymani, Anna Rohrbach, Marcus Rohrbach, Mohammad Hossein Rohban
Abstract:
Out-of-distribution (OOD) detection is crucial for ensuring the reliability and safety of machine learning models in real-world applications, where they frequently face data distributions unseen during training.Despite progress, existing methods are often vulnerable to spurious correlations that mislead models and compromise robustness. To address this, we propose SPROD, a novel prototype-based OOD detection approach that explicitly addresses the challenge posed by unknown spurious correlations.Our post-hoc method refines class prototypes to mitigate bias from spurious features without additional data or hyperparameter tuning, and is broadly applicable across diverse backbones and OOD detection settings.We conduct a comprehensive spurious correlation OOD detection benchmarking, comparing our method against existing approaches and demonstrating its superior performance across challenging OOD datasets, such as CelebA, Waterbirds, UrbanCars, Spurious Imagenet, and the newly introduced Animals MetaCoCo. On average, SPROD improves AUROC by 4.7% and FPR@95 by 9.3% over the second best.



Paperid:2982
Authors:Xinyan Liang, Shuai Li, Qian Guo, Yuhua Qian, Bingbing Jiang, Tingjin Luo, Liang Du
Abstract:
Multi-view classification (MVC) methods based on evolutionary algorithms have been widely recognized in MVC community due to their adaptive mechanisms. Fitness evaluation (FE) that aims to compute the classification performance of each individual from population, is one core step of such methods. Its accuracy directly determines the correctness of the evolutionary direction. However, there are significant differences in the information content among views involved by an individual. The adoption of a joint training strategy will limit the mining of key information in views with low information content in this process. This leads to suboptimal utilization of multi-view data, causing the model not to achieve optimal performance when converging, and thus triggering the fitness evaluation bias (FEB) problem. In view of the crucial role of FE, once FEB occurs, it will lead to the distortion of the entire evolutionary process, and the deviation will gradually accumulate with each iteration. Eventually, it will mislead the evolutionary direction and cause the model to fall into a local optimum. Regrettably, the existing evolutionary multi-view classification methods have not paid attention to this issue. To address the issue, we propose an evolutionary multi-view classification via eliminating individual fitness bias approach (EFB-EMVC) where the FEB problem is addressed by introducing evolutionary navigators for each multi-view model, thereby calibrating the evolutionary direction in real time. Experimental results on nine multi-view datasets demonstrate that this simple yet powerful strategy is effective in addressing the FEB problem and the method armed with this strategy exhibits superior performance compared to other evolutionary multi-view classification methods. (The code will be published.)



Paperid:2983
Authors:Rui Yang, Ziruo Wang, Yuntian Gu, Yitao Liang, Tongyang Li
Abstract:
Quantum computing is an emerging field recognized for the significant speedup it offers over classical computing through quantum algorithms. However, designing and implementing quantum algorithms pose challenges due to the complex nature of quantum mechanics and the necessity for precise control over quantum states. Despite the significant advancements in AI, there has been a lack of datasets specifically tailored for this purpose. In this work, we introduce QCircuitBench, the first benchmark dataset designed to evaluate AI's capability in designing and implementing quantum algorithms in the form of quantum circuit codes. Unlike using AI for writing traditional codes, this task is fundamentally more complicated due to highly flexible design space. Our key contributions include: 1. A general framework which formulates the key features of quantum algorithm design task for Large Language Models.2. Implementation for quantum algorithms from basic primitives to advanced applications, spanning 3 task suites, 23 algorithms, and 128,573 data points.3. Automatic validation and verification functions, allowing for iterative and interactive evaluation without human inspection.4. Promising potential as a training dataset through primitive fine-tuning results.We observed several interesting experimental phenomena: fine-tuning does not always outperform few-shot learning, and LLMs tend to exhibit consistent error patterns. QCircuitBench provides a comprehensive benchmark for AI-driven quantum algorithm design, while also revealing some limitations of LLMs in this domain.



Paperid:2984
Authors:Dylan Sam, Ayan Chakrabarti, Afshin Rostamizadeh, Srikumar Ramalingam, Gui Citovsky, Sanjiv Kumar
Abstract:
Measuring similarity between training examples is critical for curating high-quality and diverse pretraining datasets for language models. However, similarity is typically computed with a generic off-the-shelf embedding model that has been trained for tasks such as retrieval. Whether these embedding-based similarity metrics are well-suited for pretraining data selection remains largely unexplored.In this paper, we propose a new framework to assess the suitability of a similarity metric specifically for data curation in language model pretraining applications. Our framework's first evaluation criterion captures how well distances reflect generalization in pretraining loss between different training examples. Next, we use each embedding model to guide a standard diversity-based data curation algorithm and measure its utility by pretraining a language model on the selected data and evaluating downstream task performance. Finally, we evaluate the capabilities of embeddings to distinguish between examples from different data sources. With these evaluations, we demonstrate that standard off-the-shelf embedding models are not well-suited for the pretraining data curation setting, underperforming even remarkably simple embeddings that are extracted from models trained on the same pretraining corpus. Our experiments are performed on the Pile, for pretraining a 1.7B parameter language model on 200B tokens. We believe our analysis and evaluation framework serves as a foundation for the future design of embeddings that specifically reason about similarity in pretraining datasets.



Paperid:2985
Authors:Yujia Zhang, Xiaoyang Wu, Yixing Lao, Chengyao Wang, Zhuotao Tian, Naiyan Wang, Hengshuang Zhao
Abstract:
Humans learn abstract concepts through multisensory synergy, and once formed, such representations can often be recalled from a single modality. Inspired by this principle, we introduce Concerto, a minimalist simulation of human concept learning for spatial cognition, combining 3D intra-modal self-distillation with 2D-3D cross-modal joint embedding. Despite its simplicity, Concerto learns more coherent and informative spatial features, as demonstrated by zero-shot visualizations. It outperforms both standalone SOTA 2D and 3D self-supervised models by 14.2% and 4.8%, respectively, as well as their feature concatenation, in linear probing for 3D scene perception. With full fine-tuning, Concerto sets new SOTA results across multiple scene understanding benchmarks (e.g., 80.7% mIoU on ScanNet). We further present a variant of Concerto tailored for video-lifted point cloud spatial understanding, and a translator that linearly projects Concerto representations into CLIP’s language space, enabling open-world perception. These results highlight that Concerto emerges spatial representations with superior fine-grained geometric and semantic consistency. Code and weights will be released.



Paperid:2986
Authors:Kangjie Chen, Yingji Zhong, Zhihao Li, Jiaqi Lin, Youyu Chen, Minghan Qin, Haoqian Wang
Abstract:
3D Gaussian Splatting (3DGS) has demonstrated impressive performance in novel view synthesis under dense-view settings. However, in sparse-view scenarios, despite the realistic renderings in training views, 3DGS occasionally manifests appearance artifacts in novel views. This paper investigates the appearance artifacts in sparse-view 3DGS and uncovers a core limitation of current approaches: the optimized Gaussians are overly-entangled with one another to aggressively fit the training views, which leads to a neglect of the real appearance distribution of the underlying scene and results in appearance artifacts in novel views. The analysis is based on a proposed metric, termed Co-Adaptation Score (CA), which quantifies the entanglement among Gaussians, i.e., co-adaptation, by computing the pixel-wise variance across multiple renderings of the same viewpoint, with different random subsets of Gaussians. The analysis reveals that the degree of co-adaptation is naturally alleviated as the number of training views increases. Based on the analysis, we propose two lightweight strategies to explicitly mitigate the co-adaptation in sparse-view 3DGS: (1) random gaussian dropout; (2) multiplicative noise injection to the opacity. Both strategies are designed to be plug-and-play, and their effectiveness is validated across various methods and benchmarks. We hope that our insights into the co-adaptation effect will inspire the community to achieve a more comprehensive understanding of sparse-view 3DGS.



Paperid:2987
Authors:Yuyang Ding, Xinyu Shi, Juntao Li, Xiaobo Liang, Zhaopeng Tu, Min Zhang
Abstract:
Process reward models (PRMs) offer fine-grained, step-level evaluations that facilitate deeper reasoning processes in large language models (LLMs), proving effective in complex tasks like mathematical reasoning. However, developing PRMs is challenging due to the high cost and limited scalability of human-annotated data.Synthetic data from Monte Carlo (MC) estimation is a promising alternative but suffers from a high noise ratio, which can cause overfitting and hinder large-scale training.In this work, we conduct a preliminary study on the noise distribution in synthetic data from MC estimation, identifying that annotation models tend to both underestimate and overestimate step correctness due to limitations in their annotation capabilities. Building on these insights, we propose {\bf S}elf-Denoising Monte {\bf C}arlo {\bf An}notation (\textsc{Scan}), an efficient data synthesis and noise-tolerant learning framework.Our key findings indicate that:(1) Even lightweight models (e.g., 1.5B parameters) can produce high-quality annotations through self-denoising strategy, enabling PRMs to achieve superior performance with only 6\% the inference cost required by vanilla MC estimation.(2) With our robust learning strategy, PRMs can effectively learn from this weak supervision, achieving a 39.2 F1 score improvement (from 19.9 to 59.1) in ProcessBench.Despite using only a compact synthetic dataset, our models surpass strong baselines, including those trained on large-scale human-annotated datasets such as PRM800K.Furthermore, performance continues to improve as we scale up the synthetic data, highlighting the potential of \textsc{Scan} for scalable, cost-efficient, and robust PRM training.



Paperid:2988
Authors:Peng Wang, Xiang Liu, Peidong Liu
Abstract:
Stylizing 3D scenes instantly while maintaining multi-view consistency and faithfully resembling a style image remains a significant challenge. Current state-of-the-art 3D stylization methods typically involve computationally intensive test-time optimization to transfer artistic features into a pretrained 3D representation, often requiring dense posed input images. In contrast, leveraging recent advances in feed-forward reconstruction models, we demonstrate a novel approach to achieve direct 3D stylization in less than a second using unposed sparse-view scene images and an arbitrary style image. To address the inherent decoupling between reconstruction and stylization, we introduce a branched architecture that separates structure modeling and appearance shading, effectively preventing stylistic transfer from distorting the underlying 3D scene structure. Furthermore, we adapt an identity loss to facilitate pre-training our stylization model through the novel view synthesis task. This strategy also allows our model to retain its original reconstruction capabilities while being fine-tuned for stylization. Comprehensive evaluations, using both in-domain and out-of-domain datasets, demonstrate that our approach produces high-quality stylized 3D content that achieve a superior blend of style and scene appearance, while also outperforming existing methods in terms of multi-view consistency and efficiency.



Paperid:2989
Authors:Jialin Chen, Ziyu Zhao, Gaukhar Nurbek, Aosong Feng, Ali Maatouk, Leandros Tassiulas, Yifeng Gao, Rex Ying
Abstract:
The ubiquity of dynamic data in domains such as weather, healthcare, and energy underscores a growing need for effective interpretation and retrieval of time-series data. These data are inherently tied to domain-specific contexts, such as clinical notes or weather narratives, making cross-modal retrieval essential not only for downstream tasks but also for developing robust time-series foundation models by retrieval-augmented generation (RAG). Despite the increasing demand, time-series retrieval remains largely underexplored. Existing methods often lack semantic grounding, struggle to align heterogeneous modalities, and have limited capacity for handling multi-channel signals. To address this gap, we propose TRACE, a generic multimodal retriever that grounds time-series embeddings in aligned textual context. TRACE enables fine-grained channel-level alignment and employs hard negative mining to facilitate semantically meaningful retrieval. It supports flexible cross-modal retrieval modes, including Text-to-Timeseries and Timeseries-to-Text, effectively linking linguistic descriptions with complex temporal patterns. By retrieving semantically relevant pairs, TRACE enriches downstream models with informative context, leading to improved predictive accuracy and interpretability. Beyond a static retrieval engine, TRACE also serves as a powerful standalone encoder, with lightweight task-specific tuning that refines context-aware representations while maintaining strong cross-modal alignment. These representations achieve state-of-the-art performance on downstream forecasting and classification tasks. Extensive experiments across multiple domains highlight its dual utility, as both an effective encoder for downstream applications and a general-purpose retriever to enhance time-series models.



Paperid:2990
Authors:Kiril Vasilev, Alexandre Misrahi, Eeshaan Jain, Phil F Cheng, Petros Liakopoulos, Olivier Michielin, Michael Moor, Charlotte Bunne
Abstract:
Multimodal Large Language Models (LLMs) hold promise for biomedical reasoning, but current benchmarks fail to capture the complexity of real-world clinical workflows. Existing evaluations primarily assess unimodal, decontextualized question-answering, overlooking multi-agent decision-making environments such as Molecular Tumor Boards (MTBs). MTBs bring together diverse experts in oncology, where diagnostic and prognostic tasks require integrating heterogeneous data and evolving insights over time. Current benchmarks lack this longitudinal and multimodal complexity. We introduce CRONOS, an agentic benchmark simulating MTB-style decision-making through clinically challenging, multimodal, and longitudinal oncology questions. Ground truth annotations are validated by clinicians via a co-developed app, ensuring clinical relevance. We benchmark multiple open and closed-source LLMs and show that, even at scale, they lack reliability---frequently hallucinating, struggling with reasoning from time-resolved data, and failing to reconcile conflicting evidence or different modalities. To address these limitations, CRONOS goes beyond benchmarking by providing an agentic framework with foundation model-based tools that enhance multi-modal and longitudinal reasoning, leading to task-level performance gains of up to 6.9% and 5.8%, respectively. Overall, CRONOS offers a challenging and realistic testbed for advancing multimodal LLM reasoning, reliability, and tool-use with a focus on MTB environments in precision oncology.



Paperid:2991
Authors:Youqi WU, Jingwei Zhang, Farzan Farnia
Abstract:
Cross-modal embeddings, such as CLIP, BLIP and their variants, have achieved promising results in aligning representations across modalities. However, these embeddings often underperform compared to state-of-the-art single-modality embeddings on modality-specific tasks. In contrast, single-modality embeddings excel in their domains but lack cross-modal alignment capabilities. In this work, we focus on the problem of unifying cross-modality and single-modality embeddings to achieve the performance of modality-expert embedding within individual modalities while preserving cross-modal alignment. To this end, we propose RP-KrossFuse, a method that leverages a random projection-based Kronecker product to integrate cross-modal embeddings with single-modality embeddings. RP-KrossFuse aims to fuse the sample-pairwise similarity scores of the fused embeddings and operates efficiently in a specified kernel space and supports scalable implementations via random Fourier features for shift-invariant kernels such as the Gaussian kernel. We demonstrate the effectiveness of RP-KrossFuse through several numerical experiments, combining CLIP embeddings with uni-modal image and text embeddings. Our numerical results indicate that RP-KrossFuse achieves competitive modality-specific performance while retaining cross-modal alignment, bridging the gap between cross-modal and single-modality embeddings.



Paperid:2992
Authors:Yixiao Zhou, Xiaoqing Lyu, Hongxiang Lin, Huiying Hu, Tuo Wang
Abstract:
Link sign prediction in signed bipartite graphs, which are extensively utilized across diverse domains such as social networks and recommendation systems, has recently emerged as a pivotal challenge. However, significant space and time complexities associated with the scalability of bipartite graphs pose substantial challenges, particularly in large-scale environments. To address these issues, this paper introduces the SignFlow Bipartite Subgraph Network (SBSN), balancing sublinear training memory growth through a heuristic subgraph extraction method integrated with a novel message passing module, with optimal inference efficiency achieved via the node feature distillation module.Our subgraph sampling approach reduces the graph size by focusing on neighborhoods around target links and employs an optimized directed message passing mechanism to aggregate critical structural patterns. This mechanism allows SBSN to efficiently learn rich local structural patterns essential for accurate sign prediction. Furthermore, to overcome the inefficiency of subgraph sampling-based models during inference, SBSN incorporates a node feature distillation module after the first training stage. This module distills subgraph features into node features, enabling fast inference while retaining the rich structural information of subgraphs.Experiments reveal that SBSN shows superior performance in both medium- and large-scale datasets, efficiently managing memory and computational resources, making it a scalable solution for extensive applications.



Paperid:2993
Authors:Shiji Zhou, Tianbai Yu, Zhi Zhang, Heng Chang, Xiao Zhou, Dong Wu, Han Zhao
Abstract:
Machine unlearning (MU) aims to efficiently remove sensitive or harmful memory from a pre-trained model. The key challenge is to balance the potential tradeoff between unlearning efficacy and utility preservation, which involves forgetting undesirable information as defined while maintaining the model's original performance. One potential way to tackle this problem is to use multi-objective optimization to jointly optimize both the unlearning and utility preservation objectives. However, existing multi-objective methods only guarantee finding a Pareto-optimal solution without fine-grained control, which causes under-optimization of the unlearning objective. To this end, we first model MU as a constrained optimization problem, that is, optimizing the unlearning objective under the constraint of a bounded increase for utility loss. We then show that solving this optimization problem is equivalent to unilateral gradient surgery on the unlearning objective. To resolve the additional computational cost brought by gradient surgery, we propose an implicit gradient surgery method, which approximates the solution to the aforementioned constrained optimization problem via only one backpropagation, thereby achieving efficient utility-preserving MU. Theoretically, we provide a tight convergence analysis of the algorithm. Empirically, our extensive experiments show that the proposed algorithm achieves better tradeoff results than existing baselines.Theoretically, we provide a tight convergence analysis of the algorithm. Empirically, our extensive experiments show that the proposed algorithm achieves better tradeoff results than existing baselines.



Paperid:2994
Authors:Yiwen Tu, Pingbang Hu, Jiaqi Ma
Abstract:
Machine unlearning updates machine learning models to remove information from specific training data samples, complying with data protection regulations that allow individuals to request the removal of their personal data. Despite the recent development of numerous unlearning algorithms, reliable evaluation of these algorithms remains an open research question. In this work, we focus on membership inference attack (MIA) based evaluation, one of the most common approaches for evaluating unlearning algorithms, and address various pitfalls of existing evaluation metrics that lack theoretical understanding and reliability. Specifically, by modeling the proposed evaluation process as a cryptographic game between unlearning algorithms and MIA adversaries, the naturally-induced evaluation metric measures the data removal efficacy of unlearning algorithms and enjoys provable guarantees that existing evaluation metrics fail to satisfy. Furthermore, we propose a practical and efficient approximation of the induced evaluation metric and demonstrate its effectiveness through both theoretical analysis and empirical experiments. Overall, this work presents a novel and reliable approach to empirically evaluating unlearning algorithms, paving the way for the development of more effective unlearning techniques.



Paperid:2995
Authors:CHENGQI LI, Zhihao Shi, Yangdi Lu, Wenbo He, Xiangyu Xu
Abstract:
3D reconstruction from in-the-wild images remains a challenging task due to inconsistent lighting conditions and transient distractors. Existing methods typically rely on heuristic strategies to handle the low-quality training data, which often struggle to produce stable and consistent reconstructions, frequently resulting in visual artifacts. In this work, we propose Asymmetric Dual 3DGS, a novel framework that leverages the stochastic nature of these artifacts: they tend to vary across different training runs due to minor randomness. Specifically, our method trains two 3D Gaussian Splatting (3DGS) models in parallel, enforcing a consistency constraint that encourages convergence on reliable scene geometry while suppressing inconsistent artifacts. To prevent the two models from collapsing into similar failure modes due to confirmation bias, we introduce a divergent masking strategy that applies two complementary masks: a multi-cue adaptive mask and a self-supervised soft mask, which leads to an asymmetric training process of the two models, reducing shared error modes. In addition, to improve the efficiency of model training, we introduce a lightweight variant called Dynamic EMA Proxy, which replaces one of the two models with a dynamically updated Exponential Moving Average (EMA) proxy, and employs an alternating masking strategy to preserve divergence. Extensive experiments on challenging real-world datasets demonstrate that our method consistently outperforms existing approaches while achieving high efficiency. Codes and trained models will be released.



Paperid:2996
Authors:Kareem Amin, Sara Babakniya, Alex Bie, Weiwei Kong, Umar Syed, Sergei Vassilvitskii
Abstract:
Synthetically-generated data plays an increasingly larger role in training large language models. However, while synthetic data has been found to be useful, studies have also shown that without proper curation it can cause LLM performance to plateau, or even “collapse”, after many training iterations. In this paper, we formalize this question and develop a theoretical framework to investigate how much curation is needed in order to ensure that LLM performance continually improves. We find that the requirements are nearly minimal. We describe a training procedure that converges to an optimal LLM even if almost all of the non-synthetic training data is of poor quality. Our analysis is inspired by boosting, a classic machine learning technique that leverages a very weak learning algorithm to produce an arbitrarily good classifier. Our training procedure subsumes many recently proposed methods for training LLMs on synthetic data, and thus our analysis sheds light on why they are successful, and also suggests opportunities for future improvement. We present experiments that validate our theory, and show that dynamically focusing labeling resources on the most challenging examples --- in much the same way that boosting focuses the efforts of the weak learner --- leads to improved performance.



Paperid:2997
Authors:Jian Liu, Jing Xu, Song Guo, Jing Li, Guojingfeng, Jiaao Yu, Haohan Weng, Biwen Lei, Xianghui Yang, Zhuo Chen, Fangqi Zhu, Tao Han, Chunchao Guo
Abstract:
Abstract:Existing pretrained models for 3D mesh generation often suffer from data biases and produce low-quality results, while global reinforcement learning (RL) methods rely on object-level rewards that struggle to capture local structure details. To address these challenges, we present $\textbf{Mesh-RFT}$, a novel fine-grained reinforcement fine-tuning framework that employs Masked Direct Preference Optimization (M-DPO) to enable localized refinement via quality-aware face masking. To facilitate efficient quality evaluation, we introduce an objective topology-aware scoring system to evaluate geometric integrity and topological regularity at both object and face levels through two metrics: Boundary Edge Ratio (BER) and Topology Score (TS). By integrating these metrics into a fine-grained RL strategy, Mesh-RFT becomes the first method to optimize mesh quality at the granularity of individual faces, resolving localized errors while preserving global coherence. Experiment results show that our M-DPO approach reduces Hausdorff Distance (HD) by 24.6\% and improves Topology Score (TS) by 3.8\% over pre-trained models, while outperforming global DPO methods with a 17.4\% HD reduction and 4.9\% TS gain. These results demonstrate Mesh-RFT’s ability to improve geometric integrity and topological regularity, achieving new state-of-the-art performance in production-ready mesh generation.



Paperid:2998
Authors:Raghuveer Thirukovalluru, Rui Meng, Ye Liu, Karthikeyan K, Mingyi Su, Ping Nie, Semih Yavuz, Yingbo Zhou, Wenhu Chen, Bhuwan Dhingra
Abstract:
Contrastive learning (CL) is a prevalent technique for training embedding models, which pulls semantically similar examples (positives) closer in the representation space while pushing dissimilar ones (negatives) further apart. A key source of negatives are 'in-batch' examples, i.e., positives from other examples in the batch. Effectiveness of such models is hence strongly influenced by the size and quality of training batches. In this work, we propose 'Breaking the Batch Barrier' (B3), a novel batch construction strategy designed to curate high-quality batches for CL. Our approach begins by using a pretrained teacher embedding model to rank all examples in the dataset, from which a sparse similarity graph is constructed. A community detection algorithm is then applied to this graph to identify clusters of examples that serve as strong negatives for one another. The clusters are then used to construct batches that are rich in in-batch negatives. Empirical results on the MMEB multimodal embedding benchmark (36 tasks) demonstrate that our method sets a new state of the art, outperforming previous best methods by +1.3 and +2.9 points at the 7B and 2B model scales, respectively. Notably, models trained with \bthm\ surpass existing state-of-the-art results even with a batch size as small as 64, which is 4–16× smaller than that required by other methods.



Paperid:2999
Authors:Haoxiang You, Yilang Liu, Ian Abraham
Abstract:
Abstract:In this work, we propose a computationally efficient algorithm for visual policy learning that leverages differentiable simulation and first-order analytical policy gradients.Our approach decouple the rendering process from the computation graph, enabling seamless integration with existing differentiable simulation ecosystems without the need for specialized differentiable rendering software.This decoupling not only reduces computational and memory overhead but also effectively attenuates the policy gradient norm, leading to more stable and smoother optimization. We evaluate our method on standard visual control benchmarks using modern GPU-accelerated simulation. Experiments show that our approach significantly reduces wall-clock training time and consistently outperforms all baseline methods in terms of final returns.Notably, on complex tasks such as humanoid locomotion, our method achieves a $4\times$ improvement in final return, and successfully learns a humanoid running policy within 4 hours on a single GPU.



Paperid:3000
Authors:Wei Yang, Rui Zhong, Yiqun Chen, Chi Lu, Peng Jiang
Abstract:
Multimodal recommendation aims to integrate collaborative signals with heterogeneous content (e.g., images and text), but remains challenged by modality-specific noise, semantic inconsistency, and unstable propagation over user-item graphs. These issues are often exacerbated by naive fusion or shallow modeling strategies, leading to degraded generalization and poor robustness. While recent work has explored the frequency domain as a lens to disentangle stable versus noisy signals, most methods rely on static filtering or reweighting, lacking the ability to reason over spectral structure or adapt to modality-specific reliability. To address these challenges, we propose a unified framework for structured, frequency-aware representation learning in multimodal recommendation. Our method follows a four-stage pipeline: (i) Decompose graph-based multimodal signals into spectral bands via graph-guided transformations to isolate semantic granularity; (ii) Modulate signal reliability through spectral causal masking, suppressing spurious frequency components; (iii) Fuse complementary frequency cues using hyperspectral reasoning with low-rank cross-band interaction; and (iv) Align modality-specific spectral features via contrastive regularization for semantic and structural consistency. Experiments on three real-world benchmarks demonstrate that our method consistently outperforms strong baselines, especially under sparse and cold-start settings. Further analysis confirms that structured spectral modeling improves the robustness, interpretability, and generalizability of multimodal recommendation.



Paperid:3001
Authors:Thanh-Dat Truong, Christophe Bobda, Nitin Agarwal, Khoa Luu
Abstract:
Multimodal learning has gained much success in recent years. However, current multimodal fusion methods adopt the attention mechanism of Transformers to implicitly learn the underlying correlation of multimodal features. As a result, the multimodal model cannot capture the essential features of each modality, making it difficult to comprehend complex structures and correlations of multimodal inputs. This paper introduces a novel Multimodal Attention-based Normalizing Flow (MANGO) approach\footnote{The source code of this work will be publicly available.} to developing explicit, interpretable, and tractable multimodal fusion learning. In particular, we propose a new Invertible Cross-Attention (ICA) layer to develop the Normalizing Flow-based Model for multimodal data. To efficiently capture the complex, underlying correlations in multimodal data in our proposed invertible cross-attention layer, we propose three new cross-attention mechanisms: Modality-to-Modality Cross-Attention (MMCA), Inter-Modality Cross-Attention (IMCA), and Learnable Inter-Modality Cross-Attention (LICA). Finally, we introduce a new Multimodal Attention-based Normalizing Flow to enable the scalability of our proposed method to high-dimensional multimodal data. Our experimental results on three different multimodal learning tasks, i.e., semantic segmentation, image-to-image translation, and movie genre classification, have illustrated the state-of-the-art (SoTA) performance of the proposed approach.



Paperid:3002
Authors:Fan Chen, Zeyu Jia, Alexander Rakhlin, Tengyang Xie
Abstract:
Abstract:Reinforcement learning with outcome-based feedback faces a fundamental challenge: when rewards are only observed at trajectory endpoints, how do we assign credit to the right actions? This paper provides the first comprehensive analysis of this problem in online RL with general function approximation. We develop a provably sample-efficient algorithm achieving $\widetilde{O}({C_{\rm cov} H^3}/{\varepsilon^2})$ sample complexity, where $C_{\rm cov}$ is the coverability coefficient of the underlying MDP. By leveraging general function approximation, our approach works effectively in large or infinite state spaces where tabular methods fail, requiring only that value functions and reward functions can be represented by appropriate function classes. Our results also characterize when outcome-based feedback is statistically separated from per-step rewards, revealing an unavoidable exponential separation for certain MDPs. For deterministic MDPs, we show how to eliminate the completeness assumption, dramatically simplifying the algorithm. We further extend our approach to preference-based feedback settings, proving that equivalent statistical efficiency can be achieved even under more limited information. Together, these results constitute a theoretical foundation for understanding the statistical properties of outcome-based reinforcement learning.



Paperid:3003
Authors:Yue Zhou, Xinan He, Kaiqing Lin, Bing Fan, Feng Ding, Bin Li
Abstract:
Current AIGC detectors often achieve near-perfect accuracy on images produced by the same generator used for training but struggle to generalize to outputs from unseen generators. We trace this failure in part to latent prior bias: detectors learn shortcuts tied to patterns stemming from the initial noise vector rather than learning robust generative artifacts. To address this, we propose \textbf{On-Manifold Adversarial Training (OMAT)}: by optimizing the initial latent noise of diffusion models under fixed conditioning, we generate \emph{on-manifold} adversarial examples that remain on the generator’s output manifold—unlike pixel-space attacks, which introduce off-manifold perturbations that the generator itself cannot reproduce and that can obscure the true discriminative artifacts. To test against state-of-the-art generative models, we introduce GenImage++, a test-only benchmark of outputs from advanced generators (Flux.1, SD3) with extended prompts and diverse styles. We apply our adversarial-training paradigm to ResNet50 and CLIP baselines and evaluate across existing AIGC forensic benchmarks and recent challenge datasets. Extensive experiments show that adversarially trained detectors significantly improve cross-generator performance without any network redesign. Our findings on latent-prior bias offer valuable insights for future dataset construction and detector evaluation, guiding the development of more robust and generalizable AIGC forensic methodologies.



Paperid:3004
Authors:Ziqiao Ma, Xuweiyi Chen, Shoubin Yu, Sai Bi, Kai Zhang, Ziwen Chen, Sihan Xu, Jianing Yang, Zexiang Xu, Kalyan Sunkavalli, Mohit Bansal, Joyce Chai, Hao Tan
Abstract:
Can we scale 4D pretraining to learn a general space-time representation that reconstructs an object from a few views at some times to any view at any time? We introduce 4D-LRM, the first large-scale 4D reconstruction model that takes input from unconstrained views and timestamps and renders arbitrary novel view-time combinations. Unlike prior 4D approaches---optimization-based, geometry-based, or generative---that struggle with efficiency, generalization, or faithfulness, 4D-LRM learns a unified space-time representation and directly predicts per-pixel 4D Gaussian primitives from posed image tokens across time, enabling fast, high-quality rendering at, in principle, infinite frame rate. 4D-LRM generalizes to novel objects, interpolates across time, and handles diverse camera setups. It reconstructs 24-frame sequences in 1.5 seconds on a single A100 GPU. Our results demonstrate that scaling spatiotemporal pretraining enables accurate and efficient 4D reconstruction.



Paperid:3005
Authors:Yi Wang, Jiaze Wang, Ziyu Guo, Renrui Zhang, Donghao Zhou, Guangyong Chen, Anfeng Liu, Pheng-Ann Heng
Abstract:
Point Cloud Transformers have become a cornerstone in 3D representation for their ability to model long-range dependencies via self-attention. However, these models tend to overemphasize salient regions while neglecting other informative regions, which limits feature diversity and compromises robustness.To address this challenge, we introduce BlindFormer, a novel contrastive attention learning framework that redefines saliency by explicitly incorporating features typically neglected by the model. The proposed Attentional Blindspot Mining (ABM) suppresses highly attended regions during training, thereby guiding the model to explore its own blind spots. This redirection of attention expands the model’s perceptual field and uncovers richer geometric cues.To consolidate these overlooked features, BlindFormer employs Blindspot-Aware Joint Optimization (BJO), a joint learning objective that integrates blindspot feature alignment with the original pretext task. BJO enhances feature discrimination while preserving performance on the primary task, leading to more robust and generalizable representations.We validate BlindFormer on several challenging benchmarks and demonstrate consistent performance gains across multiple Transformer backbones. Notably, it improves Point-MAE by +13.4\% and PointGPT-S by +6.3\% on OBJ-BG under Gaussian noise. These results highlight the importance of mitigating attentional biases in 3D representation learning, revealing BlindFormer’s superior ability to handle perturbations and improve feature discrimination.



Paperid:3006
Authors:Snir Hordan, Maya Bechler-Speicher, Gur Lifshitz, Nadav Dym
Abstract:
Abstract:Spectral features are widely incorporated within Graph Neural Networks (GNNs) to improve their expressive power, or their ability to distinguish among non-isomorphic graphs. One popular example is the usage of graph Laplacian eigenvectors for positional encoding in MPNNs and Graph Transformers. The expressive power of such Spectrally-enhanced GNNs (SGNNs) is usually evaluated via the $k$-WL graph isomorphism test hierarchy and homomorphism counting. Yet, these frameworks align poorly with the graph spectra, yielding limited insight into SGNNs' expressive power. We leverage a well-studied paradigm of classifying graphs by their largest eigenvalue multiplicity to introduce an expressivity hierarchy for SGNNs. We then prove that many SGNNs are incomplete even on graphs with distinct eigenvalues. To mitigate this deficiency, we adapt rotation equivariant neural networks to the graph spectra setting to propose a method to provably improve SGNNs' expressivity on simple spectrum graphs. We empirically verify our theoretical claims via an image classification experiment on the MNIST Superpixel dataset and eigenvector canonicalization on graphs from ZINC.



Paperid:3007
Authors:Simone Carnemolla, Matteo Pennisi, Sarinda Samarasinghe, Giovanni Bellitto, Simone Palazzo, Daniela Giordano, Mubarak Shah, Concetto Spampinato
Abstract:
Understanding and explaining the behavior of machine learning models is essential for building transparent and trustworthy AI systems. We introduce DEXTER, a data-free framework that combines diffusion models and large language models to generate global, textual explanations of visual classifiers. DEXTER operates by optimizing text prompts to synthesize class-conditional images that strongly activate a target classifier. These synthetic samples are then used to elicit detailed natural language reports that describe class-specific decision patterns and biases. Unlike prior work, DEXTER enables natural language reasoning about a classifier's decision process without access to training data or ground-truth labels. We demonstrate DEXTER's flexibility across three tasks—activation maximization, slice discovery and debiasing, and bias explanation—each illustrating its ability to uncover the internal mechanisms of visual classifiers. Quantitative and qualitative evaluations, including a user study, show that DEXTER produces accurate, interpretable outputs. Experiments on ImageNet, Waterbirds, CelebA, and FairFaces confirm that DEXTER outperforms existing approaches in global model explanation and class-level bias reporting.



Paperid:3008
Authors:Jiatao Gu, Tianrong Chen, David Berthelot, Huangjie Zheng, Yuyang Wang, Ruixiang ZHANG, Laurent Dinh, Miguel Angel Bautista, Joshua Susskind, Shuangfei Zhai
Abstract:
We present STARFlow, a scalable generative model based on autoregressive flows (AFs)—a special class of normalizing flows—that achieves strong performance on high-resolution image synthesis. We first establish the theoretical universality of AFs for modeling continuous distributions. Building on this foundation, we introduce a set of architectural and algorithmic innovations that significantly enhance the scalability of normalizing flows: (1) a deep-shallow design where a deep AF block captures most of the model’s capacity, followed by a few shallow AF blocks that are computationally cheap yet contribute non-negligibly, (2) learning in the latent space of pretrained autoencoders, which proves far more effective than modeling pixels directly, and (3) a novel guidance algorithm that substantially improves sample quality. Crucially, our model remains a single, end-to-end normalizing flow, allowing exact maximum likelihood training in continuous space without discretization. \model{} achieves competitive results in both class- and text-conditional image generation, with sample quality approaching that of state-of-the-art diffusion models. To our knowledge, this is the first successful demonstration of normalizing flows at this scale and resolution.



Paperid:3009
Authors:Tue Le, Hoang Tran, Quyen Tran, Linh Ngo, Mehrtash Harandi, Trung Le
Abstract:
Leveraging the power of Large Language Models (LLMs) through preference optimization is crucial for aligning model outputs with human values. Direct Preference Optimization (DPO) has recently emerged as a simple yet effective method by directly optimizing on preference data without the need for explicit reward models. However, DPO typically relies on human-labeled preference data, which can limit its scalability. Self-Play Fine-Tuning (SPIN) addresses this by allowing models to generate their own rejected samples, reducing the dependence on human annotations. Nevertheless, SPIN uniformly applies learning signals across all tokens, ignoring the fine-grained quality variations within responses. As the model improves, rejected samples increasingly contain high-quality tokens, making the uniform treatment of tokens suboptimal. In this paper, we propose SWIFT (Self-Play Weighted Fine-Tuning), a fine-grained self-refinement method that assigns token-level importance weights estimated from a stronger teacher model. Beyond alignment, we also demonstrate that SWIFT serves as an effective knowledge distillation strategy by using the teacher not for logits matching, but for reward-guided token weighting. Extensive experiments on diverse benchmarks and settings demonstrate that SWIFT consistently surpasses both existing alignment approaches and conventional knowledge distillation methods.



Paperid:3010
Authors:Yichao Cai, Yuhang Liu, Erdun Gao, Tianjiao Jiang, Zhen Zhang, Anton van den Hengel, Javen Qinfeng Shi
Abstract:
Multimodal representation learning, exemplified by multimodal contrastive learning (MMCL) using image-text pairs, aims to learn powerful representations by aligning cues across modalities. This approach relies on the core assumption that the exemplar image-text pairs constitute two representations of an identical concept. However, recent research has revealed that real-world datasets often exhibit cross-modal misalignment. There are two distinct viewpoints on how to address this issue: one suggests mitigating the misalignment, and the other leveraging it. We seek here to reconcile these seemingly opposing perspectives, and to provide a practical guide for practitioners. Using latent variable models we thus formalize cross-modal misalignment by introducing two specific mechanisms: Selection bias, where some semantic variables are absent in the text, and perturbation bias, where semantic variables are altered---both leading to misalignment in data pairs. Our theoretical analysis demonstrates that, under mild assumptions, the representations learned by MMCL capture exactly the information related to the subset of the semantic variables invariant to selection and perturbation biases. This provides a unified perspective for understanding misalignment. Based on this, we further offer actionable insights into how misalignment should inform the design of real-world ML systems. We validate our theoretical findings via extensive empirical studies on both synthetic data and real image-text datasets, shedding light on the nuanced impact of cross-modal misalignment on multimodal representation learning.



Paperid:3011
Authors:Vineet Jain, Kusha Sareen, Mohammad Pedramfar, Siamak Ravanbakhsh
Abstract:
Abstract:Adapting a pretrained diffusion model to new objectives at inference time remains an open problem in generative modeling. Existing steering methods suffer from inaccurate value estimation, especially at high noise levels, which biases guidance. Moreover, information from past runs is not reused to improve sample quality, leading to inefficient use of compute. Inspired by the success of Monte Carlo Tree Search, we address these limitations by casting inference-time alignment as a search problem that reuses past computations. We introduce a tree-based approach that _samples_ from the reward-aligned target density by propagating terminal rewards back through the diffusion chain and iteratively refining value estimates with each additional generation. Our proposed method, Diffusion Tree Sampling (DTS), produces asymptotically exact samples from the target distribution in the limit of infinite rollouts, and its greedy variant Diffusion Tree Search (DTS*) performs a robust search for high reward samples. On MNIST and CIFAR-10 class-conditional generation, DTS matches the FID of the best-performing baseline with up to $5\times$ less compute. In text-to-image generation and language completion tasks, DTS* effectively searches for high reward samples that match best-of-N with $2\times$ less compute. By reusing information from previous generations, we get an _anytime algorithm_ that turns additional compute budget into steadily better samples, providing a scalable approach for inference-time alignment of diffusion models.



Paperid:3012
Authors:ao zhou, Jiayi Guan, Li Shen, Fan Lu, Sanqing Qu, Junqiao Zhao, Ziqiao Wang, Ya Wu, Guang Chen
Abstract:
Constrained hybrid-action reinforcement learning (RL) promises to learn a safe policy within a parameterized action space, which is particularly valuable for safety-critical applications involving discrete-continuous hybrid action spaces. However, existing hybrid-action RL algorithms primarily focus on reward maximization, which faces significant challenges for tasks involving both cost constraints and hybrid action spaces. In this work, we propose a novel Constrained Hybrid-action Policy Optimization algorithm (CHPO) to address the problems of constrained hybrid-action RL. Concretely, we rethink the limitations of hybrid-action RL in handling safe tasks with parameterized action spaces and reframe the objective of constrained hybrid-action RL by introducing the concept of Constrained Parameterized-action Markov Decision Process (CPMDP). Subsequently, we present a constrained hybrid-action policy optimization algorithm to confront the constrained hybrid-action problems and conduct theoretical analyses demonstrating that the CHPO converges to the optimal solution while satisfying safety constraints. Finally, extensive experiments demonstrate that the CHPO achieves competitive performance across multiple experimental tasks.



Paperid:3013
Authors:Yizhou Jiang, Tianren Zhang, Yihan Li, Yuqian Liu, Haichuan Gao, Ying Fang, Feng Chen
Abstract:
Spiking Neural Networks often rely on rate coding, where high-precision inference requires long time-steps, leading to significant latency and energy cost—especially for ANN-to-SNN conversions. To address this, we propose Adaptive Fission, a post-training encoding technique that selectively splits high-sensitivity neurons into groups with varying scales and weights. This enables dynamic precision and threshold allocation based on the needs of individual neurons, while incurring minimal spatial overhead. As a generalized form of population coding, it applies broadly to pretrained SNN architectures and does not require additional training or fine-tuning. Experiments on neuromorphic hardware demonstrate up to 80\% reductions in latency and power consumption without sacrificing performance.



Paperid:3014
Authors:Mengdi Liu, Xiaoxue Cheng, Zhangyang Gao, Hong Chang, Cheng Tan, Shiguang Shan, Xilin Chen
Abstract:
Designing protein sequences that fold into a target 3D structure—known as protein inverse folding—is a fundamental challenge in protein engineering. While recent deep learning methods have achieved impressive performance by recovering native sequences, they often overlook the one-to-many nature of the problem: multiple diverse sequences can fold into the same structure. This motivates the need for a generative model capable of designing diverse sequences while preserving structural consistency. To address this trade-off, we introduce ProtInvTree, the first reward-guided tree-search framework for protein inverse folding. ProtInvTree reformulates sequence generation as a deliberate, step-wise decision-making process, enabling the exploration of multiple design paths and exploitation of promising candidates through self-evaluation, lookahead, and backtracking. We propose a two-stage focus-and-grounding action mechanism that decouples position selection and residue generation. To efficiently evaluate intermediate states, we introduce a jumpy denoising strategy that avoids full rollouts. Built upon pretrained protein language models, ProtInvTree supports flexible test-time scaling by adjusting the search depth and breadth without retraining. Empirically, ProtInvTree outperforms state-of-the-art baselines across multiple benchmarks, generating structurally consistent yet diverse sequences, including those far from the native ground truth.



Paperid:3015
Authors:Yeichan Kim, Ilmun Kim, Seyoung Park
Abstract:
Abstract:Transfer learning is a key component of modern machine learning, enhancing the performance of target tasks by leveraging diverse data sources. Simultaneously, overparameterized models such as the minimum-$\ell_2$-norm interpolator (MNI) in high-dimensional linear regression have garnered much attention for their remarkable generalization capabilities, a property termed *benign overfitting*. Despite their individual importance, the intersection of transfer learning and the MNI is largely unexplored. Our research bridges this gap by proposing a novel two-step transfer MNI approach and analyzing its trade-offs. We characterize its non-asymptotic excess risk and identify conditions under which it outperforms the target-only MNI. Our analysis reveals a *free-lunch* covariate shift, where leveraging heterogeneous data offers the benefit of knowledge transfer with limited cost. To operationalize our findings, we develop a data-driven procedure to detect *informative* sources and propose an ensemble method incorporating multiple transfer MNIs. Finite-sample experiments demonstrate the robustness of our methods to model and data heterogeneity, confirming their advantage.



Paperid:3016
Authors:Ziwei Deng, mian Deng, Chenjing Liang, Zeming Gao, Chennan Ma, Chenxing Lin, Haipeng Zhang, Songzhu Mei, Siqi Shen, Cheng Wang
Abstract:
Large Language Models (LLMs) are increasingly being explored across a range of decision-making tasks. However, LLMs sometimes struggle with decision-making tasks under uncertainty that are relatively easy for humans, such as planning actions in stochastic environments. The adoption of LLMs for decision-making is impeded by uncertainty challenges, such as LLM uncertainty and environmental uncertainty. LLM uncertainty arises from the stochastic sampling process inherent to LLMs. Most LLM-based Decision-Making (LDM) approaches address LLM uncertainty through multiple reasoning chains or search trees. However, these approaches overlook environmental uncertainty, which leads to poor performance in environments with stochastic state transitions. Some recent LDM approaches deal with uncertainty by forecasting the probability of unknown variables. However, they are not designed for multi-step decision-making tasks that require interaction with the environment. To address uncertainty in LLM decision-making, we introduce PlanU, an LLM-based planning method that captures uncertainty within Monte Carlo Tree Search (MCTS). PlanU models the return of each node in the MCTS as a quantile distribution, which uses a set of quantiles to represent the return distribution. To balance exploration and exploitation during tree search, PlanU introduces an Upper Confidence Bounds with Curiosity (UCC) score which estimates the uncertainty of MCTS nodes. Through extensive experiments, we demonstrate the effectiveness of PlanU in LLM-based decision-making tasks under uncertainty.



Paperid:3017
Authors:Zhenyu Sun, Ermin Wei
Abstract:
Abstract:In this paper, we consider the general stochastic non-convex optimization problem when the sampling process follows a Markov chain. This problem exhibits its significance in capturing many real-world applications, ranging from asynchronous distributed learning to reinforcement learning. In particular, we consider the worst case where one has no prior knowledge and control of the Markov chain, meaning multiple trajectories cannot be simulated but only a single trajectory is available for algorithm design. We first provide an algorithm-independent lower bound with $\Omega(\epsilon^{-3})$ samples for any first-order methods accessing bounded variance gradient oracles to achieve $\epsilon$-approximate critical solution of the original problem. Then, we propose **Ma**rkov-**C**hain **SPIDER** (MaC-SPIDER), which leverages variance-reduced techniques, to achieve a $\mathcal{O}(\epsilon^{-3})$ upper bound. To the best of our knowledge, MaC-SPIDER is the first to achieve $\mathcal{O}(\epsilon^{-3})$ complexity when sampling from a single Markovian trajectory. And our proposed lower bound concludes its (near) optimality.



Paperid:3018
Authors:Jiarui Jiang, Wei Huang, Miao Zhang, Taiji Suzuki, Liqiang Nie
Abstract:
State-space models (SSMs), particularly Mamba, emerge as an efficient Transformer alternative with linear complexity for long-sequence modeling. Recent empirical works demonstrate Mamba's in-context learning (ICL) capabilities competitive with Transformers, a critical capacity for large foundation models. However, theoretical understanding of Mamba’s ICL remains limited, restricting deeper insights into its underlying mechanisms. Even fundamental tasks such as linear regression ICL, widely studied as a standard theoretical benchmark for Transformers, have not been thoroughly analyzed in the context of Mamba. To address this gap, we study the training dynamics of Mamba on the linear regression ICL task. By developing novel techniques tackling non-convex optimization with gradient descent related to Mamba's structure, we establish an exponential convergence rate to ICL solution, and derive a loss bound that is comparable to Transformer's. Importantly, our results reveal that Mamba can perform a variant of \textit{online gradient descent} to learn the latent function in context. This mechanism is different from that of Transformer, which is typically understood to achieve ICL through gradient descent emulation. The theoretical results are verified by experimental simulation.



Paperid:3019
Authors:Penghao Wu, Shengnan Ma, Bo Wang, Jiaheng Yu, Lewei Lu, Ziwei Liu
Abstract:
Multimodal Large Language Models (MLLMs) have shown great potential in revolutionizing Graphical User Interface (GUI) automation. However, existing GUI models mostly rely on learning from nearly error-free offline trajectories, thus lacking reflection and error recovery capabilities. To bridge this gap, we propose GUI-Reflection, a novel framework that explicitly integrates self-reflection and error correction capabilities into end-to-end multimodal GUI models throughout dedicated training stages: GUI-specific pre-training, offline supervised fine-tuning (SFT), and online reflection tuning. GUI-reflection enables self-reflection behavior emergence with fully automated data generation and learning processes without requiring any human annotation. Specifically, 1) we first propose scalable data pipelines to automatically construct reflection and error correction data from existing successful trajectories. While existing GUI models mainly focus on grounding and UI understanding ability, we propose the GUI-Reflection Task Suite to learn and evaluate reflection-oriented abilities explicitly. 2) Furthermore, we built a diverse and efficient environment for online training and data collection of GUI models on mobile devices. 3) We also present an iterative online reflection tuning algorithm leveraging the proposed environment, enabling the model to continuously enhance its reflection and error correction abilities. Our framework equips GUI agents with self-reflection and correction capabilities, paving the way for more robust, adaptable, and intelligent GUI automation, with all data, models, environments, and tools to be released publicly.



Paperid:3020
Authors:Yi-Chen Li, Zhongxiang Ling, Tao Jiang, Fuxiang Zhang, Pengyuan Wang, Lei Yuan, Zongzhang Zhang, Yang Yu
Abstract:
Learning from multi-agent expert demonstrations, known as Multi-Agent Imitation Learning (MAIL), provides a promising approach to sequential decision-making. However, existing MAIL methods including Behavior Cloning (BC) and Adversarial Imitation Learning (AIL) face significant challenges: BC suffers from the compounding error issue, while the very nature of adversarial optimization makes AIL prone to instability. In this work, we propose \textbf{M}ulti-\textbf{A}gent imitation by learning and sampling from \textbf{F}actor\textbf{I}zed \textbf{S}oft Q-function (MAFIS), a novel method that addresses these limitations for both online and offline MAIL settings. Built upon the single-agent IQ-Learn framework, MAFIS introduces the value decomposition network to factorize the imitation objective at agent level, thus enabling scalable training for multi-agent systems. Moreover, we observe that the soft Q-function implicitly defines the optimal policy as an energy-based model, from which we can sample actions via stochastic gradient Langevin dynamics. This allows us to estimate the gradient of the factorized optimization objective for continuous control tasks, avoiding the adversarial optimization between the soft Q-function and the policy required by prior work. By doing so, we obtain a tractable and \emph{non-adversarial} objective for both discrete and continuous multi-agent control. Experiments on common benchmarks including the discrete control tasks StarCraft Multi-Agent Challenge v2 (SMACv2), Gold Miner, and Multi Particle Environments (MPE), as well as the continuous control task Multi-Agent MuJoCo (MaMuJoCo), demonstrate that MAFIS achieves superior performance compared with baselines.



Paperid:3021
Authors:TAOTAO JING, Tina Chen, Renran Tian, Yaobin Chen, Joshua Domeyer, Heishiro Toyoda, Rini Sherony, Zhengming Ding
Abstract:
Accurately modeling pedestrian intention and understanding driver decision-making processes are critical for the development of safe and socially aware autonomous driving systems. However, existing datasets primarily emphasize observable behavior, offering limited insight into the underlying causal reasoning that informs human interpretation and response during traffic interactions. To address this gap, we introduce PSI, a benchmark dataset that captures the dynamic evolution of pedestrian crossing intentions from the driver’s perspective, enriched with human-annotated textual explanations that reflect the reasoning behind intention estimation and driving decision making. These annotations offer a unique foundation for developing and benchmarking models that combine predictive performance with interpretable and human-aligned reasoning. PSI supports standardized tasks and evaluation protocols across multiple dimensions, including pedestrian intention prediction, driver decision modeling, reasoning generation, and trajectory forecasting and more. By enabling causal and interpretable evaluation, PSI advances research toward autonomous systems that can reason, act, and explain in alignment with human cognitive processes.



Paperid:3022
Authors:Kai Zheng, Jianxin Wang, Jinhui Xu
Abstract:
Graph self-supervised learning (GSL) is essential for processing graph-structured data, reducing the need for manual labeling. Traditionally, this paradigm has extensively utilized Bayesian Personalized Ranking (BPR) as its primary loss function. Despite its widespread application, the theoretical analysis of its node relations evaluation have remained largely unexplored. This paper employs recent advancements in latent hyperbolic geometry to deepen our understanding of node relationships from a graph-theoretical perspective. We analyze BPR’s limitations, particularly its reliance on local connectivity through 2-hop paths, which overlooks global connectivity and the broader topological structure. To address these shortcomings, we purpose a novel loss function, BPR+, designed to encompass even-hop paths and better capture global connectivity and topological nuances. This approach facilitates a more detailed measurement of user-item relationships and improves the granularity of relationship assessments. We validate BPR+ through extensive empirical testing across five real-world datasets and demonstrate its efficacy in refining graph self-supervised learning frameworks. Additionally, we explore the application of BPR+ in drug repositioning, highlighting its potential to support pharmaceutical research and development. Our findings not only illuminate the success factors of previous methodologies but also offer new theoretical insights into this learning paradigm.



Paperid:3023
Authors:Will Merrill, Ashish Sabharwal
Abstract:
Abstract:Recent theoretical results show transformers cannot express sequential reasoning problems over long input lengths, intuitively because their computational *depth* is bounded. However, prior work treats the depth as a constant, leaving it unclear to what degree bounded depth may suffice for solving problems over short inputs, or how increasing the transformer's depth affects its expressive power. We address these questions by analyzing the expressive power of transformers whose depth can grow minimally with context length $n$. We show even highly uniform transformers with depth $\Theta(\log n)$ can express two important problems: *recognizing regular languages*, which captures state tracking abilities and was known to be expressible only by a non-standard non-uniform model of transformers, and *graph connectivity*, which underlies multi-step reasoning. Notably, both of these problems cannot be expressed by fixed-depth transformers under standard complexity conjectures, demonstrating the expressivity benefit of growing depth. Moreover, our theory quantitatively predicts how depth must grow with input length to express these problems, showing that depth scaling is more efficient than scaling width or chain-of-thought steps. Empirically, our detailed experiments designed to bridge the expressivity vs. learnability gap reveal that our theoretical depth requirements for regular language recognition match the practical depth requirements for successfully training transformers remarkably well. Thus, our results clarify precisely how depth affects transformers' reasoning capabilities, providing potential practical insights for designing models that are better at sequential reasoning.



Paperid:3024
Authors:Dexin Duan, Rui Xu, Peilin Liu, Fei Wen
Abstract:
Test-time adaptation (TTA) aims to adapt a source model to a target domain using only test data. Existing methods predominantly rely on unsupervised entropy minimization or its variants, which suffer from degeneration, leading to trivial solutions with low-entropy but inaccurate predictions. In this work, we identifyentropy-deceptive(ED) samples, instances where the model makes highly confident yet incorrect predictions, as the underlying cause of degeneration. Further, we reveal that the gradients of entropy minimization in TTA have an intrinsic low-dimensional structure, driven primarily byentropy-truthful(ET) samples whose gradients are highly correlated. In contrast, ED samples have scattered, less correlated gradients. Leveraging this observation, we show that the detrimental impact of ED samples can be suppressed by constraining model updates within the principal subspace of backward gradients. Building on this insight, we propose LCoTTA, a lifelong continual TTA method that tracks the principal subspace of gradients online and utilizes their projections onto this subspace for adaptation. Further, we provide theoretical analysis to show that the proposed subspace-based method can enhance the robustness against detrimental ED samples. Extensive experiments demonstrate that LCoTTA effectively overcomes degeneration and significantly outperforms existing methods in long-term continual adaptation scenarios. Code is available online.



Paperid:3025
Authors:Shiyuan Yin, Chenjia Bai, Zihao Zhang, Junwei Jin, Xinxin Zhang, Chi Zhang, Xuelong Li
Abstract:
Large language models (LLMs) demonstrate advanced reasoning abilities, enabling robots to understand natural language instructions and generate high-level plans with appropriate grounding. However, LLM hallucinations present a significant challenge, often leading to overconfident yet potentially misaligned or unsafe plans. While researchers have explored uncertainty estimation to improve the reliability of LLM-based planning, existing studies have not sufficiently differentiated between epistemic and intrinsic uncertainty, limiting the effectiveness of uncertainty estimation.In this paper, we present Combined Uncertainty estimation for Reliable Embodied planning (CURE), which decomposes the uncertainty into epistemic and intrinsic uncertainty, each estimated separately. Furthermore, epistemic uncertainty is subdivided into task clarity and task familiarity for more accurate evaluation. The overall uncertainty assessments are obtained using random network distillation and multi-layer perceptron regression heads driven by LLM features. We validated our approach in two distinct experimental settings: kitchen manipulation and tabletop rearrangement experiments. The results show that, compared to existing methods, our approach yields uncertainty estimates that are more closely aligned with the actual execution outcomes.



Paperid:3026
Authors:Rongyao Fang, Chengqi Duan, Kun Wang, Linjiang Huang, Hao Li, Hao Tian, Shilin Yan, Weihao Yu, Xingyu Zeng, Jifeng Dai, Xihui Liu, Hongsheng Li
Abstract:
Current image generation and editing methods primarily process textual prompts as direct inputs without explicit reasoning about visual composition or operational steps. We present Generation Chain-of-Thought (GoT), a novel paradigm that empowers a Multimodal Large Language Model (MLLM) to first generate an explicit, structured reasoning chain in natural language—detailing semantic relationships, object attributes, and, crucially, precise spatial coordinates—before any image synthesis occurs. This intermediate reasoning output directly guides the subsequent visual generation or editing process. This approach transforms conventional text-to-image generation and editing into a reasoning-guided framework that analyzes semantic relationships and spatial arrangements. We define the formulation of GoT and construct large-scale GoT datasets containing over \textbf{9M} samples with detailed reasoning chains capturing semantic-spatial relationships. To leverage the advantages of GoT, we implement a unified framework that integrates Qwen2.5-VL for reasoning chain generation with an end-to-end diffusion model enhanced by our novel Semantic-Spatial Guidance Module. Experiments show our GoT framework achieves excellent performance on both generation and editing tasks, with significant improvements over baselines. Additionally, our approach enables interactive visual generation, allowing users to explicitly modify reasoning steps for precise image adjustments. GoT pioneers a new direction for reasoning-driven visual generation and editing, producing images that better align with human intent. We will release our datasets and models to facilitate future research.



Paperid:3027
Authors:Charles London, Varun Kanade
Abstract:
Abstract:Pause tokens, simple filler symbols such as "...", consistently improve Transformer performance on both language and mathematical tasks, yet their theoretical effect remains unexplained. We provide the first formal separation result, proving that adding pause tokens to constant-depth, logarithmic-width Transformers strictly increases their computational expressivity. With bounded-precision activations, Transformers without pause tokens compute only a strict subset of $\mathsf{AC}^0$ functions, while adding a polynomial number of pause tokens enables expressing the complete class. For logarithmic-precision Transformers, we show that adding pause tokens achieves expressivity equivalent to $\mathsf{TC}^0$, matching known upper bounds. Empirically, we demonstrate that two‑layer causally masked Transformers can learn parity when supplied with pause tokens, a function that they appear unable to learn without them. Our results provide a rigorous theoretical explanation for prior empirical findings, clarify how pause tokens interact with width, depth, and numeric precision, and position them as a distinct mechanism, complementary to chain-of-thought prompting, for enhancing Transformer reasoning.



Paperid:3028
Authors:Anna Romero, Craig Russell, Alexander Krull, Virginie Uhlmann
Abstract:
The shape of objects is an important source of visual information in a wide range of applications. One of the core challenges of shape quantification is to ensure that the extracted measurements remain invariant to transformations that preserve an object’s intrinsic geometry, such as changing its size, orientation, and position in the image. In this work, we introduce ShapeEmbed, a self-supervised representation learning framework designed to encode the contour of objects in 2D images, represented as a Euclidean distance matrix, into a shape descriptor that is invariant to translation, scaling, rotation, reflection, and point indexing. Our approach overcomes the limitations of traditional shape descriptors while improving upon existing state-of-the-art autoencoder-based approaches. We demonstrate that the descriptors learned by our framework outperform their competitors in shape classification tasks on natural and biological images. We envision our approach to be of particular relevance to biological imaging applications.



Paperid:3029
Authors:Adel Javanmard, Rudrajit Das, Alessandro Epasto, Vahab Mirrokni
Abstract:
Retraining a model using its own predictions together with the original, potentially noisy labels is a well-known strategy for improving the model’s performance. While prior works have demonstrated the benefits of specific heuristic retraining schemes, the question of how to optimally combine the model's predictions and the provided labels remains largely open. This paper addresses this fundamental question for binary classification tasks. We develop a principled framework based on approximate message passing (AMP) to analyze iterative retraining procedures for two ground truth settings: Gaussian mixture model (GMM) and generalized linear model (GLM). Our main contribution is the derivation of the Bayes optimal aggregator function to combine the current model's predictions and the given labels, which when used to retrain the same model, minimizes its prediction error. We also quantify the performance of this optimal retraining strategy over multiple rounds. We complement our theoretical results by proposing a practically usable version of the theoretically-optimal aggregator function for linear probing with the cross-entropy loss, and demonstrate its superiority over baseline methods in the high label noise regime.



Paperid:3030
Authors:Antoine Collas, Ce Ju, Nicolas Salvy, Bertrand Thirion
Abstract:
Abstract:Generating realistic brain connectivity matrices is key to analyzing population heterogeneity in brain organization, understanding disease, and augmenting data in challenging classification problems. Functional connectivity matrices lie in constrained spaces—such as the set of symmetric positive definite or correlation matrices—that can be modeled as Riemannian manifolds. However, using Riemannian tools typically requires redefining core operations (geodesics, norms, integration), making generative modeling computationally inefficient. In this work, we propose DiffeoCFM, an approach that enables conditional flow matching (CFM) on matrix manifolds by exploiting pullback metrics induced by global diffeomorphisms on Euclidean spaces. We show that Riemannian CFM with such metrics is equivalent to applying standard CFM after data transformation. This equivalence allows efficient vector field learning, and fast sampling with standard ODE solvers. We instantiate DiffeoCFM with two different settings: the matrix logarithm for covariance matrices and the normalized Cholesky decomposition for correlation matrices. We evaluate DiffeoCFM on three large-scale fMRI datasets with more than $4600$ scans from $2800$ subjects (ADNI, ABIDE, OASIS‑3) and two EEG motor imagery datasets with over $30000$ trials from $26$ subjects (BNCI2014‑002 and BNCI2015‑001). It enables fast training and achieves state-of-the-art performance, all while preserving manifold constraints.



Paperid:3031
Authors:Jianyang Gu, Sam Stevens, Elizabeth Campolongo, Matthew Thompson, Net Zhang, Jiaman Wu, Andrei Kopanev, Zheda Mai, Alexander White, James Balhoff, Wasila Dahdul, Daniel Rubenstein, Hilmar Lapp, Tanya Berger-Wolf, Wei-Lun (Harry) Chao, Yu Su
Abstract:
Foundation models trained at scale exhibit remarkable emergent behaviors, learning new capabilities beyond their initial training objectives. We find such emergent behaviors in biological vision models via large-scale contrastive vision-language training. To achieve this, we first curate TreeOfLife-200M, comprising 214 million images of living organisms, the largest and most diverse biological organism image dataset to date. We then train BioCLIP-XL on TreeOfLife-200M to distinguish different species. Despite the narrow training objective, BioCLIP-XL yields extraordinary accuracy when applied to various biological visual tasks such as habitat classification and trait prediction. We identify emergent properties in the learned embedding space of BioCLIP-XL. At the inter-species level, the embedding distribution of different species aligns closely with functional and ecological meanings (e.g., beak sizes and habitats). At the intra-species level, instead of being diminished, the intra-species variations (e.g., life stages and sexes) are preserved and better separated in subspaces orthogonal to inter-species distinctions. We provide formal proof and analyses to explain why hierarchical supervision and contrastive objectives encourage these emergent properties. Crucially, our results reveal that these properties become increasingly significant with larger-scale training data, leading to a biologically meaningful embedding space.



Paperid:3032
Authors:Hong Wang, Yixuan Jiang, Jie Wang, Xinyi Li, Jian Luo, huanshuo dong
Abstract:
Operator eigenvalue problems play a critical role in various scientific fields and engineering applications, yet numerical methods are hindered by the curse of dimensionality. Recent deep learning methods provide an efficient approach to address this challenge by iteratively updating neural networks. These methods' performance relies heavily on the spectral distribution of the given operator: larger gaps between the operator's eigenvalues will improve precision, thus tailored spectral transformations that leverage the spectral distribution can enhance their performance. Based on this observation, we propose theSpectralTransformationNetwork (STNet).During each iteration, STNet uses approximate eigenvalues and eigenfunctions to perform spectral transformations on the original operator, turning it into an equivalent but easier problem.Specifically, we employ deflation projection to exclude the subspace corresponding to already solved eigenfunctions, thereby reducing the search space and avoiding converging to existing eigenfunctions. Additionally, our filter transform magnifies eigenvalues in the desired region and suppresses those outside, further improving performance.Extensive experiments demonstrate that STNet consistently outperforms existing learning-based methods, achieving state-of-the-art performance in accuracy.



Paperid:3033
Authors:Chieh Lin, Zhaoyang Lv, Songyin Wu, Zhen Xu, Thu Nguyen-Phuoc, Hung-Yu Tseng, Julian Straub, Numair Khan, Lei Xiao, Ming-Hsuan Yang, Yuheng Ren, Richard Newcombe, Zhao Dong, Zhengqin Li
Abstract:
We introduce the Deformable Gaussian Splats Large Reconstruction Model (DGS-LRM), the first feed-forward method predicting deformable 3D Gaussian splats from a monocular posed video of any dynamic scene. Feed-forward scene reconstruction has gained significant attention for its ability to rapidly create digital replicas of real-world environments. However, most existing models are limited to static scenes and fail to reconstruct the motion of moving objects. Developing a feed-forward model for dynamic scene reconstruction poses significant challenges, including the scarcity of training data and the need for appropriate 3D representations and training paradigms. To address these challenges, we introduce several key technical contributions: an enhanced large-scale synthetic dataset with ground-truth multi-view videos and dense 3D scene flow supervision; a per-pixel deformable 3D Gaussian representation that is easy to learn, supports high-quality dynamic view synthesis, and enables long-range 3D tracking; and a large transformer network that achieves real-time, generalizable dynamic scene reconstruction. Extensive qualitative and quantitative experiments demonstrate that DGS-LRM achieves dynamic scene reconstruction quality comparable to optimization-based methods, while significantly outperforming the state-of-the-art predictive dynamic reconstruction method on real-world examples. Its predicted physically grounded 3D deformation is accurate and can be readily adapted for long-range 3D tracking tasks, achieving performance on par with state-of-the-art monocular video 3D tracking methods.



Paperid:3034
Authors:Sergey Samsonov, Marina Sheshukova, Eric Moulines, Alexey Naumov
Abstract:
Abstract:In this paper we derive non-asymptotic Berry–Esseen bounds for Polyak–Ruppert averaged iterates of the Linear Stochastic Approximation (LSA) algorithm driven by the Markovian noise. Our analysis yields $O(n^{-1/4})$ convergence rates to the Gaussian limit in the Kolmogorov distance. We further establish the non-asymptotic validity of a multiplier block bootstrap procedure for constructing the confidence intervals, guaranteeing consistent inference under Markovian sampling. Our work provides the first non-asymptotic guarantees on the rate of convergence of bootstrap-based confidence intervals for stochastic approximation with Markov noise. Moreover, we recover the classical rate of order $\mathcal{O}(n^{-1/8})$ up to logarithmic factors for estimating the asymptotic variance of the iterates of the LSA algorithm.



Paperid:3035
Authors:Runyu Zhang, Arvind Raghunathan, Jeff Shamma, Na Li
Abstract:
Tools from control and dynamical systems have proven valuable for analyzing and developing optimization methods. In this paper, we establish rigorous theoretical foundations for using feedback linearization—a well-established nonlinear control technique—to solve constrained optimization problems. For equality-constrained optimization, we establish global convergence rates to first-order Karush-Kuhn-Tucker (KKT) points and uncover the close connection between the FL method and the Sequential Quadratic Programming (SQP) algorithm. Building on this relationship, we extend the FL approach to handle inequality-constrained problems. Furthermore, we introduce a momentum-accelerated feedback linearization algorithm and provide a rigorous convergence guarantee.



Paperid:3036
Authors:Bohang Sun, Yuena Lin, Tao Yang, Zhen Zhu, Zhen Yang, Gengyu Lyu
Abstract:
The Unaligned Multi-view Clustering (UMC) aims to learn a discriminative cluster structure from unaligned multi-view data, where the features of samples are not completely aligned across multiple views. Most of current methods usually prioritize employing various alignment strategies to align different sample representations and then conduct cross-view fusion for subsequent clustering. However,due to the heterogeneity of representations across different views, these alignment strategies often fail to achieve ideal view-alignment results, inevitably leading to unreliable cross-view fusion.To address this issue, we propose an alignment-free consistency fusion framework named AF-UMC, which bypasses the traditional view-alignment operation and directly extracts consistent representation from each view to perform global cross-view consistency fusion. Specifically, we first construct a cross-view consistent basis space by a cross-view reconstruction loss and a designed Structural Clarity Regularization (SCR), where autoencoders extract consistent representation of each view through projecting view-specific data to the constructed basis space. Afterwards, these extracted representations are globally pulled together for further cross-view fusion according to a designed Instance Global Contrastive Enhancement (IGCE), which makes the fused consistent representation with higher global consistency. Compared with previous methods, AF-UMC directly extracts consistent representation from each view for global fusion rather than alignment for fusion, which significantly mitigates the degraded performance caused by undesired view-alignment results while greatly reduces algorithm complexity and enhances its efficiency. Extensive experiments on various datasets demonstrate that our proposed method exhibits superior performance against other state-of-the-art methods.



Paperid:3037
Authors:Mintong Kang, Zhaorun Chen, Bo Li
Abstract:
Despite the remarkable capabilities of large language models (LLMs) across diverse applications, they remain vulnerable to generating content that violates safety regulations and policies. To mitigate these risks, LLMs undergo safety alignment; however, they can still be effectively jailbroken. Off-the-shelf guardrail models are commonly deployed to monitor generations, but these models primarily focus on detection rather than ensuring safe decoding of LLM outputs. Moreover, existing efforts lack rigorous safety guarantees, which are crucial for the universal deployment of LLMs and certifiable compliance with regulatory standards. In this paper, we propose a Claim-based Stream Decoding (CSD) algorithm coupled with a statistical risk guarantee framework using conformal analysis. Specifically, our CSD algorithm integrates a stream guardrail model to safeguard sequential claims generated by LLMs and incorporates a backtracking mechanism to revise claims flagged with high safety risks. We provide theoretical guarantees demonstrating that the CSD algorithm achieves the desired generation distribution subject to safety constraints. Furthermore, we introduce a generation risk certification framework and derive a high-probability upper bound on the safety risk of the proposed CSD algorithm. We extend our approach to online settings, where user queries arrive sequentially, and prove that our method can asymptotically control safety risk to any desired level. Empirical evaluations demonstrate the effectiveness and efficiency of the CSD algorithm compared to state-of-the-art safety decoding approaches. Additionally, we validate the soundness and tightness of the derived safety risk upper bound using realistic data in both offline and online scenarios.



Paperid:3038
Authors:Yue Wang, Qiuzhi Liu, Jiahao Xu, Tian Liang, Xingyu Chen, Zhiwei He, Linfeng Song, Dian Yu, Juntao Li, Zhuosheng Zhang, Rui Wang, Zhaopeng Tu, Haitao Mi, Dong Yu
Abstract:
Long reasoning models (LRMs) such as OpenAI's o1 and DeepSeek's R1 have demonstrated remarkable abilities in complex reasoning tasks by scaling test-time compute and exhibiting human-like deep thinking. However, we identify a phenomenon we term underthinking, where LRMs frequently switch between different reasoning thoughts without sufficiently exploring promising paths to reach a correct solution. This behavior leads to inadequate depth of reasoning and decreased performance, particularly on challenging mathematical problems. To systematically analyze this issue, we conduct experiments on three challenging test sets and two representative open-source LRMs, revealing that frequent thought switching correlates with incorrect responses. We introduce a novel metric to quantify underthinking by measuring token efficiency in incorrect answers. To address underthinking, we propose a decoding strategy with thought switching penalty (Tip) that discourages premature transitions between thoughts, encouraging deeper exploration of each reasoning path. Experimental results demonstrate that our approach improves accuracy across challenging datasets without requiring model fine-tuning. Our findings contribute to understanding reasoning inefficiencies in LRMs and offer a practical solution to enhance their problem-solving capabilities.



Paperid:3039
Authors:Wenhui Tan, Jiaze Li, Jianzhong Ju, Zhenbo Luo, Jian Luan, Ruihua Song
Abstract:
Abstract:Large Language Models (LLMs) achieve superior performance through Chain-of-Thought (CoT) reasoning, but these token-level reasoning chains are computationally expensive and inefficient.In this paper, we introduce Compressed Latent Reasoning (CoLaR), a novel framework that dynamically compresses reasoning processes in latent space through a two-stage training approach.First, during supervised fine-tuning, CoLaR extends beyond next-token prediction by incorporating an auxiliary next compressed embedding prediction objective. This process merges embeddings of consecutive tokens using a compression factor $c$ randomly sampled from a predefined range, and trains a specialized latent head to predict distributions of subsequent compressed embeddings. Second, we enhance CoLaR through reinforcement learning (RL) that leverages the latent head's non-deterministic nature to explore diverse reasoning paths and exploit more compact ones.This approach enables CoLaR to: i) **perform reasoning at a dense latent level** (i.e., silently), substantially reducing reasoning chain length, and ii) **dynamically adjust reasoning speed** at inference time by simply prompting the desired compression factor.Extensive experiments across four mathematical reasoning datasets demonstrate that CoLaR achieves 14.1% higher accuracy than latent-based baseline methods at comparable compression ratios, and reduces reasoning chain length by 53.3% with only 4.8% performance degradation compared to explicit CoT method. Moreover, when applied to more challenging mathematical reasoning tasks, our RL-enhanced CoLaR demonstrates performance gains of up to 5.4% while dramatically reducing latent reasoning chain length by 82.8%.The code and models will be released upon acceptance.



Paperid:3040
Authors:Jeonghyeon Na, Sangwon Baik, Inhee Lee, Junyoung Lee, Hanbyul Joo
Abstract:
The way humans interact with each other, including interpersonal distances, spatial configuration, and motion, varies significantly across different situations. To enable machines to understand such complex, context-dependent behaviors, it is essential to model multiple people in relation to the surrounding scene context.In this paper, we present a novel research problem to model the correlations between two people engaged in a shared interaction scenario defined by an object. We refer to this formulation as Human-Human-Object Interactions (HHOIs).To overcome the lack of dedicated datasets for HHOIs, we present a newly captured HHOIs dataset and a method to synthesize HHOI data by leveraging image generative models. Using these data sources, we train a score-based diffusion model to integrate the two modalities into a unified generative framework capable of synthesizing complete HHOIs in a single sampling process.Experimental results show that our method generates realistic multi-person scenes conditioned on both object and textual descriptions, outperforming previous approaches that focus only on single-person HOIs.Finally, we present an extension toward text-guided generation of multi-person interactions beyond dyads.



Paperid:3041
Authors:Awni Altabaa, Omar Montasser, John Lafferty
Abstract:
Chain-of-thought (CoT) supervision has emerged as a powerful empirical technique in large language models. This paper develops a statistical theory of learning under CoT supervision. Central to the theory is a quantity called the CoT information measure, which quantifies the additional discriminative power gained from the chain-of-thought for distinguishing hypotheses with different end-to-end behaviors. The main theoretical results demonstrate how CoT supervision can yield significantly faster learning rates compared to standard end-to-end supervision, with both upper bounds and information-theoretic lower bounds on sample complexity in terms of the CoT information.



Paperid:3042
Authors:Veronica Lachi, Francesco Ferrini, Antonio Longa, Bruno Lepri, Andrea Passerini, Manfred Jaeger
Abstract:
Abstract:Graph Neural Networks (GNNs) are widely used to compute representations of node pairs for downstream tasks such as link prediction. Yet, theoretical understanding of their expressive power has focused almost entirely on graph-level representations. In this work, we shift the focus to links and provide the first comprehensive study of GNN expressiveness in link representation. We introduce a unifying framework, the $k_\phi$-$k_\rho$-$m$ framework, that subsumes existing message-passing link models and enables formal expressiveness comparisons. Using this framework, we derive a hierarchy of state-of-the-art methods and offer theoretical tools to analyze future architectures. To complement our analysis, we propose a synthetic evaluation protocol comprising the first benchmark specifically designed to assess link-level expressiveness. Finally, we ask: does expressiveness matter in practice? We use a graph symmetry metric that quantifies the difficulty of distinguishing links and show that while expressive models may underperform on standard benchmarks, they significantly outperform simpler ones as symmetry increases, highlighting the need for dataset-aware model selection.



Paperid:3043
Authors:Yicong Jiang, Zheng Tracy Ke
Abstract:
Vertex hunting (VH) is the task of estimating a simplex from noisy data points and has many applications in areas such as network and text analysis. We introduce a new variant, semi-supervised vertex hunting (SSVH), in which partial information is available in the form of barycentric coordinates for some data points, known only up to an unknown transformation. To address this problem, we develop a method that leverages properties of orthogonal projection matrices, drawing on novel insights from linear algebra. We establish theoretical error bounds for our method and demonstrate that it achieves a faster convergence rate than existing unsupervised VH algorithms. Finally, we apply SSVH to two practical settings---semi-supervised network mixed membership estimation and semi-supervised topic modeling---resulting in efficient and scalable algorithms.



Paperid:3044
Authors:Longquan Dai, Wu Ming, Dejiao Xue, wang he, Jinhui Tang
Abstract:
A major challenge in using diffusion models is aligning outputs with user-defined conditions. Existing conditional generation methods fall into two major categories: classifier-based guidance, which requires differentiable target models and gradient-based correction; and classifier-free guidance, which embeds conditions directly into the diffusion model but demands expensive joint training and architectural coupling. In this work, we introduce a third paradigm: DISCrete nOise (DISCO) guidance, which replaces the continuous conditional correction term with a finite codebook of discrete noise vectors sampled from a Gaussian prior. Conditional generation is reformulated as a code selection task, and we train prediction network to choose the optimal code given the intermediate diffusion state and the conditioning input. Our approach is differentiability-free, and training-efficient, avoiding the gradient computation and architectural redundancy of prior methods. Empirical results demonstrate that DISCO achieves competitive controllability while substantially reducing resource demands, positioning it as a scalable and effective alternative for conditional diffusion generation.



Paperid:3045
Authors:Xiang Lan, Feng Wu, Kai He, Qinghao Zhao, Shenda Hong, Mengling Feng
Abstract:
Abstract:While recent multimodal large language models (MLLMs) have advanced automated ECG interpretation, they still face two key limitations: (1) insufficient multimodal synergy between ECG time series signals and ECG images, and (2) limited explainability in linking diagnoses to granular waveform evidence. We introduce GEM, the first MLLM unifying ECG time series, 12-lead ECG images and text for grounded and clinician-aligned ECG interpretation. GEM enables feature-grounded analysis, evidence-driven reasoning, and a clinician-like diagnostic process through three core innovations: a dual-encoder framework extracting complementary time series and image features, cross-modal alignment for effective multimodal understanding, and knowledge-guided instruction data generation for generating high-granularity grounding data (ECG-Grounding) linking diagnoses to measurable parameters ($e.g.$, QRS/PR Intervals). Additionally, we propose the Grounded ECG Understanding task, a clinically motivated benchmark designed to comprehensively assess the MLLM's capability in grounded ECG understanding. Experimental results on both existing and our proposed benchmarks show GEM significantly improves predictive performance (CSN $7.4\%$ $\uparrow$), explainability ($22.7\%$ $\uparrow$), and grounding ($25.3\%$ $\uparrow$), making it a promising approach for real-world clinical applications.



Paperid:3046
Authors:Zaiwei Chen
Abstract:
Abstract:This work presents the first finite-time analysis of average-reward $Q$-learning with an asynchronous implementation. A key feature of the algorithm we study is the use of adaptive stepsizes that act as local clocks for each state-action pair. We show that the mean-square error of this $Q$-learning algorithm, measured in the span seminorm, converges at a rate of $\smash{\tilde{\mathcal{O}}(1/k)}$. To establish this result, we demonstrate that adaptive stepsizes are necessary: without them, the algorithm fails to converge to the correct target. Moreover, adaptive stepsizes can be viewed as a form of implicit importance sampling that counteracts the effect of asynchronous updates.Technically, the use of adaptive stepsizes causes each $Q$-learning update to depend on the full sample history, introducing strong correlations and making the algorithm a non-Markovian stochastic approximation (SA) scheme. Our approach to overcoming this challenge involves (1) a time-inhomogeneous Markovian reformulation of non-Markovian SA, and (2) a combination of almost-sure time-varying bounds, conditioning arguments, and Markov chain concentration inequalities to break the strong correlations between the adaptive stepsizes and the iterates.



Paperid:3047
Authors:Ziqiang Cui, Yunpeng Weng, Xing Tang, Xiaokun Zhang, Shiwei Li, Peiyang Liu, Bowei He, Dugang Liu, Weihong Luo, Xiuqiang He, Chen Ma
Abstract:
Contrastive learning has shown effectiveness in improving sequential recommendation models. However, existing methods still face challenges in generating high-quality contrastive pairs: they either rely on random perturbations that corrupt user preference patterns or depend on sparse collaborative data that generates unreliable contrastive pairs. Furthermore, existing approaches typically require predefined selection rules that impose strong assumptions, limiting the model's ability to autonomously learn optimal contrastive pairs. To address these limitations, we propose a novel approach named Semantic Retrieval Augmented Contrastive Learning (SRA-CL). SRA-CL leverages the semantic understanding and reasoning capabilities of LLMs to generate expressive embeddings that capture both user preferences and item characteristics. These semantic embeddings enable the construction of candidate pools for inter-user and intra-user contrastive learning through semantic-based retrieval. To further enhance the quality of the contrastive samples, we introduce a learnable sample synthesizer that optimizes the contrastive sample generation process during model training. SRA-CL adopts a plug-and-play design, enabling seamless integration with existing sequential recommendation architectures. Extensive experiments on four public datasets demonstrate the effectiveness and model-agnostic nature of our approach. Our code is available at https://anonymous.4open.science/r/SRA-CL/.



Paperid:3048
Authors:Simin Fan, Maria Ios Glarou, Martin Jaggi
Abstract:
Abstract:The performance of large language models (LLMs) across diverse downstream applications is fundamentally governed by the quality and composition of their pretraining corpora.Existing domain reweighting algorithms primarily optimize data mixtures for a single target task, thereby resulting in models that overfit to specialized objectives while exhibiting substantial performance degradation on other benchmarks.This paper introduces $\textbf{G}$roup $\textbf{R}$obust Multi-target $\textbf{A}$daptive $\textbf{P}$r$\textbf{E}$training (GRAPE), a novel multi-source-multi-target domain reweighting framework designed to calibrate pretraining data mixtures for robust performance across multiple target tasks simultaneously.GRAPE dynamically adjusts sampling weights across source domains ($\textit{domain weights}$) while concurrently modulating $\textit{task weights}$ that quantify the relative importance of each individual target task.This adaptive process prioritizes tasks based on their learning difficulty throughout training. We formulate this interleaved reweighting mechanism as a minimax optimization problem: The inner maximization adjusts task weights leveraging group distributed-robust-optimization (DRO), where those tasks demonstrating the least improvement under the current data mixture are prioritized with higher weights; The outer minimization then optimizes domain weights to maximize loss reduction on the prioritized tasks.Experiments on $\texttt{ClimbLab}$ and $\texttt{SlimPajama}$ datasets demonstrate that GRAPE consistently outperforms baseline methods in terms of reasoning accuracies across 6 benchmarks. Furthermore, when applied to multilingual targets, GRAPE effectively identifies optimal training mixtures from mainstream languages, achieving superior language modeling capabilities across 8 low-resource target languages.



Paperid:3049
Authors:Dhruva Kashyap, Chaitanya Murti, Tanay Narshana, Pranav K Nayak, Chiranjib Bhattacharyya
Abstract:
Modifying well-trained models for tasks such as pruning or unlearning without access to training data or the original loss function is a challenging problem. While various techniques exist for tasks like pruning and classwise unlearning, they often require training data, are computationally expensive, or are architecture-specific. To address this, we investigate the fundamental question of identifying critical model components that significantly influence predictive performance without using gradients. Contrary to existing literature, we empirically observe that well-trained Transformer models exhibit Lipschitz continuity. Leveraging this insight, we theoretically demonstrate that for Lipschitz networks, the global reconstruction error is linearly bounded by local reconstruction errors for CNNs and well-trained transformers. This motivates our introduction of \emph{Subset Fidelity}, a novel metric to quantify the importance of component subsets by assessing their ability to reconstruct model outputs after weight adjustments. We further show that for uncorrelated features, selecting components with the highest individual Subset Fidelity scores is optimal. Based on this, we propose {\bf ModHiFi}, an algorithm for model modification that does not require the training data or loss function. Specifically, ModHiFi-P for structured pruning achieves 11% speed up improvement over current state of the art on ImageNet models and competitive performance on language models. For classwise unlearning, ModHiFi-U achieves complete unlearning on CIFAR-10 without fine-tuning and demonstrates competitive performance on Swin-Transformers.



Paperid:3050
Authors:JIANFENG CAI, Wengang Zhou, Zongmeng Zhang, Jiale Hong, zhannianji, Houqiang Li
Abstract:
Abstract:Multimodal large language models (MLLMs) have achieved remarkable progress in video understanding. However, hallucination, where the model generates plausible yet incorrect outputs, persists as a significant and under-addressed challenge in the video domain. Among existing solutions, activation engineering has proven successful in mitigating hallucinations in LLMs and ImageLLMs, yet its applicability to VideoLLMs remains largely unexplored. In this work, we are the first to systematically investigate the effectiveness and underlying mechanisms of activation engineering for mitigating hallucinations in VideoLLMs. We initially conduct an investigation of the key factors affecting the performance of activation engineering and find that a model’s sensitivity to hallucination depends on $\textbf{temporal variation}$ rather than task type. Moreover, selecting appropriate internal modules and dataset for activation engineering is critical for reducing hallucination. Guided by these findings, we propose a temporal-aware activation engineering framework for VideoLLMs, which adaptively identifies and manipulates hallucination-sensitive modules based on the temporal variation characteristic, substantially mitigating hallucinations without additional LLM fine-tuning. Experiments across multiple models and benchmarks demonstrate that our method markedly reduces hallucination in VideoLLMs, thereby validating the robustness of our findings.



Paperid:3051
Authors:Nisha Chandramoorthy, Adriaan de Clercq
Abstract:
A curious phenomenon observed in some dynamical generative models is the following: despite learning errors in the score function or the drift vector field, the generated samples appear to shift \emph{along} the support of the data distribution but not \emph{away} from it. In this work, we investigate this phenomenon of \emph{robustness of the support} by taking a dynamical systems approach on the generating stochastic/deterministic process. Our perturbation analysis of the probability flow reveals that infinitesimal learning errors cause the predicted density to be different from the target density only on the data manifold for a wide class of generative models. Further, what is the dynamical mechanism that leads to the robustness of the support? We show that the alignment of the top Lyapunov vectors (most sensitive infinitesimal perturbation directions) with the tangent spaces along the boundary of the data manifold leads to robustness and prove a sufficient condition on the dynamics of the generating process to achieve this alignment. Moreover, the alignment condition is efficient to compute and, in practice, for robust generative models, automatically leads to accurate estimates of the tangent bundle of the data manifold. Using a finite-time linear perturbation analysis on samples paths as well as probability flows, our work complements and extends existing works on obtaining theoretical guarantees for generative models from a stochastic analysis, statistical learning and uncertainty quantification points of view. Our results apply across different dynamical generative models, such as conditional flow-matching and score-based generative models, and for different target distributions that may or may not satisfy the manifold hypothesis.



Paperid:3052
Authors:Heli Ben-Hamu, Itai Gat, Daniel Severo, Niklas S Nolte, Brian Karrer
Abstract:
Recent masked diffusion models (MDMs) have shown competitive performance compared to autoregressive models (ARMs) for language modeling. While most literature has focused on performance enhancing sampling procedures, efficient sampling from MDMs has been scarcely explored. We make the observation that often a given sequence of partially masked tokens determines the values of multiple unknown tokens deterministically, meaning that a single prediction of a masked model holds additional information unused by standard sampling procedures. Based on this observation, we introduceEB-Sampler, a simple drop-in replacement for existing samplers, utilizing anEntropyBounded unmasking procedure that dynamically unmasks multiple tokens in one function evaluation with predefined approximate error tolerance. We formulate the EB-Sampler as part of a broad family of adaptive samplers for which we provide an error analysis that motivates our algorithmic choices. EB-Sampler accelerates sampling from current state of the art MDMs by roughly 2-3x on standard coding and math reasoning benchmarks without loss in performance. We also validate the same procedure works well on smaller reasoning tasks including maze navigation and sudoku, tasks ARMs often struggle with.



Paperid:3053
Authors:Ye Wang, Ziheng Wang, Boshen Xu, Yang Du, Kejun Lin, Zihan Xiao, Zihao Yue, Jianzhong Ju, Liang Zhang, Dingyi Yang, Xiangnan Fang, Zewen He, Zhenbo Luo, Wenxuan Wang, Junqi Lin, Jian Luan, Qin Jin
Abstract:
Temporal Video Grounding (TVG), the task of locating specific video segments based on language queries, is a core challenge in long-form video understanding. While recent Large Vision-Language Models (LVLMs) have shown early promise in tackling TVG through supervised fine-tuning (SFT), their ability to generalize remains limited. To address this, we propose a novel post-training framework that enhances the generalization capabilities of LVLMs via reinforcement learning (RL). Specifically, our contributions span three key directions: (1) Time-R1: we introduce a reasoning-guided post-training framework via RL with verifiable reward to enhance capabilities of LVLMs on the TVG task. (2) TimeRFT: we explore post-training strategies on our curated RL-friendly dataset, which trains the model to progressively comprehend more difficult samples, leading to better generalization and stable training processes. (3) TVGBench: we carefully construct a small but comprehensive and balanced benchmark suitable for LVLM evaluation, which is sourced from available public benchmarks. Extensive experiments demonstrate that Time-R1 achieves state-of-the-art performance across multiple downstream datasets using significantly less training data than prior LVLM approaches, while preserving and improving its general video understanding capabilities. Code: https://anonymous.4open.science/r/Time-R1/README.md.



Paperid:3054
Authors:Han Deng, Yuan Meng, SHIXIANG TANG, Wanli Ouyang, Xinzhu Ma
Abstract:
Competitive programming is widely used to evaluate the coding and reasoning abilities of large language models. However, the growing presence of duplicate or highly similar problems raises concerns not only about competition fairness, but also about the validity of competitive programming as a benchmark for model evaluation. We introduce a retrieval-oriented benchmark suite for competitive programming, covering four retrieval tasks—two code-centric (Text-to-Code, Code-to-Code) and two newly proposed problem-centric tasks (Problem-to-Duplicate, Simplified-to-Full)—built from a combination of automatically crawled problem–solution data and manually curated annotations. Our contribution includes both high-quality training data and temporally separated test sets for reliable evaluation. We develop two task-specialized retrievers based on this dataset: CPRetriever-Code, trained with a novel Group-InfoNCE loss for problem–code alignment, and CPRetriever-Prob, fine-tuned for problem-level similarity. Both models achieve strong results and are open-sourced for local use. Finally, we analyze LiveCodeBench and find that high-similarity problems inflate model pass rates and reduce differentiation, underscoring the need for similarity-aware evaluation in future benchmarks.



Paperid:3055
Authors:RUOFAN LIU, YICHEN PENG, Takanori Oku, Chen-Chieh Liao, Erwin Wu, Shinichi Furuya, Hideki Koike
Abstract:
Muscle coordination is fundamental when humans interact with the world. Reliable estimation of hand muscle engagement can serve as a source of internal feedback, supporting the development of embodied intelligence and dexterous skill acquisition. However, contemporary electromyography (EMG) sensing techniques either require prohibitive devices or are constrained to gross motors inherently involving large muscles. On the other hand, EMGs exhibit dependency on individual anatomical variability and task-specific contexts, resulting in limited generalization. In this work, we preliminarily investigate the latent pose-EMG correspondence using a general EMG gesture dataset. We further introduce a multimodal dataset, PianoKPM Dataset, and a hand muscle estimation framework, PianoKPM Net, to facilitate high-fidelity EMG inference. Subsequently, our approach is compared against reproducible competitive baselines. The generalization and adaptation across unseen users and tasks are evaluated by quantifying the training set scale and the included data amount.



Paperid:3056
Authors:Sen Wang, Jingyi Tian, Le Wang, Zhimin Liao, lijiayi, Huaiyi Dong, Kun Xia, Sanping Zhou, Wei Tang, Gang Hua
Abstract:
World models allow agents to simulate the consequences of actions in imagined environments for planning, control, and long-horizon decision-making. However, existing autoregressive world models struggle with visually coherent predictions due to disrupted spatial structure, inefficient decoding, and inadequate motion modeling. In response, we propose Scale-wise Autoregression with Motion PrOmpt (SAMPO), a hybrid framework that combines visual autoregressive modeling for intra-frame generation with causal modeling for next-frame generation. Specifically, SAMPO integrates temporal causal decoding with bidirectional spatial attention, which preserves spatial locality and supports parallel decoding within each scale. This design significantly enhances both temporal consistency and rollout efficiency. To further improve dynamic scene understanding, we devise an asymmetric multi-scale tokenizer that preserves spatial details in observed frames and extracts compact dynamic representations for future frames, optimizing both memory usage and model performance. Additionally, we introduce a trajectory-aware motion prompt module that injects spatiotemporal cues about object and robot trajectories, focusing attention on dynamic regions and improving temporal consistency and physical realism. Extensive experiments show that SAMPO achieves competitive performance in action-conditioned video prediction and model-based control, improving generation quality with 4.4× faster inference. We also evaluate SAMPO's zero-shot generalization and scaling behavior, demonstrating its ability to generalize to unseen tasks and benefit from larger model sizes.



Paperid:3057
Authors:Leyang Hu, Matteo Gamba, Randall Balestriero
Abstract:
Abstract:The scaling of model and data sizes has reshaped the AI landscape, making finetuning the standard protocol for adapting pretrained models to downstream tasks. However, dominant finetuning methods typically rely on weight adaptation, often lack interpretability, and depend on heuristically chosen hyperparameters. In this paper, we take a different perspective and shift the focus from weights to activation functions, viewing them through the lens of spline operators. We propose Curvature Tuning (CT), an interpretable and principled steering method that modulates a model’s decision boundary by injecting a single hyperparameter into its activation functions. We show that CT provably adjusts model decision boundary curvature and, more fundamentally, modulates the model’s hypothesis space—thereby complementing current finetuning methods, whose effect lies primarily in feature adaptation. Making this parameter trainable gives rise to a novel and highly parameter-efficient finetuning method. Empirically, CT improves both generalization and robustness: it boosts downstream accuracy of ResNet-50/152 by 7.14\%/8.46\% over linear probing and 4.64\%/1.70\% over LoRA across 12 datasets, and improves robust accuracy on the $\ell_{\infty}$ benchmark from RobustBench by 1032.64\%/1494.46\%. Our code is included in the supplementary.



Paperid:3058
Authors:Junhong Shen, Hao Bai, Lunjun Zhang, Yifei Zhou, Amrith Setlur, Peter Tong, Diego Caples, Nan Jiang, Tong Zhang, Ameet Talwalkar, Aviral Kumar
Abstract:
Test-time scaling in agentic tasks often relies on generating long reasoning traces ("think" more) before acting, but this does not allow agents to acquire new information from the environment or adapt behavior over time. In this work, we propose scaling test-time interaction, an untapped dimension for test-time scaling that increases the agent's interaction horizon to enable rich behaviors such as exploration, backtracking, and dynamic re-planning within a single rollout. To demonstrate the promise of this scaling dimension, we situate our study in the domain of web agents. We first show that even prompting-based interaction scaling can improve task success on web benchmarks non-trivially. Building on this, we introduce TTI, a curriculum-based online reinforcement learning (RL) approach that trains agents by adaptively adjusting their interaction lengths during rollout. Using a Gemma 3 12B model, TTI sets a new state-of-the-art among open-source agents trained on public data on WebVoyager and WebArena. Case studies further reveal that TTI enables agents to balance exploration and exploitation adaptively. Our results establish interaction scaling as a powerful, complementary axis to scaling per-action compute, offering new avenues for training robust and adaptive agents.



Paperid:3059
Authors:man zhou, Xuanhua He, Danfeng Hong, Bo Huang
Abstract:
Pan-sharpening aims to generate a spatially and spectrally enriched multi-spectral image by integrating complementarycross-modality information from low-resolution multi-spectral image and texture-rich panchromatic counterpart. In this work, we propose aWKV-sharing embraced random shuffle RWKV high-order modeling paradigm for pan-sharpening from Bayesian perspective, coupled with random weight manifold distribution training strategy derived from Functional theory to regularize the solution space adhering to thefollowing principles: 1) Random-shuffle RWKV. Recently, the Vision RWKV model, with its inherent linear complexity in global modeling,has inspired us to explore its untapped potential in pan-sharpening tasks. However, its attention mechanism, relying on a recurrentbidirectional scanning strategy, suffers from biased effects and demands significant processing time. To address this, we propose a novelBayesian-inspired scanning strategy called Random Shuffle, complemented by a theoretically-sound inverse shuffle to preserveinformation coordination invariance, effectively eliminating biases associated with fixed sequence scanning. The Random Shuffleapproach mitigates preconceptions in global 2D dependencies in mathematical expectation, providing the model with an unbiased prior.In line with similar spirit of Dropout, we introduce a testing methodology based on Monte Carlo averaging to ensure the model’s outputaligns more closely with expected results. 2) WKV-sharing high-order. Regarding KV’s attention score calculation in spatial mixer of RWKV, we leverage WKV-sharing mechanism to transfer KV activations across RWKV layers, achieving lower latency and improved trainability, and revisit the channel mixer in RWKV, originally a first-order weighting function, and redevelop its high-order potential by sharing the gate mechanism across RWKV layer. Comprehensive experiments across pan-sharpening benchmarks demonstrate our model’s effectiveness, consistently outperforming state-of-the-art alternatives



Paperid:3060
Authors:Chenguang Duan, Yuling Jiao, Huazhen Lin, Wensen Ma, Jerry Yang
Abstract:
Learning transferable data representations from abundant unlabeled data remains a critical challenge in machine learning. While numerous self-supervised learning methods have emerged to address this challenge, a significant class of these approaches aligns the covariance or correlation matrix with the identity matrix. Despite impressive performance across various downstream tasks, these methods often suffer from biased sample risk, leading to significant optimization shifts in mini-batch settings and hindering theoretical analysis. In this paper, we introduce a novel Adversarial Self-Supervised Representation Learning (ASSRL) for unbiased transfer learning. Our method not only outperforms existing methods across multiple benchmark datasets but also offers comprehensive end-to-end theoretical guarantees. Our analysis reveals that minimax optimization drives representations to form well-separated clusters in the embedding space, provided sufficient upstream unlabeled data. Consequently, our method delivers strong classification performance even with limited downstream labels, shedding new light on few-shot learning.



Paperid:3061
Authors:Tingzhu Bi, Yicheng Pan, Xinrui Jiang, Huize Sun, Meng Ma, Ping Wang
Abstract:
Uncovering cause-effect relationships from observational time series is fundamental to understanding complex systems. While many methods infer static causal graphs, real-world systems often exhibitdynamic causality—where relationships evolve over time. Accurately capturing these temporal dynamics requires time-resolved causal graphs. We propose UnCLe, a novel deep learning method for scalable dynamic causal discovery. UnCLe employs a pair of Uncoupler and Recoupler networks to disentangle input time series into semantic representations and learns inter-variable dependencies via auto-regressive Dependency Matrices. It estimates dynamic causal influences by analyzing datapoint-wise prediction errors induced by temporal perturbations. Extensive experiments demonstrate that UnCLe not only outperforms state-of-the-art baselines on static causal discovery benchmarks but, more importantly, exhibits a unique capability to accurately capture and represent evolving temporal causality in both synthetic and real-world dynamic systems (e.g., human motion). UnCLe offers a promising approach for revealing the underlying, time-varying mechanisms of complex phenomena.



Paperid:3062
Authors:Kinam Kim, Junha Hyung, Jaegul Choo
Abstract:
Recent advances in text-to-video diffusion models have enabled high-quality video synthesis, but controllable generation remains challenging—particularly under limited data and compute. Existing fine-tuning methods often rely on external encoders or architectural modifications, which demand large datasets and are typically restricted to spatially aligned conditioning, limiting flexibility and scalability. In this work, we introduce Temporal In-Context Fine-Tuning (TIC-FT), an efficient and versatile approach for adapting pretrained video diffusion models to diverse conditional generation tasks. Our key idea is to concatenate condition and target frames along the temporal axis and insert intermediate buffer frames with progressively increasing noise levels. These buffer frames enable smooth transitions, aligning the fine-tuning process with the pretrained model’s temporal dynamics. TIC-FT requires no architectural changes and achieves strong performance with as few as 10–30 training samples. We validate our method across a range of tasks—including image-to-video and video-to-video generation—using large-scale base models such as CogVideoX-5B and Wan-14B. Extensive experiments show that TIC-FT outperforms existing baselines in both condition fidelity and visual quality, while remaining highly efficient in both training and inference.



Paperid:3063
Authors:Vijay Veerabadran, Fanyi Xiao, Nitin Kamra, Pedro Matias, Joy Chen, Caley Drooff, Brett Roads, Riley J Williams, Ethan Henderson, Xuanyi Zhao, Kevin Carlberg, Joseph Tighe, Karl Ridgeway
Abstract:
There has recently been a surge of interest in Wearable Assistant Agents: agents embodied in a wearable form factor such as smart glasses, who can take actions toward a user’s stated goal — a high-level language-expressed command such as “where did I leave my keys?”, “Text Alice I will be late”, or “What’s the weather in Cancun?”. In this work, we consider the complementary problem of eliminating the effort required to interact with such an agent by proactively inferring the user’s goal from multimodal contextual observations. As vision-language models (VLMs) hold strong potential to ultimately solve this problem, our work focuses on creating a strong benchmark to measure progress toward this end. Given the limited prior work in this area, establishing the benchmark required collecting a novel multimodal goal-inference dataset; our dataset comprises ~30 hours of data from 363 participants across 3,482 recordings, featuring ground-truth reference goals alongside accompanying visual, audio, digital, and longitudinal contextual observations. We ran a human predictability study, where we found that humans set a strong baseline that comprises a de facto upper bound on model performance: they show multiple choice question (MCQ) accuracy of 93%, with the best VLM achieving about 84% accuracy. However, MCQ assesses discrimination, not the model’s ultimate task of generating the goal through open-ended text generation. Through a meta-evaluation, we find that a VLM judging the generated goals is as good as a human judge if it has access to a human-authored script of the video or a correct reference goal. Finally, we evaluate several families of modern vision-language models on the benchmark, showing that larger models have a significant performance advantage, but are still far from being practically useful, as they produce relevant goals only ~57% of the time. The best-performing smaller models—whose size makes them better suited to wearable applications—perform significantly worse than their counterparts, generating ~49% accuracy on the benchmark. Through a modality ablation, we show that models benefit from extra information in relevant modalities with minimal performance degradation from irrelevant modalities, but don’t gain as much when noisy modalities are included (e.g., in the case of digital context when most of the app state is irrelevant).



Paperid:3064
Authors:Younggyo Seo, Pieter Abbeel
Abstract:
Predicting a sequence of actions has been crucial in the success of recent behavior cloning algorithms in robotics. Can similar ideas improve reinforcement learning (RL)? We answer affirmatively by observing that incorporating action sequences when predicting ground-truth return-to-go leads to lower validation loss. Motivated by this, we introduce Coarse-to-fine Q-Network with Action Sequence (CQN-AS), a novel value-based RL algorithm that learns a critic network that outputs Q-values over a sequence of actions, i.e., explicitly training the value function to learn the consequence of executing action sequences. Our experiments show that CQN-AS outperforms several baselines on a variety of sparse-reward humanoid control and tabletop manipulation tasks from BiGym and RLBench.



Paperid:3065
Authors:Zinuo Li, Xian Zhang, Yongxin Guo, Mohammed Bennamoun, Farid Boussaid, Girish Dwivedi, Luqi Gong, Qiuhong Ke
Abstract:
Humans naturally understand moments in a video by integrating visual and auditory cues. For example, localizing a scene in the video like “A scientist passionately speaks on wildlife conservation as dramatic orchestral music plays, with the audience nodding and applauding” requires simultaneous processing of visual, audio, and speech signals. However, existing models often struggle to effectively fuse and interpret audio information, limiting their capacity for comprehensive video temporal understanding. To address this, we present TriSense, a triple-modality large language model designed for holistic video temporal understanding through the integration of visual, audio, and speech modalities. Central to TriSense is a Query-Based Connector that adaptively reweights modality contributions based on the input query, enabling robust performance under modality dropout and allowing flexible combinations of available inputs. To support TriSense's multimodal capabilities, we introduce TriSense-2M, a high-quality dataset of over 2 million curated samples generated via an automated pipeline powered by fine-tuned LLMs. TriSense-2M includes long-form videos and diverse modality combinations, facilitating broad generalization. Extensive experiments across multiple benchmarks demonstrate the effectiveness of TriSense and its potential to advance multimodal video analysis. Code and dataset will be publicly released.



Paperid:3066
Authors:Anirban Das, Muhammad Irtaza Khalid, Rafael Peñaloza, Steven Schockaert
Abstract:
Designing models that can learn to reason in a systematic way is an important and long-standing challenge. In recent years, a wide range of solutions have been proposed for the specific case of systematic relational reasoning, including Neuro-Symbolic approaches, variants of the Transformer architecture, and specialized Graph Neural Networks. However, existing benchmarks for systematic relational reasoning focus on an overly simplified setting, based on the assumption that reasoning can be reduced to composing relational paths. In fact, this assumption is hard-baked into the architecture of several recent models, leading to approaches that can perform well on existing benchmarks but are difficult to generalize to other settings. To support further progress in the field of systematic relational reasoning with neural networks, we introduce a new benchmark that adds several levels of difficulty, requiring models to go beyond path-based reasoning.



Paperid:3067
Authors:Moritz Stargalla, Christoph Hertrich, Daniel Reichman
Abstract:
An established measure of the expressive power of a given ReLU neural network is the number of linear regions into which it partitions the input space. There exist many different, non-equivalent definitions of what a linear region actually is. We systematically assess which papers use which definitions and discuss how they relate to each other. We then analyze the computational complexity of counting the number of such regions for the various definitions. Generally, this turns out to be an intractable problem. We prove NP- and #P-hardness results already for networks with one hidden layer and strong hardness of approximation results for two or more hidden layers. Finally, on the algorithmic side, we demonstrate that counting linear regions can at least be achieved in polynomial space for some common definitions.



Paperid:3068
Authors:Zaiyan Xu, Sushil Vemuri, Kishan Panaganti, Dileep Kalathil, Rahul Jain, Deepak Ramachandran
Abstract:
A major challenge in aligning large language models (LLMs) with human preferences is the issue of distribution shift. LLM alignment algorithms rely on static preference datasets, assuming that they accurately represent real-world user preferences. However, user preferences vary significantly across geographical regions, demographics, linguistic patterns, and evolving cultural trends. This preference distribution shift leads to catastrophic alignment failures in many real-world applications. We address this problem using the principled framework of distributionally robust optimization, and develop two novel distributionally robust direct preference optimization (DPO) algorithms, namely, Wasserstein DPO (WDPO) and Kullback–Leibler DPO (KLDPO). We characterize the sample complexity of learning the optimal policy parameters for WDPO and KLDPO. Moreover, we propose scalable gradient descent-style learning algorithms by developing suitable approximations for the challenging minimax loss functions of WDPO and KLDPO. Our empirical experiments using benchmark data sets and LLMs demonstrate the superior performance of WDPO and KLDPO in substantially improving the alignment when there is a preference distribution shift.



Paperid:3069
Authors:Qian Ma, Ruoxiang Xu, Yongqiang Cai
Abstract:
Numerous studies have demonstrated that the Transformer architecture possesses the capability for in-context learning (ICL). In scenarios involving function approximation, context can serve as a control parameter for the model, endowing it with the universal approximation property (UAP). In practice, context is represented by tokens from a finite set, referred to as a vocabulary, which is the case considered in this paper, i.e., vocabulary in-context learning (VICL). We demonstrate that VICL in single-layer Transformers, without positional encoding, does not possess the UAP; however, it is possible to achieve the UAP when positional encoding is included. Several sufficient conditions for the positional encoding are provided. Our findings reveal the benefits of positional encoding from an approximation theory perspective in the context of in-context learning.



Paperid:3070
Authors:Ruixiao Yang, Chuchu Fan
Abstract:
The Time-Dependent Traveling Salesman Problem (TDTSP) extends classical TSP with dynamic edge weights that vary with departure time, reflecting real-world scenarios like transportation networks, where travel times fluctuate due to congestion patterns. TDTSP violates symmetry, triangle inequality, and cyclic invariance properties of classical TSP, creating unique computational challenges. In this paper, we propose a neural model that encodes the time-dependent adjacency tensor, addressing asymmetry and triangle inequality violations, followed by a time-aware decoder. This architecture effectively captures the complex spatiotemporal dynamics of TDTSP. Beyond architectural innovations, our research reveals a critical evaluation insight: many practical TDTSP instances maintain the same optimal solution regardless of time-dependent edge weights. This exposes a fundamental limitation in current evaluation practices for TDTSP that rely solely on average travel time metrics across all instances. Such metrics fail to effectively distinguish between methods that genuinely capture temporal dynamics and those that merely perform well on static routing problems. Instead, we present extensive experiments on real-world datasets, evaluating our approach on both entire datasets and specifically filtered instances where temporal dependencies alter the optimal solution. Results show that our method achieves state-of-the-art average optimality gap on full instances and significant travel time reduction on instances for which time-aware routing saves time. These results demonstrate state-of-the-art ability to identify and exploit temporal dependencies while establishing new standards for evaluating routing problems with temporal dependencies.



Paperid:3071
Authors:Qianlan Yang, Xiangjun Wang, Danielle Perszyk, Yu-Xiong Wang
Abstract:
Foundation models have recently shown strong potential as web agents, capable of interpreting high-level instructions and interacting with complex web interfaces. However, existing training paradigms for these agents often rely on predefined task datasets and curated demonstrations, limiting their scalability, adaptability, and capacity for self-improvement. In this work, we introduceSelf-guided hierArchical exploration for Generalist wEb agents(SAGE), a new training framework designed to support autonomous skill acquisition through self-guided hierarchical exploration. Our method introduces a three-tier exploration strategy: a pre-exploration phase to build structural understanding of web environments, a top-level exploration strategy to generate a self-evolving curriculum of tasks from easy to hard, and a low-level exploration mechanism that combines planning-based rollouts with step-wise learning to improve policy efficiency. Together, these components form a scalable, supervision-free framework for web agent training. Experimental results on WebVoyager and WebArena demonstrate that our method significantly outperforms prior approaches, enabling foundation model agents to learn complex web tasks with greater generalization and robustness.



Paperid:3072
Authors:Rui Wang, Chen Hu, Xiaoning Song, Xiaojun Wu, Nicu Sebe, Ziheng Chen
Abstract:
Deep neural networks operating on non-Euclidean geometries have recently demonstrated impressive performance across various machine-learning applications. Several studies have extended the attention mechanism to different manifolds. However, most existing non-Euclidean attention models are tailored to specific geometries, limiting their applicability. On the other hand, recent studies show that several matrix manifolds, such as Symmetric Positive Definite (SPD), Symmetric Positive Semi-Definite (SPSD), and Grassmannian manifolds, admit gyrovector structures, which extend vector addition and scalar product into manifolds. Leveraging these properties, we propose a Gyro Attention (GyroAtt) framework over general gyrovector spaces, applicable to various matrix geometries. Empirically, we manifest GyroAtt on three gyro structures on the SPD manifold, three on the SPSD manifold, and one on the Grassmannian manifold. Extensive experiments on four electroencephalography (EEG) datasets demonstrate the effectiveness of our framework.



Paperid:3073
Authors:Yuyang Yu, Zhengwei Chen, Xuemiao Xu, Lei Zhang, Haoxin Yang, Yongwei Nie, Shengfeng He
Abstract:
3D anomaly detection in point-cloud data is critical for industrial quality control, aiming to identify structural defects with high reliability. However, current memory bank-based methods often suffer from inconsistent feature transformations and limited discriminative capacity, particularly in capturing local geometric details and achieving rotation invariance. These limitations become more pronounced when registration fails, leading to unreliable detection results. We argue that point-cloud registration plays an essential role not only in aligning geometric structures but also in guiding feature extraction toward rotation-invariant and locally discriminative representations. To this end, we propose a registration-induced, rotation-invariant feature extraction framework that integrates the objectives of point-cloud registration and memory-based anomaly detection. Our key insight is that both tasks rely on modeling local geometric structures and leveraging feature similarity across samples. By embedding feature extraction into the registration learning process, our framework jointly optimizes alignment and representation learning. This integration enables the network to acquire features that are both robust to rotations and highly effective for anomaly detection. Extensive experiments on the Anomaly-ShapeNet and Real3D-AD datasets demonstrate that our method consistently outperforms existing approaches in effectiveness and generalizability.



Paperid:3074
Authors:Sunwoo Kim, Hyunjin Hwang, Kijung Shin
Abstract:
The performance of a deep learning model on a specific task and dataset depends heavily on its neural architecture, motivating considerable efforts to rapidly and accurately identify architectures suited to the target task and dataset. To achieve this, researchers use machine learning models—typically neural architecture encoders—to predict the performance of a neural architecture. Many state-of-the-art encoders aim to capture information flow within a neural architecture, which reflects how information moves through the forward pass and backpropagation, via a specialized model structure. However, due to their complicated structures, these flow-based encoders are significantly slower to process neural architectures compared to simpler encoders, presenting a notable practical challenge. To address this, we propose FGP, a novel pre-training method for neural architecture encoding that trains an encoder to capture the information flow without requiring specialized model structures. FGP trains an encoder to reconstruct a flow surrogate, our proposed representation of the neural architecture's information flow. Our experiments show that FGP boosts encoder performance by up to 106\% in Precision@1\%, compared to the same encoder trained solely with supervised learning.



Paperid:3075
Authors:Yichen Liu, Yan Lin, Shengnan Guo, Zeyu Zhou, Youfang Lin, Huaiyu Wan
Abstract:
Vehicle GPS trajectories record how vehicles move over time, storing valuable travel semantics, including movement patterns and travel purposes. Learning travel semantics effectively and efficiently is crucial for real-world applications of trajectory data, which is hindered by two major challenges. First, travel purposes are tied to the functions of the roads and points-of-interest (POIs) involved in a trip. Such information is encoded in textual addresses and descriptions and introduces heavy computational burden to modeling. Second, real-world trajectories often contain redundant points, which harm both computational efficiency and trajectory embedding quality. To address these challenges, we propose TrajMamba, a novel approach for efficient and semantically rich vehicle trajectory learning. TrajMamba introduces a Traj-Mamba Encoder that captures movement patterns by jointly modeling both GPS and road perspectives of trajectories, enabling robust representations of continuous travel behaviors. It also incorporates a Travel Purpose-aware Pre-training procedure to integrate travel purposes into the learned embeddings without introducing extra overhead to embedding calculation. To reduce redundancy in trajectories, TrajMamba features a Knowledge Distillation Pre-training scheme to identify key trajectory points through a learnable mask generator and obtain effective compressed trajectory embeddings. Extensive experiments on two real-world datasets and three downstream tasks show that TrajMamba outperforms state-of-the-art baselines in both efficiency and accuracy.



Paperid:3076
Authors:Guillaume Vray, Devavrat Tomar, Xufeng Gao, Jean-Philippe Thiran, Evan Shelhamer, Behzad Bozorgtabar
Abstract:
This paper introducesReservoirTTA, a novel plug–in framework designed for prolonged test–time adaptation (TTA) in scenarios where the test domain continuously shifts over time, including cases where domains recur or evolve gradually. At its core, ReservoirTTA maintains a reservoir of domain-specialized models—an adaptive test-time model ensemble—that both detects new domains via online clustering over style features of incoming samples and routes each sample to the appropriate specialized model, and thereby enables domain-specific adaptation. This multi-model strategy overcomes key limitations of single model adaptation, such as catastrophic forgetting, inter-domain interference, and error accumulation, ensuring robust and stable performance on sustained non-stationary test distributions. Our theoretical analysis reveals key components that bound parameter variance and prevent model collapse, while our plug–in TTA module mitigates catastrophic forgetting of previously encountered domains. Extensive experiments on the classification corruption benchmarks, including ImageNet-C and CIFAR-10/100-C, as well as the Cityscapes→ACDC semantic segmentation task, covering recurring and continuously evolving domain shifts, demonstrate that ReservoirTTA significantly improves adaptation accuracy and maintains stable performance across prolonged, recurring shifts, outperforming state-of-the-art methods. The code will be released upon acceptance.



Paperid:3077
Authors:Bhanu Tokas, Rahul Nair, Hannah Kerner
Abstract:
Most ML datasets today contain biases. When we train models on these datasets, they often not only learn these biases but can worsen them --- a phenomenon known as bias amplification. Several co-occurrence-based metrics have been proposed to measure bias amplification in classification datasets. They measure bias amplification between a protected attribute (e.g., gender) and a task (e.g., cooking). These metrics also support fine-grained bias analysis by identifying the direction in which a model amplifies biases. However, co-occurrence-based metrics have limitations --- some fail to measure bias amplification in balanced datasets, while others fail to measure negative bias amplification. To solve these issues, recent work proposed a predictability-based metric called leakage amplification (LA). However, LA cannot identify the direction in which a model amplifies biases. We propose Directional Predictability Amplification (DPA), a predictability-based metric that is (1) directional, (2) works with balanced and unbalanced datasets, and (3) correctly identifies positive and negative bias amplification. DPA eliminates the need to evaluate models on multiple metrics to verify these three aspects. DPA also improves over prior predictability-based metrics like LA: it is less sensitive to the choice of attacker function (a hyperparameter in predictability-based metrics), reports scores within a bounded range, and accounts for dataset bias by measuring relative changes in predictability. Our experiments on well-known datasets like COMPAS (a tabular dataset), COCO, and ImSitu (image datasets) show that DPA is the most reliable metric to measure bias amplification in classification problems.



Paperid:3078
Authors:Jaewon Chu, Seunghun Lee, Hyunwoo J. Kim
Abstract:
Large language models (LLMs) have achieved remarkable success across diverse domains, due to their strong instruction-following capabilities. This raised interest in optimizing instructions for black-box LLMs, whose internal parameters are inaccessible but popular for their strong performance and ease of use. Recent approaches leverage white-box LLMs to assist instruction optimization for black-box LLMs by generating instructions from soft prompts. However, white-box LLMs often map different soft prompts to the same instruction, leading to redundant queries to the black-box model. While previous studies regarded this many-to-one mapping as a redundancy to be avoided, we reinterpret it as useful prior knowledge that can enhance the optimization performance. To this end, we introduce PREimage-informed inSTruction Optimization (PRESTO), a novel framework that leverages the preimage structure of soft prompts to improve query efficiency. PRESTO consists of three key components: (1) score sharing, which shares the evaluation score with all soft prompts in a preimage; (2) preimage-based initialization, which select initial data points that maximize search space coverage using preimage information; and (3) score consistency regularization, which enforces prediction consistency within each preimage. By leveraging preimages, PRESTO observes 14 times more scored data under the same query budget, resulting in more efficient optimization. Experimental results on 33 instruction optimization tasks demonstrate the superior performance of PRESTO.



Paperid:3079
Authors:Dar-Yen Chen, Hmrishav Bandyopadhyay, Kai Zou, Yi-Zhe Song
Abstract:
Negative guidance -- explicitly suppressing unwanted attributes -- remains a fundamental challenge in diffusion models, particularly in few-step sampling regimes. While Classifier-Free Guidance (CFG) works well in standard settings, it fails under aggressive sampling step compression due to divergent predictions between positive and negative branches. We present Normalized Attention Guidance (NAG), an efficient, training-free mechanism that applies extrapolation in attention space with L1-based normalization and refinement. NAG restores effective negative guidance where CFG collapses while maintaining fidelity. Unlike existing approaches, NAG generalizes across architectures (UNet, DiT), sampling regimes (few-step, multi-step), and modalities (image, video), functioning as a \textit{universal} plug-in with minimal computational overhead. Through extensive experimentation, we demonstrate consistent improvements in text alignment (CLIP Score), fidelity (FID, PFID), and human-perceived quality (ImageReward). Our ablation studies validate each design component, while user studies confirm significant preference for NAG-guided outputs. As a model-agnostic inference-time approach requiring no retraining, NAG provides effortless negative guidance for all modern diffusion frameworks -- pseudocode in the Appendix!



Paperid:3080
Authors:Yuheng Zhang, Dian Yu, Tao Ge, Linfeng Song, Zhichen Zeng, Haitao Mi, Nan Jiang, Dong Yu
Abstract:
Abstract:Reinforcement learning from human feedback (RLHF) has demonstrated remarkable effectiveness in aligning large language models (LLMs) with human preferences. Many existing alignment approaches rely on the Bradley-Terry (BT) model assumption, which assumes the existence of a ground-truth reward for each prompt-response pair. However, this assumption can be overly restrictive when modeling complex human preferences. In this paper, we drop the BT model assumption and study LLM alignment under general preferences, formulated as a two-player game. Drawing on theoretical insights from learning in games, we integrate optimistic online mirror descent into our alignment framework to approximate the Nash policy. Theoretically, we demonstrate that our approach achieves an $\mathcal{O}(T^{-1})$ bound on the duality gap, improving upon the previous $\mathcal{O}(T^{-1/2})$ result. Meanwhile, it enjoys a linear convergence rate in the last iterate, a property not achieved by previous methods. More importantly, we implement our method and show through experiments that it outperforms state-of-the-art RLHF algorithms across multiple representative benchmarks.



Paperid:3081
Authors:Zhuoqun Huang, Neil Marchant, Olga Ohrimenko, Benjamin Rubinstein
Abstract:
We consider the problem of certified robustness for sequence classification against edit distance perturbations. Naturally occurring inputs of varying lengths (e.g., sentences in natural language processing tasks) present a challenge to current methods that employ fixed-rate deletion mechanisms and lead to suboptimal performance. To this end, we introduce AdaptDel methods with adaptable deletion rates that dynamically adjust based on input properties. We extend the theoretical framework of randomized smoothing to variable-rate deletion, ensuring sound certification with respect to edit distance. We achieve strong empirical results in natural language tasks, observing up to 30 orders of magnitude improvement to median cardinality of the certified region, over SotA certifications.



Paperid:3082
Authors:Elena Zamaraeva, Christopher Collins, George Darling, Matthew S Dyer, Bei Peng, Rahul Savani, Dmytro Antypov, Vladimir Gusev, Judith Clymo, Paul Spirakis, Matthew Rosseinsky
Abstract:
Geometry optimization of atomic structures is a common and crucial task in computational chemistry and materials design. Following the learning to optimize paradigm, we propose a new multi-agent reinforcement learning method called Multi-Agent Crystal Structure optimization (MACS) to address the problem of periodic crystal structure optimization. MACS treats geometry optimization as a partially observable Markov game in which atoms are agents that adjust their positions to collectively discover a stable configuration. We train MACS across various compositions of reported crystalline materials to obtain a policy that successfully optimizes structures from the training compositions as well as structures of larger sizes and unseen compositions, confirming its excellent scalability and zero-shot transferability. We benchmark our approach against a broad range of state-of-the-art optimization methods and demonstrate that MACS optimizes periodic crystal structures significantly faster, with fewer energy calculations, and the lowest failure rate.



Paperid:3083
Authors:Jesse Haworth, Juo-Tung Chen, Nigel Nelson, Kim, Masoud Moghani, Chelsea Finn, Axel Krieger
Abstract:
Abstract:Robotic suturing is a prototypical long-horizon dexterous manipulation task, requiring coordinated needle grasping, precise tissue penetration, and secure knot tying. Despite numerous efforts toward end-to-end autonomy, a fully autonomous suturing pipeline has yet to be demonstrated on physical hardware. We introduce SutureBot: an autonomous suturing benchmark on the da Vinci Research Kit (dVRK), spanning needle pickup, tissue insertion, and knot tying. To ensure repeatability, we release a high-fidelity dataset comprising 1,890 suturing demonstrations. Furthermore, we propose a goal-conditioned framework that explicitly optimizes insertion-point precision, improving targeting accuracy by 80\% over a task-only baseline. To establish this task as a benchmark for dexterous imitation learning, we evaluate state-of-the-art vision-language-action (VLA) models, including $\pi_0$, GR00T N1, OpenVLA-OFT, and multitask ACT, each augmented with a high-level task-prediction policy. Autonomous suturing is a key milestone toward achieving robotic autonomy in surgery. These contributions support reproducible evaluation and development of precision-focused, long-horizon dexterous manipulation policies necessary for end-to-end suturing. Dataset is available at: \href{https://huggingface.co/datasets/jchen396/suturebot}{Hugging Face}.



Paperid:3084
Authors:Fanrui Zhang, Dian Li, Qiang Zhang, Chenjun, sinbadliu, Junxiong Lin, Jiahong Yan, Jiawei Liu, Zheng-Jun Zha
Abstract:
The rapid spread of multimodal misinformation on social media has raised growing concerns, while research on video misinformation detection remains limited due to the lack of large-scale, diverse datasets. Existing methods often overfit to rigid templates and lack deep reasoning over deceptive content. To address these challenges, we introduce FakeVV, a large-scale benchmark comprising over 100,000 video-text pairs with fine-grained, interpretable annotations. In addition, we further propose Fact-R1, a novel framework that integrates deep reasoning with collaborative rule-based reinforcement learning. Fact-R1 is trained through a three-stage process: (1) misinformation long-Chain-of-Thought (CoT) instruction tuning, (2) preference alignment via Direct Preference Optimization (DPO), and (3) Group Relative Policy Optimization (GRPO) using a novel verifiable reward function. This enables Fact-R1 to exhibit emergent reasoning behaviors comparable to those observed in advanced text-based reinforcement learning systems, but in the more complex multimodal misinformation setting. Our work establishes a new paradigm for misinformation detection, bridging large-scale video understanding, reasoning-guided alignment, and interpretable verification.



Paperid:3085
Authors:Gregory Barber, Todd Henry, Mulugeta Haile
Abstract:
We present TIDES, a text informed design approach for generating physically sound designs based on a textual description and a set of physical constraints. TIDES jointly optimizes structural (topology) and visual properties. A pre-trained text-image model is used to measure the design's visual alignment with a text prompt and a differentiable physics simulator is used to measure its physical performance. We evaluate TIDES on a series of structural optimization problems operating under different load and support conditions, at different resolutions, and experimentally in the lab by performing the 3-point bending test on 2D beam designs that are extruded and 3D printed. We find that it can jointly optimize the two objectives and return designs that satisfy engineering design requirements (compliance and density) while utilizing features specified by text.



Paperid:3086
Authors:Zijian Li, Minghao Fu, Junxian Huang, Yifan Shen, Ruichu Cai, Yuewen Sun, Guangyi Chen, Kun Zhang
Abstract:
Modeling hierarchical latent dynamics behind time series data is critical for capturing temporal dependencies across multiple levels of abstraction in real-world tasks. However, existing temporal causal representation learning methods fail to capture such dynamics, as they fail to recover the joint distribution of hierarchical latent variables from \textit{single-timestep observed variables}. Interestingly, we find that the joint distribution of hierarchical latent variables can be uniquely determined using three conditionally independent observations. Building on this insight, we propose a Causally Hierarchical Latent Dynamic (CHiLD) identification framework. Our approach first employs temporal contextual observed variables to identify the joint distribution of multi-layer latent variables. Sequentially, we exploit the natural sparsity of the hierarchical structure among latent variables to identify latent variables within each layer. Guided by the theoretical results, we develop a time series generative model grounded in variational inference. This model incorporates a contextual encoder to reconstruct multi-layer latent variables and normalize flow-based hierarchical prior networks to impose the independent noise condition of hierarchical latent dynamics. Empirical evaluations on both synthetic and real-world datasets validate our theoretical claims and demonstrate the effectiveness of CHiLD in modeling hierarchical latent dynamics.



Paperid:3087
Authors:Min Wang, Hao Yang, Qing Cheng, Jincai Huang
Abstract:
GNNs have achieved remarkable performance across a range of tasks, but their reliability under distribution shifts remains a significant challenge. In particular, energy-based OOD detection methods—which compute energy scores from GNN logits—suffer from unstable performance due to a fundamental coupling between the norm and direction of node embeddings. Our analysis reveals that this coupling leads to systematic misclassification of high-norm OOD samples and hinders reliable ID–OOD separation. Interestingly, GNNs also exhibit a desirable inductive bias known as angular clustering, where embeddings of the same class align in direction. Motivated by these observations, we propose GeoEnergy (Geometric Logit Decoupling for Energy-Based OOD Detection), a plug-and-play framework that enforces hyperspherical logit geometry by normalizing class weights while preserving embedding norms. This decoupling yields more structured energy distributions, sharper intra-class alignment, and improved calibration. GeoEnergy can be integrated into existing energy-based GNNs without retraining or architectural modification. Extensive experiments demonstrate that GeoEnergy consistently improves OOD detection performance and confidence reliability across various benchmarks and distribution shifts.



Paperid:3088
Authors:Jingfeng Wu, Pierre Marion, Peter Bartlett
Abstract:
Abstract:We study *gradient descent* (GD) with a constant stepsize for $\ell_2$-regularized logistic regression with linearly separable data. Classical theory suggests small stepsizes to ensure monotonic reduction of the optimization objective, achieving exponential convergence in $\widetilde{\mathcal{O}}(\kappa)$ steps with $\kappa$ being the condition number. Surprisingly, we show that this can be *accelerated* to $\widetilde{\mathcal{O}}(\sqrt{\kappa})$ by simply using a large stepsize---for which the objective evolves *nonmonotonically*. The acceleration brought by large stepsizes extends to minimizing the population risk for separable distributions, improving on the best-known upper bounds on the number of steps to reach a near-optimum. Finally, we characterize the largest stepsize for the local convergence of GD, which also determines the global convergence in special scenarios. Our results extend the analysis of Wu et al. (2024) from convex settings with minimizers at infinity to strongly convex cases with finite minimizers.



Paperid:3089
Authors:Donghyeon Joo, Helya Hosseini, Ramyad Hadidi, Bahar Asgari
Abstract:
We demonstrate that unstructured sparsity significantly improves KV cache compression for LLMs, enabling sparsity levels up to 70\% without compromising accuracy or requiring fine-tuning. We conduct a systematic exploration of pruning strategies and find per-token magnitude-based pruning as highly effective for both Key and Value caches under unstructured sparsity, surpassing prior structured pruning schemes. The Key cache benefits from prominent outlier elements, while the Value cache surprisingly benefits from a simple magnitude-based pruning despite its uniform distribution. KV cache size is the major bottleneck in decode performance due to high memory overhead for large context lengths. To address this, we use a bitmap-based sparse format and a custom attention kernel capable of compressing and directly computing over compressed caches pruned to arbitrary sparsity patterns, significantly accelerating memory-bound operations in decode computations and thereby compensating for the overhead of runtime pruning and compression.Our custom attention kernel coupled with the bitmap-based sparse format delivers substantial compression of KV cache and thereby enables longer context length and increased throughput. This advancement not only facilitates longer context lengths but also democratizes access to high-performing LLMs, enabling efficient deployment.



Paperid:3090
Authors:Anjie Liu, Jianhong Wang, Samuel Kaski, Jun Wang, Mengyue Yang
Abstract:
Guiding Cooperative Multi-Agent Reinforcement Learning (MARL) systems towards desirable outcomes are challenging, particularly when universal guidance on the desirable outcomes over the whole team is impractical. Furthermore, designing mechanisms to coordinate agents currently relies on empirical studies, lacking a unified perspective. To mitigate these issues, we introduce Multi-Agent Influence Diagrams (MAIDs) as a graphical tool to visualize existing coordination mechanisms. Based on MAIDs, we design a new coordination mechanism, referred to as targeted intervention, which is applied to only a single agent. In practice, we introduce a technique for causal inference, called pre-strategy intervention to implement the targeted intervention. Since MAIDs can be regarded as causal diagrams, the causal effect on desirable system outcomes is maximized, implying that the desirable outcomes are achieved. More importantly, the bundled relevance graph analysis from MAIDs is able to predict the solvability of coordination mechanisms via various MARL paradigms. In experiments, we demonstrate effectiveness of our proposed targeted intervention, and verify the result of relevance graph analysis.



Paperid:3091
Authors:Debargha Ganguly, Vikash Singh, Sreehari Sankar, Biyao Zhang, Xuecen Zhang, Srinivasan Iyengar, Xiaotian Han, Amit Sharma, Shivkumar Kalyanaraman, Vipin Chaudhary
Abstract:
Large language models (LLMs) show remarkable promise for democratizing automated reasoning by generating formal specifications. However, a fundamental tension exists: LLMs are probabilistic, while formal verification demands deterministic guarantees. This paper addresses this epistemological gap by comprehensively investigating failure modes and uncertainty quantification (UQ) in LLM-generated formal artifacts. Our systematic evaluation of five frontier LLMs reveals Satisfiability Modulo Theories (SMT) based autoformalization's domain-specific impact on accuracy (from +34.8\% on logical tasks to -44.5\% on factual ones), with known UQ techniques like the entropy of token probabilities failing to identify these errors. We introduce a probabilistic context-free grammar (PCFG) framework to model LLM outputs, yielding a refined uncertainty taxonomy. We find uncertainty signals are task-dependent (e.g., grammar entropy for logic, AUROC>0.93). Finally, a lightweight fusion of these signals enables selective verification, drastically reducing errors (14-100\%) with minimal abstention, transforming LLM-driven formalization into a reliable engineering discipline.



Paperid:3092
Authors:Qiangqiang Zhang, Chenfei Gu, Xinwei Feng, Jinhan Xie, Ting Li
Abstract:
Abstract:We propose an online conformal prediction framework under local differential privacy to address the emerging challenge of privacy-preserving uncertainty quantification in streaming data environments. Our method constructs dynamic, model-free prediction sets based on randomized binary inquiries, ensuring rigorous privacy protection without requiring access to raw data. Importantly, the proposed algorithm can be conducted in a one-pass online manner, leading to high computational efficiency and minimal storage requirements with $\mathcal{O}(1)$ space complexity, making it particularly suitable for real-time applications. The proposed framework is also broadly applicable to both regression and classification tasks, adapting flexibly to diverse predictive settings. We establish theoretical guarantees for long-run coverage at a target confidence level, ensuring statistical reliability under strict privacy constraints. Extensive empirical evaluations on both simulated and real-world datasets demonstrate that the proposed method delivers accurate, stable, and privacy-preserving predictions across a range of dynamic environments.



Paperid:3093
Authors:Weihan Xu, Yimeng Ma, Jingyue Huang, Yang Li, Wenye Ma, Taylor Berg-Kirkpatrick, Julian Mcauley, Paul Pu Liang, Hao-Wen Dong
Abstract:
Short videos are an effective tool for promoting contents and improving knowledge accessibility. While existing extractive video summarization methods struggle to produce a coherent narrative, existing abstractive methods cannot `quote' from the input videos, i.e., inserting short video clips in their outputs. In this work, we explore novel video editing models for generating shorts that feature a coherent narrative with embedded video insertions extracted from a long input video. We propose a novel retrieval-embedded generation framework that allows a large language model to quote multimodal resources while maintaining a coherent narrative. Our proposed REGen system first generates the output story script with quote placeholders using a finetuned large language model, and then uses a novel retrieval model to replace the quote placeholders by selecting a video clip that best supports the narrative from a pool of candidate quotable video clips. We examine the proposed method on the task of documentary teaser generation, where short interview insertions are commonly used to support the narrative of a documentary. Our objective evaluations show that the proposed method can effectively insert short video clips while maintaining a coherent narrative. In a subjective survey, we show that our proposed method outperforms existing abstractive and extractive approaches in terms of coherence, alignment, and realism in teaser generation.



Paperid:3094
Authors:Hang Yu, Lyumin Wu, Wenxin Zhou, Zhao Ren
Abstract:
Abstract:This paper studies nonparametric quantile regression using recurrent neural networks (RNNs) and sparse recurrent neural networks (SRNNs) to approximate the true conditional quantile function, assumed to exhibit a compositional structure referred to as the hierarchical interaction model. Specifically, we demonstrate that estimators based on RNNs and SRNNs with rectified linear unit (ReLU) activation function and properly chosen architectures achieve the optimal nonparametric convergence rate, up to a logarithmic factor, under stationary, exponentially $\boldsymbol{\beta}$-mixing observations. To establish this result, we derive approximation error bounds for functions within the hierarchical interaction model using RNNs and SRNNs. Our analysis leverages the close connections between sparse feedforward neural networks (SFNNs) and SRNNs. Numerical experiments and an empirical study on Gross Domestic Product (GDP) forecast further support the proposed theory.



Paperid:3095
Authors:Zihan Zhang, Yuxin Chen, Jason Lee, Simon Du, Lin Yang, Ruosong Wang
Abstract:
Abstract:We study deployment-efficient reward-free exploration with linear function approximation, where the goal is to explore a linear Markov Decision Process (MDP) without revealing the reward function, while minimizing the number of distinct policies implemented during learning. By ``deployment efficient'', we mean algorithms that require few policies deployed during exploration -- crucial in real-world applications where such deployments are costly or disruptive. We design a novel reinforcement learning algorithm that achieves near-optimal deployment efficiency for linear MDPs in the reward-free setting, using at most $O(H)$ exploration policies during execution (where $H$ is the horizon length), while maintaining sample complexity polynomial in feature dimension and horizon length. Unlike previous approaches with similar deployment efficiency guarantees, our algorithm's sample complexity is independent of the reachability or explorability coefficients of the underlying MDP, which can be arbitrarily small and lead to unbounded sample complexity in certain cases -- directly addressing an open problem from prior work. Our technical contributions include a data-dependent method for truncating state-action pairs in linear MDPs, efficient offline policy evaluation and optimization algorithms for these truncated MDPs, and a careful integration of these components to implement reward-free exploration with linear function approximation without sacrificing deployment efficiency.



Paperid:3096
Authors:Gang Yan, Sikai Yang, Wan Du
Abstract:
Integrating large pre-trained models into federated learning (FL) can significantly improve generalization and convergence efficiency. A widely adopted strategy freezes the pre-trained backbone and fine-tunes a lightweight task head, thereby reducing computational and communication costs. However, this partial fine-tuning paradigm introduces new security risks, making the system vulnerable to poisoned updates and backdoor attacks. To address these challenges, we propose FedRACE, a unified framework for robust FL with partially frozen models. FedRACE comprises two core components: HStat-Net, a hierarchical network that refines frozen features into compact, linearly separable representations; and DevGuard, a server-side mechanism that detects malicious clients by evaluating statistical deviance in class-level predictions modeling generalized linear models (GLMs). DevGuard further incorporates adaptive thresholding based on theoretical misclassification bounds and employs randomized majority voting to enhance detection reliability. We implement FEDRACE on the FedScale platform and evaluate it on CIFAR-100, Food-101, and Tiny ImageNet under diverse attack scenarios. FedRACE achieves a true positive rate of up to 99.3% with a false positive rate below 1.2%, while preserving model accuracy and improving generalization.



Paperid:3097
Authors:Adibvafa Fallahpour, Andrew Magnuson, Purav Gupta, Shihao Ma, Jack Naimer, Arnav Shah, Haonan Duan, Omar Ibrahim, Hani Goodarzi, Chris Maddison, Bo Wang
Abstract:
Unlocking deep, interpretable biological reasoning from complex genomic data is a paramount challenge for artificial intelligence, hindering critical scientific discovery. Existing DNA foundation models, despite their powerful sequence representation capabilities, often struggle with multi-step reasoning and lack inherent mechanisms for transparent, biologically intuitive explanations. We present BioReason, a pioneering architecture, that for the first time deeply integrates a DNA foundation model with a large language model (LLM). This novel connection empowers the LLM to directly process and reason with genomic information as a fundamental input modality, enabling a new form of multimodal biological understanding. BioReason's capacity for sophisticated, multi-step reasoning is cultivated through a regimen of supervised fine-tuning and targeted reinforcement learning, guiding the integrated system to generate logical and biologically coherent deductions. On challenging benchmarks, including KEGG-based disease pathway prediction—where BioReason improves accuracy by roughly 10 points (from 88% to 97%)—and variant effect analysis, BioReason demonstrates an average performance gain of 15% over strong single-modality baselines. A key breakthrough is BioReason's ability to reason over previously unseen biological entities and articulate its decision-making process through interpretable, step-by-step biological traces mechanistically supporting its predictions. BioReason offers a transformative approach for AI in biology, paving the way for deeper mechanistic insights and accelerated generation of testable hypotheses from genomic data.



Paperid:3098
Authors:Awa Khouna, Julien Ferry, Thibaut Vidal
Abstract:
The advent of Machine Learning as a Service (MLaaS) has heightened the trade-off between model explainability and security. In particular, explainability techniques, such as counterfactual explanations, inadvertently increase the risk of model extraction attacks, enabling unauthorized replication of proprietary models. In this paper, we formalize and characterize the risks and inherent complexity of model reconstruction, focusing on the "oracle'' queries required for faithfully inferring the underlying prediction function. We present the first formal analysis of model extraction attacks through the lens of competitive analysis, establishing a foundational framework to evaluate their efficiency. Focusing on models based on additive decision trees (e.g., decision trees, gradient boosting, and random forests), we introduce novel reconstruction algorithms that achieve provably perfect fidelity while demonstrating strong anytime performance. Our framework provides theoretical bounds on the query complexity for extracting tree-based model, offering new insights into the security vulnerabilities of their deployment.



Paperid:3099
Authors:Chengye Yu, Tianyu Wang, Zili Shao, Song Jiang
Abstract:
Abstract:Large Language Models (LLMs) for multi-turn conversations suffer from inefficiency: semantically similar queries across different user sessions trigger redundant computation and duplicate memory-intensive Key-Value (KV) caches. Existing optimizations such as prefix caching overlook semantic similarities, while typical semantic caches either ignore conversational context or are not integrated with low-level KV cache management.We propose SmartCache, a system-algorithm co-design framework that tackles this inefficiency by exploiting semantic query similarity across sessions. SmartCache leverages a Semantic Forest structure to hierarchically index conversational turns, enabling efficient retrieval and reuse of responses only when both the semantic query and conversational context match.To maintain accuracy during topic shifts, it leverages internal LLM attention scores—computed during standard prefill—to dynamically detect context changes with minimal computational overhead. Importantly, this semantic understanding is co-designed alongside the memory system: a novel two-level mapping enables transparent cross-session KV cache sharing for semantically equivalent states, complemented by a semantics-aware eviction policy that significantly improves memory utilization. This holistic approach significantly reduces redundant computations and optimizes GPU memory utilization. The evaluation demonstrates SmartCache's effectiveness across multiple benchmarks. On the CoQA and SQuAD datasets, SmartCache reduces KV cache memory usage by up to $59.1\%$ compared to prefix caching and $56.0\%$ over semantic caching, while cutting Time-to-First-Token (TTFT) by $78.0\%$ and $71.7\%$, respectively. It improves answer quality metrics, achieving $39.9\%$ higher F1 and $39.1\%$ higher ROUGE-L for Qwen-2.5-1.5B on CoQA. The Semantic-aware Tiered Eviction Policy (STEP) outperforms LRU/LFU by $29.9\%$ in reuse distance under skewed workloads.



Paperid:3100
Authors:Denizhan Kara, Tomoyoshi Kimura, Jinyang Li, Bowen He, Yizhuo Chen, Yigong Hu, Hongjue Zhao, Shengzhong Liu, Tarek Abdelzaher
Abstract:
Learning robust representations from unlabeled time series is crucial, and contrastive learning offers a promising avenue. However, existing contrastive learning approaches for time series often struggle with defining meaningful similarities, tending to overlook inherent physical correlations and diverse, sequence-varying non-stationarity. This limits their representational quality and real-world adaptability. To address these limitations, we introduce AdaTS, a novel adaptive soft contrastive learning strategy. AdaTS offers a compute-efficient solution centered on dynamic instance-wise and temporal assignments to enhance time series representations, specifically by: (i) leveraging Time-Frequency Coherence for robust physics-guided similarity measurement; (ii) preserving relative instance similarities through ordinal consistency learning; and (iii) dynamically adapting to sequence-specific non-stationarity with dynamic temporal assignments. AdaTS is designed as a pluggable module to standard contrastive frameworks, achieving up to 13.7% accuracy improvements across diverse time series datasets and three state-of-the-art contrastive frameworks while enhancing robustness against label scarcity. The code will be publicly available upon acceptance.



Paperid:3101
Authors:Xun Huang, Zhengqi Li, Guande He, Mingyuan Zhou, Eli Shechtman
Abstract:
We introduce Self Forcing, a novel training paradigm for autoregressive video diffusion models. It addresses the longstanding issue of exposure bias—where models trained on ground-truth context must generate sequences conditioned on their own imperfect outputs during inference. Unlike prior methods that denoise future frames based on ground-truth context frames, Self Forcing conditions each frame’s generation on previously self-generated outputs by performing autoregressive rollout with key-value~(KV) caching during training. This strategy enables supervision through a holistic loss at the video level that directly evaluates the quality of the entire generated sequence, rather than relying solely on traditional frame-wise objectives. To ensure training efficiency, we employ a few-step diffusion model along with a stochastic gradient truncation strategy, effectively balancing computational cost and performance. We further introduce a rolling KV cache mechanism that enables efficient autoregressive video extrapolation. Extensive experiments demonstrate that our approach achieves real-time streaming video generation with sub-second latency on a single GPU, while matching or even surpassing the generation quality of significantly slower and non-causal diffusion models.



Paperid:3102
Authors:Yinhan He, Wendy Zheng, Song Wang, Zaiyi Zheng, Yushun Dong, Yaochen Zhu, Jundong Li
Abstract:
In-Context Learning (ICL) is a technique where large language models (LLMs) leverage multiple demonstrations (i.e., examples) to perform tasks. With the recent expansion of LLM context windows, many-shot ICL (generally with more than 50 demonstrations) can lead to significant performance improvements on a variety of language tasks such as text classification and question answering.Nevertheless, ICL faces the issue of demonstration order instability (ICL-DOI), which means that performance varies significantly depending on the order of demonstrations. Moreover, ICL-DOI persists in many-shot ICL, validated by our thorough experimental investigation.Current strategies for handling ICL-DOI are not applicable to many-shot ICL due to two critical challenges: (1) Most existing methods assess demonstration order quality by first prompting the LLM, then using heuristic metrics based on the LLM's predictions. In the many-shot scenarios, these metrics without theoretical grounding become unreliable, where the LLMs struggle to effectively utilize information from long input contexts, making order distinctions less clear.(2) The requirement to examine all orders for the large number of demonstrations is computationally infeasible due to the super-exponential complexity of the order space in many-shot ICL. To tackle the first challenge, we design a demonstration order evaluation metric based on information theory for measuring order quality, which effectively quantifies the usable information gain of a given demonstration order.To address the second challenge, we propose a hierarchical demonstration order optimization method namedHIDOthat enables a more refined exploration of the order space, achieving high ICL performance without the need to evaluate all possible orders. Extensive experiments on multiple LLMs and real-world datasets demonstrate that our HIDO method consistently and efficiently outperforms other baselines. Our code project can be found at https://anonymous.4open.science/r/HIDO-B2DE/.



Paperid:3103
Authors:Kira Goldner, Taylor Lundy
Abstract:
Abstract:We initiate the study of multidimensional Bayesian utility maximization, focusing on the unit-demand setting where values are i.i.d. across both items and buyers. The seminal result of Hartline and Roughgarden '08 studies simple, information-robust mechanisms that maximize utility for $n$ i.i.d. agents and $m$ identical items via an approximation to social welfare as an upper bound, and they prove this gap between optimal utility and social welfare is $\Theta(1+\log{n/m})$ in this setting. We extend these results to the multidimensional setting. To do so, we develop simple, prior-independent, approximately-optimal mechanisms, targeting the simplest benchmark of optimal welfare. We give a $(1-1/e)$-approximation when there are more items than buyers, and a $\Theta(\log{n/m})$-approximation when there are more buyers than items, and we prove that this bound is tight in both $n$ and $m$ by reducing the i.i.d. unit-demand setting to the identical items setting. Finally, we include an extensive discussion section on why Bayesian utility maximization is a promising research direction. In particular, we characterize complexities in this setting that defy our intuition from the welfare and revenue literature, and motivate why coming up with a better benchmark than welfare is a hard problem itself.



Paperid:3104
Authors:Fengshuo Bai, Rui Zhao, Hongming Zhang, Sijia Cui, Shao Zhang, Ying Wen, Yaodong Yang, bo xu, Lei Han
Abstract:
Abstract:Preference-based reinforcement learning (PbRL) bypasses complex reward engineering by learning from human feedback. However, due to the high cost of obtaining feedback, PbRL typically relies on a limited set of preference-labeled samples. This data scarcity introduces two key inefficiencies: (1) the reward model overfits to the limited feedback, leading to poor generalization to unseen samples, and (2) the agent exploits the learned reward model, exacerbating overestimation of action values in temporal difference (TD) learning. To address these issues, we propose STAR, an efficient PbRL method that integrates preference margin regularization and policy regularization. Preference margin regularization mitigates overfitting by introducing a bounded margin in reward optimization, preventing excessive bias toward specific feedback. Policy regularization bootstraps a conservative estimate $\widehat{Q}$ from well-supported state-action pairs in the replay memory, reducing overestimation during policy learning. Experimental results show that STAR improves feedback efficiency, achieving 34.8\% higher performance in online settings and 29.7\% in offline settings compared to state-of-the-art methods. Ablation studies confirm that STAR facilitates more robust reward and value function learning. The videos of this project are released at https://sites.google.com/view/pbrl-star.



Paperid:3105
Authors:Guansheng Peng, Lining Xing, Fuyan Ma, Aldy Gunawan, Guopeng Song, Pieter Vansteenwegen
Abstract:
Recent years have witnessed a surge of interest in solving combinatorial optimization problems (COPs) using machine learning techniques. Motivated by this trend, we propose a learning-augmented exact approach for tackling an NP-hard COP, the Orienteering Problem with Time Windows, which aims to maximize the total score collected by visiting a subset of vertices in a graph within their time windows. Traditional exact algorithms rely heavily on domain expertise and meticulous design, making it hard to achieve further improvements. By leveraging deep learning models to learn effective relaxations of problem restrictions from data, our approach enables significant performance gains in an exact dynamic programming algorithm. We propose a novel graph convolutional network that predicts the directed edges defining the relaxation. The network is trained in a supervised manner, using optimal solutions as high-quality labels. Experimental results demonstrate that the proposed learning-augmented algorithm outperforms the state-of-the-art exact algorithm, achieving a 38% speedup on Solomon’s benchmark and more than a sevenfold improvement on the more challenging Cordeau’s benchmark.



Paperid:3106
Authors:Jaehyun Nam, Jinsung Yoon, Jiefeng Chen, Jinwoo Shin, Sercan Arik, Tomas Pfister
Abstract:
Agents based on large language models (LLMs) for machine learning engineering (MLE) can automatically implement ML models via code generation. However, existing approaches to build such agents often rely heavily on inherent LLM knowledge and employ coarse exploration strategies that modify the entire code structure at once. This limits their ability to select effective task-specific models and perform deep exploration within specific components, such as experimenting extensively with feature engineering options. To overcome these, we propose MLE-STAR, a novel approach to build MLE agents. MLE-STAR first leverages external knowledge by using a search engine to retrieve effective models from the web, forming an initial solution, then iteratively refines it by exploring various strategies targeting specific ML components. This exploration is guided by ablation studies analyzing the impact of individual code blocks. Furthermore, we introduce a novel ensembling method using an effective strategy suggested by MLE-STAR. Our experimental results show that MLE-STAR achieves medals in 44% of the Kaggle competitions on the MLE-bench, significantly outperforming the best alternative.



Paperid:3107
Authors:Jiawei Gu, Ziyue Qiao, Zechao Li
Abstract:
Graph Neural Networks (GNNs) are increasingly deployed in mission-critical tasks, yet they often encounter inputs that lie outside their training distribution, leading to unreliable or overconfident predictions. To address this limitation, we present RAGNOR (Robust Aggregation Graph Norm for Outlier Recognition), a post-hoc approach that leverages embedding norms for robust out-of-distribution (OOD) detection on both node-level and graph-level tasks. Unlike previous methods designed primarily for image domains, RAGNOR directly tackles the relational challenges intrinsic to graphs: local contamination by anomalous neighbors, disparate norm scales across classes or roles, and insufficient references for boundary or low-degree nodes. By combining global Z-score normalization, median-based local aggregation, and multi-hop blending, RAGNOR effectively refines raw norm signals into robust OOD scores while incurring minimal overhead and requiring no retraining of the original GNN. Experimental evaluations on multiple benchmarks demonstrate that RAGNOR not only achieves competitive or superior detection performance compared to alternative techniques, but also provides an intuitive, modular design that can be readily integrated into existing graph pipelines.



Paperid:3108
Authors:Mirali Purohit, Bimal Gajera, Vatsal Malaviya, Irish Mehta, Kunal Kasodekar, Jacob Adler, Steven Lu, Umaa Rebbapragada, Hannah Kerner
Abstract:
Foundation models have enabled rapid progress across many specialized domains by leveraging large-scale pre-training on unlabeled data, demonstrating strong generalization to a variety of downstream tasks. While such models have gained significant attention in fields like Earth Observation, their application to Mars science remains limited. A key enabler of progress in other domains has been the availability of standardized benchmarks that support systematic evaluation. In contrast, Mars science lacks such benchmarks and standardized evaluation frameworks, which have limited progress toward developing foundation models for Martian tasks. To address this gap, we introduce Mars-Bench, the first benchmark designed to systematically evaluate models across a broad range of Mars-related tasks using both orbital and surface imagery. Mars-Bench comprises 20 datasets spanning classification, segmentation, and object detection, focused on key geologic features such as craters, cones, boulders, and frost. We provide standardized, ready-to-use datasets and baseline evaluations using models pre-trained on natural images, Earth satellite data, and state-of-the-art vision-language models. Results from all analyses suggest that Mars-specific foundation models may offer advantages over general-domain counterparts, motivating further exploration of domain-adapted pre-training. Mars-Bench aims to establish a standardized foundation for developing and comparing machine learning models for Mars science.



Paperid:3109
Authors:Xiangyu Zeng, Kefan Qiu, Qingyu Zhang, Xinhao Li, Jing Wang, Jiaxin Li, Ziang Yan, Kun Tian, Meng Tian, Xinhai Zhao, Yi Wang, Limin Wang
Abstract:
Multimodal Large Language Models (MLLMs) have recently achieved remarkable progress in video understanding. However, their effectiveness in real-time streaming scenarios remains limited due to storage constraints of historical visual features and insufficient real-time spatiotemporal reasoning. To address these challenges, we propose StreamForest, a novel architecture specifically designed for streaming video understanding. Central to StreamForest is the Persistent Event Memory Forest, a memory mechanism that adaptively merges video frames into multiple event-level tree structures. This process is guided by penalty functions based on temporal distance, content similarity, and merge frequency, enabling efficient long-term memory retention under limited computational resources. To enhance real-time perception, we introduce a Fine-grained Spatiotemporal Window, which captures detailed short-term visual cues to improve current scene understanding. Additionally, we present OnlineIT, an instruction-tuning dataset tailored for streaming video tasks. OnlineIT significantly boosts MLLM performance in both real-time perception and future prediction. To evaluate generalization in practical applications, we introduce ODV-Bench, a benchmark focused on real-time streaming video understanding in autonomous driving scenarios. Experimental results demonstrate that StreamForest achieves state-of-the-art performance, with accuracies of 77.3% on StreamingBench, 60.5% on OVBench, and 55.6% on OVO-Bench. Notably, even under extreme visual token compression (limited to 1024 tokens), the model retains 96.7% of its average accuracy across eight benchmarks relative to the default setting. These results underscore the robustness, efficiency, and generalizability of StreamForest for streaming video understanding.



Paperid:3110
Authors:Ziru Wang, Mengmeng Wang, Guang Dai, Yongliu Long, Jingdong Wang
Abstract:
Although learning 3D manipulation policies from monocular RGB images is lightweight and deployment-friendly, the lack of structural information often leads to inaccurate action estimation. While explicit 3D inputs can mitigate this issue, they typically require additional sensors and introduce data acquisition overhead. An intuitive alternative is to incorporate a pre-trained depth estimator; however, this often incurs substantial inference-time cost. To address this, we propose MonoLift, a tri-level distillation framework that transfers spatial, temporal, and action-level knowledge from a depth-guided teacher to a monocular RGB student. By jointly distilling geometry-aware features, temporal dynamics, and policy behaviors during training, MonoLift enables the student model to perform 3D-aware reasoning and precise control at deployment using only monocular RGB input. Extensive experiments on both simulated and real-world manipulation tasks show that MonoLift not only outperforms existing monocular approaches but even surpasses several methods that rely on explicit 3D input, offering a resource-efficient and effective solution for vision-based robotic control.



Paperid:3111
Authors:Jiale Deng, Yanyan Shen, Ziyuan Pei, Youmin Chen, Linpeng Huang
Abstract:
Retrieval-Augmented Generation (RAG) addresses large language model (LLM) hallucinations by grounding responses in external knowledge, but its effectiveness is compromised by poor-quality retrieved contexts containing irrelevant or noisy information. While existing approaches attempt to improve performance through context selection based on predefined context quality assessment metrics, they show limited gains over standard RAG. We attribute this limitation to their failure in holistically utilizing available information (query, context list, and generator) for comprehensive quality assessment. Inspired by recent advances in data selection, we reconceptualize context quality assessment as an inference-time data valuation problem and introduce the Contextual Influence Value (CI value). This novel metric quantifies context quality by measuring the performance degradation when removing each context from the list, effectively integrating query-aware relevance, list-aware uniqueness, and generator-aware alignment. Moreover, CI value eliminates complex selection hyperparameter tuning by simply retaining contexts with positive CI values. To address practical challenges of label dependency and computational overhead, we develop a parameterized surrogate model for CI value prediction during inference. The model employs a hierarchical architecture that captures both local query-context relevance and global inter-context interactions, trained through oracle CI value supervision and end-to-end generator feedback. Extensive experiments across 8 NLP tasks and multiple LLMs demonstrate that our context selection method significantly outperforms state-of-the-art baselines, effectively filtering poor-quality contexts while preserving critical information. Code is available at the anonymous link: https://anonymous.4open.science/r/CSM-4002.



Paperid:3112
Authors:hanxue liang, Jiawei Ren, Ashkan Mirzaei, Antonio Torralba, Ziwei Liu, Igor Gilitschenski, Sanja Fidler, Cengiz Oztireli, Huan Ling, Zan Gojcic, Jiahui Huang
Abstract:
Recent advancements in static feed-forward scene reconstruction have demonstrated significant progress in high-quality novel view synthesis. However, these models often struggle with generalizability across diverse environments and fail to effectively handle dynamic content. We present BTimer (short for Bullet Timer), the first motion-aware feed-forward model for real-time reconstruction and novel view synthesis of dynamic scenes. Our approach reconstructs the full scene in a 3D Gaussian Splatting representation at a given target (‘bullet’) timestamp by aggregating information from all the context frames. Such a formulation allows BTimer to gain scalability and generalization by leveraging both static and dynamic scene datasets. Given a casual monocular dynamic video, BTimer reconstructs a bullet-time scene within 150ms while reaching state-of-the-art performance on both static and dynamic scene datasets, even compared with optimization-based approaches.



Paperid:3113
Authors:Donghao Luo, Xue Wang
Abstract:
In this paper, we pioneer the idea of implicit decomposition. And based on this idea, we propose a powerful decomposition-based enhancement framework, namely DecompNet. Our method converts the time series decomposition into an implicit process, where it can give a time series model the decomposition-related knowledge during inference, even though this model does not actually decompose the input time series. Thus, our DecompNet can enable a model to inherit the performance promotion brought by time series decomposition but will not introduce any additional inference costs, successfully enhancing the model performance while enjoying better efficiency. Experimentally, our DecompNet exhibits promising enhancement capability and compelling framework generality. Especially, it can also enhance the performance of the latest and state-of-the-art models, greatly pushing the performance limit of time series forecasting. Through comprehensive comparisons, DecompNet also shows excellent performance and efficiency superiority, making the decomposition-based enhancement framework surpass the well-recognized normalization-based frameworks for the first time.



Paperid:3114
Authors:Zhening Li, Armando Solar-Lezama, Yisong Yue, Stephan Zheng
Abstract:
We introduce a new approach toagent programming, the development of LLM-based agents.Current approaches to agent programming often entangle two aspects of agent design: the core workflow logic and the inference-time strategy (e.g., tree search). We introduceprobabilistic angelic nondeterminism(PAN),a programming model that disentangles these two concerns,allowing the programmer to describe the agent workflow andindependently experiment with different inference-time strategiesby simply changing a few inputs.We provide an implementation of PAN in Python as the EnCompass framework, which uses a Python decorator to compile agent workflow programs into a search space.We present three case studies that demonstrate how the framework lets the programmer quickly improve the reliability of an agent and easily switch between different inference-time strategies, all with little additional coding.



Paperid:3115
Authors:Tao Liu, Dafeng Zhang, Gengchen Li, Shizhuo Liu, yongqi song, Senmao Li, Shiqi Yang, Boqian Li, KAI WANG, Yaxing Wang
Abstract:
Abstract:Face aging has become a crucial task in computer vision, with applications ranging from entertainment to healthcare. However, existing methods struggle with achieving a realistic and seamless transformation across the entire lifespan, especially when handling large age gaps or extreme head poses. The core challenge lies in balancing $age\ accuracy$ and $identity\ preservation$—what we refer to as the $Age\text{-}ID\ trade\text{-}off$. Most prior methods either prioritize age transformation at the expense of identity consistency or vice versa. In this work, we address this issue by proposing a $two\text{-}pass$ face aging framework, named $Cradle2Cane$, based on few-step text-to-image (T2I) diffusion models. The first pass focuses on solving $age\ accuracy$ by introducing an adaptive noise injection ($AdaNI$) mechanism. This mechanism is guided by including prompt descriptions of age and gender for the given person as the textual condition.Also, by adjusting the noise level, we can control the strength of aging while allowing more flexibility in transforming the face.However, identity preservation is weakly ensured here to facilitate stronger age transformations.In the second pass, we enhance $identity\ preservation$ while maintaining age-specific features by conditioning the model on two identity-aware embeddings ($IDEmb$): $SVR\text{-}ArcFace$ and $Rotate\text{-}CLIP$. This pass allows for denoising the transformed image from the first pass, ensuring stronger identity preservation without compromising the aging accuracy.Both passes are $jointly\ trained\ in\ an\ end\text{-}to\text{-}end\ way\$. Extensive experiments on the CelebA-HQ test dataset, evaluated through Face++ and Qwen-VL protocols, show that our $Cradle2Cane$ outperforms existing face aging methods in age accuracy and identity consistency.Additionally, $Cradle2Cane$ demonstrates superior robustness when applied to in-the-wild human face images, where prior methods often fail. This significantly broadens its applicability to more diverse and unconstrained real-world scenarios.



Paperid:3116
Authors:Guo Zeng, Jean Honorio
Abstract:
Many online learning applications adopt the stochastic bandit problem with a linear reward model, where the unknown bandit parameter exhibits a succinct structure. We study minimax regret lower bounds which allow to know whether more efficient algorithms can be proposed. We introduce a general definition of succinctness and propose a novel framework for constructing minimax regret lower bounds based on an information-regret trade-off. When applied to entry-sparse vectors, our framework sharpens a recent lower bound by (Hao et al, NeurIPS 2020). We further apply our framework to derive novel results. To the best of our knowledge, we provide the first lower bounds for the group-sparse and low-rank matrix settings.



Paperid:3117
Authors:Xin Gao, Jiyao Liu, Guanghao Li, Yueming Lyu, Jianxiong Gao, Weichen Yu, Ningsheng Xu, Liang Wang, Caifeng Shan, Ziwei Liu, Chenyang Si
Abstract:
Recent advancements have explored text-to-image diffusion models for synthesizing out-of-distribution (OOD) samples, substantially enhancing the performance of OOD detection. However, existing approaches typically rely on perturbing text-conditioned embeddings, resulting in semantic instability and insufficient shift diversity, which limit generalization to realistic OOD. To address these challenges, we propose GOOD, a novel and flexible framework that directly guides diffusion sampling trajectories towards OOD regions using off-the-shelf in-distribution (ID) classifiers. GOOD incorporates dual-level guidance: (1) Image-level guidance based on the gradient of log partition to reduce input likelihood, drives samples toward low-density regions in pixel space. (2) Feature-level guidance, derived from k-NN distance in the classifier’s latent space, promotes sampling in feature-sparse regions. Hence, this dual-guidance design enables more controllable and diverse OOD sample generation. Additionally, we introduce a unified OOD score that adaptively combines image and feature discrepancies, enhancing detection robustness. We perform thorough quantitative and qualitative analyses to evaluate the effectiveness of GOOD, demonstrating that training with samples generated by GOOD can notably enhance OOD detection performance.



Paperid:3118
Authors:Colin Doumont, Victor Picheny, Slava Borovitskiy, Henry Moss
Abstract:
Bayesian Optimization (BO) has the potential to solve various combinatorial tasks, ranging from materials science to neural architecture search. However, BO requires specialized kernels to effectively model combinatorial domains. Recent efforts have introduced several combinatorial kernels, but the relationships among them are not well understood. To bridge this gap, we develop a unifying framework based on heat kernels, which we derive in a systematic way and express as simple closed-form expressions. Using this framework, we prove that many successful combinatorial kernels are either related or equivalent to heat kernels, and validate this theoretical claim in our experiments. Moreover, our analysis confirms and extends the results presented in Bounce: certain algorithms' performance decreases substantially when the unknown optima of the function do not have a certain structure. In contrast, heat kernels are not sensitive to the location of the optima. Lastly, we show that a fast and simple pipeline, relying on heat kernels, is able to achieve state-of-the-art results, matching or even outperforming certain slow or complex algorithms.



Paperid:3119
Authors:Haoyu Zhang, WentaoZhang, Hao Miao, Xinke Jiang, Yuchen Fang, Yifan Zhang
Abstract:
Spatio-Temporal Graph Neural Networks (STGNNs) have emerged as a powerful tool for modeling dynamic graph-structured data across diverse domains. However, they often fail to generalize in Spatio-Temporal Out-of-Distribution (STOOD) scenarios, where both temporal dynamics and spatial structures evolve beyond the training distribution. To address this problem, we propose an innovative Spatio-Temporal Retrieval-Augmented Pattern Learning framework, STRAP, which enhances model generalization by integrating retrieval-augmented learning into the STGNN continue learning pipeline.The core of STRAP is a compact and expressive pattern library that stores representative spatio-temporal patterns enriched with historical, structural, and semantic information, which is obtained and optimized during the training phase.During inference, STRAP retrieves relevant patterns from this library based on similarity to the current input and injects them into the model via a plug-and-play prompting mechanism. This not only strengthens spatio-temporal representations but also mitigates catastrophic forgetting. Moreover, STRAP introduces a knowledge-balancing objective to harmonize new information with retrieved knowledge.Extensive experiments across multiple real-world streaming graph datasets show that STRAP consistently outperforms state-of-the-art STGNN baselines on STOOD tasks, demonstrating its robustness, adaptability, and strong generalization capability without task-specific fine-tuning.



Paperid:3120
Authors:Kaitao Huang, Yan Yan, Jing-Hao Xue, Hanzi Wang
Abstract:
3D GAN inversion projects a single image into the latent space of a pre-trained 3D GAN to achieve single-shot novel view synthesis, which requires visible regions with high fidelity and occluded regions with realism and multi-view consistency. However, existing methods focus on the reconstruction of visible regions, while the generation of occluded regions relies only on the generative prior of 3D GAN. As a result, the generated occluded regions often exhibit poor quality due to the information loss caused by the low bit-rate latent code. To address this, we introduce the warping-and-inpainting strategy to incorporate image inpainting into 3D GAN inversion and propose a novel 3D GAN inversion method, WarpGAN. Specifically, we first employ a 3D GAN inversion encoder to project the single-view image into a latent code that serves as the input to 3D GAN. Then, we perform warping to a novel view using the depth map generated by 3D GAN. Finally, we develop a novel SVINet, which leverages the symmetry prior and multi-view image correspondence w.r.t. the same latent code to perform inpainting of occluded regions in the warped image. Quantitative and qualitative experiments demonstrate that our method consistently outperforms several state-of-the-art methods.



Paperid:3121
Authors:Yinan Huang, Haoteng Yin, Eli Chien, Rongzhe Wei, Pan Li
Abstract:
Learning with relational and network-structured data is increasingly vital in sensitive domains where protecting the privacy of individual entities is paramount. Differential Privacy (DP) offers a principled approach for quantifying privacy risks, with DP-SGD emerging as a standard mechanism for private model training. However, directly applying DP-SGD to relational learning is challenging due to two key factors: (i) entities often participate in multiple relations, resulting in high and difficult-to-control sensitivity; and (ii) relational learning typically involves multi-stage, potentially coupled (interdependent) sampling procedures that make standard privacy amplification analyses inapplicable. This work presents a principled framework for relational learning with formal entity-level DP guarantees. We provide a rigorous sensitivity analysis and introduce an adaptive gradient clipping scheme that modulates clipping thresholds based on entity occurrence frequency. We also extend the privacy amplification results to a tractable subclass of coupled sampling, where the dependence arises only through sample sizes. These contributions lead to a tailored DP-SGD variant for relational data with provable privacy guarantees. Experiments on fine-tuning text encoders over text-attributed network-structured relational data demonstrate the strong utility-privacy trade-offs of our approach.



Paperid:3122
Authors:Daniel Dsouza, Julia Kreutzer, Adrien Morisot, Ahmet Üstün, Sara Hooker
Abstract:
One of the most profound challenges of modern machine learning is performingwell on the long-tail of rare and underrepresented features. Large general-purposemodels are trained for many tasks, but work best on high-frequency use cases.After training, it is hard to adapt a model to perform well on specific use casesunderrepresented in the training corpus. Relying on prompt engineering or few-shotexamples to maximize the output quality on a particular test case can be frustrating,as models can be highly sensitive to small changes, react in unpredicted waysor rely on a fixed system prompt for maintaining performance. In this work, weask: Can we optimize our training protocols to both improve controllability andperformance on underrepresented use cases at inference time? We revisit the dividebetween training and inference techniques to improve long-tail performance whileproviding users with a set of control levers the model is trained to be responsiveto. We create a detailed taxonomy of data characteristics and task provenance toexplicitly control generation attributes and implicitly condition generations atinference time. We fine-tune a base model to infer these markers automatically,which makes them optional at inference time. This principled and flexible approachyields pronounced improvements in performance on examples from the long tailof the training distribution. Overall, we observe lifts of 5.7% across all tasks.However, treasure markers are particularly effective at finding difficult to obtaingains in the long-tail. We observe relative lifts of up to 14.1% on underrepresentedtasks like CodeRepair and absolute improvements of 35.3% on length instructionfollowing evaluations.



Paperid:3123
Authors:Anujraaj Goyal, Guocheng Qian, n n, Aarush Gupta, Himmy Tam, Daniil Ostashev, Ju Hu, Dhritiman Sagar, Sergey Tulyakov, Kfir Aberman, Kuan-Chieh Wang
Abstract:
Adapter-based training has emerged as a key mechanism for extending the capabilities of powerful foundation image generators, enabling personalized and stylized text-to-image synthesis.These adapters are typically trained to capture a specific target attribute, such as subject identity, using single-image reconstruction objectives. However, because the input image inevitably contains a mixture of visual factors, adapters are prone to entangling the target attribute with incidental ones such as pose, expression, or lighting. This spurious correlation problem limits generalization and undermines modularity: an identity adapter unwittingly encodes pose and expression, failing to adhere to prompts that require changing the subject's pose and expression. In this work, we uncover a simple yet effective solution: provide the very shortcuts we wish to eliminate. InShortcut-Rerouted Adapter Training, confounding factors are routed through auxiliary modules, such as ControlNet or LoRA, during training, eliminating the incentive for the adapter to internalize them. Applied to tasks like facial and full-body identity injection, our approach improves generation quality, diversity, and prompt adherence. These results point to a general design principle in the era of large models: when seeking disentangled representations, the most effective path may be to establish shortcuts for what shouldnotbe learned.



Paperid:3124
Authors:Eric Balkanski, Nicholas DeFilippis, Vasilis Gkatzelis, Xizhi Tan
Abstract:
Abstract:We study the problem of designing procurement auctions for the strategic uncapacitated facility location problem: an agency needs to procure a set of facility locations in order to open new facilities for its customers. The owner of each location has a private cost for providing access to that facility (e.g., renting it or selling it) and needs to be compensated accordingly. The goal is to design auctions that decide where to open the facilities and how much to pay the corresponding owners, aiming to minimize the total cost, i.e., the opening cost paid to the owners and the total distance between the customers and the opened facilities. We evaluate the performance of these auctions using the \emph{frugality ratio}.We first analyze the performance of the classic VCG auction in this context and prove that its frugality ratio is exactly $3$. We then leverage the learning-augmented framework and design auctions that are augmented with predictions regarding the owner's true opening costs. Specifically, we propose a family of learning-augmented auctions that achieve significant payment reductions, leading to much better frugality ratios, when the predictions are accurate. At the same time, we demonstrate that these auctions remain robust even if the predictions are arbitrarily inaccurate, and maintain reasonable frugality ratios even under adversarially inaccurate predictions. We then extend this result to provide a family of ``error-tolerant'' auctions that perform well even if the prediction is approximately accurate, and we provide bounds regarding their frugality ratio as a function of the prediction error.



Paperid:3125
Authors:Wenjie Li, Xiangyi Wang, Heng Guo, Guangwei Gao, Zhanyu Ma
Abstract:
Old-photo face restoration poses significant challenges due to compounded degradations such as breakage, fading, and severe blur. Existing pre-trained diffusion-guided methods either rely on explicit degradation priors or global statistical guidance, which struggle with localized artifacts or face color. We propose Self-Supervised Selective-Guided Diffusion (SSDiff), a framework that leverages pseudo-reference faces generated by a pre-trained diffusion model under weak guidance. These pseudo-labels exhibit structurally aligned contours and natural colors, enabling region-specific restoration via staged supervision: structural guidance applied throughout the denoising process and color refinement in later steps, aligned with the coarse-to-fine nature of diffusion. By incorporating face parsing maps and scratch masks, our method selectively restores breakage regions while avoiding identity mismatch. We further construct VintageFace, a 300-image benchmark of real old face photos with varying degradation levels. SSDiff outperforms existing GAN-based and diffusion-based methods in perceptual quality, fidelity, and regional controllability. Code and data will be released upon acceptance.



Paperid:3126
Authors:Kaichen Zhang, Yuzhong Hong, Junwei Bao, Hongfei Jiang, Yang Song, Hong Dingqian, Hui Xiong
Abstract:
Post-training plays a crucial role in refining and aligning large language models to meet specific tasks and human preferences. While recent advancements in post-training techniques, such as Group Relative Policy Optimization (GRPO), leverage increased sampling with relative reward scoring to achieve superior performance, these methods often suffer from training instability that limits their practical adoption. To address this challenge, we present Group Variance Policy Optimization (GVPO). GVPO incorporates the analytical solution to KL-constrained reward maximization directly into its gradient weights, ensuring alignment with the optimal policy. The method provides intuitive physical interpretations: its gradient mirrors the mean squared error between the central distance of implicit rewards and that of actual rewards. GVPO offers two key advantages: (1) it guarantees a unique optimal solution, exactly the KL-constrained reward maximization objective, (2) it supports flexible sampling distributions that avoids on-policy and importance sampling limitations. By unifying theoretical guarantees with practical adaptability, GVPO establishes a new paradigm for reliable and versatile LLM post-training.



Paperid:3127
Authors:Nhat Hoang-Xuan, Xiyuan Wei, Wanli Xing, Tianbao Yang, My T. Thai
Abstract:
Contrastive Language-Image Pre-training (CLIP) generates versatile multimodal embeddings for diverse applications, yet the specific information captured within these representations is not fully understood. Current explainability techniques often target specific tasks, overlooking the rich, general semantics inherent in the representations. Our objective is to reveal the concepts encoded in CLIP embeddings by learning a surrogate representation, which is expressed as a linear combination of human-understandable concepts evident in the image. Our method, which we term EXPLAIN-R, introduces a novel approach that leverages CLIP's learned instance-instance similarity to train a surrogate model that faithfully mimics CLIP's behavior. From the trained surrogate, we derive concept scores for each input image; these scores quantify the contribution of each concept and act as the explanation for the representation. Quantitative evaluations on multiple datasets demonstrate our method's superior faithfulness over the baseline. Moreover, a user study confirms that our explanations are perceived as more relevant, complete, and useful. Our work provides a novel approach for interpreting CLIP image representations, enhancing the user interpretability of representations and fostering more trustworthy AI systems.



Paperid:3128
Authors:Bochen Lyu, Xiaojing Zhang, Fangyi Zheng, He Wang, Zheng Wang, Zhanxing Zhu
Abstract:
This paper establishes a continuous time approximation, a piece-wise continuous differential equation, for the discrete Heavy-Ball~(HB) momentum method with explicit discretization error. Investigating continuous differential equations has been a promising approach for studying the discrete optimization methods. Despite the crucial role of momentum in gradient-based optimization methods, the gap between the original dynamics and the continuous time approximations due to the discretization error has not been comprehensively bridged yet. In this work, we study the HB momentum method in continuous time while putting more focus on the discretization error to provide additional theoretical tools to this area. In particular, we design a first-order piece-wise continuous differential equation, where we add a number of counter terms to account for the discretization error explicitly. As a result, we provide a continuous time model for the HB momentum method that allows the control of discretization error to arbitrary order of the learning rate. As an application, we leverage it to find a new implicit regularization of the directional smoothness and investigate the implicit bias of HB for diagonal linear networks, indicating how our results can be used in deep learning. Our theoretical findings are further supported by numerical experiments.



Paperid:3129
Authors:Binrui Shen, LiangYuan, Shengxin Zhu
Abstract:
Graph matching, typically formulated as a Quadratic Assignment Problem (QAP), seeks to establish node correspondences between two graphs. To address the NP-hardness of QAP, some existing methods adopt projection-based relaxations that embed the problem into the convex hull of the discrete domain. However, these relaxations inevitably enlarge the feasible set, introducing two sources of error: numerical scale sensitivity and geometric misalignment between the relaxed and original domains. To alleviate these errors, we propose a novel relaxation framework by reformulating the projection step as a Frobenius-regularized Linear Assignment (FRA) problem, where a tunable regularization term mitigates feasible region inflation. This formulation enables normalization-based operations to preserve numerical scale invariance without compromising accuracy. To efficiently solve FRA, we propose the Scaling Doubly Stochastic Normalization (SDSN) algorithm. Building on its favorable computational properties, we develop a theoretically grounded mixed-precision architecture to achieve substantial acceleration.Comprehensive CPU-based benchmarks demonstrate that FRAM consistently outperforms all baseline methods under identical precision settings. When combined with a GPU-based mixed-precision architecture, FRAM achieves up to 370× speedup over its CPU-FP64 counterpart, with negligible loss in solution accuracy.



Paperid:3130
Authors:Li Bai, Qingqing Ye, Xinwei Zhang, Sen Zhang, Zi Liang, Jianliang Xu, Haibo Hu
Abstract:
Machine learning models are often vulnerable to inference attacks that expose sensitive information from their training data. Shadow model technique is commonly employed in such attacks, like membership inference.However, the need for a large number of shadow models leads to high computational costs, limiting their practical applicability.Such inefficiency mainly stems from the independent training and use of these shadow models.To address this issue, we present a novel shadow pool training framework SHAPOOL, which constructs multiple shared models and trains them jointly within a single process.In particular, we leverage the Mixture-of-Experts mechanism as the shadow pool to interconnect individual models, enabling them to share some sub-networks and thereby improving efficiency.To ensure the shared models closely resemble independent models and serve as effective substitutes, we introduce three novel modules: path-choice routing, pathway regularization, and pathway alignment.These modules guarantee random data allocation for pathway learning, promote diversity among shared models, and maintain consistency with target models.We evaluate SHAPOOL in the context of various membership inference attacks and show that it significantly reduces the computational cost of shadow model construction while maintaining comparable attack performance.



Paperid:3131
Authors:Truong Buu Phan, Ashish Khisti
Abstract:
We study channel simulation and distributed matching, two fundamental problems with several applications to machine learning, using a recently introduced generalization of the standard rejection sampling (RS) algorithm known as Ensemble Rejection Sampling (ERS). For channel simulation, we propose a new coding scheme based on ERS that achieves a near-optimal coding rate. In this process, we demonstrate that standard RS can also achieve a near-optimal coding rate and generalize the result of Braverman and Garg (2014) to the continuous alphabet setting. Next, as our main contribution, we present a distributed matching lemma for ERS, which serves as the rejection sampling counterpart to the Poisson Matching Lemma (PML) introduced by Li and Anantharam (2021). Our result also generalizes a recent work on importance matching lemma (Phan et al, 2024) and, to our knowledge, is the first result on distributed matching in the family of rejection sampling schemes where the matching probability is close to PML. We demonstrate the practical significance of our approach over prior works by applying it to distributed compression. The effectiveness of our proposed scheme is validated through experiments involving synthetic Gaussian sources and distributed image compression using the MNIST dataset.



Paperid:3132
Authors:Johann Flemming Gloy, Simon Olsson
Abstract:
Abstract:Flow and diffusion-based models have emerged as powerful tools for scientific applications, particularly for sampling non-normalized probability distributions, as exemplified by Boltzmann Generators (BGs). A critical challenge in deploying these models is their reliance on sample likelihood computations, which scale prohibitively with system size $n$, often rendering them infeasible for large-scale problems. To address this, we introduce, $\textit{HollowFlow}$ a flow-based generative model leveraging a novel non-backtracking graph neural network (NoBGNN). By enforcing a block-diagonal Jacobian structure, HollowFlow likelihoods are evaluated with a constant number of backward passes in $n$, yielding speed-ups of up to $\mathcal{O}(n^2)$: a significant step towards scaling BGs to larger systems. Crucially, our framework generalizes: $\textbf{any equivariant GNN or attention-based architecture}$ can be adapted into a NoBGNN. We validate HollowFlow, by training BGs on two different systems of increasing size. For both systems the sampling and likelihood evaluation time decreases dramatically, following our theoretical scaling laws. For the larger system we obtain a $10^2\times$ speed-up, clearly illustrating the potential of HollowFlow-based approaches for high-dimensional scientific problems previously hindered by computational bottlenecks.



Paperid:3133
Authors:Jeongho Park, Daheen Kim, Cheoljun Kim, Hyungbin Park, Sangwook Kang, Gwangsu Kim
Abstract:
Abstract:We propose a new calibration method for survival models based on the Kolmogorov–Smirnov (KS) metric. Existing approaches—including conformal prediction, D-calibration, and Kaplan–Meier (KM)-based methods—often rely on heuristic binning or nonparametric estimators, which undermine their adaptability to continuous-time settings and complex model outputs. To address these limitations, we introduce a streamlined $\textit{KS metric-based post-processing}$ framework (KSP) that calibrates survival predictions without relying on discretization or KM estimation. This design enhances flexibility and broad applicability. We conduct extensive experiments on diverse real-world datasets using a variety of survival models. Empirical results demonstrate that our method consistently improves calibration performance over existing methods while maintaining high predictive accuracy. We also provide a theoretical analysis of the KS metric and discuss extensions to in-processing settings.



Paperid:3134
Authors:Hanzhen Zhao, Ding Shihong, Cong Fang, Zhouchen Lin
Abstract:
Abstract:This paper introduces PaZO, a preconditioned accelerated zeroth-order optimization algorithm for fine-tuning large language models (LLMs). First, we theoretically demonstrate the necessity of preconditioning in zeroth-order optimization, proving that zeroth-order stochastic gradient descent (ZO-SGD) alone fails to achieve the ideal convergence rate. Building on this, we propose a Preconditioned Simultaneous Perturbation Stochastic Approximation (PSPSA) and theoretical version of PaZO, and demonstrate that setting the order of preconditioner as $-1/2$ in PSPSA yields the improved convergence rate for PaZO. Moreover, we design a practical version of PaZO that stabilizes training via diagonal Hessian estimate and moving average technique. Extensive experiments on diverse downstream tasks with models like RoBERTa-large and OPT show PaZO’s effectiveness. Compared to other zeroth-order baselines, PaZO achieves better performance across models and tasks.



Paperid:3135
Authors:Xuanming Zhang, Yuxuan Chen, Min-Hsuan Yeh, Sharon Li
Abstract:
Human social interactions depend on the ability to infer others' unspoken intentions, emotions, and beliefs—a cognitive skill grounded in the psychological concept of Theory of Mind (ToM). While large language models (LLMs) excel in semantic understanding tasks, they struggle with the ambiguity and contextual nuance inherent in human communication. To bridge this gap, we introduce MetaMind, a multi-agent framework inspired by psychological theories of metacognition, designed to emulate human-like social reasoning. MetaMind decomposes social understanding into three collaborative stages: (1) a Theory-of-Mind Agent generates hypotheses user mental states (e.g., intent, emotion), (2) a Domain Agent refines these hypotheses using cultural norms and ethical constraints, and (3) a Response Agent generates contextually appropriate responses while validating alignment with inferred intent. Our framework achieves state-of-the-art performance across three challenging benchmarks, with 35.7% improvement in real-world social scenarios and 6.2% gain in ToM reasoning. Notably, it enables LLMs to match human-level performance on key ToM tasks for the first time. Ablation studies confirm the necessity of all components, which showcase the framework’s ability to balance contextual plausibility, social appropriateness, and user adaptation. This work advances AI systems toward human-like social intelligence, with applications in empathetic dialogue and culturally sensitive interactions.



Paperid:3136
Authors:Yiyou Sun, Yu Gai, Lijie Chen, Abhilasha Ravichander, Yejin Choi, Nouha Dziri, Dawn Song
Abstract:
Large language models (LLMs) frequently generate hallucinations—content that deviates from factually inaccurate or deviates from provided context—posing challenges for diagnosis. However, diagnosing the causes of hallucination is challenging due to the complex interplay of underlying causes. This paper introduces a framework to systematically understand the sources of hallucination behavior in large language models. Our key insight is that hallucinations arise when more frequent but non-factual associations outweigh faithful ones.Through theoretical and empirical analyses, we demonstrate that decoder-only transformers effectively function as subsequence embedding models, with the fully-connected layers encoding input-output associations. We propose a tracing algorithm that identifies causal subsequences by analyzing hallucination probabilities across randomized input contexts. Experiments show our method outperforms standard attribution techniques in identifying hallucination causes and is supported by evidence from the model’s training corpus. This work provides a unified perspective on hallucinations and a robust framework for their cause and analysis.



Paperid:3137
Authors:Zeting Chen, Xinyu Cai, Molei Qin, Bo An
Abstract:
Options markets represent one of the most sophisticated segments of the financial ecosystem, with prices that directly reflect market uncertainty. In this paper, we introduce the first reinforcement learning (RL) framework specifically designed for volatility trading through options, focusing on profit from the difference between implied volatility (IV) and realized volatility (RV). Our multi-agent architecture consists of an Option Position Agent (OP-Agent) responsible for volatility timing by controlling long/short volatility positions, and a Hedger Routing Agent (HR-Agent) that manages risk and maximizes path-dependent profits by selecting optimal hedging strategies with different risk preferences. Evaluating our approach using cryptocurrency options data from 2021-2024, we demonstrate superior performance on BTC and ETH, significantly outperforming traditional strategies and machine learning baselines across all profit and risk-adjusted metrics while exhibiting sophisticated trading behavior.



Paperid:3138
Authors:Maurice Kraus, Felix Divo, Devendra Singh Dhami, Kristian Kersting
Abstract:
Time series data is prevalent across numerous fields, necessitating the development of robust and accurate forecasting models. Capturing patterns both within and between temporal and multivariate components is crucial for reliable predictions. We introduce xLSTM-Mixer, a model designed to effectively integrate temporal sequences, joint time-variate information, and multiple perspectives for robust forecasting. Our approach begins with a linear forecast shared across variates, which is then refined by xLSTM blocks. They serve as key elements for modeling the complex dynamics of challenging time series data. xLSTM-Mixer ultimately reconciles two distinct views to produce the final forecast. Our extensive evaluations demonstrate its superior long-term forecasting performance compared to recent state-of-the-art methods while requiring very little memory. A thorough model analysis provides further insights into its key components and confirms its robustness and effectiveness. This work contributes to the resurgence of recurrent models in forecasting by combining them, for the first time, with mixing architectures.



Paperid:3139
Authors:Huanming Shen, Baizhou Huang, Xiaojun Wan
Abstract:
Watermarking is a promising defense against the misuse of large language models (LLMs), yet it remains vulnerable to scrubbing and spoofing attacks.This vulnerability stems from an inherent trade-off governed by watermark window size: smaller windows resist scrubbing better but are easier to reverse-engineer, enabling low-cost statistics-based spoofing attacks. This work breaks this trade-off by introducing a novel mechanism, equivalent texture keys, where multiple tokens within a watermark window can independently support the detection.Based on the redundancy, we proposeSEEK(Sub-vocabulary decomposedEquivalent tExtureKey). It achieves a Pareto improvement, increasing the resilience against scrubbing attacks without compromising robustness to spoofing. Experiments demonstrate SEEK's superiority over prior method, yielding spoofing robustness gains of +88.2\%/+92.3\%/+82.0\% and scrubbing robustness gains of +10.2\%/+6.4\%/+24.6\% across diverse dataset settings.Our anonymous code submission accompanies this paper and will be released after anonymous reviews.



Paperid:3140
Authors:Yixiu Mao, Yun Qu, Qi Wang, Xiangyang Ji
Abstract:
Offline reinforcement learning (RL) suffers from extrapolation errors induced by out-of-distribution (OOD) actions. To address this, offline RL algorithms typically impose constraints on action selection, which can be systematically categorized into density, support, and sample constraints. However, we show that each category has inherent limitations: density and sample constraints tend to be overly conservative in many scenarios, while the support constraint, though least restrictive, faces challenges in accurately modeling the behavior policy. To overcome these limitations, we propose a new neighborhood constraint that restricts action selection in the Bellman target to the union of neighborhoods of dataset actions. Theoretically, the constraint not only bounds extrapolation errors and distribution shift under certain conditions, but also approximates the support constraint without requiring behavior policy modeling. Moreover, it retains substantial flexibility and enables pointwise conservatism by adapting the neighborhood radius for each data point. In practice, we employ data quality as the adaptation criterion and design an adaptive neighborhood constraint. Building on an efficient bilevel optimization framework, we develop a simple yet effective algorithm, Adaptive Neighborhood-constrained Q learning (ANQ), to perform Q learning with target actions satisfying this constraint. Empirically, ANQ achieves state-of-the-art performance on standard offline RL benchmarks and exhibits strong robustness in scenarios with noisy or limited data.



Paperid:3141
Authors:Zhebei Shen, Qifan Yu, Juncheng Li, Wei Ji, Qizhi Chen, Siliang Tang, Yueting Zhuang
Abstract:
Recent advances in reinforcement learning (RL) methods such as Grouped Relative Policy Optimization (GRPO) have strengthened the reasoning capabilities of Large Vision-Language Models (LVLMs). However, due to the inherent entanglement between visual and textual modalities, applying GRPO to LVLMs often leads to reward convergence across different responses to the same sample as training progresses, hindering effective gradient updates and causing the enhancement of chain-of-thought reasoning to stagnate or even collapse.To address this issue, we propose a progressive instruction evolution framework, EvolvedGRPO, to gradually generate more complex questions via editing instructions in an adversarial way, progressively aligned with the model’s evolving capabilities. Specifically, we design two instruction editing strategies across modalities, incorporating incrementally increasing editing instructions and RL-based adversarial data augmentation to improve the effectiveness of model training. To address GRPO's limitations on overly difficult problems, we first train on basic subproblem versions of complex multi-modal questions in both the visual and textual modalities, progressively increasing difficulty to enable prefix-style process rewards, effectively combining the strengths of both process rewards and group-wise relative rewards. Finally, EvolvedGRPO achieves state-of-the-art performance among open-source RL models on multi-modal reasoning tasks, even approaching the closed-source GPT-4o in reasoning capabilities, and demonstrates better performance on unseen LVLM general benchmarks.



Paperid:3142
Authors:Conghui Li, Chern Hong Lim, Xin Wang
Abstract:
While deep neural networks possess the capability to perform semantic segmentation, producing a single deterministic output limits reliability in safety-critical applications, caused by uncertainty and annotation variability. To address this, stochastic segmentation models using Conditional Variational Autoencoders (CVAE), Bayesian networks, and diffusion have been explored. However, existing approaches suffer from limited latent expressiveness and interpretability. Furthermore, our experiments showed that models like Probabilistic U-Net rely excessively on high latent variance, leading to posterior collapse. This work propose a novel framework by integrating Gaussian Mixture Model (GMM) with Normalizing Flow (NF) in CVAE for stochastic segmentation. GMM structures the latent space into meaningful semantic clusters, while NF captures feature deformations with quantified uncertainty. Our method stabilizes latent distributions through constrained variance and mean ranges. Experiments on LIDC, Crack500, and Cityscapes datasets show that our approach outperformed state-of-the-art in curvilinear structure and medical image segmentation.



Paperid:3143
Authors:Jizhou Han, Shaokun Wang, Yuhang He, Chenhao Ding, Qiang Wang, Xinyuan Gao, SongLin Dong, Yihong Gong
Abstract:
Generalized Category Discovery (GCD) focuses on classifying known categories while simultaneously discovering novel categories from unlabeled data. However, previous GCD methods face challenges due to inconsistent optimization objectives and category confusion. This leads to feature overlap and ultimately hinders performance on novel categories. To address these issues, we propose the Neural Collapse-inspired Generalized Category Discovery (NC-GCD) framework. By pre-assigning and fixing Equiangular Tight Frame (ETF) prototypes, our method ensures an optimal geometric structure and a consistent optimization objective for both known and novel categories. We introduce a Consistent ETF Alignment Loss that unifies supervised and unsupervised ETF alignment and enhances category separability. Additionally, a Semantic Consistency Matcher (SCM) is designed to maintain stable and consistent label assignments across clustering iterations. Our method significantly enhancing novel category accuracy and demonstrating its effectiveness.



Paperid:3144
Authors:Zhihao Wu, Jinyu Cai, Yunhe Zhang, Jielong Lu, Zhaoliang Chen, Shuman Zhuang, Haishuai Wang
Abstract:
The convergence of multi-view learning and graph learning has propelled the emergence of multi-view graph neural networks (MvGNNs), offering unprecedented capabilities to address complex real-world data characterized by heterogeneous yet interconnected information. While existing MvGNNs exploit the potential of multi-view graphs, they often fail to harmonize the dual inductive biases critical to multi-view learning: consistency (inherent inter-view agreement) and complementarity (view-specific distinctiveness). To bridge this gap, we propose Multi-view Collaborative Graph Mixture of Experts (MvCGE), a novel framework grounded in the Mixture-of-Experts (MoE) paradigm. MvCGE establishes architectural consistency through shared parameters while preserving complementarity via layer-wise collaborative graph experts, which are dynamically activated by a graph-aware routing mechanism that adapts to the structural nuances of each view. This dual-level design is further reinforced by two novel components: a load equilibrium loss to prevent expert collapse and ensure balanced specialization, and a graph discrepancy loss based on distributional divergence to enhance inter-view complementarity. Extensive experiments on diverse datasets demonstrate MvCGE’s superiority.



Paperid:3145
Authors:Lei Shi, David Arbour, Raghavendra Addanki, Ritwik Sinha, Avi Feller
Abstract:
The growing popularity of Large Language Models (LLMs) introduces exciting opportunities for digital experimentation, as marketers increasingly combine human-generated and model-generated content to design treatments. However, the high volume and dimensionality of these treatments pose significant challenges, particularly for integration into traditional evaluation paradigms. In this paper, we propose to explicitly incorporate semantic similarity from treatment embeddings to improve statistical efficiency in the design and analysis of LLM-generated treatments at scale. Our proposed method, called double kernel representation learning, factorizes the causal effect through the inner product of the low-dimensional kernel-based representations of the treatments and user covariates. We design an alternating-minimization algorithm to efficiently implement this to learn the representation from data. Additionally, we propose an adaptive design strategy for online experimentation. We prove convergence under a low-rank factor structure and establish a sublinear regret bound for the adaptive algorithm. Finally, we demonstrate the practical performance of our approaches through experiments on semi-synthetic data.



Paperid:3146
Authors:Yijin Ren, Haifeng Xu, Qi Deng
Abstract:
This paper introduces new parameter-free first-order methods for convex optimization problems in which the objective function exhibits Hölder smoothness. Inspired by the recently proposed distance-over-gradient (DOG) technique, we propose an accelerated distance-adaptive method which achieves optimal anytime convergence rates for Hölder smooth problems without requiring prior knowledge of smoothness parameters or explicit parameter tuning. Importantly, our parameter-free approach removes the necessity of specifying target accuracy in advance, addressing a significant limitation found in the universal fast gradient methods(Nesterov,2015).We further present a parameter-free accelerated method that eliminates the need for line-search procedures and extend it to convex stochastic optimization. Preliminary experimental results highlight the effectiveness of our approach in convex nonsmooth problems and its advantages over existing parameter-free or accelerated methods.



Paperid:3147
Authors:Xiao Fei, Michail Chatzianastasis, Sarah Carneiro, Hadi Abdine, Lawrence Petalidis, Michalis Vazirgiannis
Abstract:
Predicting protein function from sequence is a central challenge in computational biology. While existing methods rely heavily on structured ontologies or similarity-based techniques, they often lack the flexibility to express structure-free functional descriptions and novel biological functions. In this work, we introduce Prot2Text-V2, a novel multimodal sequence-to-text model that generates free-form natural language descriptions of protein function directly from amino acid sequences. Our method combines a protein language model as a sequence encoder (ESM-3B) and a decoder-only language model (LLaMA-3.1-8B-Instruct) through a lightweight nonlinear modality projector.A key innovation is our Hybrid Sequence-level Contrastive Alignment Learning (H-SCALE), which improves cross-modal learning by matching mean- and std-pooled protein embeddings with text representations via contrastive loss. After the alignment phase, we apply instruction-based fine-tuning using LoRA on the decoder to teach the model how to generate accurate protein function descriptions conditioned on the protein sequence. We train Prot2Text-V2 on about 250K curated entries from SwissProt and evaluate it under low-homology conditions, where test sequences have low similarity with training samples. Prot2Text-V2 consistently outperforms traditional and LLM-based baselines across various metrics.



Paperid:3148
Authors:Ryo Umagami, Liu Yue, Xuangeng Chu, Ryuto Fukushima, Tetsuya Narita, Yusuke Mukuta, Tomoyuki Takahata, Jianfei Yang, Tatsuya Harada
Abstract:
Humans move with intention. However, existing human motion datasets and prediction models typically focus on short-term kinematics and lack explicit representations of high-level goals or social context. In this work, we introduce \textit{Intend to Move (I2M)}, a new dataset designed for intention-aware long-term human motion modeling in dynamic real-world environments. I2M contains 10.1 hours of SMPL-based 3D motion across 215 sequences, each synchronized with multi-view RGB-D video, 3D scene point clouds, and natural language descriptions of human intentions. We conduct a comprehensive analysis across eight conditioning modalities, including trajectories, semantic features, and multimodal combinations. Our results show that structured motion cues (e.g., joint trajectories) yield superior prediction accuracy, while high-level semantics such as intention text remain underutilized by current pretrained encoders. This highlights a significant gap between semantic intention representations and their grounding in physical motion. We present I2M as a benchmark to promote future research in bridging this gap through intention-grounded learning for embodied intelligence.



Paperid:3149
Authors:Sangwoo Park, Matteo Zecchin, Osvaldo Simeone
Abstract:
Abstract:Selecting artificial intelligence (AI) models, such as large language models (LLMs), from multiple candidates requires accurate performance estimation. This is ideally achieved through empirical evaluations involving abundant real-world data. However, such evaluations are costly and impractical at scale. To address this challenge, autoevaluation methods leverage synthetic data produced by automated evaluators, such as LLMs-as-judges, reducing variance but potentially introducing bias. Recent approaches have employed semi-supervised prediction-powered inference ($\texttt{PPI}$) to correct for the bias of autoevaluators. However, the use of autoevaluators may lead in practice to a degradation in sample efficiency compared to conventional methods using only real-world data. In this paper, we propose $\texttt{R-AutoEval+}$, a novel framework that provides finite-sample reliability guarantees on the model evaluation, while also ensuring an enhanced (or at least no worse) sample efficiency compared to conventional methods. The key innovation of $\texttt{R-AutoEval+}$ is an adaptive construction of the model evaluation variable, which dynamically tunes its reliance on synthetic data, reverting to conventional methods when the autoevaluator is insufficiently accurate. Experiments on the use of LLMs-as-judges for the optimization of quantization settings for the weights of an LLM, and for prompt design in LLMs confirm the reliability and efficiency of $\texttt{R-AutoEval+}$.



Paperid:3150
Authors:Jeongin Kim, Wonho Bae, YouLee Han, Giyeong Oh, Youngjae Yu, Danica J. Sutherland, Junhyug Noh
Abstract:
Semantic segmentation demands dense pixel-level annotations, which can be prohibitively expensive -- especially under extremely constrained labeling budgets. In this paper, we address the problem of low-budget active learning for semantic segmentation by proposing a novel two-stage selection pipeline. Our approach leverages a pre-trained diffusion model to extract rich multi-scale features that capture both global structure and fine details. In the first stage, we perform a hierarchical, representation-based candidate selection by first choosing a small subset of representative pixels per image using MaxHerding, and then refining these into a diverse global pool. In the second stage, we compute an entropy‐augmented disagreement score (eDALD) over noisy multi‐scale diffusion features to capture both epistemic uncertainty and prediction confidence, selecting the most informative pixels for annotation. This decoupling of diversity and uncertainty lets us achieve high segmentation accuracy with only a tiny fraction of labeled pixels. Extensive experiments on four benchmarks (ADE-Bed, CamVid, Cityscapes, and Pascal-Context) demonstrate that our method significantly outperforms existing baselines under extreme pixel‐budget regimes.



Paperid:3151
Authors:Liwei Jiang, Chai Yuanjun, Margaret Li, Mickel Liu, Raymond Fok, Maarten Sap, Yulia Tsvetkov, Nouha Dziri, Yejin Choi
Abstract:
Language models (LMs) often struggle to generate diverse, human-like creative content, raising concerns about the long-term homogenization of human thought through repeated exposure to similar outputs. Yet, scalable methods for evaluating LM output diversity remain limited—especially beyond narrow tasks like random number generation or stylized prompts. To address this gap, we introduce InfiniteChats, a large-scale dataset of 26,000 diverse, real-world open-ended user queries, along with the first comprehensive taxonomy for characterizing the full spectrum of open-ended prompts posed to LMs, comprising six top-level categories and 17 subcategories. These queries admit a wide range of plausible answers with no single ground truth. Using InfiniteChats, we present a large-scale analysis of mode collapse in LMs, manifested as redundant outputs even for inherently open-ended queries. Our study reveals a pronounced "Artificial Hivemind" effect in open-ended generation, characterized by (1) intra-model repetition, where a single model consistently generates similar responses, and (2) inter-model homogeneity, where different models produce strikingly similar outputs.InfiniteChats also includes 31,250 human annotations, across absolute ratings and pairwise preferences, with 25 independent human annotations per example. This enables fine-grained analysis of distributional preferences across annotators. Our findings show that state-of-the-art LMs, reward models, and LM judges align less with human ratings when annotators disagree or when responses are of similar quality. Overall, InfiniteChats offers the first large-scale resource for systematically studying open-endedness in LM queries, revealing critical insights to guide future research and mitigate long-term AI safety risks posed by the Artificial Hivemind.



Paperid:3152
Authors:Yuru Jiang, Wenxuan Ding, Shangbin Feng, Greg Durrett, Yulia Tsvetkov
Abstract:
We propose Sparta Alignment, an algorithm to collectively align multiple LLMs through competition and combat. To complement a single model's lack of diversity in generation and biases in evaluation, multiple LLMs form a 'sparta tribe' to compete against each other in fulfilling instructions while serving as judges for the competition of others. For each iteration, one instruction and two models are selected for a duel, the other models evaluate the two responses, and their evaluation scores are aggregated through a adapted elo-ranking based reputation system, where winners/losers of combat gain/lose weight in evaluating others. The peer-evaluated combat results then become preference pairs where the winning response is preferred over the losing one, and all models learn from these preferences at the end of each iteration. Sparta Alignment enables the self-evolution of multiple LLMs in an iterative and collective competition process. Extensive experiments demonstrate that Sparta Alignment outperforms initial models and 4 self-alignment baselines across 10 out of 12 tasks and datasets with 7.0\% average improvement. Further analysis reveals that Sparta Alignment generalizes more effectively to unseen tasks and leverages the expertise diversity of participating models to produce more logical, direct and informative outputs.



Paperid:3153
Authors:Paul Schwerdtner, Jules Berman, Benjamin Peherstorfer
Abstract:
Abstract:Deep neural networks using state space models as layers are well suited for long-range sequence tasks but can be challenging to compress after training.We introduce Hankel singular value regularization (HSVR) that encourages decay in the Hankel singular values so that the state space models become compressible. To make the proposed regularization scalable, we develop an algorithm to efficiently compute the Hankel singular values during training iterations by exploiting the specific block-diagonal structure of the system matrices that is typically used in state space models. Experiments on Long Range Arena benchmarks demonstrate that HSVR state space layers are up to 10$\times$ more compressible than standard state space layers trained without regularization while maintaining high accuracy.



Paperid:3154
Authors:Yuancheng Wang, Jiachen Zheng, Junan Zhang, Xueyao Zhang, Huan Liao, Zhizheng Wu
Abstract:
We introduceMetis, a foundation model for unified speech generation.Unlike previous task-specific or multi-task models, Metis follows a pre-training and fine-tuning paradigm. It is pre-trained on large-scale unlabeled speech data using masked generative modeling and then fine-tuned to adapt to diverse speech generation tasks.Specifically, (1) Metis utilizes two discrete speech representations: SSL tokens derived from speech self-supervised learning (SSL) features, and acoustic tokens directly quantized from waveforms. (2) Metis performs masked generative pre-training on SSL tokens, utilizing 300K hours of diverse speech data, without any additional condition. (3) Through fine-tuning with task-specific conditions, Metis achieves efficient adaptation to various speech generation tasks while supporting multimodal input, even when using limited data and trainable parameters.Experiments demonstrate that Metis can serve as a foundation model for unified speech generation: Metis outperforms state-of-the-art task-specific or multi-task systems across five speech generation tasks, including zero-shot text-to-speech, voice conversion, target speaker extraction, speech enhancement, and lip-to-speech, even with fewer than 20M trainable parameters or 300 times less training data. Audio samples are are available at https://metis-demo.github.io/.



Paperid:3155
Authors:Qitai Tan, Yiyun Chen, Mo Li, Ruiwen Gu, Yilin Su, Xiao-Ping (Steven) Zhang
Abstract:
Recent advances in deep learning have driven rapid progress in time series forecasting, yet many state-of-the-art models continue to struggle with robust performance in real-world applications, even when they achieve strong results on standard benchmark datasets. This persistent gap can be attributed to the black-box nature of deep learning architectures and the inherent limitations of current evaluation frameworks, which frequently lack the capacity to provide clear, quantitative insights into the specific strengths and weaknesses of different models, thereby complicating the selection of appropriate models for particular forecasting scenarios.To address these issues, we propose a synthetic data-driven evaluation paradigm, SynTSBench, that systematically assesses fundamental modeling capabilities of time series forecasting models through programmable feature configuration. Our framework isolates confounding factors and establishes an interpretable evaluation system with three core analytical dimensions: (1) temporal feature decomposition and capability mapping, which enables systematic evaluation of model capacities to learn specific pattern types; (2) robustness analysis under data irregularities, which quantifies noise tolerance thresholds and anomaly recovery capabilities; and (3) theoretical optimum benchmarking, which establishes performance boundaries for each pattern type—enabling direct comparison between model predictions and mathematical optima.Our experiments show that current deep learning models do not universally approach optimal baselines across all types of temporal features.



Paperid:3156
Authors:Yuyao Wang, Yu-Hung Cheng, Debarghya Mukherjee, Huimin Cheng
Abstract:
Graphon models provide a flexible nonparametric framework for estimating latent connectivity probabilities in networks, enabling a range of downstream applications such as link prediction and data augmentation. However, accurate graphon estimation typically requires a large graph, whereas in practice, one often only observes a small-sized network. One approach to addressing this issue is to adopt a transfer learning framework, which aims to improve estimation in a small target graph by leveraging structural information from a larger, related source graph. In this paper, we propose a novel method, namely GTRANS, a transfer learning framework that integrates neighborhood smoothing and Gromov-Wasserstein optimal transport to align and transfer structural patterns between graphs. To prevent negative transfer, GTRANS includes an adaptive debiasing mechanism that identifies and corrects for target-specific deviations via residual smoothing. We provide theoretical guarantees on the stability of the estimated alignment matrix and demonstrate the effectiveness of GTRANS in improving the accuracy of target graph estimation through extensive synthetic and real data experiments. These improvements translate directly to enhanced performance in downstream applications, such as the graph classification task and the link prediction task.



Paperid:3157
Authors:Deheng Yuan, Tao Guo, Zhongyi Huang
Abstract:
Abstract:Consider the communication-constrained estimation of discrete distributions under $\ell^p$ losses, where each distributed terminal holds multiple independent samples and uses limited number of bits to describe the samples. We obtain the minimax optimal rates of the problem for most parameter regimes. As a result, an elbow effect of the optimal rates at $p=2$ is clearly identified. In order to achieve the optimal rates for different parameter regimes, we introduce refinement methods and develop additional customized techniques in the estimation protocols. The general idea of the refinement methods is to first generate rough estimate by partial information and then establish refined estimate in subsequent steps guided by the rough estimate. Then customized techniques such as successive refinement, sample compression, thresholding and random hashing are leveraged to achieve the optimal rates in different parameter regimes. The optimality of the estimation protocols is shown by deriving compatible minimax lower bounds.



Paperid:3158
Authors:Vinam Arora, Divyansha Lachi, Ian Knight, Mehdi Azabou, Blake Richards, Cole Hurwitz, Joshua H Siegle, Eva Dyer
Abstract:
Identifying the functional identity of individual neurons is essential for interpreting circuit dynamics, yet remains a major challenge in large-scale in vivo recordings where anatomical and molecular labels are often unavailable. Here we introduce NuCLR, a self-supervised framework that learns context-aware representations of neuron identity by modeling each neuron's role within the broader population. NuCLR employs a spatiotemporal transformer that captures both within-neuron dynamics and across-neuron interactions, and is trained with a sample-wise contrastive objective that encourages stable, discriminative embeddings across time. Across multiple open-access datasets, NuCLR outperforms prior methods in both cell type and brain region classification. It enables zero-shot generalization to entirely new populations—without retraining or access to stimulus labels—offering a scalable approach for real-time, functional decoding of neuron identity across diverse experimental settings.



Paperid:3159
Authors:Marianne Rakic, Siyu Gai, Etienne Chollet, John Guttag, Adrian Dalca
Abstract:
A single biomedical image can be segmented in multiple valid ways, depending on the application. For instance, a brain MRI may be divided according to tissue types, vascular territories, broad anatomical regions, fine-grained anatomy, or pathology. Existing automatic segmentation models typically either (1) support only a single protocol---the one they were trained on---or (2) require labor-intensive prompting to specify the desired segmentation. We introducePancakes, a framework that, given a new image from a previously unseen domain, automatically generates multi-label segmentation maps formultipleplausible protocols, while maintaining semantic consistency across related images. In extensive experiments across seven previously unseen domains,Pancakesconsistently outperforms strong baselines, often by a wide margin, demonstrating its ability to produce diverse yet coherent segmentation maps on unseen domains.



Paperid:3160
Authors:Xiaomin Li, Zhou Yu, Zhiwei Zhang, Xupeng Chen, Ziji Zhang, Yingying Zhuang, Narayanan Sadagopan, Anurag Beniwal
Abstract:
Reasoning-enhanced large language models (RLLMs), whether explicitly trained for reasoning or prompted via chain-of-thought (CoT), have achieved state-of-the-art performance on many complex reasoning tasks. However, we uncover a surprising and previously overlooked phenomenon: explicit CoT reasoning can significantly degrade instruction-following accuracy. Evaluating 15 models on two benchmarks: IFEval (with simple, rule-verifiable constraints) and ComplexBench (with complex, compositional constraints), we consistently observe performance drops when CoT prompting is applied. Through large-scale case studies and an attention-based analysis, we identify common patterns where reasoning either helps (e.g., with formatting or lexical precision) or hurts (e.g., by neglecting simple constraints or introducing unnecessary content). We propose a metric, constraint attention, to quantify model focus during generation and show that CoT reasoning often diverts attention away from instruction-relevant tokens. To mitigate these effects, we introduce and evaluate four strategies: in-context learning, self-reflection, self-selective reasoning, and classifier-selective reasoning. Our results demonstrate that selective reasoning strategies, particularly classifier-selective reasoning, can substantially recover lost performance. To our knowledge, this is the first work to systematically expose reasoning-induced failures in instruction-following and offer practical mitigation strategies.



Paperid:3161
Authors:Parsa Mirtaheri, Ezra Edelman, Samy Jelassi, Eran Malach, Enric Boix-Adserà
Abstract:
Inference-time computation has emerged as a promising scaling axis for improving large language model reasoning. However, despite yielding impressive performance, the optimal allocation of inference-time computation remains poorly understood. A central question is whether to prioritize sequential scaling (e.g., longer chains of thought) or parallel scaling (e.g., majority voting across multiple short chains of thought). In this work, we seek to illuminate the landscape of test-time scaling by demonstrating the existence of reasoning settings where sequential scaling offers an exponential advantage over parallel scaling. These settings are based on graph connectivity problems in challenging distributions of graphs. We validate our theoretical findings with comprehensive experiments across a range of language models, including models trained from scratch for graph connectivity with different chain of thought strategies as well as large reasoning models.



Paperid:3162
Authors:Alex Fang, Hadi Pouransari, Matt Jordan, Alexander Toshev, Vaishaal Shankar, Ludwig Schmidt, Tom Gunter
Abstract:
Data filtering has become a powerful tool for improving model performance while reducing computational cost. However, as large language model compute budgets continue to grow, the limited data volume provided by heavily filtered and deduplicated datasets will become a practical constraint. In efforts to better understand how to proceed, we study model performance at various compute budgets and across multiple pre-training datasets created through data filtering and deduplication. We find that, given appropriate modifications to the training recipe, repeating existing aggressively filtered datasets for up to ten epochs can outperform training on the ten times larger superset for a single epoch across multiple compute budget orders of magnitude. While this finding relies on repeating the dataset for many epochs, we also investigate repeats within these datasets at the document level. We find that not all documents within a dataset are equal, and we can create better datasets relative to a token budget by explicitly manipulating the counts of individual documents. We conclude by arguing that even as large language models scale, data filtering remains an important direction of research.



Paperid:3163
Authors:Yiming Huang, Yajie Hao, Yuxuan Du, Jing Zhou, Xiao Yuan, Xiaoting Wang
Abstract:
Variational Quantum Algorithms (VQAs) are emerging as leading strategies with the potential to unlock practical applications and deliver significant advantages in the investigation of many-body quantum systems and quantum chemistry. A key challenge hindering the application of VQAs to large-scale problems is rooted in the no-cloning theorem in quantum mechanics, precluding standard backpropagation and leading to prohibitive quantum resource expenditure such as measurement cost.To address this challenge, we reformulate the training dynamics of VQAs as a non-linear partial differential equation and propose a novel protocol that leverages physics-informed neural networks (PINNs) to model this dynamical system efficiently. Given a small amount of training trajectory data collected from quantum devices, our protocol predicts the parameter updates of VQAs over multiple iterations on the classical side, dramatically reducing quantum resource costs. Through systematic numerical experiments, we demonstrate that our method achieves up to a 30x speedup compared to conventional methods and reduces quantum resource costs by as much as 90\% for tasks involving up to 40 qubits, including ground state preparation of different quantum systems, while maintaining competitive accuracy. Our approach complements existing techniques aimed at improving the efficiency of VQAs and further strengthens their potential for practical applications.



Paperid:3164
Authors:Chu Xu, Xinke Jiang, Rihong Qiu, Jiaran Gao, Junfeng Zhao
Abstract:
Measuring parameter importance is crucial for understanding and optimizing large language models (LLMs). Existing work predominantly focuses on pruning or probing at neuron/feature levels without fully considering the cooperative behaviors of model parameters. In this paper, we introduce a novel approach-Model Shapley to quantify parameter importance based on the Shapley value, a principled method from cooperative game theory that captures both individual and synergistic contributions among parameters. We derive a scalable second-order approximation to compute Shapley values at the parameter level, leveraging blockwise Fisher information for tractability in large-scale settings. Our method enables fine-grained differentiation of parameter importance, facilitating targeted knowledge injection and model compression. Through mini-batch Monte Carlo updates and efficient approximation of the Hessian structure, we achieve robust Shapley-based attribution with only modest computational overhead. Experimental results indicate that this cooperative game perspective enhances interpretability, guides more effective parameter-specific fine-tuning, and paves the way for continuous model improvement in various downstream tasks.



Paperid:3165
Authors:Minzhi Lin, Tianchi Xie, Mengchen Liu, Yilin Ye, Changjian Chen, Shixia Liu
Abstract:
Understanding infographic charts with design-driven visual elements (e.g., pictograms, icons) requires both visual recognition and reasoning, posing challenges for multimodal large language models (MLLMs). However, existing visual question answering benchmarks fall short in evaluating these capabilities of MLLMs due to the lack of paired plain charts and visual-element-based questions. To bridge this gap, we introduce InfoChartQA, a benchmark for evaluating MLLMs on infographic chart understanding. It includes 5,642 pairs of infographic and plain charts, each sharing the same underlying data but differing in visual presentations. We further design visual-element-based questions to capture their unique visual designs and communicative intent. Evaluation of 20 MLLMs reveals a substantial performance decline on infographic charts, particularly for visual-element-based questions related to metaphors. The paired infographic and plain charts enable fine-grained error analysis and ablation studies, which highlight new opportunities for advancing MLLMs in infographic chart understanding. We release InfoChartQA at https://github.com/CoolDawnAnt/InfoChartQA.



Paperid:3166
Authors:Ting Wei, Biao Mei, Junliang Lyu, Renquan Zhang, Feng Zhou, Yifan Sun
Abstract:
Personalized Bayesian federated learning (PBFL) handles non-i.i.d. client data and quantifies uncertainty by combining personalization with Bayesian inference. However, current PBFL methods face two main limitations: posterior inference on clients often assumes restrictive parametric forms, and server-side posterior aggregation typically relies on naive parameter averaging. To overcome these issues, we propose FedWBA, a novel PBFL method that enhances both local inference and global aggregation. At the client level, we use particle-based variational inference for nonparametric posterior representation. At the server level, we introduce particle-based Wasserstein barycenter aggregation, offering a more geometrically meaningful approach. Theoretically, we provide local and global convergence guarantees for FedWBA. Locally, we prove a KL divergence decrease lower bound per iteration for variational inference convergence. Globally, we show that the Wasserstein barycenter converges to the true parameter as the client data size increases. Empirically, experiments show that FedWBA outperforms baselines in prediction accuracy, uncertainty calibration, and convergence rate, with ablation studies confirming its robustness.



Paperid:3167
Authors:Sishun Liu, KE DENG, Xiuzhen Zhang, Yongli Ren, Yan Wang
Abstract:
Marked Temporal Point Process (MTPP) has been well studied to model the event distribution in marked event streams, which can be used to predict the mark and arrival time of the next event. However, existing studies overlook that the distribution of event marks is highly imbalanced in many real-world applications, with some marks being frequent but others rare. The imbalance poses a significant challenge to the performance of the next event prediction, especially for events of rare marks. To address this issue, we propose a thresholding method, which learns thresholds to tune the mark probability normalized by the mark's prior probability to optimize mark prediction, rather than predicting the mark directly based on the mark probability as in existing studies. In conjunction with this method, we predict the mark first and then the time. In particular, we develop a novel neural Marked Temporal Point Process (MTPP) model to support effective time sampling and estimation of mark probability without computationally expensive numerical improper integration. Extensive experiments on real-world datasets demonstrate the superior performance of our solution against various baselines for the next event mark and time prediction.



Paperid:3168
Authors:Ralf Römer, Adrian Kobras, Luca Worbis, Angela Schoellig
Abstract:
Recent advances in imitation learning (IL) with generative models, such as diffusion and flow matching, have significantly improved the capabilities of robots to perform complex, long-horizon tasks. However, distribution shifts from unseen environments or compounding action errors can still cause unpredictable and unsafe behavior, leading to task failure. Therefore, being able to predict failures of generative IL policies as early as possible during runtime is crucial for deploying robots in human-centered and safety-critical environments. The two main existing approaches for recognizing failures both have major drawbacks. Methods relying solely on out-of-distribution (OOD) detection are prone to raising false alarms as policies might generalize. Conversely, external monitoring of the robot’s interactions with its environment provides no foresight about future behavior, meaning that failures can only be detected late or in retrospect. To close the existing gap, we propose FIPER, a framework for Failure Prediction at Runtime for generative IL policies without relying on failure examples. FIPER identifies two key indicators of policy failures: 1) consecutive OOD observations and 2) persistently high uncertainty (entropy) in generated actions. We calibrate both observation- and action-based failure prediction scores on a few successful rollouts and use conformal prediction to provide statistical performance guarantees. We evaluate our framework in five simulation and real-world environments where various types of failures can occur. Our results show that FIPER better distinguishes true failures from OOD situations and predicts failures earlier and more accurately than existing methods. We thus consider FIPER an important step towards more interpretable and safer generative robot policies. Our code and data are available at this link.



Paperid:3169
Authors:Till Freihaut, Luca Viano, Volkan Cevher, Matthieu Geist, Giorgia Ramponi
Abstract:
Abstract:This paper provides the first expert sample complexity characterization for learning a Nash equilibrium from expert data in Markov Games.We show that a new quantity named the *single policy deviation concentrability coefficient* is unavoidable in the non-interactive imitation learning setting, and we provide an upper bound for behavioral cloning (BC) featuring such coefficient. BC exhibits substantial regret in games with high concentrability coefficient, leading us to utilize expert queries to develop and introduce two novel solution algorithms: MAIL-BRO and MURMAIL. The former employs a best response oracle and learns an $\varepsilon$-Nash equilibrium with $\mathcal{O}(\varepsilon^{-4})$ expert and oracle queries. The latter bypasses completely the best response oracle at the cost of a worse expert query complexity of order $\mathcal{O}(\varepsilon^{-8})$. Finally, we provide numerical evidence, confirming our theoretical findings.



Paperid:3170
Authors:Tongtong Su, Yun Liao, Fengbo Zheng
Abstract:
Self-knowledge distillation (SKD) enables single-model training by distilling knowledge from the model's own outputs, eliminating the need for a separate teacher network required in conventional distillation methods. However, current SKD methods mainly focus on replicating common features in the student model, neglecting the extraction of key features that significantly enhance student learning. Inspired by this, we devise a self-knowledge distillation framework entitled Self-Distillation training via Proximal Gradient Optimization or SDPGO, which utilizes gradient information to identify and assign greater weight to features that significantly impact classification performance, enabling the network to learn the most relevant features during training. Additionally, this manner reduces the computational cost while enhancing the model's feature representation capability. Specifically, the proposed framework refines gradient information into a dynamically changing weighting factor to evaluate the distillation knowledge via the dynamic weight adjustment scheme. Our SDPGO technique uses the gradient information to extract the features most relevant to its own learning, and assigns more weights to the features that significantly affect the classification loss, which improves the representation learning of the student model. Extensive experimental results on five image classification benchmarks, including small-scale, large-scale, and fine-grained datasets, demonstrate that the proposed framework significantly outperforms recent state-of-the-art knowledge distillation methods.



Paperid:3171
Authors:Shikun Liu, Deyu Zou, Nima Shoghi, Victor Fung, Kai Liu, Pan Li
Abstract:
In the era of foundation models, fine-tuning pre-trained models for specific downstream tasks has become crucial. This drives the need for robust fine-tuning methods to address challenges such as model overfitting and sparse labeling. Moleculargraph foundation models (MGFMs) face unique difficulties that complicate fine-tuning. These models are limited by smaller pre-training datasets and more severedata scarcity for downstream tasks, both of which require enhanced model generalization. Moreover, MGFMs must accommodate diverse objectives, including bothregression and classification tasks. To better understand and improve fine-tuningtechniques under these conditions, we classify eight fine-tuning methods into threemechanisms: weight-based, representation-based, and partial fine-tuning. Webenchmark these methods on downstream regression and classification tasks acrosssupervised and self-supervised pre-trained models in diverse labeling settings. Thisextensive evaluation provides valuable insights and informs the design of a refinedrobust fine-tuning method, ROFT-MOL. This approach combines the strengths ofsimple post-hoc weight interpolation with more complex weight ensemble fine-tuning methods, delivering improved performance across both task types whilemaintaining the ease of use inherent in post-hoc weight interpolation.



Paperid:3172
Authors:Li Kang, Xiufeng Song, Heng Zhou, Yiran Qin, Jie Yang, Zhenfei Yin, Xiaohong Liu, Philip Torr, LEI BAI
Abstract:
Coordinating multiple embodied agents in dynamic environments remains a core challenge in artificial intelligence, requiring both perception-driven reasoning and scalable cooperation strategies. While recent works have leveraged large language models (LLMs) for multi-agent planning, a few have begun to explore vision-language models (VLMs) for visual reasoning. However, these VLM-based approaches remain limited in their support for diverse embodiment types. In this work, we introduce VIKI-Bench, the first hierarchical benchmark tailored for embodied multi-agent cooperation, featuring three structured levels: agent activation, task planning, and trajectory perception. VIKI-Bench includes diverse robot embodiments, multi-view visual observations, and structured supervision signals to evaluate reasoning grounded in visual inputs. To demonstrate the utility of VIKI-Bench, we propose VIKI-R, a two-stage framework that fine-tunes a pretrained vision-language model (VLM) using Chain-of-Thought annotated demonstrations, followed by reinforcement learning under multi-level reward signals. Our extensive experiments show that VIKI-R significantly outperforms baselines method across all task levels. Furthermore, we show that reinforcement learning enables the emergence of compositional cooperation patterns among heterogeneous agents. Together, VIKI-Bench and VIKI-R offer a unified testbed and method for advancing multi-agent, visual-driven cooperation in embodied AI systems.



Paperid:3173
Authors:Zedong Liu, Shenggan Cheng, Guangming Tan, Yang You, Dingwen Tao
Abstract:
Abstract:Multimodal large language models (MLLMs) extend LLMs to handle images, videos, and audio by incorporating feature extractors and projection modules. However, these additional components—combined with complex inference pipelines and heterogeneous workloads—introduce significant inference overhead. Therefore, efficiently serving MLLMs remains a major challenge. Current tightly coupled serving architectures struggle to distinguish between mixed request types or adapt parallelism strategies to different inference stages, leading to increased time-to-first-token (TTFT) latency and poor resource utilization. To address this, we propose Elastic Multimodal Parallelism (EMP), a new serving paradigm that elastically adapts to resource heterogeneity across request types and inference stages. Building upon EMP, we develop ElasticMM, an MLLM serving system that (1) separates requests into independent modality groups with dynamic resource allocation via a modality-aware load balancer; (2) decouples inference stages and enables parallelism adjustment and adaptive scaling via elastic partition scheduling; and (3) improves inference efficiency through unified multimodal prefix caching and non-blocking encoding. Experiments on diverse real-world datasets show that ElasticMM outperforms state-of-the-art (SOTA) serving systems, reducing TTFT by up to 4.2$\times$ and achieving 3.2–4.5$\times$ higher throughput while meeting service-level objectives (SLOs).



Paperid:3174
Authors:Qijun Luo, Yifei Shen, Liangzu Peng, Dongsheng Li, Xiao Li
Abstract:
Finetuning large language models (LLMs) is a resource-intensive task for researchers in academia, with memory constraints posing a key bottleneck. A classic optimization method, block coordinate descent (BCD), significantly reduces memory cost by segmenting the trainable parameters into multiple blocks and optimizing one active block at a time while freezing the others. However, we identify that blindly applying BCD to train LLMs can be inefficient for two reasons. First, optimizing only the active block requires backpropagating through multiple deeper yet inactive blocks, resulting in wasteful computations. Second, the frozen blocks, when they are not quite close to optimality, can narrow the optimization landscape, potentially misguiding the training of the active block. To address these issues simultaneously, we propose integrating BCD withlandscape expansion, which unfreezes the inactive blocks and updates them in a cost-efficient manner during the same backpropagation as the update to the active block. Experiments on 8B and 70B models demonstrate that our proposed method surpasses memory-efficient baselines and matches Adam's downstream performance while requiring only 24 GB of memory for the 8B model and 300 GB for the 70B model.



Paperid:3175
Authors:Rongkun Zheng, Lu Qi, Xi Chen, Yi Wang, Kun Wang, Hengshuang Zhao
Abstract:
Recent advancements in reinforcement learning have significantly advanced the reasoning capabilities of multimodal large language models (MLLMs). While methods like Group Relative Policy Optimization (GRPO) and rule-based reward mechanisms show promise in text and image domains, their application to video understanding remains underexplored. This paper systematically investigates RL with GRPO for video MLLMs, proposing a framework that enhances spatio-temporal perception capacities without compromising general capabilities. Through a sophisticated reward mechanism that integrates both format and spatial-temporal rewards to guide optimization directions with limited samples effectively, we develop STAR-R1, a powerful framework that achieves state-of-the-art performance on video understanding tasks while exhibiting strong spatio-temporal reasoning abilities on video perception benchmarks. Compared to Qwen2.5-VL-7B, STAR-R1 boosts performance in tasks like temporal grounding and general video QA benchmarks (such as VideoMME and MVBench). Additionally, it significantly improves the performance on perception tasks (such as video reasoning segmentation), demonstrating STAR-R1's ability to generalize across different video perception domains while presenting an explicit reasoning process. Our findings underscore the potential of reinforcement learning for improving the video perception capacity of Video MLLMs.



Paperid:3176
Authors:Chaoyang Wang, Ashkan Mirzaei, Vidit Goel, Willi Menapace, Aliaksandr Siarohin, Michael Vasilkovsky, Ivan Skorokhodov, Vladislav Shakhrai, Sergei Korolev, Sergey Tulyakov, Peter Wonka
Abstract:
We propose the first framework capable of computing a 4D spatio-temporal grid of video frames and 3D Gaussian particles for each time step using a feed-forward architecture. Our architecture has two main components, a 4D video model and a 4D reconstruction model. In the first part, we analyze current 4D video diffusion architectures that perform spatial and temporal attention either sequentially or in parallel within a two-stream design. We highlight the limitations of existing approaches and introduce a novel fused architecture that performs spatial and temporal attention within a single layer. The key to our method is a sparse attention pattern, where tokens attend to others in the same frame, at the same timestamp, or from the same viewpoint.In the second part, we extend existing 3D reconstruction algorithms by introducing a Gaussian head, a camera token replacement algorithm, and additional dynamic layers and training. Overall, we establish a new state of the art for 4D generation, improving both visual quality and reconstruction capability.



Paperid:3177
Authors:Hao Zhang, Zhan Zhuang, Xuehao Wang, Xiaodong Yang, Yu Zhang
Abstract:
Human Activity Recognition (HAR) with wearable sensors is challenged by limited interpretability, which significantly impacts cross-dataset generalization. To address this challenge, we propose Motion-Primitive Transformer (MoPFormer), a novel self-supervised framework that enhances interpretability by tokenizing inertial measurement unit signals into semantically meaningful motion primitives and leverages a Transformer architecture to learn rich temporal representations. MoPFormer comprises two-stages. first stage is to partition multi-channel sensor streams into short segments and quantizing them into discrete "motion primitive" codewords, while the second stage enriches those tokenized sequences through a context-aware embedding module and then processes them with a Transformer encoder. The proposed MoPFormer can be pre-trained using a masked motion-modeling objective that reconstructs missing primitives, enabling it to develop robust representations across diverse sensor configurations. Experiments on six HAR benchmarks demonstrate that MoPFormer not only outperforms state-of-the-art methods but also successfully generalizes across multiple datasets. Most importantly, the learned motion primitives significantly enhance both interpretability and cross-dataset performance by capturing fundamental movement patterns that remain consistent across similar activities regardless of dataset origin.



Paperid:3178
Authors:Yin Hang, Peiwen Yuan, Xiaomin He, Yiwei Li, Jiayi Shi, Wenxiao Fan, Shaoxiong Feng, Prof. Kan
Abstract:
Abstract:Existing vision-language models often suffer from spatial hallucinations, i.e., generating incorrect descriptions about the relative positions of objects in an image. We argue that this problem mainly stems from the asymmetric properties between images and text. To enrich the spatial understanding ability of vision-language models, we propose a simple, annotation-free, plug-and-play method named {Stitch and Tell} (abbreviated as SiTe), which injects structured spatial supervision into data. It constructs stitched image–text pairs by stitching images along a spatial axis and generating spatially-aware captions or question answer pairs based on the layout of stitched image, without relying on costly advanced models or human involvement. We evaluate SiTe across three architectures including LLaVA-v1.5-7B, LLaVA-Qwen2-1.5B and HALVA-7B, two training datasets, and eight benchmarks. Experiments show that SiTe improves spatial understanding tasks such as $\text{MME}_{\text{Position}}$ (+5.50\%) and Spatial-MM (+4.19\%), while maintaining or improving performance on general vision-language benchmarks including COCO-QA (+1.02\%) and MMBench (+4.76\%). Our findings suggest that explicitly injecting spatially-aware structure into training data offers an effective way to mitigate spatial hallucinations and improve spatial understanding, while preserving general vision-language capabilities.



Paperid:3179
Authors:Jinkun Hao, Naifu Liang, Zhen Luo, Weipeng Zhong, Ran Yi, Yichen Jin, Zhaoyang Lyu, Xudong XU, Feng Zheng, Lizhuang Ma, Jiangmiao Pang
Abstract:
The ability of robots to interpret human instructions and execute manipulation tasks necessitates the availability of task-relevant tabletop scenes for training. However, traditional methods for creating these scenes rely on time-consuming manual layout design or purely randomized layouts, which are limited in terms of plausibility or alignment with the tasks. In this paper, we formulate a novel task, namely task-oriented tabletop scene generation, which poses significant challenges due to the substantial gap between high-level task instructions and the tabletop scenes. To support research on such a challenging task, we introduce \textbf{MesaTask-10K}, a large-scale dataset comprising approximately 10,700 synthetic tabletop scenes with \emph{manually crafted layouts} that ensure realistic layouts and intricate inter-object relations. To bridge the gap between tasks and scenes, we propose a \textbf{Spatial Reasoning Chain} that decomposes the generation process into object inference, spatial interrelation reasoning, and scene graph construction for the final 3D layout. We present \textbf{MesaTask}, an LLM-based framework that utilizes this reasoning chain and is further enhanced with DPO algorithms to generate physically plausible tabletop scenes that align well with given task descriptions. Exhaustive experiments demonstrate the superior performance of MesaTask compared to baselines in generating task-conforming tabletop scenes with realistic layouts.



Paperid:3180
Authors:振伟 张, Fanhua Shang, Hongying Liu, Liang Wan, Wei Feng, Yanming Hui
Abstract:
Abstract:While model quantization has become pivotal for deploying super-resolution (SR) networks on mobile devices, existing works focus on quantization methods for image super-resolution. Different from image super-resolution, the temporal error propagation, shared temporal parameterization, and temporal metric mismatch significantly degrade the performance of a video SR model. To address these problems, we propose the first quantization method, QBasicVSR, for video super-resolution. A temporal awareness adaptation post-training quantization (PTQ) framework for video super-resolution with the flow-gradient video bit adaptation and temporal shared layer bit adaptation is presented. Moreover, we put forward a novel fine-tuning method for VSR with the supervision of the full-precision model. Our method achieves extraordinary performance with state-of-the-art efficient VSR approaches, delivering up to $\times$200 faster processing speed while utilizing only 1/8 of the GPU resources. Additionally, extensive experiments demonstrate that the proposed method significantly outperforms existing PTQ algorithms on various datasets. For instance, it attains a 2.53 dB increase on the UDM10 benchmark when quantizing BasicVSR to 4-bit with 100 unlabeled video clips. The code and models will be released on GitHub.



Paperid:3181
Authors:Ruida Hu, Chao Peng, XinchenWang, Junjielong Xu, Cuiyun Gao
Abstract:
Scaling up executable code data is significant for improving language models’ software engineerin capability. The intricate nature of the process makes it labor-intensive, time-consuming and expert-knowledge-dependent to build a large number of executable code repositories, limiting the scalability of existing work based on running tests. The primary bottleneck lies in the automated building of test environments for different repositories, which is an essential yet underexplored task. To mitigate the gap, we introduce Repo2Run, the first LLM-based agent aiming at automating the building of executable test environments for any repositories at scale. Specifically, given a code repository, Repo2Run iteratively builds the Docker image, runs unit tests based on the feedback of the building, and synthesizes the Dockerfile until the entire pipeline is executed successfully. The resulting Dockerfile can then be used to create Docker container environments for running code and tests. We created a benchmark containing 420 Python repositories with unit tests for evaluation. The results illustrate that Repo2Run achieves an 86.0% success rate, outperforming SWE-agent by 77.0%. The resources of Repo2Run are available at https://anonymous.4open.science/r/Repo2Run.



Paperid:3182
Authors:Peter Súkeník, Christoph Lampert, Marco Mondelli
Abstract:
The empirical emergence of neural collapse---a surprising symmetry in the feature representations of the training data in the penultimate layer of deep neural networks---has spurred a line of theoretical research aimed at its understanding. However, existing work focuses on data-agnostic models or, when data structure is taken into account, it remains limited to multi-layer perceptrons. Our paper fills both these gaps by analyzing modern architectures in data-aware regime: we prove that global optima of deep regularized transformers and residual networks (ResNets) with LayerNorm trained with cross entropy or mean squared error loss are approximately collapsed, and the approximation gets tighter as the depth grows. More generally, we formally reduce any end-to-end large-depth ResNet or transformer training into an equivalent unconstrained features model, thus justifying its wide use in the literature even beyond data-agnostic settings. Our theoretical results are supported by experiments on computer vision and language datasets showing that, as the depth grows, neural collapse indeed becomes more prominent.



Paperid:3183
Authors:Zitao Chen, Yinjun Jia, Zitong Tian, Wei-Ying Ma, Yanyan Lan
Abstract:
Medicinal chemists often optimize drugs considering their 3D structures and designing structurally distinct molecules that retain key features, such as shapes, pharmacophores, or chemical properties. Previous deep learning approaches address this through supervised tasks like molecule inpainting or property-guided optimization. In this work, we propose a flexible zero-shot molecule manipulation method by navigating in a shared latent space of 3D molecules. We introduce a Variational AutoEncoder (VAE) for 3D molecules, named MolFLAE, which learns a fixed-dimensional, SE(3)-equivariant latent space independent of atom counts. MolFLAE encodes 3D molecules using an SE(3)-equivariant neural network into fixed number of latent nodes, distinguished by learned embeddings. The latent space is regularized, and molecular structures are reconstructed via a Bayesian Flow Network (BFN) conditioned on the encoder’s latent output. MolFLAE achieves competitive performance on standard unconditional 3D molecule generation benchmarks. Moreover, the latent space of MolFLAE enables zero-shot molecule manipulation, including atom number editing, structure reconstruction, and coordinated latent interpolation for both structure and properties. We further demonstrate our approach on a drug optimization task for the human glucocorticoid receptor, generating molecules with improved hydrophilicity while preserving key interactions, under computational evaluations. These results highlight the flexibility, robustness, and real-world utility of our method, opening new avenues for molecule editing and optimization.



Paperid:3184
Authors:Mingyang Yi, Bohan Wang
Abstract:
Recently, optimization on the Riemannian manifold has provided new insights to the optimization community. In this regard, the manifold taken as the probability measure metric space equipped with the second-order Wasserstein distance is of particular interest, since optimization on it can be linked to practical sampling processes. In general, the standard (continuous) optimization method on Wasserstein space is Riemannian gradient flow (i.e., Langevin dynamics when minimizing KL divergence). In this paper, we aim to enrich the continuous optimization methods in the Wasserstein space, by extending the gradient flow on it into the stochastic gradient descent (SGD) flow and stochastic variance reduction gradient (SVRG) flow. The two flows in Euclidean space are standard continuous stochastic methods, while their Riemannian counterparts are unexplored. By leveraging the property of Wasserstein space, we construct stochastic differential equations (SDEs) to approximate the corresponding discrete dynamics of desired Riemannian stochastic methods in Euclidean space. Then, our probability measures flows are obtained by the Fokker-Planck equation. Finally, the convergence rates of our Riemannian stochastic flows are proven, which match the results in Euclidean space.



Paperid:3185
Authors:Xingang Guo, Yaxin Li, XiangYi Kong, YILAN JIANG, Xiayu Zhao, Zhihua Gong, Yufan Zhang, Daixuan Li, Tianle Sang, Beixiao Zhu, Gregory Jun, Yingbing Huang, Yiqi Liu, Yuqi Xue, Rahul Dev Kundu, Qi Lim, Yizhou Zhao, Luke Granger, Mohamed Younis, Darioush Keivan, Nippun Sabharwal, Shreyanka Sinha, Prakhar Agarwal, Kojo Vandyck, Hanlin Mai, Zichen Wang, Aditya Venkatesh, Ayush Barik, Jiankun Yang, Chongying Yue, Jingjie He, Libin Wang, Licheng Xu, Hao Chen, Jinwen Wang, Liujun Xu, Rushabh Shetty, Ziheng Guo, Dahui Song, Manvi Jha, Weijie Liang, Weiman Yan, Bryan Zhang, Sahil Bhandary Karnoor, Jialiang Zhang, Rutva Pandya, Xinyi Gong, Mithesh Ganesh, Feize Shi, Ruiling Xu, Yifan Zhang, Yanfeng Ouyang, Lianhui Qin, Elyse Rosenbaum, Corey Snyder, Peter Seiler, Geir Dullerud, Xiaojia Zhang, Zuofu Cheng, Pavan Kumar Hanumolu, Jian Huang, Mayank Kulkarni, Mahdi Namazifar, Huan Zhang, Bin Hu
Abstract:
Modern engineering, spanning electrical, mechanical, aerospace, civil, and computer disciplines, stands as a cornerstone of human civilization and the foundation of our society. Today, industry pioneers dream of developing AI systems capable of designing and building humanity's most ambitious projects—from starships that will carry us to distant worlds to Dyson spheres that harness stellar energy. Yet engineering design represents a fundamentally different challenge for large language models (LLMs) compared to traditional textbook-style problem solving or factual question answering. Although existing benchmarks have driven progress in areas such as factual question answering, code synthesis, and scientific problem solving, real-world engineering design demands synthesis of domain knowledge, navigation of complex trade-offs, and management of the tedious processes that consume much of practicing engineers' time. Despite these shared challenges across engineering disciplines, no benchmark currently captures the unique demands of engineering design work. In this work, we introduce EngDesign, an Engineering Design benchmark that evaluates LLMs' abilities to perform practical design tasks across nine engineering domains: Operating System Design, Computer Architecture Design, Control System Design, Mechanical Systems, Structural Design, Digital Hardware Design, Analog Integrated Circuit Design, Robotics, and Signal Processing. Unlike existing benchmarks that focus on factual recall or question answering, EngDesign uniquely emphasizes LLMs’ ability to synthesize domain knowledge, reason under constraints, and generate functional, objective-oriented designs. Each task in EngDesign represents a real-world engineering design problem, accompanied by a detailed task description specifying design goals, constraints, and performance requirements. We pioneer a simulation-based evaluation paradigm where LLM-generated designs undergo rigorous testing through executable, domain-specific simulations—from circuit SPICE simulations to structural finite element analysis, from control system validation to robotic motion planning. EngDesign establishes a new benchmark paradigm that moves beyond textbook knowledge to assess genuine engineering capability and shifts evaluation from static answer checking to dynamic, simulation-driven functional verification, marking a crucial step toward realizing the vision of engineering AGI.



Paperid:3186
Authors:Kaixiang Huang, Qifeng Zhang, Jin Wang, Jingru Yang, Yang Zhou, Huan Yu, Guodong Lu, Shengfeng He
Abstract:
3D Visual Grounding (3DVG) faces persistent challenges due to coarse scene-level observations and logically inconsistent annotations, which introduce ambiguities that compromise data quality and hinder effective model supervision. To address these challenges, we introduce Refer-Judge, a novel framework that harnesses the reasoning capabilities of Multimodal Large Language Models (MLLMs) to identify and mitigate toxic data. At the core of Refer-Judge is a Jury-and-Judge Chain-of-Thought paradigm, inspired by the deliberative process of the judicial system. This framework targets the root causes of annotation noise: jurors collaboratively assess 3DVG samples from diverse perspectives, providing structured, multi-faceted evaluations. Judges then consolidate these insights using a \textit{Corroborative Refinement} strategy, which adaptively reorganizes information to correct ambiguities arising from biased or incomplete observations. Through this two-stage deliberation, Refer-Judge significantly enhances the reliability of data judgments. Extensive experiments demonstrate that our framework not only achieves human-level discrimination at the scene level but also improves the performance of baseline algorithms via data purification. Code is available in the supplementary materials.



Paperid:3187
Authors:Joey Hong, Anca Dragan, Sergey Levine
Abstract:
Large language models (LLMs) excel in tasks like question answering and dialogue, but complex tasks requiring interaction, such as negotiation and persuasion, require additional long-horizon reasoning and planning. Reinforcement learning (RL) fine-tuning can enable such planning in principle, but suffers from drawbacks that hinder scalability. In particular, multi-turn RL training incurs high memory and computational costs, which are exacerbated when training LLMs as policies. Furthermore, the largest LLMs do not expose the APIs necessary to be trained in such manner. As a result, modern methods to improve the reasoning of LLMs rely on sophisticated prompting mechanisms rather than RL fine-tuning. To remedy this, we propose a novel approach that uses goal-conditioned value functions to guide the reasoning of LLM agents, that scales even to large API-based models. These value functions predict how a task will unfold given an action, allowing the LLM agent to evaluate multiple possible outcomes, both positive and negative, to plan effectively. In addition, these value functions are trained over reasoning steps rather than full actions, to be a concise and light-weight module that facilitates decision-making in multi-turn interactions. We validate our method on tasks requiring interaction, including tool use, social deduction, and dialogue, demonstrating superior performance over both RL fine-tuning and prompting methods while maintaining efficiency and scalability.



Paperid:3188
Authors:Baiyuan Chen, Shinji Ito, Masaaki Imaizumi
Abstract:
Transformers have demonstrated exceptional performance across a wide range of domains. While their ability to perform reinforcement learning in-context has been established both theoretically and empirically, their behavior in non-stationary environments remains less understood. In this study, we address this gap by showing that transformers can achieve nearly optimal dynamic regret bounds in non-stationary settings. We prove that transformers are capable of approximating strategies used to handle non-stationary environment, and can learn the approximator in the in-context learning setup. Our experiments further show that transformers can match or even outperform existing expert algorithms in such environments.



Paperid:3189
Authors:Yamato Arai, Yuma Ichikawa
Abstract:
Layer-wise PTQ is a promising technique for compressing large language models (LLMs), due to its simplicity and effectiveness without requiring retraining. However, recent progress in this area is saturating, underscoring the need to revisit its core limitations and explore further improvements. We address this challenge by identifying a key limitation of existing layer-wise PTQ methods: the growth of quantization errors across layers significantly degrades performance, particularly in low-bit regimes. To address this fundamental issue, we propose Quantization Error Propagation (QEP), a general, lightweight, and scalable framework that enhances layer-wise PTQ by explicitly propagating quantization errors and compensating for accumulated errors. QEP also offers a tunable propagation mechanism that prevents overfitting and controls computational overhead, enabling the framework to adapt to various architectures and resource budgets. Extensive experiments on several LLMs demonstrate that QEP-enhanced layer-wise PTQ achieves substantially higher accuracy than existing methods. Notably, the gains are most pronounced in the extremely low-bit quantization regime.



Paperid:3190
Authors:Akide Liu, Zeyu Zhang, Zhexin Li, Xuehai Bai, Yizeng Han, Jiasheng Tang, Yuanjie Xing, Jichao Wu, Mingyang Yang, Weihua Chen, Jiahao He, Yuanyu He, Fan Wang, Reza Haffari, Bohan Zhuang
Abstract:
Abstract:Diffusion generative models have become the standard for producing high-quality, coherent video content, yet their slow inference speeds and high computational demands hinder practical deployment. Although both quantization and sparsity can independently accelerate inference while maintaining generation quality, naively combining these techniques in existing training-free approaches leads to significant performance degradation, as they fail to achieve proper joint optimization.We introduce FPSAttention, a novel training-aware co-design of FP8 quantization and Sparsity for video generation, with a focus on the 3D bi-directional attention mechanism. Our approach features three key innovations: 1) A unified 3D tile-wise granularity that simultaneously supports both quantization and sparsity. 2) A denoising step-aware strategy that adapts to the noise schedule, addressing the strong correlation between quantization/sparsity errors and denoising steps. 3) A native, hardware-friendly kernel that leverages FlashAttention and is implemented with optimized Hopper architecture features, enabling highly efficient execution.Trained on Wan2.1's 1.3B and 14B models and evaluated on the vBench benchmark, FPSAttention achieves a 7.09$\times$ kernel speedup for attention operations and a 4.96$\times$ end-to-end speedup for video generation compared to the BF16 baseline at 720p resolution—without sacrificing generation quality.



Paperid:3191
Authors:Eline Bovy, Caleb Probine, Marnix Suilen, Ufuk Topcu, Nils Jansen
Abstract:
Multi-environment POMDPs (ME-POMDPs) extend standard POMDPs with discrete model uncertainty. ME-POMDPs represent a finite set of POMDPs that share the same state, action, and observation spaces, but may arbitrarily vary in their transition, observation, and reward models. Such models arise, for instance, when multiple domain experts disagree on how to model a problem. The goal is to find a single policy that is robust against any choice of POMDP within the set,i.e., a policy that maximizes the worst-case reward across all POMDPs. We generalize and expand on existing work in the following way. First, we show that ME-POMDPs can be generalized to POMDPswith sets of initial beliefs, which we calladversarial-belief POMDPs(AB-POMDPs). Second, we show that any arbitrary ME-POMDP can be reduced to a ME-POMDP that only varies in its transition and reward functions or only in its observation and reward functions, while preserving (optimal) policies. We then devise exact and approximate (point-based) algorithms to compute robust policies for AB-POMDPs, and thus ME-POMDPs. We demonstrate that we can compute policies for standard POMDP benchmarks extended to the multi-environment setting.



Paperid:3192
Authors:Lisa Weijler, Sebastian Koch, Fabio Poiesi, Timo Ropinski, Pedro Hermosilla
Abstract:
Modeling the inherent hierarchical structure of 3D objects and 3D scenes is highly desirable, as it enables a more holistic understanding of environments for autonomous agents. Accomplishing this with implicit representations, such as Neural Radiance Fields, remains an unexplored challenge. Existing methods that explicitly model hierarchical structures often face significant limitations: they either require multiple rendering passes to capture embeddings at different levels of granularity, significantly increasing inference time, or rely on predefined, closed-set discrete hierarchies that generalize poorly to the diverse and nuanced structures encountered by agents in the real world. To address these challenges, we propose OpenHype, a novel approach that represents scene hierarchies using a continuous hyperbolic latent space. By leveraging the properties of hyperbolic geometry, OpenHype naturally encodes multi-scale relationships and enables smooth traversal of hierarchies through geodesic paths in latent space. Our method outperforms state-of-the-art approaches on standard benchmarks, demonstrating superior efficiency and adaptability in 3D scene understanding.



Paperid:3193
Authors:Xiaoxue Cheng, Junyi Li, Zhenduo Zhang, Xinyu Tang, Xin Zhao, Xinyu Kong, Zhiqiang Zhang
Abstract:
Large reasoning models (LRMs) have demonstrated strong performance on complex reasoning tasks, but often suffer from overthinking, generating redundant content regardless of task difficulty. Inspired by the dual process theory in cognitive science, we propose Adaptive Cognition Policy Optimization (ACPO), a reinforcement learning framework that enables LRMs to achieve efficient reasoning through adaptive cognitive allocation and dynamic system switch.ACPO incorporates two key components: (1) introducing system-aware reasoning tokens to explicitly represent the thinking modes thereby making the model's cognitive process transparent, and (2) integrating online difficulty estimation and token length budget to guide adaptive system switch and reasoning during reinforcement learning. To this end, we propose a two-stage training strategy. The first stage begins with supervised fine-tuning to cold start the model, enabling it to generate reasoning paths with explicit thinking modes. In the second stage, we apply ACPO to further enhance adaptive system switch for difficulty-aware reasoning.Experimental results demonstrate that ACPO effectively reduces redundant reasoning while adaptively adjusting cognitive allocation based on task complexity, achieving efficient hybrid reasoning.



Paperid:3194
Authors:睿涵 杨, Fanghua Ye, Jian Li, Siyu Yuan, yikai zhang, Zhaopeng Tu, Xiaolong Li, Deqing Yang
Abstract:
Large language models (LLMs) have recently transformed from text-based assistants to autonomous agents capable of planning, reasoning, and iteratively improving their actions. While numerical reward signals and verifiers can effectively rank candidate actions, they often provide limited contextual guidance. In contrast, natural language feedback better aligns with the generative capabilities of LLMs, providing richer and more actionable suggestions. However, parsing and implementing this feedback effectively can be challenging for LLM-based agents. In this work, we introduce Critique-Guided Improvement (CGI), a novel two-player framework, comprising an actor model that explores an environment and a critic model that generates detailed nature language feedback. By training the critic to produce fine-grained assessments and actionable revisions, and the actor to utilize these critiques, our approach promotes more robust exploration of alternative strategies while avoiding local optima. Experiments in three interactive environments show that CGI outperforms existing baselines by a substantial margin. Notably, even a small critic model surpasses GPT-4 in feedback quality. The resulting actor achieves state-of-the-art performance, demonstrating the power of explicit iterative guidance to enhance decision-making in LLM-based agents.



Paperid:3195
Authors:Chinmay Talegaonkar, Nikhil Gandudi Suresh, Zachary Novack, Yash Belhe, Priyanka Nagasamudra, Nicholas Antipa
Abstract:
Recent monocular metric depth estimation (MMDE) methods have made notable progress towards zero-shot generalization. However, they still exhibit a significant performance drop on out-of-distribution datasets. We address this limitation by injecting defocus blur cues at inference time into Marigold, a \textit{pre-trained} diffusion model for zero-shot, scale-invariant monocular depth estimation (MDE). Our method effectively turns Marigold into a metric depth predictor in a training-free manner. To incorporate defocus cues, we capture two images with a small and a large aperture from the same viewpoint. To recover metric depth, we then optimize the metric depth scaling parameters and the noise latents of Marigold at inference time using gradients from a loss function based on the defocus-blur image formation model. We compare our method against existing state-of-the-art zero-shot MMDE methods on a self-collected real dataset, showing quantitative and qualitative improvements.



Paperid:3196
Authors:Zhenqing Ling, Daoyuan Chen, Liuyi Yao, Qianli Shen, Yaliang Li, Ying Shen
Abstract:
Fine-tuning large language models (LLMs) using diverse datasets is crucial for enhancing their overall performance across various domains.In practical scenarios, existing methods based on modeling the mixture proportions of data composition often struggle with data whose domain labels are missing, imprecise or non-normalized, while methods based on data selection usually encounter difficulties in balancing multi-domain performance.To address these challenges, in this work, we investigate the role of data diversity in enhancing the overall abilities of LLMs by empirically constructing contrastive data pools and theoretically deriving explanations. Building upon the insights gained, we propose a new method that gives the LLM a dual identity: an output model to cognitively probe and select data based on diversity reward, as well as an input model to be tuned with the selected data.Extensive experiments show that the proposed method notably boosts performance across domain-undetermined data and a series of foundational downstream tasks when applied to various advanced LLMs. We release our code and hope this study can shed light on the understanding of data diversity and advance feedback-driven data-model co-design for LLMs.



Paperid:3197
Authors:Weijian Ma, Shizhao Sun, Ruiyu Wang, Jiang Bian
Abstract:
A Computer-Aided Design (CAD) model encodes an object in two coupled forms: a \emph{parametric constructions sequence} and its resulting \emph{visible geometric shape}.During iterative design, adjustments to the geometric shape inevitably require synchronized edits to the underlying parametric sequence, called \emph{geometry-driven parametric CAD editing}.The task calls for 1) preserving the original sequence’s structure, 2) ensuring each edit's semantic validity, and 3) maintaining high shape fidelity to the target shape, all under scarce editing data triplets.We present \emph{CADMorph}, an iterative \emph{plan–generate–verify} framework that combines two pretrained models during inference: a \emph{parameter-to-shape} (P2S) latent diffusion model and a \emph{masked-parameter-prediction} (MPP) model.In the planning stage, cross-attention maps from the P2S model pinpoint the segments that need modification and offer editing masks. The MPP model then infills these masks with semantically valid edits in the generation stage. During verification, the P2S model embeds each candidate sequence in shape-latent space, measures its distance to the target shape, and selects the closest one. The three stages thus tackle structure preservation, semantic validity, and shape fidelity respectively. Besides, both P2S and MPP models are trained without triplet data, bypassing the data-scarcity bottleneck.CADMorph surpasses GPT-4o and specialized CAD baselines, and supports downstream applications such as iterative editing and reverse-engineering enhancement.



Paperid:3198
Authors:Yash Jhaveri, Harley Wiltzer, Patrick Shafto, Marc Bellemare, David Meger
Abstract:
In the pursuit of finding an optimal policy, reinforcement learning (RL) methods generally ignore the properties of learned policies apart from their expected return. Thus, even when successful, it is difficult to characterizewhichpolicies will be learned andwhatthey will do. In this work, we present a theoretical framework for policy optimization that guarantees convergence to aparticularoptimal policy, via vanishing entropy regularization and a noveltemperature decoupling mechanism. Our approach realizes an interpretable, diversity-preserving optimal policy as the regularization temperature vanishes and ensures the convergence of policy derived objects—value functions andreturn distributions. In a particular instance of our method, for example, the realized policy samples all optimal actions uniformly. Inspired by these insights, we provide a convergent algorithm for control in distributional RL, addressing an open problem in distributional RL theory and bringing stability to optimal return distribution estimation.



Paperid:3199
Authors:Jia Nuo Liew, Shenghan Lin, Bowen Chen, Wei Zhang, Xiaowei Zhu, Wei Zhang, Xiaolin Hu
Abstract:
Abstract:The ability to extract oriented edges from visual input is a core computation across animal vision systems. Orientation maps, long associated with the layered architecture of the mammalian visual cortex, systematically organise neurons by their preferred edge orientation. Despite lacking cortical structures, the $\textit{ Drosophila melanogaster}$ brain contains feature-selective neurons and exhibits complex visual detection capacity, raising the question of whether map-like vision representations can emerge without cortical infrastructure. We integrate a complete fruit fly brain connectome with biologically grounded spiking neuron models to simulate neuroprocessing in the fly visual system. By driving the network with oriented stimuli and analysing downstream responses, we show that coherent orientation maps can emerge from purely connectome-constrained dynamics. These results suggest that species of independent origin could evolve similar visual structures, offering new insights into the evolution of visual processing and bio-inspired AI architectures.



Paperid:3200
Authors:Junqi You, Chieh Lin, Weijie Lyu, Zhengbo Zhang, Ming-Hsuan Yang
Abstract:
Abstract:Recent advances in 3D scene reconstruction enable real-time viewing in virtual and augmented reality. To support interactive operations for better immersiveness, such as moving or editing objects, 3D scene inpainting methods are proposed to repair or complete the altered geometry. To support users in interacting (such as moving or editing objects) with the scene for the next level of immersiveness, 3D scene inpainting methods are developed to repair the altered geometry. However, current approaches rely on lengthy and computationally intensive optimization, making them impractical for real-time or online applications. We propose InstaInpaint, a reference-based feed-forward framework that produces 3D-scene inpainting from a 2D inpainting proposal within 0.4 seconds. We develop a self-supervised masked-finetuning strategy to enable training of our custom large reconstruction model (LRM) on the large-scale dataset. Through extensive experiments, we analyze and identify several key designs that improve generalization, textural consistency, and geometric correctness. InstaInpaint achieves a 1000$\times$ speed-up from prior methods while maintaining a state-of-the-art performance across two standard benchmarks. Moreover, we show that InstaInpaint generalizes well to flexible downstream applications such as object insertion and multi-region inpainting.



Paperid:3201
Authors:Yuanchuan Guo, Jun Liu, Huimin Cheng, Ying Ma
Abstract:
As spatial transcriptomics (ST) datasets increasingly span multiple adjacent or replicated slices, effective joint analysis across slices is needed to reconstruct tissue structures and identify consistent spatial gene expression patterns. This requires resolving spatial correspondences between slices while capturing shared transcriptomic features, two tasks that are typically addressed in isolation. Multi-slice analysis remains challenging due to physical distortions, technical variability, and batch effects. To address these challenges, we introduce Joint Alignment and Deep Embedding for multi-slice ST (JADE), a unified computational framework that simultaneously learns spot-wise alignments and shared low-dimensional embeddings across tissue slices. Unlike existing methods, JADE adopts a roundtrip framework in which each iteration alternates between alignment and embedding refinement. To infer alignment, we employ attention mechanisms that dynamically assess and weight the importance of different embedding dimensions, allowing the model to focus on the most alignment-relevant features while suppressing noise. To the best of our knowledge, JADE is the first method that jointly optimizes alignment and representation learning in a shared latent space, enabling robust multi-slice integration. We demonstrate that JADE outperforms existing alignment and embedding methods across multiple evaluation metrics in the 10x Visium human dorsolateral prefrontal cortex (DLPFC) and Stereo-seq axolotl brain datasets. By bridging spatial alignment and feature integration, JADE provides a scalable and accurate solution for cross-slice analysis of ST data.



Paperid:3202
Authors:Haotian Luo, Haiying He, Yibo Wang, Jinluan Yang, Rui Liu, Naiqiang Tan, Xiaochun Cao, Dacheng Tao, Li Shen
Abstract:
Recently, long-thought reasoning models achieve strong performance on complex reasoning tasks, but often incur substantial inference overhead, making efficiency a critical concern. Our empirical analysis reveals that the benefit of using Long-CoT varies across problems: while some problems require elaborate reasoning, others show no improvement—or even degraded accuracy. This motivates adaptive reasoning strategies that tailor reasoning depth to the input. However, prior work primarily reduces redundancy within long reasoning paths, limiting exploration of more efficient strategies beyond the Long-CoT paradigm. To address this, we propose a novel two-stage framework for adaptive and efficient reasoning. First, we construct a hybrid reasoning model by merging long and short CoT models to enable diverse reasoning styles. Second, we apply bi-level preference training to guide the model to select suitable reasoning styles (group-level), and prefer concise and correct reasoning within each style group (instance-level). Experiments demonstrate that our method significantly reduces inference costs compared to other baseline approaches, while maintaining performance. Notably, on five mathematical datasets, the average length of reasoning is reduced by more than 50\%, highlighting the potential of adaptive strategies to optimize reasoning efficiency in large language models.



Paperid:3203
Authors:Wei Pang, Kevin Qinghong Lin, Xiangru Jian, Xi He, Philip Torr
Abstract:
Academic poster generation is a crucial yet challenging task in scientific communication, requiring the compression of long-context interleaved documents into a single, visually coherent page. To address this challenge, we introduce Paper2Poster, the first benchmark and metric suite for poster generation, which pairs recent conference papers with author-designed posters and evaluates outputs on (i) Visual Quality—semantic alignment with human posters, (ii) Textual Coherence—language fluency, (iii) Holistic Assessment—six fine-grained aesthetic and informational criteria scored by a VLM-as-judge, and notably (iv) PaperQuiz—the poster’s ability to convey core paper content as measured by VLMs answering generated quizzes. Building on this benchmark, we propose PosterAgent, a top‐down, visual‐in‐the‐loop multi‐agent pipeline: the (a) Parser distills the paper into a structured asset library; the (b) Planner aligns text–visual pairs into a binary‐tree layout that preserves reading order and spatial balance; and the (c) Painter–Commenter loop refines each panel by executing rendering code and using VLM feedback to eliminate overflow and ensure alignment.In our comprehensive evaluation, we find that GPT‐4o outputs—though visually appealing at first glance—often exhibit noisy text and poor PaperQuiz scores; We find that reader engagement is the primary aesthetic bottleneck, as human‐designed posters rely largely on visual semantics to convey meaning.Our fully open‐source Paper2Poster pipeline outperforms GPT‐4o–based systems across nearly all metrics while consuming 87 \% fewer tokens. These findings chart clear directions for the next generation of fully automated poster‐generation models.



Paperid:3204
Authors:Zhenjie Mao, Yang Yuhuan, Chaofan Ma, Dongsheng Jiang, Jiangchao Yao, Ya Zhang, Yanfeng Wang
Abstract:
Referring Image Segmentation (RIS) aims to segment the target object in an image given a natural language expression. While recent methods leverage pre-trained vision backbones and more training corpus to achieve impressive results, they predominantly focus on simple expressions—short, clear noun phrases like “red car” or “left girl”. This simplification often reduces RIS to a key word/concept matching problem, limiting the model’s ability to handle referential ambiguity in expressions. In this work, we identify two challenging real-world scenarios: object-distracting expressions, which involve multiple entities with contextual cues, and category-implicit expressions, where the object class is not explicitly stated. To address the challenges, we propose a novel framework, SaFiRe, which mimics the human two-phase cognitive process—first forming a global understanding, then refining it through detail-oriented inspection. This is naturally supported by Mamba’s scan-then-update property, which aligns with our phased design and enables efficient multi-cycle refinement with linear complexity. We further introduce aRefCOCO, a new benchmark designed to evaluate RIS models under ambiguous referring expressions. Extensive experiments on both standard and proposed datasets demonstrate the superiority of SaFiRe over state-of-the-art baselines.



Paperid:3205
Authors:Wenjun Ding, Jingling Liu, Lixing Chen, Xiu Su, Tao Sun, Fan Wu, Zhe Qu
Abstract:
Meta-learning has achieved significant advancements, with generalization emerging as a key metric for evaluating meta-learning algorithms. While recent studies have mainly focused on training strategies, data-split methods, and tightening generalization bounds, they often ignore the impact of inner-levels on generalization. To bridge this gap, this paper focuses on several prominent meta-learning algorithms and establishes two generalization analytical frameworks for them based on their inner-processes: the Gradient Descent Framework (GDF) and the Proximal Descent Framework (PDF). Within these frameworks, we introduce two novel algorithmic stability definitions and derive the corresponding generalization bounds. Our findings reveal a trade-off of inner-levels under GDF, whereas PDF exhibits a beneficial relationship. Moreover, we highlight the critical role of the meta-objective function in minimizing generalization error. Inspired by this, we propose a new, simplified meta-objective function definition to enhance generalization performance. Many real-world experiments support our findings and show the improvement of the new meta-objective function.



Paperid:3206
Authors:Nan Li, Yonghui Su, Lianbo Ma
Abstract:
Network quantization effectively reduces both memory footprints and inference time of deep neural networks, enabling their deployment on resource-constrained devices. To fully utilize the multiple bit-width arithmetic operations of the hardware, mixed-precision quantization (MPQ) is developed to assign different bit-widths to each layer. However, the quantization policy obtained by existing MPQ methods struggles to achieve the objectives of efficiency and generalization simultaneously. In this paper, we propose an efficient and generalizable MPQ based on topological entropy (GMPQ-TE). Specifically, topological entropy, derived from topological data analysis, effectively measures the quantization sensitivity of each layer by using the minibatch of data with the same label. Furthermore, we observe that topological entropy remains consistent across various datasets and shows a strong correlation with both quantized model accuracy and bit-width. Thus, MPQ is formulated as a single-pass linear programming problem, obtaining a generalizable quantization policy in a few seconds (11s on MobileNet-V2). Extensive experiments show that the quantization policy obtained on CIFAR-10 can generalize to ImageNet and PASCAL VOC. GMPQ-TE achieves a competitive accuracy-complexity trade-off compared to state-of-the-art MPQ methods.



Paperid:3207
Authors:Lei Cao, Hao Zhang, Chunyu Li, Jiayi Ma
Abstract:
This study proposes PDFuse, a robust, general training-free image fusion framework built on pre-trained latent diffusion models with projection–manifold regularization. By redefining fusion as a diffusion inference process constrained by multiple source images, PDFuse can adapt to varied image modalities and produce high-fidelity outputs utilizing the diffusion prior. To ensure both source consistency and full utilization of generative priors, we develop novel projection–manifold regularization, which consists of two core mechanisms. On the one hand, the Multi-source Information Consistency Projection (MICP) establishes a projection system between diffusion latents and source images, solved efficiently via conjugate gradients to inject multi-source information into the inference. On the other hand, the Latent Manifold-preservation Guidance (LMG) aligns the latent distribution of diffusion variables with that of the sources, guiding generation to respect the model’s manifold prior. By alternating these mechanisms, PDFuse strikes an optimal balance between fidelity and generative quality, achieving superior fusion performance across diverse tasks. Moreover, PDFuse constructs a canonical interference operator set. It synergistically incorporates it into the aforementioned dual mechanisms, effectively leveraging generative priors to address various degradation issues during the fusion process without requiring clean data for supervising training. Extensive experimental evidence substantiates that PDFuse achieves highly competitive performance across diverse image fusion tasks.



Paperid:3208
Authors:Tianteng Gu, Bei Liu, Bo Xiao, Ke Zeng, Jiacheng Liu, Yanmin Qian
Abstract:
Pruning is a widely used technique to compress large language models (LLMs) by removing unimportant weights, but it often suffers from significant performance degradation—especially under semi-structured sparsity constraints. Existing pruning methods primarily focus on estimating the importance of individual weights, which limits their ability to preserve critical capabilities of the model. In this work, we propose a new perspective: rather than merely selecting which weights to prune, we first redistribute parameter importance to make the model inherently more amenable to pruning. By minimizing the information entropy of normalized importance scores, our approach concentrates importance onto a smaller subset of weights, thereby enhancing pruning robustness. We instantiate this idea through DenoiseRotator, which applies learnable orthogonal transformations to the model’s weight matrices. Our method is model-agnostic and can be seamlessly integrated with existing pruning techniques such as Magnitude, SparseGPT, and Wanda. Evaluated on LLaMA3, Qwen2.5, and Mistral models under 50% unstructured and 2:4 semi-structured sparsity, DenoiseRotator consistently improves perplexity and zero-shot accuracy. For instance, on LLaMA3-70B pruned with SparseGPT at 2:4 semi-structured sparsity, DenoiseRotator reduces the perplexity gap to the dense model by 58%, narrowing the degradation from 8.1 to 3.4 points.



Paperid:3209
Authors:Jongchan Park, Mingyu Park, Donghwan Lee
Abstract:
Abstract:Offline reinforcement learning (RL) aims to learn a policy from a static dataset without further interactions with the environment. Collecting sufficiently large datasets for offline RL is exhausting since this data collection requires colossus interactions with environments and becomes tricky when the interaction with the environment is restricted. Hence, how an agent learns the best policy with a minimal static dataset is a crucial issue in offline RL, similar to the sample efficiency problem in online RL. In this paper, we propose a simple yet effective plug-and-play pretraining method to initialize a feature of a $Q$-network to enhance data efficiency in offline RL. Specifically, we introduce a shared $Q$-network structure that outputs predictions of the next state and $Q$-value. We pretrain the shared $Q$-network through a supervised regression task that predicts a next state and trains the shared $Q$-network using diverse offline RL methods. Through extensive experiments, we empirically demonstrate that our method enhances the performance of existing popular offline RL methods on the D4RL, Robomimic and V-D4RL benchmarks. Furthermore, we show that our method significantly boosts data-efficient offline RL across various data qualities and data distributions trough D4RL and ExoRL benchmarks. Notably, our method adapted with only 10\% of the dataset outperforms standard algorithms even with full datasets.



Paperid:3210
Authors:Remco Leijenaar, Hamidreza Kasaei
Abstract:
Learning semantically meaningful representations from unstructured 3D point clouds remains a central challenge in computer vision, especially in the absence of large-scale labeled datasets. While masked point modeling (MPM) is widely used in self-supervised 3D learning, its reconstruction-based objective can limit its ability to capture high-level semantics. We propose AsymDSD, an Asymmetric Dual Self-Distillation framework that unifies masked modeling and invariance learning through prediction in the latent space rather than the input space. AsymDSD builds on a joint embedding architecture and introduces several key design choices: an efficient asymmetric setup, disabling attention between masked queries to prevent shape leakage, multi-mask sampling, and a point cloud adaptation of multi-crop. AsymDSD achieves state-of-the-art results on ScanObjectNN (90.53\%) and further improves to 93.72\% when pretrained on 930k shapes, surpassing prior methods.



Paperid:3211
Authors:Zhiyuan Wang, Jinwoo Go, Byung-Jun Yoon, Nathan Urban, Xiaoning Qian
Abstract:
In recent developments in scientific machine learning (SciML), neural surrogate solvers for partial differential equations (PDEs) have become powerful tools for accelerating scientific computation for various science and engineering applications. However, training neural PDE solvers often demands a large amount of high-fidelity PDE simulation data, which are expensive to generate. Active learning (AL) offers a promising solution by adaptively selecting training data from the PDE settings--including parameters, initial and boundary conditions--that are expected to be most informative to help reduce this data burden. In this work, we introduce PaPQS, a Plug-and-Play Query Synthesis AL framework that synthesizes informative PDE settings directly in the continuous design space. PaPQS optimizes the Expected Information Gain (EIG) while encouraging batch diversity, enabling model-aware exploration of the design space via backpropagation through the neural PDE solution trajectories. The framework is applicable to general PDE systems and surrogate architectures, and can be seamlessly integrated with existing AL strategies. Extensive experiments across different PDE systems demonstrate that our AL framework, PaPQS, consistently improves sample efficiency over existing AL baselines.



Paperid:3212
Authors:Federico Berto, Chuanbo Hua, Laurin Luttmann, Jiwoo Son, Junyoung Park, Kyuree Ahn, Changhyun Kwon, Lin Xie, Jinkyoo Park
Abstract:
Combinatorial optimization problems involving multiple agents are notoriously challenging due to their NP-hard nature and the necessity for effective agent coordination. Despite advancements in learning-based methods, existing approaches often face critical limitations, including suboptimal agent coordination, poor generalization, and high computational latency. To address these issues, we propose PARCO (Parallel AutoRegressive Combinatorial Optimization), a general reinforcement learning framework designed to construct high-quality solutions for multi-agent combinatorial tasks efficiently. To this end, PARCO integrates three key novel components: (1) transformer-based communication layers to enable effective agent collaboration during parallel solution construction, (2) a multiple pointer mechanism for low-latency, parallel agent decision-making, and (3) priority-based conflict handlers to resolve decision conflicts via learned priorities. We evaluate PARCO in multi-agent vehicle routing and scheduling problems, where our approach outperforms state-of-the-art learning methods, demonstrating strong generalization ability and remarkable computational efficiency. We make our source code publicly available to foster future research.



Paperid:3213
Authors:Di Liu, Meng Chen, Baotong Lu, Huiqiang Jiang, Zhenhua Han, Qianxi Zhang, Qi Chen, Chengruidong Zhang, Bailu Ding, Kai Zhang, Chen Chen, Fan Yang, Yuqing Yang, Lili Qiu
Abstract:
Transformer-based Large Language Models (LLMs) have become increasingly important. However, scaling LLMs to longer contexts incurs slow inference speed and high GPU memory consumption for caching key-value (KV) vectors. This paper presents RetrievalAttention, a training-free approach to both accelerate the decoding phase and reduce GPU memory consumption by pre-building KV vector indexes for fixed contexts and maintaining them in CPU memory for efficient retrieval. Unlike conventional KV cache methods, RetrievalAttention integrate approximate nearest neighbor search (ANNS) indexes into attention computation. We observe that off-the-shelf ANNS techniques often fail due to the out-of-distribution (OOD) nature of query and key vectors in attention mechanisms. RetrievalAttention overcomes this with an attention-aware vector index. Our evaluation shows RetrievalAttention achieves near full attention accuracy while accessing only 1-3\% of the data, significantly reducing inference costs. Remarkably, RetrievalAttention enables LLMs with 8B parameters to handle 128K tokens on a single NVIDIA RTX4090 (24GB), achieving a decoding speed of 0.107 seconds per token.



Paperid:3214
Authors:Pengyi Li, Jianye Hao, Hongyao Tang, Jinbin Qiao, YAN ZHENG
Abstract:
The integration of evolutionary algorithms (EAs) with reinforcement learning (RL) has shown superior performance compared to standalone methods. However, previous research focuses on exploration in policy parameter space, while overlooking the reward function search. To bridge this gap, we propose LaRes, a novel hybrid framework that achieves efficient policy learning through reward function search. LaRes leverages large language models (LLMs) to generate the reward function population, guiding RL in policy learning. The reward functions are evaluated by the policy performance and improved through LLMs. To improve sample efficiency, LaRes employs a shared experience buffer that collects experiences from all policies, with each experience containing rewards from all reward functions. Upon reward function updates, the rewards of experiences are relabeled, enabling efficient use of historical data. Furthermore, we introduce a Thompson sampling-based selection mechanism that enables more efficient elite interaction. To prevent policy collapse when improving reward functions, we propose a reward scaling and parameter constraint mechanism to efficiently coordinate reward search with policy learning. Across 16 robot manipulation tasks, LaRes outperforms other strong RL and ERL baselines guided by expert rewards in terms of both sample efficiency and final performance.



Paperid:3215
Authors:Rui Li, Quanyu Dai, Zeyu Zhang, Xiaohe Bo, Zihang Tian, Xu Chen, Zhenhua Dong, Ruiming Tang
Abstract:
Current Large Language Models (LLMs) are confronted with overwhelming information volume when comprehending long-form documents. This challenge raises the imperative of a cohesive memory module, which can elevate vanilla LLMs into autonomous reading agents. Despite the emergence of some heuristic approaches, a systematic design principle remains absent. To fill this void, we draw inspiration from Jean Piaget’s Constructivist Theory, illuminating three traits of the agentic memory—structured schemata, flexible assimilation, and dynamic accommodation. This blueprint forges a clear path toward a more robust and efficient memory system for LLM-based reading comprehension. To this end, we develop CAM, a prototype implementation of Constructivist Agentic Memory that simultaneously embodies the structurality, flexibility, and dynamicity. At its core, CAM is endowed with an incremental overlapping clustering algorithm for structured memory development, supporting both coherent hierarchical summarization and online batch integration. During inference, CAM adaptively explores the memory structure to activate query-relevant information for contextual response, akin to the human associative process. Compared to existing approaches, our design demonstrates dual advantages in both performance and efficiency across diverse long-text reading comprehension tasks, including question answering, query-based summarization, and claim verification.



Paperid:3216
Authors:Guancheng Wan, Xiaoran Shang, Guibin Zhang, Jinhe Bi, Yuxin Wu, Liangtao Zheng, Xin Lin, Yue Liu, Yanbiao Ma, Wenke Huang, Bo Du
Abstract:
Abstract:Robust Federated Graph Learning (FGL) provides an effective decentralized framework for training Graph Neural Networks (GNNs) in noisy-label environments. However, the subtlety of noise during training presents formidable obstacles for developing robust FGL systems. Previous robust FL approaches neither adequately constrain edge-mediated error propagation nor account for intra-class topological differences. At the client level, we innovatively demonstrate that hyperspherical embedding can effectively capture graph structures in a fine-grained manner. Correspondingly, our method effectively addresses the aforementioned issues through fine-grained hypersphere alignment. Moreover, we uncover undetected noise arising from localized perspective constraints and propose the geometric-aware hyperspherical purification module at the server level. Combining both level strategies, we present our robust FGL framework,**HYPERION**, which operates all components within a unified hyperspherical space. **HYPERION** demonstrates remarkable robustness across multiple datasets, for instance, achieving a 29.7\% $\uparrow$ F1-macro score with 50\%-pair noise on Cora. The code is available for anonymous access at \url{https://anonymous.4open.science/r/Hyperion-NeurIPS/}.



Paperid:3217
Authors:Ahmet Karadeniz, Dimitrios Mallis, Danila Rukhovich, Kseniya Cherenkova, Anis Kacem, Djamila Aouada
Abstract:
Computer-Aided Design (CAD) plays a foundational role in modern manufacturing and product development, often requiring designers to modify or build upon existing models. Converting 3D scans into parametric CAD representations—a process known as CAD reverse engineering—remains a significant challenge due to the high precision and structural complexity of CAD models. Existing deep learning-based approaches typically fall into two categories: bottom-up, geometry-driven methods, which often fail to produce fully parametric outputs, and top-down strategies, which tend to overlook fine-grained geometric details. Moreover, current methods neglect an essential aspect of CAD modeling: sketch-level constraints. In this work, we introduce a novel approach to CAD reverse engineering inspired by how human designers manually perform the task. Our method leverages multi-plane cross-sections to extract 2D patterns and capture fine parametric details more effectively. Our method enables the reconstruction of detailed and editable CAD models, outperforming state-of-the-art methods and, for the first time, incorporating sketch constraints directly into the reconstruction process.



Paperid:3218
Authors:Eric Michaud, Asher Parker-Sartori, Max Tegmark
Abstract:
We study the problem of creating strong, yet narrow, AI systems. While recent AI progress has been driven by the training of large general-purpose foundation models, the creation of smaller models specialized for narrow domains could be valuable for both efficiency and safety. In this work, we explore two challenges involved in creating such systems, having to do with basic properties of how neural networks learn and structure their representations. The first challenge regards when it is possible to train narrow models from scratch. Through experiments on a synthetic task, we find that it is sometimes necessary to train networks on a wide distribution of data to learn certain narrow skills within that distribution. This effect arises when skills depend on each other hierarchically, and training on a broad distribution introduces a curriculum which substantially accelerates learning. The second challenge regards how to transfer particular skills from large general models into small specialized models. We find that model skills are often not perfectly localized to a particular set of prunable components. However, we find that methods based on pruning can still outperform distillation. We investigate the use of a regularization objective to align desired skills with prunable components while unlearning unnecessary skills.



Paperid:3219
Authors:Basile Confavreux, William Dorrell, Nishil Patel, Andrew Saxe
Abstract:
Synaptic plasticity is widely considered to be crucial to the brain’s ability to learn throughout life. Decades of theoretical work have therefore been invested in deriving and designing biologically plausible learning rules capable of granting various memory abilities to neural networks. Most of these theoretical approaches optimize directly for a desired memory function; but this procedure can lead to complex, finely-tuned rules, rendering them brittle to perturbations and difficult to implement in practice. Instead, we build on recent work that automatically discovers large numbers of candidate plasticity rules operating in recurrent spiking neural networks. Surprisingly, despite the fact that these rules are selected solely to achieve network stabilization, we observe across a range of network models---feedforward, recurrent; rate and spiking---that almost all these rules endow the network with memory - seemingly by accident. To understand this phenomenon, we study an analytic toy model. We show that memory arises from the degeneracy of weight matrices that stabilize a network: where the network lands in this space of stable weights depends on its past inputs---that is, memory. Even simple Hebbian plasticity rules can utilize this degeneracy, creating a zoo of memory abilities with various lifetimes. In practice, the larger the network and the more co-active plasticity rules in the system, the stronger the memory-by-accident phenomenon becomes. Overall, our findings suggest that activity-silent memory is a near-unavoidable consequence of stabilization. Simple forms of memory, such as familiarity or novelty detection, appear to be widely available resources for plastic brain networks, suggesting that they could form the raw materials that were later sculpted into higher-order cognitive abilities.



Paperid:3220
Authors:Ahmed Boughdiri, Clément Berenfeld, Julie Josse, Erwan Scornet
Abstract:
Generalization methods offer a powerful solution to one of the key drawbacks of randomized controlled trials (RCTs): their limited representativeness. By enabling the transport of treatment effect estimates to target populations subject to distributional shifts, these methods are increasingly recognized as the future of meta-analysis, the current gold standard in evidence-based medicine.Yet most existing approaches focus on the risk difference, overlooking the diverse range of causal measures routinely reported in clinical research. Reporting multiple effect measures—both absolute (e.g., risk difference, number needed to treat) and relative (e.g., risk ratio, odds ratio)—is essential to ensure clinical relevance, policy utility, and interpretability across contexts.To address this gap, we propose a unified framework for transporting a broad class of first-moment population causal effect measures under covariate shift. We provide identification results under two conditional exchangeability assumptions, derive both classical and semiparametric estimators, and evaluate their performance through theoretical analysis, simulations, and real-world applications. Our analysis shows the specificity of different causal measures and thus the interest of studying them all: for instance, two common approaches (one-step, estimating equation) lead to similar estimators for the risk difference but to two distinct estimators for the odds ratio.



Paperid:3221
Authors:Li Tenghui, Guoxu Zhou, Xuyang Zhao, Yuning Qiu, Qibin Zhao
Abstract:
Abstract:As the length of input text grows, the key-value (KV) cache in LLMs imposes prohibitive GPU memory costs and limits long‐context inference on resource‐constrained devices. Existing approaches, such as KV quantization and pruning, reduce memory usage but suffer from numerical precision loss or suboptimal retention of key-value pairs. We introduce Low Rank Query and Key attention (LRQK), a two‐stage framework that jointly decomposes the full‐precision query and key matrices into compact rank-$r$ factors during the prefill stage, and then uses these low-dimensional projections to compute proxy attention scores in $\mathcal{O}(lr)$ time at each decode step. By selecting only the top-$k$ tokens and a small fixed set of recent tokens, LRQK employs a hierarchical GPU–CPU cache with a hit-and-miss mechanism that transfers only missing full-precision KV pairs, thereby preserving exact attention outputs while reducing CPU–GPU data movement. Extensive experiments on the RULER and LongBench benchmarks with LLaMA-3-8B and Qwen2.5-7B demonstrate that LRQK matches or surpasses leading sparse-attention methods in long context settings, while delivering significant memory savings with minimal loss in accuracy.



Paperid:3222
Authors:Mohamad Hakam Shams Eddin, Yikui Zhang, Stefan Kollet, Jürgen Gall
Abstract:
Abstract:Recent deep learning approaches for river discharge forecasting have improved the accuracy and efficiency in flood forecasting, enabling more reliable early warning systems for risk management. Nevertheless, existing deep learning approaches in hydrology remain largely confined to local-scale applications and do not leverage the inherent spatial connections of bodies of water. Thus, there is a strong need for new deep learning methodologies that are capable of modeling spatio-temporal relations to improve river discharge and flood forecasting for scientific and operational applications. To address this, we present RiverMamba, a novel deep learning model that is pretrained with long-term reanalysis data and that can forecast global river discharge and floods on a $0.05^\circ$ grid up to 7 days lead time, which is of high relevance in early warning. To achieve this, RiverMamba leverages efficient Mamba blocks that enable the model to capture global-scale channel network routing and enhance its forecast capability for longer lead times. The forecast blocks integrate ECMWF HRES meteorological forecasts, while accounting for their inaccuracies through spatio-temporal modeling. Our analysis demonstrates that RiverMamba delivers reliable predictions of river discharge, including extreme floods across return periods and lead times, surpassing both operational AI- and physics-based models.



Paperid:3223
Authors:Zhengping Jiang, Anqi Liu, Ben Van Durme
Abstract:
Language model outputs are not always reliable, thus prompting research into how to adapt model responses based on uncertainty. Common approaches include: abstention, where models refrain from generating responses when uncertain; and linguistic calibration, where models hedge their statements using uncertainty quantifiers. However, abstention can withhold valuable information, while linguistically calibrated responses are often challenging to leverage in downstream tasks. We propose a unified view, Conformal Linguistic Calibration (CLC), which reinterprets linguistic calibration as answer set prediction. First we present a framework connecting abstention and linguistic calibration through the lens of linguistic pragmatics. We then describe an implementation of CLC that allows for controlling the level of imprecision in model responses. Results demonstrate our method produces calibrated outputs with conformal guarantees on factual accuracy. Further, our approach enables fine-tuning models to perform uncertainty-aware adaptive claim rewriting, offering a controllable balance between factuality and specificity.



Paperid:3224
Authors:Xiaotian Qiao, Ke Xu, Xianglong Yang, Ruijie Dong, Xiaofang Xia, Jiangtao Cui
Abstract:
Shadow characteristics are of great importance for scene understanding.Traditionally, shadow regions are considered as binary masks, often resulting in imprecise detection and suboptimal performance in tasks like shadow removal.We demonstrate that such an assumption oversimplifies light-object interactions in the scene, as the scene details under either hard or soft shadows remain visible to a certain degree.Based on this insight, we aim to reformulate the shadow detection paradigm from the opacity perspective, and introduce a new fine-grained shadow detection method.In particular, given the input image, we first propose a shadow opacity augmentation module to generate realistic images with varied shadow opacities.We then introduce a shadow feature separation module to learn the shadow position and opacity representations separately, followed by an opacity mask prediction module that fuses these representations and predicts fine-grained shadow detection results.In addition, we construct a new dataset with opacity-annotated shadow masks across varied scenarios.Extensive experiments demonstrate that our method outperforms the baselines qualitatively and quantitatively, enhancing downstream shadow applications like shadow removal and editing.The dataset and code will be made publicly available.



Paperid:3225
Authors:Yuchen Ma, Dennis Frauen, Jonas Schweisthal, Stefan Feuerriegel
Abstract:
Estimating treatment effects is crucial for personalized decision-making in medicine, but this task faces unique challenges in clinical practice. At training time, models for estimating treatment effects are typically trained on well-structured medical datasets that contain detailed patient information. However, at inference time, predictions are often made using textual descriptions (e.g., descriptions with self-reported symptoms), which are incomplete representations of the original patient information. In this work, we make three contributions. (1)We show that the discrepancy between the data available during training time and inference time can lead to biased estimates of treatment effects. We formalize this issue as an inference time text confounding problem, where confounders are fully observed during training time but only partially available through text at inference time. (2)To address this problem, we propose a novel framework for estimating treatment effects that explicitly accounts for inference time text confounding. Our framework leverages large language models together with a custom doubly robust learner to mitigate biases caused by the inference time text confounding. (3)Through a series of experiments, we demonstrate the effectiveness of our framework in real-world applications.



Paperid:3226
Authors:Ghada Sokar, Pablo Samuel Castro
Abstract:
Scaling deep reinforcement learning in pixel-based environments presents a significant challenge, often resulting in diminished performance. While recent works have proposed algorithmic and architectural approaches to address this, the underlying cause of the performance drop remains unclear. In this paper, we identify the connection between the output of the encoder (a stack of convolutional layers) and the ensuing dense layers as the main underlying factor limiting scaling capabilities; we denote this connection as thebottleneck, and we demonstrate that previous approaches implicitly target this bottleneck. As a result of our analyses, we present global average pooling as a simple yet effective way of targeting the bottleneck, thereby avoiding the complexity of earlier approaches.



Paperid:3227
Authors:Jiaqi Wang, Zhiguang Cao, Peng Zhao, Rui Cao, YuBin Xiao, YUAN JIANG, You Zhou
Abstract:
Abstract:The rise of smart manufacturing under Industry 4.0 introduces mass customization and dynamic production, demanding more advanced and flexible scheduling techniques. The flexible job-shop scheduling problem (FJSP) has attracted significant attention due to its complex constraints and strong alignment with real-world production scenarios. Current deep reinforcement learning (DRL)-based approaches to FJSP predominantly employ constructive methods. While effective, they often fall short of reaching (near-)optimal solutions. In contrast, improvement-based methods iteratively explore the neighborhood of initial solutions and are more effective in approaching optimality. However, the flexible machine allocation in FJSP poses significant challenges to the application of this framework, including accurate state representation, effective policy learning, and efficient search strategies. To address these challenges, this paper proposes a $\textbf{M}$emory-enhanced $\textbf{I}$mprovement $\textbf{S}$earch framework with he$\textbf{t}$erogeneous gr$\textbf{a}$ph $\textbf{r}$epresentation—$\textit{MIStar}$. It employs a novel heterogeneous disjunctive graph that explicitly models the operation sequences on machines to accurately represent scheduling solutions. Moreover, a memory-enhanced heterogeneous graph neural network (MHGNN) is designed for feature extraction, leveraging historical trajectories to enhance the decision-making capability of the policy network. Finally, a parallel greedy search strategy is adopted to explore the solution space, enabling superior solutions with fewer iterations. Extensive experiments on synthetic data and public benchmarks demonstrate that $\textit{MIStar}$ significantly outperforms both traditional handcrafted improvement heuristics and state-of-the-art DRL-based constructive methods.



Paperid:3228
Authors:Stephane Gaubert, Julien Grand-Clément, Ricardo Katz
Abstract:
Abstract:We investigate refinements of the mean-payoff criterion in two-player zero-sum perfect-information stochastic games. A strategy is *Blackwell optimal* if it is optimal in the discounted game for all discount factors sufficiently close to $1$. The notion of *$d$-sensitive optimality* interpolates between mean-payoff optimality (corresponding to the case $d=-1$) and Blackwell optimality ($d=\infty$). The *Blackwell threshold* $\alpha_{\sf bw} \in [0,1[$ is the discount factor above which all optimal strategies in the discounted game are guaranteed to be Blackwell optimal. The *$d$-sensitive threshold* $\alpha_{\sf d} \in [0,1[$ is defined analogously. Bounding $\alpha_{\sf bw}$ and $\alpha_{\sf d}$ are fundamental problems in algorithmic game theory, since these thresholds control the complexity for computing Blackwell and $d$-sensitive optimal strategies, by reduction to discounted games which can be solved in $O\left((1-\alpha)^{-1}\right)$ iterations. We provide the first bounds on the $d$-sensitive threshold $\alpha_{\sf d}$ beyond the case $d=-1$, and we establish improved bounds for the Blackwell threshold $\alpha_{\sf bw}$. This is achieved by leveraging separation bounds on algebraic numbers, relying on Lagrange bounds and more advanced techniques based on Mahler measures and multiplicity theorems.



Paperid:3229
Authors:Vaclav Voracek, Francesco Orabona
Abstract:
Abstract:The construction of confidence intervals for the mean of a bounded random variable is a classical problem in statistics with numerous applications in machine learning and virtually all scientific fields.In particular, obtaining the tightest possible confidence intervals is vital every time the sampling of the random variables is expensive.The current state-of-the-art method to construct confidence intervals is by using betting algorithms. This is a very successful approach for deriving optimal confidence sequences, even matching the rate of law of iterated logarithms. However, in the fixed horizon setting, these approaches are either sub-optimal or based on heuristic solutions with strong empirical performance but without a finite-time guarantee. Hence, no betting-based algorithm guaranteeing the optimal $\mathcal{O}(\sqrt{\frac{\sigma^2\log\frac1\delta}{n}})$ width of the confidence intervals are known.This work bridges this gap. We propose a betting-based algorithm to compute confidence intervals that empirically outperforms the competitors. Our betting strategy uses the optimal strategy in every step (in a certain sense), whereas the standard betting methods choose a constant strategy in advance. Leveraging this fact results in strict improvements even for classical concentration inequalities, such as the ones of Hoeffding or Bernstein.Moreover, we also prove that the width of our confidence intervals is optimal up to an $1+o(1)$ factor diminishing with $n$.



Paperid:3230
Authors:Haoran Zhou, Gim Hee Lee
Abstract:
Recent advancements in foundation models for 2D vision have enabled more effective analysis of dynamic scenes from monocular videos. However, despite their strong generalization capabilities, these models often lack 3D consistency, leading to spatial misalignment and temporal flickering that limit their effectiveness in dynamic 3D environments. In this paper, we present Motion4D, a novel framework that addresses these challenges by integrating 2D priors from foundation models into a unified 4D scene representation. Our method features a two-part iterative optimization framework: 1) Sequential optimization which updates motion and semantic fields in consecutive stages to maintain local consistency, and 2) Global optimization which jointly refines all attributes for long-term coherence. To enhance motion accuracy, we introduce a 3D confidence map that dynamically adjusts the supervision of motion priors, and an adaptive resampling process that inserts new Gaussians into under-represented regions based on per-pixel RGB and semantic errors. Furthermore, we enhance semantic coherence through an iterative refinement process that corrects 2D semantic inconsistencies by iteratively optimizing the semantic fields and updating prompts of SAM2. Extensive evaluations demonstrate that our Motion4D significantly outperforms both 2D foundation models and existing 3D representation methods in dynamic scene understanding tasks including video object segmentation and point-based tracking. Our code will be open-sourced on acceptance.



Paperid:3231
Authors:Feng Yao, Zilong Wang, Liyuan Liu, Junxia Cui, Li Zhong, Xiaohan Fu, Haohui Mai, Viswanathan Krishnan, Jianfeng Gao, Jingbo Shang
Abstract:
Code generation with large language models (LLMs), often termed vibe coding, is increasingly adopted in production but fails to ensure code quality, particularly in security (e.g., SQL injection vulnerabilities) and maintainability (e.g., missing type annotations). Existing methods, such as supervised fine-tuning and rule-based post-processing, rely on labor-intensive annotations or brittle heuristics, limiting their scalability and effectiveness. We propose REAL (Reinforcement rEwards from Automated anaLysis), a reinforcement learning framework that trains LLMs to generate production-quality code using program analysis–guided feedback. Specifically, REAL integrates two automated signals: (1) static analyzers detecting security and maintainability defects and (2) unit tests ensuring functional correctness. Unlike prior work, our framework is prompt-agnostic and reference-free, enabling scalable supervision without manual intervention. Experiments across multiple datasets and model scales demonstrate that REAL outperforms state-of-the-art methods in simultaneous assessments of functionality and code quality. Our work bridges the gap between rapid prototyping and production-ready code, enabling LLMs to deliver both speed and quality.



Paperid:3232
Authors:Chuanchao Zang, Xiangtao Meng, Wenyu Chen, Tianshuo Cong, Zha Yaxing, Dong Qi, Zheng Li, Shanqing Guo
Abstract:
The open-source release of large language models (LLMs) enables malicious users to create unauthorized derivative models at low cost, posing significant threats to intellectual property (IP) and market stability. Existing IP protection methods either require access to model parameters or are vulnerable to fine-tuning attacks. To fill this gap, we propose ErrorTrace, a robust and black-box traceability mechanism for protecting LLM IP. Specifically, ErrorTrace leverages the unique error patterns of model families by mapping and analyzing their distinct error spaces, enabling robust and efficient IP protection without relying on internal parameters or specific query responses. Experimental results show that ErrorTrace achieves a traceability accuracy of 0.8518 for 27 base models when the suspect model is not included in ErrorTrace's training set, outperforming the baseline by 0.2593. Additionally,ErrorTrace successfully tracks 34 fine-tuned, pruned and merged models across various scenarios, demonstrating its broad applicability and robustness.



Paperid:3233
Authors:Samuele Bortolotti, Emanuele Marconato, Paolo Morettin, Andrea Passerini, Stefano Teso
Abstract:
Concept-based Models are neural networks that learn a concept extractor to map inputs to high-level concepts and an inference layer to translate these into predictions. Ensuring these modules produce interpretable concepts and behave reliably in out-of-distribution is crucial, yet the conditions for achieving this remain unclear. We study this problem by establishing a novel connection between Concept-based Models and reasoning shortcuts (RSs), a common issue where models achieve high accuracy by learning low-quality concepts, even when the inference layer is fixed and provided upfront. Specifically, we extend RSs to the more complex setting of Concept-based Models and derive theoretical conditions for identifying both the concepts and the inference layer. Our empirical results highlight the impact of RSs and show that existing methods, even combined with multiple natural mitigation strategies, often fail to meet these conditions in practice.



Paperid:3234
Authors:Dennis Wei, Inkit Padhi, Soumya Ghosh, Amit Dhurandhar, Karthikeyan Natesan Ramamurthy, Maria Chang
Abstract:
Training data attribution (TDA) is concerned with understanding model behavior in terms of the training data. This paper draws attention to the common setting where one has access only to the final trained model, and not the training algorithm or intermediate information from training. To serve as a gold standard for TDA in this "final-model-only" setting, we proposefurther training, with appropriate adjustment and averaging, to measure the sensitivity of the given model to training instances. We then unify existing gradient-based methods for TDA by showing that they all approximate the further training gold standard in different ways. We investigate empirically the quality of these gradient-based approximations to further training, for tabular, image, and text datasets and models. We find that the approximation quality of first-order methods is sometimes high but decays with the amount of further training. In contrast, the approximations given by influence function methods are more stable but surprisingly lower in quality.



Paperid:3235
Authors:Daniel Wurgaft, Ekdeep S Lubana, Core Francisco Park, Hidenori Tanaka, Gautam Reddy, Noah Goodman
Abstract:
In-Context Learning (ICL) enables language models to adapt to novel tasks, significantly expanding their generality. Prior work has documented a rich phenomenology of ICL in specific, simplified settings—particularly the emergence of multiple solutions as task diversity and the number of observed samples increase. However, it remains unclear why these phenomena arise and whether they can be captured under a unified framework. Motivated by this, and drawing on common assumptions about simplicity bias and neural scaling, we develop a hierarchical Bayesian account that accurately predicts a Transformer’s generalization behavior in standard ICL settings—without requiring access to its parameters. Our framework views Transformer training as a process of estimating evidence for different solutions, and inference-time behavior as a posterior-weighted average over these solutions’ predictions. This perspective naturally implies a tradeoff between loss and complexity among candidate solutions, offering a unified explanation for prior empirical findings and yielding novel predictions. For instance, we show that sublinear sample efficiency during pretraining can induce a superlinear trend in the timescale for transitioning between solutions as task diversity scales. Together, our results advance a unified theory of ICL grounded in tradeoffs between solution loss and complexity in model space.



Paperid:3236
Authors:Sai Niranjan Ramachandran, Manish Krishan Lal, Suvrit Sra
Abstract:
Abstract:We analyze how the sampling dynamics of distributions evolve in score-based diffusion models using \emph{cross-fluctuations}, a centered-moment statistic from statistical physics. Specifically, we show that starting from an unbiased isotropic normal distribution, samples undergo sharp, discrete transitions, eventually forming distinct events of a desired distribution while progressively revealing finer structure. As this process is reversible, these transitions also occur in reverse, where intermediate states progressively merge, tracing a path back to the initial distribution. We demonstrate that these transitions can be detected as discontinuities in $n^{\text{th}}$-order cross-fluctuations. For variance-preserving SDEs, we derive a closed-form for these cross-fluctuations that is efficiently computable for the reverse trajectory, thus tying cross-fluctuation dynamics to event formation within the desired distribution. We find that detecting these transitions directly boosts sampling efficiency, accelerates class-conditional and rare-class generation, and improves two zero-shot tasks--image classification and style transfer--without expensive grid search or retraining. We also show that this viewpoint unifies classical coupling and mixing from finite Markov chains with continuous dynamics while extending to stochastic SDEs and non-Markovian samplers. Our framework therefore bridges discrete Markov chain theory, phase analysis, and modern generative modeling.



Paperid:3237
Authors:Niraj Kumar, Tyler Chen, Akshay Seshadri, Mattia Jacopo Villani, Pradeep Niroula, Shouvanik Chakrabarti, Pranav Deshpande, Romina Yalovetzky, Archan Ray, Marco Pistoia
Abstract:
Shapley values have emerged as a critical tool for explaining which features impact the decisions made by machine learning models.However, computing exact Shapley values is difficult, generally requiring an exponential (in the feature dimension) number of model evaluations. To address this, many model-agnostic randomized estimators have been developed, the most influential and widely used being the KernelSHAP method (Lundberg & Lee, 2017). While related estimators such as unbiased KernelSHAP (Covert & Lee, 2021) and LeverageSHAP (Musco & Witter, 2025) are known to satisfy theoretical guarantees, bounds for KernelSHAP have remained elusive. We describe a broad and unified framework that encompasses KernelSHAP and related estimators constructed using both with and without replacement sampling strategies. We then prove strong non-asymptotic theoretical guarantees that apply to all estimators from our framework. This provides, to the best of our knowledge, the first theoretical guarantees for KernelSHAP and sheds further light on tradeoffs between existing estimators. Through comprehensive benchmarking on small and medium dimensional datasets for Decision-Tree models, we validate our approach against exact Shapley values, consistently achieving low mean squared error with modest sample sizes. Furthermore, we make specific implementation improvements to enable scalability of our methods to high-dimensional datasets. Our methods, tested on datasets such MNIST and CIFAR10, provide consistently better results compared to the KernelSHAP library.



Paperid:3238
Authors:Yuxin Liu, Zhenghao (Mark) Peng, Xuanhao Cui, Bolei Zhou
Abstract:
Scenario-based testing is essential for validating the performance of autonomous driving (AD) systems. However, such testing is limited by the scarcity of long-tailed, safety-critical scenarios in existing datasets collected in the real world. To tackle the data issue, we propose the Adv-BMT framework, which augments real-world scenarios with diverse and realistic adversarial interactions. The core component of Adv-BMT is a bidirectional motion transformer (BMT) model to perform inverse traffic motion predictions, which takes the last frame of the scenario as input and reconstruct the traffic in the inverse of chronological order, till the initial time step. The Adv-BMT framework is a two-stage pipeline: it first conducts adversarial initializations and then inverse motion predictions. Different from previous work, we do not need any collision data for pretraining and are still able to generate realistic and diverse collision interactions. Our experimental results validate the quality of generated collision scenarios by Adv-BMT: training in our augmented dataset would reduce episode collision rates by 20\% compared to previous work. The code will be made available.



Paperid:3239
Authors:Yaorui Shi, Sihang Li, Chang Wu, ZHIYUAN LIU, Junfeng Fang, Hengxing Cai, An Zhang, Xiang Wang
Abstract:
Large language models have demonstrated impressive reasoning capabilities but are inherently limited by their knowledge reservoir.Retrieval-augmented reasoning mitigates this limitation by allowing LLMs to query external resources, but existing methods often retrieve irrelevant or noisy information, hindering accurate reasoning.In this paper, we proposeAutoRefine, a reinforcement learning post-training framework that adopts a new ``search-and-refine-during-think'' paradigm.AutoRefine introduces explicit knowledge refinement steps between successive search calls, enabling the model to iteratively filter, distill, and organize evidence before generating an answer.Furthermore, we incorporate tailored retrieval-specific rewards alongside answer correctness rewards using group relative policy optimization.Experiments on single-hop and multi-hop QA benchmarks demonstrate that AutoRefine significantly outperforms existing approaches, particularly in complex, multi-hop reasoning scenarios.Detailed analysis shows that AutoRefine issues frequent, higher-quality searches and synthesizes evidence effectively.Code is availablehere.



Paperid:3240
Authors:Max Möbus, Björn Braun, Christian Holz
Abstract:
Explaining black-box models for time series data is critical for the wide-scale adoption of deep learning techniques across domains such as healthcare. Recently, explainability methods for deep time series models have seen significant progress by adopting saliency methods that perturb masked segments of time series to uncover their importance towards the prediction of black-box models. Thus far, such methods have been largely restricted to regular time series. Irregular time series, however, sampled at irregular time intervals and potentially with missing values, are the dominant form of time series in various critical domains (e.g., hospital records). In this paper, we conduct the first evaluation of saliency methods for the interpretation of irregular time series models. We first translate techniques for regular time series into the continuous time realm of irregular time series and show under which circumstances such techniques are still applicable. However, existing perturbation techniques neglect the timing and structure of observed data, e.g., informative missingness when data is not missing at random. Thus, we propose Contimask, a simple framework to also apply non-differentiable perturbations, such as simulating that parts of the data had not been observed using NeuroEvolution. Doing so, we are first to successfully detect saliency that is independent of the value of observed data and achieve significant improvements at interpreting irregular time series models on real-world data where 90% of the data is missing.



Paperid:3241
Authors:Octave Mariotti, Zhipeng Du, Yash Bhalgat, Oisin Mac Aodha, Hakan Bilen
Abstract:
Semantic correspondence (SC) aims to establish semantically meaningful matches across different instances of an object category. We illustrate how recent supervised SC methods remain limited in their ability to generalize beyond sparsely annotated training keypoints, effectively acting as keypoint detectors. To address this, we propose a novel approach for learning dense correspondences by lifting 2D keypoints into a canonical 3D space using monocular depth estimation. Our method constructs a continuous canonical manifold that captures object geometry without requiring explicit 3D supervision or camera annotations. Additionally, we introduce SPair-U, an extension of SPair-71k with novel keypoint annotations, to better assess generalization. Experiments not only demonstrate that our model significantly outperforms supervised baselines on unseen keypoints, highlighting its effectiveness in learning robust correspondences, but that unsupervised baselines outperform supervised counterparts when generalized across different datasets.



Paperid:3242
Authors:Xingyu Chen, Shihao Ma, Runsheng Lin, Jiecong Lin, Bo Wang
Abstract:
Designing regulatory DNA sequences that achieve precise cell-type-specific gene expression is crucial for advancements in synthetic biology, gene therapy and precision medicine. Although transformer-based language models (LMs) can effectively capture patterns in regulatory DNA, their generative approaches often struggle to produce novel sequences with reliable cell-specific activity. Here, we introduce regCon, a novel constrained reinforcement learning (RL) framework tailored for designing regulatory DNA sequences with controllable cell-type specificity. By formulating regulatory sequence design as a biologically informed constrained optimization problem, we apply RL to autoregressive genomic LMs, enabling the models to iteratively refine sequences that maximize regulatory activity in targeted cell types while constraining off-target effects. Our evaluation on human promoters and enhancers demonstrates that regCon consistently outperforms existing generative and RL-based approaches, generating high-fitness regulatory sequences and achieving state-of-the-art cell-type specificity. Moreover, regCon-generated sequences capture key cell-type-specific transcription factor binding sites (TFBS), short DNA motifs recognized by regulatory proteins that control gene expression, demonstrating the biological plausibility of the generated sequences.



Paperid:3243
Authors:Guy Bar-Shalom, Fabrizio Frasca, Yaniv Galron, Yftah Ziser, Haggai Maron
Abstract:
Abstract:Detecting hallucinations in Large Language Model-generated text is crucial for their safe deployment. While probing classifiers show promise, they operate on isolated layer–token pairs and are LLM-specific, limiting their effectiveness and hindering cross-LLM applications. In this paper, we introduce a novel approach to address these shortcomings. We build on the natural sequential structure of activation data in both axes (layers $\times$ tokens) and advocate treating full activation tensors akin to images. We design ACT-ViT, a Vision Transformer-inspired model that can be effectively and efficiently applied to activation tensors and supports training on data from multiple LLMs simultaneously. Through comprehensive experiments encompassing diverse LLMs and datasets, we demonstrate that ACT-ViT consistently outperforms traditional probing techniques while remaining extremely efficient for deployment. In particular, we show that our architecture benefits substantially from multi-LLM training, achieves strong zero-shot performance on unseen datasets, and can be transferred effectively to new LLMs through fine-tuning.



Paperid:3244
Authors:Hossein Goli, Michael Gimelfarb, Nathan de Lara, Haruki Nishimura, Masha Itkina, Florian Shkurti
Abstract:
Off-policy evaluation (OPE) estimates the performance of a target policy using offline data collected from a behavior policy, and is crucial in domains such as robotics or healthcare where direct interaction with the environment is costly or unsafe. Existing OPE methods are ineffective for high-dimensional, long-horizon problems, due to exponential blow-ups in variance from importance weighting or compounding errors from learned dynamics models. To address these challenges, we propose STITCH-OPE, a model-based generative framework that leverages denoising diffusion for long-horizon OPE in high-dimensional state and action spaces. Starting with a diffusion model pre-trained on the behavior data, STITCH-OPE generates synthetic trajectories from the target policy by guiding the denoising process using the score function of the target policy. STITCH-OPE proposes two technical innovations that make it advantageous for OPE: (1) prevents over-regularization by subtracting the score of the behavior policy during guidance, and (2) generates long-horizon trajectories by stitching partial trajectories together end-to-end. We provide a theoretical guarantee that under mild assumptions, these modifications result in an exponential reduction in variance versus long-horizon trajectory diffusion. Experiments on the D4RL and OpenAI Gym benchmarks show substantial improvement in mean squared error, correlation, and regret metrics compared to state-of-the-art OPE methods.



Paperid:3245
Authors:Liliang Ren, Congcong Chen, Haoran Xu, Young Jin Kim, Adam Atkinson, Zheng Zhan, Jiankai Sun, Baolin Peng, Liyuan Liu, Shuohang Wang, Hao Cheng, Jianfeng Gao, Weizhu Chen, yelong shen
Abstract:
Recent advances in language modeling have demonstrated the effectiveness of Stat Space Models (SSMs) for efficient sequence modeling. While hybrid SSM architectures such as Samba and the decoder-decoder architecture, YOCO, have shown promising performance gains over Transformers, prior works have not investigated the efficiency potential of representation sharing between SSM layers. In this paper, we introduce the Gated Memory Unit (GMU), a simple yet effective mechanism for efficient memory sharing across layers, and apply it to create a decoder-hybrid-decoder architecture, SambaY, through integrating GMUs into the cross-decoder of YOCO. SambaY significantly enhances decoding efficiency, preserves linear pre-filling time complexity, and boosts long-context performance, all while eliminating the need for explicit positional encoding. Through extensive scaling experiments, we demonstrate that our architecture exhibits a significantly lower irreducible loss compared to a strong YOCO baseline, indicating superior performance scalability under large-scale compute regimes. Our largest model enhanced with Differential Attention, Phi4-mini-Flash-Reasoning, achieves comparable performance to Phi4-mini-Reasoning on reasoning tasks such as Math500, AIME24, and GPQA Diamond, while delivering up to 10× higher decoding throughput on 2K-length prompts with 32K generation length under the vLLM inference framework.



Paperid:3246
Authors:Kristiyan Sakalyan, Alessandro Palma, Filippo Guerranti, Fabian Theis, Stephan Günnemann
Abstract:
Understanding the evolution of cellular microenvironments is essential for deciphering tissue development and disease progression. While spatial transcriptomics now enables high-resolution mapping of tissue organization across space and time, current techniques that analyze cellular evolution operate at the single-cell level, overlooking critical spatial relationships. We introduce NicheFlow, a flow-based generative model that infers the temporal trajectory of cellular microenvironments across sequential spatial slides. By representing local cell neighborhoods as point clouds, NicheFlow jointly models the evolution of cell states and coordinates using optimal transport and Variational Flow Matching. Our approach successfully recovers both global spatial architecture and local microenvironment composition across diverse spatio-temporal datasets, from embryonic to brain development.



Paperid:3247
Authors:Francesco Cozzi, Marco Pangallo, Alan Perotti, André Panisson, Corrado Monti
Abstract:
Agent-Based Models (ABMs) are powerful tools for studying emergent properties in complex systems. In ABMs, agent behaviors are governed by local interactions and stochastic rules. However, these rules are ad hoc and, in general, non-differentiable, limiting the use of gradient-based methods for optimization, and thus integration with real-world data. We propose a novel framework to learn a differentiable surrogate of any ABM by observing its generated data. Our method combines diffusion models to capture behavioral stochasticity and graph neural networks to model agent interactions. Distinct from prior surrogate approaches, our method introduces a fundamental shift: rather than approximating system-level outputs, it models individual agent behavior directly, preserving the decentralized, bottom-up dynamics that define ABMs. We validate our approach on two ABMs (Schelling's segregation model and a Predator-Prey ecosystem) showing that it replicates individual-level patterns and accurately forecasts emergent dynamics beyond training. Our results demonstrate the potential of combining diffusion models and graph learning for data-driven ABM simulation.



Paperid:3248
Authors:Shixin Li, Xiaojing Ma, Zewei Li, Xiaofan Bai, Pingyi Hu, Dongmei Zhang, Bin Zhu
Abstract:
Adversarial attacks threaten the reliability of deep neural networks, particularly in black-box settings where transferability is essential. However, existing transfer-based attacks often fail when the target model’s architecture or training diverges from the surrogate, due to decision-boundary variation and representation drift. We introduce CORTA, a consensus-robust transfer attack that explicitly models these two sources of transfer failure as parameter and representation perturbations on the surrogate model. We formalize transferability as a distributionally robust optimization (DRO) problem over an uncertainty set of plausible targets, and provide efficient first-order approximations with theoretical guarantees. CORTA enforces consensus misclassification by jointly regularizing parameter sensitivity and promoting robustness to feature blending on the surrogate. Extensive experiments on ImageNet and CIFAR-100 show that CORTA consistently outperforms state-of-the-art transfer-based black-box attacks, including ensemble methods, across both convolutional and transformer architectures. For example, when transferring from ResNet-18 to Swin-B on CIFAR-100, CORTA achieves a 19.1\% higher transfer success rate than the strongest baseline. Our approach establishes a new benchmark for robust black-box adversarial evaluation.



Paperid:3249
Authors:HUANLEI GUO, Song LIU, Bingyi Jing
Abstract:
Molecular docking is a fundamental technique in structure-based drug discovery, essential for predicting the binding modes of protein-ligand complexes. Traditional docking methods, while reliable, suffer from slow computational speeds and limited adaptability. Recent deep learning (DL) approaches have significantly improved docking efficiency and accuracy. However, these methods struggle with generating conformations that are both chemically consistent and physically plausible due to limited incorporation of physical priors. To overcome these challenges, we propose ForceFM, a novel force-guided model that integrates a force-guided network into the generation process, steering ligand poses toward low-energy, chemically realistic conformations. Force guidance also halves inference cost compared with the unguided approaches. Importantly, replacing the guiding potential with diverse energy functions—including Vina, Glide, Gnina, and Confscore—preserves or improves performance, underscoring the method’s generality and robustness. These results highlight ForceFM's ability to set new standards in docking accuracy and physical consistency, surpassing the limitations of previous methods.



Paperid:3250
Authors:Zhenyu Pan, Yucheng Lu, Han Liu
Abstract:
We present MetaFind, a scene-aware multi-modal retrieval framework designed to enhance scene generation in the metaverse by retrieving 3D assets from large-scale repositories. MetaFind addresses two core challenges: (i) inconsistent asset retrieval that overlooks spatial, semantic, and stylistic constraints, and (ii) the absence of a standardized retrieval paradigm specifically tailored for 3D asset retrieval, as existing approaches predominantly rely on general-purpose 3D shape representation models. Our key innovation is a retrieval mechanism that enhances both spatial reasoning and style consistency by jointly modeling object-level features (including appearance) and scene-level layout structures. Methodologically, MetaFind introduces a plug-and-play layout encoder that captures both spatial relationships and object appearance features, ensuring retrieved 3D assets are contextually and stylistically coherent with the existing scene. The framework supports iterative scene construction by continuously adapting retrieval results to current scene updates. Empirical evaluations demonstrate the improved spatial and stylistic consistency of MetaFind in various retrieval tasks compared to baseline methods.



Paperid:3251
Authors:Xuechen Zhang, Zijian Huang, Yingcong Li, Chenshun Ni, Jiasi Chen, Samet Oymak
Abstract:
Abstract:Small language models (SLMs) struggle to learn complex reasoning behaviors, especially when high-quality traces are scarce or difficult to learn from. A typical approach for training such models combines a supervised fine-tuning (SFT) stage, often to distill reasoning capabilities from a larger model, followed by a reinforcement learning (RL) stage such as Group Relative Policy Optimization (GRPO). In this paper, we investigate the fundamental limitations of this SFT + RL paradigm and propose methods to overcome them. Using a toy student-expert model over Markov chains, we demonstrate that the SFT + RL strategy can fail completely when (1) the expert's traces are too difficult for the small model to express, or (2) the small model's initialization achieves exponentially sparse rewards as task complexity grows. To address these, we introduce BREAD, a GRPO variant that bridges SFT and RL via partial expert guidance and branch rollouts. When self-generated traces fail, BREAD adaptively inserts short expert prefixes/hints, allowing the small model to complete the rest of the reasoning path, and ensuring that each update includes at least one successful trace. This mechanism both densifies the reward signal and induces a natural learning curriculum. BREAD requires fewer than 40\% of ground-truth traces, consistently outperforming standard GRPO while speeding up the training by about 3$\times$. Importantly, we find that BREAD helps the model solve problems that are otherwise unsolvable by the SFT + RL strategy, highlighting how branch rollouts and expert guidance can aid SLM reasoning.



Paperid:3252
Authors:BaoLinh Tran, Van Vu
Abstract:
Abstract:The matrix recovery (completion) problem, a central problem in data science, involves recovering a matrix $A$ from a relatively small random set of entries. While such a task is generally impossible, it has been shown that one can recover $A$ exactly in polynomial time, with high probability, under three basic and necessary assumptions: (1) the rank of $A$ is very small compared to its dimensions (low rank), (2) $A$ has delocalized singular vectors (incoherence), and (3) the sample size is sufficiently large. Various algorithms address this task, including convex optimization by Candes, Recht, and Tao (2009, 2010), alternating projection by Hardt and Wooters (2014), and low-rank approximation with gradient descent by Keshavan, Montanari, and Oh (2009, 2010). In applications, Candes and Plan (2009) noted that it is more realistic to assume noisy observations. In such cases, the above approaches provide approximate recovery with small root mean square error, which is difficult to convert into exact recovery. Recently, results by Abbe et al. (2017) and Bhardwaj et al. (2023) on approximation in the infinity norm showed that one can recover $A$ even in the noisy case, provided $A$ has bounded precision. However, beyond the three basic assumptions, they either required that the condition number of $A$ be small (2017) or that the gaps between consecutive singular values be large (2023). These additional assumptions conflict, with one requiring singular values to be close together and the other suggesting they should be far apart. It is thus natural to conjecture that neither is necessary. In this paper, we demonstrate that this is indeed the case. We propose a simple algorithm for exact recovery of noisy data, relying solely on the three basic assumptions. The core step of the algorithm is a straightforward truncated singular value decomposition, which is highly efficient. To analyze the algorithm, we prove a new infinity norm version of the classical Davis-Kahan perturbation theorem, improving an earlier result in (2023). Our proof employs a combinatorial contour integration argument and is entirely distinct from all previous approaches.



Paperid:3253
Authors:Periklis Mantenoglou, RISHI HAZRA, Pedro Zuidberg Dos Martires, Luc De Raedt
Abstract:
Owing to their reasoning capabilities, large language models (LLMs) have been used to perform planning with natural language. However, LLMs have largely been tested on unconstrained planning domains. In order to deploy them in real-world settings where adherence to constraints, in particular safety constraints, is critical, we need to evaluate their performance on constrained planning tasks. We introduce LEXICON — a natural language-based (LEXI) constrained (CON) planning benchmark, consisting of a suite of environments, that can be used to evaluate in a principled fashion the planning capabilities of LLMs. The core idea behind LEXICON is to take existing planning environments and impose temporal constraints on the states. These constrained problems are then translated into natural language and given to an LLM to solve. A key feature of LEXICON is its extensibility. That is, the set of supported environments can be extended with new (unconstrained) environment generators, for which temporal constraints are constructed automatically. This renders LEXICON future-proof: the hardness of the generated planning problems can be increased as the planning capabilities of LLMs improve. Our experiments reveal that the performance of state-of-the-art LLMs, including reasoning models like o3 and R1, dramatically deteriorates as the degree of constrainedness of the planning tasks increases.



Paperid:3254
Authors:Chaeyun Jang, Deukhwan Cho, Hyungi Lee, Seanie Lee, Juho Lee
Abstract:
Recently, Large Language Models (LLMs) have been increasingly used to support various decision-making tasks, assisting humans in making informed decisions. However, when LLMs confidently provide incorrect information, it can lead humans to make suboptimal decisions. To prevent LLMs from generating incorrect information on topics they are unsure of and to improve the accuracy of generated content, prior works have proposed Retrieval Augmented Generation (RAG), where external documents are referenced to generate responses. However, previous RAG methods focus only on retrieving documents most relevant to the input query, without specifically aiming to ensure that the human user's decisions are well-calibrated. To address this limitation, we propose a novel retrieval method called Calibrated Retrieval-Augmented Generation (CalibRAG), which ensures that decisions informed by RAG are well-calibrated. Then we empirically validate that CalibRAG improves calibration performance as well as accuracy, compared to other baselines across various datasets.



Paperid:3255
Authors:Efthymia Tsamoura, Kaifu Wang, Dan Roth
Abstract:
Abstract:We study one of the most popular problems in **neurosymbolic learning** (NSL), that of learning neural classifiers given only the result of applying a symbolic component $\sigma$ to the gold labels of the elements of a vector $\mathbf x$. The gold labels of the elements in $\mathbf x$ are unknown to the learner. We make multiple contributions, theoretical and practical, to address a problem that has not been studied so far in this context, that of characterizing and mitigating *learning imbalances*, i.e., major differences in the errors that occur when classifying instances of different classes (aka **class-specific risks**). Our theoretical reveals a unique phenomenon: that $\sigma$ can greatly impact learning imbalances. This result sharply contrasts with previous research on supervised and weakly supervised learning, which only studies learning imbalances under data imbalances. On the practical side, we introduce a technique for estimating the marginal of the hidden gold labels using weakly supervised data. Then, we introduce algorithms that mitigate imbalances at training and testing time by treating the marginal of the hidden labels as a constraint. We demonstrate the effectiveness of our techniques using strong baselines from NSL and long-tailed learning, suggesting performance improvements of up to 14\%.



Paperid:3256
Authors:Akiyoshi Tomihari, Ryo Karakida
Abstract:
The theoretical understanding of self-attention (SA) has been steadily progressing. A prominent line of work studies a class of SA layers that admit an energy function decreased by state updates. While it provides valuable insights into inherent biases in signal propagation, it often relies on idealized assumptions or additional constraints not necessarily present in standard SA. Thus, to broaden our understanding, this work aims to relax these energy constraints and provide an energy-agnostic characterization of inference dynamics by dynamical systems analysis.In more detail, we first consider relaxing the symmetry and single-head constraints traditionally required in energy-based formulations. Next, to investigate more general SA architectures capable of oscillatory dynamics without necessarily admitting an energy function, we analyze the Jacobian matrix of the state. We reveal that normalization layers effectively normalize the Jacobian's complex eigenvalues, forcing the dynamics close to a critical state. This significantly enhances inference performance. Furthermore, we utilize the Jacobian perspective to develop regularization methods for training and a pseudo-energy for monitoring inference dynamics.



Paperid:3257
Authors:Daeun Kyung, Hyunseung Chung, Seongsu Bae, Jiho Kim, Jae Ho Sohn, Taerim Kim, Soo Kyung Kim, Edward Choi
Abstract:
Doctor-patient consultations require multi-turn, context-aware communication tailored to diverse patient personas. Training or evaluating doctor LLMs in such settings requires realistic patient interaction systems. However, existing simulators often fail to reflect the full range of personas seen in clinical practice. To address this, we introduce PatientSim, a patient simulator that generates realistic and diverse patient personas for clinical scenarios, grounded in medical expertise.PatientSim operates using: 1) clinical profiles, including symptoms and medical history, derived from real-world data in the MIMIC-ED and MIMIC-IV datasets, and 2) personas defined by four axes: personality, language proficiency, medical history recall level, and cognitive confusion level, resulting in 37 unique combinations.We evaluated eight LLMs for factual accuracy and persona consistency. The top-performing open-source model, Llama 3.3, was validated by four clinicians to confirm the robustness of our framework.As an open-source, customizable platform, PatientSim provides a reproducible and scalable solution that can be customized for specific training needs. Offering a privacy-compliant environment, it serves as a robust testbed for evaluating medical dialogue systems across diverse patient presentations and shows promise as an educational tool for healthcare.



Paperid:3258
Authors:Dongping Liao, Xitong Gao, Cheng-Zhong Xu
Abstract:
The increasing emphasis on privacy and data security has driven the adoption of federated learning (FL), a decentralized approach to train machine learning models without sharing raw data. Prompt learning (PL), which fine-tunes prompt embeddings of pretrained models, offers significant advantages in FL settings by reducing computational costs and communication overheads while leveraging the strong performance and generalization capabilities of vision-language models such as CLIP. To date, the reliable evaluation of federated prompt learning (FPL) algorithms for vision tasks remains an understudied area in current research. In this work, we introduce a comprehensive benchmark, named FLiP, to evaluate FPL algorithms. FLiP assesses the performance of 13 state-of-the-art centralized and FPL methods across 3 FL protocols and 12 open datasets, considering 6 distinct evaluation scenarios. Our findings demonstrate that PL maintains strong generalization performance in both in-distribution and out-of-distribution settings with minimal resource consumption. This work highlights the suitable application scenarios of each FPL algorithm in various environments characterized by data scarcity, unseen classes, and cross-domain distributional shifts. All benchmarks and code are publicly available to facilitate further research in this domain.



Paperid:3259
Authors:Mengxi Xiao, Mang Ye, Ben Liu, Xiaofen Zong, He Li, Jimin Huang, Qianqian Xie, Min Peng
Abstract:
The application of AI in psychiatric diagnosis faces significant challenges, including the subjective nature of mental health assessments, symptom overlap across disorders, and privacy constraints limiting data availability. To address these issues, we present MoodAngels, the first specialized multi-agent framework for mood disorder diagnosis. Our approach combines granular-scale analysis of clinical assessments with a structured verification process, enabling more accurate interpretation of complex psychiatric data. Complementing this framework, we introduce MoodSyn, an open-source dataset of 1,173 synthetic psychiatric cases that preserves clinical validity while ensuring patient privacy. Experimental results demonstrate that MoodAngels outperforms conventional methods, with our baseline agent achieving 12.3\% higher accuracy than GPT-4o on real-world cases, and our full multi-agent system delivering further improvements. Together, these contributions provide both an advanced diagnostic tool and a critical research resource for computational psychiatry, bridging important gaps in AI-assisted mental health assessment.



Paperid:3260
Authors:Haoqi Wu, Wei Dai, Ming Xu, Wang Li, Qiang Yan
Abstract:
Diffusion Models have gained significant popularity due to their remarkable capabilities in image generation, albeit at the cost of intensive computation requirement. Meanwhile, despite their widespread deployment in inference services such as Midjourney, concerns about the potential leakage of sensitive information in uploaded user prompts have arisen. Existing solutions either fail to strike an effective balance between utility and efficiency, or lack rigorous privacy guarantees. To bridge this gap, we propose ObCLIP, a plug-and-play safeguard that enables oblivious cloud-device hybrid generation scheme.By oblivious, each input prompt is transformed into a set of semantically similar candidate prompts that differ only in sensitive attributes (e.g., gender, ethnicity). The cloud server processes all candidate prompts without knowing which one is the real one, thus preventing any prompt leakage. To mitigate server cost, only a small portion of denoising steps is performed upon the large cloud model. The resulting intermediate latents are then transmitted back to the device, which selects the targeted latent and completes the remaining denoising using a small local model to obtain the final image. Additionally, we analyze and incorporate several cache-based accelerations that leverage temporal and batch redundancy, effectively reducing computation cost with minimal utility degradation. Extensive experiments across multiple datasets demonstrate that ObCLIP provides rigorous privacy and comparable utility to large cloud models with slightly increased server computation.



Paperid:3261
Authors:Zhiyang Xun, Shivam Gupta, Eric Price
Abstract:
Abstract:Given a noisy linear measurement $y = Ax + \xi$ of a distribution $p(x)$, and a good approximation to the prior $p(x)$, when can we sample from the posterior $p(x \mid y)$? Posterior sampling provides an accurate and fair framework for tasks such as inpainting, deblurring, and MRI reconstruction, and several heuristics attempt to approximate it. Unfortunately, approximate posterior sampling is computationally intractable in general.To sidestep this hardness, we focus on (local or global) log-concave distributions $p(x)$. In this regime, Langevin dynamics yields posterior samples when the exact scores of $p(x)$ are available, but it is brittle to score--estimation error, requiring an MGF bound (sub‑exponential error). By contrast, in the unconditional setting, diffusion models succeed with only an $L^2$ bound on the score error. We prove that combining diffusion models with an *annealed* variant of Langevin dynamics achieves conditional sampling in polynomial time using merely an $L^4$ bound on the score error.



Paperid:3262
Authors:Zhenyu Li, Kehai Chen, Yunfei Long, Xuefeng Bai, Yaoyin Zhang, Xuchen Wei, Juntao Li, Min Zhang
Abstract:
Large Language Models (LLMs) have demonstrated remarkable instruction-following capabilities across various applications. However, their performance in multilingual settings lacks systematic investigation, with existing evaluations lacking fine-grained constraint analysis across diverse linguistic contexts. We introduceXIFBench, a comprehensive constraint-based benchmark for evaluating multilingual instruction-following abilities of LLMs, comprising 558 instructions with 0-5 additional constraints across five categories (Content,Style,Situation,Format, andNumerical) in six languages spanning different resource levels. To support reliable and consistent cross-lingual evaluation, we implement three methodological innovations: cultural accessibility annotation, constraint-level translation validation, and requirement-based evaluation using English requirements as semantic anchors across languages. Extensive experiments with various LLMs not only quantify performance disparities across resource levels but also provide detailed insights into how language resources, constraint categories, instruction complexity, and cultural specificity influence multilingual instruction-following.



Paperid:3263
Authors:Yulin Li, Haokun GUI, Ziyang Fan, Junjie Wang, Bin Kang, BIN CHEN, Zhuotao Tian
Abstract:
Recent advances in Video Large Language Models (VLLMs) have achieved remarkable video understanding capabilities, yet face critical efficiency bottlenecks due to quadratic computational growth with lengthy visual token sequences of long videos. While existing keyframe sampling methods can improve temporal modeling efficiency, additional computational cost is introduced before feature encoding, and the binary frame selection paradigm is found suboptimal. Therefore, in this work, we proposeDynamicToken compression via LLM-guidedKeyframe prior (DyToK), a training-free paradigm that enables dynamic token compression by harnessing VLLMs' inherent attention mechanisms. Our analysis reveals that VLLM attention layers naturally encoding query-conditioned keyframe priors, by which DyToK dynamically adjusts per-frame token retention ratios, prioritizing semantically rich frames while suppressing redundancies. Extensive experiments demonstrate that DyToK achieves state-of-the-art efficiency-accuracy tradeoffs. DyToK shows plug-and-play compatibility with existing compression methods, such as VisionZip and FastV, attaining 2.5x faster inference while preserving accuracy across multiple VLLMs, such as LLaVA-OneVision and Qwen2.5-VL. Code and models will be made publicly available.



Paperid:3264
Authors:Chi-Pin Huang, Yueh-Hua Wu, Min-Hung Chen, Frank Wang, Fu-En Yang
Abstract:
Vision-language-action (VLA) reasoning tasks require agents to interpret multimodal instructions, perform long-horizon planning, and act adaptively in dynamic environments. Existing approaches typically train VLA models in an end-to-end fashion, directly mapping inputs to actions without explicit reasoning, which hinders their ability to plan over multiple steps or adapt to complex task variations. In this paper, we propose ThinkAct, a dual-system framework that bridges high-level reasoning with low-level action execution via reinforced visual latent planning. ThinkAct trains a multimodal LLM to generate embodied reasoning plans guided by reinforcing action-aligned visual rewards based on goal completion and trajectory consistency. These reasoning plans are compressed into a visual plan latent that conditions a downstream action model for robust action execution on target environments. Extensive experiments on embodied reasoning and robot manipulation benchmarks demonstrate that ThinkAct enables few-shot adaptation, long-horizon planning, and self-correction behaviors in complex embodied AI tasks.



Paperid:3265
Authors:Xinji Mai, Haotian Xu, Xing W, Weinong Wang, Yingying Zhang, Wenqiang Zhang
Abstract:
Large Language Models (LLMs) often struggle with mathematical reasoning tasks requiring precise, verifiable computation. While Reinforcement Learning (RL) from outcome-based rewards enhances text-based reasoning, understanding how agents autonomously learn to leverage external tools like code execution remains crucial. We investigate RL from outcome-based rewards for Tool-Integrated Reasoning, ZeroTIR, training base LLMs to spontaneously generate and execute Python code for mathematical problems without supervised tool-use examples. Our central contribution is we demonstrate that as RL training progresses, key metrics scale predictably. Specifically, we observe strong positive correlations where increased training steps lead to increases in the spontaneous code execution frequency, the average response length, and, critically, the final task accuracy. This suggests a quantifiable relationship between computational effort invested in training and the emergence of effective, tool-augmented reasoning strategies. We implement a robust framework featuring a decoupled code execution environment and validate our findings across standard RL algorithms and frameworks. Experiments show ZeroTIR significantly surpasses non-tool ZeroRL baselines on challenging math benchmarks. Our findings provide a foundational understanding of how autonomous tool use is acquired and scales within Agent RL, offering a reproducible benchmark for future studies. Code is released at \href{https://github.com/Anonymize-Author/AgentRL}{https://github.com/Anonymize-Author/AgentRL}.



Paperid:3266
Authors:Kaiqi Jiang, Jeremy Cohen, Yuanzhi Li
Abstract:
The study of Neural Tangent Kernels (NTKs) in deep learning has drawn increasing attention in recent years. NTKs typically actively change during training and are related to feature learning. In parallel, recent work on Gradient Descent (GD) has found a phenomenon called Edge of Stability (EoS), in which the largest eigenvalue of the NTK oscillates around a value inversely proportional to the step size. However, although follow-up works have explored the underlying mechanism of such eigenvalue behavior in depth, the understanding of the behavior of the NTKeigenvectorsduring EoS is still missing. This paper examines the dynamics of NTK eigenvectors during EoS in detail. Across different architectures, we observe that larger learning rates cause the leading eigenvectors of the final NTK, as well as the full NTK matrix, to have greater alignment with the training target. We then study the underlying mechanism of this phenomenon and provide a theoretical analysis for a two-layer linear network. Our study enhances the understanding of GD training dynamics in deep learning.



Paperid:3267
Authors:Rui Jiao, Hanlin Wu, Wenbing Huang, Yuxuan Song, Yawen Ouyang, Yu Rong, Tingyang Xu, Pengju Wang, Hao Zhou, Wei-Ying Ma, Jingjing Liu, Yang Liu
Abstract:
Metal-Organic Frameworks (MOFs) have attracted considerable attention due to their unique properties including high surface area and tunable porosity, and promising applications in catalysis, gas storage, and drug delivery. Structure prediction for MOFs is a challenging task, as these frameworks are intrinsically periodic and hierarchically organized, where the entire structure is assembled from building blocks like metal nodes and organic linkers. To address this, we introduce MOF-BFN, a novel generative model for MOF structure prediction based on Bayesian Flow Networks (BFNs). Given the local geometry of building blocks, MOF-BFN jointly predicts the lattice parameters, as well as the positions and orientations of all building blocks within the unit cell. In particular, the positions are modelled in the fractional coordinate system to naturally incorporate the periodicity. Meanwhile, the orientations are modeled as unit quaternions sampled from learned Bingham distributions via the proposed Bingham BFN, enabling effective orientation generation on the 4D unit hypersphere. Experimental results demonstrate that MOF-BFN achieves state-of-the-art performance across multiple tasks, including structure prediction, geometric property evaluation, and de novo generation, offering a promising tool for designing complex MOF materials.



Paperid:3268
Authors:Yu Huang, Zixin Wen, Aarti Singh, Yuejie Chi, Yuxin Chen
Abstract:
Abstract:The ability to reason lies at the core of artificial intelligence, and tough problems usually call for deeper, more extended reasoning to tackle. A crucial question about AI reasoning is whether models can extrapolate from learned reasoning patterns to solve harder tasks that require longer chain-of-thoughts (CoT).In this work, we present a theoretical analysis of transformers trained via gradient descent on synthetic data for various state tracking tasks, revealing how length-generalizable reasoning can emerge.Specifically, we prove that 1) for tasks with simple algebraic structure, e.g., composition in a cyclic group, transformers learn a solution pattern that generalizes to polynomial-length CoT; 2) for more complex tasks like composition in a symmetric group, a recursive self-training curriculum enables transformers to bootstrap longer reasoning and achieve polynomial generalization.Intuitively, our results demonstrate that transformers can learn sequential reasoning skills that scale with problem complexity. Moreover, we provide the first provable optimization-based guarantee for constant-depth transformers on the state tracking problems beyond circuit complexity class $\mathsf{TC}^0$: we prove that they can learn problems in $\mathsf{NC}^1$, which goes beyond prior results that consider problems within $\mathsf{TC}^0$, unless the famous conjecture $\mathsf{TC}^0 \neq \mathsf{NC}^1$ is false.



Paperid:3269
Authors:Naghmeh Ghanooni, Dennis Wagner, Waleed Mustafa, Sophie Burkhardt, Anthony Lin, Marius Kloft
Abstract:
Neural networks generally prefer simple and easy to learn features. When these features are spuriously correlated with the labels, the network's performance can suffer, particularly for underrepresented classes or concepts. Self-supervised representation learning methods, such as contrastive learning, are especially prone to this issue, often resulting in worse performance of downstream tasks. We empirically link this limitation to the spectrum of the feature matrix and provide theoretical support of this phenomenon.We propose a novel framework for learning high-rank feature representations, enabling the extraction of more meaningful and diverse features. Our approach achieves this without relying on ground-truth labels or any additional information. We demonstrate the applicability of the proposed framework using SimCLR, a prominent contrastive learning paradigm, and highlight its potential for broader use across representation learning methodologies.



Paperid:3270
Authors:Zhuomeng Zhang, Fangqi Li, Hanyi Wang, Shi-Lin Wang
Abstract:
One prerequisite for secure and reliable artificial intelligence services is tracing the copyright of backend deep neural networks. In the black-box scenario, the copyright of deep neural networks can be traced by their fingerprints, i.e., their outputs on a series of fingerprinting triggers. The performance of deep neural network fingerprints is usually evaluated in robustness, leaving the accuracy of copyright tracing among a large number of models with a lismited number of triggers intractable. This fact challenges the application of deep neural network fingerprints as the cost of queries is becoming a bottleneck. This paper studies the performance of deep neural network fingerprints from an information theoretical perspective. With this new perspective, we demonstrate that copyright tracing can be more accurate and efficient by using triggers with the largest marginal mutual information. Extensive experiments demonstrate that our method can be seamlessly incorporated into any existing fingerprinting scheme to facilitate the copyright tracing of deep neural networks.



Paperid:3271
Authors:UZAIR AKBAR, Niki Kilbertus, Hao Shen, Krikamol Muandet, Bo Dai
Abstract:
The technique of data augmentation (DA) is often used in machine learning for regularization purposes to better generalize under i.i.d. settings. In this work, we make a case for the use of DA beyond just the i.i.d. setting, but for generalization across interventions as well by presenting a unifying framework with topics in causal inference. Specifically, we argue that when the outcome generating mechanism is invariant to our choice of DA, then such augmentations can effectively be thought of as interventions on the treatment generating mechanism itself. This can potentially help to reduce the amount of bias in our estimation of causal effects arising from hidden confounders. In the presence of such unobserved confounding we typically make use of instrumental variables (IVs) -- sources of treatment randomization that are conditionally independent of the outcome. However, IVs may not be as readily available as DA for many applications, which is the main motivation behind this work. By appropriately regularizing IV based estimators, we introduce the concept of IV-like (IVL) regression for when treatment randomization sources may carry no information about the outcome and the possibility of its use for improving predictive performance across treatment interventions and reducing confounding bias. Finally, we cast parameterized DA as a IVL regression problem and show that when used in composition can simulate a worst-case application of such DA, further improving performance on causal estimation and generalization tasks beyond what simple DA may offer. This is shown both theoretically for the population case and via simulation experiments for the finite sample case using a simple linear example. We also present real data experiments to support our case.



Paperid:3272
Authors:Yilmazcan Ozyurt, Tunaberk Almaci, Stefan Feuerriegel, Mrinmaya Sachan
Abstract:
We introduce ExRec, a general framework for personalized exercise recommendation with semantically-grounded knowledge tracing. Our method builds on the observation that existing exercise recommendation approaches simulate student performance via knowledge tracing (KT) but they often overlook two key aspects: (a) the semantic content of questions and (b) the sequential, structured progression of student learning. To address this, our ExRec presents an end-to-end pipeline, from annotating the KCs of questions and learning their semantic representations to training KT models and optimizing several reinforcement learning (RL) methods. Moreover, we improve standard Q-learning-based continuous RL methods via a tailored model-based value estimation (MVE) approach that directly leverages the components of KT model in estimating cumulative knowledge improvement. We validate the effectiveness of our ExRec using various RL methods across four real-world tasks with different educational goals in online math learning. We further show that ExRec generalizes robustly to new, unseen questions and that it produces interpretable student learning trajectories. Together, our findings highlight the promise of KT-guided RL for effective personalization in education.



Paperid:3273
Authors:Subhojyoti Mukherjee, Viet Lai, Raghavendra Addanki, Ryan Rossi, Seunghyun Yoon, Trung Bui, Anup B. Rao, Jayakumar Subramanian, Branislav Kveton
Abstract:
Question answering (QA) agents automatically answer questions posed in natural language. In this work, we learn to ask clarifying questions in QA agents. The key idea in our method is to simulate conversations that contain clarifying questions and learn from them using reinforcement learning (RL). To make RL practical, we propose and analyze offline RL objectives that can be viewed as reward-weighted supervised fine-tuning (SFT) and easily optimized in large language models. Our work stands in a stark contrast to recently proposed methods, based on SFT and direct preference optimization, which have additional hyper-parameters and do not directly optimize rewards. We compare to these methods empirically and report gains in both optimized rewards and language quality.



Paperid:3274
Authors:Sang Keun Choe, Hwijeen Ahn, Juhan Bae, Kewen Zhao, Youngseog Chung, Adithya Pratapa, Willie Neiswanger, Emma Strubell, Teruko Mitamura, Jeff Schneider, Eduard Hovy, Roger Grosse, Eric Xing
Abstract:
Large language models (LLMs) are trained on a vast amount of human-written data, but data providers often remain uncredited. In response to this issue, data valuation (or data attribution), which quantifies the contribution or value of each data to the model output, has been discussed as a potential solution. Nevertheless, applying existing data valuation methods to recent LLMs and their vast training datasets has been largely limited by prohibitive compute and memory costs. In this work, we focus on influence functions, a popular gradient-based data valuation method, and significantly improve its scalability with an efficient gradient projection strategy called LoGra that leverages the gradient structure in backpropagation. We then provide a theoretical motivation of gradient projection approaches to influence functions to promote trust in the data valuation process. Lastly, we lower the barrier to implementing data valuation systems by introducing LogIX, a software package that can transform existing training code into data valuation code with minimal effort. In our data valuation experiments, LoGra achieves competitive accuracy against more expensive baselines while showing up to 6,500x improvement in throughput and 5x reduction in GPU memory usage when applied to Llama3-8B-Instruct and the 1B-token dataset.



Paperid:3275
Authors:Louis Kerner, Michel Meintz, Bihe Zhao, Franziska Boenisch, Adam Dziedzic
Abstract:
State-of-the-art text-to-image models like Infinity generate photorealistic images at an unprecedented speed. These models operate in a bitwise autoregressive manner over a discrete set of tokens that is practically infinite in size. However, their impressive generative power comes with a growing risk: as their outputs increasingly populate the Internet, they are likely to be scraped and reused as training data-potentially by the very same models. This phenomenon has been shown to lead to model collapse, where repeated training on generated content, especially from the models' own previous versions, causes a gradual degradation in performance. A promising mitigation strategy is watermarking, which embeds human-imperceptible yet detectable signals into generated images-enabling the identification of generated content. In this work, we introduce BitMark, a robust bitwise watermarking framework for Infinity. Our method embeds a watermark directly at the bit level of the token stream across multiple scales (also referred to as resolutions) during Infinity's image generation process. Our bitwise watermark subtly influences the bits to preserve visual fidelity and generation speed while remaining robust against a spectrum of removal techniques. Furthermore, it exhibits high radioactivity, i.e., when watermarked generated images are used to train another image generative model, this second model's outputs will also carry the watermark. The radioactive traces remain detectable even when only fine-tuning diffusion or image autoregressive models on images watermarked with our BitMark. Overall, our approach provides a principled step toward preventing model collapse in image generative models by enabling reliable detection of generated outputs.



Paperid:3276
Authors:Kornel Howil, Joanna Waczynska, Piotr Borycki, Tadeusz Dziarmaga, Marcin Mazur, Przemysław Spurek
Abstract:
Gaussian Splatting (GS) has recently emerged as an efficient representation for rendering 3D scenes from 2D images and has been extended to images, videos, and dynamic 4D content. However, applying style transfer to GS-based representations, especially beyond simple color changes, remains challenging. In this work, we introduce CLIPGaussians, the first unified style transfer framework that supports text- and image-guided stylization across multiple modalities: 2D images, videos, 3D objects, and 4D scenes. Our method operates directly on Gaussian primitives and integrates into existing GS pipelines as a plug-in module, without requiring large generative models or retraining from scratch. CLIPGaussians approach enables joint optimization of color and geometry in 3D and 4D settings, and achieves temporal coherence in videos, while preserving a model size. We demonstrate superior style fidelity and consistency across all tasks, validating CLIPGaussians as a universal and efficient solution for multimodal style transfer.



Paperid:3277
Authors:Wei Shen, Jiawei Zhang, Minhui Huang, Cong Shen
Abstract:
Abstract:We study bilevel optimization problems where the lower-level problems are strongly convex and have coupled linear constraints. To overcome the potential non-smoothness of the hyper-objective and the computational challenges associated with the Hessian matrix, we utilize penalty and augmented Lagrangian methods to reformulate the original problem as a single-level one. Especially, we establish a strong theoretical connection between the reformulated function and the original hyper-objective by characterizing the closeness of their values and derivatives. Based on this reformulation, we propose a single-loop, first-order algorithm for linearly constrained bilevel optimization (SFLCB). We provide rigorous analyses of its non-asymptotic convergence rates, showing an improvement over prior double-loop algorithms -- form $O(\epsilon^{-3}\log(\epsilon^{-1}))$ to $O(\epsilon^{-3})$. The experiments corroborate our theoretical findings and demonstrate the practical efficiency of the proposed SFLCB algorithm.



Paperid:3278
Authors:Zukang Xu, Xing Hu, Dawei Yang, Qiang Wu
Abstract:
Large language models (LLMs) have demonstrated remarkable performance across a wide range of natural language processing tasks. However, their exponentially increasing parameters pose significant challenges for deployment on resource-constrained devices. Vector Quantization (VQ) shows great promise for low-bit quantization (e.g., 2 to 4 bits), but existing work faces two key challenges: unconstrained direction error and suboptimal bit allocation. In this paper, we propose RSAVQ, a novel VQ framework to enhance extremely low-bit quantization for LLMs. RSAVQ introduces two geometry-driven innovations that effectively mitigate above limitations: (1) Error Direction Sensitivity Guidance (EDSG), which leverages the Fisher information matrix (FIM)-induced Riemannian metric to project quantization errors onto low-sensitivity directions in the parameter space. Specifically, this projection is performed along the negative natural gradient direction, which effectively suppresses error expansion. (2) Weight Channel Sensitivity Guidance (WCSG) , which constructs a channel-wise sensitivity metric via FIM curvature analysis to dynamically guide bit resource allocation. The approach facilitates a globally optimal quantization solution within prescribed bit constraints. Experiments demonstrate that RSAVQ outperforms existing methods for LLMs. For example, in 2-bit quantization of LLaMA-3 8B, RSAVQ leads baselines like VPTQ and QuIP# by 0.4 in perplexity (PPL) and 1.5 in zero-shot accuracy. This work offers a practical solution for constrained environments and a theoretical bridge between information geometry and the quantization of neural networks, advancing efficient deep learning.



Paperid:3279
Authors:Ling Yang, Ye Tian, Bowen Li, Xinchen Zhang, Ke Shen, Yunhai Tong, Mengdi Wang
Abstract:
While prior unified multimodal foundation models have explored architectural unification or objective alignment across modalities, most focus on pretraining paradigms, leaving post-training strategies—particularly in non-autoregressive settings—underexplored.We present MMaDA, a systematic investigation into the design space of unified diffusion-based multimodal foundation models. Our approach introduces three key innovations: First, a unified diffusion architecture with a shared probabilistic formulation and modality-agnostic design eliminates modality-specific components. Second, a mixed long chain-of-thought (CoT) finetuning strategy curates a unified CoT format across modalities, aligning reasoning processes between textual and visual domains to enable cold-start training for the final reinforcement learning (RL) stage. Third, we propose UniGRPO, a unified policy-gradient-based RL algorithm tailored for diffusion foundation models, which employs diversified reward modeling to unify post-training across reasoning and generation tasks.Experiments demonstrate that MMaDA achieves state-of-the-art performance across multiple domains—including language reasoning, multimodal understanding, and text-to-image generation—highlighting its strong generalization capability as a unified multimodal foundation model. Our work bridges the gap between pretraining and post-training in unified diffusion architectures, offering a holistic framework for future research.



Paperid:3280
Authors:Gavin McCracken, Gabriela Moisescu-Pareja, Vincent Létourneau, Doina Precup, Jonathan Love
Abstract:
Abstract:We propose a testable universality hypothesis, asserting that seemingly disparate neural network solutions observed in the simple task of modular addition actually reflect a common abstract algorithm. While prior work interpreted variations in neuron-level representations as evidence for distinct algorithms, we demonstrate---through multi-level analyses spanning neurons, neuron clusters, and entire networks---that multilayer perceptrons and transformers universally implement the abstract algorithm we call the approximate Chinese Remainder Theorem. Crucially, we introduce approximate cosets and show that neurons activate exclusively on them. Furthermore, our theory works for deep neural networks (DNNs). It predicts that universally learned solutions in DNNs with trainable embeddings or more than one hidden layer require only $\mathcal{O}(\log n)$ features, a result we empirically confirm. This work thus provides the first theory‑backed interpretation of \textit{multilayer} networks solving modular addition. It advances generalizable interpretability and opens a testable universality hypothesis for group multiplication beyond modular addition.



Paperid:3281
Authors:Jiazi Bu, Pengyang Ling, Yujie Zhou, Pan Zhang, Tong Wu, Xiaoyi Dong, Yuhang Zang, Yuhang Cao, Dahua Lin, Jiaqi Wang
Abstract:
Text-to-image (T2I) diffusion/flow models have drawn considerable attention recently due to their remarkable ability to deliver flexible visual creations. Still, high-resolution image synthesis presents formidable challenges due to the scarcity and complexity of high-resolution content. Recent approaches have investigated training-free strategies to enable high-resolution image synthesis with pre-trained models. However, these techniques often struggle with generating high-quality visuals and tend to exhibit artifacts or low-fidelity details, as they typically rely solely on the endpoint of the low-resolution sampling trajectory while neglecting intermediate states that are critical for preserving structure and synthesizing finer detail. To this end, we present HiFlow, a training-free and model-agnostic framework to unlock the resolution potential of pre-trained flow models. Specifically, HiFlow establishes a virtual reference flow within the high-resolution space that effectively captures the characteristics of low-resolution flow information, offering guidance for high-resolution generation through three key aspects: initialization alignment for low-frequency consistency, direction alignment for structure preservation, and acceleration alignment for detail fidelity. By leveraging such flow-aligned guidance, HiFlow substantially elevates the quality of high-resolution image synthesis of T2I models and demonstrates versatility across their personalized variants. Extensive experiments validate HiFlow's capability in achieving superior high-resolution image quality over state-of-the-art methods.



Paperid:3282
Authors:Benjamin Walker, Lingyi Yang, Nicola Muca Cirone, Cristopher Salvi, Terry Lyons
Abstract:
Abstract:Structured Linear Controlled Differential Equations (SLiCEs) provide a unifying framework for sequence models with structured, input-dependent state-transition matrices that retain the maximal expressivity of dense matrices whilst being cheaper to compute. The framework encompasses existing architectures, such as input-dependent block-diagonal linear recurrent neural networks and DeltaNet's diagonal-plus-low-rank structure, as well as two novel variants based on sparsity and the Walsh-Hadamard transform. We prove that, unlike the diagonal state-transition matrices of S4 and Mamba, SLiCEs employing block-diagonal, sparse, or Walsh-Hadamard matrices match the maximal expressivity of dense matrices. Empirically, SLiCEs solve the $A_5$ state-tracking benchmark with a single layer, achieve best-in-class length generalisation on regular language tasks among parallel models, and match the state-of-the-art performance of log neural controlled differential equations on six multivariate time-series classification datasets while cutting the average time per training step by a factor of twenty.



Paperid:3283
Authors:Runhan Shi, Letian Chen, Gufeng Yu, Yang Yang
Abstract:
Abstract:Chemical reaction prediction remains a fundamental challenge in organic chemistry, where existing machine learning models face two critical limitations: sensitivity to input permutations (molecule/atom orderings) and inadequate modeling of substructural interactions governing reactivity. These shortcomings lead to inconsistent predictions and poor generalization to real-world scenarios. To address these challenges, we propose ReaDISH, a novel reaction prediction model that learns permutation-invariant representations while incorporating interaction-aware features. It introduces two innovations: (1) symmetric difference shingle encoding, which computes molecular shingle differences to capture reaction-specific structural changes while eliminating order sensitivity; and (2) geometry-structure interaction attention, a mechanism that models intra- and inter-molecular interactions at the shingle level. Extensive experiments demonstrate that ReaDISH improves reaction prediction performance across diverse benchmarks. It shows enhanced robustness with an average improvement of 8.76\% on R$^2$ under permutation perturbations.



Paperid:3284
Authors:Ruyue Liu, Rong Yin, Xiangzhen Bo, Xiaoshuai Hao, Yong Liu, Jinwen Zhong, Can Ma, Weiping Wang
Abstract:
Large-scale pre-trained models have revolutionized Natural Language Processing (NLP) and Computer Vision (CV), showcasing remarkable cross-domain generalization abilities. However, in graph learning, models are typically trained on individual graph datasets, limiting their capacity to transfer knowledge across different graphs and tasks. This approach also heavily relies on large volumes of annotated data, which presents a significant challenge in resource-constrained settings. Unlike NLP and CV, graph-structured data presents unique challenges due to its inherent heterogeneity, including domain-specific feature spaces and structural diversity across various applications. To address these challenges, we propose a novel structure-aware self-supervised learning method for Text-Attributed Graphs (SSTAG). By leveraging text as a unified representation medium for graph learning, SSTAG bridges the gap between the semantic reasoning of Large Language Models (LLMs) and the structural modeling capabilities of Graph Neural Networks (GNNs). Our approach introduces a dual knowledge distillation framework that co-distills both LLMs and GNNs into structure-aware multilayer perceptrons (MLPs), enhancing the scalability of large-scale TAGs. Additionally, we introduce an in-memory mechanism that stores typical graph representations, aligning them with memory anchors in an in-memory repository to integrate invariant knowledge, thereby improving the model’s generalization ability. Extensive experiments demonstrate that SSTAG outperforms state-of-the-art models on cross-domain transfer learning tasks, achieves exceptional scalability, and reduces inference costs while maintaining competitive performance.



Paperid:3285
Authors:Hao Luo, Zihao Yue, Wanpeng Zhang, Yicheng Feng, Sipeng Zheng, Deheng Ye, Zongqing Lu
Abstract:
Recent advances in large multimodal models (LMMs) have significantly advanced video comprehension, yet their performance remains limited in first-person scenarios. The interactive nature of egocentric videos is critical for applications like embodied intelligence, but introduces complex visual contexts that conventional models struggle to capture. To bridge this gap, we introduce OpenMMEgo withinnovations across three dimensions: data, model, and training strategy. To provide rich spatiotemporal visual knowledge, we curate a large-scale, high-quality dataset named OME10M, comprising over 8.2M egocentric video QA pairs synthesized from Ego4D series. We also establish OMEBench, a comprehensive benchmark for rigorous egocentric understanding assessment. To alleviate the frequent viewpoint shifts inherent in egocentric videos, we implement semantic-aware visual token compression. Further, a curriculum learning strategy is complemented to foster stable learning across various data complexities. OpenMMEgo consistently improves the performance of LMMs on egocentric benchmarks without sacrificing general video understanding performance. Notably, Qwen2.5-VL tuned with OpenMMEgo substantially outperforms other models of the same size in egocentric video understanding. All components will be open-sourced.



Paperid:3286
Authors:areeb ahmad, Abhinav Joshi, Ashutosh Modi
Abstract:
In the quest for interpretability of LLMs, circuit discovery has emerged as a powerful framework. It identifies a computational subgraph of models that can replicate a model's behavior on a certain task. Existing methods operate at the standard component level granularity (e.g., attention heads, MLPs), potentially overlooking more fine-grained computational structure. We propose a framework that treats singular vector pairs of the augmented query–key, value–output, and MLP projection matrices as the atomic units of inspection of the model's behavior. By applying singular value decomposition (SVD) to the weight matrices of attention and MLP layers, we uncover orthogonal functional directions within each component that independently contribute to task behavior. These directions define a compositional basis over which distinct computations can occur in parallel, even within a single attention head. Our framework can replicate GPT-2 behavior faithfully on tasks like indirect object identification. We also show that some of these directions act as a control knob for concepts which corresponds to meaningful subtasks within tasks. Our approach establishes a more precise foundation for automated interpretability that better aligns with the underlying low-rank structure of the transformer weights.



Paperid:3287
Authors:Zhengrui Ma, Yang Feng, Chenze Shao, Fandong Meng, Jie Zhou, Min Zhang
Abstract:
We introduce SLED, an alternative approach to speech language modeling by encoding speech waveforms into continuous latent representation sequences and modeling them autoregressively using an energy distance objective in this continuous space. By bypassing the reliance on residual vector quantization, it avoids discretization errors and eliminates the need for the complicated hierarchical architectures common in existing speech language models. It simplifies the overall modeling pipeline while preserving the richness of speech information and maintaining inference efficiency. Empirical results demonstrate that SLED achieves strong performance in both zero-shot and streaming speech synthesis, showing its potential for broader applications in general-purpose speech language modeling.



Paperid:3288
Authors:Zi Yang, Lei Qiu, FANG LYU, Ming Zhong, Zhilei Chai, Haojie Zhou, Huimin Cui, Xiaobing Feng
Abstract:
Compiler optimization is essential for improving program performance, yet modern compilers still depend on manually crafted transformation rules over intermediate representations (IRs). As compilers grow in complexity, maintaining these rule-based optimizations becomes increasingly labor-intensive and difficult to scale. Recent advances in large language models (LLMs) offer a promising alternative, but their effectiveness in compiler optimization remains limited—primarily due to the lack of IR-oriented datasets that expose models to diverse transformation samples in real-world scenarios (optimization-sensitive samples), hindering LLMs from learning rich and generalizable optimization strategies.In this paper, we introduce IR-OptSet, the first public optimization-sensitive dataset for advancing LLM-based IR optimizers. It comprises 170K LLVM IR samples from open-source repositories across 8 representative optimization domains. IR-OptSet defines two core tasks: Code Analysis and Optimized Code Generation, and provides tools for correctness verification, performance evaluation, and dataset expansion. In our experiments, fine-tuning three representative LLMs on IR-OptSet leads to significant accuracy improvements across both tasks. Moreover, the LLM fine-tuned with IR-OptSetoutperforms traditional compiler with the -O3 optionin 64 test cases in terms of performance. Further analysis reveals that IR-OptSet provides greater transformation diversity and representativeness than three widely used IR-oriented datasets, highlighting its potential to drive model-based IR optimization. IR-OptSet is publicly available athttps://huggingface.co/datasets/YangziResearch/IR-OptSet.



Paperid:3289
Authors:Seok-Ju Hahn, Junghye Lee
Abstract:
Synthetic data generation is an appealing solution for a data-centric and privacy-preserving federated learning system. However, it is also appalling to adopt in practice, due to the high computational and communication costs of generative models, such as diffusion models. To reconcile these challenges, we define the task of collaborative synthetic data generation as a cooperative sampling procedure from the mixture of decentralized distributions. We propose to implement this by leveraging the composition-friendly nature of diffusion models, building upon a theoretically grounded step-size condition of the unadjusted Langevin algorithm that ensures both compatibility with diffusion sampling and convergence to the target density. As a result, we can directly obtain refined synthetic data across clients, without exchanging model parameters. We empirically demonstrate that our framework reduces communication overhead while significantly improving generation quality, even with differential privacy guarantees.



Paperid:3290
Authors:Ziyu Xiong, Yichi Zhang, Foyez Alauddin, Chu Xin Cheng, Joon An, Mohammad Seyedsayamdost, Ellen Zhong
Abstract:
Nuclear Magnetic Resonance (NMR) spectroscopy is a cornerstone technique for determining the structures of small molecules and is especially critical in the discovery of novel natural products and clinical therapeutics. Yet, interpreting NMR spectra remains a time-consuming, manual process requiring extensive domain expertise. We introduce NMRDiff3D, an end-to-end framework that directly predicts the structure of a given molecule from its 1D NMR spectra and chemical formula. We frame structure elucidation as conditional generation from an atomic diffusion model built on a scalable non-equivariant transformer architecture. To scale to the complex chemical groups found in natural products, we curate SpectraNP, a dataset of synthetic 1D NMR spectra for 111,193 natural products. NMRDiff3D accomplishes the challenging task of correctly predicting the structures of complex natural products with an accuracy of >70%. This work takes a significant step toward solving the grand challenge of automating small-molecule structure elucidation and highlights the potential of deep learning in accelerating molecular discovery.



Paperid:3291
Authors:Yize Wu, KE GAO, Ling Li, Yanjun Wu
Abstract:
Speculative decoding is an effective and lossless method for Large Language Model (LLM) inference acceleration. It employs a smaller model to generate a draft token sequence, which is then verified by the original base model. In multi-GPU systems, inference latency can be further reduced through tensor parallelism (TP), while the optimal TP size of the draft model is typically smaller, leading to GPU idling during the drafting stage. We observe that such inefficiency stems from the conventional sequential execution of layers, which is seemingly natural but actually unnecessary. Therefore, we propose EasySpec, a layer-parallel speculation strategy that optimizes the efficiency of multi-GPU utilization. EasySpec breaks the inter-layer data dependency in the draft model, enabling multiple layers to run simultaneously across multiple devices as ``fuzzy'' speculation. After each drafting-and-verification iteration, the draft model’s key-value cache is calibrated in a single forward pass, preventing long-term fuzzy-error accumulation at minimal additional latency. EasySpec is a training-free and plug-in method. We evaluated EasySpec on several mainstream open-source LLMs, using smaller versions of models from the same series as drafters. The results demonstrate that EasySpec can achieve a peak speedup of 4.17x compared to vanilla decoding, while preserving the original distributions of the base LLMs. Specifically, the drafting stage can be accelerated by up to 1.62x with a maximum accuracy drop of only 7\%.



Paperid:3292
Authors:Bailey Trang, Parham Saremi, Alan Wang, Fangrui Huang, Zahra TehraniNasab, Amar Kumar, Tal Arbel, Fei-Fei Li, Ehsan Adeli
Abstract:
Capturing diversity is crucial in conditional and prompt-based image generation, particularly when conditions contain uncertainty that can lead to multiple plausible outputs. To generate diverse images reflecting this diversity, traditional methods often modify random seeds, making it difficult to discern meaningful differences between samples, or diversify the input prompt, which is limited in verbally interpretable diversity. We propose \modelnamenospace, a novel conditional image generation framework, applicable to any pretrained conditional generative model, that addresses inherent condition/prompt uncertainty and generates diverse plausible images. \modelname is based on a simple yet effective idea: decomposing the input condition into diverse latent representations, each capturing an aspect of the uncertainty and generating a distinct image. First, we integrate a latent graph, parameterized by Generative Flow Networks (GFlowNets), into the prompt representation computation. Second, leveraging GFlowNets' advanced graph sampling capabilities to capture uncertainty and output diverse trajectories over the graph, we produce multiple trajectories that collectively represent the input condition, leading to diverse condition representations and corresponding output images. Evaluations on natural image and medical image datasets demonstrate \modelnamenospace’s improvement in both diversity and fidelity across image synthesis, image generation, and counterfactual generation tasks.



Paperid:3293
Authors:Marzieh Nilipour, Mohammad Abam, Davoud Kareshki, Masoud Seddighin, MohammadHossein Paydar
Abstract:
Abstract:We investigate metric distortion in distributed voting, where voters are divided into groups that each select a local representative, and a final winner is chosen from these representatives. This model captures real-world systems such as the U.S. presidential elections, where state-level decisions determine the national outcome. Our analysis focuses on four fundamental cost-minimization objectives introduced by Anshelevich et al.~\cite{anshelevich2022distortion}: the average of averages ($\avgavg$), average of maxima ($\avgmax$), maximum of averages ($\maxavg$), and maximum of maxima ($\maxmax$). We provide improved bounds for both deterministic and randomized rules which provides a near-complete picture of metric distortion in distributed voting. For deterministic mechanisms, we tighten the distortion gaps: we establish a tight lower bound of 5 for $\maxavg$ (improving the previous $2+\sqrt{5}$), reduce the upper bound for $\avgmax$ from 11 to 9, and show a tight upper bound of 3 for $\maxmax$ (improving the previous bound of 5). For randomized rules, we study two classes: (i) $\randdet$, where only the second stage is randomized, and (ii) $\randrand$, where both stages may be randomized. For $\randdet$, we establish tight bounds for all objectives: $5-2/k$ for $\avgavg$, $3$ for $\avgmax$ and $\maxmax$, $5$ for $\maxavg$. For $\randrand$, we show tight bounds of $3$ for $\maxavg$ and $\maxmax$, along with nearly tight bounds for $\avgavg$ and $\avgmax$ in the ranges $[3-2/n, 3-2/(kn^*)]$ and $[3-2/n, 3]$, respectively, where $n$ is the number of voters, $n^*$ is the size of the largest group, and k is the number of groups.



Paperid:3294
Authors:Yue Gong, Chuan Lei, Xiao Qin, Kapil Vaidya, Balakrishnan Narayanaswamy, Tim Kraska
Abstract:
Text-to-SQL systems translate natural language (NL) questions into SQL queries, enabling non-technical users to interact with structured data. While large language models (LLMs) have shown promising results on the text-to-SQL task, they often produce semantically incorrect yet syntactically valid queries, with limited insight into their reliability. We propose SQLens, an end-to-end framework for fine-grained detection and correction of semantic errors in LLM-generated SQL. SQLens integrates error signals from both the underlying database and the LLM to identify potential semantic errors within SQL clauses. It further leverages these signals to guide query correction. Empirical results on two public benchmarks show that SQLens outperforms the best LLM-based self-evaluation method by 25.78% in F1 for error detection, and improves execution accuracy of out-of-the-box text-to-SQL systems by up to 20%. Our code with full documentation is available at https://anonymous.4open.science/r/sqlens-7DE6/.



Paperid:3295
Authors:Chaohao Yuan, Maoji Wen, Ercan KURUOGLU, Yang Liu, Jia Li, Tingyang Xu, Deli Zhao, Hong Cheng, Yu Rong
Abstract:
To enhance the generalization ability of graph neural networks (GNNs) in learning and simulation physical dynamics, a series of equivariant GNNs have been developed to incorporate the symmetric inductive bias. However, the existing methods do not take into account the non-stationarity nature of physical dynamics, where the joint distribution changes over time. Moreover, previous approaches for modeling non-stationary time series typically involve normalizing the data, which disrupts the symmetric assumption inherent in physical dynamics. To model the non-stationary physical dynamics while preserving the symmetric inductive bias, we introduce a Non-Stationary Equivariant Graph Neural Network (NS-EGNN) to capture the non-stationarity in physical dynamics while preserving the symmetric property of the model. Specifically, NS-EGNN employs Fourier Transform on segments of physical dynamics to extract time-varying frequency information from the trajectories. It then uses the first and second-order differences to mitigate non-stationarity, followed by pooling for future predictions. Through capturing varying frequency characteristics and alleviate the linear and quadric trend in the raw physical dynamics, NS-EGNN better models the temporal dependencies in the physical dynamics. NS-EGNN has been applied on various types of physical dynamics, including molecular, motion and protein dynamics. In various scenario, NS-EGNN consistently surpasses the performance of existing state-of-the-art algorithms, underscoring its effectiveness.



Paperid:3296
Authors:Mirela Ostrek, Michael Black, Justus Thies
Abstract:
Synthesizing high-quality 3D head textures is crucial for gaming, virtual reality, and digital humans. Achieving seamless 360° textures typically requires expensive multi-view datasets with precise tracking. However, traditional methods struggle without back-view data or precise geometry, especially for human heads, where even minor inconsistencies disrupt realism. We introduce HairFree, an unsupervised texturing framework guided by textual descriptions and 2D diffusion priors, producing high-consistency 360° bald head textures—including non-human skin with fine details—without any texture, back-view, bald, non-human, or synthetic training data. We fine-tune a diffusion prior on a dataset of mostly frontal faces, conditioned on predicted 3D head geometry and face parsing. During inference, HairFree uses precise skin masks and 3D FLAME geometry as input conditioning, ensuring high 3D consistency and alignment. We synthesize the full 360° texture by first generating a frontal RGB image aligned to the 3D FLAME pose and mapping it to UV space. As the virtual camera moves, we inpaint and merge missing regions. A built-in semantic prior enables precise region separation—particularly for isolating and removing hair—allowing seamless integration with various assets like customizable 3D hair, eyeglasses, jewelry, etc. We evaluate HairFree quantitatively and qualitatively, demonstrating its superiority over state-of-the-art 3D head avatar generation methods. Code \& model will be released for reproducibility.



Paperid:3297
Authors:Ali Zafari, Xi Chen, Shirin Jalali
Abstract:
Zero-shot denoising aims to denoise observations without access to training samples or clean reference images. This setting is particularly relevant in practical imaging scenarios involving specialized domains such as medical imaging or biology. In this work, we propose the Zero-Shot Neural Compression Denoiser (ZS-NCD), a novel denoising framework based on neural compression. ZS-NCD treats a neural compression network as an untrained model, optimized directly on patches extracted from a single noisy image. The final reconstruction is then obtained by aggregating the outputs of the trained model over overlapping patches. Thanks to the built-in entropy constraints of compression architectures, our method naturally avoids overfitting and does not require manual regularization or early stopping. Through extensive experiments, we show that ZS-NCD achieves state-of-the-art performance among zero-shot denoisers for both Gaussian and Poisson noise, and generalizes well to both natural and non-natural images. Additionally, we provide new finite-sample theoretical results that characterize upper bounds on the achievable reconstruction error of general maximum-likelihood compression-based denoisers. These results further establish the theoretical foundations of compression-based denoising.



Paperid:3298
Authors:Mingzhuo Lin, Jianping Luo
Abstract:
Neural Architecture Search (NAS) aims to identify high-performance networks within a defined search space. Training-free metrics have been proposed to estimate network performance without actual training, reducing NAS deployment costs. However, individual training-free metrics often capture only partial architectural features, and their estimation capabilities are different in various tasks. Combining multiple training-free metrics has been explored to enhance scalability across tasks. Yet, these methods typically optimize global metric combinations over the entire search space, overlooking the varying sensitivities of different architectures to specific metrics, which may limit the final architectures' performance. To address these challenges, we propose the Per-Architecture Training-Free Metric Optimization NAS (PO-NAS) algorithm. This algorithm: (a) Integrates multiple training-free metrics as auxiliary scores, dynamically optimizing their combinations using limited real-time training data, without relying on benchmarks; (b) Individually optimizes metric combinations for each architecture; (c) Integrates an evolutionary algorithm that leverages efficient predictions from surrogate models, enhancing search efficiency in large search spaces. Notably, PO-NAS combines the efficiency of training-free search with the robust performance of training-based evaluations. Extensive experiments demonstrate the effectiveness of our approach. Our code has been made publicly available at https://anonymous.4open.science/r/PO-NAS-2953.



Paperid:3299
Authors:Amirmohammad Izadi, Mohammadali Banayeeanzade, Fatemeh Askari, Ali Rahimiakbar, Mohammad Vahedi, Hosein Hasani, Mahdieh Soleymani
Abstract:
Despite progress in Vision-Language Models (VLMs), their capacity for visual reasoning is often limited by the \textit{binding problem}: the failure to reliably associate perceptual features with their correct visual referents. This limitation underlies persistent errors in tasks such as counting, visual search, scene description, and spatial relationship understanding. A key factor is that current VLMs process visual features largely in parallel, lacking mechanisms for spatially grounded, serial attention.This paper introduces a simple yet effective intervention: augmenting visual inputs with low-level spatial structures (e.g., horizontal lines) and pairing this with a textual prompt that encourages sequential, spatially-aware parsing. We empirically demonstrate substantial performance improvements across core visual reasoning tasks. Specifically, our method improves GPT-4o visual search accuracy by 25.00\%, increases counting accuracy by 26.83\%, reduces edit distance error in scene description by 0.32, and enhances performance on spatial relationship tasks by 9.50\% on a a 2D synthetic dataset.Furthermore, we find that the visual modification is essential for these gains; purely textual strategies, including Chain-of-Thought prompting, are insufficient and can even degrade performance. Our method enhances binding only with a single-query inference, underscoring the importance of visual input design over purely linguistically-based approaches. These findings suggest that low-level visual structuring is a powerful and underexplored direction for improving compositional visual reasoning and could serve as a general strategy for enhancing VLM performance on spatially grounded tasks.



Paperid:3300
Authors:Zemin Huang, Zhiyang Chen, Zijun Wang, Tiancheng Li, Guo-Jun Qi
Abstract:
We introduce the Diffusion Chain of Lateral Thought (DCoLT), a reasoning framework for diffusion language models. DCoLT treats each intermediate step in the reverse diffusion process as a latent "thinking" action and optimizes the entire reasoning trajectory to maximize the reward on the correctness of the final answer with outcome-based Reinforcement Learning (RL).Unlike traditional Chain-of-Thought (CoT) methods that follow a causal, linear thinking process, DCoLT allows bidirectional, non-linear reasoning with no strict rule on grammatical correctness amid its intermediate steps of thought.We implement DCoLT on two representative Diffusion Language Models (DLMs). First, we choose SEDD as a representative continuous-time discrete diffusion model, where its concrete score derives a probabilistic policy to maximize the RL reward over the entire sequence of intermediate diffusion steps. We further consider the discrete-time masked diffusion language model -- LLaDA, and find that the order to predict and unmask tokens plays an essential role to optimize its RL action resulting from the ranking-based Unmasking Policy Module (UPM) defined by the Plackett-Luce model.Experiments on both math and code generation tasks show that using only public data and 16 H800 GPUs, DCoLT-reinforced DLMs outperform other DLMs trained by SFT or RL or even both. Notably, DCoLT-reinforced LLaDA boosts its reasoning accuracy by +9.8%, +5.7%, +11.4%, +19.5% on GSM8K, MATH, MBPP, and HumanEval.



Paperid:3301
Authors:Jun Yang, Cheng-Chi Wang, Bogdan "Bo" Stoica, Kexin Pei
Abstract:
Large Language Models (LLMs) have been increasingly used to optimize code efficiency. Evaluating their effectiveness and further suggesting optimization opportunities often rely on high-quality tests to demonstrate the performance bottlenecks presented in the program. However, existing approaches rely on a limited set of hand-curated inputs or LLM-generated uninteresting length-stressing tests, failing to reveal more nuanced optimization opportunities. We present WEDGE, a framework for generating performance-stressing input given the program under test. WEDGE synthesizes explicit performance-characterizing constraints in the form of branch conditions to partition the programs' execution space into performance-specific regions. When integrated with the coverage-guided fuzzer, reaching different regions introduces explicit rewards for test generation to explore1 inefficient implementations. Our evaluation shows that WEDGE introduces a significant slowdown compared to the tests in CodeContests and those claimed to be optimized by existing approaches. From the utility perspective, integrating our tests substantially improves the existing code optimization approaches that1 rely on test-driven execution feedback. We release PERFFORGE, the performance1 tests generated by WEDGE, to benchmark future approaches for efficient code1 generation at https://github.com/elmerjfudd/wedge.



Paperid:3302
Authors:Wei Wu, Xuan Tan, Yan Peng, Ling Chen, FangFang Li, Chuan Luo
Abstract:
Abstract:Signed networks can reflect more complex connections through positive and negative edges, and cost-effective signed network sketching can significantly benefit an important link sign prediction task in the era of big data. Existing signed network embedding algorithms mainly learn node representation in the Graph Neural Network (GNN) framework with the balance theory. However, the node-wise representation learning methods either limit the representational power because they primarily rely on node pairwise relationship in the network, or suffer from severe efficiency issues. Recent research has explored simplicial complexes to capture higher-order interactions and integrated them into GNN frameworks. Motivated by that, we propose EdgeSketch+, a simple and effective edge embedding algorithm beyond traditional node-centric modeling that directly represents edges as low-dimensional vectors without transitioning from node embeddings. The proposed approach maintains a good balance between accuracy and efficiency by exploiting the Locality Sensitive Hashing (LSH) technique to swiftly capture the higher-order information derived from the simplicial complex in a manner of no learning processes. Experiments show that EdgeSketch+ matches state-of-the-art accuracy while significantly reducing runtime, achieving speedups of up to $546.07\times$ compared to GNN-based methods.



Paperid:3303
Authors:Samuel Lavoie, Michael Noukhovitch, Aaron Courville
Abstract:
We argue that diffusion models' success in modeling complex distributions is, for the most part, coming from their conditioning. This paper investigates the representation used to condition diffusion models from the perspective that ideal representations should improve modeling the data distribution, be easy to generate, and be compositional to allow generalizing outside the training distribution. We introduce Discrete Latent Code (DLC), an image representation derived from Simplicial Embeddings trained with a self-supervised learning objective. DLCs are sequences of discrete tokens, as opposed to the standard continuous image embeddings. They are easy to generate and their compositionality enables sampling of novel images beyond the training distribution.Diffusion models trained with DLCsimprove generation fidelity, establishing a new state-of-the-art for unconditional image generation on ImageNet. Additionally, we show that composing DLCs allows the image generator to produce interesting out-of-distribution samples that coherently combine the semantics of images in diverse ways.Finally, we showcase how DLCs can enable text-to-image generation by leveraging large-scale pretrained language models. Using only 9M image-caption pairs, we efficiently finetune a text diffusion model to generate novel DLCs that produces samples outside of the data distribution used to train the image generator.



Paperid:3304
Authors:Daksh Mittal, Ang Li, Tzu-Ching Yen, C. Guetta, Hongseok Namkoong
Abstract:
Autoregressive models have emerged as a powerful framework for modeling exchangeable sequences---i.i.d. observations when conditioned on some latent factor---enabling direct modeling of uncertainty from missing data (rather than a latent). Motivated by the critical role posterior inference plays as a subroutine in decision-making (e.g., active learning, bandits), we study the inferential and architectural inductive biases that are most effective for exchangeable sequence modeling. For the inference stage, we highlight a fundamental limitation of the prevalent single-step generation approach: its inability to distinguish between epistemic and aleatoric uncertainty. Instead, a long line of works in Bayesian statistics advocates for multi-step autoregressive generation; we demonstrate this "correct approach" enables superior uncertainty quantification that translates into better performance on downstream decision-making tasks. This naturally leads to the next question: which architectures are best suited for multi-step inference? We identify a subtle yet important gap between recently proposed Transformer architectures for exchangeable sequences (Müller et al., 2022; Nguyen & Grover, 2022; Ye & Namkoong, 2024), and prove that they in fact cannot guarantee exchangeability despite introducing significant computational overhead. Through empirical evaluation, we find that these custom architectures can significantly underperform compared to standard causal masking, highlighting the need for new architectural innovations in Transformer-based modeling of exchangeable sequences.



Paperid:3305
Authors:Erich Liang, Roma Bhattacharjee, Sreemanti Dey, Rafael Moschopoulos, Caitlin Wang, Michel Liao, Grace Tan, Andrew Wang, Karhan Kayan, Stamatis Alexandropoulos, Jia Deng
Abstract:
Accurately tracking camera intrinsics is crucial for achieving 3D understanding from 2D video. However, most 3D algorithms assume that camera intrinsics stay constant throughout a video, which is often not true for many real-world in-the-wild videos. A major obstacle in this field is a lack of dynamic camera intrinsics benchmarks--existing benchmarks typically offer limited diversity in scene content and intrinsics variation, and none provide per-frame intrinsic changes for consecutive video frames. In this paper, we present ChangeIn, a real-world benchmark that provides per-frame ground truth intrinsics annotations for videos with dynamic intrinsics. Compared to prior benchmarks, ChangeIn captures a wider range of intrinsic variations and scene diversity, featuring 143K+ annotated frames from 386 high-resolution indoor and outdoor videos with dynamic camera intrinsics. To ensure accurate per-frame intrinsics, we build a comprehensive look-up table of calibration experiments and extend the Kalibr toolbox to improve its accuracy and robustness. Using our benchmark, we evaluate existing baseline methods for predicting camera intrinsics and find that most struggle to achieve accurate predictions on videos with dynamic intrinsics.



Paperid:3306
Authors:Yaniv Nikankin, Dana Arad, Yossi Gandelsman, Yonatan Belinkov
Abstract:
Vision-Language models (VLMs) show impressive abilities to answer questions on visual inputs (e.g., counting objects in an image), yet demonstrate higher accuracies when performing an analogous task on text (e.g., counting words in a text).We investigate this accuracy gap by identifying and comparing the circuits---the task-specific computational sub-graphs---in different modalities.We show that while circuits are largely disjoint between modalities, they implement relatively similar functionalities: the differences lie primarily in processing modality-specific data positions (an image or a text sequence).Zooming in on the image data representations, we observe they become aligned with the higher-performing analogous textual representations only towards later layers, too late in processing to effectively influence subsequent positions. To overcome this, we patch the representations of visual data tokens from later layers back into earlier layers. In experiments with multiple tasks and models, this simple intervention closes a third of the performance gap between the modalities, on average.Our analysis sheds light on the multi-modal performance gap in VLMs and suggests a training-free approach for reducing it.



Paperid:3307
Authors:Mincheol Park, Heeji Won, Won Woo Ro, Suhyun Kim
Abstract:
Test-time adaptation (TTA) aims to preserve model performance under distribution shifts. Yet, most existing methods rely on entropy minimization for confident predictions. This paper re-examines the sufficiency of entropy minimization by analyzing its dual relationship with energy. We view energy as a proxy for likelihood, where lower energy indicates higher observability under the learned distribution. We uncover that entropy and energy are tightly associated, controlled by the model’s confidence or ambiguity, and show that simultaneous reduction of both is essential. Importantly, we reveal that entropy minimization alone neither ensures energy reduction nor supports reliable likelihood estimation, and it requires explicit discriminative guidance to reach zero entropy. To combat these problems, we propose a twofold solution. First, we introduce a likelihood-based objective grounded in energy-based models, which reshape the energy landscape to favor test samples. For stable and scalable training, we adopt sliced score matching—a sampling-free, Hessian-insensitive approximation of Fisher divergence. Second, we enhance entropy minimization with a cross-entropy that treats the predicted class as a target to promote discriminability. By counterbalancing entropy and energy through the solution of multi-objective optimization, our unified TTA, ReTTA, outperforms existing entropy- or energy-based approaches across diverse distribution shifts.



Paperid:3308
Authors:Senqiao Yang, Junyi Li, Xin Lai, Jinming Wu, Wei Li, Zejun MA, Bei Yu, Hengshuang Zhao, Jiaya Jia
Abstract:
Recent advancements in vision-language models (VLMs) have improved performance by increasing the number of visual tokens, which are often significantly longer than text tokens.However, we observe that most real-world scenarios do not require such an extensive number of visual tokens. While the performance drops significantly in a small subset of OCR-related tasks, models still perform accurately in most other general VQA tasks with only 1/4 resolution.Therefore, we propose to dynamically process distinct samples with different resolutions, and present a new paradigm for visual token compression, namely, VisionThink.It starts with a downsampled image and smartly decides whether it is sufficient for problem solving. Otherwise, the model could output a special token to request the higher-resolution image. Compared to existing Efficient VLM methods that compress tokens using fixed pruning ratios or thresholds, VisionThink autonomously decides whether to compress tokens case by case. As a result, it demonstrates strong fine-grained visual understanding capability on OCR-related tasks, and meanwhile saves substantial visual tokens on simpler tasks.We adopt reinforcement learning and propose the LLM-as-Judge strategy to successfully apply RL to general VQA tasks. Moreoever, we carefully design a reward function and penalty mechanism to achieve a stable and reasonable image resize call ratio.Extensive experiments demonstrate the superiority, efficiency, and effectiveness of our method.All our code and data will be open-sourced.



Paperid:3309
Authors:Xinglin Wang, Yiwei Li, Shaoxiong Feng, Peiwen Yuan, Yueqi Zhang, Jiayi Shi, Chuyi Tan, Boyuan Pan, Yao Hu, Prof. Kan
Abstract:
Test-Time Scaling (TTS) improves the performance of Large Language Models (LLMs) by using additional inference-time computation to explore multiple reasoning paths through search. Yet how to allocate a fixed rollout budget most effectively during search remains underexplored, often resulting in inefficient use of compute at test time. To bridge this gap, we formulate test-time search as a resource allocation problem and derive the optimal allocation strategy that maximizes the probability of obtaining a correct solution under a fixed rollout budget. Within this formulation, we reveal a core limitation of existing search methods: solution-level allocation tends to favor reasoning directions with more candidates, leading to theoretically suboptimal and inefficient use of compute. To address this, we propose Direction-Oriented Resource Allocation (DORA), a provably optimal method that mitigates this bias by decoupling direction quality from candidate count and allocating resources at the direction level. To demonstrate DORA’s effectiveness, we conduct extensive experiments on challenging mathematical reasoning benchmarks including MATH500, AIME2024, and AIME2025. The empirical results show that DORA consistently outperforms strong baselines with comparable computational cost, achieving state-of-the-art accuracy. We hope our findings contribute to a broader understanding of optimal TTS for LLMs.



Paperid:3310
Authors:Wanjia Zhao, Jiaqi Han, Siyi Gu, Mingjian Jiang, James Zou, Stefano Ermon
Abstract:
Geometric diffusion models have shown remarkable success in molecular dynamics and structure generation. However, efficiently fine-tuning them for downstream tasks with varying geometric controls remains underexplored. In this work, we propose an SE(3)-equivariant adapter framework (GeoAda) that enables flexible and parameter-efficient fine-tuning for controlled generative tasks without modifying the original model architecture. GeoAda introduces a structured adapter design: control signals are first encoded through coupling operators, then processed by a trainable copy of selected base model layers, and finally projected back via decoupling operators followed by an equivariant zero-initialized convolution. By fine-tuning only these lightweight adapter modules, GeoAda preserves the model’s geometric consistency while mitigating overfitting and catastrophic forgetting. We theoretically prove that the proposed adapters maintain SE(3)-equivariance, ensuring that the geometric inductive biases of the pretrained diffusion model remain intact during adaptation. We demonstrate the wide applicability of \method across diverse geometric control types, including frame control, global control, subgraph control, and a broad range of application domains such as particle dynamics, molecular dynamics, human motion prediction, and molecule generation. Empirical results show that GeoAda achieves state-of-the-art fine-tuning performance while preserving original task accuracy, whereas other baselines experience significant performance degradation due to overfitting and catastrophic forgetting.



Paperid:3311
Authors:Wenbo Hu, Yining Hong, Yanjun Wang, Leison Gao, Zibu Wei, Xingcheng Yao, Nanyun Peng, Yonatan Bitton, Idan Szpektor, Kai-Wei Chang
Abstract:
Humans excel at performing complex tasks by leveraging long-term memory across temporal and spatial experiences. In contrast, current Large Language Models (LLMs) struggle to effectively plan and act in dynamic, multi-room 3D environments. We posit that part of this limitation is due to the lack of proper 3D spatial-temporal memory modeling in LLMs. To address this, we first introduce 3DMem-Bench, a comprehensive benchmark comprising over 26,000 trajectories and 2,892 embodied tasks, question-answering and captioning, designed to evaluate an agent's ability to reason over long-term memory in 3D environments.Second, we propose 3DLLM-Mem, a novel dynamic memory management and fusion model for embodied spatial-temporal reasoning and actions in LLMs. Our model uses working memory tokens, which represents current observations, as queries to selectively attend to and fuse the most useful spatial and temporal features from episodic memory, which stores past observations and interactions. Our approach allows the agent to focus on task-relevant information while maintaining memory efficiency in complex, long-horizon environments.Experimental results demonstrate that 3DLLM-Mem achieves state-of-the-art performance across various tasks, outperforming the strongest baselines by 16.5\% in success rate on 3DMem-Bench's most challenging in-the-wild embodied tasks.



Paperid:3312
Authors:Jiyan Qiu, Lyulin Kuang, Guan Wang, Yichen Xu, Leiyao Cui, Shaotong Fu, Yixin Zhu, Rita Zhang
Abstract:
Vehicle aerodynamics optimization is fundamental to automotive engineering, drag reduction, noise minimization, and vehicle body stability through complex fluid dynamics simulations. Traditional approaches rely on computationally expensive Computational Fluid Dynamics (CFD) simulations that limit design exploration or simplified models that compromise accuracy. Machine learning methods offer promising alternatives but require high-fidelity training data that has been largely unavailable in the public domain. The gap between academic machine learning research and industrial CFD applications remains unbridged due to the absence of datasets meeting rigorous engineering standards. Here we present DrivAerStar, a comprehensive and reproducible dataset of 12,000 high-precision automotive CFD simulations, created by 3 basic rear designs and 20 fine-tuned Computer Aided Design (CAD) parameters by the Free Form Deformation (FFD) algorithm, with all configurations simulated using the industry-standard STAR-CCM+® software. Unlike existing datasets, DrivAerStar provides complete engineering data that has been thoroughly validated against wind tunnel experiments with discrepancies below 5%, including aerodynamic coefficients, surface pressures, and velocity fields. Our benchmarks demonstrate that machine learning models trained on this dataset achieve industrial-grade prediction accuracy while reducing computational costs by orders of magnitude. This dataset establishes a foundation for data-driven aerodynamic design methodologies that can transform automotive development processes. Beyond automotive applications, DrivAerStar represents a paradigm for integrating high-fidelity industrial-grade physics-based simulations with artificial intelligence, potentially extending to diverse engineering disciplines where computational constraints currently limit design optimization.



Paperid:3313
Authors:Zhichao Wang, Xinhai Chen, Qinglin Wang, Xiang Gao, Qingyang Zhang, Menghan Jia, Xiang Zhang, Jie Liu
Abstract:
Abstract:Partial differential equations (PDEs) form the mathematical foundation for modeling physical systems in science and engineering, where numerical solutions demand rigorous accuracy-efficiency tradeoffs. Mesh movement techniques address this challenge by dynamically relocating mesh nodes to rapidly-varying regions, enhancing both simulation accuracy and computational efficiency. However, traditional approaches suffer from high computational complexity and geometric inflexibility, limiting their applicability, and existing supervised learning-based approaches face challenges in zero-shot generalization across diverse PDEs and mesh topologies.In this paper, we present an $\textbf{U}$nsupervised and $\textbf{G}$eneralizable $\textbf{M}$esh $\textbf{M}$ovement $\textbf{N}$etwork (UGM2N). We first introduce unsupervised mesh adaptation through localized geometric feature learning, eliminating the dependency on pre-adapted meshes. We then develop a physics-constrained loss function, M-Uniform loss, that enforces mesh equidistribution at the nodal level. Experimental results demonstrate that the proposed network exhibits equation-agnostic generalization and geometric independence in efficient mesh adaptation. It demonstrates consistent superiority over existing methods, including robust performance across diverse PDEs and mesh geometries, scalability to multi-scale resolutions and guaranteed error reduction without mesh tangling.



Paperid:3314
Authors:XINGYANG LI, Muyang Li, Tianle Cai, Haocheng Xi, Shuo Yang, Yujun Lin, Lvmin Zhang, Songlin Yang, Jinbo Hu, Kelly Peng, Maneesh Agrawala, Ion Stoica, Kurt Keutzer, Song Han
Abstract:
Abstract:Recent advances in diffusion models have enabled high-quality video generation, but the additional temporal dimension significantly increases computational costs, making training and inference on long videos prohibitively expensive. In this paper, we identify a phenomenon we call \textit{Spatiotemporal Energy Decay} in video diffusion models: post-softmax attention scores decrease as spatial and temporal distance between tokens increase, akin to the physical decay of signal or waves over space and time in nature. Motivated by this, we propose \textit{\method}, a scalable sparse attention mechanism with $\mathcal{O}(n \log n)$ complexity that translates energy decay into exponentially decaying compute density. \method employs a simple, static attention mask where each token attends to spatially nearby tokens, with the attention window size shrinking with temporal distance. Moreover, \method allows pre-trained video diffusion models to extend their generation length with minimal fine-tuning with LoRA. Extensive experiments show that \method maintains video quality across Wan2.1-14B, HunyuanVideo, achieving up to a 1.9× speedup over full attention. With minimal fine-tuning, it enables video generation up to 4× longer while reducing training costs by up to 4.4× compared to direct fine-tuning and accelerating inference by 3.7× compared to full attention inference. Code and models will be released upon publication.



Paperid:3315
Authors:Jiaxin Wu, Chenglong Pang, Guangxiong Chen, Jie Zhao
Abstract:
Graph Convolutional Networks (GCNs) based on linear aggregation have been widely applied across various domains due to their exceptional performance. To enhance performance, these networks often utilize the graph Laplacian norm to suppress the propagation of information from first-order neighbors. However, this approach may dilute valuable interaction information and make the model slowly learn sparse interaction relationships from neighbors, which increases training time and negatively affects performance. To address these issues, we introduce BoostGCN, a novel linear GCN model that focuses on amplifying significant interactions with first-order neighbors, which enables the model to accurately and quickly capture significant relationships. BoostGCN has relatively fixed parameters, making it user-friendly. Experiments on four real-world datasets demonstrate that BoostGCN outperforms existing state-of-the-art GCN models in both performance and efficiency.



Paperid:3316
Authors:Seong Hyeon Park, Jinwoo Shin
Abstract:
Understanding the 3D world from 2D monocular videos is a crucial ability for AI.Recently, to tackle this underdetermined task, end-to-end 3D geometry priors have been sought after, such as pre-trained point map models at scale.These models enable robust 3D understanding from casually taken videos, providing accurate object shapes disentangled from uncertain camera parameters.However, they still struggle when affected by object deformation and dynamics, failing to establish consistent correspondence over the frames.Furthermore, their architectures are typically limited to pairwise frame processing, which is insufficient for capturing complex motion dynamics over extended sequences.To address these limitations, we introduce Track3R, a novel framework that integrates a new architecture and task to jointly predict point map and motion trajectories across multiple frames from video input.Specifically, our key idea is modeling two disentangled trajectories for each point: one representing object motion and the other camera poses.This design not only can enable understanding of the 3D object dynamics, but also facilitates the learning of more robust priors for 3D shapes in dynamic scenes.In our experiments, Track3R demonstrates significant improvements in a joint point mapping and 3D motion estimation task for dynamic scenes, such as 25.8% improvements in the motion estimation, and 15.7% in the point mapping accuracy.



Paperid:3317
Authors:Kai Qi, Fan Wang, Zhewen Dong, Jian Sun
Abstract:
Transformers have demonstrated effectiveness in solving partial differential equations (PDEs). However, extending them to solve PDEs on complex geometries remains a challenge. In this work, we propose SpiderSolver, a geometry-aware transformer that introduces spiderweb tokenization for handling complex domain geometry and irregularly discretized points. Our method partitions the irregular spatial domain into spiderweb-like patches, guided by the domain boundary geometry. SpiderSolver leverages a coarse-grained attention mechanism to capture global interactions across spiderweb tokens and a fine-grained attention mechanism to refine feature interactions between the domain boundary and its neighboring interior points. We evaluate SpiderSolver on PDEs with diverse domain geometries across five datasets, including cars, airfoils, blood flow in the human thoracic aorta, as well as canonical cases governed by the Navier-Stokes and Darcy flow equations. Experimental results demonstrate that SpiderSolver consistently achieves state-of-the-art performance across different datasets and metrics, with better generalization ability in the OOD setting. The source codes will be released.



Paperid:3318
Authors:Chendi Qian, Christopher Morris
Abstract:
Linear and quadratic optimization are crucial in numerous real-world applications, from training machine learning models to integer-linear optimization. Recently, learning-to-optimize methods (L2O) for linear (LP) or quadratic programming (QPs) using message-passing graph neural networks (MPNNs) have gained traction, promising lightweight, data-driven proxies for solving such optimization problems. For example, they replace the costly computation of strong branching scores in branch-and-bound solvers, requiring solving many such optimization problems. However, robust L2O MPNNs remain challenging in data-scarce settings, especially when addressing complex optimization problems such as QPs. This work introduces a principled approach to data augmentation tailored for QPs via MPNNs. Our method leverages theoretically justified data augmentation techniques to generate diverse yet optimality-preserving instances. Furthermore, we integrate these augmentations into a self-supervised learning framework based on contrastive learning, thereby pretraining MPNNs for enhanced performance on L2O tasks. Extensive experiments demonstrate that our approach improves generalization in supervised scenarios and facilitates effective transfer learning to related optimization problems.



Paperid:3319
Authors:Zhe Li, Xiang Bai, Jieyu Zhang, Zhuangzhe Wu, Che Xu, Ying Li, Chengkai Hou, Shanghang Zhang
Abstract:
Constructing accurate digital twins of articulated objects is essential for robotic simulation training and embodied AI world model building, yet historically requires painstaking manual modeling or multi-stage pipelines. In this work, we propose \textbf{URDF-Anything}, an end-to-end automatic reconstruction framework based on a 3D multimodal large language model (MLLM). URDF-Anything utilizes an autoregressive prediction framework based on point-cloud and text multimodal input to jointly optimize geometric segmentation and kinematic parameter prediction. It implements a specialized [SEG] token mechanism that interacts directly with point cloud features, enabling fine-grained part-level segmentation while maintaining consistency with the kinematic parameter predictions.Experiments on both simulated and real-world datasets demonstrate that our method significantly outperforms existing approaches regarding geometric segmentation (mIoU 17\% improvement), kinematic parameter prediction (average error reduction of 29\%), and physical executability (surpassing baselines by 50\%). Notably, our method exhibits excellent generalization ability, performing well even on objects outside the training set. This work provides an efficient solution for constructing digital twins for robotic simulation, significantly enhancing the sim-to-real transfer capability.



Paperid:3320
Authors:Yinqi Li, Jiahe Zhao, Hong Chang, RuiBing Hou, Shiguang Shan, Xilin Chen
Abstract:
Abstract:Contrastive Language-Image Pre-training (CLIP) has become a foundation model and has been applied to various vision and multimodal tasks. However, recent works indicate that CLIP falls short in distinguishing detailed differences in images and shows suboptimal performance on dense-prediction and vision-centric multimodal tasks. Therefore, this work focuses on improving existing CLIP models, aiming to capture as many visual details in images as possible. We find that a specific type of generative models, unCLIP, provides a suitable framework for achieving our goal. Specifically, unCLIP trains an image generator conditioned on the CLIP image embedding. In other words, it inverts the CLIP image encoder. Compared to discriminative models like CLIP, generative models are better at capturing image details because they are trained to learn the data distribution of images. Additionally, the conditional input space of unCLIP aligns with CLIP's original image-text embedding space. Therefore, we propose to invert unCLIP (dubbed un$^2$CLIP) to improve the CLIP model. In this way, the improved image encoder can gain unCLIP's visual detail capturing ability while preserving its alignment with the original text encoder simultaneously. We evaluate our improved CLIP across various tasks to which CLIP has been applied, including the challenging MMVP-VLM benchmark, the dense-prediction open-vocabulary segmentation task, and multimodal large language model tasks. Experiments show that un$^2$CLIP significantly improves the original CLIP and previous CLIP improvement methods.



Paperid:3321
Authors:WEI-KAI CHANG, Rajiv Khanna
Abstract:
As deep learning models continue to scale, the growing computational demands have amplified the need for effective coreset selection techniques. Coreset selection aims to accelerate training by identifying small, representative subsets of data that approximate the performance of the full dataset. Among various approaches, gradient-based methods stand out due to their strong theoretical underpinnings and practical benefits, particularly under limited data budgets. However, these methods face challenges such as naïve stochastic gradient descent (SGD) acting as a surprisingly strong baseline and the breakdown of representativeness due to loss curvature mismatches over time.In this work, we propose a novel framework that addresses these limitations. First, we establish a connection between posterior sampling and loss landscapes, enabling robust coreset selection even in high-data-corruption scenarios. Second, we introduce a smoothed loss function based on posterior sampling onto the model weights, enhancing stability and generalization while maintaining computational efficiency. We also present a novel convergence analysis for our sampling-based coreset selection method. Finally, through extensive experiments, we demonstrate how our approach achieves faster training and enhanced generalization across diverse datasets than the current state of the art.



Paperid:3322
Authors:Buyun Liang, Liangzu Peng, Jinqi Luo, Darshan Thaker, Kwan Ho Ryan Chan, Rene Vidal
Abstract:
Large Language Models (LLMs) are increasingly deployed in high-risk domains where trustworthy outputs are essential. Despite their strong performance, state-of-the-art LLMs can still produce hallucinations, raising concerns about their reliability. To better understand this failure mode, prior works have explored eliciting hallucinations via adversarial attack methods. However, existing attacks often generate prompts that are unlikely to occur in real-world interactions, either by appending gibberish characters or modifying the underlying meaning of the original prompt, thus providing little insight into how hallucinations may occur. In this work, we propose an adversarial attack method that searches for Semantically Equivalent and Coherent Attacks (SECA) to realistically elicit hallucinations. Our approach formulates the attack as a constrained optimization problem over the input prompt space under semantic equivalence and coherence constraints. To avoid searching for successful attack prompts combinatorially, our proposed solver consists of two models: An LLM-based proposer that generates diverse rephrasings of a given prompt and an LLM-based feasibility checker that verifies that the generated attack prompts satisfy linguistic conditions such as mutual entailment and information retention.Our experiments on the MMLU dataset show that SECA can generate attacks that achieve a higher attack success rate and lower constraint violations than existing attack methods. By applying our gradient-free method to a wide range of open-source and commercial, gradient-inaccessible LLMs, our study highlights the sensitivity of LLMs to realistic and plausible prompt variations.



Paperid:3323
Authors:Fangtong Sun, Congyu Li, Ke Yang, Hanwen Yu, Yuchen Pan, Xichuan Zhang, Yiying Li
Abstract:
Low-light vision remains a fundamental challenge in computer vision due to severe illumination degradation, which significantly affects the performance of downstream tasks such as detection and segmentation. While recent state-of-the-art methods have improved performance through invariant feature learning modules, they still fall short due to incomplete modeling of low-light conditions. Therefore, we revisit low-light image formation and extend the classical Lambertian model to better characterize low-light conditions. By shifting our analysis to the frequency domain, we theoretically prove that the frequency-domain channel ratio can be leveraged to extract illumination-invariant features via a structured filtering process. We then propose a novel and end-to-end trainable module named \textbf{F}requency-domain \textbf{R}adial \textbf{B}asis \textbf{Net}work (\textbf{FRBNet}), which integrates the frequency-domain channel ratio operation with a learnable frequency domain filter for the overall illumination-invariant feature enhancement. As a plug-and-play module, FRBNet can be integrated into existing networks for low-light downstream tasks without modifying loss functions. Extensive experiments across various downstream tasks demonstrate that FRBNet achieves superior performance, including +2.2 mAP for dark object detection and +2.9 mIoU for nighttime segmentation. Code is available at \href{https://anonymous.4open.science/r/FRBNet_Anony}{FRBNet_anony}.



Paperid:3324
Authors:Duo Zheng, shijia Huang, Yanyang Li, Liwei Wang
Abstract:
Previous research has explored the use of Multimodal Large Language Models (MLLMs) for 3D scene understanding by treating scenes as videos. These methods typically rely on explicit dense 3D inputs, such as point clouds or reconstructed Bird's-Eye View (BEV) maps. In this work, we take an innovative step forward by questioning whether we can improve MLLMs' 3D spatial understanding and reasoning capability by directly learning from videos with 3D vision geometry priors. We introduce a simple yet effective approach: the Video-3D Geometry LLM (VG LLM). Our method utilizes a 3D visual geometry encoder to extract 3D geometry prior information from the input video sequences. This information is integrated with visual tokens and fed into the MLLM. Our approach significantly boosts the 3D perception and reasoning capabilities of MLLMs, demonstrating notable improvements in various 3D scene understanding tasks and spatial reasoning benchmarks, all learned directly from videos. Remarkably, without relying on any explicit 3D data inputs, our 4B model achieves competitive performance with prior state-of-the-art methods, even outperforming the Gemini-1.5-Pro on the VSI-Bench.



Paperid:3325
Authors:Andreas Opedal, Yanick Zengaffinen, Haruki Shirakami, Clemente Pasti, Mrinmaya Sachan, Abulhair Saparov, Ryan Cotterell, Bernhard Schölkopf
Abstract:
Much recent work has focused on how language models can be adapted to solve reasoning problems through search methods. To solve a task in a realistic setting, an agent often needs to search over many facts, many of which are irrelevant to said task, and do so in an efficient manner. In this paper, we investigate how language models perform in such a setting, focusing on whether they can successfully prove theorems, if they take unnecessary proof steps, and which form of search algorithm they employ. We consider a proof system for GSM-like arithmetic reasoning problems and evaluate language models on problems with a large space of arithmetic facts that are irrelevant to the task. Beyond answer accuracy, we also verify the full proof generated by the language model. This is done by measuring how well the language model output matches the most efficient ground-truth proof, which is a provably unique normal form under the proof system. We find that both standard and reasoning-oriented models perform worse on problems that have been augmented with irrelevant facts of different types. Even when they correctly prove the goal theorem, they are often inefficient, using more test-time compute than necessary. Our analysis suggests that their algorithm is closer to depth-first than to breadth-first search, and that they use information from the query as a heuristic, but that this heuristic still sometimes leads them down irrelevant paths.



Paperid:3326
Authors:Florian Sestak, Artur Toshev, Andreas Fürst, Günter Klambauer, Andreas Mayr, Johannes Brandstetter
Abstract:
Generative models are spearheading recent progress in deep learning, showcasing strong promise for trajectory sampling in dynamical systems as well. However, whereas latent space modeling paradigms have transformed image and video generation, similar approaches are more difficult for most dynamical systems. Such systems -- from chemical molecule structures to collective human behavior -- are described by interactions of entities, making them inherently linked to connectivity patterns, entity conservation, and the traceability of entities over time. Our approach, LaM-SLidE (Latent Space Modeling of Spatial Dynamical Systems via Linked Entities), bridges the gap between: (1) keeping the traceability of individual entities in a latent system representation, and (2) leveraging the efficiency and scalability of recent advances in image and video generation, where pre-trained encoder and decoder enable generative modeling directly in latent space. The core idea of LaM-SLidE is the introduction of identifier representations (IDs) that enable the retrieval of entity properties and entity composition from latent system representations, thus fostering traceability.Experimentally, across different domains, we show that LaM-SLidE performs favorably in terms of speed, accuracy, and generalizability. Code is available at https://anonymous.4open.science/r/lam-slide-B38B.



Paperid:3327
Authors:Mengyu Ye, Jun Suzuki, Tatsuro Inaba, Tatsuki Kuribayashi
Abstract:
Recent interpretability work on large language models (LLMs) is increasingly dominated by a feature discovery approach with the help of proxy modules, where the quality of features learned by, e.g., sparse auto-encoders (SAEs), has been evaluated.This paradigm naturally raises a critical question --- how much better features such proxies have indeed discovered, especially compared to those already represented within the original model parameters --- and unfortunately, such a comparison has little been made so far.In this work, we revisit the interpretability of feature vectors stored in feed-forward (FF) layers, given the perspective of FF as key-value memories, with modern interpretability benchmarks.Our extensive evaluation revealed that SAE and FFs exhibits a similar range of interpretability, although SAEs displayed an observable but minimal improvement in some aspects. Furthermore, in certain aspects, surprisingly, even vanilla FFs yielded better interpretability scores than the SAEs, and features discovered in SAEs and FFs diverged.These bring questions about the advantage of SAEs from both perspectives of feature quality and faithfulness, compared to directly interpreting FF feature vectors.



Paperid:3328
Authors:Kunyu Wang, Xueyang Fu, Yuanfei Bao, Chengjie Ge, Chengzhi Cao, Wei Zhai, Zheng-Jun Zha
Abstract:
Continual Test-Time Adaptation (CTTA) aims to online adapt a pre-trained model to changing environments during inference. Most existing methods focus on exploiting target data, while overlooking another crucial source of information, the pre-trained weights, which encode underutilized domain-invariant priors. This paper takes the geometric attributes of pre-trained weights as a starting point, systematically analyzing three key components: magnitude, absolute angle, and pairwise angular structure. We find that the pairwise angular structure remains stable across diverse corrupted domains and encodes domain-invariant semantic information, suggesting it should be preserved during adaptation. Based on this insight, we propose PAID (Pairwise Angular Invariant Decomposition), a prior-driven CTTA method that decomposes weight into magnitude and direction, and introduces a learnable orthogonal matrix via Householder reflections to globally rotate direction while preserving the pairwise angular structure. During adaptation, only the magnitudes and the orthogonal matrices are updated. PAID achieves consistent improvements over recent SOTA methods on four widely used CTTA benchmarks, demonstrating that preserving pairwise angular structure offers a simple yet effective principle for CTTA.



Paperid:3329
Authors:Vishnu Vinod, Krishna Pillutla, Abhradeep Guha Thakurta
Abstract:
Abstract:As major progress in LLM-based long-form text generation enables paradigms such as retrieval-augmented generation (RAG) and inference-time scaling, safely incorporating private information into the generation remains a critical open question. We present InvisibleInk, a highly scalable long-form text generation framework satisfying rigorous differential privacy guarantees with respect to the sensitive references. It interprets sampling from the LLM's next-token-distribution as the exponential mechanism over the LLM logits with two innovations. First, we reduce the privacy cost by isolating and clipping only the sensitive information in the model logits (relative to the public logits). Second, we improve text quality by sampling from a small superset of the top-$k$ private tokens. Empirical evaluations demonstrate a consistent $8\times$ reduction in computation cost over state-of-the-art baselines to generate long-form private text of the same utility across privacy levels. In summary, InvisibleInk is able to generate private long-form text at less than $10$ times the computation cost of non-private generation.



Paperid:3330
Authors:Liang Peng, Boxi Wu, Haoran Cheng, Yibo Zhao, Xiaofei He
Abstract:
Direct preference optimization (DPO) is an effective method for aligning generative models with human preferences and has been successfully applied to fine‑tune text‑to‑image diffusion models. Its practical adoption, however, is hindered by a labor‑intensive pipeline that first produces a large set of candidate images and then requires humans to rank them pairwise. We address this bottleneck with self‑supervised direct preference optimization, a new paradigm that removes the need for any pre‑generated images or manual ranking. During training, we create preference pairs on the fly through self‑supervised image transformations, allowing the model to learn from fresh and diverse comparisons at every iteration. This online strategy eliminates costly data collection and annotation while remaining plug‑and‑play for any text‑to‑image diffusion method. Surprisingly, the on‑the‑fly pairs produced by the proposed method not only match but exceed the effectiveness of conventional DPO, which we attribute to the greater diversity of preferences sampled during training. Extensive experiments with Stable Diffusion 1.5 and Stable Diffusion XL confirm that our method delivers substantial gains. The codes are available in the supplementary material.



Paperid:3331
Authors:Francesco Orabona, Ryan D'Orazio
Abstract:
The Polyak stepsize has been proven to be a fundamental stepsize in convex optimization, giving near optimal gradient descent rates across a wide range of assumptions.The universality of the Polyak stepsize has also inspired many stochastic variants,with theoretical guarantees and strong empirical performance. Despite the many theoretical results, our understanding of the convergence properties and shortcomings of the Polyak stepsize or its variants is both incomplete and fractured across different analyses. We propose a new, unified, and simple perspective for the Polyak stepsize and its variants as gradient descent on a surrogate loss. We show that each variant is equivalent to minimize a surrogate function with stepsizes that adapt to a guaranteed local curvature.Our general surrogate loss perspective is then used to provide a unified analysis of existing variants across different assumptions.Moreover, we show a number of negative results proving that the non-convergence results in some of the upper bounds is indeed real.



Paperid:3332
Authors:Zaifeng Pan, AJJKUMAR DAHYALAL PATEL, Zhengding Hu, Yipeng Shen, Yue Guan, Wan-Lu Li, Lianhui Qin, Yida Wang, Yufei Ding
Abstract:
Large language model (LLM) based agentic workflows have become a popular paradigm for coordinating multiple specialized agents to solve complex tasks. To improve serving efficiency, existing LLM systems employ prefix caching to reuse key-value (KV) tensors corresponding to agents' fixed prompts, thereby avoiding redundant computation across repeated invocations. However, current systems typically evict KV caches using a Least Recently Used (LRU) policy, which fails to anticipate future agent usage and often discards KV caches shortly before their reuse. This leads to frequent cache misses and substantial recomputation or swap- ping overhead. We present KVFlow, a workflow-aware KV cache management framework tailored for agentic workloads. KVFlow abstracts the agent execution schedule as an Agent Step Graph and assigns each agent a steps-to-execution value that estimates its temporal proximity to future activation. These values guide a fine-grained eviction policy at the KV node level, allowing KVFlow to preserve entries likely to be reused and efficiently manage shared prefixes in tree-structured caches. Moreover, KVFlow introduces a fully overlapped KV prefetching mecha- nism, which proactively loads required tensors from CPU to GPU in background threads for agents scheduled in the next step, thereby avoiding cache miss stalls during generation. Compared to SGLang with hierarchical radix cache, KVFlow achieves up to 1.83× speedup for single workflows with large prompts, and up to 2.19× speedup for scenarios with many concurrent workflows.



Paperid:3333
Authors:Hanlin Yu, Berfin Inal, Georgios Arvanitidis, Søren Hauberg, Francesco Locatello, Marco Fumero
Abstract:
Neural models learn representations of high dimensional data which lie on low dimensional manifolds. Multiple factors, including stochasticities in the training process, model architectures, and additional inductive biases, may induce different representations, even when learning the same task on the same data. However, it has recently been shown that when a latent structure is shared between distinct latent spaces, relative distances between representations can be preserved, up to distortions. Building on this idea, we demonstrate that exploiting the differential-geometric structure of latent spaces of neural models, it is possible to capturepreciselythe transformations between representational spaces trained on similar data distributions. Specifically, we assume that distinct neural models parametrize approximately the same underlying manifold, we introduce a representation based on thepullback metric, that captures the intrinsic structure of the latent space, while scaling efficiently to large models. We validate experimentally our method on model stitching and retrieval tasks, covering autoencoders and vision foundation discriminative models, across diverse architectures, datasets and pretraining schemes.



Paperid:3334
Authors:Boyu Li, Peiqi Duan, Zhaojun Huang, Xinyu Zhou, Yifei Xia, Boxin Shi
Abstract:
Dense metric depth estimation has witnessed great developments in recent years. While single-image-based methods have demonstrated commendable performance in certain circumstances, they may encounter challenges regarding scale ambiguities and visual illusions in real world. Traditional depth-from-focus methods are constrained by low sampling rates during data acquisition. In this paper, we introduce a novel approach to enhance dense metric depth estimation by fusing events with image foundation models via a prompting approach. Specifically, we build Event-based Differential Focus Volumes (EDFVs) using events triggered through focus sweeping, which are subsequently transformed into sparse metric depth maps. These maps are then utilized for prompting dense depth estimation via our proposed Event-based Depth Prompting Network (EDPN). We further construct synthetic and real-captured datasets to facilitate the training and evaluation of both frame-based and event-based methods. Quantitative and qualitative results, including both in-domain and zero-shot experiments, demonstrate the superior performance of our method compared to existing approaches.



Paperid:3335
Authors:Arina Medvedeva, Edoardo Balzani, Alex Williams, Stephen Keeley
Abstract:
The Poisson Generalized Linear Model (GLM) is a foundational tool for analyzing neural spike train data.However, standard implementations rely on discretizing spike times into binned count data, limiting temporal resolution and scalability. Here, we develop stochastic optimization methods and polynomial approximations to the continuous-time analog of these models, and show them to be advantageous over their discrete-time counterparts. Further, we propose using a set of exponentially scaled Laguerre polynomials as an orthogonal temporal basis, which improves filter identification and yields closed-form integral solutions under the polynomial approximation. Applied to both synthetic and real spike-time data from rodent hippocampus, our methods demonstrate superior accuracy and scalability compared to traditional binned GLMs, enabling functional connectivity inference in large-scale neural recordings that are temporally precise on the order of synaptic dynamical timescales. We provide open-source implementations of both MC and PA estimators, optimized for GPU acceleration, to facilitate adoption in the neuroscience community.



Paperid:3336
Authors:Shukai Gong, YIYANG FU, Fengyuan Ran, Quyu Kong, Feng Zhou
Abstract:
Abstract:We propose TPP-SD, a novel approach that accelerates Transformer temporal point process (TPP) sampling by adapting speculative decoding (SD) techniques from language models. By identifying the structural similarities between thinning algorithms for TPPs and speculative decoding for language models, we develop an efficient sampling framework that leverages a smaller draft model to generate multiple candidate events, which are then verified by the larger target model. TPP-SD maintains the same output distribution as autoregressive sampling while achieving significant acceleration. Experiments on both synthetic and real datasets demonstrate that our approach produces samples from identical distributions as standard methods, but with 2-6$\times$ speedup. Our ablation studies analyze the impact of hyperparameters such as draft length and draft model size on sampling efficiency. TPP-SD bridges the gap between powerful Transformer TPP models and the practical need for rapid sequence generation.



Paperid:3337
Authors:Jiaming Ma, Binwu Wang, Qihe Huang, Guanjun Wang, Pengkun Wang, Zhengyang Zhou, Yang Wang
Abstract:
The stable periodic patterns present in time series data serve as the foundation for long-term forecasting. However, existing models suffer from limitations such as continuous and chaotic input partitioning, as well as weak inductive biases, which restrict their ability to capture such recurring structures. In this paper, we propose MoFo, which interprets periodicity as both the correlation of period-aligned time steps and the trend of period-offset time steps. We first design period-structured patches—2D tensors generated through discrete sampling—where each row contains only period-aligned time steps, enabling direct modeling of periodic correlations. Period-offset time steps within a cycle are aligned in columns. To capture trends across these offset time steps, we introduce a period-aware modulator. This modulator introduces an adaptive strong inductive bias through a regulated relaxation function, encouraging the model to generate attention coefficients that align with periodic trends. This function is end-to-end trainable, enabling the model to adaptively capture the distinct periodic patterns across diverse datasets. Extensive empirical results on popular benchmark datasets demonstrate that MoFo achieves competitive performance compared to 17 advanced baselines, while offering up to 14x memory efficiency gain and 10x faster training speed.



Paperid:3338
Authors:Jeffrey Willette, Heejun Lee, Sung Ju Hwang
Abstract:
The attention mechanism of a transformer has a quadratic complexity, leading to high inference costs and latency for long sequences. However, attention matrices are mostly sparse, which implies that many entries may be omitted from computation for efficient inference. Sparse attention inference methods aim to reduce this computational burden; however, they also come with a troublesome performance degradation. We discover that one reason for this degradation is that the sparse calculation induces a distributional shift in the attention outputs. The distributional shift causes decoding-time queries to fail to align well with the appropriate keys from the prefill stage, leading to a drop in performance. We propose a simple, novel, and effective procedure for correcting this distributional shift, bringing the distribution of sparse attention outputs closer to that of quadratic attention. Our method can be applied on top of any sparse attention method, and results in an average 36\%pt performance increase, recovering 88\% of quadratic attention accuracy on the 131K RULER benchmark when applied on top of sliding window attention with sink tokens while only adding a small overhead. Our method can maintain approximately 98.5\% sparsity over full quadratic attention, making our model 32 times faster than Flash Attention 2 when processing 1M token prefills.



Paperid:3339
Authors:Jiawei Huang, Minming Li, Hu Ding
Abstract:
Contrastive self-supervised learning has emerged as a powerful paradigm for extracting meaningful representations without labels. While effective at capturing broad categorical distinctions, current methods often struggle to preserve the fine-grained and hierarchical relationships inherent in real-world data. From the perspective of semantic alignment, conventional contrastive learning aligns representations to semantic structure at a global level, treating the entire embedding space uniformly and frequently overlooking rich local structural information. In this paper, we propose \emph{Adaptive Multi-scale Affinity alignment (AMA-alignment)}, a framework that introduces localized contrastive objectives and a dynamic multi-scale optimization strategy to adaptively identify and refine poorly aligned regions within the embedding space. Although our model is inherently more complex due to its \emph{multi-scale} and \emph{adaptive} design, we provide the theoretical guarantees indicating that its convergence rate remains comparable to that of standard smooth non-convex optimization. We conduct a set of experiments on diverse benchmarks to show that AMA-alignment can effectively preserve hierarchical structure; moreover, AMA-alignment also outperforms existing contrastive methods on a range of downstream tasks.



Paperid:3340
Authors:Azadeh Motamedi, Jae-Mo Kang, Il-Min Kim
Abstract:
Vision–Language Models (VLMs) have recently attracted considerable attention in Federated Learning (FL) due to their strong and robust performance. In particular, few-shot adaptation with pre-trained VLMs like CLIP enhances the performance of downstream tasks. However, existing methods still suffer from substantial communication overhead, high local computational demands, and suboptimal performance under non-IID user data. To simultaneously address all those limitations, we propose NormFit, a lightweight solution that selectively fine-tunes only a very small portion of the model parameters, specifically only the Pre-LayerNorm parameters of the vision encoder within a VLM. Overcoming the existing tradeoff between performance and communication/computation efficiency in few-shot FL, NormFit sets a new benchmark by simultaneously achieving superior accuracy and substantially reduced communication and computational demands. Theoretically, we show that NormFit yields a considerably smaller generalization gap compared to tuning all LayerNorm parameters. Importantly, NormFit can function effectively as a standalone solution or integrate seamlessly with existing few-shot fine-tuning methods to further enhance their performance. Notably, NormFit offers implementation simplicity, achieving these improvements without any algorithmic modifications, changes to the underlying model architecture, or the addition of external parameters. Code is provided as the supplementary material.



Paperid:3341
Authors:Zecheng Wang, Chunshan Li, Yupeng Zhang, Han Liu, Bingning Wang, Dianhui Chu, Dianbo Sui
Abstract:
Abstract:Direct Preference Optimization (DPO) is a widely used preference optimization algorithm in large language model (LLM) alignment, which reparameterizes the reward function in reinforcement learning with human feedback (RLHF) without requiring a separate reward model. However, during the DPO training process, when a large negative gradient is applied to low-confidence samples, LLMs with a softmax output head tend to squeeze the confidence in the model's output distribution towards the highest-confidence sentence, which may lead to a decrease in the confidence of both preference and non-preference samples, while increasing the confidence of unrelated tokens.This phenomenon becomes more complex in reasoning tasks. In this work, focusing on reasoning tasks, we propose VPO, a negative gradient constraint method for human non-preference samples based on $\mathcal{V}$-usable information.By using $\mathcal{V}$-usable information to measure the similarity between preference pairs and selectively constrain the negative gradient, VPO can alleviate the squeezing effect of DPO, enhance alignment with the generation objective, and maintain the model's ability to distinguish between preference and non-preference samples.We compare VPO with DPO and its latest variants on mathematical reasoning tasks using the LLama 3.1 and Qwen 2.5 series, including both Base and Instruct models.Our results demonstrate that VPO consistently and significantly outperforms existing methods. Specifically, on Qwen2.5-7B-Base, VPO achieves 7.80\% and 13.25\% improvement over DPO on MATH500 and AMC23, respectively. We also conduct ablation experiments and in-depth analysis on VPO to explain its effectiveness and rationale.



Paperid:3342
Authors:Lu Gao, Wenlan Chen, Daoyuan Wang, Cheng Liang, Fei Guo
Abstract:
Cross-modal representation learning aims to extract semantically aligned representations from heterogeneous modalities such as images and text. Existing multimodal VAE-based models often suffer from limited capability to align heterogeneous modalities or lack sufficient structural constraints to clearly separate the modality-specific and shared factors. In this work, we propose a novel framework, termedDisentangledCross-Modal Representation Learning withEnhancedMutual Supervision (DCMEM). Specifically, our model disentangles the common and distinct information across modalities and regularizes the shared representation learned from each modality in a mutually supervised manner. Moreover, we incorporate the information bottleneck principle into our model to ensure that the shared and modality-specific factors encode exclusive yet complementary information. Notably, our model is designed to be trainable on both complete and partial multimodal datasets with a valid Evidence Lower Bound. Extensive experimental results demonstrate significant improvements of our model over existing methods on various tasks including cross-modal generation, clustering, and classification.



Paperid:3343
Authors:Zihan Wang, Seungjun Lee, Gim Hee Lee
Abstract:
Vision-and-Language Navigation (VLN) is a core task where embodied agents leverage their spatial mobility to navigate in 3D environments toward designated destinations based on natural language instructions. Recently, video-language large models (Video-VLMs) with strong generalization capabilities and rich commonsense knowledge have shown remarkable performance when applied to VLN tasks. However, these models still encounter the following challenges when applied to real-world 3D navigation: 1) Insufficient understanding of 3D geometry and spatial semantics; 2) Limited capacity for large-scale exploration and long-term environmental memory; 3) Poor adaptability to dynamic and changing environments.To address these limitations, we propose Dynam3D, a dynamic layered 3D representation model that leverages language-aligned, generalizable, and hierarchical 3D representations as visual input to train 3D-VLM in navigation action prediction. Given posed RGB-D images, our Dynam3D projects 2D CLIP features into 3D space and constructs multi-level 3D patch-instance-zone representations for 3D geometric and semantic understanding with a dynamic and layer-wise update strategy. Our Dynam3D is capable of online encoding and localization of 3D instances, and dynamically updates them in changing environments to provide large-scale exploration and long-term memory capabilities for navigation. By leveraging large-scale 3D-language pretraining and task-specific adaptation, our Dynam3D sets new state-of-the-art performance on VLN benchmarks including R2R-CE, REVERIE-CE and NavRAG-CE under monocular settings. Furthermore, experiments for pre-exploration, lifelong memory, and real-world robot validate the effectiveness of practical deployment.



Paperid:3344
Authors:Qiang Xiang, Shuang Sun, Binglei Li, Dejia Song, Huaxia Li, Yibo Chen, Xu Tang, Yao Hu, Junping Zhang
Abstract:
Diffusion models have demonstrated remarkable capabilities in generating high-quality images. Recent advancements in Layout-to-Image (L2I) generation have leveraged positional conditions and textual descriptions to facilitate precise and controllable image synthesis. Despite overall progress, current L2I methods still exhibit suboptimal performance.Therefore, we propose InstanceAssemble, a novel architecture that incorporates layout conditions via instance-assembling attention, enabling position control with bounding boxes (bbox) and multimodal content control including texts and additional visual content. Our method achieves flexible adaption to existing DiT-based T2I models through light-weighted LoRA modules. Additionally, we propose a Layout-to-Image benchmark, Denselayout, a comprehensive benchmark for layout-to-image generation, containing 5k images with 90k instances in total. We further introduce Layout Grounding Score (LGS), an interpretable evaluation metric to more precisely assess the accuracy of L2I generation. Experiments demonstrate that our InstanceAssemble method achieves state-of-the-art performance under complex layout conditions, while exhibiting strong compatibility with diverse style LoRA modules.



Paperid:3345
Authors:Maximilian Egger, Rawad Bitar, Antonia Wachter-Zeh, Nir Weinberger, Deniz Gunduz
Abstract:
Abstract:Federated Learning (FL) incurs high communication costs in both uplink and downlink. The literature largely focuses on lossy compression of model updates in deterministic FL. In contrast, stochastic (Bayesian) FL considers distributions over parameters, enabling uncertainty quantification, better generalization, and, crucially, inherent communication-regularized training through a mirror-descent structure. In this paper, we consider both uplink and downlink communication in stochastic FL, and propose a communication framework based on remote source generation. Employing Minimal Random Coding (MRC) for remote generation, we allow the server and the clients to sample from local and global posteriors (sources), respectively, rather than transmitting locally sampled updates. The framework encompasses communication-regularized local optimization and principled compression of model updates, leveraging gradually updated prior distributions as side information. Through extensive simulations, we show that our method achieves $5-32\times$ reduction in total communication cost while preserving accuracy. We further analyze the communication cost, refining existing MRC bounds and enabling precise quantification of uplink and downlink trade-offs. We also extend our method to conventional FL via stochastic quantization and prove a contraction property for the biased MRC compressor to facilitate convergence analysis.



Paperid:3346
Authors:Qinsi Wang, Jinghan Ke, Hancheng Ye, Yueqian Lin, Yuzhe Fu, Jianyi Zhang, Kurt Keutzer, Chenfeng Xu, Yiran Chen
Abstract:
Abstract:Current Reinforcement Fine-tuning (RFT) paradigms for Large Language Models (LLMs) suffer from sample inefficiency due to the redundant exposure of identical queries under uniform data sampling. While previous work has explored curriculum learning via heuristic difficulty metrics, these strategies exhibit limitations by neglecting the intrinsic learning signals generated by the model itself, thus leading to suboptimal training regimes. In this paper, we identify a model-inherent signal termed *angle concentration* that effectively reflects an LLM's capacity to learn from specific data. We theoretically and empirically demonstrate a correlation between the angular distribution of token hidden state vectors and the resulting gradient, revealing a learning preference for data exhibiting higher angle concentration. Inspired by this finding, we propose GAIN-RL, a Gradient-driven Angle-Informed Navigated RL framework. By leveraging the model's intrinsic angle concentration signal, GAIN-RL dynamically selects training data in each epoch, ensuring consistently impactful gradient updates and thus significantly enhancing overall training efficiency. Empirical evaluations show that GAIN-RL (GRPO) achieves over a 2.5$\times$ acceleration in training efficiency across diverse mathematical and coding tasks and varying model scales. Furthermore, GAIN-RL (GRPO)'s efficient sampling yields data-efficient training, achieving better performance with half the original data compared to vanilla GRPO with full training data.



Paperid:3347
Authors:Xixi Wan, AIHUA ZHENG, Bo Jiang, Beibei Wang, Chenglong Li, Jin Tang
Abstract:
Multi-modal object Re-IDentification (ReID) has gained considerable attention with the goal of retrieving specific targets across cameras using heterogeneous visual data sources. At present, multi-modal object ReID faces two core challenges: (1) learning robust features under fine-grained local noise caused by occlusion, frame loss, and other disruptions; and (2) effectively integrating heterogeneous modalities to enhance multi-modal representation. To address the above challenges, we propose a robust approach named Uncertainty-Guided Graph model for multi-modal object ReID (UGG-ReID). UGG-ReID is designed to mitigate noise interference and facilitate effective multi-modal fusion by estimating both local and sample-level epistemic uncertainty and explicitly modeling their dependencies. Specifically, we first propose the Gaussian patch-graph representation model that leverages uncertainty to quantify fine-grained local cues and capture their structural relationships. This process boosts the expressiveness of modal-specific information, ensuring that the generated embeddings are both more informative and robust. Subsequently, we design an uncertainty-guided mixture of experts strategy that dynamically routes samples to experts exhibiting low uncertainty. This strategy effectively suppresses noise-induced instability, leading to enhanced robustness. Meanwhile, we design an uncertainty-guided routing to strengthen the multi-modal interaction, improving the performance. UGG-ReID is comprehensively evaluated on five representative multi-modal object ReID datasets, encompassing diverse spectral modalities. Experimental results show that the proposed method achieves excellent performance on all datasets and is significantly better than current methods in terms of noise immunity. Our code will be made public upon acceptance.



Paperid:3348
Authors:Xingyu Liu, Pengfei Ren, Qi Qi, Haifeng Sun, Zirui Zhuang, Jing Wang, Jianxin Liao, Jingyu Wang
Abstract:
Recent advances in hand–object interaction modeling have employed implicit representations, such as Signed Distance Functions (SDF), Neural Radiance Fields (NeRF), and 3D Gaussian Splatting (3DGS), to reconstruct hands and objects with arbitrary topology and photorealistic detail. However, these methods often rely on dense 3D surface annotations, or are tailored to short clips constrained in motion trajectories and scene contexts, limiting their generalization to diverse environments and movement patterns. In this work, we present HOGS, an adaptively perceptive 3DGS-based framework for generalizable hand–object modeling from unconstrained monocular RGB images. By integrating photometric cues from the visual modality with the physically grounded structure of 3D Gaussians, HOGS disentangles inherent geometry from transient lighting and motion-induced appearance changes. This endows implicit representations with the ability to generalize to unseen environments and dynamic motion patterns. Experiments on two challenging datasets demonstrate that HOGS outperforms state-of-the-art methods in monocular hand-object reconstruction and photo-realistic rendering.



Paperid:3349
Authors:Yiren Song, Cheng Liu, Mike Zheng Shou
Abstract:
Diffusion models have advanced image stylization significantly, yet two core challenges persist: (1) maintaining consistent stylization in complex scenes, particularly identity, composition, and fine details, and (2) preventing style degradation in image-to-image pipelines with style LoRAs. GPT-4o's exceptional stylization consistency highlights the performance gap between open-source methods and proprietary models. To bridge this gap, we propose \textbf{OmniConsistency}, a universal consistency plugin leveraging large-scale Diffusion Transformers (DiTs). OmniConsistency contributes: (1) an in-context consistency learning framework trained on aligned image pairs for robust generalization; (2) a two-stage progressive learning strategy decoupling style learning from consistency preservation to mitigate style degradation; and (3) a fully plug-and-play design compatible with arbitrary style LoRAs under the Flux framework. Extensive experiments show that OmniConsistency significantly enhances visual coherence and aesthetic quality, achieving performance comparable to commercial state-of-the-art model GPT-4o.



Paperid:3350
Authors:Lujian Yao, Siming Zheng, Xinbin Yuan, Zhuoxuan Cai, Pu Wu, Jinwei Chen, Bo Li, Peng-Tao Jiang
Abstract:
Traditional photography composition approaches are dominated by 2D cropping-based methods. However, these methods fall short when scenes contain poorly arranged subjects. Professional photographers often employ perspective adjustment as a form of 3D recomposition, modifying the projected 2D relationships between subjects while maintaining their actual spatial positions to achieve better compositional balance. Inspired by this artistic practice, we propose photography perspective composition (PPC), extending beyond traditional cropping-based methods. However, implementing the PPC faces significant challenges: the scarcity of perspective transformation datasets and undefined assessment criteria for perspective quality. To address these challenges, we present three key contributions: (1) An automated framework for building PPC datasets through expert photographs. (2) A video generation approach that demonstrates the transformation process from suboptimal to optimal perspectives. (3) A perspective quality assessment (PQA) model constructed based on human performance. Our approach is concise and requires no additional prompt instructions or camera trajectories, helping and guiding ordinary users to enhance their composition skills.



Paperid:3351
Authors:Leitian Tao, Xuefeng Du, Sharon Li
Abstract:
Reward modeling, crucial for aligning large language models (LLMs) with human preferences, is often bottlenecked by the high cost of preference data. Existing textual data synthesis methods are computationally expensive. We propose a novel framework for synthesizing preference data directly in the LLM's latent embedding space. Our method employs a Variational Autoencoder (VAE) to learn a structured latent representation of response embeddings. By performing controlled perturbations in this latent space and decoding back to the embedding space, we efficiently generate diverse, semantically consistent synthetic preference pairs, bypassing costly text generation and annotation. We provide theoretical guarantees that our synthesized pairs approximately preserve original preference ordering and improve reward model generalization. Empirically, our latent-space synthesis significantly outperforms text-based augmentation on standard benchmarks, achieving superior results while being 18× faster in generation and using a 16,000× smaller model. Our work offers a scalable and effective alternative for enhancing reward modeling through efficient data augmentation.



Paperid:3352
Authors:Peiwen Yuan, Yiwei Li, Shaoxiong Feng, Xinglin Wang, Yueqi Zhang, Jiayi Shi, Chuyi Tan, Boyuan Pan, Yao Hu, Prof. Kan
Abstract:
LLM-as-Benchmark-Generator methods have been widely studied as a supplement to human annotators for scalable evaluation, while the potential biases within this paradigm remain underexplored. In this work, we systematically define and validate the phenomenon of inflated performance in models evaluated on their self-generated benchmarks, referred to as self-bias, and attribute it to sub-biases arising from question domain, language style, and wrong labels.On this basis, we propose Silencer, a general framework that leverages the heterogeneity between multiple generators at both the sample and benchmark levels to neutralize bias and generate high-quality, self-bias-silenced benchmark. Experimental results across various settings demonstrate that Silencer can suppress self-bias to near zero, significantly improve evaluation effectiveness of the generated benchmark (with an average improvement from 0.655 to 0.833 in Pearson correlation with high-quality human-annotated benchmark), while also exhibiting strong generalizability.



Paperid:3353
Authors:Lorenzo Guerra, Thomas Chapuis, Guillaume Duc, Pavlo Mozharovskyi, Van-Tam Nguyen
Abstract:
Detecting intrusions in network traffic is a challenging task, particularly under limited supervision and constantly evolving attack patterns. While recent works have leveraged graph neural networks for network intrusion detection, they often decouple representation learning from anomaly detection, limiting the utility of the embeddings for identifying attacks. We propose GraphIDS, a self-supervised intrusion detection model that unifies these two stages by learning local graph representations of normal communication patterns through a masked autoencoder. An inductive graph neural network embeds each flow with its local topological context to capture typical network behavior, while a Transformer‑based encoder-decoder reconstructs these embeddings, implicitly learning global co-occurrence patterns via self-attention without requiring explicit positional information. During inference, flows with unusually high reconstruction errors are flagged as potential intrusions. This end-to-end framework ensures that embeddings are directly optimized for the downstream task, facilitating the recognition of malicious traffic. On diverse NetFlow benchmarks, GraphIDS achieves up to 99.98% PR‑AUC and 99.61% macro F1-score, outperforming baselines by 5–25 percentage points.



Paperid:3354
Authors:guojun lei, Rong Zhang, Chi Wang, Tianhang Liu, Hong Li, Zhiyuan Ma, Weiwei Xu
Abstract:
Recent advancements in video generation models have enabled the creation of diverse and realistic videos, with promising applications in advertising and film production. However, as one of the essential tasks of video generation models, video concept transfer remains significantly challenging.Existing methods generally model video as an entirety, leading to limited flexibility and precision when solely editing specific regions or concepts. To mitigate this dilemma, we propose a novel architecture UniTransfer, which introduces both spatial and diffusion timestep decomposition in a progressive paradigm, achieving precise and controllable video concept transfer. Specifically, in terms of spatial decomposition, we decouple videos into three key components: the foreground subject, the background, and the motion flow. Building upon this decomposed formulation, we further introduce a dual-to-single-stream DiT-based architecture for supporting fine-grained control over different components in the videos. We also introduce a self-supervised pretraining strategy based on random masking to enhance the decomposed representation learning from large-scale unlabeled video data. Inspired by the Chain-of-Thought reasoning paradigm, we further revisit the denoising diffusion process and propose a Chain-of-Prompt (CoP) mechanism to achieve the timestep decomposition. We decompose the denoising process into three stages of different granularity and leverage large language models (LLMs) for stage-specific instructions to guide the generation progressively. We also curate an animal-centric video dataset called OpenAnimal to facilitate the advancement and benchmarking of research in video concept transfer. Extensive experiments demonstrate that our method achieves high-quality and controllable video concept transfer across diverse reference images and scenes, surpassing existing baselines in both visual fidelity and editability.



Paperid:3355
Authors:Deep Patel, Emmanouil-Vasileios Vlatakis-Gkaragkounis
Abstract:
Many emerging applications—such as adversarial training, AI alignment, and robust optimization—can be framed as zero-sum games between neural nets, with von Neumann–Nash equilibria (NE) capturing the desirable system behavior. While such games often involve non-convex non-concave objectives, empirical evidence shows that simple gradient methods frequently converge, suggesting a hidden geometric structure. In this paper, we provide a theoretical framework that explains this phenomenon through the lens of \emph{hidden convexity} and \emph{overparameterization}. We identify sufficient conditions spanning initialization, training dynamics, and network width—that guarantee global convergence to a NE in a broad class of non-convex min-max games. To our knowledge, this is the first such result for games that involve two-layer neural networks. Technically, our approach is twofold: (a) we derive a novel path-length bound for alternating gradient-descent-ascent scheme in min-max games; and (b) we show that games with hidden convex–concave geometry reduce to settings satisfying two-sided Polyak–Łojasiewicz (PL) and smoothness conditions, which hold with high probability under overparameterization, using tools from random matrix theory.



Paperid:3356
Authors:Yang Han, Pengyu Wang, Kai Yu, xin chen, Lu Chen
Abstract:
Mass spectrometry (MS) plays a critical role in molecular identification, significantly advancing scientific discovery. However, structure elucidation from MS data remains challenging due to the scarcity of annotated spectra. While large-scale pretraining has proven effective in addressing data scarcity in other domains, applying this paradigm to mass spectrometry is hindered by the complexity and heterogeneity of raw spectral signals.To address this, we propose MS-BART, a unified modeling framework that maps mass spectra and molecular structures into a shared token vocabulary, enabling cross-modal learning through large-scale pretraining on reliably computed fingerprint–molecule datasets. Multi-task pretraining objectives further enhance MS-BART's generalization by jointly optimizing denoising and translation task. The pretrained model is subsequently transferred to experimental spectra through finetuning on fingerprint predictions generated with MIST, a pre-trained spectral interpretation model, thereby enhancing robustness to real-world spectral variability. While finetuning alleviates the distributional difference, MS-BART still suffers molecular hallucination and requires further alignment. We therefore introduce a chemical feedback mechanism that guides the model toward generating molecules closer to the reference structure.Extensive evaluations demonstrate that MS-BART achieves state-of-the-art performance across 10/12 key metrics on MassSpecGym and NPLIB1, while comprehensive ablation studies systematically validate the model's effectiveness and robustness.Code atAnonymous Link.



Paperid:3357
Authors:Ansong Ni, Ruta Desai, Yang Li, Xinjie Lei, Dong Wang, Jiemin Zhang, Jane Yu, Ramya Raghavendra, Gargi Ghosh, Shang-Wen Li, Asli Celikyilmaz
Abstract:
With increasingly powerful large language models (LLMs) and LLM-based agents tackling an ever-growing list of tasks, we envision a future where numerous LLM agents work seamlessly with other AI agents and humans to solve complex problems and enhance daily life. To achieve these goals, LLM agents must develop collaborative skills such as effective persuasion, assertion and disagreement, which are often overlooked in the prevalent single-turn training and evaluation of LLMs. In this work, we present Collaborative Reasoner (Coral), a framework to evaluate and improve the collaborative reasoning abilities of language models. In particular, tasks and metrics in Coral necessitate agents to disagree with incorrect solutions, convince their partners of a correct solution, and ultimately agree as a team to commit to a final solution, all through a natural multi-turn conversation. Through comprehensive evaluation on six collaborative reasoning tasks covering domains of coding, math, scientific QA and social reasoning, we show that current models cannot effectively collaborate due to undesirable social behaviors, collapsing even on problems that they can solve singlehandedly. To improve the collaborative reasoning capabilities of LLMs, we propose a self-play method to generate synthetic multi-turn preference data and further train the language models to be better collaborators. Experiments with Llama-3.1, Ministral and Qwen-2.5 models show that our proposed self-improvement approach consistently outperforms finetuned chain-of-thought performance of the same base model, yielding gains up to 16.7% absolute. Human evaluations show that the models exhibit more effective disagreement and produce more natural conversations after training on our synthetic interaction data.



Paperid:3358
Authors:Jiarui Fang, Jinzhe Pan, Aoyu Li, Xibo Sun, WANG Jiannan
Abstract:
Abstract:This paper presents PipeFusion, an innovative parallel methodology to tackle the high latency issues associated with generating high-resolution images using diffusion transformers (DiTs) models. PipeFusion partitions images into patches and the model layers across multiple GPUs. It employs a patch-level pipeline parallel strategy to orchestrate communication and computation efficiently. By capitalizing on the high similarity between inputs from successive diffusion steps, PipeFusion reuses one-step stale feature maps to provide context for the current pipeline step. This approach notably reduces communication costs compared to existing DiTs inference parallelism, including tensor parallel, sequence parallel and DistriFusion. PipeFusion also exhibits superior memory efficiency, because it can distribute model parameters across multiple devices, making it more suitable for DiTs with large parameter sizes, such as Flux.1. Experimental results demonstrate that PipeFusion achieves state-of-the-art performance on 8$\times$L40 PCIe GPUs for Pixart, Stable-Diffusion 3 and Flux.1 models.



Paperid:3359
Authors:Wei Feng, Zongyuan Ge
Abstract:
Generalized Category Discovery (GCD) aims to leverage labeled samples from known categories to cluster unlabeled data that may include both known and unknown categories. While existing methods have achieved impressive results under standard conditions, their performance often deteriorates in the presence of distribution shifts. In this paper, we explore a more realistic task: Domain-Shifted Generalized Category Discovery (DS_GCD), where the unlabeled data includes not only unknown categories but also samples from unknown domains. To tackle this challenge, we propose a \textbf{\underline{F}}requency-guided Gene\textbf{\underline{r}}alized Cat\textbf{\underline{e}}gory Discov\textbf{\underline{e}}ry framework (Free) that enhances the model's ability to discover categories under distributional shift by leveraging frequency-domain information. Specifically, we first propose a frequency-based domain separation strategy that partitions samples into known and unknown domains by measuring their amplitude differences. We then propose two types of frequency-domain perturbation strategies: a cross-domain strategy, which adapts to new distributions by exchanging amplitude components across domains, and an intra-domain strategy, which enhances robustness to intra-domain variations within the unknown domain. Furthermore, we extend the self-supervised contrastive objective and semantic clustering loss to better guide the training process. Finally, we introduce a clustering-difficulty-aware resampling technique to adaptively focus on harder-to-cluster categories, further enhancing model performance. Extensive experiments demonstrate that our method effectively mitigates the impact of distributional shifts across various benchmark datasets and achieves superior performance in discovering both known and unknown categories.



Paperid:3360
Authors:Tomas Soucek, Sylvestre-Alvise Rebuffi, Pierre Fernandez, Nikola Jovanović, Hady Elsahar, Valeriu Lacatusu, Tuan Tran, Alexandre Mourachko
Abstract:
Recent years have seen a surge in interest in digital content watermarking techniques due to the explosion of generative models, as well as increased legal pressure. With an ever-growing percentage of AI-generated content available online, watermarking plays an increasingly important role in ensuring content authenticity and attribution at scale. There have been many works assessing the robustness of watermarking to removal attacks, yet, watermark forging, the scenario when a watermark is stolen from genuine content and applied to malicious content, remains underexplored. In this work, we investigate watermark forging in the context of widely used post-hoc image watermarking. Our contributions are as follows. First, we introduce a preference model to assess whether an image is watermarked. The model is trained using a ranking loss on purely procedurally generated images without any need for real watermarks. Second, we show the capability of this model to remove and forge watermarks by optimizing the input image via backpropagation. This technique requires only a single watermarked image and works without knowledge of the watermarking model, making our attack much simpler and more practical than attacks introduced in related work. Third, we evaluate our proposed method on a variety of post-hoc image watermarking models, demonstrating that our approach can effectively forge watermarks, questioning the security of current watermarking approaches. Our code and further resources will be made publicly available.



Paperid:3361
Authors:Leonie Bossemeyer, Samuel Heinrich, Grant Van Horn, Oisin Mac Aodha
Abstract:
Fine-grained visual recognition, essential in many expert domains, often requires specialists years of dedicated training. Modeling the progression of such expertize remains challenging, and accurately inferring a learner’s knowledge state is a key step toward understanding visual learning. We introduce CleverBirds, a large-scale knowledge tracing benchmark for fine-grained bird recognition. Collected by the citizen-science platform eBird, it offers insight into how individuals acquire expertize in complex fine-grained classification tasks. More than 40,000 participants have engaged in the quiz, answering over 17 million multiple-choice questions spanning 10,000+ bird species, with long-range learning patterns across an average of 400+ questions per participant. We release this dataset to support the development and evaluation of new methods for visual knowledge tracing. We show that tracking learners' knowledge is challenging, especially across participant subgroups and question types, with different forms of contextual information offering varying degrees of predictive benefit. CleverBirds is among the largest benchmarks of its kind, offering a substantially higher number of learnable concepts. With it, we hope to enable new avenues for studying the development of visual expertize over time and across individuals.



Paperid:3362
Authors:Yilang Zhang, Bingcong Li, Georgios Giannakis
Abstract:
Low-Rank Adaptation (LoRA) lowers the computational and memory overhead of fine-tuning large models by updating a low-dimensional subspace of the pre-trained weight matrix. Albeit efficient, LoRA exhibits suboptimal convergence and noticeable performance degradation, due to inconsistent and imbalanced weight updates induced by its nonunique low-rank factorizations. To overcome these limitations, this article identifies the optimal low-rank factorization per step that minimizes an upper bound on the loss. The resultant refactored low-rank adaptation (RefLoRA) method promotes a flatter loss landscape, along with consistent and balanced weight updates, thus speeding up stable convergence. Extensive experiments evaluate RefLoRA on natural language understanding, and commonsense reasoning tasks with popular large language models including DeBERTaV3, LLaMA-7B, LLaMA2-7B and LLaMA3-8B. The numerical tests corroborate that RefLoRA converges faster, outperforms various benchmarks, and enjoys negligible computational overhead compared to state-of-the-art LoRA variants.



Paperid:3363
Authors:Weipeng Zhong, Peizhou Cao, Yichen Jin, Luo Li, Wenzhe Cai, Jingli Lin, Hanqing Wang, Zhaoyang Lyu, Tai WANG, Xudong XU, Bo Dai, Jiangmiao Pang
Abstract:
The advancement of Embodied AI heavily relies on large-scale, simulatable 3D scene datasets characterized by scene diversity and realistic layouts. However, existing datasets typically suffer from limitations in diversity or simulatability, sanitized layouts lacking small items, and severe object collisions. To address these shortcomings, we introduce \textbf{OmniScenes}, a novel large-scale simulatable indoor scene dataset comprising approximately 40,000 diverse scenes by integrating three disparate scene sources, \ie, real-world scans, procedurally generated scenes, and designer-created scenes, including 1.96M objects and 800k CAD models that cover 15 common scene types and 288 object classes, resulting in complex layouts that have most-ever 41.5 objects per region in average. Our comprehensive data processing pipeline ensures simulatability by creating real-to-sim replicas for real-world scans, achieves realistic layouts by preserving small items, and enhances interactivity by incorporating interactive objects and resolving collisions. We demonstrate the value of OmniScenes with two benchmark applications: scene layout generation and point-goal navigation. Both show the new challenges posed by the complex and realistic layouts. More importantly, OmniScenes paves the way for scaling up the model training for both tasks, making the generation and navigation in such complex scenes possible. We commit to open-sourcing the data, models, and benchmarks to benefit the community.



Paperid:3364
Authors:Jingyuan He, Jiongnan Liu, Vishan Oberoi, Bolin Wu, Mahima Jagadeesh Patel, Kangrui Mao, Chuning Shi, I-Ta Lee, Arnold Overwijk, Chenyan Xiong
Abstract:
Recommender systems are among the most impactful AI applications, interacting with billions of users every day, guiding them to relevant products, services, or information tailored to their preferences.However, the research and development of recommender systems are hindered by existing datasets that fail to capture realistic user behaviors and inconsistent evaluation settings that lead to ambiguous conclusions.This paper introduces the \textbf{O}pen \textbf{R}ecommendation \textbf{B}enchmark for Reproducible Research with H\textbf{I}dden \textbf{T}ests (\textbf{ORBIT}), a unified benchmark for consistent and realistic evaluation of recommendation models. ORBIT offers a standardized evaluation framework of public datasets with reproducible splits and transparent settings for its public leaderboard. Additionally, ORBIT introduces a new webpage recommendation task, ClueWeb-Reco, featuring web browsing sequences from 87 million webpages derived from real, user-consented, and privacy-guaranteed browsing data. ClueWeb-Reco aligns with modern recommendation scenarios and is reserved as the hidden test part of our leaderboard to challenge recommendation models' generalization ability. ORBIT measures 12 representative recommendation models on its public benchmark and introduce a prompted LLM baseline on the ClueWeb-Reco hidden test.Our benchmark results reflect general improvements of recommender systems on the public datasets, with variable individual performances.The results on the hidden test reveal the limitations of existing approaches in large-scale webpage recommendation and highlight the potential for improvements with LLM integrations.ORBIT benchmark and leaderboard are available at \url{https://www.open-reco-bench.ai}.



Paperid:3365
Authors:Zeyuan Yin, Xiaoming Liu
Abstract:
Abstract:Recent advances in 3D Gaussian diffusion models suffer from time-intensive denoising and post-denoising processing due to the massive number of Gaussian primitives, resulting in slow generation and limited scalability along sampling trajectories.To improve the efficiency of 3D diffusion models, we propose $\textbf{TRIM}$ ($\textbf{T}$rajectory $\textbf{R}$eduction and $\textbf{I}$nstance $\textbf{M}$ask denoising), a training-free approach that incorporates both temporal and spatial trimming strategies, to accelerate inference without compromising output quality while supporting the inference-time scaling for Gaussian diffusion models.Instead of scaling denoising trajectories in a costly end-to-end manner, we develop a lightweight selector model to evaluate latent Gaussian primitives derived from multiple sampled noises, enabling early trajectory reduction by selecting candidates with high-quality potential.Furthermore, we introduce instance mask denoising to prune learnable Gaussian primitives by filtering out redundant background regions, reducing inference computation at each denoising step.Extensive experiments and analysis demonstrate that TRIM significantly improves both the efficiency and quality of 3D generation.



Paperid:3366
Authors:Ling Yang, Xinchen Zhang, Ye Tian, Shiyi Zhang, Chenming Shang, Minghao Xu, Wentao Zhang, Bin CUI
Abstract:
The remarkable success of the autoregressive paradigm has made significant advancement in Multimodal Large Language Models (MLLMs), with powerful models like Show-o, Transfusion and Emu3 made notable strides in unified image understanding and generation. For the first time, we uncover a common phenomenon: the understanding capability of MLLMs is usually stronger than their generative capability, with a significant gap between them. Building on this insight, we propose HermesFlow, a simple and general framework designed to seamlessly bridge the gap between understanding and generation in MLLMs. Specifically, we take the homologous data as input to curate homologous preference data of both understanding and generation. Through Pair-DPO and self-play iterative optimization, HermesFlow effectively aligns multimodal understanding and generation using homologous preference data. Extensive experiments demonstrate the significant superiority of our approach over prior methods, particularly in narrowing the gap between multimodal understanding and generation. These findings highlight the potential of HermesFlow as a general alignment framework for next-generation multimodal foundation models.



Paperid:3367
Authors:Tonghan Wang, Yanchen Jiang, David Parkes
Abstract:
Abstract:Differentiable economics---the use of deep learning for auction design---has driven progress in multi-item auction design with additive or unit-demand valuations. However, progress in combinatorial auctions (CAs), even for the simplest and yet important single bidder case, remains limited due to the exponential growth of the bundle space with the number of items. We address this challenge by introducing a deep menu-based mechanism, which is, to our knowledge, the first dominant-strategy incentive compatible (DSIC) and revenue-optimizing single-bidder CA. Our idea is to generate a bundle distribution through an ordinary differential equation (ODE) applied to a tractable initial distribution. Our method learns suitable ODE-based transforms, one for each menu element, to optimize expected revenue. Our method achieves 1.11$-$2.23$\times$ higher revenue than baselines on standard CA testbeds and scales up to 150 items. Relative to menu-learning baselines that we introduce, our method also reduces training iterations by 3.6$-$9.5$\times$ and cuts training time by about 80\% in settings with 50 and 100 items.



Paperid:3368
Authors:Jitesh Jain, Zhengyuan Yang, Humphrey Shi, Jianfeng Gao, Jianwei Yang
Abstract:
In recent times, the standard practice for developing MLLMs is to feed features from vision encoder(s) into the LLM and train with natural language supervision. This approach often causes models to lean towards language comprehension and undermine the rich visual perception signals present in the data, which are critical for tasks involving spatial reasoning in the domain of embodied AI and robotics. Is it possible to optimize both at the same time? In this work, we propose VisPer-LM, the first approach that infuses visual perception knowledge from expert vision encoders into the LLM's (of an MLLM) hidden representations. We start by investigating MLLMs trained solely with natural language supervision and identify a positive correlation between the quality of visual representations within these models and their downstream performance. Given this insight, we formulate the objective during the pretraining stage in MLLMs as a coupled optimization of predictive visual embedding and next (text) token prediction. Moreover, through extensive probing, we observe improved visual representation quality due to embedding optimization, underscoring the effectiveness of our probing setup. We demonstrate that our VisPer-LM outperforms the single and multi-encoder baselines, proving our approach's superiority over explicitly feeding the corresponding features to the LLM. In particular, VisPer-LM boosts performance by an average margin of up to 2.5% on various benchmarks, with a notable improvement of 8.7% on the Depth task in CV-Bench.



Paperid:3369
Authors:Xinyan Su, Zhiheng Zhang, Jiyan Qiu, Zhaojuan Yue, Jun Li
Abstract:
Influence Maximization (IM) is to identify the seed set to maximize information dissemination in a network. Elegant IM algorithms could naturally extend to cases where each node is equipped with a specific weight, reflecting individual gains to measure the node’s importance. In these prevailing literatures, they typically assume such individual gains remain constant throughout the cascade process and are solvable through explicit formulas based on the node’s characteristics and network topology. However, this assumption is not always feasible due to two reasons: 1) Unobservability: The individual gains of each node are primarily evaluated by the difference between the outputs in the activated and non-activated states. In practice, we can only observe one of these states, with the other remaining unobservable post-propagation. 2) Environmental sensitivity: In addition to the node’s inherent properties, individual gains are also sensitive to the activation status of surrounding nodes, which is dynamic during iteration, even when the network topology remains static. To address these challenges, we extend the consideration of IM to a broader scenario with dynamic node individual gains, leveraging causality techniques. In our paper, we introduce a Causal Influence Maximization (CauIM) framework and develop two algorithms, G-CauIM and A-CauIM, where the latter incorporates a novel acceleration technique. Theoretically, for CauIM, we establish the generalized lower bound of influence spread and provide robustness analysis. Empirically, in synthetic and real-world experiments, we demonstrate the effectiveness and robustness of CauIM.



Paperid:3370
Authors:Diana Cai, Robert Gower, David Blei, Lawrence Saul
Abstract:
Abstract:We introduce a highly expressive yet distinctly tractable family for black box variational inference (BBVI). Each member of this family is a weighted product of experts (PoE), and each weighted expert in the product is proportional to a multivariate $t$-distribution. These products of experts can model distributions with skew, heavy tails, and multiple modes, but to use them for BBVI, we must be able to sample from their densities. We show how to do this by reformulating these products of experts as latent variable models with auxiliary Dirichlet random variables. These Dirichlet variables emerge from a Feynman identity, originally developed for loop integrals in quantum field theory, that expresses the product of multiple fractions (or in our case, $t$-distributions) as an integral over the simplex. We leverage this simplicial latent space to draw weighted samples from these products of experts---samples which BBVI then uses to find the PoE that best approximates a target density. Given a collection of experts, we derive an iterative procedure to optimize the exponents that determine their geometric weighting in the PoE. At each iteration, this procedure minimizes a regularized Fisher divergence to match the scores of the variational and target densities at a batch of samples drawn from the current approximation. This minimization reduces to a convex quadratic program, and we prove under general conditions that these updates converge exponentially fast to a nearly optimal weighting of experts. We conclude by evaluating this approach on a variety of synthetic and real-world target distributions.



Paperid:3371
Authors:Adrian Bulat, Yassine Ouali, Georgios Tzimiropoulos
Abstract:
This work aims to compress the vision tokens of an LVLM into a representation that is simultaneously suitable for (a) generative and (b) discriminative tasks, (c) is nearly lossless, and (d) storage-efficient. To this end, we propose C&C, a novel compression method that leverages the LVLM itself for task-agnostic visual token compression.Unlike prior methods that perform token reduction on-the-fly, our approach offloads computation to a dedicated, upfront indexing stage, effectively decoupling compression from generation. This enables learning more powerful representations for generation during inference. At the core of C&C is a ``double-forward pass'' training strategy. During the first forward pass, the LLM (of the LVLM) creates a bottleneck by compressing the dense visual tokens into a few summary tokens. Subsequently, the second forward pass processes the language instruction(s) alongside the summary tokens, used as a direct replacement for the image ones. The training of C&C is guided by two key losses: an autoregressive loss applied after the second pass that provides a direct optimization objective for reconstructing the original information flow, and a contrastive loss applied after the first pass to bolster the representational strength of the summary tokens, particularly for discriminative tasks. Moreover, we propose stage-specific adapters for further enhancing performance. C&C produces highly informative compressed representations. An in-depth ablation study confirms the efficacy of our approach. For generative tasks, we achieve a 2x higher compression rate without compromising capabilities, setting a new state-of-the-art. For discriminative tasks, we establish new state-of-the-art results on image retrieval and compositionality benchmarks.



Paperid:3372
Authors:Zixian Yang, Sushil Varma, Lei Ying
Abstract:
Abstract:We study a two-sided market, wherein, price-sensitive heterogeneous customers and servers arrive and join their respective queues.A compatible customer-server pair can then be matched by the platform, at which point, they leave the system.Our objective is to design pricing and matching algorithms that maximize the platform's profit, while maintaining reasonable queue lengths.As the demand and supply curves governing the price-dependent arrival rates may not be known in practice, we design a novel online-learning-based pricing policy and establish its near-optimality. In particular, we prove a tradeoff among three performance metrics: $\tilde{O}(T^{1-\gamma})$ regret, $\tilde{O}(T^{\gamma/2})$ average queue length, and $\tilde{O}(T^{\gamma})$ maximum queue length for $\gamma \in (0, 1/6]$, significantly improving over existing results (Yang & Ying, 2024). Moreover, barring the permissible range of $\gamma$, we show that this trade-off between regret and average queue length is optimal up to logarithmic factors under a class of policies, matching the optimal one as in (Varma et al., 2023) which assumes the demand and supply curves to be known.Our proposed policy has two noteworthy features: a dynamic component that optimizes the tradeoff between low regret and small queue lengths; and a probabilistic component that resolves the tension between obtaining useful samples for fast learning and maintaining small queue lengths.



Paperid:3373
Authors:Félix Lefebvre, Gael Varoquaux
Abstract:
Many machine learning tasks can benefit from external knowledge. Large knowledge graphs store such knowledge, and embedding methods can be used to distill it into ready-to-use vector representations for downstream applications. For this purpose, current models have however two limitations: they are primarily optimized for link prediction, via local contrastive learning, and they struggle to scale to the largest graphs due to GPU memory limits. To address these, we introduce SEPAL: a Scalable Embedding Propagation ALgorithm for large knowledge graphs designed to produce embeddings for downstream tasks.The key idea of SEPAL is to optimize embeddings only on a small core of entities, and then use message passing to propagate embeddings outside the core, ensuring global embedding alignment.We evaluate SEPAL on 7 large-scale knowledge graphs and 46 downstream machine learning tasks. Our results show that SEPAL significantly outperforms previous methods on downstream tasks. In addition, SEPAL scales up its base embedding model, enabling fitting huge knowledge graphs on commodity hardware.



Paperid:3374
Authors:Murtaza Nazir, Matthew Finlayson, John Morris, Xiang Ren, Swabha Swayamdipta
Abstract:
Language model inversion seeks to recover hidden prompts using only language model outputs. This capability has implications for security and accountability in language model deployments, such as leaking private information from an API-protected language model’s system message. We propose a new method – prompt inversion from logprob sequences (PILS) – that recovers hidden prompts by gleaning clues from the model’s next-token probabilities over the course of multiple generation steps. Our method is enabled by a key insight: The vector-valued outputs of a language model occupy a low-dimensional subspace. This enables us to losslessly compress the full next-token probability distribution over multiple generation steps using a linear map, allowing more output information to be used for inversion. Our approach yields massive gains over previous state-of-the-art methods for recovering hidden prompts, achieving 2–3.5 times higher exact recovery rates across test sets, in one case increasing the recovery rate from 17% to 60%. Our method also exhibits surprisingly good generalization behavior; for instance, an inverter trained on 16 generations steps gets 5–27% higher prompt recovery when we increase the number of steps to 32 at test time. Furthermore, we demonstrate strong performance of our method on the more challenging task of recovering hidden system messages. We also analyze the role of verbatim repetition in prompt recovery and propose a new method for cross-family model transfer for logit-based inverters. Our findings suggest that next-token probabilities are a considerably more vulnerable attack surface for inversion attacks than previously known.



Paperid:3375
Authors:Binxu Wang, Cengiz Pehlevan
Abstract:
We develop an analytical framework for understanding how the learned distribution evolves during diffusion model training. Leveraging the Gaussian equivalence principle, we derived exact solutions for the gradient-flow dynamics of weights in one or two layer linear or linear convolutional denoiser settings with arbitrary data, where linear networks converge along principal components, and convolutional networks converge along Fourier modes. Remarkably, these solutions allow us to derive the generated distribution in closed-form and its KL-divergence through training. These analytical results expose a pronounced \emph{spectral bias}, i.e. for both weights and generated distributions, the convergence time of a mode follows an inverse power law of its variance. Empirical experiments on both Gaussian and natural image datasets demonstrate that the power-law spectral bias—remain robust even when using deeper or convolutional architectures. Our results underscore the importance of the data covariance in dictating the order and rate at which diffusion models learn different modes of the data, providing potential explanations of why earlier stopping could lead to incorrect details in image generative model.



Paperid:3376
Authors:Jennifer Mickel, Maria De-Arteaga, Liu Leqi, Kevin Tian
Abstract:
To recognize and mitigate the harms of generative AI systems, it is crucial to consider who is represented in the outputs of generative AI systems and how people are represented. A critical gap emerges when naively improving who is represented, as this does not imply bias mitigation efforts have been applied to address how people are represented. We critically examined this by investigating gender representation in occupation across state-of-the-art large language models. We first show evidence suggesting that over time, there have been interventions to models altering the resulting gender distribution, and we find that women are more represented than men when models are prompted to generate biographies or personas. We then develop a methodology for understanding representational differences between groups. We use this methodology to demonstrate that representational biases persist in how different genders are represented by examining statistically significant word differences across genders. This results in a proliferation of these representational differences, some of which are associated with representational harms and stereotypes, despite existing interventions to increase female representation, which can reinforce these harms.



Paperid:3377
Authors:Jinwei Gan, Zifeng Cheng, Zhiwei Jiang, Cong Wang, Yafeng Yin, Xiang Luo, Yuchen Fu, Qing Gu
Abstract:
Large language models (LLMs) have achieved remarkable performance across many generation tasks.However, effectively aligning them with desired behaviors remains a significant challenge.Activation steering is an effective and cost-efficient approach that directly modifies the activations of LLMs during the inference stage, aligning their responses with the desired behaviors and avoiding the high cost of fine-tuning.Existing methods often rigidly intervene at the first token of all responses, causing the LLM's replies to deviate from their original output and leading to homogenized and less informative responses.To this end, we propose aFlexibleActivationSteering withBacktracking (FASB) framework that flexibly determines whether intervention is needed, when to intervene, and the intervention strength by probing the internal states.The backtracking mechanism enables the correction of deviated tokens and helps steer the model toward the desired behavior.Extensive experiments on the TruthfulQA dataset and six multiple-choice datasets demonstrate that our method outperforms baselines.



Paperid:3378
Authors:Rafael Elberg, Felipe del Río, Mircea Petrache, Denis Parra
Abstract:
Compositional generalization—the ability to interpret novel combinations of familiar elements—is a hallmark of human cognition and language. Despite recent advances, deep neural networks still struggle to acquire this property reliably. In this work, we introduce CELEBI (Compressive-Expressive Language Emergence through a discrete Bottleneck and Iterated learning), a novel self-supervised framework for inducing compositionality in learned representations from pre-trained models, through a reconstruction-based communication game between a sender and a receiver. Building on theories of language emergence, we integrate three mechanisms that jointly promote compressibility, expressivity, and efficiency in the emergent language. First, interactive decoding incentivizes intermediate reasoning by requiring the receiver to produce partial reconstructions after each symbol. Second, a reconstruction-based imitation phase, inspired by iterated learning, trains successive generations of agents to imitate reconstructions rather than messages, enforcing a tighter communication bottleneck. Third, pairwise distance maximization regularizes message diversity by encouraging high distances between messages, with formal links to entropy maximization. Our method significantly improves both the efficiency and compositionality of the learned messages on the Shapes3D and MPI3D datasets, surpassing prior discrete communication frameworks in both reconstruction accuracy and topographic similarity. This work provides new theoretical and empirical evidence for the emergence of structured, generalizable communication protocols from simplicity-based inductive biases.



Paperid:3379
Authors:Rikuto Kotoge, Zheng Chen, Tasuku Kimura, Yasuko Matsubara, Takufumi Yanagisawa, Haruhiko Kishima, Yasushi Sakurai
Abstract:
Dynamic GNNs, which integrate temporal and spatial features in Electroencephalography (EEG) data, have shown great potential in automating seizure detection.However, fully capturing the underlying dynamics necessary to represent brain states, such as seizure and non-seizure, remains a non-trivial task and presents two fundamental challenges.First, most existing dynamic GMN methods are built on temporally fixed static graphs, which fail to reflect the evolving nature of brain connectivity during seizure progression. Second, current efforts to jointly model temporal signals and graph structures and, more importantly, their interactions remain nascent, often resulting in inconsistent performance.To address these challenges, we present the first theoretical analysis of these two problems, demonstrating the effectiveness and necessity of explicit dynamic modeling and time-then-graph dynamic GNN method.Building on these insights, we propose EvoBrain, a novel seizure detection model that integrates a two-stream Mamba architecture with a GCN enhanced by Laplacian Positional Encoding, following neurological insights.Moreover, EvoBrain incorporates explicitly dynamic graph structures, allowing both nodes and edges to evolve over time.Our contributions include (a) a theoretical analysis proving the expressivity advantage of explicit dynamic modeling and time-then-graph over other approaches, (b) a novel and efficient model that significantly improves AUROC by 23\% and F1 score by 30\%, compared with the dynamic GNN baseline, and (c) broad evaluation of our method on the challenging early seizure prediction tasks.The source code is available at:\url{https://anonymous.4open.science/r/EvoBrain-FBC5}



Paperid:3380
Authors:Yu Guo, Shengfeng He, Yuxu Lu, Haonan An, Yihang Tao, Huilin Zhu, Jingxian Liu, Yuguang Fang
Abstract:
Maritime object detection is essential for navigation safety, surveillance, and autonomous operations, yet constrained by two key challenges: the scarcity of annotated maritime data and poor generalization across various maritime attributes (e.g., object category, viewpoint, location, and imaging environment). To address these challenges, we propose Neptune-X, a data-centric generative-selection framework that enhances training effectiveness by leveraging synthetic data generation with task-aware sample selection. From the generation perspective, we develop X-to-Maritime, a multi-modality-conditioned generative model that synthesizes diverse and realistic maritime scenes. A key component is the Bidirectional Object-Water Attention module, which captures boundary interactions between objects and their aquatic surroundings to improve visual fidelity. To further improve downstream tasking performance, we propose Attribute-correlated Active Sampling, which dynamically selects synthetic samples based on their task relevance. To support robust benchmarking, we construct the Maritime Generation Dataset, the first dataset tailored for generative maritime learning, encompassing a wide range of semantic conditions. Extensive experiments demonstrate that our approach sets a new benchmark in maritime scene synthesis, significantly improving detection accuracy, particularly in challenging and previously underrepresented settings.



Paperid:3381
Authors:Zhiqin Yang, Yonggang Zhang, Chenxin Li, Yiu-ming Cheung, Bo Han, Yixuan Yuan
Abstract:
Federated Learning (FL) confronts a significant challenge known as data heterogeneity, which impairs model performance and convergence. Existing methods have made notable progress in addressing this issue. However, improving performance in certain heterogeneity scenarios remains an overlooked question:How robust are these methods to deploy under diverse heterogeneity scenarios?To answer this, we conduct comprehensive evaluations across varied heterogeneity scenarios, showing that most existing methods exhibit limited robustness. Meanwhile, insights from these experiments highlight that sharing statistical information can mitigate heterogeneity by enabling clients to update with a global perspective. Motivated by this, we proposeFedGPS(FederatedGoal-PathSynergy), a novel framework that seamlessly integrates statistical distribution and gradient information from others. Specifically, FedGPS statically modifies each client’s learning objective to implicitly model the global data distribution using surrogate information, while dynamically adjusting local update directions with non-self gradient information at every update step. Extensive experiments show that FedGPS outperforms existing methods across diverse heterogeneity scenarios, validating the effectiveness and robustness of FedGPS under different datasets.



Paperid:3382
Authors:Jiani Liu, Zhiyuan Wang, Zeliang Zhang, Chao Huang, Susan Liang, Yunlong Tang, Chenliang Xu
Abstract:
Vision Transformers (ViTs) have demonstrated impressive performance across a range of applications, including many safety-critical tasks. Many previous studies have observed that adversarial examples crafted on ViTs exhibit higher transferability than those crafted on CNNs, indicating that ViTs contain structural characteristics favorable for transferable attacks. In this work, we take a further step to deeply investigate the role of computational redundancy brought by its unique characteristics in ViTs and its impact on adversarial transferability. Specifically, we identify two forms of redundancy, including the data-level and model-level, that can be harnessed to amplify attack effectiveness. Building on this insight, we design a suite of techniques, including attention sparsity manipulation, attention head permutation, clean token regularization, ghost MoE diversification, and learn to robustify before the attack. A dynamic online learning strategy is also proposed to fully leverage these operations to enhance the adversarial transferability. Extensive experiments on the ImageNet-1k dataset validate the effectiveness of our approach, showing that our methods significantly outperform existing baselines in both transferability and generality across diverse model architectures, including different variants of ViTs and mainstream Vision Large Language Models (VLLMs).



Paperid:3383
Authors:Diqi He, Xuehao Gao, Hao Li, Dingwen Zhang, Junwei Han
Abstract:
The Zero-shot Vision-and-Language Navigation in Continuous Environments (VLN-CE) task requires agents to navigate previously unseen 3D environments using natural language instructions, without any scene-specific training. A critical challenge in this setting lies in ensuring agents’ actions align with both spatial structure and task intent over long-horizon execution. Existing methods often fail to achieve robust navigation due to a lack of structured decision-making and insufficient integration of feedback from previous actions. To address these challenges, we propose STRIDER (Instruction-Aligned Structural Decision Space Optimization), a novel framework that systematically optimizes the agent’s decision space by integrating spatial layout priors and dynamic task feedback. Our approach introduces two key innovations: 1) a Structured Waypoint Generator that constrains the action space through spatial structure, and 2) a Task-Alignment Regulator that adjusts behavior based on task progress, ensuring semantic alignment throughout navigation. Extensive experiments on the R2R-CE and RxR-CE benchmarks demonstrate that STRIDER significantly outperforms strong SOTA across key metrics; in particular, it improves Success Rate (SR) from 29\% to 35\%, a relative gain of 20.7\%. Such results highlight the importance of spatially constrained decision-making and feedback-guided execution in improving navigation fidelity for zero-shot VLN-CE.



Paperid:3384
Authors:Jiayi Yuan, Hao Li, Xinheng Ding, Wenya Xie, Yu-Jhe Li, Wentian Zhao, Kun Wan, Jing Shi, Xia Hu, Zirui Liu
Abstract:
Abstract:Large Language Models (LLMs) are now integral across various domains and have demonstrated impressive performance. Progress, however, rests on the premise that benchmark scores are both accurate and reproducible. We demonstrate that the reproducibility of LLM performance is fragile: changing system configuration such as evaluation batch size, GPU count, and GPU version can introduce significant difference in the generated responses. This issue is especially pronounced in reasoning models, where minor rounding differences in early tokens can cascade into divergent chains of thought, ultimately affecting accuracy. For instance, under bfloat16 precision with greedy decoding, a reasoning model like DeepSeek-R1-Distill-Qwen-7B can exhibit up to $\textbf{9\\% accuracy variance due to differences in GPU count, type, and evaluation batch size used.}$We trace the root cause of this variability to the non-associative nature of floating-point arithmetic under limited numerical precision. Our work presents the first systematic investigation into how numerical precision affects reproducibility in LLM inference. Through carefully controlled experiments across various hardware, software, and precision settings, we quantify when and how model outputs diverge.Our analysis reveals that floating-point precision—while critical for reproducibility—is often neglected in evaluation practices. Inspired by this, we develop a lightweight inference pipeline, dubbed $\textit{LayerCast}$, that stores weights in 16-bit precision but performs all computations in FP32, balancing memory efficiency with numerical stability.



Paperid:3385
Authors:Yanping Fu, Xinyuan Liu, Tianyu Li, Yike Ma, Yucheng Zhang, Feng Dai
Abstract:
Abstract:Topology reasoning, which unifies perception and structured reasoning, plays a vital role in understanding intersections for autonomous driving. However, its performance heavily relies on the accuracy of lane detection, particularly at connected lane endpoints. Existing methods often suffer from lane endpoints deviation, leading to incorrect topology construction. To address this issue, we propose TopoPoint, a novel framework that explicitly detects lane endpoints and jointly reasons over endpoints and lanes for robust topology reasoning. During training, we independently initialize point and lane query, and proposed Point-Lane Merge Self-Attention to enhance global context sharing through incorporating geometric distances between points and lanes as an attention mask . We further design Point-Lane Graph Convolutional Network to enable mutual feature aggregation between point and lane query. During inference, we introduce Point-Lane Geometry Matching algorithm that computes distances between detected points and lanes to refine lane endpoints, effectively mitigating endpoint deviation. Extensive experiments on the OpenLane-V2 benchmark demonstrate that TopoPoint achieves state-of-the-art performance in topology reasoning (48.8 on OLS). Additionally, we propose DET$_p$ to evaluate endpoint detection, under which our method significantly outperforms existing approaches (52.6 v.s. 45.2 on DET$_p$). The codes will be released soon.



Paperid:3386
Authors:Jiaxin Huang, Runnan Chen, Ziwen Li, Zhengqing Gao, Xiao He, Yandong Guo, Mingming Gong, Tongliang Liu
Abstract:
Reasoning segmentation aims to segment target objects in complex scenes based on human intent and spatial reasoning. While recent multimodal large language models (MLLMs) have demonstrated impressive 2D image reasoning segmentation, adapting these capabilities to 3D scenes remains underexplored. In this paper, we introduce MLLM-For3D, a simple yet effective framework that transfers knowledge from 2D MLLMs to 3D scene understanding. Specifically, we utilize MLLMs to generate multi-view pseudo-segmentation masks and corresponding text embeddings, then unproject 2D masks into 3D space and align them with the text embeddings. The primary challenge lies in the absence of 3D context and spatial consistency across multiple views, causing the model to hallucinate objects that do not exist and fail to target objects consistently. Training the 3D model with such irrelevant objects leads to performance degradation. To address this, we first filter irrelevant views using token attention. With these reliable pseudo-labels, we develop a token-for-Query approach for multimodal semantic alignment, enabling consistent identification of the same object across different views. Moreover, we introduce a spatial consistency strategy to enforce that segmentation masks remain coherent in the 3D space, effectively capturing the geometry of the scene. Extensive evaluations of various challenging indoor scene benchmarks demonstrate that, even without labeled 3D training data, MLLM-For3D outperforms existing 3D reasoning segmentation methods, effectively interpreting user intent, understanding 3D scenes, and reasoning about spatial relationships.



Paperid:3387
Authors:Yiyuan Pan, Yunzhe XU, Zhe Liu, Hesheng Wang
Abstract:
Visual navigation is essential for autonomous agents, yet data scarcity poses a fundamental bottleneck that severely impairs the generalization of learned policies to unseen scenarios. Existing visual navigation agents depend on complex predictive or reasoning modules that become counterproductive and data-hungry in such limited-data regimes. This paper introduces NeuRO, a novel hybrid framework that pioneers the integration of networks with downstream task-based optimization to tackle this critical problem. NeuRO addresses core difficulties in this integration: (i) it empowers the network to translate inherently unreliable visual predictions under data scarcity into calibrated convex uncertainty sets, which then directly inform and constrain the downstream optimization problem, using Partially Input Convex Neural Networks (PICNNs) via a conformal method; and (ii) it reformulates the partially observable task as a generalizable robust optimization problem to effectively leverage these uncertainty-aware representations to derive robust policies. Extensive experiments on both unordered and sequential MultiON tasks demonstrate that NeuRO establishes state-of-the-art performance, particularly in generalization to unseen environments. Our work thus presents a significant advancement for developing robust, generalizable autonomous agents.



Paperid:3388
Authors:Mianchu Wang, Giovanni Montana
Abstract:
Retrosynthesis planning aims to decompose target molecules into available building blocks, forming a synthesis tree where each internal node represents an intermediate compound and each leaf ideally corresponds to a purchasable reactant. However, this tree becomes invalid if any leaf node is not a valid building block, making the planning process vulnerable to the "weakest link" in the synthetic route. Existing methods often optimise for average performance across branches, failing to account for this worst-case sensitivity. In this paper, we reframe retrosynthesis as a worst-path optimisation problem within tree-structured Markov Decision Processes (MDPs). We prove that this formulation admits a unique optimal solution and offers monotonic improvement guarantees. Building on this insight, we introduce Interactive Retrosynthesis Planning (InterRetro), a method that interacts with the tree MDP, learns a value function for worst-path outcomes, and improves its policy through self-imitation, preferentially reinforcing past decisions with high estimated advantage. Empirically, InterRetro achieves state-of-the-art results, solving 100\% of targets on the Retro*-190 benchmark, shortening synthetic routes by 4.9\%, and achieving promising performance using only 10\% of the training data - representing a significant advance in computational retrosynthesis planning.



Paperid:3389
Authors:Fanxu Meng, Pingzhi Tang, Zengwei Yao, Xing Sun, Muhan Zhang
Abstract:
Modern large-language models often face communication bottlenecks on current hardware rather than computational limitations. Multi-head latent attention (MLA) addresses this by compressing the key-value cache using low-rank matrices, while the absorbed operation prevents the KV cache from reverting to its original size, significantly boosting both training and inference speed.Despite the success of DeepSeek V2/V3/R1, most model vendors have heavily invested in optimizing GQA-based models and therefore lack strong incentives to retrain MLA-based models from scratch.In this paper, we introduce TransMLA, a framework that seamlessly converts any GQA-based pre-trained model (e.g., LLaMA, Qwen, Mixtral) into an MLA-based model. For the first time, our method enablesdirect conversion of these models into a format compatible with DeepSeek's codebase, allowing them to fully leverage DeepSeek-specific optimizations such as vLLM and SGlang.By compressing 93% of the KV cache in LLaMA-2-7B, we achieve an10.6x speedupwith an 8K context length while maintaining meaningful output. Moreover, the model requires only6B tokensfor fine-tuning to recover comparable performance across multiple benchmarks.TransMLA provides a practical path for migrating GQA-based models to the MLA structure, and when combined with DeepSeek’s advanced optimizations—such as FP8 quantization and Multi-Token Prediction—further inference acceleration can be achieved.



Paperid:3390
Authors:Chongjun Tu, Lin Zhang, pengtao chen, Peng Ye, Xianfang Zeng, Wei Cheng, Gang Yu, Tao Chen
Abstract:
Multimodal Large Language Models (MLLMs) have shown impressive video content understanding capabilities but struggle with fine-grained motion comprehension. To comprehensively assess the motion understanding ability of existing MLLMs, we introduce FAVOR-Bench, which comprises 1,776 videos from both ego-centric and third-person perspectives and enables assessment through both close-ended and open-ended tasks. For close-ended evaluation, we carefully design 8,184 multiple-choice question-answer pairs spanning six distinct sub-tasks. For open-ended evaluation, we employ the GPT-assisted evaluation and develop a novel cost-efficient LLM-free assessment method, where the latter can enhance benchmarking interpretability and accessibility. Comprehensive experiments with21 state-of-the-art MLLMs reveal significant limitations in their ability to comprehend and describe detailed temporal dynamics in video motions. To alleviate this limitation, we further build FAVOR-Train, a dataset of 17,152 videos with fine-grained motion annotations. Finetuning Qwen2.5-VL on FAVOR-Train yields consistent improvements on motion-related tasks across TVBench, MotionBenchand our FAVOR-Bench. Our assessment results demonstrate that the proposed FAVOR-Bench and FAVOR-Train provide valuable tools for the community to develop more powerful video understanding models.



Paperid:3391
Authors:Div Garg, Diego Caples, Andis Draguns, Nikil Ravi, Pranav Putta, Naman Garg, Prannay Hebbar, Youngchul Joo, Jindong Gu, Charles London, Christian Schroeder de Witt, Sumeet Motwani
Abstract:
We introduce REAL, a benchmark and framework for multi-turn agent evaluations on deterministic simulations of real-world websites. REAL comprises high-fidelity, deterministic replicas of 11 widely-used websites across domains such as e-commerce, travel, communication, and professional networking. We also release a benchmark consisting of 112 practical tasks that mirror everyday complex user interactions requiring both accurate information retrieval and state-changing actions. All interactions occur within this fully controlled setting, eliminating safety risks and enabling robust, reproducible evaluation of agent capability and reliability. Our novel evaluation framework combines programmatic checks of website state for action-based tasks with rubric-guided LLM-based judgments for information retrieval. The framework supports both open-source and proprietary agent systems through a flexible evaluation harness that accommodates black-box commands within browser environments, allowing research labs to test agentic systems without modification. Our empirical results show that frontier language models achieve at most a 41% success rate on REAL, highlighting critical gaps in autonomous web navigation and task completion capabilities. Our framework supports easy integration of new tasks, reproducible evaluation, and scalable post-training data generation, marking a significant step forward in evaluating and advancing agent capabilities.



Paperid:3392
Authors:Tianle Pu, Zijie Geng, Haoyang Liu, Shixuan Liu, Jie Wang, Li Zeng, Chao Chen, Changjun Fan
Abstract:
Abstract:Mixed-Integer Linear Programming (MILP) is a fundamental and powerful framework for modeling complex optimization problems across diverse domains.Recently, learning-based methods have shown great promise in accelerating MILP solvers by predicting high-quality solutions.However, most existing approaches are developed and evaluated in single-domain settings, limiting their ability to generalize to unseen problem distributions.This limitation poses a major obstacle to building scalable and general-purpose learning-augmented solvers.To address this challenge, we introduce RoME, a domain-Robust Mixture-of-Experts (MoE) framework for predicting MILP solutions across domains.RoME dynamically routes problem instances to specialized experts based on learned task embeddings.The model is trained using a two-level distributionally robust optimization strategy: inter-domain to mitigate global shifts across domains, and intra-domain to enhance local robustness by introducing perturbations on task embeddings.We reveal that cross-domain training not only enhances the model's generalization capability to unseen domains but also improves performancewithin each individual domain by encouraging the model to capture more general intrinsic combinatorial patterns.Specifically, a single RoME model trained on three domains achieves an average improvement of $67.7\%$ then evaluated on five diverse domains.We further test the pretrained model on MIPLIB in a zero-shot setting, demonstrating its ability to deliver measurable performance gains on challenging real-world instances where existing learning-based approaches often struggle to generalize.



Paperid:3393
Authors:Shixuan Wang, Jingwen Ye, Xinchao Wang
Abstract:
Recent advances in generative models trained on large-scale datasets have enabled high-quality 3D synthesis across various domains. However, these models also raise critical privacy concerns. Unlike 2D image synthesis, where risks typically involve the leakage of visual features or identifiable patterns, 3D generation introduces additional challenges, as reconstructed shapes, textures, and spatial structures may inadvertently expose proprietary designs, biometric data, or other sensitive geometric information.This paper presents the first exploration of machine unlearning in 3D generation tasks. We investigate different unlearning objectives, including re-targeting and partial unlearning, and propose a novel framework that does not require full supervision of the unlearning target. To enable a more efficient unlearning process, we introduce a skip-acceleration mechanism, which leverages the similarity between multi-view generated images to bypass redundant computations. By establishing coherence across viewpoints during acceleration, our framework not only reduces computation but also enhances unlearning effectiveness, outperforming the non-accelerated baseline in both accuracy and efficiency.We conduct extensive experiments on the typical 3D generation models (Zero123 and Zero123XL), demonstrating that our approach achieves a 30\% speedup, while effectively unlearning target concepts without compromising generation quality. Our framework provides a scalable and practical solution for privacy-preserving 3D generation, ensuring responsible AI deployment in real-world applications.



Paperid:3394
Authors:Andrea Montanari, Pierfrancesco Urbani
Abstract:
Abstract:Understanding the inductive bias and generalization properties of large overparametrized machine learning models requires to characterize the dynamics of the training algorithm. We study the learning dynamics of large two-layer neural networks via dynamical mean field theory, a well established technique of non-equilibrium statistical physics.We show that, for large network width, the training dynamics exhibits a separation of timescales which implies:$(i)$ The emergence of a slow time scale associated with the growth in Gaussian/Rademacher complexity of the network;$(ii)$ Inductive bias towards small complexity if the initialization has small enough complexity;$(iii)$ A dynamical decoupling between feature learning and overfitting regimes;$(iv)$ A non-monotone behavior of the test error, associated `feature unlearning' regime at large times.



Paperid:3395
Authors:Siyang Jiang, Rui Fang, Hsi-Wen Chen, Wei Ding, Guoliang Xing, Ming-syan Chen
Abstract:
Abstract:Few-shot learning (FSL) aims to classify unseen examples (query set) into labeled data (support set) through low-dimensional embeddings. However, the diversity and unpredictability of environments and capture devices make FSL more challenging in real-world applications. In this paper, we propose $\textit{Dual Support Query Shift (DSQS)}$, a novel challenge in FSL that integrates two key issues: inter-set shifts (between support and query sets) and intra-set shifts (within each set), which significantly hinder model performance. To tackle these challenges, we introduce a $\textit{DUal ALignment framework (DUAL)}$, which core insight is that clean features can improve optimal transportation (OT) alignment. Firstly, DUAL leverages a robust embedding function enhanced by a repairer network trained with perturbed and adversarially generated "hard" examples to obtain clean features. Additionally, it incorporates a two-stage OT approach with a negative entropy regularizer, aligning support set instances, minimizing intra-class distances, and using query data as anchor nodes for achieving effective distribution alignment. We provide a theoretical bound of DUAL and experimental results on three image datasets, compared against 10 state-of-the-art baselines, showing that DUAL achieves a remarkable average performance improvement of $18.39\%$. Our code is available at https://anonymous.4open.science/r/CODE-F558/.



Paperid:3396
Authors:Aadyot Bhatnagar, Sarthak Jain, Joel Beazer, Samuel Curran, Alexander Hoffnagle, Kyle Ching, Michael Martyn, Stephen Nayfach, Jeffrey Ruffolo, Ali Madani
Abstract:
Generative protein language models (PLMs) are powerful tools for designing proteins purpose-built to solve problems in medicine, agriculture, and industrial processes.Recent work has trained ever larger language models, but there has been little systematic study of the optimal training distributions and the influence of model scale on the sequences generated by PLMs.We introduce the ProGen3 family of sparse generative PLMs, and we develop compute-optimal scaling laws to scale up to a 46B-parameter model pre-trained on 1.5T amino acid tokens.ProGen3's pre-training data is sampled from an optimized data distribution over the PPA v1, a carefully curated dataset of 3.4B full-length proteins. We evaluate for the first time in the wet lab the influence of model scale on the sequences generated by PLMs, and we find that larger models generate viable proteins for a much wider diversity of protein families.Finally, we find both computationally and experimentally that larger models are more responsive to alignment with laboratory data, resulting in improved protein fitness prediction and sequence generation capabilities.These results indicate that larger PLMs like ProGen3-46B trained on larger, well-curated datasets are powerful foundation models that push the frontier of protein design.



Paperid:3397
Authors:Shristi Das Biswas, Arani Roy, Kaushik Roy
Abstract:
Abstract:As Text-to-Image models continue to evolve, so does the risk of generating unsafe, copyrighted, or privacy-violating content. Existing safety interventions - ranging from training data curation and model fine-tuning to inference-time filtering and guidance - often suffer from incomplete concept removal, susceptibility to jail-breaking, computational inefficiency, or collateral damage to unrelated capabilities. In this paper, we introduce CURE, a training-free concept unlearning framework that operates directly in the weight space of pre-trained diffusion models, enabling fast, interpretable, and highly specific suppression of undesired concepts. At the core of our method is the Spectral Eraser, a closed-form, orthogonal projection module that identifies discriminative subspaces using Singular Value Decomposition over token embeddings associated with the concepts to forget and retain. Intuitively, the Spectral Eraser identifies and isolates features unique to the undesired concept while preserving safe attributes. This operator is then applied in a single step update to yield an edited model in which the target concept is effectively unlearned - without retraining, supervision, or iterative optimization. To balance the trade-off between filtering toxicity and preserving unrelated concepts, we further introduce an Expansion Mechanism for spectral regularization which selectively modulates singular vectors based on their relative significance to control the strength of forgetting. All the processes above are in closed-form, guaranteeing extremely efficient erasure in only $2$ seconds. Benchmarking against prior approaches, CURE achieves a more efficient and thorough removal for targeted artistic styles, objects, identities, or explicit content, with minor damage to original generation ability and demonstrates enhanced robustness against red-teaming. Code will be released.



Paperid:3398
Authors:Quang Truong, Zhikai Chen, Mingxuan Ju, Tong Zhao, Neil Shah, Jiliang Tang
Abstract:
Relational databases underpin critical infrastructure across a wide range of domains, yet the design of generalizable pre-training strategies for learning from relational databases remains an open challenge due to task heterogeneity. Specifically, there exist infinitely many possible downstream tasks, as tasks are defined based on relational schema graphs, temporal dependencies, and SQL-defined label logics. An effective pre-training framework is desired to take these factors into account in order to obtain task-aware representations. By incorporating knowledge of the underlying distribution that drives label generation, downstream tasks can benefit from relevant side-channel information. To bridge this gap, we introduce Task Vector Estimation (TVE), a novel pre-training framework that constructs predictive supervisory signals via set-based aggregation over schema traversal graphs, explicitly modeling next-window relational dynamics. We formalize our approach through an information-theoretic lens, demonstrating that task-informed representations retain more relevant signals than those obtained without task priors. Extensive experiments on the RelBench benchmark show that TVE consistently outperforms traditional pre-training baselines. Our findings advocate for pre-training objectives that encode task heterogeneity and temporal structure as design principles for predictive modeling on relational databases.



Paperid:3399
Authors:Hao Zheng, Shiyu Liang, Yuting Zheng, Chaofan Sun, LEI BAI, Enhui Liao
Abstract:
Reconstructing ocean surface nutrients from sparse observations is critical for understanding long-term biogeochemical cycles. Most prior work focuses on reconstructing atmospheric fields and treat the reconstruction problem as image inpainting, assuming smooth, single-scale dynamics. In contrast, nutrient transport follows advection–diffusion dynamics under nonstationary, multiscale ocean flow. This mismatch leads to instability, as small errors in unresolved eddies can propagate through time and distort nutrient predictions.To address this, we introduce NUTS, a two-scale reconstruction model that decouples large-scale transport and mesoscale variability. The homogenized solver captures stable, coarse-scale advection under filtered flow. A refinement module then restores mesoscale detail conditioned on the residual eddy field.NUTS is stable, interpretable, and robust to mesoscale perturbations, with theoretical guarantees from homogenization theory. NUTS outperforms all data-driven baselines in global reconstruction and achieves site-wise accuracy comparable to numerical models. On real observations, NUTS reduces NRMSE by 79.9\% for phosphate and 19.3\% for nitrate over the best baseline. Ablation studies validate the effectiveness of each module.



Paperid:3400
Authors:Petros Prastakos, Kayhan Behdin, Rahul Mazumder
Abstract:
Abstract:Sparse variable selection improves interpretability and generalization in high-dimensional learning by selecting a small subset of informative features. Recent advances in Mixed Integer Programming (MIP) have enabled solving large-scale non-private sparse regression—known as Best Subset Selection (BSS)—with millions of variables in minutes. However, extending these algorithmic advances to the setting of Differential Privacy (DP) has remained largely unexplored. In this paper, we introduce two new differentially private estimators for sparse variable selection, levering modern MIP techniques. Our framework is general and applies broadly to problems like sparse regression or classification, and we provide theoretical support recovery guarantees in the case of BSS. Inspired by the exponential mechanism, we develop structured sampling procedures that efficiently explore the non-convex objective landscape, avoiding the exhaustive combinatorial search in the exponential mechanism. We complement our theoretical findings with extensive numerical experiments, using both least squares and hinge loss for our objective function, and demonstrate that our methods achieve state-of-the-art empirical support recovery, outperforming competing algorithms in settings with up to $p=10^4$.



Paperid:3401
Authors:Danning Xie, Mingwei Zheng, Xuwei Liu, Jiannan Wang, Chengpeng Wang, Lin Tan, Xiangyu Zhang
Abstract:
Large language models (LLMs) have been widely adopted across diverse domains of software engineering, such as code generation, program repair, and vulnerability detection. These applications require understanding beyond surface-level code patterns: value propagation, control flow, and interdependence between program elements. However, existing benchmarks primarily evaluate end-to-end outcomes, such as whether code is correctly repaired or generated, leaving the models' ability of program semantic reasoning underexplored.This work presents CoRe, a high-quality, human-verified benchmark designed to evaluate LLMs on fundamental static analysis tasks. CoRe includes 10 task instances spanning data dependency, control dependency, and information flow across programs written in C/C++, Java, and Python. To ensure semantic diversity and reasoning complexity, we propose a semantics-aware diverse sampling strategy that selects targets and task instances based on structural coverage and dependency depth. We evaluate 10 state-of-the-art LLMs and show that, while they perform well at identifying dependencies, models still struggle with tasks that require deeper semantic understanding and multi-step reasoning.We further conduct qualitative analyses to uncover key challenges, such as complex control structures and backward dependency patterns, offering insights into improving LLMs’ code reasoning capabilities.



Paperid:3402
Authors:Kun Xiang, Heng Li, Terry Jingchen Zhang, Yinya Huang, Zirong Liu, Peixin Qu, Jixi He, Jiaqi Chen, Yu-Jie Yuan, Jianhua Han, Hang Xu, Hanhui Li, Mrinmaya Sachan, Xiaodan Liang
Abstract:
We present SeePhys, a large-scale multimodal benchmark for LLM reasoning grounded in physics questions ranging from middle school to PhD qualifying exams. The benchmark covers 7 fundamental domains spanning the physics discipline, incorporating 21 categories of highly heterogeneous diagrams. In contrast to prior works where visual elements mainly serve auxiliary purposes, our benchmark features a substantial proportion of vision-essential problems (75%) that mandate visual information extraction for correct solutions. Through extensive evaluation, we observe that even the most advanced visual reasoning models (e.g., Gemini-2.5-pro and o4-mini) achieve sub-60% accuracy on our benchmark. These results reveal fundamental challenges in current large language models' visual understanding capabilities, particularly in: (i) establishing rigorous coupling between diagram interpretation and physics reasoning, and (ii) overcoming their persistent reliance on textual cues as cognitive shortcuts.Project Page: github.com/SeePhys/seephys-projectHugging Face: huggingface.co/datasets/SeePhys/SeePhys



Paperid:3403
Authors:Yiju Guo, Wenkai Yang, Zexu Sun, Ning Ding, Zhiyuan Liu, Yankai Lin
Abstract:
Large language models (LLMs) have demonstrated significant improvements in contextual understanding. However, their ability to attend to truly critical information during long-context reasoning and generation still falls behind the pace. Specifically, our preliminary experiments reveal that certain misleading patterns can misdirect the model’s attention during inference, and removing these patterns substantially improves reasoning accuracy. We attribute this phenomenon to spurious correlations in the training data, which obstruct the model’s capacity to infer authentic causal instruction–response relationships. This phenomenon may induce redundant reasoning processes, potentially resulting in significant inference overhead and, more critically, the generation of erroneous or suboptimal responses. To mitigate this, we introduce a two-stage framework called Causal Attention Distillation (CAD) leveraging intervention-based inference to disentangle confounding factors. In the first stage, CAD employs gradient-based comparisons with an advanced teacher to automatically identify confounding tokens based on causal relationships in the training corpus. Then, in the second stage, it masks these tokens during distillation to enact intervention, aligning the student’s attention with the teacher’s focus distribution on truly critical context tokens. Experimental results demonstrate that CAD not only achieves an absolute improvement in various mathematical reasoning and code generation benchmarks but also effectively suppresses attention to confounding tokens during inference, yielding a more interpretable and reliable reasoning model.



Paperid:3404
Authors:Xinran Wang, Jin Du, Azal Khan, qi le, Enmao Diao, Jiawei Zhou, Jie Ding, Ali Anwar
Abstract:
Large language models can generate rare but catastrophic outputs, such as harmful conversations or insecure code. Existing Reinforcement Learning from Human Feedback (RLHF) typically maximizes average reward, leaving high-risk tail events insufficiently controlled. We introduce Quantile‑Guided Alignment (QA), a framework that allows users to specify desired improvements at any quantile—individually or across multiple reward dimensions—thus shifting the distribution of outputs with finer control toward safer, more desirable outcomes. The method extends standard RLHF via an augmented reward formulation that enforces quantile constraints. Experiments on conversation and code‐generation tasks show that quantile alignment significantly enhances quality at targeted tails while maintaining overall performance. The results position QA as a principled route to risk‑calibrated language models with tail‑focused alignment.



Paperid:3405
Authors:Tingting Dan, Zhihao Fan, Guorong Wu
Abstract:
Abstract:Alzheimer’s disease (AD) is marked by cognitive decline along with the widespread of tau aggregates across the brain cortex. Due to the challenges of imaging pathology spreading flows *in vivo*, however, quantitative analysis on the cortical pathways of tau propagation and its interaction with the cascade of amyloid-beta (A$\beta$) plaques lags behind the experimental insights of underlying pathophysiological mechanisms. To address this challenge, we present a physics-informed neural network, empowered by mean-field theory, to uncover the biologically meaningful spreading pathways of tau aggregates between two longitudinal snapshots. Following the notion of `prion-like' mechanism in AD, we first formulate the dynamics of tau propagation as a mean-field game (MFG), where the spread of tau aggregate at each location (aka. agent) depends on the collective behavior of the surrounding agents as well as the potential field formed by amyloid burden. Given the governing equation of propagation dynamics, MFG reaches an equilibrium that allows us to model the evolution of tau aggregates as an optimal transport with the lowest cost in *Wasserstein* space. By leveraging the variational primal-dual structure in MFG, we propose a *Wasserstein*-1 Lagrangian generative adversarial network (GAN), in which a Lipschitz critic seeks the appropriate transport cost at the population level and a generator parameterizes the flow fields of optimal transport across individuals. Additionally, we incorporate a symbolic regression module to derive an explicit formulation capturing the A$\beta$-tau crosstalk. Experimental results on public neuroimaging datasets demonstrate that our explainable deep model not only yields precise and reliable predictions of future tau progression for unseen new subjects but also provides a new window to uncover new understanding of pathology propagation in AD through learning-based approaches.



Paperid:3406
Authors:Xukai Zhang, Shuliang Wang, Guangyin Jin, Ziqiang Yuan, Hanning Yuan, Sijie Ruan
Abstract:
Air pollution remains one of the most critical environmental challenges globally, posing severe threats to public health, ecological sustainability, and climate governance. While existing physics-based and data-driven models have made progress in air quality forecasting, they often struggle to jointly capture the complex spatiotemporal dynamics and ensure spatial continuity of pollutant distributions. In this study, we introduce CTENet, a novel chemical transport deep learning model that embeds the Advection-Diffusion-Reaction equation into a Physics-Informed Neural Network (PINN) framework using an Eulerian representation to model the spatiotemporal evolution of pollutants. Extensive experiments on two real-world datasets demonstrate that CTENet consistently outperforms state-of-the-art (SOTA) baselines, achieving a remarkable RMSE improvement of 45.8% on the USA dataset and 21.0% on the China dataset. The implementation code is publicly available at: https://anonymous.4open.science/r/CTENet



Paperid:3407
Authors:Mingfeng Li, Qiang Zhang, Weijie Zheng, Benjamin Doerr
Abstract:
Abstract:Recent breakthroughs in the analysis of multi-objective evolutionary algorithms (MOEAs) are mathematical runtime analyses of those algorithms which are intensively used in practice. So far, most of these results show the same performance as previously known for simple algorithms like the GSEMO. The few results indicating advantages of the popular MOEAs share the same shortages: They consider the performance for the problem of computing the full Pareto front, (of some algorithms enriched with newly invented mechanisms,) and this on newly designed benchmarks. In this work, we overcome these shortcomings by analyzing how existing popular MOEAs approximate the Pareto front of the established LargeFront benchmark. We prove that all popular MOEAs, including NSGA-II (sequential version), NSGA-III, SMS-EMOA, and SPEA2, only need an expected time of $O(n^2 \log n)$ function evaluations to compute an additive $\varepsilon$-approximation of the Pareto front of the LargeFront benchmark. This contrasts with the already proven exponential runtime (with high probability) of the GSEMO on the same task. % (which we also extend to any dominance-only black-box MOEA). This result is the first mathematical runtime analysis showing and explaining the superiority of popular MOEAs over simple ones like the GSEMO for the central task of computing good approximations to the Pareto front.



Paperid:3408
Authors:Yuyang Huang, Yabo Chen, Junyu Zhou, Wenrui Dai, XIAOPENG ZHANG, Junni Zou, Hongkai Xiong, Qi Tian
Abstract:
Source-free domain adaptation (SFDA) is a challenging tasks that tackles domain shifts using only a pre-trained source model and unlabeled target data. Existing SFDA methods are restricted by the fundamental limitation of source-target domain discrepancy. Non-generation SFDA methods suffer from unreliable pseudo-labels in challenging scenarios with large domain discrepancies, while generation-based SFDA methods are evidently degraded due to enlarged domain discrepancies in creating pseudo-source data. To address this limitation, we propose a novel generation-based framework named Diffusion-Driven Progressive Target Manipulation (DPTM) that leverages unlabeled target data as references to reliably generate and progressively refine a pseudo-target domain for SFDA. Specifically, we divide the target samples into a trust set and a non-trust set based on the reliability of pseudo-labels to sufficiently and reliably exploit their information. For samples from the non-trust set, we develop a manipulation strategy to semantically transform them into the newly assigned categories, while simultaneously maintaining them in the target distribution via a latent diffusion model. Furthermore, we design a progressive refinement mechanism that progressively reduces the domain discrepancy between the pseudo-target domain and the real target domain via iterative refinement. Experimental results demonstrate that DPTM outperforms existing methods by a large margin and achieves state-of-the-art performance on four prevailing SFDA benchmark datasets with different scales. Remarkably, DPTM can significantly enhance the performance by up to 22.7\% in scenarios with large source-target gaps.



Paperid:3409
Authors:Jiyang Zheng, Siqi Pan, Yu Yao, Zhaoqing Wang, Dadong Wang, Tongliang Liu
Abstract:
Text-to-Audio-Video (T2AV) generation aims to produce temporally and semantically aligned visual and auditory content from natural language descriptions. While recent progress in text-to-audio and text-to-video models has improved generation quality within each modality, jointly modeling them remains challenging due to incomplete and asymmetric correspondence: audio often reflects only a subset of the visual scene, and vice versa. Naively enforcing full alignment introduces semantic noise and temporal mismatches. To address this, we propose a novel framework that performs selective cross-modal alignment through a learnable masking mechanism, enabling the model to isolate and align only the shared latent components relevant to both modalities. This mechanism is integrated into an adaptation module that interfaces with pretrained encoders and decoders from latent video and audio diffusion models, preserving their generative capacity with reduced training overhead. Theoretically, we show that our masked objective provably recovers the minimal set of shared latent variables across modalities. Empirically, our method achieves state-of-the-art performance on standard T2AV benchmarks, demonstrating significant improvements in audiovisual synchronization and semantic consistency.



Paperid:3410
Authors:Md Mostafijur Rahman, Radu Marculescu
Abstract:
U‑shaped networks output logits at multiple spatial scales, each capturing a different blend of coarse context and fine detail. Yet, training still treats these logits in isolation—either supervising only the final, highest‑resolution logits or applying deep supervision with identical loss weights at every scale—without exploringmixed‑scalecombinations. Consequently, the decoder output misses the complementary cues that arise only when coarse and fine predictions are fused. To address this issue, we introduce LoMix (Logits Mixing), a Neural Architecture Search (NAS)‑inspired, differentiable plug-and-play module thatcreatesmixed‑scale outputs andlearnshow exactly each of them should guide the training process. More precisely, LoMix mixes the multi-scale decoder logits with four lightweight fusion operators: addition, multiplication, concatenation, and attention-based weighted fusion, yielding a rich set of synthetic “mutant’’ maps. Every original or mutant map is given a softplus loss weight that is co‑optimized with network parameters, mimicking a one‑step architecture search that automatically discovers the most useful scales, mixtures, and operators. Plugging LoMix into recent U-shaped architectures (i.e., PVT‑V2‑B2 backbone with EMCAD decoder) on Synapse 8‑organ dataset improves DICE by +3.59% over single‑output supervision, +2.38% over uniform deep supervision, and +1.44% over equally weighted additive fusion, all withzeroinference overhead. When training data are scarce (e.g., one or two labeled scans, 5% of the trainset), the advantage grows to +9.23%, underscoring LoMix’s data efficiency. Experiments across four popular benchmarks, including stronger results on the Synapse 13‑organ segmentation, and multiple backbones confirm that learnable weighted mixed‑scale fusion generalizes broadly while remaining data efficient, fully interpretable, and overhead-free at inference.Code and trained weights will be made public upon acceptance of the paper.



Paperid:3411
Authors:Ziyi Wu, Anil Kag, Ivan Skorokhodov, Willi Menapace, Ashkan Mirzaei, Igor Gilitschenski, Sergey Tulyakov, Aliaksandr Siarohin
Abstract:
Direct Preference Optimization (DPO) has recently been applied as a post‑training technique for text-to-video diffusion models.To obtain training data, annotators are asked to provide preferences between two videos generated from independent noise.However, this approach prohibits fine-grained comparisons, and we point out that it biases the annotators towards low-motion clips as they often contain fewer visual artifacts.In this work, we introduce DenseDPO, a method that addresses these shortcomings by making three contributions.First, we create each video pair for DPO by denoising corrupted copies of a ground truth video.This results in aligned pairs with similar motion structures while differing in local details, effectively neutralizing the motion bias.Second, we leverage the resulting temporal alignment to label preferences on short segments rather than entire clips, yielding a denser and more precise learning signal.With only one‑third of the labeled data, DenseDPO greatly improves motion generation over vanilla DPO, while matching it in text alignment, visual quality, and temporal consistency.Finally, we show that DenseDPO unlocks automatic preference annotation using off-the-shelf Vision Language Models (VLMs): GPT accurately predicts segment-level preferences similar to task-specifically fine-tuned video reward models, and DenseDPO trained on these labels achieves performance close to using human labels.



Paperid:3412
Authors:Yuxi Liu, Min Liu, Yi Tang, Shuai Jiang, Yaonan Wang
Abstract:
Existing NAS methods for semantic segmentation typically apply uniform optimization to all candidate networks (paths) within a one-shot supernet. However, the concurrent existence of both promising and suboptimal paths often results in inefficient weight updates and gradient conflicts. This issue is particularly severe in semantic segmentation due to its complex multi-branch architectures and large search space, which further degrade the supernet's ability to accurately evaluate individual paths and identify high-quality candidates. To address this issue, we propose Dynamic Path Selection (DPS), a selective training strategy that leverages multiple performance proxies to guide path optimization. DPS follows a stage-wise paradigm, where each phase emphasizes a different objective: early stages prioritize convergence, the middle stage focuses on expressiveness, and the final stage emphasizes a balanced combination of expressiveness and generalization. At each stage, paths are selected based on these criteria, concentrating optimization efforts on promising paths, thus facilitating targeted and efficient model updates. Additionally, DPS integrates a dynamic stage scheduler and a diversity-driven exploration strategy, which jointly enable adaptive stage transitions and maintain structural diversity among selected paths. Extensive experiments demonstrate that, under the same search space, DPS can discover efficient models with strong generalization and superior performance.



Paperid:3413
Authors:Sk Tanzir Mehedi, Raja Jurdak, Chadni Islam, Gowri Sankar Ramachandran
Abstract:
Securing software supply chains is a growing challenge due to the inadequacy of existing datasets in capturing the complexity of next-gen attacks, such as multiphase malware execution, remote access activation, and dynamic payload generation. Existing datasets, which rely on metadata inspection and static code analysis, are inadequate for detecting such attacks. This creates a critical gap because these datasets do not capture what happens during and after a package is installed. To address this gap, we present QUT-DV25, a dynamic analysis dataset specifically designed to support and advance research on detecting and mitigating supply chain attacks within the Python Package Index (PyPI) ecosystem. This dataset captures install and post-install-time traces from 14,271 Python packages, of which 7,127 are malicious. The packages are executed in an isolated sandbox environment using an extended Berkeley Packet Filter (eBPF) kernel and user-level probes. It captures 36 real-time features, that includes system calls, network traffic, resource usages, directory access patterns, dependency logs, and installation behaviors, enabling the study of next-gen attack vectors. ML analysis using the QUT-DV25 dataset identified four malicious PyPI packages previously labeled as benign, each with thousands of downloads. These packages deployed covert remote access and multi-phase payloads, were reported to PyPI maintainers, and subsequently removed. This highlights the practical value of QUT-DV25, as it outperforms reactive, metadata, and static datasets, offering a robust foundation for developing and benchmarking advanced threat detection within the evolving software supply chain ecosystem.



Paperid:3414
Authors:Licong Lin, Song Mei
Abstract:
Abstract:Contrastive learning---a modern approach to extract useful representations from unlabeled data by training models to distinguish similar samples from dissimilar ones---has driven significant progress in foundation models. In this work, we develop a new theoretical framework for analyzing data augmentation-based contrastive learning, with a focus on SimCLR as a representative example. Our approach is based on the concept of \emph{approximate sufficient statistics}, which we extend beyond its original definition in~\cite{oko2025statistical} for contrastive language-image pretraining (CLIP) using KL-divergence. We generalize it to equivalent forms and general $f$-divergences, and show that minimizing SimCLR and other contrastive losses yields encoders that are approximately sufficient. Furthermore, we demonstrate that these near-sufficient encoders can be effectively adapted to downstream regression and classification tasks, with performance depending on their sufficiency and the error induced by data augmentation in contrastive learning. Concrete examples in linear regression and topic classification are provided to illustrate the broad applicability of our results.



Paperid:3415
Authors:Man Ho Lam, Chaozheng Wang, Jen-Tse Huang, Michael R Lyu
Abstract:
Large Language Models (LLMs) have demonstrated remarkable performance on various code-related tasks, but their reliability in code comprehension under perturbations remains insufficiently studied.We present CodeCrash, a comprehensive stress-testing benchmark with 1,279 questions collected from two established datasets, CruxEval and LiveCodeBench, to evaluate 17 LLMs across input and output prediction tasks under structural and textual perturbations.Our evaluation reveals that LLMs heavily rely on embedded natural language cues for reasoning and understanding, resulting in over 13 percentage point degradation in output prediction tasks and leading to superficial reasoning shortcuts in input prediction tasks.Furthermore, we identify a progressive relationship between reasoning depth and model reliability, where explicit step-by-step execution improves robustness, and internal reasoning models achieve even greater robustness but expose critical vulnerabilities.The self-reflection mechanism increases token usage by 2-3 times, reduces confidence, and even leads to severe reasoning collapse under reasoning-level textual perturbations.CodeCrash provides a rigorous benchmark for evaluating robustness in code understanding, guiding future research toward more reliable and resilient LLMs in code reasoning.



Paperid:3416
Authors:Bowen Gao, Yanwen Huang, Yiqiao Liu, Wenxuan Xie, Bowei He, Haichuan Tan, Wei-Ying Ma, Ya-Qin Zhang, Yanyan Lan
Abstract:
Structure-guided molecular generation is pivotal in early-stage drug discovery, enabling the design of compounds tailored to specific protein targets. However, despite recent advances in 3D generative modeling, particularly in improving docking scores, these methods often produce rare and intrinsically irrational molecular structures that deviate from drug-like chemical space. To quantify this issue, we propose a novel metric, the Molecule Reasonable Ratio (MRR), which measures structural rationality and reveals a critical gap between existing models and real-world approved drugs. To address this, we introduce the Collaborative Intelligence Drug Design (CIDD) framework, the first approach to unify the 3D interaction modeling capabilities of generative models with the general knowledge and reasoning power of large language models (LLMs). By leveraging LLM-based Chain-of-Thought reasoning, CIDD generates molecules that not only bind effectively to protein pockets but also exhibit strong structural drug-likeness, rationality, and synthetic accessibility. On the CrossDocked2020 benchmark, CIDD consistently improves drug-likeness metrics, including QED, SA, and MRR, across different base generative models, while maintaining competitive binding affinity. Notably, it raises the combined success rate (balancing drug-likeness and binding) from 15.72% to 34.59%, more than doubling previous results. These findings demonstrate the value of integrating knowledge reasoning with geometric generation to advance AI-driven drug design.



Paperid:3417
Authors:Yuanpei Gao, Qi Yan, Yan Leng, Renjie Liao
Abstract:
While deep learning methods have achieved strong performance in time series prediction, their black-box nature and inability to explicitly model underlying stochastic processes often limit their generalization to non-stationary data, especially in the presence of abrupt changes. In this work, we introduce Neural MJD, a neural network based non-stationary Merton jump diffusion (MJD) model. Our model explicitly formulates forecasting as a stochastic differential equation (SDE) simulation problem, combining a time-inhomogeneous Itô diffusion to capture non-stationary stochastic dynamics with a time-inhomogeneous compound Poisson process to model abrupt jumps. To enable tractable learning, we introduce a likelihood truncation mechanism that caps the number of jumps within small time intervals and provide a theoretical error bound for this approximation. Additionally, we propose an Euler-Maruyama with restart solver, which achieves a provably lower error bound in estimating expected states and reduced variance compared to the standard solver. Experiments on both synthetic and real-world datasets demonstrate that Neural MJD consistently outperforms state-of-the-art deep learning and statistical learning methods.



Paperid:3418
Authors:Zhe Hu, Jing Li, Zhongzhu Pu, Hou Pong (Ken) Chan, Yu Yin
Abstract:
Vision Language Models exhibited immense potential for embodied AI, yet they often lack the sophisticated situational reasoning required for complex decision-making. This paper shows that VLMs can achieve surprisingly strong decision-making performance when visual scenes are represented merely as text-only descriptions, suggesting foundational reasoning can be effectively learned from language. Motivated by this insight, we propose Praxis-VLM, a reasoning VLM for vision-grounded decision-making. Praxis-VLM employs the GRPO algorithm on textual scenarios to instill robust reasoning capabilities, where models learn to evaluate actions and their consequences. These reasoning skills, acquired purely from text, successfully transfer to multimodal inference with visual inputs, significantly reducing reliance on scarce paired image-text training data. Experiments across diverse decision-making benchmarks demonstrate that Praxis-VLM substantially outperforms standard supervised fine-tuning, exhibiting superior performance and generalizability. Further analysis confirms that our models engage in explicit and effective reasoning, underpinning their enhanced performance and adaptability.



Paperid:3419
Authors:Qian Chen, Linxin Yang, Akang Wang, Xiaodong Luo, Yin Zhang
Abstract:
The combination of linear transformations and nonlinear activation functions forms the foundation of most modern deep neural networks, enabling them to approximate highly complex functions. This paper explores the introduction of quadratic transformations to further increase the nonlinearity of the model, with the aim of enhancing the performance of existing architectures. To minimize the additional parameters and computational burden, we propose a lightweight quadratic enhancer that leverages matrix decomposition, weight sharing, and sparsification techniques. This approach introduces only a minimal and negligible increase in parameters and forward computation, while still yielding substantial improvements in model performance. We evaluate the effectiveness of the proposed method across three tasks: text classification, image classification, and fine-tuning large language models (LLMs). In all tasks, our approach demonstrates significant performance gains.



Paperid:3420
Authors:Michal Kmicikiewicz, Vincent Fortuin, Ewa Szczurek
Abstract:
Designing protein sequences of both high fitness and novelty is a challenging task in data-efficient protein engineering. Exploration beyond wild-type neighborhoods often leads to biologically implausible sequences or relies on surrogate models that lose fidelity in novel regions. Here, we propose ProSpero, an active learning framework in which a frozen pre-trained generative model is guided by a surrogate updated from oracle feedback. By integrating fitness-relevant residue selection with biologically-constrained Sequential Monte Carlo sampling, our approach enables exploration beyond wild-type neighborhoods while preserving biological plausibility. We show that our framework remains effective even when the surrogate is misspecified.ProSpero consistently outperforms or matches existing methods across diverse protein engineering tasks, retrieving sequences of both high fitness and novelty.



Paperid:3421
Authors:Jialiang Kang, Han Shu, Wenshuo Li, Yingjie Zhai, Xinghao Chen
Abstract:
Abstract:Speculative decoding is a widely adopted technique for accelerating inference in large language models (LLMs), yet its application to vision-language models (VLMs) remains underexplored, with existing methods achieving only modest speedups ($<1.5\times$). This gap is increasingly significant as multimodal capabilities become central to large-scale models. We hypothesize that large VLMs can effectively filter redundant image information layer by layer without compromising textual comprehension, whereas smaller draft models struggle to do so. To address this, we introduce Vision-Aware Speculative Decoding (ViSpec), a novel framework tailored for VLMs. ViSpec employs a lightweight vision adaptor module to compress image tokens into a compact representation, which is seamlessly integrated into the draft model's attention mechanism while preserving original image positional information. Additionally, we extract a global feature vector for each input image and augment all subsequent text tokens with this feature to enhance multimodal coherence. To overcome the scarcity of multimodal datasets with long assistant responses, we curate a specialized training dataset by repurposing existing datasets and generating extended outputs using the target VLM with modified prompts. Our training strategy mitigates the risk of the draft model exploiting direct access to the target model's hidden states, which could otherwise lead to shortcut learning when training solely on target model outputs. Extensive experiments validate ViSpec, achieving, to our knowledge, the first substantial speedup in VLM speculative decoding.



Paperid:3422
Authors:Zijun Chen, Shengbo Wang, Nian Si
Abstract:
Abstract:Motivated by practical applications where stable long-term performance is critical—such as robotics, operations research, and healthcare—we study the problem of distributionally robust (DR) average-reward reinforcement learning. We propose two algorithms that achieve near-optimal sample complexity. The first reduces the problem to a DR discounted Markov decision process (MDP), while the second, Anchored DR Average-Reward MDP, introduces an anchoring state to stabilize the controlled transition kernels within the uncertainty set. Assuming the nominal MDP is uniformly ergodic, we prove that both algorithms attain a sample complexity of $\widetilde{O}\left(|\mathbf{S}||\mathbf{A}| t_{\mathrm{mix}}^2\varepsilon^{-2}\right)$ for estimating the optimal policy as well as the robust average reward under KL and $f_k$-divergence-based uncertainty sets, provided the uncertainty radius is sufficiently small. Here, $\varepsilon$ is the target accuracy, $|\mathbf{S}|$ and $|\mathbf{A}|$ denote the sizes of the state and action spaces, and $t_{\mathrm{mix}}$ is the mixing time of the nominal MDP. This represents the first finite-sample convergence guarantee for DR average-reward reinforcement learning. We further validate the convergence rates of our algorithms through numerical experiments.



Paperid:3423
Authors:Jason Yang, Wenda Chu, Daniel Khalil, Raul Astudillo, Bruce Wittmann, Frances Arnold, Yisong Yue
Abstract:
Protein fitness optimization involves finding a protein sequence that maximizes desired quantitative properties in a combinatorially large design space of possible sequences. Recent developments in steering protein generative models (e.g diffusion models, language models) offer a promising approach. However, by and large, past studies have optimized surrogate rewards and/or utilized large amounts of labeled data for steering, making it unclear how well existing methods perform and compare to each other in real-world optimization campaigns where fitness is measured by low-throughput wet-lab assays. In this study, we explore fitness optimization using small amounts (hundreds) of labeled sequence-fitness pairs and comprehensively evaluate strategies such as classifier guidance and posterior sampling for guiding generation from different discrete diffusion models of protein sequences. We also demonstrate how guidance can be integrated into adaptive sequence selection akin to Thompson sampling in Bayesian optimization, showing that plug-and-play guidance strategies offer advantages compared to alternatives such as reinforcement learning with protein language models.



Paperid:3424
Authors:Peixian Ma, Xialie Zhuang, Chengjin Xu, Xuhui Jiang, Ran Chen, Jian Guo
Abstract:
Natural Language to SQL (NL2SQL) enables intuitive interactions with databases by transforming natural language queries into structured SQL statements. Despite recent advancements in enhancing human-computer interaction within database applications, significant challenges persist, particularly regarding the inference performance in complex scenarios involving multi-table joins and nested queries. Current methodologies primarily utilize supervised fine-tuning (SFT) to train the NL2SQL model, which may limit adaptability and interpretability in new environments (e.g., finance and healthcare).In order to enhance the reasoning performance of the NL2SQL model in the above complex situations, we introduce SQL-R1, a novel NL2SQL reasoning model trained by the reinforcement learning (RL) algorithms.We design a specialized RL-based reward function tailored for NL2SQL tasks and discussed the impact of cold start and synthetic data on the effectiveness of intensive training. In addition, we achieve competitive accuracy using only a tiny amount of synthetic NL2SQL data for augmented training and further explore data engineering for RL. In existing experiments, SQL-R1 achieves execution accuracy of 88.6\% and 67.1\% on the benchmark Spider and BIRD, respectively.



Paperid:3425
Authors:Enjun Du, Xunkai Li, Tian Jin, Zhihan Zhang, Rong-Hua Li, Guoren Wang
Abstract:
The era of foundation models has revolutionized AI research, yet Graph Foundation Models (GFMs) remain constrained by the scarcity of large-scale graph corpora. Traditional graph data synthesis techniques primarily focus on simplistic structural operations, lacking the capacity to generate semantically rich nodes with meaningful textual attributes—a critical limitation for real-world applications. While large language models (LLMs) demonstrate exceptional text generation capabilities, their direct application to graph synthesis is impeded by context window limitations, hallucination phenomena, and structural consistency challenges. To address these issues, we introduce \textbf{GraphMaster}—the first multi-agent framework specifically designed for graph data synthesis in data-limited environments. GraphMaster orchestrates four specialized LLM agents (Manager, Perception, Enhancement, and Evaluation) that collaboratively optimize the synthesis process through iterative refinement, ensuring both semantic coherence and structural integrity. To rigorously evaluate our approach, we create new data-limited “Sub” variants of six standard graph benchmarks, specifically designed to test synthesis capabilities under realistic constraints. Additionally, we develop a novel interpretability assessment framework that combines human evaluation with a principled Grassmannian manifold-based analysis, providing both qualitative and quantitative measures of semantic coherence. Experimental results demonstrate that GraphMaster significantly outperforms traditional synthesis methods across multiple datasets, establishing a strong foundation for advancing GFMs in data-scarce environments.



Paperid:3426
Authors:Dongkeun Kim, Minsu Cho, Suha Kwak
Abstract:
Social interactions often emerge from subtle, fine-grained cues such as facial expressions, gaze, and gestures. However, existing methods for social interaction detection overlook such nuanced cues and primarily rely on holistic representations of individuals. Moreover, they directly detect social groups without explicitly modeling the underlying interactions between individuals. These drawbacks limit their ability to capture localized social signals and introduce ambiguity when group configurations should be inferred from social interactions grounded in nuanced cues. In this work, we propose a part-aware bottom-up group reasoning framework for fine-grained social interaction detection. The proposed method infers social groups and their interactions using body part cues and their interpersonal relations. Our model first detects individuals and enhances their features using part-aware features, and then infers group configuration by associating individuals via similarity-based reasoning, which considers not only spatial relations but also subtle social cues that signal interactions, leading to more accurate group inference. Experiments on the NVI dataset demonstrate that our method outperforms prior methods, achieving the new state of the art. Our codebase will be open to the public for research purposes.



Paperid:3427
Authors:Haydn Jones, Natalie Maus, Josh magnus Ludan, Maggie Huan, Jiaming Liang, Marcelo Der Torossian Torres, Jiatao Liang, Zachary Ives, Yoseph Barash, Cesar de la Fuente-Nunez, Jacob Gardner, Mark Yatskar
Abstract:
AI-driven discovery can greatly reduce design time and enhance new therapeutics’ effectiveness. Models using simulators explore broad design spaces but risk violating implicit constraints due to a lack of experimental priors. For example, in a new analysis we performed on a diverse set of models on the GuacaMol benchmark using supervised classifiers, over 60% of molecules proposed had high probability of being mutagenic. In this work, we introduce Medex, a dataset of priors for design problems extracted from literature describing compounds used in lab settings. It is constructed with LLM pipelines for discovering therapeutic entities in relevant paragraphs and summarizing information in concise fair-use facts. Medex consists of 36.4 million pairs of natural language facts, and appropriate entity representations (i.e. SMILES or refseq IDs). To demonstrate the potential of the data, we train LLM, CLIP, and LLava architectures to reason jointly about text and design targets and evaluate on tasks from the Therapeutic Data Commons (TDC). Medex is highly effective for creating models with strong priors: in supervised prediction problems that use our data as pretraining, our best models with 15M learnable parameters outperform larger 2B TxGemma on both regression and classification TDC tasks and perform comparably to 9B models on average. Models built with Medex can be used as constraints while optimizing for novel molecules in GuacaMol, result19 ing in proposals that are safer and nearly as effective. We release our dataset at huggingface.co/datasets/medexanon/Medex, and will provide expanded versions as available literature grows.



Paperid:3428
Authors:Lianghe Shi, Meng Wu, Huijie Zhang, Zekai Zhang, Molei Tao, Qing Qu
Abstract:
The widespread use of diffusion models has led to an abundance of AI-generated data, raising concerns about model collapse---a phenomenon in which recursive iterations of training on synthetic data lead to performance degradation. Prior work primarily characterizes this collapse via variance shrinkage or distribution shift, but these perspectives miss practical manifestations of model collapse. This paper identifies a transition from generalization to memorization during model collapse in diffusion models, where models increasingly replicate training data instead of generating novel content during iterative training on synthetic samples. This transition is directly driven by the declining entropy of the synthetic training data produced in each training cycle, which serves as a clear indicator of model degradation. Motivated by this insight, we propose an entropy-based data selection strategy to mitigate the transition from generalization to memorization and alleviate model collapse. Empirical results show that our approach significantly enhances visual quality and diversity in recursive generation, effectively preventing collapse.



Paperid:3429
Authors:Junhao Shen, Haiteng Zhao, Yuzhe Gu, Songyang Gao, Kuikun Liu, Haian Huang, Jianfei Gao, Dahua Lin, Wenwei Zhang, Kai Chen
Abstract:
Enhancing large vision-language models (LVLMs) with visual slow-thinking reasoning is crucial for solving complex multimodal tasks. However, since LVLMs are mainly trained with vision-language alignment, it is difficult to adopt on-policy reinforcement learning (RL) to develop the slow thinking ability because the rollout space is restricted by its initial abilities. Off-policy RL offers a way to go beyond the current policy, but directly distilling trajectories from external models may cause visual hallucinations due to mismatched visual perception abilities across models. To address these issues, this paper proposesSOPHIA, a simple and scalableSemi-Off-Policy RL for vision-language slow-tHInking reAsoning. SOPHIA builds a semi-off-policy behavior model by combining on-policy visual understanding from a trainable LVLM with off-policy slow-thinking reasoning from a language model, assigns outcome-based rewards to reasoning, and propagates visual rewards backward. Then LVLM learns slow-thinking reasoning ability from the obtained reasoning trajectories using propagated rewards via off-policy RL algorithms. Extensive experiments with InternVL2.5 and InternVL3.0 with 8B and 38B sizes show the effectiveness of SOPHIA. Notably, SOPHIA improves InternVL3.0-38B by 8.50\% in average, reaching state-of-the-art performance among open-source LVLMs on multiple multimodal reasoning benchmarks, and even outperforms some closed-source models (e.g., GPT-4.1) on the challenging MathVision and OlympiadBench, achieving 49.08\% and 49.95\% pass@1 accuracy, respectively. Analysis shows SOPHIA outperforms supervised fine-tuning and direct on-policy RL methods, offering a better policy initialization for further on-policy training.



Paperid:3430
Authors:Young-Jae Park, Minseok Seo, Hae-Gon Jeon
Abstract:
Accurate video forecasting enables autonomous vehicles to anticipate hazards, robotics and surveillance systems to predict human intent, and environmental models to issue timely warnings for extreme weather events. However, existing methods remain limited: transformers rely on global attention with quadratic complexity, making them impractical for high-resolution, long-horizon video prediction, while convolutional and recurrent networks suffer from short-range receptive fields and vanishing gradients, losing key information over extended sequences. To overcome these challenges, we introduce VideoTitans, the first architecture to adapt the gradient-driven Titans memory—originally designed for language modelling to video prediction. VideoTitans integrates three core ideas: (i) a sliding-window attention core that scales linearly with sequence length and spatial resolution, (ii) an episodic memory that dynamically retains only informative tokens based on a gradient-based surprise signal, and (iii) a small set of persistent tokens encoding task-specific priors that stabilize training and enhance generalization. Extensive experiments on Moving-MNIST, Human3.6M, TrafficBJ and WeatherBench benchmarks show that VideoTitans consistently reduces computation (FLOPs) and achieves competitive visual fidelity compared to state-of-the-art recurrent, convolutional, and efficient-transformer methods. Comprehensive ablations confirm that each proposed component contributes significantly. Code, checkpoints, and demonstration videos will be publicly available to ensure reproducibility and promote further research.



Paperid:3431
Authors:tal gonen, Itai Pemper, Ilan Naiman, Nimrod Berman, Omri Azencot
Abstract:
Generative modeling of time series is a central challenge in time series analysis, particularly under data-scarce conditions. Despite recent advances in generative modeling, a comprehensive understanding of how state-of-the-art generative models perform under limited supervision remains lacking. In this work, we conduct the first large-scale study evaluating leading generative models in data-scarce settings, revealing a substantial performance gap between full-data and data-scarce regimes. To close this gap, we propose a unified diffusion-based generative framework that can synthesize high-fidelity time series across diverse domains using just a few examples. Our model is pretrained on a large, heterogeneous collection of time series datasets, enabling it to learn generalizable temporal representations. It further incorporates architectural innovations such as dynamic convolutional layers for flexible channel adaptation and dataset token conditioning for domain-aware generation. Without requiring abundant supervision, our unified model achieves state-of-the-art performance in few-shot settings—outperforming domain-specific baselines across a wide range of subset sizes. Remarkably, it also surpasses all baselines even when tested on full datasets benchmarks, highlighting the strength of pretraining and cross-domain generalization. We hope this work encourages the community to revisit few-shot generative modeling as a key problem in time series research and pursue unified solutions that scale efficiently across domains.



Paperid:3432
Authors:Jiayuan Ding, Jianhui Lin, Shiyu Jiang, Yixin Wang, Ziyang Miao, Zhaoyu Fang, Jiliang Tang, Min Li, Xiaojie Qiu
Abstract:
Foundation models (FMs) have shown great promise in single-cell genomics, yet current approaches, such as scGPT, Geneformer, and scFoundation, rely on centralized training and language modeling objectives that overlook the tabular nature of single-cell data and raise significant privacy concerns. We present TABULA, a foundation model designed for single-cell transcriptomics, which integrates a novel tabular modeling objective and federated learning framework to enable privacy-preserving pretraining across decentralized datasets. TABULA directly models the cell-by-gene expression matrix through column-wise gene reconstruction and row-wise cell contrastive learning, capturing both gene-level relationships and cell-level heterogeneity without imposing artificial gene sequence order. Extensive experiments demonstrate the effectiveness of TABULA: despite using only half the pretraining data, TABULA achieves state-of-the-art performance across key tasks, including gene imputation, perturbation prediction, cell type annotation, and multi-omics integration. It is important to note that as public single-cell datasets continue to grow, TABULA provides a scalable and privacy-aware foundation that not only validates the feasibility of federated tabular modeling but also establishes a generalizable framework for training future models under similar privacy-preserving settings.



Paperid:3433
Authors:Heekang Song, Wan Choi
Abstract:
Abstract:In this paper, we propose an optimally structured gradient coding scheme to mitigate the straggler problem in distributed learning. Conventional gradient coding methods often assume homogeneous straggler models or rely on excessive data replication, limiting performance in real-world heterogeneous systems. To address these limitations, we formulate an optimization problem minimizing residual error while ensuring unbiased gradient estimation by explicitly considering individual straggler probabilities. We derive closed-form solutions for optimal encoding and decoding coefficients via Lagrangian duality and convex optimization, and propose data allocation strategies that reduce both redundancy and computational load. We also analyze convergence behavior for $\lambda$-strongly convex and $\mu$-smooth loss functions. Numerical results show that our approach significantly reduces the impact of stragglers and accelerates convergence compared to existing methods.



Paperid:3434
Authors:Jincheol Yang, Jaemin Choi, Matti Zinke, Suk-Ju Kang
Abstract:
Abstract:Monocular depth estimation has advanced significantly with foundation models like Depth Anything, leveraging large-scale transformer architectures for the superior generalization. However, the deployment on resource-constrained devices remains challenging due to the high computation and memory requirement. Existing quantization methods, such as post-training quantization and quantization-aware training, often face trade-offs between efficiency and accuracy, or require extensive labeled data for retraining.To address these limitations, we propose Quantization with Self-Compensating Auxiliary for Monocular Depth Estimation (QSCA), a novel framework for 4-bit post-training quantization of Monocular depth estimation models. Our method integrates a lightweight Self-Compensating Auxiliary (SCA) module into both transformer encoder and decoder blocks, enabling the quantized model to recover from performance degradation without requiring ground truth.This design enables fast adaptation while preserving structural and spatial consistency in predicted depth maps.To our knowledge, this is the first framework to successfully apply 4-bit quantization across all layers of large-scale monocular depth estimation models.Experimental results demonstrate that QSCA significantly improves quantized depth estimation performance. On the NYUv2 dataset, it achieves an 11\% improvement in $\delta_1$ accuracy over existing post-training quantization methods.



Paperid:3435
Authors:Hongjin Qian, Zheng Liu
Abstract:
Augmenting large language models (LLMs) with external retrieval has become a standard method to address their inherent knowledge cutoff limitations. However, traditional retrieval-augmented generation methods employ static, pre-inference retrieval strategies, making them inadequate for complex tasks involving ambiguous, multi-step, or evolving information needs. Recent advances in test-time scaling techniques have demonstrated significant potential in enabling LLMs to dynamically interact with external tools, motivating the shift toward adaptive inference-time retrieval.Inspired by Information Foraging Theory (IFT), we propose InForage, a reinforcement learning framework that formalizes retrieval-augmented reasoning as a dynamic information-seeking process. Unlike existing approaches, InForage explicitly rewards intermediate retrieval quality, encouraging LLMs to iteratively gather and integrate information through adaptive search behaviors. To facilitate training, we construct a human-guided dataset capturing iterative search and reasoning trajectories for complex, real-world web tasks. Extensive evaluations across general question answering, multi-hop reasoning tasks, and a newly developed real-time web QA dataset demonstrate InForage's superior performance over baseline methods. These results highlight InForage's effectiveness in building robust, adaptive, and efficient reasoning agents. We provide all codes and datasets in the supplementary materials.



Paperid:3436
Authors:Taha Entesari, Arman Hatami, Rinat Khaziev, Anil Ramakrishna, Mahyar Fazlyab
Abstract:
Large Language Models (LLMs) deployed in real-world settings increasingly face the need to unlearn sensitive, outdated, or proprietary information. Existing unlearning methods typically formulate forgetting and retention as a regularized trade-off, combining both objectives into a single scalarized loss. This often leads to unstable optimization and degraded performance on retained data, especially under aggressive forgetting. We propose a new formulation of LLM unlearning as a constrained optimization problem: forgetting is enforced via a novel logit-margin flattening loss that explicitly drives the output distribution toward uniformity on a designated forget set, while retention is preserved through a hard constraint on a separate retain set. Compared to entropy-based objectives, our loss is softmax-free, numerically stable, and maintains non-vanishing gradients, enabling more efficient and robust optimization. We solve the constrained problem using a scalable primal-dual algorithm that exposes the trade-off between forgetting and retention through the dynamics of the dual variable. Evaluations on the TOFU and MUSE benchmarks across diverse LLM architectures demonstrate that our approach consistently matches or exceeds state-of-the-art baselines, effectively removing targeted information while preserving downstream utility.



Paperid:3437
Authors:Nuo Chen, Zehua Li, Keqin Bao, Junyang Lin, Dayiheng Liu
Abstract:
Building robust and general reasoning ability is a central goal in the development of large language models (LLMs). Recent efforts increasingly turn to code as a rich training source, given its inherent logical structure and diverse reasoning paradigms—such as divide-and-conquer, topological ordering, and enumeration. However, reasoning in code is often expressed implicitly and entangled with syntactic or implementation noise, making direct training on raw code suboptimal. To address this, we introduce TraceMind, a large-scale corpus of 2.6 million samples that transforms code execution into explicit, step-by-step chain-of-thought style rationales, which we call Chain of Execution (CoE). The corpus spans domains including mathematics, classical algorithms and algorithmic competition, and is enriched with variable-tracing questions and code rewritings to enhance logical granularity and code diversity. We evaluate Tracepile using three training setups—continue-pretraining, instruction tuning after pretraining, and two-stage finetuning. Experiments across four base models (LLaMA 3, LLaMA 3.1, Qwen-2.5, and Qwen-2.5 Coder) and 20 benchmarks covering math, code, logic, and algorithms demonstrate consistent improvements. Notably, Tracepile boosts LLaMA3-8B by 9.2\% on average across nine math datasets and delivers clear gains on LiveCodeBench, CRUX, and Zebra Logic under two-stage finetuning.



Paperid:3438
Authors:Yuanfan Li, Yunwen Lei, Zheng-Chu Guo, Yiming Ying
Abstract:
Abstract:Recent advances have significantly improved our understanding of the generalization performance of gradient descent (GD) methods in deep neural networks. A natural and fundamental question is whether GD can achieve generalization rates comparable to the minimax optimal rates established in the kernel setting. Existing results either yield suboptimal rates of $O(1/\sqrt{n})$, or focus on networks with smooth activation functions, incurring exponential dependence on network depth $L$. In this work, we establish optimal generalization rates for GD with deep ReLU networks by carefully trading off optimization and generalization errors, achieving only polynomial dependence on depth. Specifically, under the assumption that the data are NTK separable from the margin $\gamma$, we prove an excess risk rate of $\widetilde{O}(L^4 (1 + \gamma L^2) / (n \gamma^2))$, which aligns with the optimal SVM-type rate $\widetilde{O}(1 / (n \gamma^2))$ up to depth-dependent factors.A key technical contribution is our novel control of activation patterns near a reference model, enabling a sharper Rademacher complexity bound for deep ReLU networks trained with gradient descent.



Paperid:3439
Authors:Wenlin Zhang, Xiangyang Li, Kuicai Dong, Yichao Wang, Pengyue Jia, Xiaopeng Li, Yingyi Zhang, Derong Xu, Zhaocheng Du, Huifeng Guo, Ruiming Tang, Xiangyu Zhao
Abstract:
Retrieval-augmented generation (RAG) enhances large language models (LLMs) by integrating external knowledge, yet traditional RAG systems struggle with static workflows and limited adaptability for complex, multistep reasoning tasks. Agentic RAG systems, such as DeepResearch, address these issues through dynamic retrieval, iterative context refinement, and adaptive workflows. However, recent methods like Search-R1, which rely on outcome-based reinforcement learning, face challenges such as low exploration efficiency, gradient conflict, and sparse reward signals. To tackle these limitations, we introduce ReasonRAG, a novel method that leverages RAG-ProGUIDE—a high-quality dataset providing fine-grained, process-level rewards for query generation, evidence extraction, and answer generation. By employing process-supervised reinforcement learning, ReasonRAG enhances LLMs’ autonomous capabilities in search, query generation, evidence extraction, and answer synthesis. Experimental results show that ReasonRAG, utilizing RAG-ProGUIDE, outperforms existing approaches like Search-R1 and traditional RAG systems, achieving superior performance on five benchmark datasets with only 5k training instances—significantly fewer than the 90k required by Search-R1. Our code is available at https://anonymous.4open.science/r/ReasonRAG-B442.



Paperid:3440
Authors:Suqin Yuan, Lei Feng, Bo Han, Tongliang Liu
Abstract:
Sample selection is a prevalent approach in learning with noisy labels, aiming to identify confident samples for training. Although existing sample selection methods have achieved decent results by reducing the noise rate of the selected subset, they often overlook that not all mislabeled examples harm the model's performance equally. In this paper, we demonstrate that mislabeled examples correctly predicted by the model early in the training process are particularly harmful to model performance. We refer to these examples as Mislabeled Easy Examples (MEEs). To address this, we propose Early Cutting, which introduces a recalibration step that employs the model's later training state to re-select the confident subset identified early in training, thereby avoiding misleading confidence from early learning and effectively filtering out MEEs. Experiments on the CIFAR, WebVision, and full ImageNet-1k datasets demonstrate that our method effectively improves sample selection and model performance by reducing MEEs.



Paperid:3441
Authors:Haoru Tan, Xiuzhe Wu, Sitong Wu, Shaofeng Zhang, Yanfeng Chen, Xingwu Sun, Jeanne Shen, Xiaojuan Qi
Abstract:
Reinforcement fine-tuning (RFT) is essential for enhancing the reasoning and generalization capabilities of large language models, but its success heavily relies on the quality of the training data. While data selection has been extensively studied in supervised learning, its role in reinforcement learning, particularly during the RFT stage, remains largely underexplored. In this work, we introduce RFT-DA, a data selection framework designed to identify and prioritize valuable training examples (i.e., assign high RFT-DA scores) while filtering out noisy or detrimental ones, thereby enhancing both the efficiency and effectiveness of reinforcement fine-tuning.At the core of RFT-DA is a sample-level ablation analysis that quantifies the importance of each training example by measuring how its removal affects the final training reward, offering a direct estimate of its contribution to model learning. To ensure scalability, we propose a first-order approximation of the RFT-DA score by backtracking through the optimization process and applying temporal differentiation to the sample-wise influence term, along with a first-order Taylor approximation to adjacent time steps. This yields a lightweight, gradient-based estimator that evaluates the alignment between an individual sample’s gradient and the average gradient direction of all training samples, where a higher degree of alignment implies greater training utility. Extensive experiments demonstrate that RFT-DA consistently improves reward performance and accelerates convergence in reinforcement fine-tuning. Notably, in mathematical reasoning tasks, we only require about 20\% of the data selected based on our data ablation score to achieve more stable training and superior results compared to using the entire dataset. The trained 7B model matches or even outperforms the o1-mini across various benchmarks.



Paperid:3442
Authors:Advik Basani, Xiao Zhang
Abstract:
LLMs have demonstrated impressive capabilities across various natural language processing tasks yet remain vulnerable to prompts, known as jailbreak attacks, carefully designed to bypass safety guardrails and elicit harmful responses. Traditional methods rely on manual heuristics that suffer from limited generalizability. Despite being automatic, optimization-based attacks often produce unnatural jailbreak prompts that can be easily detected by safety filters or require high computational costs due to discrete token optimization. This paper introducesGenerative Adversarial Suffix Prompter(GASP), a novel automated framework that can efficiently generate human-readable jailbreak prompts in a fully black-box setting. In particular, GASP leverages latent Bayesian optimization to craft adversarial suffixes by efficiently exploring continuous latent spaces, gradually optimizing the suffix generator to improve attack efficacy while balancing prompt coherence via a targeted iterative refinement procedure. Through comprehensive experiments, we show that GASP can produce natural adversarial prompts, significantly improving jailbreak success, reducing training times, and accelerating inference speed, thus making it an efficient and scalable solution for red-teaming LLMs.



Paperid:3443
Authors:Hongtao Xu, Wenting Shen, Yuanxin Wei, Ang Wang, Guo Runfan, Tianxing Wang, Yong Li, Mingzhen Li, Weile Jia
Abstract:
Long-context supervised fine-tuning (Long-SFT) plays a vital role in enhancing the performance of large language models (LLMs) on long-context tasks. To smoothly adapt LLMs to long-context scenarios, this process typically entails training on mixed datasets containing both long and short sequences. However, this heterogeneous sequence length distribution poses significant challenges for existing training systems, as they fail to simultaneously achieve high training efficiency for both long and short sequences, resulting in sub-optimal end-to-end system performance in Long-SFT.In this paper, we present a novel perspective on data scheduling to address the challenges posed by the heterogeneous data distributions in Long-SFT. We propose Skrull, a dynamic data scheduler specifically designed for efficient long-SFT. Through dynamic data scheduling, Skrull balances the computation requirements of long and short sequences, improving overall training efficiency. Furthermore, we formulate the scheduling process as a joint optimization problem and thoroughly analyze the trade-offs involved. Based on those analysis, Skrull employs a lightweight scheduling algorithm to achieve near-zero cost online scheduling in Long-SFT. Finally, we implement Skrull upon DeepSpeed, a state-of-the-art distributed training system for LLMs. Experimental results demonstrate that Skrull outperforms DeepSpeed by 3.76x on average (up to 7.54x) in real-world long-SFT scenarios.



Paperid:3444
Authors:Sen Bai, Chunqi Yang, Xin Bai, Xin Zhang, Zhengang Jiang
Abstract:
Binary (0-1) integer programming (BIP) is pivotal in scientific domains requiring discrete decision-making. As the advance of AI computing, recent works explore neural network-based solver for integer linear programming (ILP) problems. Yet, they lack scalability for tackling nonlinear challenges. To handle nonlinearities, state-of-the-art Branch-and-Cut solvers employ linear relaxations, leading to exponential growth in auxiliary variables and severe computation limitations. To overcome these limitations, we propose BIPNN, an unsupervised learning framework to solve BIP problems via hypergraph neural networks (HyperGNN). Specifically, (i) BIPNN reformulates BIPs-constrained, discrete, and nonlinear (sin, log, exp) optimization problems-into unconstrained, differentiable, and polynomial loss functions. The reformulation stems from the observation of a precise one-to-one mapping between polynomial BIP objectives and hypergraph structures, enabling the unsupervised training of HyperGNN to optimize BIP problems in an end-to-end manner. On this basis, (ii) we propose a GPU-accelerated and continuous-annealing-enhanced training pipeline for BIPNN. The pipeline enables BIPNN to optimize large-scale nonlinear terms in BIPs fully in parallel via straightforward gradient descent, thus significantly reducing the training cost while ensuring the generation of discrete, high-quality solutions. Extensive experiments on synthetic and real-world datasets highlight the superiority of our approach.



Paperid:3445
Authors:Woomin Song, Sai Muralidhar Jayanthi, Srikanth Ronanki, Kanthashree Mysore Sathyendra, Jinwoo Shin, Aram Galstyan, Shubham Katiyar, Sravan Babu Bodapati
Abstract:
As large language models increasingly gain popularity in real-world applications, processing extremely long contexts, often exceeding the model’s pre-trained context limits, has emerged as a critical challenge. While existing approaches to efficient long-context processing show promise, recurrent compression-based methods struggle with information preservation, whereas random access approaches require substantial memory resources. We introduce REFORM, a novel inference framework that efficiently handles long contexts through a two-phase approach. First, it incrementally processes input chunks while maintaining a compressed KV cache, constructs cross-layer context embeddings, and utilizes early exit strategy for improved efficiency. Second, it identifies and gathers essential tokens via similarity matching and selectively recomputes the KV cache. Compared to baselines, REFORM achieves over 50% and 27% performance gains on RULER and BABILong respectively at 1M context length. It also outperforms baselines on ∞-Bench, RepoEval, and MM-NIAH, demonstrating flexibility across diverse tasks and domains. Additionally, REFORM reduces inference time by 30% and peak memory usage by 5%, achieving both efficiency and superior performance.



Paperid:3446
Authors:Yuhao Sun, Zhenyi Zhang, Zihan Wang, Tiejun Li, Peijie Zhou
Abstract:
Recovering the dynamics from a few snapshots of a high-dimensional system is a challenging task in statistical physics and machine learning, with important applications in computational biology. Many algorithms have been developed to tackle this problem, based on frameworks such as optimal transport and the Schrödinger bridge. A notable recent framework is Regularized Unbalanced Optimal Transport (RUOT), which integrates both stochastic dynamics and unnormalized distributions. However, since most existing methods do not explicitly enforce optimality conditions, their solutions often struggle to satisfy the principle of least action and meet challenges to converge in a stable and reliable way. To address these issues, we propose Variational RUOT (Var-RUOT), a new framework to solve the RUOT problem. By incorporating the optimal necessary conditions for the RUOT problem into both the parameterization of the search space and the loss function design, Var-RUOT only needs to learn a scalar field to solve the RUOT problem and can search for solutions with lower action. We also examined the challenge of selecting a growth penalty function in the widely used Wasserstein-Fisher-Rao metric and proposed a solution that better aligns with biological priors in Var-RUOT. We validated the effectiveness of Var-RUOT on both simulated data and real single-cell datasets. Compared with existing algorithms, Var-RUOT can find solutions with lower action while exhibiting faster convergence and improved training stability.



Paperid:3447
Authors:Hritam Basak, Zhaozheng Yin
Abstract:
We address the problem of language-guided 3D affordance prediction, a core capability for embodied agents interacting with unstructured environments. Existing methods often rely on fixed affordance categories or require external expert prompts, limiting their ability to generalize across different objects and interpret multi-step instructions. In this work, we introduce \textit{ViSPLA}, a novel iterative self-prompting framework that leverages the intrinsic geometry of predicted masks for continual refinement. We redefine affordance detection as a language-conditioned segmentation task: given a 3D point cloud and language instruction, our model predicts a sequence of refined affordance masks, each guided by differential geometric feedback including Laplacians, normal derivatives, and curvature fields. This feedback is encoded into visual prompts that drive a multi-stage refinement decoder, enabling the model to self-correct and adapt to complex spatial structures. To further enhance precision and coherence, we introduce Implicit Neural Affordance Fields, which define continuous probabilistic regions over the 3D surface without additional supervision. Additionally, our Spectral Convolutional Self-Prompting module operates in the frequency domain of the point cloud, enabling multi-scale refinement that captures both coarse and fine affordance structures. Extensive experiments demonstrate that \textit{ViSPLA} achieves state-of-the-art results on both seen and unseen objects on two benchmark datasets. Our framework establishes a new paradigm for open-world 3D affordance reasoning by unifying language comprehension with low-level geometric perception through iterative refinement.



Paperid:3448
Authors:Marzieh Ajirak, Oded Bein, Ellen Bowen, Dora Kanellopoulos, Avital Falk, FAITH GUNNING, Nili Solomonov, Logan Grosenick
Abstract:
We propose a unified framework for adaptive routing in multitask, multimodal prediction settings where data heterogeneity and task interactions vary across samples. Motivated by applications in psychotherapy where structured assessments and unstructured clinician notes coexist with partially missing data and correlated outcomes, we introduce a routing-based architecture that dynamically selects modality processing pathways and task-sharing strategies on a per-sample basis. Our model defines multiple modality paths including raw and fused representations of text and numeric features and learns to route each input through the most informative expert combination. Task-specific predictions are produced by shared or independent heads depending on the routing decision, and the entire system is trained end-to-end. We evaluate the model on both synthetic data and real-world psychotherapy notes predicting depression and anxiety outcomes. Our experiments show that our method consistently outperforms fixed multitask or single-task baselines, and that the learned routing policy provides interpretable insights into modality relevance and task structure. This addresses critical challenges in personalized healthcare by enabling per-subject adaptive information processing that accounts for data heterogeneity and task correlations. Applied to psychotherapy, this framework could improve mental health outcomes, enhance treatment assignment precision, and increase clinical cost-effectiveness through personalized intervention strategies.



Paperid:3449
Authors:Justin Lazarow, Kai Kang, Afshin Dehghan
Abstract:
We revisit scene-level 3D object detection as the output of an object-centric framework capable of both localization and mapping using 3D oriented boxes as the underlying geometric primitive. While existing 3D object detection approaches operate globally and implicitly rely on the a priori existence of metric camera poses, our method, Rooms from Motion (RfM) operates on a collection of un-posed images. By replacing the standard 2D keypoint-based matcher of structure-from-motion with an object-centric matcher based on image-derived 3D boxes, we estimate metric camera poses, object tracks, and finally produce a global, semantic 3D object map. When a priori pose is available, we can significantly improve map quality through optimization of global 3D boxes against individual observations. RfM shows strong localization performance and subsequently produces maps of higher quality than leading point-based and multi-view 3D object detection methods on CA-1M and ScanNet++, despite these global methods relying on overparameterization through point clouds or dense volumes. Rooms from Motion achieves a general, object-centric representation which not only extends the work of Cubify Anything to full scenes but also allows for inherently sparse localization and parametric mapping proportional to the number of objects in a scene.



Paperid:3450
Authors:Eva Xie, Stefan Mihalas, Łukasz Kuśmierz
Abstract:
Growing evidence suggests that synaptic weights in the brain follow heavy-tailed distributions, yet most theoretical analyses of recurrent neural networks (RNNs) assume Gaussian connectivity. We systematically study the activity of RNNs with random weights drawn from biologically plausible Lévy alpha-stable distributions. While mean-field theory for the infinite system predicts that the quiescent state is always unstable---implying ubiquitous chaos---our finite-size analysis reveals a sharp transition between quiescent and chaotic dynamics. We theoretically predict the gain at which the finite system transitions from quiescent to chaotic dynamics, and validate it through simulations. Compared to Gaussian networks, finite heavy-tailed RNNs exhibit a broader gain regime near the edge of chaos, namely, a slow transition to chaos. However, this robustness comes with a tradeoff: heavier tails reduce the Lyapunov dimension of the attractor, indicating lower effective dimensionality. Our results reveal a biologically aligned tradeoff between the robustness of dynamics near the edge of chaos and the richness of high-dimensional neural activity. By analytically characterizing the transition point in finite-size networks---where mean-field theory breaks down---we provide a tractable framework for understanding dynamics in realistically sized, heavy-tailed neural circuits.



Paperid:3451
Authors:Meng-Hao Guo, Xuanyu Chu, Qianrui Yang, Zhe-Han Mo, Yiqing Shen, Pei-lin Li, Xinjie Lin, Jinnian Zhang, Xin-Sheng Chen, Yi Zhang, Kiyohiro Nakayama, Zhengyang Geng, Houwen Peng, Han Hu, Shi-min Hu
Abstract:
The rapid advancement of native multi-modal models and omni-models, exemplified by GPT-4o, Gemini and o3 with their capability to process and generate content across modalities such as text and images, marks a significant milestone in the evolution of intelligence. Systematic evaluation of their multi-modal output capabilities in visual thinking process (a.k.a., multi-modal chain of thought, M-CoT) becomes critically important. However, existing benchmarks for evaluating multi-modal models primarily focus on assessing multi-modal inputs and text-only reasoning process while neglecting the importance of reasoning through multi-modal outputs. In this paper, we present a benchmark, dubbed as RBench-V, designed to assess models’ vision-indispensable reasoning. To conduct RBench-V, we carefully hand-pick 803 questions covering math, physics, counting and games. Unlike problems in previous benchmarks, which typically specify certain input modalities, RBench-V presents problems centered on multi-modal outputs, which require image manipulation, such as generating novel images and constructing auxiliary lines to support reasoning process. We evaluate numerous open- and closed-source models on RBench-V, including o3, Gemini 2.5 pro, Qwen2.5-VL, etc. Even the best-performing model, o3, achieves only 25.8% accuracy on RBench-V, far below the human score of 82.3%, which shows current models struggle to leverage multi-modal reasoning. Data and code are available at https://evalmodels.github.io/rbenchv.



Paperid:3452
Authors:Jialong Zhou, Xiao Yang, Lichao Wang
Abstract:
The emergence of large language models (LLMs) enables the development of intelligent agents capable of engaging in complex and multi-turn dialogues. However, multi-agent collaboration face critical safety challenges, such as hallucination amplification and error injection and propagation. This paper presents GUARDIAN, a unified method for detecting and mitigating multiple safety concerns in GUARDing Intelligent Agent collaboratioNs. By modeling the multi-agent collaboration process as a discrete-time temporal attributed graph, GUARDIAN explicitly captures the propagation dynamics of hallucinations and errors. The unsupervised encoder-decoder architecture incorporating an incremental training paradigm, learns to reconstruct node attributes and graph structures from latent embeddings, enabling the identification of anomalous nodes and edges with unparalleled precision. Moreover, we introduce a graph abstraction mechanism based on the Information Bottleneck Theory, which compresses temporal interaction graphs while preserving essential patterns. Extensive experiments demonstrate GUARDIAN's effectiveness in safeguarding LLM multi-agent collaborations against diverse safety vulnerabilities, achieving state-of-the-art accuracy with efficient resource utilization.



Paperid:3453
Authors:Chen Xiong, Pin-Yu Chen, Tsung-Yi Ho
Abstract:
Recent advances in Large Language Models (LLMs) have spurred transformative applications in various domains, ranging from open-source to proprietary LLMs. However, jailbreak attacks, which aim to break safety alignment and user compliance by tricking the target LLMs into answering harmful and risky responses, are becoming an urgent concern. The practice of red-teaming for LLMs is to proactively explore potential risks and error-prone instances before the release of frontier AI technology. This paper proposes an agentic workflow to automate and scale the red-teaming process of LLMs through the Composition-of-Principles (CoP) framework, where human users provide a set of red-teaming principles as instructions to an AI agent to automatically orchestrate effective red-teaming strategies and generate jailbreak prompts. Distinct from existing red-teaming methods, our CoP framework provides a unified and extensible framework to encompass and orchestrate human-provided red-teaming principles to enable the automated discovery of new red-teaming strategies. When tested against leading LLMs, CoP reveals unprecedented safety risks by finding novel jailbreak prompts and improving the best-known single-turn attack success rate by up to 13.8 times.



Paperid:3454
Authors:Yibin Wang, li zhimin, Yuhang Zang, Chunyu Wang, Qinglin Lu, Cheng Jin, Jiaqi Wang
Abstract:
Recent advances in multimodal Reward Models (RMs) have shown significant promise in delivering reward signals to align vision models with human preferences. However, current RMs are generally restricted to providing direct responses or engaging in shallow reasoning processes with limited depth, often leading to inaccurate reward signals. We posit that incorporating explicit long chains of thought (CoT) into the reward reasoning process can significantly strengthen their reliability and robustness. Furthermore, we believe that once RMs internalize CoT reasoning, their direct response accuracy can also be improved through implicit reasoning capabilities. To this end, this paper proposes UnifiedReward-Think, the first unified multimodal CoT-based reward model, capable of multi-dimensional, step-by-step long-chain reasoning for both visual understanding and generation reward tasks. Specifically, we adopt an exploration-driven reinforcement fine-tuning approach to elicit and incentivize the model's latent complex reasoning ability: (1) We first use a small amount of image generation preference data to distill the reasoning process of GPT-4o, which is then used for the model's cold start to learn the format and structure of CoT reasoning. (2) Subsequently, by leveraging the model's prior knowledge and generalization capabilities, we prepare large-scale unified multimodal preference data to elicit the model's reasoning process across various vision tasks. During this phase, correct reasoning outputs are retained for rejection sampling to refine the model (3) while incorrect predicted samples are finally used for Group Relative Policy Optimization (GRPO) based reinforcement fine-tuning, enabling the model to explore diverse reasoning paths and optimize for correct and robust solutions. Extensive experiments confirm that incorporating long CoT reasoning significantly enhances the accuracy of reward signals. Notably, after mastering CoT reasoning, the model exhibits implicit reasoning capabilities, allowing it to surpass existing baselines even without explicit reasoning traces.



Paperid:3455
Authors:Zhuoran Qiao, Feizhi Ding, Thomas Dresselhaus, Mia Rosenfeld, Xiaotian Han, Owen Howell, Aniketh Iyengar, Stephen Opalenski, Anders Christensen, Sai Krishna Sirumalla, Fred Manby, Thomas Miller, Matthew Welborn
Abstract:
Biomolecular structure determination is essential to a mechanistic understanding of diseases and the development of novel therapeutics. Machine-learning-based structure prediction methods have made significant advancements by computationally predicting protein and bioassembly structures from sequences and molecular topology alone. Despite substantial progress in the field, challenges remain to deliver structure prediction models to real-world drug discovery. Here, we present NeuralPLexer3 -- a physics-inspired flow-based generative model that achieves state-of-the-art prediction accuracy on key biomolecular interaction types and improves training and sampling efficiency compared to its predecessors and alternative methodologies. Examined through existing and new benchmarks, NeuralPLexer3 excels in areas crucial to structure-based drug design, including blind docking, physical validity, and ligand-induced protein conformational changes.



Paperid:3456
Authors:Jiahuan Zhou, Chao Zhu, Zhenyu Cui, Zichen Liu, Xu Zou, Gang Hua
Abstract:
Continual Test-Time Adaptation (CTTA) aims to quickly fine-tune the model during the test phase so that it can adapt to multiple unknown downstream domain distributions without pre-acquiring downstream domain data. To this end, existing advanced CTTA methods mainly reduce the catastrophic forgetting of historical knowledge caused by irregular switching of downstream domain data by restoring the initial model or reusing historical models. However, these methods are usually accompanied by serious insufficient learning of new knowledge and interference from potentially harmful historical knowledge, resulting in severe performance degradation. To this end, we propose a class-aware domain Knowledge Fusion and Fission method for continual test-time adaptation, called KFF, which adaptively expands and merges class-aware domain knowledge in old and new domains according to the test-time data from different domains, where discriminative historical knowledge can be dynamically accumulated. Specifically, considering the huge domain gap within streaming data, a domain Knowledge FIssion (KFI) module is designed to adaptively separate new domain knowledge from a paired class-aware domain prompt pool, alleviating the impact of negative knowledge brought by old domains that are distinct from the current domain. Besides, to avoid the cumulative computation and storage overheads from continuously fissioning new knowledge, a domain Knowledge FUsion (KFU) module is further designed to merge the fissioned new knowledge into the existing knowledge pool with minimal cost, where a greedy knowledge dynamic merging strategy is designed to improve the compatibility of new and old knowledge while keeping the computational efficiency.



Paperid:3457
Authors:Junting Chen, Haotian Liang, Lingxiao Du, Weiyun Wang, Mengkang Hu, Yao Mu, Wenhai Wang, Jifeng Dai, Ping Luo, Wenqi Shao, Lin Shao
Abstract:
The rapid progress of navigation, manipulation, and vision models has made mobile manipulators capable in many specialized tasks. However, the open-world mobile manipulation (OWMM) task remains a challenge due to the need for generalization to open-ended instructions and environments, as well as the systematic complexity to integrate high-level decision making with low-level robot control based on both global scene understanding and current agent state.To address this complexity, we propose a novel multi-modal agent architecture that maintains multi-view scene frames and agent states for decision-making and controls the robot by function calling.A second challenge is the hallucination from domain shift. To enhance the agent performance, we further introduce an agentic data synthesis pipeline for the OWMM task to adapt the VLM model to our task domain with instruction fine-tuning.We highlight our fine-tuned OWMM-VLM as the first dedicated foundation model for mobile manipulators with global scene understanding, robot state tracking, and multi-modal action generation in a unified model. Through extensive experiments, we demonstrate that our model achieves state-of-the-art performance compared to other models.The project page is at https://owmm-vlm-project.github.io



Paperid:3458
Authors:Nan Wang, Lixing Xiao, Yuantao Chen, Weiqing Xiao, Pierre Merriaux, Lei Lei, Ziyang Yan, Saining Zhang, Shaocong Xu, chongjie Ye, Bohan Li, Zhaoxi Chen, Tianfan Xue, Hao Zhao
Abstract:
Neural rendering techniques, including NeRF and Gaussian Splatting (GS), rely on photometric consistency to produce high-quality reconstructions. However, in real-world scenarios, it is challenging to guarantee perfect photometric consistency in acquired images. Appearance codes have been widely used to address this issue, but their modeling capability is limited, as a single code is applied to the entire image. Recently, the bilateral grid was introduced to perform pixel-wise color mapping, but it is difficult to optimize and constrain effectively. In this paper, we propose a novel multi-scale bilateral grid that unifies appearance codes and bilateral grids. We demonstrate that this approach significantly improves geometric accuracy in dynamic, decoupled autonomous driving scene reconstruction, outperforming both appearance codes and bilateral grids. This is crucial for autonomous driving, where accurate geometry is important for obstacle avoidance and control. Our method shows strong results across four datasets: Waymo, NuScenes, Argoverse, and PandaSet. We further demonstrate that the improvement in geometry is driven by the multi-scale bilateral grid, which effectively reduces floaters caused by photometric inconsistency.



Paperid:3459
Authors:Yuting Huang, Ziquan Fang, Zhihao Zeng, Lu Chen, Yunjun Gao
Abstract:
Abstract:Spatio-temporal prediction plays a crucial role in intelligent transportation, weather forecasting, and urban planning. While integrating multi-modal data has shown potential for enhancing prediction accuracy, key challenges persist: (i) inadequate fusion of multi-modal information, (ii) confounding factors that obscure causal relations, and (iii) high computational complexity of prediction models. To address these challenges, we propose E$^2$-CSTP, an Effective and Efficient Causal multi-modal Spatio-Temporal Prediction framework. E$^2$-CSTP leverages cross-modal attention and gating mechanisms to effectively integrate multi-modal data. Building on this, we design a dual-branch causal inference approach: the primary branch focuses on spatio-temporal prediction, while the auxiliary branch mitigates bias by modeling additional modalities and applying causal interventions to uncover true causal dependencies. To improve model efficiency, we integrate GCN with the Mamba architecture for accelerated spatio-temporal encoding. Extensive experiments on 4 real-world datasets show that E$^2$-CSTP significantly outperforms 9 state-of-the-art methods, achieving up to 9.66% improvements in accuracy as well as 17.37%-56.11% reductions in computational overhead. All code and data are publicly available at https://anonymous.4open.science/r/E2-CSTP.



Paperid:3460
Authors:Jianing Chen, Zehao Li, Yujun Cai, Hao Jiang, Chengxuan Qian, Juyuan Kang, Shuqin Gao, Honglong Zhao, Tianlu Mao, Yucheng Zhang
Abstract:
Reconstructing dynamic 3D scenes from monocular videos remains a fundamental challenge in 3D vision. While 3D Gaussian Splatting (3DGS) achieves real-time rendering in static settings, extending it to dynamic scenes is challenging due to the difficulty of learning structured and temporally consistent motion representations. This challenge often manifests as three limitations in existing methods: redundant Gaussian updates, insufficient motion supervision, and weak modeling of complex non-rigid deformations. These issues collectively hinder coherent and efficient dynamic reconstruction. To address these limitations, we propose HAIF-GS, a unified framework that enables structured and consistent dynamic modeling through sparse anchor-driven deformation. It first identifies motion-relevant regions via an Anchor Filter to suppresses redundant updates in static areas. A self-supervised Induced Flow-Guided Deformation module induces anchor motion using multi-frame feature aggregation, eliminating the need for explicit flow labels. To further handle fine-grained deformations, a Hierarchical Anchor Propagation mechanism increases anchor resolution based on motion complexity and propagates multi-level transformations. Extensive experiments on synthetic and real-world benchmarks validate that HAIF-GS significantly outperforms prior dynamic 3DGS methods in rendering quality, temporal coherence, and reconstruction efficiency.



Paperid:3461
Authors:Ye Tian, Angela McCarthy, Gabriel Gomide, Nancy Liddle, Jedrzej Golebka, Royce Chen, Jeff Liebmann, Kaveri Thakoor
Abstract:
Medical imaging super-resolution is critical for improving diagnostic utility and reducing costs, particularly for low-cost modalities such as portable Optical Coherence Tomography (OCT). We propose OCTDiff, a Bridged Diffusion Model designed to enhance image resolution and quality from portable OCT devices. Our image-to-image diffusion framework addresses key challenges in the conditional generation process of denoising diffusion probabilistic models (DDPMs). A novel component, Adaptive Noise Aggregation (ANA), is introduced to improve denoising dynamics within the reverse diffusion. Additionally, we integrate Multi-Scale Cross-Attention (MSCA) into the U-Net backbone to capture local dependencies across spatial resolutions. To address overfitting on small clinical datasets and to preserve fine structural details essential for retinal diagnostics, we design a customized loss function guided by clinical quality scores. OCTDiff outperforms convolutional baselines and standard DDPMs, achieving state-of-the-art performance on clinical portable OCT datasets. Our model and its downstream applications can potentially be generalizable to other medical imaging modalities and revolutionize the current workflow of ophthalmic diagnostics. The code will be made available upon acceptance.



Paperid:3462
Authors:Chenfei Gu, Qiangqiang Zhang, Ting Li, Jinhan Xie, Niansheng Tang
Abstract:
The growing prevalence of streaming data and increasing concerns over data privacy pose significant challenges for traditional nonparametric regression methods, which are often ill-suited for real-time, privacy-aware learning. In this paper, we tackle these issuesby first proposing a novel one-pass online functional stochastic gradient descent algorithm that leverages the Huber loss (H-FSGD), to improve robustness against outliers and heavy-tailed errors in dynamic environments. To further accommodate privacy constraints, we introduce a locally differentially private extension, Private H-FSGD (PH-FSGD), designed to real-time, privacy-preserving estimation. Theoretically, we conduct a comprehensive non-asymptotic convergence analysis of the proposed estimators, establishing finite-sample guarantees and identifying optimal step size schedules that achieve optimal convergence rates. In particular, we provide practical insights into the impact of key hyperparameters, such as step size and privacy budget, on convergence behavior. Extensive experiments validate our theoretical findings, demonstrating that our methods achieve strong robustness and privacy protection without sacrificing efficiency.



Paperid:3463
Authors:Wei Zhuo, Zhaohuan Zhan, Han Yu
Abstract:
Federated learning (FL) on graph-structured data typically faces non-IID challenges, particularly in scenarios where each client holds a distinct subgraph sampled from a global graph. In this paper, we introduce Federated learning with Auxiliary projections (FedAux), a personalized subgraph FL framework that learns to align, compare, and aggregate heterogeneously distributed local models without sharing raw data or node embeddings. In FedAux, each client jointly trains (i) a local GNN and (ii) a learnable auxiliary projection vector (APV) that differentiably projects node embeddings onto a 1D space. A soft-sorting operation followed by a lightweight 1D convolution refines these embeddings in the ordered space, enabling the APV to effectively capture client-specific information. After local training, these APVs serve as compact signatures that the server uses to compute inter‑client similarities and perform similarity‑weighted parameter mixing, yielding personalized models while preserving cross‑client knowledge transfer. Moreover, we provide rigorous theoretical analysis to establish the convergence and rationality of our design. Empirical evaluations across diverse graph benchmarks demonstrate that FedAux substantially outperforms existing baselines in both accuracy and personalization performance. The code is anonymously available here.



Paperid:3464
Authors:Adrian Hill, Neal McKee, Stefan Bluecher, Johannes Maeß, Klaus-Robert Müller
Abstract:
Explaining complex machine learning models is a fundamental challenge when developing safe and trustworthy deep learning applications. To date, a broad selection of explainable AI (XAI) algorithms exist. One popular choice is SmoothGrad, which has been conceived to alleviate the well-known shattered gradient problem by smoothing gradients through convolution. SmoothGrad proposes to solve this high-dimensional convolution integral by sampling -- typically approximating the convolution with limited precision. Higher numbers of samples would amount to higher precision in approximating the convolution but also to higher computing demand, therefore in practice only few samples are used in SmoothGrad. In this work we propose a well founded novel methodSmoothDiffto resolve this tradeoff yielding aspeedup of over two orders of magnitude. Specifically,SmoothDiffleverages automatic differentiation to decompose the expected values of Jacobians across a network architecture, directly targeting only the non-linearities responsible for shattered gradients and making it easy to implement. We demonstrate SmoothDiff's excellent speed and performance in a number of experiments and benchmarks. Thus, SmoothDiff greatly enhances the usability (quality and speed) of SmoothGrad -- a popular workhorse of XAI.



Paperid:3465
Authors:Jiaxin Huang, Ziwen Li, Hanlue Zhang, Runnan Chen, Zhengqing Gao, Xiao He, Yandong Guo, Wenping Wang, Tongliang Liu, Mingming Gong
Abstract:
The integration of language and 3D perception is critical for embodied AI and robotic systems to perceive, understand, and interact with the physical world. Spatial reasoning, a key capability for understanding spatial relationships between objects, remains underexplored in current 3D vision-language research. Existing datasets often mix semantic cues (e.g., object name) with spatial context, leading models to rely on superficial shortcuts rather than genuinely interpreting spatial relationships. To address this gap, we introduce Surprise3D, a novel dataset designed to evaluate language-guided spatial reasoning segmentation in complex 3D scenes. Surprise3D consists of more than 200k vision language pairs across 900+ detailed indoor scenes from ScanNet++ v2, including more than 2.8k unique object classes. The dataset contains 89k+ human-annotated spatial queries deliberately crafted without object name, thereby mitigating shortcut biases in spatial understanding. These queries comprehensively cover various spatial reasoning skills, such as relative position, narrative perspective, parametric perspective, and absolute distance reasoning. Initial benchmarks demonstrate significant challenges for current state-of-the-art expert 3D visual grounding methods and 3D-LLMs, underscoring the necessity of our dataset and the accompanying 3D Spatial Reasoning Segmentation (3D-SRS) benchmark suite. Surprise3D and 3D-SRS aim to facilitate advancements in spatially aware AI, paving the way for effective embodied interaction and robotic planning.



Paperid:3466
Authors:Changguang WU, Jiangxin Dong, Chengjian Li, Jinhui Tang
Abstract:
We presentPlenodium(plenoptic medium), an effective and efficient 3D representation framework capable of jointly modeling both objects and participating media.In contrast to existing medium representations that rely solely on view-dependent modeling, our novel plenoptic medium representation incorporates both directional and positional information through spherical harmonics encoding, enabling highly accurate underwater scene reconstruction.To address the initialization challenge in degraded underwater environments, we propose the pseudo-depth Gaussian complementation to augment COLMAP-derived point clouds with robust depth priors.In addition, a depth ranking regularized loss is developed to optimize the geometry of the scene and improve the ordinal consistency of the depth maps.Extensive experiments on real-world underwater datasets demonstrate that our method achieves significant improvements in 3D reconstruction.Furthermore, we conduct a simulated dataset with ground truth and the controllable scattering medium to demonstrate the restoration capability of our method in underwater scenarios. Our code is available at: https://anonymous.4open.science/r/plenodium-1119.



Paperid:3467
Authors:Yide Qiu, Tong Zhang, Xing Cai, Hui Yan, Zhen Cui
Abstract:
The non-Euclidean geometry inherent in graph structures fundamentally impedes cross-graph knowledge transfer. Drawing inspiration from texture transfer in computer vision, we pioneer topological primitives as transferable semantic units for graph structural knowledge. To address three critical barriers - the absence of specialized benchmarks, aligned semantic representations, and systematic transfer methodologies - we present G²SN-Transfer, a unified framework comprising: (i) TopoGraph-Mapping that transforms non-Euclidean graphs into transferable sequences via topological primitive distribution dictionaries; (ii) G²SN, a dual-stream architecture learning text-topology aligned representations through contrastive alignment; and (iii) AdaCross-Transfer, a data-adaptive knowledge transfer mechanism leveraging cross-attention for both full-parameter and parameter-frozen scenarios. Particularly, G²SN is a dual-stream sequence network driven by ordinary differential equations, and our theoretical analysis establishes the convergence guarantee of G²SN. We construct STA-18, the first large-scale benchmark with aligned topological primitive-text pairs across 18 diverse graph datasets. Comprehensive evaluations demonstrate that G²SN achieves state-of-the-art performance on four structural learning tasks (average 3.2\% F1-score improvement), while our transfer method yields consistent enhancements across 13 downstream tasks (5.2\% average gains) including 10 large-scale graph datasets. The datasets and code are available at https://anonymous.4open.science/r/UGSKT-C10E/.



Paperid:3468
Authors:Guy Moss, Leah Muhle, Reinhard Drews, Jakob H Macke, Cornelius Schröder
Abstract:
Simulation-based inference (SBI) is an established approach for performing Bayesian inference on scientific simulators. SBI so far works best on low-dimensional parametric models. However, it is difficult to infer function-valued parameters, which frequently occur in disciplines that model spatiotemporal processes such as the climate and earth sciences. Here, we introduce an approach for efficient posterior estimation, using a Fourier Neural Operator (FNO) architecture with a flow matching objective. We show that our approach, FNOPE, can perform inference of function-valued parameters at a fraction of the simulation budget of state of the art methods. In addition, FNOPE supports posterior evaluation at arbitrary discretizations of the domain, as well as simultaneous estimation of vector-valued parameters. We demonstrate the effectiveness of our approach on several benchmark tasks and a challenging spatial inference task from glaciology. FNOPE extends the applicability of SBI methods to new scientific domains by enabling the inference of function-valued parameters.



Paperid:3469
Authors:Liuyuan Jiang, Quan Xiao, Lisha Chen, Tianyi Chen
Abstract:
Bilevel optimization, a hierarchical optimization paradigm, has gained significant attention in a wide range of practical applications, notably in the fine-tuning of generative models. However, due to the nested problem structure, most existing algorithms require either the Hessian vector calculation or the nested loop updates, which are computationally inefficient in large language model (LLM) fine-tuning. In this paper, building upon the fully first-order penalty-based approach, we propose an efficient value function-free (\textsf{PBGD-Free}) algorithm that eliminates the loop of solving the lower-level problem and admits fully single-loop updates. Inspired by the landscape analysis of representation learning-based LLM fine-tuning problem, we propose a relaxed flatness condition for the upper-level function and prove the convergence of the proposed value-function-free algorithm. We test the performance of the proposed algorithm in various applications and demonstrate its superior computational efficiency over the state-of-the-art bilevel methods.



Paperid:3470
Authors:Sang-Hyun Lee
Abstract:
Reinforcement learning (RL) typically assumes repetitive resets to provide an agent with diverse and unbiased experiences. These resets require significant human intervention and result in poor training efficiency in real-world settings. Autonomous RL (ARL) addresses this challenge by jointly training forward and reset policies. While recent ARL algorithms have shown promise in reducing human intervention, they assume narrow support over the distributions of initial or goal states and rely on task-specific knowledge to identify irreversible states. In this paper, we propose a robust and scalable ARL algorithm, called RSA, that enables an agent to handle diverse initial and goal states and to avoid irreversible states without task-specific knowledge. RSA generates a curriculum by identifying informative states based on the learning progress of an agent. We hypothesize that informative states are neither overly difficult nor trivially easy for the agent being trained. To detect and avoid irreversible states without task-specific knowledge, RSA encodes the behaviors exhibited in those states rather than the states themselves. Experimental results demonstrate that RSA outperforms existing ARL algorithms with fewer manual resets in both reversible and irreversible environments.



Paperid:3471
Authors:Zhiqi Pang, Lingling Zhao, Junjie Wang, Chunyu Wang
Abstract:
Visible-infrared person re-identification (VI-ReID) aims to match visible and infrared images of the same individual. Supervised VI-ReID (SVI-ReID) methods have achieved promising performance under the guidance of manually annotated identity labels. However, the substantial annotation cost severely limits their scalability in real-world applications. As a result, unsupervised VI-ReID (UVI-ReID) methods have attracted increasing attention. These methods typically rely on pseudo-labels generated by clustering and matching algorithms to replace manual annotations. Nevertheless, the quality of pseudo-labels is often difficult to guarantee, and low-quality pseudo-labels can significantly hinder model performance improvements. To address these challenges, we explore the use of attribute arrays extracted by a large vision-language model (LVLM) to enhance VI-ReID, and propose a novel LVLM-driven attribute-aware modeling (LVLM-AAM) approach. Specifically, we first design an attribute-aware reliable labeling strategy, which refines intra-modality clustering results based on image-level attributes and improves inter-modality matching by grouping clusters according to cluster-level attributes. Next, we develop an explicit-implicit attribute fusion module, which integrates explicit and implicit attributes to obtain more fine-grained identity-related text features. Finally, we introduce an attribute-aware contrastive learning module, which jointly leverages static and dynamic text features to promote modality-invariant feature learning. Extensive experiments conducted on VI-ReID datasets validate the effectiveness of the proposed LVLM-AAM and its individual components. LVLM-AAM not only significantly outperforms existing unsupervised methods but also surpasses several supervised methods.



Paperid:3472
Authors:Hongyi Jin, Zijun Ding, Dung Daniel Ngo, Steven Wu
Abstract:
In recent years, multicalibration has emerged as a desirable learning objective for ensuring that a predictor is calibrated across a rich collection of overlapping subpopulations. Existing approaches typically achieve multicalibration by discretizing the predictor's output space and iteratively adjusting its output values. However, this discretization approach departs from the standard empirical risk minimization (ERM) pipeline, introduces rounding error \hjedit{and additional sensitive hyperparameter}, and may distort the predictor’s outputs in ways that hinder downstream decision-making.In this work, we propose a discretization-free multicalibration method that directly optimizes an empirical risk objective over an ensemble of depth-two decision trees. Our ERM approach can be implemented using off-the-shelf tree ensemble learning methods such as \hjedit{LightGBM}. Our algorithm provably achieves multicalibration, provided that the data distribution satisfies a technical condition we term as loss saturation. Across multiple datasets, our empirical evaluation shows that this condition is always met in practice. Our discretization-free algorithm consistently matches or outperforms existing multicalibration approaches---even when evaluated using a discretization-based multicalibration metric that shares its discretization granularity with the baselines.



Paperid:3473
Authors:Jiacong Chen, Qingyu Mao, Youneng Bao, Xiandong MENG, Fanyang Meng, Ronggang Wang, Yongsheng Liang
Abstract:
3D Gaussian Splatting (3DGS) has emerged as a high-fidelity and efficient paradigm for online free-viewpoint video (FVV) reconstruction, offering viewers rapid responsiveness and immersive experiences. However, existing online methods face challenge in prohibitive storage requirements primarily due to point-wise modeling that fails to exploit the motion properties. To address this limitation, we propose a novel Compact Gaussian Streaming (ComGS) framework, leveraging the locality and consistency of motion in dynamic scene, that models object-consistent Gaussian point motion through keypoint-driven motion representation. By transmitting only the keypoint attributes, this framework provides a more storage-efficient solution. Specifically, we first identify a sparse set of motion-sensitive keypoints localized within motion regions using a viewspace gradient difference strategy. Equipped with these keypoints, we propose an adaptive motion-driven mechanism that predicts a spatial influence field for propagating keypoint motion to neighboring Gaussian points with similar motion. Moreover, ComGS adopts an error-aware correction strategy for key frame reconstruction that selectively refines erroneous regions and mitigates error accumulation without unnecessary overhead. Overall, ComGS achieves a remarkable storage reduction of over 159 × compared to 3DGStream and 14 × compared to the SOTA method QUEEN, while maintaining competitive visual fidelity and rendering speed. Our code will be released.



Paperid:3474
Authors:Yuandong Tian
Abstract:
Abstract:We prove rich algebraic structures of the solution space for 2-layer neural networks with quadratic activation and $L_2$ loss, trained on reasoning tasks in Abelian group (e.g., modular addition). Such a rich structure enables analytical construction of global optimal solutions from partial solutions that only satisfy part of the loss, despite its high nonlinearity. We coin the framework as CoGS (Composing Global Solutions}). Specifically, we show that the weight space over different numbers of hidden nodes of the 2-layer network is equipped with a semi-ring algebraic structure, and the loss function to be optimized consists of monomial potentials, which are ring homomorphisms, allowing partial solutions to be composed into global ones by ring addition and multiplication. Our experiments show that around $95\%$ of the solutions obtained by gradient descent match exactly our theoretical constructions. Although the global solutions constructed only required a small number of hidden nodes, our analysis on gradient dynamics shows that overparameterization asymptotically decouples training dynamics and is beneficial. We further show that training dynamics favors simpler solutions under weight decay, and thus high-order global solutions such as perfect memorization are unfavorable. The anonymous repo is https://anonymous.4open.science/r/anon_submission-73B3/



Paperid:3475
Authors:Jiaqi Wei, Hao Zhou, Xiang Zhang, Di Zhang, Zijie Qiu, Noah Wei, Jinzhe Li, Wanli Ouyang, Siqi Sun
Abstract:
Retrieval-augmented generation (RAG) has become a widely adopted paradigm for enabling knowledge-grounded large language models (LLMs). However, standard RAG pipelines often fail to ensure that model reasoning remains consistent with the evidence retrieved, leading to factual inconsistencies or unsupported conclusions.In this work, we reinterpret RAG as \textit{Retrieval-Augmented Reasoning} and identify a central but underexplored problem: \textit{Reasoning Misalignment}—the divergence between an LLM's internal reasoning trajectory and the evidential constraints provided by retrieval. To address this issue, we propose \textsc{AlignRAG}, a novel iterative framework grounded in \textit{Critique-Driven Alignment (CDA)}. At the heart of \textsc{AlignRAG} lies a \textit{contrastive critique synthesis} mechanism that generates retrieval-sensitive critiques while mitigating self-bias. This mechanism trains a dedicated retrieval-augmented \textit{Critic Language Model (CLM)} using labeled critiques that distinguish between evidence-aligned and misaligned reasoning. Alignment signals for supervision are obtained through self-supervised or externally guided labeling strategies.The resulting CLM is explicitly optimized for evidence sensitivity, enabling it to detect and revise reasoning errors during inference without relying solely on self-generated feedback. Empirical evaluations show that our 8B-parameter CLM improves performance over the Self-Refine baseline by \textbf{12.1\%} on out-of-domain tasks and outperforms a standard 72B-parameter CLM by \textbf{2.2\%}, while remaining compatible with existing RAG architectures as a \textit{plug-and-play} module.Moreover, \textsc{AlignRAG} demonstrates strong robustness under both informative and noisy retrieval scenarios, validating its ability to \textit{thrive} even when retrieval quality degrades. Overall, \textsc{AlignRAG} offers a principled solution for aligning model reasoning with retrieved evidence, substantially improving the factual reliability and robustness of RAG systems. Our source code is provided at an anonymous \href{https://anonymous.4open.science/r/RAG-55E7/}{link.}



Paperid:3476
Authors:Weitong Zhang, Chengqi Zang, Bernhard Kainz
Abstract:
Large Language Models exhibit logical inconsistency across multi-turn inference processes, undermining correctness in complex inferential tasks. Challenges arise from ensuring that outputs align with both factual correctness and human intent. Approaches like single-agent reflection and multi-agent debate frequently prioritize consistency, but at the expense of accuracy. To address this problem, we propose a novel game-theoretic consensus mechanism that enables LLMs to self-check their outputs during the decoding stage of output generation. Our method models the decoding process as a multistage Bayesian Decoding Game, where strategic interactions dynamically converge to a consensus on the most reliable outputs without human feedback or additional training. Remarkably, our game design allows smaller models to outperform much larger models through game mechanisms (e.g., 78.1 LLaMA13B vs. 76.6 PaLM540B). As a model-agnostic method, our approach consistently improves even the latest models, enhancing DeepSeek-7B's performance on MMLU by 12.4%. Our framework effectively balances correctness and consistency, demonstrating that properly designed game-theoretic mechanisms can significantly enhance the self-verification capabilities of language models across various tasks and model architectures.



Paperid:3477
Authors:Marianna Nezhurina, Tomer Porian, Giovanni Puccetti, Tommie Kerssies, Romain Beaumont, Mehdi Cherti, Jenia Jitsev
Abstract:
Abstract:In studies of transferable learning, scaling laws are obtained for various important foundation models to predict their properties and performance at larger scales. We show here how scaling law derivation can also be used for model and dataset comparison, allowing to decide which procedure is to be preferred for pre-training. For the first time, full scaling laws based on dense measurements across wide span of model and samples seen scales are derived for two important language-vision learning procedures, CLIP and MaMMUT, that use either contrastive only or contrastive and captioning text generative loss. Ensuring sufficient prediction accuracy for held out points, we use derived scaling laws to compare both models, obtaining evidence for MaMMUT stronger improvement with scale and better sample efficiency than standard CLIP. To strengthen validity of the comparison, we show scaling laws for various downstream tasks, looking at classification, retrieval, and segmentation, and for two different open datasets, DataComp-1.4B and Re-LAION-1.4B, observing consistently same trends. We show that comparison can be also performed when deriving scaling laws with constant learning rate schedule that reduces compute cost. Accurate derivation of scaling laws provides thus means to perform model and dataset comparison across scale span that avoids misleading conclusions based on measurements from single reference scales only, paving the road for systematic comparison and improvement of open foundation models and datasets for their creation. We release all the pre-trained models with their intermediate checkpoints, including openMammut-L/14, which achieves $80.3$% zero-shot ImageNet-1k accuracy, trained on 12.8B samples from DataComp-1.4B. Code and data to reproduce findings will be released at \url{https://anonymous.4open.science/anonymize/open_clip_scaling-76D2}



Paperid:3478
Authors:Hyeonggeun Han, Sehwan Kim, Hyungjun Joo, Sangwoo Hong, Jungwoo Lee
Abstract:
Despite their impressive generative capabilities, text-to-image diffusion models often memorize and replicate training data, prompting serious concerns over privacy and copyright. Recent work has attributed this memorization to an attraction basin—a region where applying classifier-free guidance (CFG) steers the denoising trajectory toward memorized outputs—and has proposed deferring CFG application until the denoising trajectory escapes this basin. However, such delays often result in non-memorized images that are poorly aligned with the input prompts, highlighting the need to promote earlier escape so that CFG can be applied sooner in the denoising process. In this work, we show that the initial noise sample plays a crucial role in determining when this escape occurs. We empirically observe that different initial samples lead to varying escape times. Building on this insight, we propose two mitigation strategies that adjust the initial noise—either collectively or individually—to find and utilize initial samples that encourage earlier basin escape. These approaches significantly reduce memorization while preserving image-text alignment.



Paperid:3479
Authors:Yanbo Wang, Zixiang Xu, Yue Huang, Gao Chujie, Siyuan Wu, Jiayi Ye, Pin-Yu Chen, Xiuying Chen, Xiangliang Zhang
Abstract:
Large Language Models (LLMs) often struggle to maintain their original performance when faced with semantically coherent but task-irrelevant contextual information. Although prior studies have explored this issue using fixed-template or retrieval-based distractions, such static methods show limited effectiveness against contemporary models. To address this problem, we propose a dynamic distraction generation framework based on tree search, where the generation process is guided by model behavior. Without modifying the original question or answer, the method efficiently produces challenging adaptive distractions across multiple datasets, enabling systematic stress testing of LLMs’ contextual robustness. Experiments on four benchmarks demonstrate that the generated distractions lead to an average performance drop of over 45\% for mainstream models. Further comparisons of mitigation strategies show that prompt-based optimization methods yield limited gains, whereas post-training approaches (e.g., DPO) significantly enhance the model's contextual robustness. The results indicate that these issues do not stem from knowledge deficits in LLMs, but from a fundamental inability to maintain consistent reasoning under contextual distraction, posing a major challenge to the reliability of LLMs in real-world applications.



Paperid:3480
Authors:Vittorio Giammarino, Ruiqi Ni, Ahmed Qureshi
Abstract:
Offline Goal-Conditioned Reinforcement Learning (GCRL) holds great promise for domains such as autonomous navigation and locomotion, where collecting interactive data is costly and unsafe. However, it remains challenging in practice due to the need to learn from datasets with limited coverage of the state-action space and to generalize across long-horizon tasks. To improve on these challenges, we propose a Physics-informed (Pi) regularized loss for value learning, derived from the Eikonal Partial Differential Equation (PDE) and which induces a geometric inductive bias in the learned value function. Unlike generic gradient penalties that are primarily used to stabilize training, our formulation is grounded in continuous-time optimal control and encourages value functions to align with cost-to-go structures. The proposed regularizer is broadly compatible with temporal-difference-based value learning and can be integrated into existing Offline GCRL algorithms. When combined with Hierarchical Implicit Q-Learning (HIQL), the resulting method, Physics-informed HIQL (Pi-HIQL), yields significant improvements in both performance and generalization, with pronounced gains in stitching regimes and large-scale navigation tasks. Code is available at link.



Paperid:3481
Authors:Zhengliang Shi, Lingyong Yan, Dawei Yin, Suzan Verberne, Maarten Rijke, Zhaochun Ren
Abstract:
Large language models (LLMs) have been widely integrated into information retrieval to advance traditional techniques. However, effectively enabling LLMs to seek accurate knowledge in complex tasks remains a challenge due to the complexity of multi-hop queries as well as the irrelevant retrieved content. To address these limitations, we propose ExSearch, an agentic search framework, where the LLM learns to retrieve useful information as the reasoning unfolds through a self-incentivized process. At each step, the LLM decides what to retrieve (thinking), triggers an external retriever (search), and extracts fine-grained evidence (recording) to support next-step reasoning. To enable LLM with this capability, we adopts a Generalized Expectation-Maximization algorithm. In the E-step, the LLM generates multiple search trajectories and assigns an importance weight to each; the M-step trains the LLM on them with a re-weighted loss function. This creates a self-incentivized loop, where the LLM iteratively learns from its own generated data, progressively improving itself for search. We further theoretically analyze this training process, establishing convergence guarantees. Extensive experiments on four knowledge-intensive benchmarks show that ExSearchS substantially outperforms baselines, e.g., +7.8% improvement on exact match score. Motivated by these promising results, we introduce ExSearch-Zoo, an extension that extends our method to broader scenarios, to facilitate future work.



Paperid:3482
Authors:haowei hua, Wanyu Lin
Abstract:
Crystal structures are defined by the periodic arrangement of atoms in 3D space, inherently making them equivariant to SO(3) group. A fundamental requirement for crystal property prediction is that the model's output should remain invariant to arbitrary rotational transformations of the input structure. One promising strategy to achieve this invariance is to align the given crystal structure into a canonical orientation with appropriately computed rotations, or called frames.However, existing work either only considers a global frame or solely relies on more advanced local frames based on atoms' local structure. A global frame is too coarse to capture the local structure heterogeneity of the crystal, while local frames may inadvertently disrupt crystal symmetry, limiting their expressivity.In this work, we revisit the frame design problem for crystalline materials and propose a novel approach to construct expressive {\bf S}ymmetry Preserving Frames, dubbed as SPFrame, for modeling crystal structures. Specifically, this local-global associative frame constructs invariant local frames rather than equivariant ones, thereby preserving the symmetry of the crystal. In parallel, it integrates global structural information to construct an equivariant global frame to enforce SO(3) invariance.Extensive experimental results demonstrate that SPFrame consistently outperforms traditional frame construction techniques and existing crystal property prediction baselines across multiple benchmark tasks.



Paperid:3483
Authors:Hongyi Zhou, Jin Zhu, Pingfan Su, Kai Ye, Ying Yang, Shakeel Gavioli-Akilagun, Chengchun Shi
Abstract:
We study the problem of determining whether a piece of text has been authored by a human or by a large language model (LLM). Existing state of the art logits-based detectors make use of statistics derived from the log-probability of the observed text evaluated using the distribution function of a given source LLM. However, relying solely on log probabilities can be sub-optimal. In response, we introduce AdaDetectGPT -- a novel classifier that adaptively learns a witness function from training data to enhance the performance of logits-based detectors. We provide statistical guarantees on its true positive rate, false positive rate, true negative rate and false negative rate. Extensive numerical studies show AdaDetectGPT nearly uniformly improves the state-of-the-art method in various combination of datasets and LLMs, and the improvement can reach up to 58%.



Paperid:3484
Authors:Bo-Wen Yin, Jiao-Long Cao, Xuying Zhang, Yuming Chen, Ming-Ming Cheng, Qibin Hou
Abstract:
Recent research on representation learning has proved the merits of multi-modal clues for robust semantic segmentation. Nevertheless, a flexible pretrain-and-finetune pipeline for multiple visual modalities remains unexplored. In this paper, we propose a novel multi-modal learning framework, termed OmniSegmentor. It has two key innovations: 1) Based on ImageNet, we assemble a large-scale dataset for multi-modal pretraining, called OmniSegmentor, which contains five popular visual modalities; 2) We provide an efficient pretraining manner to endow the model with the capacity to encode different modality information in the OmniSegmentor. For the first time, we introduce a universal multi-modal pretraining framework that consistently amplifies the model's perceptual capabilities across various scenarios, regardless of the arbitrary combination of the involved modalities. Remarkably, our OmniSegmentor achieves new state-of-the-art records on a wide range of multi-modal semantic segmentation datasets, including NYU Depthv2, EventScape, MFNet, DeLiVER, SUNRGBD, and KITTI-360. Data, model checkpoints, and source code will be made publicly available.



Paperid:3485
Authors:Fangrui Zhu, Hanhui Wang, Yiming Xie, Jing Gu, Tianye Ding, Jianwei Yang, Huaizu Jiang
Abstract:
Unlocking spatial reasoning in Large Multimodal Models (LMMs) is crucial for enabling intelligent interaction with 3D environments. While prior efforts often rely on explicit 3D inputs or specialized model architectures, we ask: can LMMs reason about 3D space using only structured 2D representations derived from perception?In this work, we introduce Struct2D, a perception-guided prompting framework that combines bird’s-eye-view (BEV) images with object marks and object-centric metadata, optionally incorporating egocentric keyframes when needed. Using Struct2D, we conduct an in-depth zero-shot analysis of closed-source LMMs (e.g., GPT-4o) and find that they exhibit surprisingly strong spatial reasoning abilities when provided with projected 2D inputs, effectively handling tasks such as relative direction estimation and route planning.Motivated by these findings, we construct a large-scale instructional tuning dataset, \textbf{Struct2D-Set}, using an automated pipeline that generates fine-grained QA pairs grounded in 3D indoor scenes. We then fine-tune an open-source LMM (Qwen2.5VL) using Struct2D-Set, relying on noisy 3D perception rather than ground-truth annotations. Despite this, the tuned model achieves strong performance across multiple spatial reasoning benchmarks, including 3D question answering, captioning, and object grounding, spanning eight diverse reasoning categories.Our approach demonstrates that structured 2D inputs can effectively bridge perception and language reasoning in LMMs—without requiring explicit 3D representations as input. We will release both our code and dataset to support future research.



Paperid:3486
Authors:Kaihua Chen, Tarasha Khurana, Deva Ramanan
Abstract:
We explore novel-view synthesis for dynamic scenes from monocular videos. Prior approaches rely on costly test-time optimization of 4D representations or do not preserve scene geometry when trained in a feed-forward manner. Our approach is based on three key insights: (1) covisible pixels (that are visible in both the input and target views) can be rendered by first reconstructing the dynamic 3D scene and rendering the reconstruction from the novel-views and (2) hidden pixels in novel views can be ``inpainted" with feed-forward 2D video diffusion models. Notably, our video inpainting diffusion model (CogNVS) can be self-supervised from 2D videos, allowing us to train it on a large corpus of in-the-wild videos. This in turn allows for (3) CogNVS to be applied zero-shot to novel test videos via test-time finetuning. We empirically verify that CogNVS outperforms almost all prior art for novel-view synthesis of dynamic scenes from monocular videos.



Paperid:3487
Authors:Chenhui Xu, Dancheng Liu, Amir Nassereldine, Jinjun Xiong
Abstract:
Physics‑Informed Neural Networks (PINNs) often exhibit “failure modes” in which the PDE residual loss converges while the solution error stays large, a phenomenon traditionally blamed on local optima separated from the true solution by steep loss barriers.We challenge this understanding by demonstrate that the real culprit is insufficient arithmetic precision: with standard FP32, the L‑BFGS optimizer prematurely satisfies its convergence test, freezing the network in a spurious failure phase.Simply upgrading to FP64 rescues optimization, enabling vanilla PINNs to solve PDEs without any failure modes.These results reframe PINN failure modes as precision‑induced stalls rather than inescapable local minima and expose a three‑stage training dynamic—un‑converged, failure, success—whose boundaries shift with numerical precision.Our findings emphasize that rigorous arithmetic precision is the key to dependable PDE solving with neural networks.Our code is available at Supplementary Material.



Paperid:3488
Authors:Jinyoung Park, Jeehye Na, Jinyoung Kim, Hyunwoo J. Kim
Abstract:
Recent works have demonstrated the effectiveness of reinforcement learning (RL)-based post-training for enhancing the reasoning capabilities of large language models (LLMs). In particular, Group Relative Policy Optimization (GRPO) has shown impressive success using a PPO-style reinforcement algorithm with group-based normalized rewards. However, GRPO has been less explored in Video Large Language Models (VideoLLMs). In this paper, we explore GRPO and identify two problems that deteriorate the effective learning: (1) reliance on safeguards, and (2) vanishing advantage. To mitigate these challenges, we propose DeepVideo-R1, a video large language model trained with Reg-GRPO (Regressive GRPO) and difficulty-aware data augmentation. Reg-GRPO reformulates the GRPO loss function into a regression task that directly predicts the advantage in GRPO, eliminating the need for heuristic safeguards such as the clipping and min functions. It aligns VideoLLMs with advantages, providing effective guidance. The difficulty-aware data augmentation strategy augments input prompts/videos to generate samples at solvable difficulty levels, enabling diverse reward signals. Experimental results show that our approach significantly improves video reasoning performance across multiple benchmarks. Our codes are included in the supplement.



Paperid:3489
Authors:Jiayi Kuang, Haojing Huang, Yinghui Li, Xinnian Liang, Zhikun Xu, Yangning Li, Xiaoyu Tan, Chao Qu, Meishan Zhang, Ying Shen, Philip S Yu
Abstract:
Large Language Models (LLMs) have demonstrated outstanding performance in mathematical reasoning capabilities. However, we argue that current large-scale reasoning models primarily rely on scaling up training datasets with diverse mathematical problems and long thinking chains, which raises questions about whether LLMs genuinely acquire mathematical concepts and reasoning principles or merely remember the training data. In contrast, humans tend to break down complex problems into multiple fundamental atomic capabilities. Inspired by this, we propose a new paradigm for evaluating mathematical atomic capabilities. Our work categorizes atomic abilities into two dimensions: (1) field-specific abilities across four major mathematical fields, algebra, geometry, analysis, and topology, and (2) logical abilities at different levels, including conceptual understanding, forward multi-step reasoning with formal math language, and counterexample-driven backward reasoning. We propose corresponding training and evaluation datasets for each atomic capability unit, and conduct extensive experiments about how different atomic capabilities influence others, to explore the strategies to elicit the required specific atomic capability. Evaluation and experimental results on advanced models show many interesting discoveries and inspirations about the different performances of models on various atomic capabilities and the interactions between atomic capabilities. Our findings highlight the importance of decoupling mathematical intelligence into atomic components, providing new insights into model cognition and guiding the development of training strategies toward a more efficient, transferable, and cognitively grounded paradigm of "atomic thinking".



Paperid:3490
Authors:Zhi Jing, Siyuan Yang, Jicong Ao, Ting Xiao, Yu-Gang Jiang, Chenjia Bai
Abstract:
For robotic manipulation, existing robotics datasets and simulation benchmarks predominantly cater to robot-arm platforms. However, for humanoid robots equipped with dual arms and dexterous hands, simulation tasks and high-quality demonstrations are notably lacking. Bimanual dexterous manipulation is inherently more complex, as it requires coordinated arm movements and hand operations, making autonomous data collection challenging. This paper presents HumanoidGen, an automated task creation and demonstration collection framework that leverages atomic dexterous operations and LLM reasoning to generate relational constraints. Specifically, we provide spatial annotations for both assets and dexterous hands based on the atomic operations, and perform an LLM planner to generate a chain of actionable spatial constraints for arm movements based on object affordances and scenes. To further improve planning ability, we employ a variant of Monte Carlo tree search to enhance LLM reasoning for long-horizon tasks and insufficient annotation. In experiments, we create a novel benchmark with augmented scenarios to evaluate the quality of the collected data. The results show that the performance of the 2D and 3D diffusion policies can scale with the generated dataset.



Paperid:3491
Authors:Pei Peng, Ming-Kun Xie, Hang Hao, Tong Jin, Sheng-Jun Huang
Abstract:
Object–context shortcuts remain a persistent challenge in vision‑language models, undermining zero‑shot reliability when test-time scenes diverge from familiar training co-occurrences. We recast this issue as a causal inference problem and ask: Would the prediction remain if the object appeared in a different environment? To answer it at inference time, we estimate object and background expectations within CLIP’s representation space, and synthesize counterfactual embeddings by recombining object features with diverse alternative contexts sampled from external datasets, batch neighbors, or text-derived descriptions. By estimating the Total Direct Effect and simulating intervention, we further subtract background‑only activation, preserving beneficial object–context interactions while mitigating hallucinated scores. Without retraining or prompt design, our method substantially improves both worst-group and average accuracy on context-sensitive benchmarks, establishing a new zero‑shot state of the art. Beyond performance, our framework provides a lightweight representation-level counterfactual approach, offering a practical causal avenue for debiased and reliable multimodal reasoning.



Paperid:3492
Authors:Xiaochen Zhao, Chengting Yu, Kairong Yu, Lei Liu, Aili Wang
Abstract:
Spiking Neural Networks (SNNs) exhibit exceptional energy efficiency on neuromorphic hardware due to their sparse activation patterns. However, conventional training methods based on surrogate gradients and Backpropagation Through Time (BPTT) not only lag behind Artificial Neural Networks (ANNs) in performance, but also incur significant computational and memory overheads that grow linearly with the temporal dimension. To enable high-performance SNN training under limited computational resources, we propose a enhanced self-distillation framework, jointly optimized with rate-based backpropagation. Specifically, the firing rates of intermediate SNN layers are projected onto lightweight ANN branches, and high-quality knowledge generated by the model itself is used to optimize substructures through the ANN pathways. Unlike traditional self-distillation paradigms, we observe that low-quality self-generated knowledge may hinder convergence. To address this, we decouple the teacher signal into reliable and unreliable components, ensuring that only reliable knowledge is used to guide the optimization of the model. Extensive experiments on CIFAR-10, CIFAR-100, CIFAR10-DVS, and ImageNet demonstrate that our method reduces training complexity while achieving high-performance SNN training. For instance, on CIFAR-100, it reduces memory consumption by 75.80% and training time by 23.30% compared to BPTT, while improving accuracy by 1.85%. Notably, the proposed self-distillation framework also shows strong adaptability when applied to ANNs.



Paperid:3493
Authors:Antoine Moulin, Gergely Neu, Luca Viano
Abstract:
Abstract:We study the problem of offline imitation learning in Markov decision processes (MDPs), where the goal is to learn a well-performing policy given a dataset of state-action pairs generated by an expert policy. Complementing a recent line of work on this topic that assumes that the expert policy belongs to a tractable class of known policies, we approach this problem from a new angle and leverage another type of structural assumption about the environment. Specifically, for the class of linear $Q^\pi$-realizable MDPs, we introduce a new algorithm called primal-dual offline imitation learning (PDOIL), which is guaranteed to match the performance of any expert up to an additive error $\varepsilon$ with access to $\mathcal{O}(\varepsilon^{-2})$ samples. Moreover, we extend this result to possibly non-linear $Q^\pi$-realizable MDPs at the cost of a worst sample complexity of order $\mathcal{O}(\varepsilon^{-4})$. Finally, our analysis suggests a new loss function for training critic networks from expert data in deep imitation learning. Empirical evaluations on standard benchmarks demonstrate that the neural net implementation of PDOIL is superior to behavior cloning and competitive with state-of-the-art algorithms.



Paperid:3494
Authors:Fabian Immel, Jan-Hendrik Pauls, Richard Schwarzkopf, Frank Bieder, Jonas Merkert, Christoph Stiller
Abstract:
Autonomous vehicles rely on detailed and accurate environmental information to operate safely.High definition (HD) maps offer a promising solution, but their high maintenance cost poses a significant barrier to scalable deployment. This challenge is addressed by online HD map construction methods, which generate local HD maps from live sensor data.However, these methods are inherently limited by the short perception range of onboard sensors. To overcome this limitation and improve general performance, recent approaches have explored the use of standard definition (SD) maps as prior, which are significantly easier to maintain.We propose SDTagNet, the first online HD map construction method that fully utilizes the information of widely available SD maps, like OpenStreetMap, to enhance far range detection accuracy. Our approach introduces two key innovations.First, in contrast to previous work, we incorporate not only polyline SD map data with manually selected classes, but additional semantic information in the form of textual annotations.In this way, we enrich SD vector map tokens with NLP-derived features, eliminating the dependency on predefined specifications or exhaustive class taxonomies.Second, we introduce a point-level SD map encoder together with orthogonal element identifiers to uniformly integrate all types of map elements.Experiments on Argoverse 2 and nuScenes show that this boosts map perception performance by up to +5.9 mAP (+45%) w.r.t. map construction without priors and up to +3.2 mAP (+20%) w.r.t. previous approaches that already use SD map priors.



Paperid:3495
Authors:Xinyi Wang, Xun Yang, Yanlong Xu, Yuchen Wu, Zhen Li, Na Zhao
Abstract:
Effective human–agent collaboration in physical environments requires understanding not only what to act upon, but also where the actionable elements are and how to interact with them. Existing approaches often operate at the object level or disjointedly handle fine-grained affordance reasoning, lacking coherent, instruction-driven grounding and reasoning. In this work, we introduce a new task: Fine-grained 3D Embodied Reasoning, which requires an agent to predict, for each referenced affordance element in a 3D scene, a structured triplet comprising its spatial location, motion type, and motion axis, based on a task instruction. To solve this task, we propose AffordBot, a novel framework that integrates Multimodal Large Language Models (MLLMs) with a tailored chain-of-thought (CoT) reasoning paradigm. To bridge the gap between 3D input and 2D-compatible MLLMs, we render surround-view images of the scene and project 3D element candidates into these views, forming a rich visual representation aligned with the scene geometry. Our CoT pipeline begins with an active perception stage, prompting the MLLM to select the most informative viewpoint based on the instruction, before proceeding with step-by-step reasoning to localize affordance elements and infer plausible interaction motions. Evaluated on the SceneFun3D dataset, AffordBot achieves state-of-the-art performance, demonstrating strong generalization and physically grounded reasoning with only 3D point cloud input and MLLMs.



Paperid:3496
Authors:Yixiao Wang, Zishan Shao, Ting Jiang, Aditya Devarakonda
Abstract:
Abstract:We present a novel enhanced cyclic coordinate descent (ECCD) framework for solving generalized linear models with elastic net constraints that reduces training time in comparison to existing state-of-the-art methods.We redesign the CD method by performing a Taylor expansion around the current iterate to avoid nonlinear operations arising in the gradient computation.By introducing this approximation we are able to unroll the vector recurrences occurring in the CD method and reformulate the resulting computations into more efficient batched computations.We show empirically that the recurrence can be unrolled by a tunable integer parameter, $s$, such that $s > 1$ yields performance improvements without affecting convergence, whereas $s = 1$ yields the original CD method.A key advantage of ECCD is that it avoids the convergence delay and numerical instability exhibited by block coordinate descent.Finally, we implement our proposed method in C++ using Eigen to accelerate linear algebra computations.Comparison of our method against existing state-of-the-art solvers show performance improvements up $13\times$ for regularization path variant on benchmark datasets obtained from the LIBSVM repository.



Paperid:3497
Authors:Bin Lei, Weitai Kang, Zijian Zhang, Winson Chen, Xi Xie, Shan Zuo, Mimi Xie, Ali Payani, Mingyi Hong, Yan Yan, Caiwen Ding
Abstract:
Abstract:This paper introduces \textsc{InfantAgent-Next}, a generalist agent capable of interacting with computers in a multimodal manner, encompassing text, images, audio, and video.Unlike existing approaches that either build intricate workflows around a single large model or only provide workflow modularity, our agent integrates tool-based and pure vision agents within a highly modular architecture, enabling different models to collaboratively solve decoupled tasks in a step-by-step manner. Our generality is demonstrated by our ability to evaluate not only pure vision-based real-world benchmarks (i.e., OSWorld), but also more general or tool-intensive benchmarks (e.g., GAIA and SWE-Bench).Specifically,weachieve $\mathbf{7.27\\%}$ accuracy on OSWorld, higher than Claude-Computer-Use. Codes and evaluation scripts are included in the supplementary material and will be released as open-source.



Paperid:3498
Authors:Yiren Lu, Yunlai Zhou, Yiran Qiao, Chaoda Song, Tuo Liang, Jing Ma, Huan Wang, Yu Yin
Abstract:
Open-vocabulary querying in 3D space is crucial for enabling more intelligent perception in applications such as robotics, autonomous systems, and augmented reality. However, most existing methods rely on 2D pixel-level parsing, leading to multi-view inconsistencies and poor 3D object retrieval. Moreover, they are limited to static scenes and struggle with dynamic scenes due to the complexities of motion modeling. In this paper, we propose Segment then Splat, a 3D-aware open vocabulary segmentation approach for both static and dynamic scenes based on Gaussian Splatting. Segment then Splat reverses the long established approach of ``segmentation after reconstruction" by dividing Gaussians into distinct object sets before reconstruction. Once the reconstruction is complete, the scene is naturally segmented into individual objects, achieving true 3D segmentation. This approach not only eliminates Gaussian-object misalignment issues in dynamic scenes but also accelerates the optimization process, as it eliminates the need for learning a separate language field. After optimization, a CLIP embedding is assigned to each object to enable open-vocabulary querying. Extensive experiments demonstrate the effectiveness of our proposed method in both static and dynamic scenarios.



Paperid:3499
Authors:Ziqi Pang, Yu-Xiong Wang
Abstract:
The fundamental challenge of long video understanding, e.g., question answering, lies in the extensive number of frames, making it infeasible to densely understand the local details while comprehensively digest the global contexts, especially within a limited context length. To address this problem, our insight is to process short video segments individually and combine these segment-level analyses into a final response. This intuition is noted in the well-established MapReduce principle in big data processing and is naturally compatible with inference scaling at the system level. Motivated by this, we propose \name (pronounced as "mister video"), a long video understanding framework adopting the MapReduce principle. We define the standard operations of MapReduce in a long video understanding context: the Map steps conduct independent and sequence-parallel dense perception on short video segments, covering local details, while the Reduce steps comprehensively aggregate the segment-level results into an answer with global contexts. Thanks to the low cost and convenience of building video agents, we instantiate such Map and Reduce operations as an effective video agent capable of attending to local details and global contexts. Based on such abilities, we further introduce two critical yet previously under-explored long video understanding designs: (a) consistent character/object names in the captions, benefiting the reasoning of actions and stories across long horizons; (b) question intention analysis, which changes the key-frame retrieval in previous video agents to localizing the relevant information via jointly reasoning the whole video contexts and questions. Our \name achieves a >7% accuracy improvement on the challenging LVBench over state-of-the-art video agents and vision-language models (VLMs) and demonstrates a clear advantage on multiple long video benchmarks, highlighting the potential of the MapReduce principle.



Paperid:3500
Authors:Seungyong Moon, Bumsoo Park, Hyun Oh Song
Abstract:
While language models have shown remarkable performance across diverse tasks, they still encounter challenges in complex reasoning scenarios. Recent research suggests that language models trained on linearized search traces toward solutions, rather than solely on the final solutions, exhibit improved generalization, despite the search traces being potentially noisy or suboptimal. However, relying on such imperfect traces can result in inefficient use of test-time compute. To address this, we propose guided reinforced self-training (Guided-ReST), a fine-tuning algorithm designed to improve the model’s capability for effective search during inference. The key insight behind Guided-ReST is that optimal solutions can serve as valuable step-by-step landmarks to guide the model’s search process. Based on this insight, we introduce a novel data generation method that seamlessly incorporates optimal solutions into the model’s search procedure, enabling the generation of high-quality search traces. By fine-tuning the model on these search traces, we effectively distill improved search strategies into the model. Our method significantly improves the search capabilities of language models on arithmetic and mathematical reasoning tasks, including Countdown, MATH-500, and AMC23.



Paperid:3501
Authors:zhangkai wu, Xuhui Fan, Hongyu Wu, Longbing Cao
Abstract:
Pre-trained diffusion models are commonly used to generate clean data (e.g., images) from random noises, effectively forming pairs of noises and corresponding clean images. Distillation on these pre-trained models can be viewed as the process of constructing advanced trajectories within the pair to accelerate sampling. For instance, consistency model distillation develops consistent projection functions to regulate trajectories, although sampling efficiency remains a concern. Rectified flow method enforces straight trajectories to enable faster sampling, yet relies on numerical ODE solvers, which may introduce approximation errors. In this work, we bridge the gap between the consistency model and the rectified flow method by proposing a Straight Consistent Trajectory~(SCoT) model. SCoT enjoys the benefits of both approaches for fast sampling, producing trajectories with consistent and straight properties simultaneously. These dual properties are strategically balanced by targeting two critical objectives: (1) regulating the gradient of SCoT's mapping to a constant, (2) ensuring trajectory consistency. Extensive experimental results demonstrate the effectiveness and efficiency of SCoT. The source code is available at an anonymous repository: \url{https://anonymous.4open.science/r/SCoT-04F9/README.md}.



Paperid:3502
Authors:Eric Zhao, Pranjal Awasthi, Nika Haghtalab
Abstract:
Finetuning provides a scalable and cost-effective means of customizing language models for specific tasks or response styles, with greater reliability than prompting or in-context learning. In contrast, the conventional wisdom is that injecting knowledge via finetuning results in brittle performance and poor generalization. We argue that the dichotomy of "task customization" (e.g., instruction tuning) and "knowledge injection" (e.g., teaching new facts) is a distinction without a difference. We instead identify concrete factors that explain the heterogeneous effectiveness observed with finetuning. To this end, we conduct a large-scale experimental study of finetuning the frontier Gemini v1.5 model family on a spectrum of datasets that are artificially engineered to interpolate between the strengths and failure modes of finetuning. Our findings indicate that question-answer training data formats provide much stronger knowledge generalization than document/article-style training data, numerical information can be harder for finetuning to retain than categorical information, and models struggle to apply finetuned knowledge during multi-step reasoning even when trained on similar examples---all factors that render ``knowledge injection'' to be especially difficult, even after controlling for considerations like data augmentation and information volume. On the other hand, our findings also indicate that it is not fundamentally more difficult to finetune information about a real-world event than information about writing style.



Paperid:3503
Authors:Alireza Aghasi, Nicholas F. Marshall, Saeid Pourmand, Wyatt Whiting
Abstract:
We propose a novel randomized algorithm for constructing binary neural networks with tunable accuracy. This approach is motivated by hyperdimensional computing (HDC)---a brain-inspired paradigm that leverages high-dimensional vector representations, offering efficient hardware implementation and robustness to model corruptions. Unlike traditional low-precision methods that use quantization, we consider binary embeddings of data as points in the hypercube equipped with the Hamming distance. We propose a novel family of floating-point neural networks, G-Nets, which are general enough to mimic standard network layers. Each floating-point G-Net has a randomized binary embedding, an embedded hyperdimensional (EHD) G-Net, that retains the accuracy of its floating-point counterparts, with theoretical guarantees, due to the concentration of measure. Empirically, our binary models match convolutional neural network accuracies and outperform prior HDC models by large margins, for example, we achieve almost 30\% higher accuracy on CIFAR-10 compared to prior HDC models. G-Nets are a theoretically justified bridge between neural networks and randomized binary neural networks, opening a new direction for constructing robust binary/quantized deep learning models. Our implementation is available at \url{https://github.com/GNet2025/GNet}.



Paperid:3504
Authors:Hafez Ghaemi, Eilif B. Muller, Shahab Bakhtiari
Abstract:
Current self-supervised algorithms commonly rely on transformations such as data augmentation and masking to learn visual representations. This is achieved by enforcing invariance or equivariance with respect to these transformations after encoding two views of an image. This dominant two-view paradigm often limits the flexibility of learned representations for downstream adaptation by creating performance trade-offs between high-level invariance-demanding tasks such as image classification and more fine-grained equivariance-related tasks. In this work, we proposes \emph{seq-JEPA}, a world modeling framework that introduces architectural inductive biases into joint-embedding predictive architectures to resolve this trade-off. Without relying on dual equivariance predictors or loss terms, seq-JEPA simultaneously learns two architecturally segregated representations: one equivariant to specified transformations and another invariant to them. To do so, our model processes short sequences of different views (observations) of inputs. Each encoded view is concatenated with an embedding of the relative transformation (action) that produces the next observation in the sequence. These view-action pairs are passed through a transformer encoder that outputs an aggregate representation. A predictor head then conditions this aggregate representation on the upcoming action to predict the representation of the next observation. Empirically, seq-JEPA demonstrates strong performance on both equivariant and invariant benchmarks without sacrificing one for the other. Furthermore, it excels at tasks that inherently require aggregating a sequence of observations, such as path integration across actions and predictive learning across eye movements.



Paperid:3505
Authors:Chunyu Wei, Wenji Hu, Xingjia Hao, Xin Wang, Yifan Yang, Yunhai Wang, Yang Tian, Yueguo Chen
Abstract:
Large Language Models (LLMs) face significant limitations when applied to large-scale graphs, struggling with context constraints and inflexible reasoning. We introduce GraphChain, a novel framework enabling LLMs to analyze large graphs by orchestrating dynamic sequences of specialized tools, mimicking human exploratory processes. GraphChain incorporates two core technical contributions: (1) Progressive Graph Distillation, a reinforcement learning approach that learns to generate tool sequences balancing task relevance and intermediate state compression, thereby overcoming LLM context limitations. (2) Structure-aware Test-Time Adaptation (STTA), a mechanism using a lightweight, self-supervised adapter conditioned on graph spectral properties to efficiently adapt a frozen LLM policy to diverse graph structures via soft prompts without retraining. Experiments show GraphChain significantly outperforms prior methods, enabling scalable and adaptive LLM-driven graph analysis.



Paperid:3506
Authors:Giuseppe Castiglione, Christopher L Buckley, Ivor Simpson
Abstract:
Deep neural networks trained with gradient descent exhibit varying rates of learning for different patterns. However, the complexity of fitting models to data makes direct elucidation of the dynamics of learned patterns challenging. To circumvent this, many works have opted to characterize phases of learning through summary statistics known as order parameters. In this work, we propose a unifying framework for constructing order parameters based on the Neural Tangent Kernel (NTK), in which the relationship with the data set is more transparent. In particular, we derive a local approximation of the NTK for a class of deep regression models (SIRENs) trained to reconstruct natural images. In so doing, we analytically connect three seemingly distinct phase transitions: the emergence of wave patterns in residuals (a novel observation), loss rate collapse, and NTK alignment. Our results provide a dynamical perspective on the observed biases of SIRENs, and deep image regression models more generally.



Paperid:3507
Authors:Guo Chen, Zhiqi Li, Shihao Wang, Jindong Jiang, Yicheng Liu, Lidong Lu, De-An Huang, Wonmin Byeon, Matthieu Le, Max Ehrlich, Tong Lu, Limin Wang, Bryan Catanzaro, Jan Kautz, Andrew Tao, Zhiding Yu, Guilin Liu
Abstract:
We introduce Eagle2.5, a frontier vision-language model (VLM) for long-context multimodal learning. Our work addresses the challenges in long video comprehension and high-resolution image understanding, introducing a generalist framework for both tasks. The proposed training framework incorporates Automatic Degrade Sampling and Image Area Preservation, two techniques that preserve contextual integrity and visual details. The framework also includes numerous efficiency optimizations in the pipeline for long-context data training. Finally, we propose Eagle-Video-110K, a novel dataset that integrates both story-level and clip-level annotations, facilitating long-video understanding. Eagle2.5 demonstrates substantial improvements on long-context multimodal benchmarks, providing a robust solution to the limitations of existing VLMs. Notably, our best model Eagle2.5-8B achieves 72.4\% on Video-MME with 512 input frames, matching the results of top-tier commercial model such as GPT-4o and large-scale open-source models like Qwen2.5-VL-72B and InternVL2.5-78B.



Paperid:3508
Authors:Brian Zheng, Alisa Liu, Orevaoghene Ahia, Jonathan Hayase, Yejin Choi, Noah Smith
Abstract:
Modern tokenizers employ deterministic algorithms to map text into a single ``canonical" token sequence, yet the same string can be encoded as many non-canonical tokenizations using the language model vocabulary, including tokenizing by character. In this paper, we investigate the robustness of LMs to input encoded with non-canonical tokenizations entirely unseen during training. Surprisingly, when evaluated across 20 benchmarks, we find that instruction-tuned models retain up to 93.4\% of their original performance when given a randomly sampled tokenization, and 90.8\% with character-level tokenization. We find that overall stronger models tend to be more robust, and that robustness diminishes as the tokenization departs farther from the canonical form. Motivated by these results, we identify settings where non-canonical tokenization schemes can \textit{improve} performance, finding that character‑level segmentation improves string manipulation and code understanding tasks by up to 15\%, and right‑aligned digit grouping enhances large‑number arithmetic by over 33\%. Finally, we investigate the source of this robustness, finding that it arises in the instruction-tuning phase. We provide evidence that both base and post-trained models grasp the semantics of non-canonical tokenizations (perceiving them as containing misspellings). However, base models try to mimic the imagined mistakes and degenerate into nonsensical output, while post-trained models are committed to fluent responses. Overall, our findings suggest that models are less committed to their tokenizer than previously believed, and highlight the promise of intervening on tokenization at inference time to boost language model performance.



Paperid:3509
Authors:Alan Arazi, Eilam Shapira, Roi Reichart
Abstract:
While deep learning has achieved remarkable success across many domains, it has historically underperformed on tabular learning tasks, which remain dominated by gradient boosting decision trees (GBDTs). However, recent advancements are paving the way for Tabular Foundation Models, which can leverage real-world knowledge and generalize across diverse datasets, particularly when the data contains free-text. Although incorporating language model capabilities into tabular tasks has been explored, most existing methods utilize static, target-agnostic textual representations, limiting their effectiveness. We introduce TabSTAR: a Foundation Tabular Model with Semantically Target-Aware Representations. TabSTAR is designed to enable transfer learning on tabular data with textual features, with an architecture free of dataset-specific parameters. It unfreezes a pretrained text encoder and takes as input target tokens, which provide the model with the context needed to learn task-specific embeddings. TabSTAR achieves state-of-the-art performance for both medium- and large-sized datasets across known benchmarks of classification tasks with text features, and its pretraining phase exhibits scaling laws in the number of datasets, offering a pathway for further performance improvements.



Paperid:3510
Authors:Mumin Jia, Jairo Diaz-Rodriguez
Abstract:
Human cognition is punctuated by abrupt, spontaneous shifts between topics—driven by emotional, contextual, or associative cues—a phenomenon known as spontaneous thought in neuroscience. In contrast, self-attention-based models rely on structured patterns over their inputs to predict each next token, lacking spontaneity. Motivated by this distinction, we characterizespontaneous topic changesin self-attention architectures and reveal divergences fromspontaneous human thought. First, we establish theoretical results under a simplified, single-layer self-attention model with suitable conditions by defining a topic as a set of Token Priority Graphs (TPGs). Specifically, we demonstrate that (1) the model maintains the priority order of tokens related to the input topic, (2) a spontaneous topic change can occur only if lower-priority tokens outnumber all higher-priority tokens of the input topic, and (3) unlike human cognition, the longer context length or the more ambiguous input topic does not increase the likelihood of spontaneous change. Second, we empirically validate that the effect of input length or topic ambiguity persists in modern, state-of-the-art LLMs, underscoring a fundamental disparity between human cognition and AI behavior in the context of spontaneous topic changes. To the best of our knowledge, no prior work has explored these questions with a focus so closely aligned to human thought.



Paperid:3511
Authors:Liyan Tang, Grace Kim, Xinyu Zhao, Thom Lake, Wenxuan Ding, Fangcong Yin, Prasann Singhal, Manya Wadhwa, Zeyu Liu, Zayne Sprague, Ramya Namuduri, Bodun Hu, Juan Rodriguez, Puyuan Peng, Greg Durrett
Abstract:
Chart understanding presents a unique challenge for large vision-language models (LVLMs), as it requires the integration of sophisticated textual and visual reasoning capabilities. However, current LVLMs exhibit a notable imbalance between these skills, falling short on visual reasoning that is difficult to perform in text. We conduct a case study using a synthetic dataset solvable only through visual reasoning and show that model performance degrades significantly with increasing visual complexity, while human performance remains robust. We then introduceChartMuseum, a new Chart Question Answering (QA) benchmark containing 1,162 expert-annotated questions spanning multiple reasoning types, curated from real-world charts across 184 sources, specifically built to evaluate complex visual and textual reasoning. Unlike prior chart understanding benchmarks---where frontier models perform similarly and near saturation---our benchmark exposes a substantial gap between model and human performance, while effectively differentiating model capabilities: although humans achieve 93% accuracy, the best-performing model Gemini-2.5-Pro attains only 63.0%, and the leading open-source LVLM Qwen2.5-VL-72B-Instruct achieves only 38.5%. Moreover, on questions requiring primarily visual reasoning,allmodels experience a 35%-55% performance drop from text-reasoning-heavy question performance. Lastly, our qualitative error analysis reveals specific categories of visual reasoning that are challenging for current LVLMs. Both ChartMuseum and the evaluation code are available athttps://github.com/Liyan06/ChartMuseum.



Paperid:3512
Authors:Jiaming Ma, Binwu Wang, Guanjun Wang, Kuo Yang, Zhengyang Zhou, Pengkun Wang, Xu Wang, Yang Wang
Abstract:
The effectiveness of Spatiotemporal Graph Convolutional Networks (STGCNs) critically hinges on the quality of the underlying graph topology. While end-to-end adaptive graph learning methods have demonstrated promising results in capturing latent spatiotemporal dependencies, they often suffer from high computational complexity and limited expressive capacity. In this paper, we propose MAGE for efficient spatiotemporal forecasting. We first conduct a theoretical analysis demonstrating that the ReLU activation function employed in existing methods amplifies edge-level noise during graph topology learning, thereby compromising the fidelity of the learned graph structures. To enhance model expressiveness, we introduce a sparse yet balanced mixture-of-experts strategy, where each expert perceives the unique underlying graph through kernel-based functions and operates with linear complexity relative to the number of nodes. The sparsity mechanism ensures that each node interacts exclusively with compatible experts, while the balancing mechanism promotes uniform activation across all experts, enabling diverse and adaptive graph representations. Furthermore, we theoretically establish that a single graph convolution using the learned graph in MAGE is mathematically equivalent to multiple convolutional steps under conventional graphs. We evaluate MAGE against 14 state-of-the-art baselines on 17 real-world spatiotemporal datasets. MAGE achieves SOTA performance on 94% (48/51) of the evaluation metrics. Notably, on the SD dataset, MAGE achieves an impressive 5.15% performance improvement, while also improving memory efficiency by 10X and training efficiency by 20X.



Paperid:3513
Authors:Qijun Luo, Mengqi Li, Lei Zhao, Xiao Li
Abstract:
Abstract:Training language models on long sequence data is a demanding requirement for enhancing the model's capability on complex tasks, e.g., long-chain reasoning. However, as the sequence length scales up, the memory cost for storing activation values becomes huge during the Backpropagation (BP) process, even with the application of gradient checkpointing technique. To tackle this challenge, we propose a *memory-efficient* and *exact* BP method called **StreamBP**, which performs a linear decomposition of the chain rule along the sequence dimension in a layer-wise manner, significantly reducing the memory cost of activation values and logits. The proposed method is applicable to common objectives such as SFT, GRPO, and DPO. Moreover, by leveraging the causal structure of LLMs, StreamBP reduces the computational FLOPs and allows faster BP speed compared to the gradient checkpointing baseline. StreamBP scales up the maximum sequence length of BP by $2.8-5.5 \times$ larger, and reduces BP time cost by approximately $30$% compared to gradient checkpointing. We further develop a communication-efficient distributed StreamBP to effectively support multi-GPU training and broaden its applicability.



Paperid:3514
Authors:Juntong Wang, Xiyuan Wang, Muhan Zhang
Abstract:
Common Neighbors (CNs) and their higher-order variants are important pairwise features widely used in state-of-the-art link prediction methods. However, existing methods often struggle with the repetition across different orders of CNs and fail to fully leverage their potential. We identify that these limitations stem from two key issues: redundancy and over-smoothing in high-order common neighbors. To address these challenges, we design orthogonalization to eliminate redundancy between different-order CNs and normalization to mitigate over-smoothing. By combining these two techniques, we propose Orthogonal Common Neighbor (OCN), a novel approach that significantly outperforms the strongest baselines by an average of 7.7\% on popular link prediction benchmarks. A thorough theoretical analysis is provided to support our method. Ablation studies also verify the effectiveness of our orthogonalization and normalization techniques.



Paperid:3515
Authors:Kenta Niwa, Yuki Takezawa, Guoqiang Zhang, W. Kleijn
Abstract:
Abstract:For communication-efficient decentralized learning, it is essential to employ dynamic graphs designed to improve the expected spectral gap by reducing deviations from global averaging. The $1$-peer exponential graph demonstrates its finite-time convergence property--achieved by maximizing the expected spectral gap--but only when the number of nodes $n$ is a power of two. However, its efficiency across any $n$ and the commutativity of mixing matrices remain unexplored. We delve into the principles underlying the $1$-peer exponential graph to explain its efficiency across any $n$ and leverage them to develop new dynamic graphs. We propose two new dynamic graphs: the $k$-peer exponential graph and the null-cascade graph. Notably, the null-cascade graph achieves finite-time convergence for any $n$ while ensuring commutativity. Our experiments confirm the effectiveness of these new graphs, particularly the null-cascade graph, in most test settings.



Paperid:3516
Authors:Sikai Bai, Jie ZHANG, Haoxi Li, Zicong Hong, Song Guo
Abstract:
Abstract:Despite the significant breakthrough of Mixture-of-Experts (MoE), the increasing scale of these MoE models presents huge memory and storage challenges. Existing MoE pruning methods, which involve reducing parameter size with a uniform sparsity across all layers, often lead to suboptimal outcomes and performance degradation due to varying expert redundancy in different MoE layers. To address this, we propose a non-uniform pruning strategy, dubbed Differentiable Expert Pruning (DiEP), which adaptively adjusts pruning rates at the layer level while jointly learning inter-layer importance, effectively capturing the varying redundancy across different MoE layers. By transforming the global discrete search space into a continuous one, our method handles exponentially growing non-uniform expert combinations, enabling adaptive gradient-based pruning. Extensive experiments on five advanced MoE models demonstrate the efficacy of our method across various NLP tasks. Notably, \textbf{DiEP} retains around 92\% of original performance on Mixtral 8$\times$7B with only half the experts, outperforming other pruning methods by up to 7.1% on the challenging MMLU dataset.



Paperid:3517
Authors:Chenglong Wang, Yang Gan, Hang Zhou, Chi Hu, Yongyu Mu, Kai Song, MuRun Yang, Bei Li, Tong Xiao, Chunliang Zhang, Tongran Liu, JingBo Zhu, Zhengtao Yu
Abstract:
Recent advances in diffusion language models (DLMs) have presented a promising alternative to traditional autoregressive large language models (LLMs). However, DLMs still lag behind LLMs in reasoning performance, especially as the number of denoising steps decreases. Our analysis reveals that this shortcoming arises primarily from the independent generation of masked tokens across denoising steps, which fails to capture the token correlation. In this paper, we define two types of token correlation: intra-sequence correlation and inter-sequence correlation, and demonstrate that enhancing these correlations improves reasoning performance. To this end, we propose a Multi-Reward Optimization (MRO) approach, which encourages DLMs to consider the token correlation during the denoising process. More specifically, our MRO approach leverages test-time scaling, reject sampling, and reinforcement learning to directly optimize the token correlation with multiple elaborate rewards. Additionally, we introduce group step and importance sampling strategies to mitigate reward variance and enhance sampling efficiency. Through extensive experiments, we demonstrate that MRO not only improves reasoning performance but also achieves significant sampling speedups while maintaining high performance on reasoning benchmarks.



Paperid:3518
Authors:Hsiang Hsu, Pradeep Niroula, Zichang He, Ivan Brugere, Freddy Lecue, Richard Chen
Abstract:
Machine unlearning offers a practical alternative to avoid full model re-training by approximately removing the influence of specific user data. While existing methods certify unlearning via statistical indistinguishability from re-trained models, these guarantees do not naturally extend to model outputs when inputs are adversarially perturbed. In particular, slight perturbations of forget samples may still be correctly recognized by the unlearned model---even when a re-trained model fails to do so---revealing a novel privacy risk: information about the forget samples may persist in their local neighborhood. In this work, we formalize this vulnerability as residual knowledge and show that it is inevitable in high-dimensional settings. To mitigate this risk, we propose a fine-tuning strategy, named RURK, that penalizes the model’s ability to re-recognize perturbed forget samples. Experiments on vision benchmarks with deep neural networks demonstrate that residual knowledge is prevalent across existing unlearning methods and that our approach effectively prevents residual knowledge.



Paperid:3519
Authors:Utkarsh Chavan, Prashant Trivedi, Nandyala Hemachandra
Abstract:
Abstract:Multi-agent systems (MAS) are central to applications such as swarm robotics and traffic routing, where agents must coordinate in a decentralized manner to achieve a common objective. Stochastic Shortest Path (SSP) problems provide a natural framework for modeling decentralized control in such settings. While the problem of learning in SSP has been extensively studied in single-agent settings, the decentralized multi-agent variant remains largely unexplored. In this work, we take a step towards addressing that gap. We study decentralized multi-agent SSPs (Dec-MASSPs) under linear function approximation, where the transition dynamics and costs are represented using linear models. Applying novel symmetry-based arguments, we identify the structure of optimal policies. Our main contribution is the first regret lower bound for this setting based on the construction of hard-to-learn instances for any number of agents, $n$. Our regret lower bound of $\Omega(\sqrt{K})$, over $K$ episodes, highlights the inherent learning difficulty in Dec-MASSPs. These insights clarify the learning complexity of decentralized control and can further guide the design of efficient learning algorithms in multi-agent systems.



Paperid:3520
Authors:Divya Jyoti Bajpai, Manjesh Kumar Hanawal
Abstract:
Abstract:Early-Exit Deep Neural Networks enable adaptive inference by allowing prediction at intermediary layers, significantly reducing computational costs and latency. Most of the early exit strategies greedily exit a sample at an intermediary layer if the confidence in class prediction exceeds a predefined threshold that is set using a static validation set. This is problematic as the model might be overconfident in a wrong class. Also, they are not robust to distribution shifts encountered in deployment, which can undermine model trustworthiness and accuracy. To address these challenges, we propose UAT that adapts the threshold for exit decisions using a Multi-Armed Bandit framework, enabling online, unsupervised adjustment of exit decisions. UAT makes decisions based on a new reward function that assesses predictive certainty and its reliability to balance computational efficiency and prediction quality while penalizing unnecessary late exits. We provide guarantees on risk achieved by UAT and validate its performance on diverse tasks spanning vision-language understanding, text generation, and classification. Our framework demonstrates consistent improvements in speedup $(1.70-2.10\times)$ with a minimal performance drop $(<2)$\% as compared to full model performance.



Paperid:3521
Authors:Haonan Duan, Stephen Lu, Caitlin F Harrigan, Nishkrit Desai, Jiarui Lu, Michał Koziarski, Leonardo Cotta, Chris Maddison
Abstract:
Designing experiments and result interpretations are core scientific competencies, particularly in biology, where researchers perturb complex systems to uncover the underlying systems. Recent efforts to evaluate the scientific capabilities of large language models (LLMs) fail to test these competencies because wet-lab experimentation is prohibitively expensive: in expertise, time and equipment. We introduce SciGym, a first-in-class benchmark that assesses LLMs' iterative experiment design and analysis abilities in open-ended scientific discovery tasks. SciGym overcomes the challenge of wet-lab costs by running a dry lab of biological systems. These models, encoded in Systems Biology Markup Language, are efficient for generating simulated data, making them ideal testbeds for experimentation on realistically complex systems. We evaluated six frontier LLMs on 137 small systems, and released a total of 350 systems at https://huggingface.co/datasets/h4duan/scigym-sbml. Our evaluation shows that while more capable models demonstrated superior performance, all models' performance declined significantly as system complexity increased, suggesting substantial room for improvement in the scientific capabilities of LLM agents.



Paperid:3522
Authors:Julius Vetter, Manuel Gloeckler, Daniel Gedon, Jakob H Macke
Abstract:
Simulation-based inference (SBI) offers a flexible and general approach to performing Bayesian inference: In SBI, a neural network is trained on synthetic data simulated from a model and used to rapidly infer posterior distributions for observed data. A key goal for SBI is to achieve accurate inference with as few simulations as possible, especially for expensive simulators. In this work, we address this challenge by repurposing recent probabilistic foundation models for tabular data: We show how tabular foundation models---specifically TabPFN---can be used as pre-trained autoregressive conditional density estimators for SBI. We propose Neural Posterior Estimation with Prior-data Fitted Networks (NPE-PF) and show that it is competitive with current SBI approaches in terms of accuracy for both benchmark tasks and two complex scientific inverse problems. Crucially, it often substantially outperforms them in terms of simulation efficiency, sometimes requiring orders of magnitude fewer simulations. NPE-PF eliminates the need for selecting and training an inference network and tuning its hyperparameters. We also show that it exhibits superior robustness to model misspecification and can be scaled to simulation budgets that exceed the context size limit of TabPFN.NPE-PF provides a new direction for SBI, where training-free, general-purpose inference models offer efficient, easy-to-use, and flexible solutions for a wide range of stochastic inverse problems.



Paperid:3523
Authors:Davide Maran, Csaba Szepesvari
Abstract:
Abstract:We study the problem of controlling worst-case errors in misspecified linear regression under the random design setting, where the regression function is estimated via (penalized) least-squares. This setting arises naturally in value function approximation for bandit algorithms and reinforcement learning.Our first main contribution is the observation that the amplification of the misspecification error when using least-squares is governed by the \emph{Lebesgue constant}, a classical quantity from approximation theory that depends on the choice of the feature subspace and the covariate distribution.We also show that this dependence on the misspecification error is tight for least-squares regression: in general, no method minimizing the empirical squared loss can improve it substantially. As a second contribution, we propose a method that augments the original feature set with auxiliary features designed to reduce the error amplification. For this method we prove an oracle inequality that shows that the method successfully competes with an ``oracle'' that knows the best way of using the auxiliary features to reduce error amplification.As an illustration, when the domain is a real interval and the features are monomials, we prove that in the limit as $d\to\infty$, our method reduces the amplification factor to $O(1)$. Note that without our method, least-squares with the monomials (and in fact polynomials) will suffers a worst-case error of order $\Omega(d)$ times the one of the best uniform linear approximator.



Paperid:3524
Authors:Antoine Moulin, Emmanuel Esposito, Dirk van der Hoeven
Abstract:
Abstract:We consider the problem setting of prediction with expert advice with possibly heavy-tailed losses, i.e.\ the only assumption on the losses is an upper bound on their second moments, denoted by $\theta$. We develop adaptive algorithms that do not require any prior knowledge about the range or the second moment of the losses. Existing adaptive algorithms have what is typically considered a lower-order term in their regret guarantees. We show that this lower-order term, which is often the maximum of the losses, can actually dominate the regret bound in our setting. Specifically, we show that even with small constant $\theta$, this lower-order term can scale as $\sqrt{KT}$, where $K$ is the number of experts and $T$ is the time horizon. We propose adaptive algorithms with improved regret bounds that avoid the dependence on such a lower-order term and guarantee $\mathcal{O}(\sqrt{\theta T\log(K)})$ regret in the worst case, and $\mathcal{O}(\theta \log(KT)/\Delta_{\min})$ regret when the losses are sampled i.i.d.\ from some fixed distribution, where $\Delta_{\min}$ is the difference between the mean losses of the second best expert and the best expert. Additionally, when the loss function is the squared loss, our algorithm also guarantees improved regret bounds over prior results.



Paperid:3525
Authors:Wei Huang, Andi Han, Yujin Song, Yilan Chen, Denny Wu, Difan Zou, Taiji Suzuki
Abstract:
The capacity of deep learning models is often large enough to both learn the underlying statistical signal and overfit to noise in the training set. This noise memorization can be harmful especially for data with a low signal-to-noise ratio (SNR), leading to poor generalization. Inspired by prior observations that label noise provides implicit regularization that improves generalization, in this work, we investigate whether introducing label noise to the gradient updates can enhance the test performance of neural network (NN) in the low SNR regime. Specifically, we consider training a two-layer NN with a simple label noise gradient descent (GD) algorithm, in an idealized signal-noise data setting. We prove that adding label noise during training suppresses noise memorization, preventing it from dominating the learning process; consequently, label noise GD enjoys rapid signal growth while the overfitting remains controlled, thereby achieving good generalization despite the low SNR. In contrast, we also show that NN trained with standard GD tends to overfit to noise in the same low SNR setting and establish a non-vanishing lower bound on its test error, thus demonstrating the benefit of introducing label noise in gradient-based training.



Paperid:3526
Authors:Razaib Tariq, Minji Heo, Shahroz Tariq, Simon Woo
Abstract:
Deepfake detection remains a pressing challenge, particularly in real-world settings where smartphone-captured media from digital screens often introduces Moiré artifacts that can distort detection outcomes. This study systematically evaluates state-of-the-art (SOTA) deepfake detectors on Moiré-affected videos—an issue that has received little attention. We collected a dataset of 12,832 videos, spanning 35.64 hours, from Celeb-DF, DFD, DFDC, UADFV, and FF++ datasets, capturing footage under diverse real-world conditions, including varying screens, smartphones, lighting setups, and camera angles. To further examine the influence of Moiré patterns on deepfake detection, we conducted additional experiments using our DeepMoiréFake, referred to as (DMF) dataset, and two synthetic Moiré generation techniques. Across 15 top-performing detectors, our results show that Moiré artifacts degrade performance by as much as 25.4\%, while synthetically generated Moiré patterns lead to a 21.4\% drop in accuracy. Surprisingly, demoiréing methods, intended as a mitigation approach, instead worsened the problem, reducing accuracy by up to 16\%. These findings underscore the urgent need for detection models that can robustly handle Moiré distortions alongside other real-world challenges, such as compression, sharpening, and blurring. By introducing the DMF dataset, we aim to drive future research toward closing the gap between controlled experiments and practical deepfake detection.



Paperid:3527
Authors:Jing-An Sun, Hang Fan, Junchao Gong, Ben Fei, Kun Chen, Fenghua Ling, zhangwenlong, Wanghan Xu, Li Yan, Pierre Gentine, LEI BAI
Abstract:
Data assimilation (DA) aims to estimate the full state of a dynamical system by combining partial and noisy observations with a prior model forecast, commonly referred to as the background. In atmospheric applications, the problem is fundamentally ill-posed due to the sparsity of observations relative to the high-dimensional state space. Traditional methods address this challenge by simplifying background priors to regularize the solution, which are empirical and require continual tuning for application. Inspired by alignment techniques in text-to-image diffusion models, we propose Align-DA, which formulates DA as a generative process and uses reward signals to guide—replacing manual tuning with data-driven alignment. Specifically, we train a score-based model in the latent space to approximate the background-conditioned prior, and align it using three complementary reward signals for DA: (1) assimilation accuracy, (2) forecast skill initialized from the assimilated state, and (3) physical consistency of the analysis fields. Experiments with multiple reward signals demonstrate consistent improvements in analysis quality across different evaluation metrics and observation-guidance strategies. These results show that preference alignment, implemented as a soft constraint, can automatically adapt complex priors tailored to DA, offering a promising new direction for advancing the field.



Paperid:3528
Authors:Yong Liu, Zirui Zhu, Chaoyu Gong, Minhao Cheng, Cho-Jui Hsieh, Yang You
Abstract:
While fine-tuning large language models (LLMs) for specific tasks often yields impressive results, it comes at the cost of memory inefficiency due to back-propagation in gradient-based training. Memory-efficient Zeroth-order (MeZO) optimizers, recently proposed to address this issue, only require forward passes during training, making them more memory-friendly. However, compared with exact gradients, ZO-based gradients usually exhibit an estimation error, which can significantly hurt the optimization process, leading to slower convergence and suboptimal solutions. In addition, we find that the estimation error will hurt more when adding to large weights instead of small weights. Based on this observation, this paper introduces Sparse MeZO, a novel memory-efficient zeroth-order optimization approach that applies ZO only to a carefully chosen subset of parameters. We propose a simple yet effective parameter selection scheme that yields significant performance gains with Sparse-MeZO. Additionally, we develop a memory-optimized implementation for sparse masking, ensuring the algorithm requires only inference-level memory consumption, allowing Sparse-MeZO to fine-tune LLaMA-30b on a single A100 GPU. Experimental results illustrate that Sparse-MeZO consistently improves both performance and convergence speed over MeZO without any overhead. For example, it achieves a 9% absolute accuracy improvement and 3.5x speedup over MeZO on the RTE task.



Paperid:3529
Authors:Seo Hyun Kim, Sunwoo Hong, Youngrok Park, Hojung Jung, Se-Young Yun
Abstract:
Abstract:Masked diffusion models have demonstrated competitive results on various tasks including language generation. However, due to its multiple refinement process, the inference is often bottlenecked by slow and static sampling speed. To overcome this problem, we introduce `KL-Adaptive Stability Sampling' (KLASS), a fast yet effective sampling method that exploits token-level KL divergence to identify stable, high-confidence predictions. By unmasking multiple tokens in each iteration without any additional model training, our approach speeds up generation significantly while maintaining sample quality. We test our sampler on diverse domains including $2\times$ faster in wall-clock time while improving performance in reasoning benchmarks (GSM8K and MATH) compared to standard greedy decoding, achieving state-of-the-art result among diffusion sampler.



Paperid:3530
Authors:Zeyu Zhang, Quanyu Dai, Luyu Chen, Zeren Jiang, Rui Li, Jieming Zhu, Xu Chen, Yi Xie, Zhenhua Dong, Ji-Rong Wen
Abstract:
LLM-based agents have been widely applied as personal assistants, capable of memorizing information from user messages and responding to personal queries. However, there still lacks an objective and automatic evaluation on their memory capability, largely due to the challenges in constructing reliable questions and answers (QAs) according to user messages. In this paper, we propose MemSim, a Bayesian simulator designed to automatically construct reliable QAs from generated user messages, simultaneously keeping their diversity and scalability. Specifically, we introduce the Bayesian Relation Network (BRNet) and a causal generation mechanism to mitigate the impact of LLM hallucinations on factual information, facilitating the automatic creation of an evaluation dataset. Based on MemSim, we generate a dataset in the daily-life scenario, named MemDaily, and conduct extensive experiments to assess the effectiveness of our approach. We also provide a benchmark for evaluating different memory mechanisms in LLM-based agents with the MemDaily dataset.



Paperid:3531
Authors:Wenyu Zhu, Jianhui Wang, Bowen Gao, Yinjun Jia, Haichuan Tan, Ya-Qin Zhang, Wei-Ying Ma, Yanyan Lan
Abstract:
Virtual screening (VS) is a critical component of modern drug discovery, yet most existing methods—whether physics-based or deep learning-based—are developed around {\em holo} protein structures with known ligand-bound pockets. Consequently, their performance degrades significantly on {\em apo} or predicted structures such as those from AlphaFold2, which are more representative of real-world early-stage drug discovery, where pocket information is often missing. In this paper, we introduce an alignment-and-aggregation framework to enable accurate virtual screening under structural uncertainty. Our method comprises two core components: (1) a tri-modal contrastive learning module that aligns representations of the ligand, the \textit{holo} pocket, and cavities detected from structures, thereby enhancing robustness to pocket localization error; and (2) a cross-attention based adapter for dynamically aggregating candidate binding sites, enabling the model to learn from activity data even without precise pocket annotations. We evaluated our method on a newly curated benchmark of \textit{apo} structures, where it significantly outperforms state-of-the-art methods in blind apo setting, improving the early enrichment factor (EF1\%) from 11.75 to 37.19. Notably, it also maintains strong performance on \textit{holo} structures. These results demonstrate the promise of our approach in advancing first-in-class drug discovery, particularly in scenarios lacking experimentally resolved protein-ligand complexes.



Paperid:3532
Authors:Yiqun Chen, Lingyong Yan, Weiwei Sun, Xinyu Ma, Yi Zhang, Shuaiqiang Wang, Dawei Yin, Yiming Yang, Jiaxin Mao
Abstract:
Retrieval-augmented generation (RAG) is widely utilized to incorporate external knowledge into large language models, thereby enhancing factuality and reducing hallucinations in question-answering (QA) tasks. A standard RAG pipeline consists of several components, such as query rewriting, document retrieval, document filtering, and answer generation. However, these components are typically optimized separately through supervised fine-tuning, which can lead to misalignments between the objectives of individual components and the overarching aim of generating accurate answers. Although recent efforts have explored using reinforcement learning (RL) to optimize specific RAG components, these approaches often focus on simple pipelines with only two components or do not adequately address the complex interdependencies and collaborative interactions among the modules. To overcome these limitations, we propose treating the complex RAG pipeline with multiple components as a multi-agent cooperative task, in which each component can be regarded as an RL agent. Specifically, we present MMOA-RAG, \textbf{M}ulti-\textbf{M}odule joint \textbf{O}ptimization \textbf{A}lgorithm for \textbf{RAG}, which employs multi-agent reinforcement learning to harmonize all agents' goals toward a unified reward, such as the F1 score of the final answer. Experiments conducted on various QA benchmarks demonstrate that MMOA-RAG effectively boost the overall performance of the pipeline and outperforms existing baselines. Furthermore, comprehensive ablation studies validate the contributions of individual components and demonstrate MMOA-RAG can be adapted to different RAG pipelines and benchmarks. The code of MMOA-RAG is on \url{https://anonymous.4open.science/r/MMOA-RAG-2DCC}.



Paperid:3533
Authors:Wenbo Li, Yan Xu, Mingde Yao, Fengjie Liang, Jiankai Sun, Menglu Wang, Guofeng Zhang, Linjiang Huang, Hongsheng Li
Abstract:
Novel view synthesis has been significantly advanced by the recent techniques, \eg, 3D Gaussian Splatting (3DGS). However, most current methods still require accurate camera pose estimation by Structure from Motion (SfM), otherwise, the performance would degrade significantly, especially in textureless indoor scenes. In this paper, we propose an efficient, pose-free 3DGS framework that can jointly optimize the 3D representation and camera poses. This framework enables robust reconstruction in textureless indoor regions where SfM generally fails. To enhance robustness in indoor scenes, where textureless regions and abrupt camera motions prevail, we introduce a local-to-global optimization scheme that progressively builds a global 3D representation from local estimations. Specifically, in the local stage, we construct a local neural scene representation supervised by depth foundation models to estimate camera pose and depth in textureless environments. In the global stage, the estimated local geometries are progressively merged into the global scene representation, and the camera poses are further refined. To address the abrupt camera motions in indoor scenes, we propose to decompose the rotation and translation components during pose optimization based on the local plane assumption and constrain the cross-frame consistency. Extensively benchmarked on public datasets, our method demonstrates state-of-the-art performance in both camera pose estimation and 3D scene reconstruction. Code will be public upon acceptance.



Paperid:3534
Authors:Shiting Xiao, Rishabh Kabra, Yuhang Li, Donghyun Lee, Joao Carreira, Priyadarshini Panda
Abstract:
The ability to segment objects based on open-ended language prompts remains a critical challenge, requiring models to ground textual semantics into precise spatial masks while handling diverse and unseen categories. We present OpenWorldSAM, a framework that extends the prompt-driven Segment Anything Model v2 (SAM2) to open-vocabulary scenarios by integrating multi-modal embeddings extracted from a lightweight vision-language model (VLM). Our approach is guided by four key principles: i) Unified prompting: OpenWorldSAM supports a diverse range of prompts, including category-level and sentence-level language descriptions, providing a flexible interface for various segmentation tasks. ii) Efficiency: By freezing the pre-trained components of SAM2 and the VLM, we train only 4.5 million parameters on the COCO-stuff dataset, achieving remarkable resource efficiency. iii) Instance Awareness: We enhance the model's spatial understanding through novel positional tie-breaker embeddings and cross-attention layers, enabling effective segmentation of multiple instances. iiv) Generalization: OpenWorldSAM exhibits strong zero-shot capabilities, generalizing well on unseen categories and an open vocabulary of concepts without additional training. Extensive experiments demonstrate that OpenWorldSAM achieves state-of-the-art performance in open-vocabulary semantic, instance, and panoptic segmentation across multiple benchmarks, including ADE20k, PASCAL, ScanNet, and SUN-RGBD.



Paperid:3535
Authors:Yusheng Zhao, Qixin Zhang, Xiao Luo, Weizhi Zhang, Zhiping Xiao, Wei Ju, Philip S Yu, Ming Zhang
Abstract:
Large language models (LLMs) have been used in many zero-shot learning problems, with their strong generalization ability. Recently, adopting LLMs in text-attributed graphs (TAGs) has drawn increasing attention. However, the adoption of LLMs faces two major challenges: limited information on graph structure and unreliable responses. LLMs struggle with text attributes isolated from the graph topology. Worse still, they yield unreliable predictions due to both information insufficiency and the inherent weakness of LLMs (e.g., hallucination). Towards this end, this paper proposes a novel method named Dynamic Text Bundling Supervision (DENSE) that queries LLMs with bundles of texts to obtain bundle-level labels and uses these labels to supervise graph neural networks. Specifically, we sample a set of bundles, each containing a set of nodes with corresponding texts of close proximity. We then query LLMs with the bundled texts to obtain the label of each bundle. Subsequently, the bundle labels are used to supervise the optimization of graph neural networks, and the bundles are further refined to exclude noisy items. To justify our design, we also provide theoretical analysis of the proposed method. Extensive experiments across ten datasets validate the effectiveness of the proposed method.



Paperid:3536
Authors:Siyan Zhao, Devaansh Gupta, Qinqing Zheng, Aditya Grover
Abstract:
Recent large language models (LLMs) have demonstrated strong reasoning capabilities that benefits from online reinforcement learning (RL).These capabilities have primarily been demonstrated within the left-to-right autoregressive (AR) generation paradigm. In contrast, non-autoregressive paradigms based on diffusion generate text in a coarse-to-fine manner. Although recent diffusion-based large language models (dLLMs) have achieved competitive language modeling performance compared to their AR counterparts, it remains unclear if dLLMs can also leverage recent advances in LLM reasoning.To this end, we propose, a framework to adapt pre-trained masked dLLMs into reasoning models via a combination of supervised finetuning (SFT) and RL.Specifically, we develop and extend techniques to improve reasoning in pretrained dLLMs: (a) we utilize a masked SFT technique to distill knowledge and instill self-improvement behavior directly from existing datasets, and (b) we introduce a novel critic-free, policy-gradient based RL algorithm called diffu-GRPO, the first integration of policy gradient methods to masked dLLMs. Through empirical studies, we investigate the performance of different post-training recipes on multiple mathematical and planning benchmarks. We find that d1 yields the best performance and significantly improves performance of a state-of-the-art dLLM.



Paperid:3537
Authors:Bastien Dubail, Stefan Stojanovic, Alexandre Proutiere
Abstract:
Low-rank structure is a common implicit assumption in many modern reinforcement learning (RL) algorithms. For instance, reward-free and goal-conditioned RL methods often presume that the successor measure admits a low-rank representation. In this work, we challenge this assumption by first remarking that the successor measure itself is not low-rank. Instead, we demonstrate that a low-rank structure naturally emerges in the shifted successor measure, which captures the system dynamics after bypassing a few initial transitions. We provide finite-sample performance guarantees for the entry-wise estimation of a low-rank approximation of the shifted successor measure from sampled entries. Our analysis reveals that both the approximation and estimation errors are primarily governed by the so-called spectral recoverability of the corresponding matrix. To bound this parameter, we derive a new class of functional inequalities for Markov chains that we call Type II Poincaré inequalities and from which we can quantify the amount of shift needed for effective low-rank approximation and estimation. This analysis shows in particular that the required shift depends on decay of the high-order singular values of the shifted successor measure and is hence typically small in practice. Additionally, we establish a connection between the necessary shift and the local mixing properties of the underlying dynamical system, which provides a natural way of selecting the shift. Finally, we validate our theoretical findings with experiments, and demonstrate that shifting the successor measure indeed leads to improved performance in goal-conditioned RL.



Paperid:3538
Authors:Chongyu Fan, Jiancheng Liu, Licong Lin, Jinghan Jia, Ruiqi Zhang, Song Mei, Sijia Liu
Abstract:
This work studies the problem of large language model (LLM) unlearning, aiming to remove unwanted data influences (e.g., copyrighted or harmful content) while preserving model utility. Despite the increasing demand for unlearning, a technically-grounded optimization framework is lacking. Gradient ascent (GA)-type methods, though widely used, are suboptimal as they reverse the learning process without controlling optimization divergence (i.e., deviation from the pre-trained state), leading to risks of model collapse. Negative preference optimization (NPO) has been proposed to address this issue and is considered one of the state-of-the-art LLM unlearning approaches. In this work, we revisit NPO and identify another critical issue: reference model bias. This bias arises from using the reference model (i.e., the model prior to unlearning) to assess unlearning success, which can lead to a misleading impression of the true data-wise unlearning effectiveness. Specifically, it could cause (a) uneven allocation of optimization power across forget data with varying difficulty levels, and (b) ineffective gradient weight smoothing during the early stages of unlearning optimization. To overcome these challenges, we propose a simple yet effective unlearning optimization framework, called SimNPO, showing that simplicity—removing the reliance on a reference model (through the lens of simple preference optimization)—benefits unlearning. We provide deeper insights into SimNPO's advantages, including an analysis based on mixtures of Markov chains. Extensive experiments further validate its efficacy on benchmarks like TOFU, MUSE, and WMDP.



Paperid:3539
Authors:Heyuan Li, Kenkun Liu, Lingteng Qiu, Qi Zuo, Keru Zheng, Zilong Dong, Xiaoguang Han
Abstract:
Tri-plane-like representations have been widely adopted in 3D-aware GANs for head image synthesis and other 3D object/scene modeling tasks due to their efficiency. However, querying features via Cartesian coordinate projection often leads to feature entanglement, which results in mirroring artifacts. A recent work, SphereHead, attempted to address this issue by introducing spherical tri-planes based on a spherical coordinate system. While it successfully mitigates feature entanglement, SphereHead suffers from uneven mapping between the square feature maps and the spherical planes, leading to inefficient feature map utilization during rendering and difficulties in generating fine image details.Moreover, both tri-plane and spherical tri-plane representations share a subtle yet persistent issue: feature penetration across convolutional channels can cause interference between planes, particularly when one plane dominates the others (see Fig. 1). These challenges collectively prevent tri-plane-based methods from reaching their full potential. In this paper, we systematically analyze these problems for the first time and propose innovative solutions to address them. Specifically, we introduce a novel hybrid-plane (hy-plane for short) representation that combines the strengths of both planar and spherical planes while avoiding their respective drawbacks. We further enhance the spherical plane by replacing the conventional theta-phi warping with a novel near-equal-area warping strategy, which maximizes the effective utilization of the square feature map. In addition, our generator synthesizes a single-channel unified feature map instead of multiple feature maps in separate channels, thereby effectively eliminating feature penetration. With a series of technical improvements, our hy-plane representation enables our method, HyPlaneHead, to achieve state-of-the-art performance in full-head image synthesis.



Paperid:3540
Authors:Ryan Punamiya, Dhruv Patel, Patcharapong Aphiwetsa, Pranav Kuppili, Lawrence Zhu, Simar Kareer, Judy Hoffman, Danfei Xu
Abstract:
Egocentric human experience data presents a vast resource for scaling up end-to-end imitation learning for robotic manipulation. However, significant domain gaps in visual appearance, sensor modalities, and kinematics between human and robot impede knowledge transfer. This paper presents EgoBridge, a unified co-training framework that explicitly aligns the policy latent spaces between human and robot data using domain adaptation. Through a measure of discrepancy on the joint policy latent features and actions based on Optimal Transport (OT), we learn observation representations that not only align between the human and robot domain but also preserve the action-relevant information critical for policy learning. EgoBridge achieves a significant absolute policy success rate improvement by 44% over human-augmented cross-embodiment baselines in three real-world single-arm and bimanual manipulation tasks. EgoBridge also generalizes to new objects, scenes, and tasks seen only in human data, where baselines fail entirely. Videos and additional information can be found at https://ego-bridge.github.io/



Paperid:3541
Authors:Adam Klivans, Konstantinos Stavropoulos, Kevin Tian, Arsen Vasilyan
Abstract:
Abstract:Inspired by recent work on learning with distribution shift, we give ageneral outlier removal algorithm called *iterative polynomialfiltering* and show a number of striking applications for supervisedlearning with contamination:(1) We show that any function class that can be approximated bylow-degree polynomials with respect to a hypercontractive distributioncan be efficiently learned under bounded contamination (alsoknown as *nasty noise*). This is a surprising resolution to alongstanding gap between the complexity of agnostic learning andlearning with contamination, as it was widely believed that low-degreeapproximators only implied tolerance to label noise.(2) For any function class that admits the (stronger) notion ofsandwiching approximators, we obtain near-optimal learning guaranteeseven with respect to heavy additive contamination, where far more than$1/2$ of the training set may be added adversarially. Priorrelated work held only for regression and in a list-decodable setting.(3) We obtain the first efficient algorithms for tolerant testablelearning of functions of halfspaces with respect to any fixedlog-concave distribution. Even the non-tolerant case for a singlehalfspace in this setting had remained open.These results significantly advance our understanding of efficientsupervised learning under contamination, a setting that has been muchless studied than its unsupervised counterpart.



Paperid:3542
Authors:Yuxuan Wang, Ming Yang, Weishuai Zeng, Yu Zhang, Xinrun Xu, Haobin Jiang, Gang Ding, Zongqing Lu
Abstract:
Achieving general agile whole-body control on humanoid robots remains a major challenge due to diverse motion demands and data conflicts. While existing frameworks excel in training single motion-specific policies, they struggle to generalize across highly varied behaviors due to conflicting control requirements and mismatched data distributions. In this work, we propose BUMBLEBEE (BB), an expert-generalist learning framework that combines motion clustering and sim-to-real adaptation to overcome these challenges. BB first leverages an autoencoder-based clustering method to group behaviorally similar motions using motion features and motion descriptions. Expert policies are then trained within each cluster and refined with real-world data through iterative delta action modeling to bridge the sim-to-real gap. Finally, these experts are distilled into a unified generalist controller that preserves agility and robustness across all motion types. Experiments on two simulations and a real humanoid robot demonstrate that BB achieves state-of-the-art general whole-body control, setting a new benchmark for agile, robust, and generalizable humanoid performance in the real world.



Paperid:3543
Authors:Wenhang Shi, Yiren Chen, Shuqing Bian, Xinyi Zhang, Kai Tang, Pengfei Hu, Zhe Zhao, WEI LU, Xiaoyong Du
Abstract:
Prompt engineering is crucial for leveraging the full potential of large language models (LLMs). While automatic prompt optimization offers a scalable alternative to costly manual design, generating effective prompts remains challenging. Existing methods often struggle to stably generate improved prompts, leading to low efficiency, and overlook that prompt optimization easily gets trapped in local optima. Addressing this, we propose GRACE, a framework that integrates two synergistic strategies: Gated Refinement and Adaptive Compression, achieving Efficient prompt optimization. The gated refinement strategy introduces a feedback regulation gate and an update rejection gate, which refine update signals to produce stable and effective prompt improvements. When optimization stagnates, the adaptive compression strategy distills the prompt’s core concepts, restructuring the optimization trace and opening new paths. By strategically introducing information loss through refinement and compression, GRACE delivers substantial gains in performance and efficiency. In extensive experiments on 11 tasks across three practical domains, including BIG-Bench Hard (BBH), domain-specific, and general NLP tasks, GRACE achieves significant average relative performance improvements of 4.7\%, 4.4\% and 2.7\% over state-of-the-art methods, respectively. Further analysis shows that GRACE achieves these gains using only 25\% of the prompt generation budget required by prior methods, highlighting its high optimization efficiency and low computational overhead. We will release the code upon paper acceptance.



Paperid:3544
Authors:Adam J. Eisen, Mitchell Ostrow, Sarthak Chandra, Leo Kozachkov, Earl Miller, Ila Fiete
Abstract:
Biological function arises through the dynamical interactions of multiple subsystems, including those between brain areas, within gene regulatory networks, and more. A common approach to understanding these systems is to model the dynamics of each subsystem and characterize communication between them. An alternative approach is through the lens of control theory: how the subsystems control one another. This approach involves inferring the directionality, strength, and contextual modulation of control between subsystems. However, methods for understanding subsystem control are typically linear and cannot adequately describe the rich contextual effects enabled by nonlinear complex systems. To bridge this gap, we devise a data-driven nonlinear control-theoretic framework to characterize subsystem interactions via the Jacobian of the dynamics. We address the challenge of learning Jacobians from time-series data by proposing the JacobianODE, a deep learning method that leverages properties of the Jacobian to directly estimate it for arbitrary dynamical systems from data alone. We show that JacobianODEs outperform existing Jacobian estimation methods on challenging systems, including high-dimensional chaos. Applying our approach to a multi-area recurrent neural network (RNN) trained on a working memory selection task, we show that the “sensory” area gains greater control over the “cognitive” area over learning. Furthermore, we leverage the JacobianODE to directly control the trained RNN, enabling precise manipulation of its behavior. Our work lays the foundation for a theoretically grounded and data-driven understanding of interactions among biological subsystems.



Paperid:3545
Authors:Han Chen, Hans-Georg Müller
Abstract:
The functional modeling of samples of distributions is a major challenge since distributions do not form a vector space. While various approaches exist for univariate distributions, including transformations to a Hilbert space, there are far fewer tools available for multivariate distributions. We introduce a transformation approach to map multivariate distributions to a Hilbert space by a map that proceeds via a Wasserstein slicing approach and can be inverted. Then tools from functional data analysis, including functional principal component analysis, functional modes of variation and functional regression can be adopted in the Hilbert space and the invertibility of the map makes it possible to obtain the distributions that correspond to the resulting functional objects in the Hilbert space. Convergence results for these Hilbert space representations can be obtained for a general class of such maps. The proposed method is illustrated by modeling joint distributions of systolic and diastolic blood pressure data.



Paperid:3546
Authors:Raghu Vamshi Hemadri, Jitendra Bhandari, Andre Nakkab, Johann Knechtel, Badri Gopalan, Ramesh Narayanaswamy, Ramesh Karri, Siddharth Garg
Abstract:
Modern chip design is complex, and there is a crucial need for early-stage prediction of key design-quality metrics like timing and routing congestion directly from Verilog code (a commonly used programming language for hardware design). It is especially important yet complex to predict individual lines of code that cause timing violations or downstream routing congestion. Prior works have tried approaches like converting Verilog into an intermediate graph representation and using LLM embeddings alongside other features to predict module-level quality, but did not consider line-level quality prediction. We propose VeriLoC, the first method that predicts design quality directly from Verilog at both the line- and module-level. To this end, VeriLoC leverages recent Verilog code-generation LLMs to extract local line-level and module-level embeddings, and trains downstream classifiers/regressors on concatenations of these embeddings. VeriLoC achieves high F1-scores of 0.86-0.95 for line-level congestion and timing prediction, and reduces the mean average percentage error from 14%-18% for SOTA methods down to only 4%. We believe that VeriLoC embeddings and insights from our work will also be of value for other predictive and optimization tasks for complex hardware design.



Paperid:3547
Authors:Yueqi Zhang, Peiwen Yuan, Yiwei Li, Shaoxiong Feng, Xinglin Wang, Jiayi Shi, Chuyi Tan, Boyuan Pan, Yao Hu, Prof. Kan
Abstract:
Human–AI conversation frequently relies on quoting earlier text—“check it with the formula I just highlighted”—yet today’s large language models (LLMs) lack an explicit mechanism for locating and exploiting such spans. We formalise the challenge as span-conditioned generation, decomposing each turn into the dialogue history, a set of token-offset quotation spans, and an intent utterance. Building on this abstraction, we introduce a quotation-centric data pipeline that automatically synthesises task-specific dialogues, verifies answer correctness through multi-stage consistency checks, and yields both a heterogeneous training corpus and the first benchmark covering five representative scenarios. To meet the benchmark’s zero-overhead and parameter-efficiency requirements, we propose QuAda, a lightweight training-based method that attaches two bottleneck projections to every attention head, dynamically amplifying or suppressing attention to quoted spans at inference time while leaving the prompt unchanged and updating < 2.8% of backbone weights. Experiments across models show that QuAda is suitable for all scenarios and generalises to unseen topics, offering an effective, plug-and-play solution for quotation-aware dialogue.



Paperid:3548
Authors:Sasan Sharifipour, Constantino Álvarez Casado, Mohammad Sabokrou, Miguel Bordallo Lopez
Abstract:
Training deep learning models for point cloud prediction tasks such as shape completion and generation depends critically on loss functions that measure discrepancies between predicted and ground-truth point sets. Commonly used functions such as Chamfer Distance (CD), HyperCD, and InfoCD rely on nearest-neighbor assignments, which often induce many-to-one correspondences, leading to point congestion in dense regions and poor coverage in sparse regions. These losses also involve non-differentiable operations due to index selection, which may affect gradient-based optimization. Earth Mover Distance (EMD) enforces one-to-one correspondences and captures structural similarity more effectively, but its cubic computational complexity limits its practical use. We propose the Adaptive Probabilistic Matching Loss (APML), a fully differentiable approximation of one-to-one matching that leverages Sinkhorn iterations on a temperature-scaled similarity matrix derived from pairwise distances. We analytically compute the temperature to guarantee a minimum assignment probability, eliminating manual tuning. APML achieves near-quadratic runtime, comparable to Chamfer-based losses, and avoids non-differentiable operations. When integrated into state-of-the-art architectures (PoinTr, PCNNet) on ShapeNet benchmarks and on a spatio‑temporal Transformer (CSI2PC) that \textit{generates} 3‑D human point clouds from WiFi‑CSI measurements, APM loss yields faster convergence, superior spatial distribution, especially in low-density regions, and improved or on-par quantitative performance without additional hyperparameter search. The code is available at: https://github.com/apm-loss/apml.



Paperid:3549
Authors:Dionysis Arvanitakis, Vaidehi Srinivas, Aravindan Vijayaraghavan
Abstract:
Abstract:Tensor decomposition is a canonical non-convex optimization problem that is computationally challenging, and yet important due to applications in factor analysis and parameter estimation of latent variable models. In practice, scalable iterative methods, particularly Alternating Least Squares (ALS), remain the workhorse for tensor decomposition despite the lack of global convergence guarantees. A popular approach to tackle challenging non-convex optimization problems is overparameterization--- on input an $n \times n \times n$ tensor of rank $r$, the algorithm can output a decomposition of potentially rank $k$ (potentially larger than $r$). On the theoretical side, overparameterization for iterative methods is challenging to reason about and requires new techniques. The work of Wang et al., (NeurIPS 2020) makes progress by showing that a variant of gradient descent globally converges when overparameterized to $k=O(r^{7.5} \log n)$. Our main result shows that overparameterization provably enables global convergence of ALS: on input a third order $n \times n \times n$ tensor with a decomposition of rank $r \ll n$, ALS overparameterized with rank $k=O(r^2)$ achieves global convergence with high probability under random initialization. Moreover our analysis also gives guarantees for the more general low-rank approximation problem. The analysis introduces new techniques for understanding iterative methods in the overparameterized regime based on new matrix anticoncentration arguments.



Paperid:3550
Authors:Yan-Shuo Liang, Jiarui Chen, Wu-Jun Li
Abstract:
Continual learning (CL), which requires the model to learn multiple tasks sequentially, is crucial for language models (LMs). Recently, low-rank adaptation (LoRA), one of the most representative parameter-efficient fine-tuning (PEFT) methods, has gained increasing attention in CL of LMs. However, most existing CL methods based on LoRA typically expand a new LoRA branch to learn each new task and force the new and old LoRA branches to contribute equally to old tasks, potentially leading to forgetting. In this work, we propose a new method, called gated integration of low-rank adaptation (GainLoRA), for CL of LMs. GainLoRA expands a new LoRA branch for each new task and introduces gating modules to integrate the new and old LoRA branches. Furthermore, GainLoRA leverages the new gating module to minimize the contribution from the new LoRA branch to old tasks, effectively mitigating forgetting and improving the model's overall performance. Experimental results on CL benchmarks demonstrate that GainLoRA outperforms existing state-of-the-art methods.



Paperid:3551
Authors:Eunbyeol Cho, Jiyoun Kim, Minjae Lee, Sungjin Park, Edward Choi
Abstract:
Electronic Health Records (EHR) are time-series relational databases that record patient interactions and medical events over time, serving as a critical resource for healthcare research and applications. However, privacy concerns and regulatory restrictions limit the sharing and utilization of such sensitive data, necessitating the generation of synthetic EHR datasets.Unlike previous EHR synthesis methods—which typically generate medical records consisting of expert-chosen features (e.g., a few vital signs, structured codes only)—we introduce RawMed, the first framework to synthesize multi-table, time-series EHR data that closely resembles raw EHRs.Using text-based representation and compression techniques, RawMed captures complex structures and temporal dynamics with minimal preprocessing. We also propose a new evaluation framework for multi-table time-series synthetic EHRs, assessing distributional similarity, inter-table relationships, temporal dynamics, and privacy. Validated on two open-source EHR datasets, RawMed outperforms baseline models in fidelity and utility. Source code will be publicly released.



Paperid:3552
Authors:Zeyu Zhu, KE LIANG, Lingyuan Meng, Meng Liu, Suyuan Liu, Renxiang Guan, Miaomiao Li, Wanwei Liu, Xinwang Liu
Abstract:
Spatial transcriptomics (ST) technologies provide gene expression measurements with spatial resolution, enabling the dissection of tissue structure and function. A fundamental challenge in ST analysis is clustering spatial spots into coherent functional regions. While existing models effectively integrate expression and spatial signals, they largely overlook sequence-level biological priors encoded in the DNA sequences of expressed genes. To bridge this gap, we propose SAINT (Sequence-Aware Integration for Nucleotide-informed Transcriptomics), a unified framework that augments spatial representation learning with nucleotide-derived features. We construct sequence-augmented datasets across 14 tissue sections from three widely used ST benchmarks (DLPFC, HBC, and MBA), retrieving reference DNA sequences for each expressed gene and encoding them using a pretrained Nucleotide Transformer. For each spot, gene-level embeddings are aggregated via expression-weighted and attention-based pooling, then fused with spatial-expression representations through a late fusion module. Extensive experiments demonstrate that SAINT consistently improves clustering performance across multiple datasets. Experiments validate the superiority, effectiveness, sensitivity, and transferability of our framework, confirming the complementary value of incorporating sequence-level priors into spatial transcriptomics clustering.



Paperid:3553
Authors:Can Yaras, Alec Xu, Pierre Abillama, Changwoo Lee, Laura Balzano
Abstract:
Abstract:Transformers have achieved state-of-the-art performance across various tasks, but suffer from a notable quadratic complexity in sequence length due to the attention mechanism. In this work, we propose MonarchAttention -- a novel approach to sub-quadratic attention approximation via Monarch matrices, an expressive class of structured matrices. Based on the variational form of softmax, we describe an efficient optimization-based algorithm to compute an approximate projection of softmax attention onto the class of Monarch matrices with $\Theta(N\sqrt{N} d)$ computational complexity and $\Theta(Nd)$ memory/IO complexity. Unlike previous approaches, MonarchAttention is both (1) transferable, yielding minimal performance loss with no additional training, even when replacing every attention layer of the transformer, and (2) hardware-efficient, utilizing the highest-throughput tensor core units on modern GPUs. With optimized kernels, MonarchAttention achieves substantial speed-ups in wall-time over FlashAttention-2: $1.4\times$ for shorter sequences $(N=256)$, $4.5\times$ for medium-length sequences $(N=4K)$, and $8.2\times$ for longer sequences $(N=16K)$. We demonstrate the quality of MonarchAttention on diverse tasks and architectures in vision and language problems, showing that it flexibly and accurately approximates softmax attention in a variety of contexts.



Paperid:3554
Authors:Sihyung Park, Wenbin Lu, Shu Yang
Abstract:
Truncation by death, a prevalent challenge in critical care, renders traditional dynamic treatment regime (DTR) evaluation inapplicable due to ill-defined potential outcomes. We introduce a principal stratification-based method, focusing on the always-survivor value function. We derive a semiparametrically efficient, multiply robust estimator for multi-stage DTRs, demonstrating its robustness and efficiency. Empirical validation and an application to electronic health records showcase its utility for personalized treatment optimization.



Paperid:3555
Authors:Chen Li, Huiying Xu, Changxin Gao, Zeyu Wang, Yun Liu, Xinzhong Zhu
Abstract:
Single-Domain Generalized Object Detection (SDGOD), as a cutting-edge research topic in computer vision, aims to enhance model generalization capability in unseen target domains through single-source domain training. Current mainstream approaches attempt to mitigate domain discrepancies via data augmentation techniques. However, due to domain shift and limited domain‑specific knowledge, models tend to fall into the pitfall of spurious correlations. This manifests as the model's over-reliance on simplistic classification features (e.g., color) rather than essential domain-invariant representations like object contours. To address this critical challenge, we propose the Cauvis (Causal Visual Prompts) method. First, we introduce a Cross-Attention Prompts module that mitigates bias from spurious features by integrating visual prompts with cross-attention. To address the inadequate domain knowledge coverage and spurious feature entanglement in visual prompts for single-domain generalization, we propose a dual-branch adapter that disentangles causal-spurious features while achieving domain adaptation via high-frequency feature extraction. Cauvis achieves state-of-the-art performance with 15.9–31.4\% gains over existing domain generalization methods on SDGOD datasets, while exhibiting significant robustness advantages in complex interference environments.



Paperid:3556
Authors:Jane Lee, Baturay Saglam, Spyridon Pougkakiotis, Amin Karbasi, Dionysis Kalogerias
Abstract:
Constrained optimization provides a common framework for dealing with conflicting objectives in reinforcement learning (RL). In most of these settings, the objectives (and constraints) are expressed though the expected accumulated reward. However, this formulation neglects risky or even possibly catastrophic events at the tails of the reward distribution, and is often insufficient for high-stakes applications in which the risk involved in outliers is critical. In this work, we propose a framework for risk-aware constrained RL, which exhibits per-stage robustness properties jointly in reward values and time using optimized certainty equivalents (OCEs). Our framework ensures an exact equivalent to the original constrained problem within a parameterized strong Lagrangian duality framework under appropriate constraint qualifications, and yields a simple algorithmic recipe which can be wrapped around standard RL solvers, such as PPO. Lastly, we establish the convergence of the proposed algorithm and verify the risk-aware properties of our approach through several numerical experiments.



Paperid:3557
Authors:Giuseppe Bruno, Federico Pasqualotto, Andrea Agazzi
Abstract:
Abstract:In this paper, we study the evolution of tokens through the depth of encoder-only transformer models at inference time by modeling them as a system of particles interacting in a mean-field way and studying the corresponding dynamics. More specifically, we consider this problem in the moderate interaction regime, where the number $N$ of tokens is large and the inverse temperature parameter $\beta$ of the model scales together with $N$. In this regime, the dynamics of the system displays a multiscale behavior: a fast phase, where the token empirical measure collapses on a low-dimensional space, an intermediate phase, where the measure further collapses into clusters, and a slow one, where such clusters sequentially merge into a single one. We provide a rigorous characterization of the limiting dynamics in each of these phases and prove convergence in the above mentioned limit, exemplifying our results with some simulations.



Paperid:3558
Authors:Kaibo Wang, Jianda Mao, Tong Wu, Yang Xiang
Abstract:
Classifier-Free Guidance (CFG) is an essential component of text-to-image diffusion models, and understanding and advancing its operational mechanisms remain a central focus of research. Existing approaches stem from divergent theoretical interpretations, thereby limiting the design space and obscuring key design choices. To address this, we propose a unified perspective that reframes conditional guidance as fixed point iterations, seeking to identify a golden path where latents produce consistent outputs under both conditional and unconditional generation. We demonstrate that CFG and its variants constitute a special case of single-step short-sighted iteration, which is theoretically proven to exhibit inefficiency. To this end, we introduce Foresight Guidance (FSG), which prioritizes solving longer-interval subproblems in early diffusion stages with increased iterations. Extensive experiments across diverse datasets and model architectures validate the superiority of FSG over state-of-the-art methods in both image quality and computational efficiency. Our work offers novel perspectives for unlocking the potential of conditional guidance and adaptive design.



Paperid:3559
Authors:Yuchen Li, Chaoran Feng, Zhenyu Tang, Kaiyuan Deng, Wangbo Yu, Yonghong Tian, Li Yuan
Abstract:
We introduce GS2E (Gaussian Splatting to Event Generation), a large-scale synthetic event dataset designed for high-fidelity event vision tasks, captured from real-world sparse multi-view RGB images. Existing event datasets are often synthesized from dense RGB videos, which typically suffer from limited viewpoint diversity and geometric inconsistency, or rely on expensive, hard-to-scale hardware setups. GS2E addresses these limitations by first reconstructing photorealistic static scenes using 3D Gaussian Splatting, followed by a novel, physically-informed event simulation pipeline. This pipeline integrates adaptive trajectory interpolation with physically-consistent event contrast threshold modeling. As a result, it generates temporally dense and geometrically consistent event streams under diverse motion and lighting conditions, while maintaining strong alignment with the underlying scene structure. Experimental results on event-based 3D reconstruction highlight GS2E’s superior generalization capabilities and its practical value as a benchmark for advancing event vision research.



Paperid:3560
Authors:Shanchuan Lin, Ceyuan Yang, Hao He, Jianwen Jiang, Yuxi Ren, Xin Xia, Yang Zhao, Xuefeng Xiao, Lu Jiang
Abstract:
Existing large-scale video generation models are computationally intensive, preventing adoption in real-time and interactive applications. In this work, we propose autoregressive adversarial post-training (AAPT) to turn a pre-trained latent video diffusion model intoa real-time, interactive, streaming video generator. Our model autoregressively generates a latent frame at a time using a single neural function evaluation (1NFE). The model can stream the result to the user in real time and receive interactive responses as control to generate the next latent frame. Unlike existing approaches, our method explores adversarial training as an effective paradigm for autoregressive generation. This allows us to design a more efficient architecture for one-step generation and to train the model in a student-forcing way to mitigate error accumulation. The adversarial approach also enables us to train the model for long-duration generation fully utilizing the KV cache. As a result, our 8B model achieves real-time, 24fps, nonstop, streaming video generation at 736x416 resolution on a single H100, or 1280x720 on 8xH100 up to a minute long (1440 frames).



Paperid:3561
Authors:Pai Liu, Lingfeng Zhao, Shivangi Agarwal, Jinghan Liu, Audrey Huang, Philip Amortila, Nan Jiang
Abstract:
Holdout validation and hyperparameter tuning from data is a long-standing problem in offline reinforcement learning (RL). A standard framework is to use off-policy evaluation (OPE) methods to evaluate and select the policies, but OPE either incurs exponential variance (e.g., importance sampling) or has hyperparameters on their own (e.g., FQE and model-based). In this work we focus on hyperparameter tuning for OPE itself, which is even more under-investigated. Concretely, we select among candidate value functions ("model-free") or dynamics ("model-based") to best assess the performance of a target policy. We develop: (1) new model-free and model-based selectors with theoretical guarantees, and (2) a new experimental protocol for empirically evaluating them. Compared to the model-free protocol in prior works, our new protocol allows for more stable generation of candidate value functions, better control of misspecification, and evaluation of model-free and model-based methods alike. We exemplify the protocol on Gym-Hopper, and find that our new model-free selector, LSTD-Tournament, demonstrates promising empirical performance.



Paperid:3562
Authors:Jaebin Lee, Hankook Lee
Abstract:
In machine learning, effective modeling requires a holistic consideration of how to encode inputs, make predictions (i.e., decoding), and train the model. However, in time-series forecasting, prior work has predominantly focused on encoder design, often treating prediction and training as separate or secondary concerns. In this paper, we propose TimePerceiver, a unified encoder-decoder forecasting framework that is tightly aligned with an effective training strategy. To be specific, we first generalize the forecasting task to include diverse temporal prediction objectives such as extrapolation, interpolation, and imputation. Since this generalization requires handling input and target segments that are arbitrarily positioned along the temporal axis, we design a novel encoder-decoder architecture that can flexibly perceive and adapt to these varying positions. For encoding, we introduce a set of latent bottleneck representations that can interact with all input segments to jointly capture temporal and cross-channel dependencies. For decoding, we leverage learnable queries corresponding to target timestamps to effectively retrieve relevant information. Extensive experiments demonstrate that our framework consistently and significantly outperforms prior state-of-the-art baselines across a wide range of benchmark datasets.



Paperid:3563
Authors:Yulei Qin, Gang Li, Zongyi Li, Zihan Xu, Yuchen Shi, Zhekai Lin, Xiao Cui, Ke Li, Xing Sun
Abstract:
Existing large language models (LLMs) face challenges of following complex instructions, especially when multiple constraints are present and organized in paralleling, chaining, and branching structures. One intuitive solution, namely chain-of-thought (CoT), is expected to universally improve capabilities of LLMs. However, we find that the vanilla CoT exerts a negative impact on performance due to its superficial reasoning pattern of simply paraphrasing the instructions. It fails to peel back the compositions of constraints for identifying their relationship across hierarchies of types and dimensions. To this end, we propose a systematic method to boost LLMs in dealing with complex instructions via incentivizing reasoning for test-time compute scaling. First, we stem from the decomposition of complex instructions under existing taxonomies and propose a reproducible data acquisition method. Second, we exploit reinforcement learning (RL) with verifiable rule-centric reward signals to cultivate reasoning specifically for instruction following. We address the shallow, non-essential nature of reasoning under complex instructions via sample-wise contrast for superior CoT enforcement. We also exploit behavior cloning of experts to facilitate steady distribution shift from fast-thinking LLMs to skillful reasoners. Extensive evaluations on seven comprehensive benchmarks confirm the validity of the proposed method, where a 1.5B LLM achieves 11.74% gains with performance comparable to a 8B LLM. Codes and data are available at https://anonymous.4open.science/r/IRAIF-B3A0/README.md



Paperid:3564
Authors:Yanggan Gu, Yuanyi Wang, Zhaoyi Yan, Yiming Zhang, Qi Zhou, Fei Wu, Hongxia Yang
Abstract:
Model fusion combines multiple Large Language Models (LLMs) with different strengths into a more powerful, integrated model through lightweight training methods. Existing works on model fusion focus primarily on supervised fine-tuning (SFT), leaving preference alignment (PA) —a critical phase for enhancing LLM performance—largely unexplored.The current few fusion methods on PA phase, like WRPO, simplify the process by utilizing only response outputs from source models while discarding their probability information. To address this limitation, we propose InfiFPO, a preference optimization method for implicit model fusion.InfiFPO replaces the reference model in Direct Preference Optimization (DPO) with a fused source model that synthesizes multi-source probabilities at the sequence level, circumventing complex vocabulary alignment challenges in previous works and meanwhile maintaining the probability information.By introducing probability clipping and max-margin fusion strategies, InfiFPO enables the pivot model to align with human preferences while effectively distilling knowledge from source models.Comprehensive experiments on 11 widely-used benchmarks demonstrate that InfiFPO consistently outperforms existing model fusion and preference optimization methods. When using Phi-4 as the pivot model, InfiFPO improves its average performance from 79.95 to 83.33 on 11 benchmarks, significantly improving its capabilities in mathematics, coding, and reasoning tasks.Our code is available at https://anonymous.4open.science/r/InfiFPO-1190.



Paperid:3565
Authors:ZIhui Cheng, Qiguang Chen, Xiao Xu, Jiaqi Wang, Weiyun Wang, Hao Fei, Yidong Wang, Alex Jinpeng Wang, Zhi Chen, Wanxiang Che, Libo Qin
Abstract:
Abstract:Large Vision-Language Models (LVLMs) have achieved significant success in multimodal tasks, with multimodal chain-of-thought (MCoT) further enhancing performance and interpretability. Recent MCoT methods fall into two categories: (i) Textual-MCoT (T-MCoT), which takes multimodal input and produces textual output; and (ii) Interleaved-MCoT (I-MCoT), which generates interleaved image-text outputs. Despite advances in both approaches, the mechanisms driving these improvements are not fully understood. To fill this gap, we first reveal that MCoT boosts LVLMs by incorporating $\textit{visual thoughts}$, which convey image information to the reasoning process regardless of the MCoT format, depending only on clarity and conciseness of expression. Furthermore, to explore visual thoughts systematically, we define four distinct forms of visual thought expressions and analyze them comprehensively. Our findings demonstrate that these forms differ in clarity and conciseness, yielding varying levels of MCoT improvement. Additionally, we explore the internal nature of visual thoughts, finding that visual thoughts serve as intermediaries between the input image and reasoning to deeper transformer layers, enabling more advanced visual information transmission. We hope that the visual thoughts can inspire further breakthroughs for future MCoT research.



Paperid:3566
Authors:Chen Wang, Chuhao Chen, Yiming Huang, Zhiyang Dou, Yuan Liu, Jiatao Gu, Lingjie Liu
Abstract:
Existing video generation models excel at producing photo-realistic videos from text or images, but often lack physical plausibility and 3D controllability. To overcome these limitations, we introduce PhysCtrl, a novel framework for physics-grounded image-to-video generation with physical parameters and force control. At its core is a generative physics network that learns the distribution of physical dynamics across four materials (elastic, sand, plasticine, and rigid) via a diffusion model conditioned on physics parameters and applied forces. We represent physical dynamics as 3D point trajectories and train on a large-scale synthetic dataset of 550K animations generated by physics simulators. We enhance the diffusion model with a novel spatiotemporal attention block that emulates particle interactions and incorporates physics-based constraints during training to enforce physical plausibility. Experiments show that PhysCtrl generates realistic, physics-grounded motion trajectories which, when used to drive image-to-video models, yield high-fidelity, controllable videos that outperform existing methods in both visual quality and physical plausibility. Our code, model and data will be made publicly available upon publication.



Paperid:3567
Authors:Siyu Xu, Yunke Wang, Chenghao Xia, Dihao Zhu, Tao Huang, Chang Xu
Abstract:
Vision-Language-Action (VLA) models have demonstrated strong multi-modal reasoning capabilities, enabling direct action generation from visual perception and language instructions in an end-to-end manner. However, their substantial computational cost poses a challenge for real-time robotic control, where rapid decision-making is essential. This paper introduces VLA-Cache, a training-free inference acceleration method that reduces computational overhead by adaptively caching and reusing static visual tokens across frames. Exploiting the temporal continuity in robotic manipulation, VLA-Cache identifies minimally changed tokens between adjacent frames and reuses their cached key-value representations, thereby circumventing redundant computations. Additionally, to maintain action precision, VLA-Cache selectively re-computes task-relevant tokens that are environmentally sensitive, ensuring the fidelity of critical visual information. To further optimize efficiency, we introduce a layer adaptive token reusing strategy that dynamically adjusts the reuse ratio based on attention concentration across decoder layers, prioritizing critical tokens for recomputation. Extensive experiments on two simulation platforms (LIBERO and SIMPLER) and a real-world robotic system demonstrate that VLA-Cache achieves up to 1.7× speedup in CUDA latency and a 15% increase in control frequency, with negligible loss on task success rate. The manipulation videos are available at the following anonymous link: https://anonymous-5408-neurips-2025.glitch.me/.



Paperid:3568
Authors:Jiawen Yu, Hairuo Liu, Qiaojun Yu, Jieji Ren, Ce Hao, Haitong Ding, Guangyu Huang, Guofan Huang, Yan Song, Panpan Cai, Cewu Lu, Wenqiang Zhang
Abstract:
Abstract:Vision-Language-Action (VLA) models have advanced general-purpose robotic manipulation by leveraging pretrained visual and linguistic representations. However, they struggle with contact-rich tasks that require fine-grained control involving force, especially under visual occlusion or dynamic uncertainty. To address these limitations, we propose \textbf{ForceVLA}, a novel end-to-end manipulation framework that treats external force sensing as a first-class modality within VLA systems. ForceVLA introduces \textbf{FVLMoE}, a force-aware Mixture-of-Experts fusion module that dynamically integrates pretrained visual-language embeddings with real-time 6-axis force feedback during action decoding. This enables context-aware routing across modality-specific experts, enhancing the robot's ability to adapt to subtle contact dynamics. We also introduce \textbf{ForceVLA-Data}, a new dataset comprising synchronized vision, proprioception, and force-torque signals across five contact-rich manipulation tasks. ForceVLA improves average task success by 23.2\% over strong $\pi_0$-based baselines, achieving up to 80\% success in tasks such as plug insertion. Our approach highlights the importance of multimodal integration for dexterous manipulation and sets a new benchmark for physically intelligent robotic control. Code and data will be released at https://sites.google.com/view/forcevla2025/.



Paperid:3569
Authors:Daniel Bolya, Po-Yao Huang, Peize Sun, Jang Hyun Cho, Andrea Madotto, Chen Wei, Tengyu Ma, Jiale Zhi, Jathushan Rajasegaran, Hanoona Bangalath, Junke Wang, Marco Monteiro, Hu Xu, Shiyu Dong, Nikhila Ravi, Shang-Wen Li, Piotr Dollar, Christoph Feichtenhofer
Abstract:
We introduce Perception Encoder (PE), a family of state-of-the-art vision encoders for image and video understanding. Traditionally, vision encoders have relied on a variety of pretraining objectives, each excelling at different downstream tasks. Surprisingly, after scaling a carefully tuned image pretraining recipe and refining with a robust video data engine, we find that contrastive vision-language training alone can produce strong, general embeddings for all of these downstream tasks. There is only one caveat: these embeddings are hidden within the intermediate layers of the network. To draw them out, we introduce two alignment methods: language alignment for multimodal language modeling, and spatial alignment for dense prediction. Together, our PE family of models achieves state-of-the-art results on a wide variety of tasks, including zero-shot image and video classification and retrieval; document, image, and video Q&A and spatial tasks such as detection, tracking, and depth estimation. To foster further research, we will release our models, code, and novel dataset of synthetically and human-annotated videos.



Paperid:3570
Authors:Jongseo Lee, Wooil Lee, Gyeong-Moon Park, Seong Tae Kim, Jinwoo Choi
Abstract:
Effective explanations of video action recognition models should disentangle how movements unfold over time from the surrounding spatial context. However, existing methods—based on saliency—produce entangled explanations, making it unclear whether predictions rely on motion or spatial context. Language-based approaches offer structure but often fail to explain motions due to their tacit nature—intuitively understood but difficult to verbalize. To address these challenges, we propose Disentangled Action aNd Context concept-based Explainable (DANCE) video action recognition, a framework that predicts actions through disentangled concept types: motion dynamics, objects, and scenes. We define motion dynamics concepts as human pose sequences. We employ a large language model to automatically extract object and scene concepts. Built on an ante-hoc concept bottleneck design, DANCE enforces prediction through these concepts. Experiments on four datasets—KTH, Penn Action, HAA500, and UCF101—demonstrate that DANCE significantly improves explanation clarity with competitive performance. Through a user study, we validate the superior interpretability of DANCE. Experimental results also show that DANCE is beneficial for model debugging, editing, and failure analysis.



Paperid:3571
Authors:Yuran Wang, Ruihai Wu, Yue Chen, Jiarui Wang, Jiaqi Liang, Ziyu Zhu, Haoran Geng, Jitendra Malik, Pieter Abbeel, Hao Dong
Abstract:
Garment manipulation is a critical challenge due to the diversity in garment categories, geometries, and deformations. Despite this, humans can effortlessly handle garments, thanks to the dexterity of our hands. However, existing research in the field has struggled to replicate this level of dexterity, primarily hindered by the lack of realistic simulations of dexterous garment manipulation. Therefore, we propose DexGarmentLab, the first environment specifically designed for dexterous (especially bimanual) garment manipulation, which features large-scale high-quality 3D assets for 15 task scenarios, and refines simulation techniques tailored for garment modeling to reduce the sim-to-real gap. Previous data collection typically relies on teleoperation or training expert reinforcement learning (RL) policies, which are labor-intensive and inefficient. In this paper, we leverage garment structural correspondence to automatically generate a dataset with diverse trajectories using only a single expert demonstration, significantly reducing manual intervention. However, even extensive demonstrations cannot cover the infinite states of garments, which necessitates the exploration of new algorithms. To improve generalization across diverse garment shapes and deformations, we propose a Hierarchical gArment-manipuLation pOlicy (HALO). It first identifies transferable affordance points to accurately locate the manipulation area, then generates generalizable trajectories to complete the task. Through extensive experiments and detailed analysis of our method and baseline, we demonstrate that HALO consistently outperforms existing methods, successfully generalizing to previously unseen instances even with significant variations in shape and deformation where others fail. Our project page is available at: https://dexgarmentlab-review.github.io/.



Paperid:3572
Authors:Zhuo Cao, Heming Du, Bingqing Zhang, Xin Yu, Xue Li, Sen Wang
Abstract:
Abstract:Existing Moment retrieval (MR) methods focus on Single-Moment Retrieval (SMR). However, one query can correspond to multiple relevant moments in real-world applications. This makes the existing methods insufficient for video temporal grounding. By revisiting the gap between current MR tasks and real-world applications, we introduce a high-quality datasets called QVHighlights Multi-Moment Dataset (QV-M$^2$), along with new evaluation metrics tailored for multi-moment retrieval (MMR). QV-M$^2$ consists of 2,212 annotations covering 5,522 video segments. Building on existing efforts in MMR, we propose a framework called FlashMMR. Specifically, we propose a Multi-moment Post-verification module to refine the moment boundaries. We introduce constrained temporal adjustment and subsequently leverage a verification module to re-evaluate the candidate segments. Through this sophisticated filtering pipeline, low-confidence proposals are pruned, and robust multi-moment alignment is achieved. We retrain and evaluate 6 existing MR methods on QV-M$^2$ and QVHighlights under both SMR and MMR settings. Results show that QV-M$^2$ serves as an effective benchmark for training and evaluating MMR models, while FlashMMR provides a strong baseline. Specifically, on QV-M$^2$, it achieves improvements over prior SOTA method by 3.00\% on G-mAP, 2.70\% on mAP@3+tgt, and 2.56\% on mR@3. The proposed benchmark and method establish a foundation for advancing research in more realistic and challenging video temporal grounding scenarios.



Paperid:3573
Authors:Tongtong Liang, Dan Qiao, Yu-Xiang Wang, Rahul Parhi
Abstract:
We study the implicit bias of flatness / low (loss) curvature and its effects on generalization in two-layer overparameterized ReLU networks with multivariate inputs---a problem well motivated by the minima stability and edge-of-stability phenomena in gradient-descent training. Existing work either requires interpolation or focuses only on univariate inputs. This paper presents new and somewhat surprising theoretical results for multivariate inputs. On two natural settings (1) generalization gap for flat solutions, and (2) mean-squared error (MSE) in nonparametric function estimation by stable minima, we prove upper and lower bounds, which establish that while flatness does imply generalization, the resulting rates of convergence necessarily deteriorate exponentially as the input dimension grows. This gives an exponential separation between the flat solutions vis-à-vis low-norm solutions (i.e., weight decay), which knowingly do not suffer from the curse of dimensionality. In particular, our minimax lower bound construction, based on a novel packing argument with boundary-localized ReLU neurons, reveals how flat solutions can exploit a kind of ''pseudo-sparsity'' where neurons rarely activate, but with high weight magnitudes. This leads to poor performance in high dimensions. We corroborate these theoretical findings with extensive numerical simulations. To the best of our knowledge, our analysis provides the first systematic explanation for why flat minima may fail to generalize in high dimensions.



Paperid:3574
Authors:Haohan Chi, Huan-ang Gao, Ziming Liu, Jianing Liu, Chenyu Liu, Jinwei Li, Kaisen Yang, Yangcheng Yu, Zeda Wang, Wenyi Li, Leichen Wang, Xingtao HU, HAO SUN, Hang Zhao, Hao Zhao
Abstract:
Vision-Language-Action (VLA) models for autonomous driving show promise but falter in unstructured corner case scenarios, largely due to a scarcity of targeted benchmarks. To address this, we introduce Impromptu VLA. Our core contribution is the Impromptu VLA Dataset: over 80,000 meticulously curated video clips, distilled from over 2M source clips sourced from 8 open-source large-scale datasets. This dataset is built upon our novel taxonomy of four challenging unstructured categories and features rich, planning-oriented question-answering annotations and action trajectories. Crucially, experiments demonstrate that VLAs trained with our dataset achieve substantial performance gains on established benchmarks—improving closed-loop NeuroNCAP scores and collision rates, and reaching near state-of-the-art L2 accuracy in open-loop nuScenes trajectory prediction. Furthermore, our Q&A suite serves as an effective diagnostic, revealing clear VLM improvements in perception, prediction, and planning. Our code, data and models are available at https://anonymous.4open.science/r/Impromptu-VLA-54ED/



Paperid:3575
Authors:Yuqian Yuan, Ronghao Dang, long li, Wentong Li, Dian Jiao, Xin Li, Deli Zhao, Fan Wang, Wenqiao Zhang, Jun Xiao, Yueting Zhuang
Abstract:
The emergence of multimodal large language models (MLLMs) has driven breakthroughs in egocentric vision applications. These applications necessitate persistent, context-aware understanding of objects, as users interact with tools in dynamic and cluttered environments. However, existing embodied benchmarks primarily focus on static scene exploration, emphasizing object's appearance and spatial attributes while neglecting the assessment of dynamic changes arising from users' interactions.capabilities in object-level spatiotemporal reasoning required for real-world interactions.To address this gap, we introduce EOC-Bench, an innovative benchmark designed to systematically evaluate object-centric embodied cognition in dynamic egocentric scenarios.Specially, EOC-Bench features 3,277 meticulously annotated QA pairs categorized into three temporal categories: Past, Present, and Future, covering 11 fine-grained evaluation dimensions and 3 visual object referencing types.To ensure thorough assessment, we develop a mixed-format human-in-the-loop annotation frameworkBased on EOC-Bench, we conduct comprehensive evaluations of various proprietary, open-source, and object-level MLLMs. EOC-Bench serves as a crucial tool for advancing the embodied object cognitive capabilities of MLLMs, establishing a robust foundation for developing reliable core models for embodied systems.All data and evaluation codes will be made publicly available.



Paperid:3576
Authors:Yuqian Cheng, Zhuo Chen, Qian Lin
Abstract:
The physics-informed neural networks (PINNs) are widely applied in solving differential equations. However, few studies have discussed their consistency. In this paper, we consider the consistency of PINNs when applied to second-order elliptic equations with Dirichlet boundary conditions. We first provide the necessary and sufficient condition for the consistency of the physics-informed kernel gradient flow algorithm, and then as a direct corollary, when the neural network is sufficiently wide, we obtain a necessary and sufficient condition for the consistency of PINNs based on the neural tangent kernel theory. We also estimate the non-asymptotic loss bounds of physics-informed kernel gradient flow and PINN under suitable stronger assumptions. Finally, these results inspires us to construct a notable pathological example where the PINN method is inconsistent.



Paperid:3577
Authors:Bosong Huang, Ming Jin, Yuxuan Liang, Johan Barthelemy, Debo Cheng, Qingsong Wen, Chenghao Liu, Shirui Pan
Abstract:
Explaining time series classification models is crucial, particularly in high-stakes applications such as healthcare and finance, where transparency and trust play a critical role. Although numerous time series classification methods have identified key subsequences, known as shapelets, as core features for achieving state-of-the-art performance and validating their pivotal role in classification outcomes, existing post-hoc time series explanation (PHTSE) methods primarily focus on timestep-level feature attribution. These explanation methods overlook the fundamental prior that classification outcomes are predominantly driven by key shapelets. To bridge this gap, we present ShapeX, an innovative framework that segments time series into meaningful shapelet-driven segments and employs Shapley values to assess their saliency. At the core of ShapeX lies the Shapelet Describe-and-Detect (SDD) framework, which effectively learns a diverse set of shapelets essential for classification. We further demonstrate that ShapeX produces explanations which reveal causal relationships instead of just correlations, owing to the atomicity properties of shapelets. Experimental results on both synthetic and real-world datasets demonstrate that ShapeX outperforms existing methods in identifying the most relevant subsequences, enhancing both the precision and causal fidelity of time series explanations.



Paperid:3578
Authors:ChengAo Shen, Wenchao Yu, Ziming Zhao, Dongjin Song, Wei Cheng, Haifeng Chen, Jingchao Ni
Abstract:
Time series, typically represented as numerical sequences, can also be transformed into images and texts, offering multi-modal views (MMVs) of the same underlying signal. These MMVs can reveal complementary patterns and enable the use of powerful pre-trained large models, such as large vision models (LVMs), for long-term time series forecasting (LTSF). However, as we identified in this work, applying LVMs to LTSF poses an inductive bias towards "forecasting periods". To harness this bias, we propose DMMV, a novel decomposition-based multi-modal view framework that leverages trend-seasonal decomposition and a novel backcast-residual based adaptive decomposition to integrate MMVs for LTSF. Comparative evaluations against 14 state-of-the-art (SOTA) models across diverse datasets show that DMMV outperforms single-view and existing multi-modal baselines, achieving the best mean squared error (MSE) on 6 out of 8 benchmark datasets. The code for reproducing our work is available in the supplementary material.



Paperid:3579
Authors:Masahiro Kaneko, Timothy Baldwin
Abstract:
Abstract:Adversarial attacks by malicious users that threaten the safety of large language models (LLMs) can be viewed as attempts to infer a \emph{target property} $T$ that is unknown when an instruction is issued, and becomes knowable only after the model's reply is observed. Examples of target properties $T$ include the binary flag that triggers an LLM's harmful response or rejection, and the degree to which information deleted by unlearning can be restored, both elicited via adversarial instructions. The LLM reveals an \emph{observable signal} $Z$ that potentially leaks hints for attacking through a response containing answer tokens, thinking process tokens, or logits.Yet the scale of information leaked remains anecdotal, leaving auditors without principled guidance and defenders blind to the transparency--risk trade-off.We fill this gap with an information-theoretic framework\footnote{We plan to release the code submitted as supplementary material after our paper is accepted.} that computes how much information can be safely disclosed, and enables auditors to gauge how close their methods come to the fundamental limit.Treating the mutual information $I(Z;T)$ between the observation $Z$ and the target property $T$ as the leaked bits per query, we show that achieving error $\varepsilon$ requires at least $\log(1/\varepsilon)/I(Z;T)$ queries, scaling linearly with the inverse leak rate and only logarithmically with the desired accuracy.Thus, even a modest increase in disclosure collapses the attack cost from quadratic to logarithmic in terms of the desired accuracy.Experiments on seven LLMs across system-prompt leakage, jailbreak, and relearning attacks corroborate the theory: exposing answer tokens alone requires about a thousand queries, adding logits cuts this to about a hundred, and revealing the full thinking process trims it to a few dozen.Our results provide the first principled yardstick for balancing transparency and security when deploying LLMs.



Paperid:3580
Authors:Vighnesh Subramaniam, David Mayo, Colin Conwell, Tomaso Poggio, Boris Katz, Brian Cheung, Andrei Barbu
Abstract:
We demonstrate that architectures which traditionally are considered to be ill-suited for a task can be trained using inductive biases from another architecture. We call a network untrainable when they overfit, underfit, or converge to poor results even when tuning their hyperparameters. For example, fully connected networks overfit on object recognition while deep convolutional networks without residual connections underfit. The traditional answer is to change the architecture to impose some inductive bias, although the nature of that bias is unknown. We introduce guidance, where a guide network steers a target network using a neural distance function. The target minimizes its task loss plus a layerwise representational similarity against the frozen guide. If the guide is trained, this transfers over the architectural prior and knowledge of the guide to the target. If the guide is untrained, this transfers over only part of the architectural prior of the guide. We show that guidance prevents FCN overfitting on ImageNet, narrows the vanilla RNN–Transformer gap, boosts plain CNNs toward ResNet accuracy, and aids Transformers on RNN-favored tasks. We further identify that guidance-driven initialization alone can mitigate FCN overfitting. Our method provides a mathematical tool to investigate priors and architectures, and in the long term, could automate architecture design.



Paperid:3581
Authors:Kanghao Chen, Zixin Zhang, Guoqiang Liang, Lutao Jiang, Zeyu Wang, Yingcong Chen
Abstract:
Recent advancements in low-light video enhancement (LLVE) have increasingly leveraged both RGB and event cameras to improve video quality under challenging conditions. However, existing approaches share two key drawbacks. First, they are tuned for steady low-light scenes, so their performance drops when illumination varies. Second, they assume every sensing modality is always available, while real systems may lose or corrupt one of them. These limitations make the methods brittle in dynamic, real-world settings. In this paper, we propose EVDiffuser, a novel framework for consistent LLVE that integrates RGB and event data through a modality-adaptive diffusion pipeline. By harnessing the powerful priors of video diffusion models, EVDiffuser enables consistent video enhancement and generalization to diverse scenarios under varying illumination, where RGB or events may even be absent. Specifically, we first design a modality-agnostic conditioning mechanism based on a diffusion pipeline by treating the two modalities as optional conditions, which is fine-tuned using augmented and integrated datasets. Furthermore, we introduce a modality-adaptive guidance rescaling that dynamically adjusts the contribution of each modality according to sensor-specific characteristics. Additionally, we establish a benchmark that accounts for varying illumination and diverse real-world scenarios, facilitating future research on consistent event-guided LLVE. Our experiments demonstrate state-of-the-art performance across challenging scenarios (i.e., varying illumination) and sensor-based settings (e.g., event-only, RGB-only), highlighting the generalization of our framework.



Paperid:3582
Authors:Zihan Zheng, Zerui Cheng, Zeyu Shen, Shang Zhou, Kaiyuan Liu, Hansen He, Dongruixuan Li, Stanley Wei, Hangyi Hao, Jianzhu Yao, Peiyao Sheng, Zixuan Wang, Wenhao Chai, Aleksandra Korolova, Peter Henderson, Sanjeev Arora, Pramod Viswanath, Jingbo Shang, Saining Xie
Abstract:
Recent reports claim that large language models (LLMs) now outperform elite humans in competitive programming. Drawing on knowledge from a group of medalists in international algorithmic contests, we revisit this claim, examining how LLMs differ from human experts and where limitations still remain. We introduce LiveCodeBench Pro, a benchmark composed of problems from Codeforces, ICPC, and IOI that are continuously updated to reduce the likelihood of data contamination. A team of Olympiad medalists annotates every problem for algorithmic categories and conducts a line-by-line analysis of failed model-generated submissions. Using this new data and benchmark, we find that frontier models still have significant limitations: without external tools, the best model achieves only 53\% pass@1 on medium-difficulty problems and 0\% on hard problems, domains where expert humans still excel. We also find that LLMs succeed at implementation-heavy problems but struggle with nuanced algorithmic reasoning and complex case analysis, often generating confidently incorrect justifications. High performance appears largely driven by implementation precision and tool augmentation, not superior reasoning. LiveCodeBench Pro thus highlights the significant gap to human grandmaster levels, while offering fine-grained diagnostics to steer future improvements in code-centric LLM reasoning.



Paperid:3583
Authors:Yu-Hu Yan, Peng Zhao, Zhi-Hua Zhou
Abstract:
In this work, we study offline convex optimization with smooth objectives, where the classical Nesterov's Accelerated Gradient (NAG) method achieves the optimal accelerated convergence. Extensive research has aimed to understand NAG from various perspectives. A recent line of work approaches this from the viewpoint of online learning and online-to-batch conversion, emphasizing the role ofoptimistic algorithmsfor acceleration. In this work, we contribute to this perspective by proposing noveloptimistic online-to-batch conversionsthat incorporate optimism theoretically into the analysis, thereby significantly simplifying the algorithm design while preserving the optimal convergence rates. Specifically, we demonstrate the effectiveness of our conversions through the following results:(i)when combined with simple online gradient descent, our optimistic conversion achieves the optimal accelerated convergence;(ii)our conversion also applies to strongly convex objectives, achieving the optimal accelerated convergence rate;(iii)our optimistic conversion can achieve universality to smoothness --- applicable to both smooth and non-smooth objectives without requiring knowledge of the smoothness parameter --- and remains efficient as non-universal methods by using only one gradient query in each iteration. Finally, we explain the effectiveness of our optimistic online-to-batch conversions by a precise correspondence with NAG.



Paperid:3584
Authors:Yingxuan Yang, Huacan Chai, Shuai Shao, Yuanyi Song, Siyuan Qi, Renting Rui, Weinan Zhang
Abstract:
The rapid advancement of Large Language Models (LLMs) has catalyzed the development of multi-agent systems, where multiple LLM-based agents collaborate to solve complex tasks. However, existing systems predominantly rely on centralized coordination, which introduces scalability bottlenecks, limits adaptability, and creates single points of failure. Additionally, concerns over privacy and proprietary knowledge sharing hinder cross-organizational collaboration, leading to siloed expertise. To address these challenges, we propose AgentNet, a decentralized, Retrieval-Augmented Generation (RAG)-based framework that enables LLM-based agents to autonomously evolve their capabilities and collaborate efficiently in a Directed Acyclic Graph (DAG)-structured network. Unlike traditional multi-agent systems that depend on static role assignments or centralized control, AgentNet allows agents to specialize dynamically, adjust their connectivity, and route tasks without relying on predefined workflows.AgentNet’s core design is built upon several key innovations: (1) Fully Decentralized Paradigm: Removing the central orchestrator, allowing agents to coordinate and specialize autonomously, fostering fault tolerance and emergent collective intelligence. (2) Dynamically Evolving Graph Topology: Real-time adaptation of agent connections based on task demands, ensuring scalability and resilience.(3) Adaptive Learning for Expertise Refinement: A retrieval-based memory system that enables agents to continuously update and refine their specialized skills.By eliminating centralized control, AgentNet enhances fault tolerance, promotes scalable specialization, and enables privacy-preserving collaboration across organizations. Through decentralized coordination and minimal data exchange, agents can leverage diverse knowledge sources while safeguarding sensitive information. Experimental results demonstrate that AgentNet outperforms traditional centralized multi-agent systems, significantly improving efficiency, adaptability, and scalability in dynamic environments, making it a promising foundation for next-generation autonomous, privacy-respecting multi-agent ecosystems.



Paperid:3585
Authors:Wangkai Li, Rui Sun, Huayu Mai, Tianzhu Zhang
Abstract:
Semantic segmentation suffers from significant performance degradation when the trained network is applied to a different domain. To address this issue, unsupervised domain adaptation (UDA) has been extensively studied. Despite the effectiveness of selftraining techniques in UDA, they still overlook the explicit modelingof domain-shared feature extraction.In this paper, we propose DiDA, an unsupervised domain bridging approach for semantic segmentation. DiDA consists of two key modules: (1) Degradation-based Intermediate Domain Construction, which creates continuous intermediate domains through simple image degradation operations to encourage learning domain-invariant features as domain differences gradually diminish; (2) Semantic Shift Compensation, which leverages a diffusion encoder to disentangle and compensate for semantic shift information with degraded time-steps, preserving discriminative representations in the intermediate domains.As a plug-and-play solution, DiDA supports various degradation operations and seamlessly integrates with existing UDA methods. Extensive experiments on multiple domain adaptive semantic segmentation benchmarks demonstrate that DiDA consistently achieves significant performance improvements across all settings.



Paperid:3586
Authors:Zhuohao Yu, Xingru Jiang, Weizheng Gu, Yidong Wang, Qingsong Wen, Shikun Zhang, Wei Ye
Abstract:
Watermarking LLM-generated text is critical for content attribution and misinformation prevention. However, existing methods compromise text quality, require white-box model access and logit manipulation —limitations that exclude API-based models and multilingual / domain-specific use cases. We proposeSAEMark, a multi-bit watermarking method that embeds personalized signatures without altering logits. SAEMark utilizesSparse Autoencoder (SAE)to extract semantic features of generated texts and selects outputs matching watermark key-derived feature distributions via rejection sampling. This approach operates post-generation, preserving text quality while naturally generalizing to multilingual scenarios including programming languages.Experiments across 4 datasets shows SAEMark's consistent accuracy, highest text quality among 9 baselines. SAEMark establishes a new paradigm: secure, personalizable watermarks that work out-of-the-box for all languages and closed-source LLMs for ethical and trustworthy AI systems.



Paperid:3587
Authors:WANG Yun, Qiaole Dong, Yongjian Zhang, Tin Lun Lam, Yanwei Fu, Dapeng Wu, Junjie Hu
Abstract:
Temporally consistent depth estimation from stereo video is critical for real-world applications such as augmented reality, where inconsistent depth estimation disrupts the immersion of users.Despite its importance, this task remains challenging due to the difficulty in modeling long-term temporal consistency in a computationally efficient manner.Previous methods attempt to address this by aggregating spatio-temporal information but face a fundamental trade-off: limited temporal modeling provides only modest gains, whereas capturing long-range dependencies significantly increases computational cost.To address this limitation, we introduce a memory buffer for modeling long-range spatio-temporal consistency while achieving efficient dynamic stereo matching.Inspired by the two-stage decision-making process in humans, we propose a Pick-and-Play Memory (PPM) construction module for dynamic Stereo matching, dubbed as PPMStereo. PPM consists of a pick process that identifies the most relevant frames and a play process that weights the selected frames adaptively for spatio-temporal aggregation.This two-stage collaborative process maintains a compact yet highly informative memory buffer while achieving temporally consistent information aggregation.Extensive experiments validate the effectiveness of PPMStereo, demonstrating state-of-the-art performance in both accuracy and temporal consistency.



Paperid:3588
Authors:Ye Sun, Hao Zhang, Henghui Ding, Tiehua Zhang, Xingjun Ma, Yu-Gang Jiang
Abstract:
Achieving fine-grained spatio-temporal understanding in videos remains a major challenge for current Video Large Multimodal Models (Video LMMs). Addressing this challenge requires mastering two core capabilities: video referring understanding, which captures the semantics of video regions, and video grounding, which segments object regions based on natural language descriptions.However, most existing approaches tackle these tasks in isolation, limiting progress toward unified, referentially grounded video interaction. We identify a key bottleneck in the lack of high-quality, unified video instruction data and a comprehensive benchmark for evaluating referentially grounded video chat.To address these challenges, we contribute in three core aspects: dataset, model, and benchmark.First, we introduce SAMA-239K, a large-scale dataset comprising 15K videos specifically curated to enable joint learning of video referring understanding, grounding, and multi-turn video chat.Second, we propose the SAMA model, which incorporates a versatile spatio-temporal context aggregator and a Segment Anything Model to jointly enhance fine-grained video comprehension and precise grounding capabilities.Finally, we establish SAMA-Bench, a meticulously designed benchmark consisting of 5,067 questions from 522 videos, to comprehensively evaluate the integrated capabilities of Video LMMs in multi-turn, spatio-temporal referring understanding and grounded dialogue.Extensive experiments and benchmarking results show that SAMA not only achieves strong performance on SAMA-Bench but also sets a new state-of-the-art on general grounding benchmarks, while maintaining highly competitive performance on standard visual understanding benchmarks.



Paperid:3589
Authors:Soumya Suvra Ghosal, Souradip Chakraborty, Avinash Reddy, Yifu Lu, Mengdi Wang, Dinesh Manocha, Furong Huang, Mohammad Ghavamzadeh, Amrit Singh Bedi
Abstract:
Recent trends in test-time scaling for reasoning models (e.g., OpenAI o1, DeepSeek R1) have led to a popular belief that extending thinking traces using prompts like “Wait” or “Let me rethink” can improve performance. This raises a natural question: Does thinking more at test-time truly lead to better reasoning? To answer this question, we perform a detailed empirical study across models and benchmarks, which reveals a consistent pattern of initial performance improvements from additional thinking followed by a decline, due to "overthinking". To understand this non-monotonic trend, we consider a simple probabilistic model, which reveals that additional thinking increases output variance—creating an illusion of improved reasoning while ultimately undermining precision. Thus, observed gains from "more thinking" are not true indicators of improved reasoning, but artifacts stemming from the connection between model uncertainty and evaluation metric. This suggests that test-time scaling through extended thinking is not an effective way to utilize the inference thinking budget. Recognizing these limitations, we introduce an alternative test-time scaling approach, parallel thinking, inspired by Best-of-N sampling. Our method generates multiple independent reasoning paths within the same inference budget and selects the most consistent response via majority vote, achieving up to 20% higher accuracy compared to extended thinking. This provides a simple yet effective mechanism for test-time scaling of reasoning models.



Paperid:3590
Authors:Dandan Liang, Jianing Zhang, Evan Chen, Zhe Li, Rui Li, Haibo Yang
Abstract:
Abstract:Split Federated Learning (SFL) enables scalable training on edge devices by combining the parallelism of Federated Learning (FL) with the computational offloading of Split Learning (SL). Despite its great success, SFL suffers significantly from the well-known straggler issue in distributed learning systems. This problem is exacerbated by the dependency between Split Server and clients: the Split Server side model update relies on receiving activations from clients. Such synchronization requirement introduces significant time latency, making straggler a critical bottleneck to the scalability and efficiency of the system. To mitigate this problem, we propose *MU-SplitFed*, a straggler-resilient SFL algorithm that decouples training progress from straggler delays via a simple yet effective unbalanced update mechanism. By enabling the server to perform $\tau$ local updates per client round, *MU-SplitFed* achieves convergence rate $\mathcal{O}(\sqrt{d/(\tau T)})$, showing a linear reduction in communication round by a factor of $\tau$. Experiments demonstrate that *MU-SplitFed* consistently outperforms baseline methods with the presence of stragglers and effectively mitigates their impact through adaptive tuning of $\tau$.



Paperid:3591
Authors:Vittorio Erba, Emanuele Troiani, Lenka Zdeborová, Florent Krzakala
Abstract:
Abstract:We study the high-dimensional asymptotics of empirical risk minimization (ERM) in over-parametrized two-layer neural networks with quadratic activations trained on synthetic data. We derive sharp asymptotics for both training and test errors by mapping the $\ell_2$-regularized learning problem to a convex matrix sensing task with nuclear norm penalization. This reveals that capacity control in such networks emerges from a low-rank structure in the learned feature maps. Our results characterize the global minima of the loss and yield precise generalization thresholds, showing how the width of the target function governs learnability. This analysis bridges and extends ideas from spin-glass methods, matrix factorization, and convex optimization and emphasizes the deep link between low-rank matrix sensing and learning in quadratic neural networks.



Paperid:3592
Authors:Hao Chen, Guanxi Lu, Yasuyuki Okoshi, Zhiwen Mo, Masato Motomura, Hongxiang Fan
Abstract:
Abstract:Test-time scaling (TTS) has proven effective in enhancing the reasoning capabilities of large language models (LLMs). Verification plays a key role in TTS, simultaneously influencing (1) reasoning performance and (2) compute efficiency, due to the quality and computational cost of verification. In this work, we challenge the conventional paradigms of verification, and make the first attempt toward systematically investigating the impact of verification granularity—that is, how frequently the verifier is invoked during generation, beyond verifying only the final output or individual generation steps.To this end, we introduce Variable Granularity Search (VG-Search), a unified algorithm that generalizes beam search and Best-of-N sampling via a tunable granularity parameter $g$. Extensive experiments with VG-Search under varying compute budgets, generator-verifier configurations, and task attributes reveal that dynamically selecting $g$ can improve the compute efficiency and scaling behavior. Building on these findings, we propose adaptive VG-Search strategies that achieve accuracy gains of up to 3.1\% over Beam Search and 3.6\% over Best-of-N, while reducing FLOPs by over 52\%. We will open-source the code to support future research.



Paperid:3593
Authors:Abhishek Roy, Geelon So, Yian Ma
Abstract:
Abstract:In multi-objective optimization, a single decision vector must balance the trade-offs between many objectives. Solutions achieving an optimal trade-off are said to be Pareto optimal---these are decision vectors for which improving any one objective must come at a cost to another. Many decisions can be Pareto optimal, so when the decision maker can choose only one, this raises questions of which solution to pick and how. We formulate the problem as *Pareto-constrained optimization*, where the goal is to optimize a preference function constrained to a set of Pareto optimal solutions. This constrained optimization problem poses significant challenges: not only is the constraint set defined implicitly, but it is also generally non-convex and non-smooth, even when the objectives are strongly convex. We propose a reformulation of the problem where the constraint set is redefined in terms of an appropriately defined manifold. This reformulation allows us to introduce a clearer and more accurate notion of optimality and stationarity, improving upon existing definitions in the literature. We provide an algorithm with a last-iterate convergence rate of $O( K^{-1/2})$ to this notion of stationarity when the preference function is Lipschitz smooth and when the objective functions are strongly convex and Lipschitz smooth. Moreover, motivated by practical applications like Reinforcement Learning with Human Feedback (RLHF), we extend this algorithm to tackle the case where the preference function can only be queried via dueling feedback.



Paperid:3594
Authors:Shakir Yousefi, Andreas Plesner, Till Aczel, Roger Wattenhofer
Abstract:
Abstract:Modern neural networks exhibit state-of-the-art performance on many existing benchmarks, but their high computational requirements and energy usage cause researchers to explore more efficient solutions for real-world deployment. Logic gate networks (LGNs) learns a large network of logic gates for efficient image classification. However, learning a network that can solve a simple problem like CIFAR-10 can take days to weeks to train. Even then, almost half of the network remains unused, causing a \emph{discretization gap}. This discretization gap hinders real-world deployment of LGNs, as the performance drop between training and inference negatively impacts accuracy. We inject Gumbel noise with a straight-through estimator during training to significantly speed up training, improve neuron utilization, and decrease the discretization gap. We theoretically show that this results from implicit Hessian regularization, which improves the convergence properties of LGNs. We train networks $4.5 \times$ faster in wall-clock time, reduce the discretization gap by 98\%, and reduce the number of unused gates by 100\%.



Paperid:3595
Authors:Utkarsh Utkarsh, Pengfei Cai, Alan Edelman, Rafael Gomez-Bombarelli, Christopher Rackauckas
Abstract:
Deep generative models have recently been applied to physical systems governed by partial differential equations (PDEs), offering scalable simulation and uncertainty-aware inference. However, enforcing physical constraints, such as conservation laws (linear and nonlinear) and physical consistencies, remains challenging. Existing methods often rely on soft penalties or architectural biases that fail to guarantee hard constraints. In this work, we propose Physics-Constrained Flow Matching (PCFM), a zero-shot inference framework that enforces arbitrary nonlinear constraints in pretrained flow-based generative models. PCFM continuously guides the sampling process through physics-based corrections applied to intermediate solution states, while remaining aligned with the learned flow and satisfying physical constraints. Empirically, PCFM outperforms both unconstrained and constrained baselines on a range of PDEs, including those with shocks, discontinuities, and sharp features, while ensuring exact constraint satisfaction at the final solution. Our method provides a general framework for enforcing hard constraints in both scientific and general-purpose generative models, especially in applications where constraint satisfaction is essential.



Paperid:3596
Authors:Joséphine Raugel, Jérémy Rapin, Stéphane d'Ascoli, Valentin Wyart, Jean-Remi King
Abstract:
Recent studies suggest that the representations learned by large language models (LLMs) are partially aligned to those of the human brain. However, whether this representational alignment arises from a similar sequence of computations remains elusive. In this study, we explore this question by examining temporally-resolved brain signals of participants listening to 10 hours of an audiobook. We study these neural dynamics jointly with a benchmark encompassing 17 LLMs varying in size and architecture type. Our analyses reveal that LLMs and the brain generate representations in a similar order: specifically, activations in the initial layers of LLMs tend to best align with early brain responses, while the deeper layers of LLMs tend to best align with later brain responses. This brain-LLM alignment is consistent across transformers and recurrent architectures. However, its emergence depends on both model size and context length. Overall, the alignment between LLMs and the brain provides novel elements supporting a partial convergence between biological and artificial neural networks.



Paperid:3597
Authors:Valérie Costa, Thomas Fel, Ekdeep S Lubana, Bahareh Tolooshams, Demba Ba
Abstract:
Motivated by the hypothesis that neural network representations encode abstract, interpretable features as linearly accessible and approximately orthogonal directions, sparse autoencoders (SAEs) have become a popular tool in the interpretability literature. However, recent work has revealed properties of model representations that fall outside this hypothesis, showing evidence of hierarchical, nonlinear, and multi-dimensional features. This mismatch raises a fundamental question: can standard SAEs recover meaningful features when the latent geometry of representations deviates from linear or quasi-orthogonal assumptions? If not, what would it mean to design an architecture whose inductive bias aligns with the actual structure observed in modern representations? To explore this, we adopt a construction-based approach and introduce MP-SAE as a direct implementation of a revised structural assumption: conditional orthogonality, which has emerged from recent empirical studies. By reinterpreting the Matching Pursuit (MP) algorithm from sparse coding through this new lens, we develop an SAE that unrolls its encoder into a sequence of residual-guided steps, enabling it to capture hierarchical and nonlinearly accessible features. Comparing this architecture with existing SAEs across synthetic and natural data settings, we find that (i) hierarchical concepts induce conditionally orthogonal features, which standard SAEs fail to capture accurately, and (ii) the nonlinear encoding mechanism of MP-SAE recovers highly meaningful features. This allows us to uncover shared structure in seemingly divergent representation spaces, such as those found in vision-language models, and challenges the assumption that useful features are solely linearly accessible. We also show that the sequential encoder of MP-SAE provides an additional benefit: adaptive sparsity during inference, which may be of independent interest. Ultimately, our work reflects a broader principle: interpretability should begin with the observed structure of neural representations, allowing appropriate methods to emerge from inductive biases that are aligned with that structure.



Paperid:3598
Authors:Chen Zhao, Ci En, Yunzhe Xu, Tiehan Fan, Shanyan Guan, Yanhao Ge, Jian Yang, Ying Tai
Abstract:
Ultra-high-resolution (UHR) text-to-image (T2I) generation has seen notable progress. However, two key challenges remain : 1) the absence of a large-scale high-quality UHR T2I dataset, and (2) the neglect of tailored training strategies for fine-grained detail synthesis in UHR scenarios. To tackle the first challenge, we introduce \textbf{UltraHR-100K}, a high-quality dataset of 100K UHR images with rich captions, offering diverse content and strong visual fidelity. Each image exceeds 3K resolution and is rigorously curated based on detail richness, content complexity, and aesthetic quality. To tackle the second challenge, we propose a frequency-aware post-training method that enhances fine-detail generation in T2I diffusion models. Specifically, we design (i) \textit{Detail-Oriented Timestep Sampling (DOTS)} to focus learning on detail-critical denoising steps, and (ii) \textit{Soft-Weighting Frequency Regularization (SWFR)}, which leverages Discrete Fourier Transform (DFT) to softly constrain frequency components, encouraging high-frequency detail preservation. Extensive experiments on our proposed UltraHR-eval4K benchmarks demonstrate that our approach significantly improves the fine-grained detail quality and overall fidelity of UHR image generation.



Paperid:3599
Authors:Ilias Diakonikolas, Giannis Iakovidis, Daniel Kane, Lisheng Ren
Abstract:
Abstract:We study the complexity of learning real-valued Multi-Index Models (MIMs) under the Gaussian distribution. A $K$-MIM is a function $f:\mathbb{R}^d\to \mathbb{R}$ that depends only on the projection of its input onto a $K$-dimensional subspace. We give a general algorithm for PAC learning a broad class of MIMs with respect to the square loss, even in the presence of adversarial label noise. Moreover, we establish a nearly matching Statistical Query (SQ) lower bound, providing evidence that the complexity of our algorithm is qualitatively optimal as a function of the dimension. Specifically, we consider the class of bounded variation MIMs with the property that degree at most $m$ distinguishing moments exist with respect to projections onto any subspace. In the presence of adversarial label noise, the complexity of our learning algorithm is $d^{O(m)}2^{\mathrm{poly}(K/\epsilon)}$. For the realizable and independent noise settings, our algorithm incurs complexity $d^{O(m)}2^{\mathrm{poly}(K)}(1/\epsilon)^{O(K)}$. To complement our upper bound, we show that if for some subspace degree-$m$ distinguishing moments do not exist, then any SQ learner for the corresponding class of MIMs requires complexity $d^{\Omega(m)}$. As an application, we give the first efficient learner for the class of positive-homogeneous $L$-Lipschitz $K$-MIMs. The resulting algorithm has complexity $\mathrm{poly}(d) 2^{\mathrm{poly}(KL/\epsilon)}$. This gives a new PAC learning algorithm for Lipschitz homogeneous ReLU networks with complexity independent of the network size, removing the exponential dependence incurred in prior work.



Paperid:3600
Authors:Kichang Yang, Seonjun Kim, Minjae Kim, Nairan Zhang, Chi Zhang, Youngki Lee
Abstract:
Edge deployment of large vision-language models (VLMs) increasingly relies on flash-based weight offloading, where activation sparsification is used to reduce I/O overhead. However, conventional sparsification techniques are model-centric—they select neurons based solely on activation magnitude, ignoring flash storage behavior. In this paper, we propose \textsc{Neuron Chunking}, a structured sparsification method that jointly considers activation importance and flash I/O efficiency. We introduce a contiguity-based latency model that accurately predicts the cost of arbitrary access patterns, and a multi-scale chunk selection algorithm that favors contiguous chunks with a high importance-to-latency ratio. Our approach significantly improves I/O efficiency, achieving up to 4.58× and 5.85× speedups over existing sparsification methods on Jetson Orin Nano and Jetson AGX Orin, respectively.



Paperid:3601
Authors:Yimu Wang, Mozhgan Nasr Azadani, Sean Sedwards, Krzysztof Czarnecki
Abstract:
Improving the visual understanding ability of vision-language models (VLMs) is crucial for enhancing their performance across various tasks. While using multiple pretrained visual experts has shown great promise, it often incurs significant computational costs during training and inference. To address this challenge, we propose HAWAII, a novel framework that distills knowledge from multiple visual experts into a single vision encoder, enabling it to inherit the complementary strengths of several experts with minimal computational overhead. To mitigate conflicts among different teachers and switch between different teacher-specific knowledge, instead of using a fixed set of adapters for multiple teachers, we propose to use teacher-specific Low-Rank Adaptation (LoRA) adapters with a corresponding router. Each adapter is aligned with a specific teacher, avoiding noisy guidance during distillation. To enable efficient knowledge distillation, we propose fine-grained and coarse-grained distillation. At the fine-grained level, token importance scores are employed to emphasize the most informative tokens from each teacher adaptively. At the coarse-grained level, we summarize the knowledge from multiple teachers and transfer it to the student using a set of general-knowledge LoRA adapters with a router. Extensive experiments on various vision-language tasks demonstrate the superiority of HAWAII, compared to the popular open-source VLMs.



Paperid:3602
Authors:Laura Kopf, Nils Feldhus, Kirill Bykov, Philine L Bommer, Anna Hedström, Marina Höhne, Oliver Eberle
Abstract:
Automated interpretability research aims to identify concepts encoded in neural network features to enhance human understanding of model behavior. Current feature description methods face two critical challenges: limited robustness and the flawed assumption that each neuron encodes only a single concept (monosemanticity), despite growing evidence that neurons are often polysemantic. This assumption restricts the expressiveness of feature descriptions and limits their ability to capture the full range of behaviors encoded in model internals. To address this, we introduce Polysemantic FeatuRe Identification and Scoring Method (PRISM), a novel framework that captures the inherent complexity of neural network features. Unlike prior approaches that assign a single description per feature, PRISM provides more nuanced descriptions for both polysemantic and monosemantic features. Through extensive benchmarking against existing methods, we demonstrate that our approach produces more accurate and faithful feature descriptions, improving both overall description quality (via a description score) and the ability to capture distinct concepts when polysemanticity is present (via a polysemanticity score).



Paperid:3603
Authors:David Heineman, Valentin Hofmann, Ian Magnusson, Yuling Gu, Noah Smith, Hanna Hajishirzi, Kyle Lo, Jesse Dodge
Abstract:
Developing large language models is expensive and often involves making decisions with small experiments, typically by evaluating on large, multi-task evaluation suites. In this work, we analyze specific properties which make a benchmark more reliable and useful for such decisions, and interventions to design higher-quality evaluation benchmarks. We introduce two key metrics that show differences in current benchmarks: signal, a benchmark’s ability to separate better models from worse models, and noise, a benchmark’s sensitivity to random variability between training steps. We demonstrate that benchmarks with a better signal-to-noise ratio are more reliable when making decisions at small scale, and those with less noise have lower scaling law prediction error. These results suggest that improving signal or noise will lead to more useful benchmarks, so we introduce four interventions designed to directly affect signal or noise. For example, we propose that switching to a metric that has better signal and noise (e.g., perplexity rather than accuracy) leads to better reliability and scaling law error. We also find that filtering noisy benchmarks such that they have better signal-to-noise ratio leads to more reliable evaluations. We also find that averaging the output of a model's checkpoints to reduce noise leads to consistent improvements. We conclude by recommending that those creating new benchmarks, or selecting which existing benchmarks to use, aim for high signal and low noise. We use 30 benchmarks for these experiments, and 465 open-weight language models from 60M to 32B parameters, resulting in a new, publicly available dataset of 50K evaluation benchmark results, totaling 200M instances.



Paperid:3604
Authors:Yiheng Du, Aditi Krishnapriyan
Abstract:
Computationally resolving the spatiotemporal evolution of fluid flows is central for many engineering applications. A key challenge arises from turbulence modeling, due to the complex interactions across a wide range of scales, especially in fluid domains with complicated geometries. Fully resolving large-scale turbulence through direct numerical simulation (DNS) is computationally prohibitive, motivating machine learning alternatives. In this work, we propose EddyFormer, a Transformer-based spectral-element model designed for large-scale turbulence simulations. EddyFormer decomposes the prediction into grid-scale and subgrid-scale fields, using attention mechanisms for global dependencies and convolutions for local dynamics. We train EddyFormer on isotropic turbulence, demonstrating that it achieves accuracy of DNS at 256^3 resolution while providing a 30x speedup over DNS, significantly outperforming existing ML baselines in both accuracy and efficiency. Benchmarks on diverse turbulent flows further validate its effectiveness, resolving flow where other models fail and generalizing well to unseen domains.



Paperid:3605
Authors:Weixin Chen, Han Zhao
Abstract:
Neural Probabilistic Circuits (NPCs), a new class of concept bottleneck models, comprise an attribute recognition model and a probabilistic circuit for reasoning. By integrating the outputs from these two modules, NPCs produce compositional and interpretable predictions. While offering enhanced interpretability and high performance on downstream tasks, the neural-network-based attribute recognition model remains a black box. This vulnerability allows adversarial attacks to manipulate attribute predictions by introducing carefully crafted subtle perturbations to input images, potentially compromising the final predictions. In this paper, we theoretically analyze the adversarial robustness of NPC and demonstrate that it only depends on the robustness of the attribute recognition model and is independent of the robustness of the probabilistic circuit. Moreover, we propose RNPC, the first robust neural probabilistic circuit against adversarial attacks on the recognition module. RNPC introduces a novel class-wise integration for inference, ensuring a robust combination of outputs from the two modules. Our theoretical analysis demonstrates that RNPC exhibits provably improved adversarial robustness compared to NPC. Empirical results on image classification tasks show that RNPC achieves superior adversarial robustness compared to existing concept bottleneck models while maintaining high accuracy on benign inputs.



Paperid:3606
Authors:Anda Tang, Yiming Dong, Yutao Zeng, zhou Xun, Zhouchen Lin
Abstract:
Abstract:The expanding computational costs and limited resources underscore the critical need for budgeted-iteration training, which aims to achieve optimal learning within predetermined iteration budgets. While learning rate schedules fundamentally govern the performance of different networks and tasks, particularly in budgeted-iteration scenarios, their design remains largely heuristic, lacking theoretical foundations. In addition, the optimal learning rate schedule requires extensive trial-and-error selection, making the training process inefficient. In this work, we propose the Unified Budget-Aware (UBA) schedule, a theoretically grounded learning rate schedule that consistently outperforms commonly-used schedules among diverse architectures and tasks under different constrained training budgets. First, we bridge the gap by constructing a novel training budget-aware optimization framework, which explicitly accounts for the robustness to landscape curvature variations. From this framework, we derive the UBA schedule, controlled by a single hyper-parameter $\varphi$ that provides a trade-off between flexibility and simplicity, eliminating the need for per-network numerical optimization. Moreover, we establish a theoretical connection between $\varphi$ and the condition number, adding interpretation and justification to our approach. Besides, we prove the convergence for different values of $\varphi$. We offer practical guidelines for $\varphi$ selection via theoretical analysis and empirical results. Extensive experimental results show that UBA $\textit{consistently surpasses}$ the commonly-used schedules across diverse vision and language tasks, spanning network architectures (e.g., ResNet, OLMo) and scales, under different training-iteration budgets.



Paperid:3607
Authors:Yifei Zhang, Hao Zhu, Junhao Dong, Haoran Shi, Ziqiao Meng, Piotr Koniusz, Han Yu
Abstract:
Abstract:Parameter-efficient fine-tuning (PEFT) methods are essential for adapting large foundation models to specific tasks without prohibitive storage costs. However, current approaches like LoRA treat each layer's adaptations independently, overlooking cross-layer relationships and resulting in parameter counts that scale linearly with model depth. We present a theoretical analysis showing that skip connections in transformers induce smooth gradient propagation across layers, leading to weight adaptations that naturally concentrate in low-frequency spectral components—particularly in the cross-layer dimension. Empirical analysis confirms this insight, revealing that nearly 70\% of adaptation energy concentrates in low frequencies. Building on this finding, we propose CrossSpectra, which parameterizes all attention weight adaptations (Q, K, V) across layers as a unified 3D tensor and decomposes it using sparse spectral coefficients. By employing $k_1$ non-zero coefficients within each layer's frequency space and truncating to $k_2$ frequencies across layers, CrossSpectra achieves parameter requirements of $O(k_1 k_2)$ compared to LoRA's $O(Lrd)$, where $L$ is the layer count and $r$ is the rank. Extensive experiments demonstrate that CrossSpectra matches or exceeds baseline performance while using 275× fewer parameters than LoRA—achieving 0.36\% of LoRA's parameters when fine-tuning LLaMA-7B on instruction-following tasks. Our approach shows that leveraging architectural properties through spectral analysis can yield dramatic efficiency gains in parameter-efficient fine-tuning.



Paperid:3608
Authors:Jingyi Tian, Le Wang, Sanping Zhou, Sen Wang, lijiayi, Gang Hua
Abstract:
Learning generalizable robotic manipulation policies remains a key challenge due to the scarcity of diverse real-world training data. While recent approaches have attempted to mitigate this through self-supervised representation learning, most either rely on 2D vision pretraining paradigms such as masked image modeling, which primarily focus on static semantics or scene geometry, or utilize large-scale video prediction models that emphasize 2D dynamics, thus failing to jointly learn the geometry, semantics, and dynamics required for effective manipulation. In this paper, we present DynaRend, a representation learning framework that learns 3D-aware and dynamics-informed triplane features via masked reconstruction and future prediction using differentiable volumetric rendering. By pretraining on multi-view RGB-D video data, DynaRend jointly captures spatial geometry, future dynamics, and task semantics in a unified triplane representation. The learned representations can be effectively transferred to downstream robotic manipulation tasks via action value map prediction. We evaluate DynaRend on two challenging benchmarks, RLBench and Colosseum, as well as in real-world robotic experiments, demonstrating substantial improvements in policy success rate, generalization to environmental perturbations, and real-world applicability across diverse manipulation tasks.



Paperid:3609
Authors:Yujing Sun, Lingchen Sun, Shuaizheng Liu, Rongyuan Wu, Zhengqiang ZHANG, Lei Zhang
Abstract:
It is a challenging problem to reproduce rich spatial details while maintaining temporal consistency in real-world video super-resolution (Real-VSR), especially when we leverage pre-trained generative models such as stable diffusion (SD) for realistic details synthesis. Existing SD-based Real-VSR methods often compromise spatial details for temporal coherence, resulting in suboptimal visual quality. We argue that the key lies in how to effectively extract the degradation-robust temporal consistency priors from the low-quality (LQ) input video and enhance the video details while maintaining the extracted consistency priors.To achieve this, we propose a Dual LoRA Learning (DLoRAL) paradigm to train an effective SD-based one-step diffusion model, achieving realistic frame details and temporal consistency simultaneously.Specifically, we introduce a Cross-Frame Retrieval (CFR) module to aggregate complementary information across frames, and train a Consistency-LoRA (C-LoRA) to learn robust temporal representations from degraded inputs.After consistency learning, we fix the CFR and C-LoRA modules and train a Detail-LoRA (D-LoRA) to enhance spatial details while aligning with the temporal space defined by C-LoRA to keep temporal coherence. The two phases alternate iteratively for optimization, collaboratively delivering consistent and detail-rich outputs. During inference, the two LoRA branches are merged into the SD model, allowing efficient and high-quality video restoration in a single diffusion step. Experiments show that DLoRAL achieves strong performance in both accuracy and speed. Code and models will be released.



Paperid:3610
Authors:Matteo Ceriscioli, Karthika Mohan
Abstract:
Recent work [Richens and Everitt, 2024] has shown that agents robust to distribution shifts learn a causal model of their environment. However, these results rely on the assumption of no mediation, i.e., that an agent's actions do not affect their environment, which can be restrictive in many real-world settings. For example, a robot in an industrial plant might interact with tools, move through space, and transform products to complete its task. In this work, we extend the theoretical foundations of robust agency by proving that agents capable of adapting to distribution shifts must learn the underlying causal relationships even in the presence of mediation. We introduce an algorithm for eliciting Causal Influence Diagrams from robust agents using optimal policy oracles, with the flexibility to incorporate prior causal knowledge and demonstrate its effectiveness in mediated single-agent scenarios and multi-agent environments. We identify conditions under which the presence of a single robust agent is sufficient to recover the full causal model and derive optimal policies for other agents in the same environment. Finally, we demonstrate how to apply these results to sequential decision-making tasks modeled as Partially Observable Markov Decision Processes (POMDPs).



Paperid:3611
Authors:Yihong Luo, Tianyang Hu, Weijian Luo, Kenji Kawaguchi, Jing Tang
Abstract:
This paper addresses the challenge of achieving high-quality and fast image generation that aligns with complex human preferences. While recent advancements in diffusion models and distillation have enabled rapid generation, the effective integration of reward feedback for improved abilities like controllability and preference alignment remains a key open problem. Existing reward-guided post-training approaches targeting accelerated few-step generation often deem diffusion distillation losses indispensable.However, in this paper, we identify an interesting yet fundamental paradigm shift: as conditions become more specific, well-designed reward functions emerge as the primary driving force in training strong, few-step image generative models. Motivated by this insight, we introduce Reward-Instruct, a novel and surprisingly simple reward-centric approach for converting pre-trained base diffusion models into reward-enhanced few-step generators. Unlike existing methods, Reward-Instruct does not rely on expensive yet tricky diffusion distillation losses. Instead, it iteratively updates the few-step generator's parameters by directly sampling from a reward-tilted parameter distribution. Such a training approach entirely bypasses the need for expensive diffusion distillation losses, making it favorable to scale in high image resolutions. Despite its simplicity, Reward-Instruct yields surprisingly strong performance. Our extensive experiments on text-to-image generation have demonstrated that Reward-Instruct achieves state-of-the-art results in visual quality and quantitative metrics compared to distillation-reliant methods, while also exhibiting greater robustness to the choice of reward function.



Paperid:3612
Authors:Guancheng Wan, Fengyuan Ran, Ruikang Zhang, Wenke Huang, Xuankun Rong, Guibin Zhang, Yuxin Wu, Bo Du, Mang Ye
Abstract:
Abstract:Graph neural networks (GNNs) have achieved remarkable success in various domains but typically rely on centralized, static graphs, which limits their applicability in distributed, evolving environments. To address this limitation, we define the task of Federated Continual Graph Learning (FCGL), a paradigm for incremental learning on dynamic graphs distributed across decentralized clients. Existing methods, however, neither preserve graph topology during task transitions nor mitigate parameter conflicts in server‐side aggregation. To overcome these challenges, we introduce **MOTION**, a generalizable FCGL framework that integrates two complementary modules: the Graph Topology‐preserving Multi‐Sculpt Coarsening (G‐TMSC) module, which maintains the structural integrity of past graphs through a multi‐expert, similarity‐guided fusion process, and the Graph‐Aware Evolving Parameter Adaptive Engine (G‐EPAE) module, which refines global model updates by leveraging a topology‐sensitive compatibility matrix. Extensive experiments on real‐world datasets show that our approach improves average accuracy (AA) by an average of 30\% $\uparrow$ over the FedAvg baseline across five datasets while maintaining a negative $\downarrow$ average forgetting (AF) rate, significantly enhancing generalization and robustness under FCGL settings. The code is available for anonymous access at https://anonymous.4open.science/r/MOTION.



Paperid:3613
Authors:Pulkit Gopalani, Wei Hu
Abstract:
Training Transformers on algorithmic tasks frequently demonstrates an intriguingabrupt learningphenomenon: an extended performance plateau followed by a sudden, sharp improvement. This work investigates the underlying mechanisms for such dynamics, primarily in shallow Transformers. We reveal that during the plateau, the model often develops an interpretablepartial solutionwhile simultaneously exhibiting a strongrepetition biasin their outputs. This output degeneracy is accompanied byinternal representation collapse, where hidden states across different tokens become nearly parallel. We further identify the slow learning of optimal attention maps as a key bottleneck. Hidden progress in attention configuration during the plateau precedes the eventual rapid convergence, and directly intervening on attention significantly alters plateau duration and the severity of repetition bias and representational collapse. We validate that these identified phenomena—repetition bias and representation collapse—are not artifacts of toy setups but also manifest in the early pre-training stage of large language models like Pythia and OLMo.



Paperid:3614
Authors:Jiahui Zhang, Wenjie Du, Yang Wang
Abstract:
Masked Graph Autoencoders (MGAEs) have gained significant attention recently. Their proxy tasks typically involve random corruption of input graphs followed by reconstruction. However, in the molecular domain, two main issues arise: the predetermined mask ratio and reconstruction objectives can lead to suboptimal performance or negative transfer due to overly simplified or complex tasks, and these tasks may deviate from chemical priors. To tackle these challenges, we propose Dynamic and Chemical Constraints (DyCC) for MGAEs. This includes a masking strategy called GIBMS, which preserves essential semantic information during graph masking while adaptively adjusting the mask ratio and content for each molecule. Additionally, we introduce a Soft Label Generator (SLG) that reconstructs masked tokens as learnable prototypes (soft labels) rather than hard labels. These components adhere to chemical constraints and allow dynamic variation of proxy tasks during training. We integrate the model-agnostic DyCC into various MGAEs and conduct comprehensive experiments, demonstrating significant performance improvements. Our code is available at \url{https://anonymous.4open.science/r/DyCC-E3EC/}.



Paperid:3615
Authors:Jiarui Gan, R Majumdar, Debmalya Mandal, Goran Radanovic
Abstract:
Abstract:We study a stochastic principal-agent model. A principal and an agent interact in a stochastic environment, each privy to observations about the state not available to the other. The principal has the power of commitment, both to elicit information from the agent and to signal her own information. The players communicate with each other and then select actions independently. Both players are {\em far-sighted}, aiming to maximize their total payoffs over the entire time horizon. We consider both the computation and learning of the principal's optimal policy. The key challenge lies in enabling {\em history-dependent} policies, which are essential for achieving optimality in this model but difficult to cope with because of the exponential growth of possible histories as the size of the model increases; explicit representation of history-dependent policies is infeasible as a result.To address this challenge, we develop algorithmic techniques based on the concept of {\em inducible value set}. The techniques yield an efficient algorithm that computes an $\epsilon$-approximate optimal policy in time polynomial in $1/\epsilon$. We also present an efficient learning algorithm for an episodic reinforcement learning setting with unknown transition probabilities. The algorithm achieves sublinear regret $\widetilde{\mathcal{O}}(T^{2/3})$ for both players over $T$ episodes.



Paperid:3616
Authors:Qishuai Wen, Zhiyuan Huang, Chun-Guang Li
Abstract:
Attention mechanisms in Transformer have gained significant empirical success, however, their mathematical derivation essence remains unclear. Meanwhile, the self-attention in Transformer suffers computation burdens due to quadratic complexity. Different from the prior work on addressing the interpretability or efficiency issue separately, in this paper, we propose a unified optimization objective, which compresses all inputs by contracting a few representatives of them, to derive an attention mechanism that tackles both issues mentioned above simultaneously. Specifically, we mathematically derive an interpretable and efficient self-attention operator by unfolding the gradient-based optimization steps on the proposed objective, which contracts the representatives and broadcast the contractions back to all the inputs to yield compact and structured representation. We thus refer to it as Contract-and-Broadcast Self-Attention (CBSA). To be more specific, given a fixed number of representatives, the computational overhead of CBSA scales linearly with respect to the number of input tokens. Moreover, by specifying different sets of representatives, we can derive its more efficient variants with sacrificed expression capacity. In particular, we demonstrate that: a) full attention derives from CBSA when the inputs are self-expressed; b) channel attention derives from CBSA when the representatives are fixed and orthogonal. We conduct extensive experiments on both synthetic and real world data, and the experimental results demonstrate the effectiveness of our proposed CBSA.



Paperid:3617
Authors:Zeyu Wang, Yidan Song, Shihao Qin, Shanqing Yu, Yujin Huang, Qi Xuan, Xin Zheng
Abstract:
Privacy and security concerns are becoming increasingly critical for recommender systems, as model extraction attack provides an effective way to probe system robustness by replicating the model’s recommendation logic — potentially exposing sensitive user preferences and proprietary algorithmic knowledge.Despite the promising performance of existing model extraction methods, they still face two key challenges: \textit{unrealistic assumptions} on the requirement of accessible member or surrogate data and \textit{generalization problem} where surrogate model architecture constraints lead to overfitting on generated data.To tackle these challenges, in this paper, we first thoroughly analyze how the architecture of surrogate models influences extraction attack performance, highlighting the superior effectiveness of the graph convolution architecture. Based on this, we propose a novel Data-free Black-box Graph convolution-based Recommender Model Extraction method, dubbed DBGRME. Specifically, DBGRME contains: (1) an interaction generator to alleviate the need for member data requirements in a data-free scenario; and (2) a generalization-aware graph convolution-based surrogate model to capture diverse and complex recommender interaction patterns for mitigating the overfitting issue.Experimental results on various datasets and victim models demonstrate the superiority of our attack in data-free scenarios (e.g., surpassing PTQ data-require methods with 17.4\% improvement on LightGCN). Code is available: \url{https://anonymous.4open.science/r/DBGRME-2CB4/}.



Paperid:3618
Authors:Zhi-Kai Chen, Jun-Peng Jiang, Han-Jia Ye, De-Chuan Zhan
Abstract:
Autoregressive (AR) image generation models can produce high-fidelity images but often struggle with slow inference due to their token-by-token, sequential decoding. Speculative decoding, which employs a draft model to approximate the AR model’s output, offers a promising way to reduce inference time. While this technique has been successfully applied to accelerate text-based AR models without sacrificing output quality, its application to image generation remains largely unexplored. Directly adapting this method to images is challenging because of the substantially larger sampling space, which complicates alignment between speculative and target model predictions, and the inadequate use of two-dimensional spatial information, which limits the exploitation of local image dependencies. To address these obstacles, we propose Spatial Speculative Decoding, a novel approach that leverages the inherent two-dimensional structure of images to guide a speculative model toward more accurate predictions and faster token generation. Experimental results on multiple text-to-image benchmarks demonstrate a 1.71× speedup over standard AR models, while preserving both image fidelity and diversity.



Paperid:3619
Authors:Hyeong Kyu Choi, Jerry Zhu, Sharon Li
Abstract:
Multi-Agent Debate (MAD) has emerged as a promising paradigm for improving the performance of large language models through collaborative reasoning. Despite recent advances, the key factors driving MAD’s effectiveness remain unclear. In this work, we disentangle MAD into two key components–Majority Voting and inter-agent Debate–and assess their respective contributions. Through extensive experiments across seven NLP benchmarks, we find that Majority Voting alone accounts for most of the performance gains typically attributed to MAD. To explain this, we propose a theoretical framework that models debate as a stochastic process. We prove that it induces a martingale over agents’ belief trajectories, implying that debate alone does not improve expected correctness. Guided by these insights, we demonstrate that targeted interventions, by biasing the belief update toward correction, can meaningfully enhance debate effectiveness. Overall, our findings suggest that while MAD has potential, simple ensembling methods remain strong and more reliable alternatives in many practical settings. Code will be released for reproducible research.



Paperid:3620
Authors:Yihuai Hong, Yiran Zhao, Wei Tang, Yang Deng, Yu Rong, Wenxuan Zhang
Abstract:
Over time, a growing wave of large language models from various series has been introduced to the community. Researchers are striving to maximize the performance of language models with constrained parameter sizes. However, from a microscopic perspective, there has been limited research on how to better store knowledge in model parameters, particularly within MLPs, to enable more effective utilization of this knowledge by the model. In this work, we analyze twenty publicly available open-source large language models to investigate the relationship between their strong performance and the way knowledge is stored in their corresponding MLP parameters. Our findings reveal that as language models become more advanced and demonstrate stronger knowledge capabilities, their parameters exhibit increased specialization. Specifically, parameters in the MLPs tend to be more focused on encoding similar types of knowledge. We experimentally validate that this specialized distribution of knowledge contributes to improving the efficiency of knowledge utilization in these models. Furthermore, by conducting causal training experiments, we confirm that this specialized knowledge distribution plays a critical role in improving the model's efficiency in leveraging stored knowledge.



Paperid:3621
Authors:Wisdom Ikezogwo, Kevin M. Zhang, Saygin Seyfioglu
Abstract:
Multi-modal models are data hungry. While datasets with natural images are abundant, medical image datasets can not afford the same luxury. To enable representation learning for medical images at scale, we turn to YouTube, a platform with a large reservoir of open-source medical pedagogical videos. We curate MedicalNarratives, a dataset 4.7M medical image-text pairs, with 1M samples containing dense annotations in the form of traces and bounding boxes. Similar to think-aloud studies where instructors speak while hovering their mouse cursor movements over relevant image regions, 1M images in MedicalNarratives contains localized mouse traces in image pixels, creating a spatial association between the text and pixels. To evaluate the utility of MedicalNarratives, we train GenMedClip with a CLIP-like objective using our dataset spanning 12 medical domains. GenMedClip outperforms previous state-of-the-art models on all 12 domains on a newly constructed medical imaging benchmark. Data, demo, code, and models will be made available.



Paperid:3622
Authors:Ching Chang, Jeehyun Hwang, Yidan Shi, Haixin Wang, Wei Wang, Wen-Chih Peng, Tien-Fu Chen
Abstract:
Time series data in real-world applications such as healthcare, climate modeling, and finance are often irregular, multimodal, and messy, with varying sampling rates, asynchronous modalities, and pervasive missingness.However, existing benchmarks typically assume clean, regularly sampled, unimodal data, creating a significant gap between research and real-world deployment.We introduce Time-IMM, a dataset specifically designed to capture cause-driven irregularity in multimodal multivariate time series.Time-IMM represents nine distinct types of time series irregularity, categorized into trigger-based, constraint-based, and artifact-based mechanisms.Complementing the dataset, we introduce IMM-TSF, a benchmark library for forecasting on irregular multimodal time series, enabling asynchronous integration and realistic evaluation.IMM-TSF includes specialized fusion modules, including a timestamp-to-text fusion module and a multimodality fusion module, which support both recency-aware averaging and attention-based integration strategies.Empirical results demonstrate that explicitly modeling multimodality on irregular time series data leads to substantial gains in forecasting performance.Time-IMM and IMM-TSF provide a foundation for advancing time series analysis under real-world conditions.The dataset is publicly available at \url{https://www.kaggle.com/datasets/blacksnail789521/time-imm/data}, and the benchmark library can be accessed at \url{https://anonymous.4open.science/r/IMMTSF_NeurIPS2025}.



Paperid:3623
Authors:Thieu Vo, Duy-Tung Pham, An Nguyen The, Hoang Tran, Nhan-Phu Chung, Xin Tong, Tan Nguyen
Abstract:
This paper investigates the dynamical properties of tokens in pre-trained transformer models and explores their application to improving Transformers. To this end, we analyze the dynamical system governing the continuous-time limit of the pre-trained model and characterize the asymptotic behavior of its solutions. Specifically, we characterize when tokens move closer to or farther from one another over time, depending on the model parameters. We provide sufficient conditions, based on these parameters, to identify scenarios where tokens either converge to zero or diverge to infinity. Unlike prior works, our conditions are broader in scope and more applicable to real-world models. Furthermore, we investigate how different forms of positional encoding - specifically absolute and rotary - affect these dynamical regimes. Empirical evidence reveals that the convergence scenario adversely impacts model performance. Motivated by these insights, we propose simple refinements to Transformer architectures that mitigate convergence behavior in models with absolute or rotary positional encoding. These findings support theoretical foundations and design principles for improving Transformer models.



Paperid:3624
Authors:Shuang Zeng, Xinyuan Chang, Mengwei Xie, Xinran Liu, Yifan Bai, Zheng Pan, Mu Xu, Xing Wei
Abstract:
Visual language models (VLMs) have attracted increasing interest in autonomous driving due to their powerful reasoning capabilities. However, existing VLMs typically utilize discrete text Chain-of-Thought (CoT) tailored to the current scenario, which essentially represents highly abstract and symbolic compression of visual information, potentially leading to spatio-temporal relationship ambiguity and fine-grained information loss. Is autonomous driving better modeled on real-world simulation and imagination than on pure symbolic logic? In this paper, we propose a spatio-temporal CoT reasoning method that enables models to think visually. First, VLM serves as a world model to generate unified image frame for predicting future world states: where perception results (e.g., lane divider and 3D detection) represent the future spatial relationships, and ordinary future frame represent the temporal evolution relationships.This spatio-temporal CoT then serves as intermediate reasoning steps, enabling the VLM to function as an inverse dynamics model for trajectory planning based on current observations and future predictions. To implement visual generation in VLMs, we propose a unified pretraining paradigm integrating visual generation and understanding, along with a progressive visual CoT enhancing autoregressive image generation. Extensive experimental results demonstrate the effectiveness of the proposed method, advancing autonomous driving towards visual reasoning.



Paperid:3625
Authors:Hossein Jafarinia, Danial Hamdi, Amirhossein Alamdar, Elahe Zahiri, Soroush Vafaie Tabar, Alireza Alipanah, Nahal Mirzaie, Saeed Razavi, Amir Najafi, Mohammad Hossein Rohban
Abstract:
Multiple Instance Learning (MIL) is a standard weakly supervised approach for Whole Slide Image (WSI) classification, where performance hinges on both feature representation and MIL-pooling strategies. Recent research has predominantly focused on Transformer-based architectures adapted for WSIs. However, we argue that this trend faces a fundamental limitation: data scarcity. In typical settings, Transformer models yield only marginal gains without access to large-scale datasets—resources that are virtually inaccessible to all but a few well-funded research labs. This limitation motivates the exploration of simpler, non-attention-based methods in conjunction with unsupervised representation learning. Building on this motivation, we first present a series of theoretical results characterizing classical MIL-pooling strategies, revealing trade-offs between \emph{generalization} and \emph{sensitivity}. Grounded in this analysis, we introduce MaxSoft—a novel MIL-pooling function that enables flexible control over this trade-off, allowing adaptation to specific tasks and datasets. To further tackle real-world deployment challenges such as specimen heterogeneity, we propose PerPatch augmentation—a simple yet effective technique that enhances model robustness. Empirically, MaxSoft achieves state-of-the-art performance in low-data regimes across four major benchmarks, often matching or surpassing large-scale foundation models. When combined with PerPatch augmentation, this performance is further improved through increased robustness.



Paperid:3626
Authors:Ali Hariri, Alvaro Arroyo, Alessio Gravina, Moshe Eliasof, Carola-Bibiane Schönlieb, Davide Bacciu, Xiaowen Dong, Kamyar Azizzadenesheli, Pierre Vandergheynst
Abstract:
ChebNet, one of the earliest spectral GNNs, has largely been overshadowed by Message Passing Neural Networks (MPNNs), which gained popularity for their simplicity and effectiveness in capturing local graph structure. Despite their success, MPNNs are limited in their ability to capture long-range dependencies between nodes. This has led researchers to adapt MPNNs throughrewiringor make use ofGraph Transformers, which compromise the computational efficiency that characterized early spatial message passing architectures, and typically disregard the graph structure. Almost a decade after its original introduction, we revisit ChebNet to shed light on its ability to model distant node interactions. We find that out-of-box, ChebNet already shows competitive advantages relative to classical MPNNs and GTs on long-range benchmarks, while maintaining good scalability properties for high-order polynomials. However, we uncover that this polynomial expansion leads ChebNet to an unstable regime during training. To address this limitation, we cast ChebNet as a stable and non-dissipative dynamical system, which we coin Stable-ChebNet. Our Stable-ChebNet model allows for stable information propagation, and has controllable dynamics which do not require the use of eigendecompositions, positional encodings, or graph rewiring. Across several benchmarks, Stable-ChebNet achieves near state-of-the-art performance.



Paperid:3627
Authors:wenlong deng, Yi Ren, Muchen Li, Danica J. Sutherland, Xiaoxiao Li, Christos Thrampoulidis
Abstract:
Reinforcement learning (RL) has become popular in enhancing the reasoning capabilities of large language models (LLMs), with Group Relative Policy Optimization (GRPO) emerging as a widely used algorithm in recent systems. Despite GRPO's widespread adoption, we identify a previously unrecognized phenomenon we term Lazy Likelihood Displacement (LLD), wherein the likelihood of correct responses marginally increases or even decreases during training. This behavior mirrors a recently discovered misalignment issue in Direct Preference Optimization (DPO), attributed to the influence of negative gradients. We provide a theoretical analysis of GRPO’s learning dynamic, identifying the source of LLD as the naive penalization of all tokens in incorrect responses with the same strength. To address this, we develop a method called NTHR, which downweights penalties on tokens contributing to the LLD. Unlike prior DPO-based approaches, NTHR takes advantage of GRPO’s group-based structure, using correct responses as anchors to identify influential tokens. Experiments on math reasoning benchmarks demonstrate that NTHR effectively mitigates LLD, yielding consistent performance gains across models ranging from 0.5B to 3B parameters.



Paperid:3628
Authors:Liangyu Zhong, Fabio Rosenthal, Joachim Sicking, Fabian Hüger, Thorsten Bagdonat, Hanno Gottschalk, Leo Schwinn
Abstract:
While Multimodal Large Language Models (MLLMs) offer great perception and reasoning capabilities for image-text input, fine-grained Visual Question Answering (VQA) focusing on small details still remains a challenge. Although visual cropping techniques seem promising, recent approaches have several limitations: the need for task-specific fine-tuning, low efficiency due to uninformed exhaustive search, or incompatibility with efficient attention implementations. We address these shortcomings by proposing a training-free visual cropping method, dubbed FOCUS, that leverages MLLM-internal representations to guide the search for the most relevant image region. This is accomplished in four steps: first, we identify the target object(s) in the prompt; second, we compute an object relevance map using the key-value (KV) cache; third, we propose and rank relevant image regions based on the map; and finally, we perform the fine-grained VQA task using the top-ranked region. As a result of this informed search strategy, our method achieves strong performance across four fine-grained VQA datasets and two types of MLLM. It outperforms three existing visual cropping methods in both accuracy and efficiency, and matches the best-performing baseline, ZoomEye, with 3 - 6.5 x higher efficiency. Finally, we perform an ablation study to assess the impact of key design choices.We plan to release our code upon acceptance.



Paperid:3629
Authors:Donghao Sun, Xi Wang, Xu Yang, Kun Wei, Cheng Deng
Abstract:
Continual learning from unlabeled data streams while effectively combating catastrophic forgetting poses an intractable challenge. Traditional methods predominantly rely on visual clustering techniques to generate pseudo labels, which are frequently plagued by problems such as noise and suboptimal quality, profoundly affecting the impact on the model evolution. To surmount these obstacles, we introduce an innovative approach that synergistically combines both visual and textual information to generate dual space hybrid pseudo labels for reliable model continual evolution. Specifically, by harnessing the capabilities of large multimodal models, we initially generate generalizable text descriptions for a few representative samples. These descriptions then undergo a `Coarse to Fine' refinement process to capture the subtle nuances between different data points, significantly enhancing the semantic accuracy of the descriptions. Simultaneously, a novel cross-modal hybrid approach seamlessly integrates these fine-grained textual descriptions with visual features, thereby creating a more robust and reliable supervisory signal. Finally, such descriptions are employed to alleviate the catastrophic forgetting issue via a semantic alignment distillation, which capitalizes on the stability inherent in language knowledge to effectively prevent the model from forgetting previously learned information. Comprehensive experiments conducted on a variety of benchmarks demonstrate that our proposed method attains state-of-the-art performance, and ablation studies further substantiate the effectiveness and superiority of the proposed method.



Paperid:3630
Authors:Elvis Dohmatob
Abstract:
Self-attention has emerged as a fundamental component driving the success of modern transformer architectures, which power large language models and various applications. However, a theoretical understanding of how such models actually work is still under active development. The recent work of (Marion et al., 2025) introduced the so-called "single-location regression" problem, which can provably be solved by a simplified self-attention layer but not by linear models, thereby demonstrating a striking functional separation. A rigorous analysis of self-attention with softmax for this problem is challenging due to the coupled nature of the model. In the present work, we use ideas from the classical random energy model in statistical physics to analyze softmax self-attention on the single-location problem. Our analysis yields exact analytic expressions for the population risk in terms of the overlaps between the learned model parameters and those of an oracle. Moreover, we derive a detailed description of the gradient descent dynamics for these overlaps and prove that, under broad conditions, the dynamics converge to the unique oracle attractor. Our work not only advances our understanding of self-attention but also provides key theoretical ideas that are likely to find use in further analyses of even more complex transformer architectures.



Paperid:3631
Authors:Ning Zhang, Henry Kenlay, Li Zhang, Mihai Cucuringu, Xiaowen Dong
Abstract:
Graph convolutional neural networks (GCNNs) have emerged as powerful tools for analyzing graph-structured data, achieving remarkable success across diverse applications. However, the theoretical understanding of the stability of these models, i.e., their sensitivity to small changes in the graph structure, remains in rather limited settings, hampering the development and deployment of robust and trustworthy models in practice. To fill this gap, we study how small perturbations in the graph topology affect GCNN outputs and propose a novel formulation for analyzing model stability. Unlike prior studies that focus only on worst-case perturbations, our distribution-aware formulation characterizes output perturbations across a broad range of input data. This way, our framework enables, for the first time, a probabilistic perspective on the interplay between the statistical properties of the node data and perturbations in the graph topology. We conduct extensive experiments to validate our theoretical findings and demonstrate their benefits over existing baselines, in terms of both representation stability and adversarial attacks on downstream tasks. Our results demonstrate the practical significance of the proposed formulation and highlight the importance of incorporating data distribution into stability analysis.



Paperid:3632
Authors:Hongtao Huang, Chengkai Huang, Junda Wu, Tong Yu, Julian Mcauley, Lina Yao
Abstract:
Forecasting multi-step user behavior trajectories requires reasoning over structured preferences across future actions, a challenge overlooked by traditional sequential recommendation. This problem is critical for applications such as personalized commerce and adaptive content delivery, where anticipating a user’s complete action sequence enhances both satisfaction and business outcomes. We identify an essential limitation of existing paradigms: their inability to capture global, listwise dependencies among sequence items.To address this, we formulate User Behavior Trajectory Prediction (UBTP) as a new task setting that explicitly models long-term user preferences. We introduce Listwise Preference Diffusion Optimization (LPDO), a diffusion-based training framework that directly optimizes structured preferences over entire item sequences. LPDO incorporates a Plackett–Luce supervision signal and derives a tight variational lower bound aligned with listwise ranking likelihoods, enabling coherent preference generation across denoising steps while overcoming the independent-token assumption of prior diffusion methods.To rigorously evaluate multi-step prediction quality, we propose the task-specific metric: Sequential Match (SeqMatch), which measures exact trajectory agreement, and adopt Perplexity (PPL), which assesses probabilistic fidelity. Extensive experiments on real-world user behavior benchmarks demonstrate that LPDO outperforms state-of-the-art baselines, establishing a new benchmark for structured preference learning with diffusion models.



Paperid:3633
Authors:Asal Mehradfar, Xuzhe Zhao, Yilun Huang, Emir Ceyani, Yankai Yang, Shihao han, Hamidreza Aghasi, Salman Avestimehr
Abstract:
Designing analog circuits from performance specifications is a complex, multi-stage process encompassing topology selection, parameter inference, and layout feasibility. We introduce FALCON, a unified machine learning framework that enables fully automated, specification-driven analog circuit synthesis through topology selection and layout-constrained optimization. Given a target performance, FALCON first selects an appropriate circuit topology using a performance-driven classifier guided by human design heuristics. Next, it employs a custom, edge-centric graph neural network trained to map circuit topology and parameters to performance, enabling gradient-based parameter inference through the learned forward model. This inference is guided by a differentiable layout cost, derived from analytical equations capturing parasitic and frequency-dependent effects, and constrained by design rules. We train and evaluate FALCON on a large-scale custom dataset of 1M analog mm-wave circuits, generated and simulated using Cadence Spectre across 20 expert-designed topologies. Through this evaluation, FALCON demonstrates >99\% accuracy in topology inference, <10\% relative error in performance prediction, and efficient layout-aware design that completes in under 1 second per instance. Together, these results position FALCON as a practical and extensible foundation model for end-to-end analog circuit design automation.



Paperid:3634
Authors:Tam Le, Truyen Nguyen, Hideitsu Hino, Kenji Fukumizu
Abstract:
Abstract:We investigate optimal transport (OT) for measures on graph metric spaces with different total masses. To mitigate the limitations of traditional $L^p$ geometry, Orlicz-Wasserstein (OW) and generalized Sobolev transport (GST) employ \emph{Orlicz geometric structure}, leveraging convex functions to capture nuanced geometric relationships and remarkably contribute to advance certain machine learning approaches. However, both OW and GST are restricted to measures with equal total mass, limiting their applicability to real-world scenarios where mass variation is common, and input measures may have noisy supports, or outliers. To address unbalanced measures, OW can either incorporate mass constraints or marginal discrepancy penalization, but this leads to a more complex two-level optimization problem. Additionally, GST provides a scalable yet rigid framework, which poses significant challenges to extend GST to accommodate nonnegative measures. To tackle these challenges, in this work we revisit the entropy partial transport (EPT) problem. By exploiting Caffarelli \& McCann's insights, we develop a novel variant of EPT endowed with Orlicz geometric structure, called \emph{Orlicz-EPT}. We establish theoretical background to solve Orlicz-EPT using a binary search algorithmic approach. Especially, by leveraging the dual EPT and the underlying graph structure, we formulate a novel regularization approach that leads to the proposed \emph{Orlicz-Sobolev transport} (OST). Notably, we demonstrate that OST can be efficiently computed by simply solving a univariate optimization problem, in stark contrast to the intensive computation needed for Orlicz-EPT. Building on this, we derive geometric structures for OST and draw its connections to other transport distances. We empirically illustrate that OST is several-order faster than Orlicz-EPT. Furthermore, we show preliminary evidence on the advantages of OST for measures on a graph in document classification and topological data analysis.



Paperid:3635
Authors:Moritz Haas, Sebastian Bordt, Ulrike Luxburg, Leena Chennuru Vankadara
Abstract:
The dominant paradigm for training large-scale vision and language models is He initialization and a single global learning rate (standard parameterization, SP). Despite its practical success, standard parametrization remains poorly understood from a theoretical perspective: Existing infinite-width theory would predict instability under large learning rates and vanishing feature learning under stable learning rates. However, empirically optimal learning rates consistently decay much slower than theoretically predicted. By carefully studying neural network training dynamics, we demonstrate that this discrepancy is not fully explained by finite-width phenomena such as catapult effects or a lack of alignment between weights and incoming activations. We instead show that the apparent contradiction can be fundamentally resolved by taking the loss function into account: In contrast to Mean Squared Error (MSE) loss, we prove that under cross-entropy (CE) loss, an intermediatecontrolled divergenceregime emerges, where logits diverge but loss, gradients, and activations remain stable. Stable training under large learning rates enables persistent feature evolution at scale in all hidden layers, which is crucial for the practical success of SP. In experiments across optimizers (SGD, Adam), architectures (MLPs, GPT) and data modalities (vision, language), we validate that neural networks operate in this controlled divergence regime under CE loss but not under MSE loss. Our empirical evidence suggests that width-scaling considerations are surprisingly useful for predicting empirically optimal learning rate exponents. Finally, our analysis clarifies the effectiveness and limitations of recently proposed layerwise learning rate scalings for standard initialization.



Paperid:3636
Authors:Xuhui Liao, qiyu wang, Zhiqiang Liang, Liwei Xiao, Junjie Chen
Abstract:
Inverse protein folding addresses the challenge of designing amino acid sequences that fold into a predetermined tertiary structure, bridging geometric and evolutionary constraints to advance protein engineering. Inspired by the pivotal role of multiple sequence alignments (MSAs) in structure prediction models like AlphaFold, we hypothesize that structural alignments can provide an informative prior for inverse folding. In this study, we introduce DualMPNN, a dual-stream message passing neural network that leverages structurally homologous templates to guide amino acid sequence design of predefined query structures. DualMPNN processes the query and template proteins via two interactive branches, coupled through alignment-aware cross-stream attention mechanisms that enable exchange of geometric and co-evolutionary signals. Comprehensive evaluations across on CATH 4.2, TS50 and T500 benchmarks demonstrate DualMPNN achieves state-of-the-art recovery rates of 65.51%, 70.99%, and 70.37%, significantly outperforming base model ProteinMPNN by 15.64%, 16.56%, 12.29%, respectively. Further template quality analysis and structural foldability assessment underscore the value of structural alignment priors for protein design.



Paperid:3637
Authors:Jiahui Zhang, Yurui Chen, Yueming Xu, Ze Huang, Jilin Mei, Chunhui Chen, Yanpeng Zhou, Yu-Jie Yuan, Xinyue Cai, Guowei Huang, Xingyue Quan, Hang Xu, Li Zhang
Abstract:
Recent advances in LVLMs have improved vision-language understanding, but they still struggle with spatial perception, limiting their ability to reason about complex 3D scenes. Unlike previous approaches that incorporate 3D representations into models to improve spatial understanding, we aim to unlock the potential of VLMs by leveraging spatially relevant image data. To this end, we introduce a novel 2D spatial data generation and annotation pipeline built upon scene data with 3D ground-truth. This pipeline enables the creation of a diverse set of spatial tasks, ranging from basic perception tasks to more complex reasoning tasks. Leveraging this pipeline, we construct SPAR-7M, a large-scale dataset generated from thousands of scenes across multiple public datasets. In addition, we introduce SPAR-Bench, a benchmark designed to offer a more comprehensive evaluation of spatial capabilities compared to existing spatial benchmarks, supporting both single-view and multi-view inputs. Training on both SPAR-7M and large-scale 2D datasets enables our models to achieve state-of-the-art performance on 2D spatial benchmarks. Further fine-tuning on 3D task-specific datasets yields competitive results, underscoring the effectiveness of our dataset in enhancing spatial reasoning.



Paperid:3638
Authors:Rui Chu, Bingyin Zhao, Hanling Jiang, Shuchin Aeron, Yingjie Lao
Abstract:
Recent research shows that large language models (LLMs) are vulnerable to hijacking attacks under the scenario of in-context learning (ICL) where LLMs demonstrate impressive capabilities in performing tasks by conditioning on a sequence of in-context examples (ICEs) (i.e., prompts with task-specific input-output pairs). Adversaries can manipulate the provided ICEs to steer the model toward attacker-specified outputs, effectively ''hijacking'' the model's decision-making process. Unlike traditional adversarial attacks targeting single inputs, hijacking attacks in LLMs aim to subtly manipulate the initial few examples to influence the model's behavior across a range of subsequent inputs, which requires distributed and stealthy perturbations. However, existing approaches overlook how to effectively allocate the perturbation budget across ICEs. We argue that fixed budgets miss the potential of dynamic reallocation to improve attack success while maintaining high stealthiness and text quality. In this paper, we propose BAM-ICL, a novelbudgetedadversarialmanipulation hijacking attack framework for in-context learning. We also consider a more practical yet stringent scenario where ICEs arrive sequentially and only the current ICE can be perturbed. BAM-ICL mainly consists of two stages: In the offline stage, where we assume the adversary has access to data drawn from the same distribution as the target task, we develop a global gradient-based attack to learn optimal budget allocations across ICEs. In the online stage, where ICEs arrive sequentially, perturbations are generated progressively according to the learned budget profile. We evaluate BAM-ICL on diverse LLMs and datasets. The experimental results demonstrate that it achieves superior attack success rates and stealthiness, and the adversarial ICEs are highly transferable to other models.



Paperid:3639
Authors:Xin Zhang, Mingxin Li, Yanzhao Zhang, Dingkun Long, Yongqi Li, Yinghui Li, Pengjun Xie, Meishan Zhang, Wenjie Li, Min Zhang, Philip S Yu
Abstract:
Searching over semi-structured data with natural language (NL) queries has attracted sustained attention, enabling broader audiences to access information easily. As more applications, such as LLM agents and RAG systems, emerge to search and interact with semi-structured data, two major challenges have become evident: (1) the increasing diversity of domains and schema variations, making domain-customized solutions prohibitively costly; (2) the growing complexity of NL queries, which combine both exact field matching conditions and fuzzy semantic requirements, often involving multiple fields and implicit reasoning. These challenges make formal language querying or keyword-based search insufficient. In this work, we explore neural retrievers as a unified non-formal querying solution by directly index semi-structured collections and understand NL queries. We employ LLM-based automatic evaluation and build a large-scale semi-structured retrieval benchmark (SSRB) using LLM generation and filtering, containing 14M semi-structured objects from 99 different schemas across 6 domains, along with 8,485 test queries that combine both exact and fuzzy matching conditions. Our systematic evaluation of popular retrievers shows that current state-of-the-art models could achieve acceptable performance, yet they still lack precise understanding of matching constraints. While by in-domain training of dense retrievers, the performance can be significantly improved. We believe that our SSRB could serve as a valuable resource for future research in this area, and we hope to inspire further exploration of semi-structured retrieval with complex queries.



Paperid:3640
Authors:Xiyao Wang, Zhengyuan Yang, Chao Feng, Yuhang Zhou, Xiaoyu Liu, Yongyuan Liang, Ming Li, Ziyi Zang, Linjie Li, Chung-Ching Lin, Kevin Lin, Furong Huang, Lijuan Wang
Abstract:
Reinforcement learning (RL) has shown great effectiveness for fine-tuning large language models (LLMs) using tasks that are challenging yet easily verifiable, such as math reasoning or code generation. However, extending this success to visual perception in vision–language models (VLMs) has been impeded by the scarcity of vision-centric tasks that are simultaneously challenging and unambiguously verifiable. To this end, we introduce \textbf{ViCrit} (\textit{Visual Caption Hallucination Critic}), an RL proxy task that trains VLMs to localize a subtle, synthetic visual hallucination injected into paragraphs of human-written image captions. Starting from a 200-word captions, we inject a single, subtle visual description error—altering a few words on objects, attributes, counts, or spatial relations—and task the model to pinpoint the corrupted span given the image and the modified caption. This formulation preserves the full perceptual difficulty while providing a binary, exact-match reward that is easy to compute and unambiguous. Models trained with the \textbf{ViCrit Task} exhibit substantial gains across a variety of VL benchmarks. Crucially, the improvements transfer beyond natural-image training data to abstract image reasoning and visual math, showing promises of learning to perceive rather than barely memorizing seen objects. To facilitate evaluation, we further introduce \textbf{ViCrit-Bench}, a category-balanced diagnostic benchmark that systematically probes perception errors across diverse image domains and error types. Together, our results demonstrate that fine-grained hallucination criticism is an effective and generalizable objective for enhancing visual perception in VLMs.



Paperid:3641
Authors:Tianqi Luo, Chuhan Huang, Leixian Shen, Boyan Li, Shuyu Shen, Wei Zeng, Nan Tang, Yuyu Luo
Abstract:
Text-to-Visualization (Text2VIS) enables users to create visualizations from natural language queries, making data insights more accessible. However, Text2VIS faces challenges in interpreting ambiguous queries, as users often express their visualization needs in imprecise language. To address this challenge, we introduce nBench 2.0, a new benchmark designed to evaluate Text2VIS systems in scenarios involving ambiguous queries. nvBench 2.0 includes 7,878 natural language queries and 24,076 corresponding visualizations, derived from 780 tables across 153 domains. It is built using a controlled ambiguity-injection pipeline that generates ambiguous queries through a reverse-generation workflow. By starting with unambiguous seed visualizations and selectively injecting ambiguities, the pipeline yields multiple valid interpretations for each query, with each ambiguous query traceable to its corresponding visualization through step-wise reasoning paths.We evaluate various Large Language Models (LLMs) on their ability to perform ambiguous Text2VIS tasks using nBench 2.0. We also propose Step-Text2Vis, an LLM-based model trained on nvBench 2.0, which enhances performance in ambiguous scenarios through step-wise preference optimization. Our results show that Step-Text2Vis outperforms all baselines, setting a new state-of-the-art for ambiguous Text2VIS tasks. Our source code and data are available at https://nvbench2.github.io/



Paperid:3642
Authors:Matvei Popov, Peter Robicheaux, Anish Madan, Isaac Robinson, Joseph Nelson, Deva Ramanan, Neehar Peri
Abstract:
Vision-language models (VLMs) trained on internet-scale data achieve remarkable zero-shot detection performance on common objects like car, truck, and pedestrian. However, state-of-the-art models still struggle to generalize to out-of-distribution classes, tasks and imaging modalities not typically found in their pre-training. Rather than simply re-training VLMs on more visual data, we argue that one should align VLMs to new concepts with annotation instructions containing a few visual examples and rich textual descriptions. To this end, we introduce Roboflow100-VL, a large-scale collection of 100 multi-modal object detection datasets with diverse concepts not commonly found in VLM pre-training. We evaluate state-of-the-art models on our benchmark in zero-shot, few-shot, semi-supervised, and fully-supervised settings, allowing for comparison across data regimes. Notably, we find that VLMs like GroundingDINO and Qwen2.5-VL achieve less than 2% zero-shot accuracy on challenging medical imaging datasets within Roboflow100-VL, demonstrating the need for few-shot concept alignment. Our code and dataset are available on GitHub and Roboflow.



Paperid:3643
Authors:Sijia Chen, Xiaomin Li, mengxue zhang, Eric Jiang, Qingcheng Zeng, Chen-Hsiang Yu
Abstract:
Large language models (LLMs) are increasingly deployed in medical contexts, raising critical concerns about safety, alignment, and susceptibility to adversarial manipulation. While prior benchmarks assess model refusal capabilities for harmful prompts, they often lack clinical specificity, graded harmfulness levels, and coverage of jailbreak-style attacks. We introduce CARES (Clinical Adversarial Robustness and Evaluation of Safety), a benchmark for evaluating LLM safety in healthcare. CARES includes over 18,000 prompts spanning eight medical safety principles, four harm levels, and four prompting styles—direct, indirect, obfuscated, and role-play—to simulate both malicious and benign use cases. We propose a three-way response evaluation protocol (Accept, Caution, Refuse) and a fine-grained Safety Score metric to assess model behavior. Our analysis reveals that many state-of-the-art LLMs remain vulnerable to jailbreaks that subtly rephrase harmful prompts, while also over-refusing safe but atypically phrased queries. Finally, we propose a mitigation strategy using a lightweight classifier to detect jailbreak attempts and steer models toward safer behavior via reminder-based conditioning. CARES provides a rigorous framework for testing and improving medical LLM safety under adversarial and ambiguous conditions.



Paperid:3644
Authors:Jonas Elsborg, Luca Thiede, Alan Aspuru-Guzik, Tejs Vegge, Arghya Bhowmik
Abstract:
We present the Electronic Tensor Reconstruction Algorithm (ELECTRA) - an equivariant model for predicting electronic charge densities using floating orbitals. Floating orbitals are a long-standing concept in the quantum chemistry community that promises more compact and accurate representations by placing orbitals freely in space, as opposed to centering all orbitals at the position of atoms. Finding the ideal placement of these orbitals requires extensive domain knowledge, though, which thus far has prevented widespread adoption. We solve this in a data-driven manner by training a Cartesian tensor network to predict the orbital positions along with orbital coefficients. This is made possible through a symmetry-breaking mechanism that is used to learn position displacements with lower symmetry than the input molecule while preserving the rotation equivariance of the charge density itself. Inspired by recent successes of Gaussian Splatting in representing densities in space, we are using Gaussian orbitals and predicting their weights and covariance matrices. Our method achieves a state-of-the-art balance between computational efficiency and predictive accuracy on established benchmarks.



Paperid:3645
Authors:Lequan Lin, Dai Shi, Andi Han, Feng Chen, Qiuzheng Chen, Jiawen Li, Zhaoyang Li, Jiyuan Zhang, Zhenbang Sun, Junbin Gao
Abstract:
Supervised learning relies on high-quality labeled data, but obtaining such data through human annotation is both expensive and time-consuming. Recent work explores using large language models (LLMs) for annotation, but LLM-generated labels still fall short of human-level quality. To address this problem, we propose the Annotation with Critical Thinking (ACT) data pipeline, where LLMs serve not only as annotators but also as judges to critically identify potential errors. Human effort is then directed towards reviewing only the most "suspicious" cases, significantly improving the human annotation efficiency. Our major contributions are as follows: (1) ACT is applicable to a wide range of domains, including natural language processing (NLP), computer vision (CV), and multimodal understanding, by leveraging multimodal-LLMs (MLLMs). (2) Through empirical studies, we derive 7 insights on how to enhance annotation quality while efficiently reducing the human cost, and then translate these findings into user-friendly guidelines. (3) We theoretically analyze how to modify the loss function so that models trained on ACT data achieve similar performance to those trained on fully human-annotated data. Our experiments show that the performance gap can be reduced to less than 2% on most benchmark datasets while saving up to 90% of human costs.



Paperid:3646
Authors:Xiao An, Jiaxing Sun, Zihan Gui, Wei He
Abstract:
The rapid advancement of Large Vision-Language Models (VLMs), both general-domain models and those specifically tailored for remote sensing, has demonstrated exceptional perception and reasoning capabilities in Earth observation tasks. However, a benchmark for systematically evaluating their capabilities in this domain is still lacking. To bridge this gap, we propose CHOICE, an extensive benchmark designed to objectively evaluate the hierarchical remote sensing capabilities of VLMs. Focusing on 2 primary capability dimensions essential to remote sensing: perception and reasoning, we further categorize 6 secondary dimensions and 23 leaf tasks to ensure a well-rounded assessment coverage. CHOICE guarantees the quality of all 10,507 problems through a rigorous process of data collection from 50 globally distributed cities, question construction and quality control. The newly curated data and the format of multiple-choice questions with definitive answers allow for an objective and straightforward performance assessment. Our evaluation of 3 proprietary and 21 open-source VLMs highlights their critical limitations within this specialized context. We hope that CHOICE will serve as a valuable resource and offer deeper insights into the challenges and potential of VLMs in the field of remote sensing. Code and dataset are available atthis https URL.



Paperid:3647
Authors:Hongbo Liu, Jingwen He, Yi Jin, Dian Zheng, Yuhao Dong, Fan Zhang, Ziqi Huang, Yinan He, Weichao Chen, Yu Qiao, Wanli Ouyang, Shengjie Zhao, Ziwei Liu
Abstract:
Recent Vision-Language Models (VLMs) have shown strong performance in general-purpose visual understanding and reasoning, but their ability to comprehend the visual grammar of movie shots remains underexplored and insufficiently evaluated. To bridge this gap, we present \textbf{ShotBench}, a dedicated benchmark for assessing VLMs’ understanding of cinematic language. ShotBench includes 3,049 still images and 500 video clips drawn from more than 200 films, with each sample annotated by trained annotators or curated from professional cinematography resources, resulting in 3,608 high-quality question-answer pairs. We conduct a comprehensive evaluation of over 20 state-of-the-art VLMs across eight core cinematography dimensions. Our analysis reveals clear limitations in fine-grained perception and cinematic reasoning of current VLMs. To improve VLMs capability in cinematography understanding, we construct a large-scale multimodal dataset, named ShotQA, which contains about 70k Question-Answer pairs derived from movie shots. Besides, we propose ShotVL and train this VLM model with a two-stage training strategy, integrating both supervised fine-tuning and Group Relative Policy Optimization (GRPO). Experimental results demonstrate that our model achieves substantial improvements, surpassing all existing strongest open-source and proprietary models evaluated on ShotBench, establishing a new state-of-the-art performance.



Paperid:3648
Authors:Changkun Liu, Bin Tan, Zeran Ke, Shangzhan Zhang, Jiachen Liu, Ming Qian, Nan Xue, Yujun Shen, Tristan Braud
Abstract:
Abstract:This paper addresses metric 3D reconstruction of indoor scenes by exploiting their inherent geometric regularities with compact representations. Using planar 3D primitives -- a well-suited representation for man-made environments -- we introduce PLANA3R, a pose-free framework for metric $\underline{Plana}$r $\underline{3}$D $\underline{R}$econstruction from unposed two-view images. Our approach employs Vision Transformers to extract a set of sparse planar primitives, estimate relative camera poses, and supervise geometry learning via planar splatting, where gradients are propagated through high-resolution rendered depth and normal maps of primitives. Unlike prior feedforward methods that require 3D plane annotations during training, PLANA3R learns planar 3D structures without explicit plane supervision, enabling scalable training on large-scale stereo datasets using only depth and normal annotations. We validate PLANA3R on multiple indoor-scene datasets with metric supervision and demonstrate strong generalization to out-of-domain indoor environments across diverse tasks under metric evaluation protocols, including 3D surface reconstruction, depth estimation, and relative pose estimation. Furthermore, by formulating with planar 3D representation, our method emerges with the ability for accurate plane segmentation.



Paperid:3649
Authors:Zhen Zhang, Xuehai He, Weixiang Yan, Ao Shen, Chenyang Zhao, Xin Wang
Abstract:
Human cognition typically involves thinking through abstract, fluid concepts rather than strictly using discrete linguistic tokens. Current Large Language Models (LLMs), however, are constrained to reasoning within the boundaries of human language, processing discrete token embeddings that represent fixed points in semantic space. This discrete constraint restricts the expressive power and upper potential of such reasoning models, often causing incomplete exploration of reasoning paths, as standard Chain-of-Thought (CoT) methods rely on sampling one token per step. In this work, we introduce Soft Thinking, a training-free method that emulates human-like ``soft'' reasoning by generating abstract concept tokens in a continuous concept space. These concept tokens are created by the probability-weighted mixture of token embeddings, which span the continuous concept space, enabling smooth transitions and richer representations that transcend traditional discrete boundaries. In essence, each generated concept token encapsulates multiple meanings from related discrete tokens, implicitly exploring various reasoning paths to converge effectively toward the correct answer. Empirical evaluations on diverse mathematical and coding benchmarks consistently demonstrate the effectiveness and efficiency of Soft Thinking, improving pass@1 accuracy by up to 2.48 points while simultaneously reducing token usage by up to 22.4\% compared to standard CoT. Qualitative analysis further reveals that Soft Thinking outputs remain highly interpretable and readable, highlighting the potential of Soft Thinking to break the inherent limits of discrete language-based reasoning.



Paperid:3650
Authors:Ye Liu, Zihan Ji, Hongmin Cai
Abstract:
Multimodal representation learning is critical for a wide range of applications, such as multimodal sentiment analysis. Current multimodal representation learning methods mainly focus on the multimodal alignment or fusion strategies, such that the complementary and consistent information among heterogeneous modalities can be fully explored. However, they mistakenly treat the uncertainty noise within each modality as the complementary information, failing to simultaneously leverage both consistent and complementary information while eliminating the aleatoric uncertainty within each modality. To address this issue, we propose a plug-and-play feature causality decomposition method for multimodal representation learning from causality perspective, which can be integrated into existing models with no affects on the original model structures. Specifically, to deal with the heterogeneity and consistency, according to whether it can be aligned with other modalities, the unimodal feature is first disentangled into two parts: modality-invariant (the synergistic information shared by all heterogeneous modalities) and modality-specific part. To deal with complementarity and uncertainty, the modality-specific part is further decomposed into unique and redundant features, where the redundant feature is removed and the unique feature is reserved based on the backdoor-adjustment. The effectiveness of noise removal is supported by causality theory. Finally, the task-related information, including both synergistic and unique components, is further fed to the original fusion module to obtain the final multimodal representations. Extensive experiments show the effectiveness of our proposed strategies.



Paperid:3651
Authors:Michelle Yuan, Khushbu Pahwa, Shuaichen Chang, Mustafa Kaba, MONICA SUNKARA, Jiarong Jiang, Xiaofei Ma, Yi Zhang
Abstract:
Designing effective agentic systems requires the seamless composition and integration of agents, tools, and models within dynamic and uncertain environments. Most of the existing methods rely on static, semantic retrieval approaches for the tool or agent discovery. However, effective reuse and composition of existing components remain challenging due to incomplete capability descriptions and the limitations of such retrieval methods as they overlook component capability, cost, and real-time utility; which are crucial factors for adaptive and efficient system composition. To address these challenges, we introduce a structured, automated framework for agentic system composition, inspired by the knapsack problem. Our framework enables a composer agent to systematically identify, select, and assemble an optimal set of agentic components by jointly considering performance, budget constraints, and compatibility. By dynamically testing candidate components and modeling their utility in real-time, our approach streamlines the assembly of agentic systems, facilitating scalable, adaptive and cost-effective reuse of resources while overcoming the bottlenecks of conventional discovery techniques. Empirical evaluation with Claude 3.5 Sonnet v2 across five benchmarking datasets shows that our online knapsack based composer consistently lies on the Pareto frontier, achieving higher success rates at significantly lower component costs compared to our baselines. In the single-agent setup, we observe a success rate improvement of up to 31.6% in comparison to the top performing retrieval baseline. Our online knapsack composer achieves comparable success rates to the setting wherein agent is equipped with all the tools in the tool inventory, but with 98.5% reduction in costs. Our approach also applies to the composition of multi-agent systems, boosting the success rate from 37% (top performance among the retrieval baselines) to 87%, when agents are selected from an agent inventory of 100+ agents. These results demonstrate that our approach significantly outperforms retrieval-based approaches. The substantial performance gap confirms the robust adaptability of our method across diverse domains and budget constraints.



Paperid:3652
Authors:Yuxuan Song, Zhe Zhang, Yu Pei, Jingjing Gong, Qiying Yu, Zheng Zhang, Mingxuan Wang, Hao Zhou, Jingjing Liu, Wei-Ying Ma
Abstract:
Generative modeling of discrete variables is challenging yet crucial for applications in natural language processing and biological sequence design. We introduce the Shortlisting Model (\method), a novel simplex-based diffusion model inspired by progressive candidate pruning. SLM operates on simplex centroids, reducing generation complexity and enhancing scalability. Additionally, SLM incorporates a flexible implementation of classifier-free guidance, enhancing unconditional generation performance. Extensive experiments on DNA promoter and enhancer design, protein design, character-level and large-vocabulary language modeling demonstrate the competitive performance and strong potential of SLM. Our code can be found at https://anonymous.4open.science/r/SLM-pro-1208



Paperid:3653
Authors:Yuang Qi, Na Zhao, Qiyi Yao, Benlong Wu, Weiming Zhang, Nenghai Yu, Kejiang Chen
Abstract:
Recent provably secure linguistic steganography (PSLS) methods rely on mainstream autoregressive language models (ARMs) to address historically challenging tasks, that is, to disguise covert communication as ``innocuous'' natural language communication. However, due to the characteristic of sequential generation of ARMs, the stegotext generated by ARM-based PSLS methods will produce serious error propagation once it changes, making existing methods unavailable under an active tampering attack. To address this, we propose a robust provably secure linguistic steganography with diffusion language models (DMs). Unlike ARMs, DMs can generate text in partial parallel manner, allowing us to find robust positions for steganographic embedding that can be combined with error-correcting codes. Furthermore, we introduce an error correction strategies, including pseudo-random error correction and neighborhood search correction, during steganographic extraction. Theoretical proof and experimental results demonstrate that our method is secure and robust. It can resist token ambiguity in stegotext segmentation and, to some extent, withstand token-level attacks of insertion, deletion, and substitution.



Paperid:3654
Authors:Gabriel Franco, Mark Crovella
Abstract:
The attention mechanism plays a central role in the computations performed by transformer-based models, and understanding the reasons why heads attend to specific tokens can aid in interpretability of language models. Although considerable work has shown that models construct low-dimensional feature representations, little work has explicitly tied low-dimensional features to the attention mechanism itself. In this paper we work to bridge this gap by presenting methods for identifying attention-causal communication, meaning low-dimensional features that are written into and read from tokens, and that have a provable causal relationship to attention patterns. Our methods do not make use of counterfactual inputs to the model, but rather work by joint analysis of model weights and model representations to isolate signals -- attention-causal features. These signals serve as communication channels from upstream model components to the attention head. We show that by identifying the signals present when a model processes a prompt, we can perform circuit discovery that is more precise and efficient than previous methods. Further, we show that signals can uncover unexplored mechanisms at work in the model, including a surprising degree of global coordination across attention heads.



Paperid:3655
Authors:Zhongwei Wan, Alex Dou, Che Liu, Yu Zhang, Dongfei Cui, Qinjian Zhao, Hui Shen, Jing Xiong, Yi Xin, Yifan Jiang, Chaofan Tao, Yangfan He, Mi Zhang, Shen Yan
Abstract:
Multimodal large language models (MLLMs) have shown promising capabilities in reasoning tasks, yet still struggle significantly with complex problems requiring explicit self-reflection and self-correction, especially compared to their unimodal text-based counterparts. Existing reflection methods are simplistic and struggle to generate meaningful, instructive feedback, as the reasoning ability and knowledge limits of pre-trained models are largely fixed during initial training. To overcome these challenges, we propose \textit{multimodal \textbf{S}elf-\textbf{R}eflection enhanced reasoning with Group Relative \textbf{P}olicy \textbf{O}ptimization} \textbf{SRPO}, a two-stage reflection-aware reinforcement learning (RL) framework explicitly designed to enhance multimodal LLM reasoning. In the first stage, we construct a high-quality, reflection-focused dataset under the guidance of an advanced MLLM, which generates reflections based on initial responses to help the policy model to learn both reasoning and self-reflection. In the second stage, we introduce a novel reward mechanism within the GRPO framework that encourages concise and cognitively meaningful reflection while avoiding redundancy. Extensive experiments across multiple multimodal reasoning benchmarks—including MathVista, MathVision, Mathverse, and MMMU-Pro—using Qwen-2.5-VL-7B and Qwen-2.5-VL-32B demonstrate that SRPO significantly outperforms state-of-the-art models, achieving notable improvements in both reasoning accuracy and reflection quality.



Paperid:3656
Authors:Larkin Liu, Jalal Etesami
Abstract:
We explore the use of expert-guided bandit learning, which we refer to as online mixture-of-experts (OMoE). In this setting, given a context, a candidate committee of experts must determine how to aggregate their outputs to achieve optimal results in terms of aggregate accuracy. We propose three algorithms to address this problem. The first algorithm combines aggregate voting with UCB-driven successive elimination, efficiently pruning suboptimal exploration actions. The second algorithm employs an online weighted-majority-voting mechanism, leveraging the respective voting power of each expert proportional to their predictive power. We derive theoretical guarantees for the regret properties in the bandit setting under ideal circumstances, and empirical results are provided accordingly. As a modern study on applications, these methods are applied to the online fine-tuning of a set of expert large language models (LLMs), where after each response, the generative LLM dynamically reweights its set of experts and/or selects the optimal committee of experts to generate the most accurate response. Our results introduce new methodologies and no-regret guarantees for combining multiple experts to improve on the performance of the an aggregate model overall.



Paperid:3657
Authors:Aloni Cohen
Abstract:
Are there any conditions under which a generative model’s outputs are guaranteed not to infringe the copyrights of its training data? This is the question of "provable copyright protection" first posed by Vyas, Kakade, and Barak [ICML 2023]. They definenear access-freeness (NAF)and propose it as sufficient for protection. This paper revisits the question and establishes new foundations for provable copyright protection---foundations that are firmer both technically and legally. First, we show that NAF alone does not prevent infringement. In fact, NAF models can enable verbatim copying, a blatant failure of copy protection that we dub beingtainted. Then, we introduce ourblameless copy protection frameworkfor defining meaningful guarantees, and instantiate it withclean-room copy protection. Clean-room copy protection allows a user to control their risk of copying by behaving in a way that is unlikely to copy in a counterfactual "clean-room setting." Finally, we formalize a common intuition about differential privacy and copyright by proving that DP implies clean-room copy protection when the dataset isgolden, a copyright deduplication requirement.



Paperid:3658
Authors:Chenshuang Zhang, Kang Zhang, Joon Son Chung, In So Kweon, Junmo Kim, Chengzhi Mao
Abstract:
Distinguishing visually similar objects by their motion remains a critical challenge in computer vision. Although supervised trackers show promise, contemporary self-supervised trackers struggle when visual cues become ambiguous, limiting their scalability and generalization without extensive labeled data. We find that pre-trained video diffusion models inherently learn motion representations suitable for tracking without task-specific training. This ability arises because their denoising process isolates motion in early, high-noise stages, distinct from later appearance refinement. Capitalizing on this discovery, our self-supervised tracker significantly improves performance in distinguishing visually similar objects, a underexplored failure point for existing methods. Our method achieves up to a 6-point improvement over recent self-supervised approaches on established benchmarks and our newly introduced tests focused on tracking visually similar items. Visualizations confirm that these diffusion-derived motion representations enable robust tracking of even identical objects across challenging viewpoint changes and deformations.



Paperid:3659
Authors:Xiaogang Jia, Qian Wang, Anrui Wang, Han Wang, Balázs Gyenes, Emiliyan Gospodinov, Xinkai Jiang, Ge Li, Hongyi Zhou, Weiran Liao, Xi Huang, Maximilian Beck, Moritz Reuss, Rudolf Lioutikov, Gerhard Neumann
Abstract:
Robotic manipulation systems benefit from complementary sensing modalities, where each provides unique environmental information.Point clouds capture detailed geometric structure, while RGB images provide rich semantic context. Current point cloud methods struggle to capture fine-grained detail, especially for complex tasks, which RGB methods lack geometric awareness, which hinders their precision and generalization. We introduce PointMapPolicy, a novel approach that conditions diffusion policies on structured grids of points without downsampling. The resulting data type makes it easier to extract shape and spatial relationships from observations, and can be transformed between reference frames. Yet due to their structure in a regular grid, we enable the use of established computer vision techniques directly to 3D data. Using xLSTM as a backbone, our model efficiently fuses the point maps with RGB data for enhanced multi-modal perception.Through extensive experiments on the RoboCasa and CALVIN benchmarks and real robot evaluations, we demonstrate that our method achieves state-of-the-art performance across diverse manipulation tasks. The overview and demos are available on our project page: https://point-map.github.io/Point-Map/



Paperid:3660
Authors:Kuan Zhang, Chengliang Chai, Jingzhe Xu, Chi Zhang, Han Han, Ye Yuan, Guoren Wang, Lei Cao
Abstract:
Abstract:Recent studies indicate that deep neural networks degrade in generalization performance under noisy supervision. Existing methods focus on isolating clean subsets or correcting noisy labels, facing limitations such as high computational costs, heavy hyperparameter tuning process, and coarse-grained optimization. To address these challenges, we propose a novel two-stage noisy learning framework that enables instance-level optimization through a dynamically weighted loss function, avoiding hyperparameter tuning. To obtain stable and accurate information about noise modeling, we introduce a simple yet effective metric, termed $\textit{wrong event}$, which dynamically models the cleanliness and difficulty of individual samples while maintaining computational costs. Our framework first collects $\textit{wrong event}$ information and builds a strong base model. Then we perform noise-robust training on the base model, using a probabilistic model to handle the $\textit{wrong event}$ information of samples. Experiments on six synthetic and real-world LNL benchmarks demonstrate our method surpasses state-of-the-art methods in performance, achieves a nearly 75\% reduction in storage and computational time, strongly improving model scalability. Our code is available at https://anonymous.4open.science/r/IDO-noisy_label/.



Paperid:3661
Authors:Parvin Nazari, Bojian Hou, Davoud Ataee Tarzanagh, Li Shen, George Michailidis
Abstract:
Online bilevel optimization (OBO) is a powerful framework for machine learning problems where both outer and inner objectives evolve over time, requiring dynamic updates. Current OBO approaches rely on deterministic \textit{window-smoothed} regret minimization, which may not accurately reflect system performance when functions change rapidly. In this work, we introduce a novel search direction and show that both first- and zeroth-order (ZO) stochastic OBO algorithms leveraging this direction achieve sublinear {stochastic bilevel regret without window smoothing}. Beyond these guarantees, our framework enhances efficiency by: (i) reducing oracle dependence in hypergradient estimation, (ii) updating inner and outer variables alongside the linear system solution, and (iii) employing ZO-based estimation of Hessians, Jacobians, and gradients. Experiments on online parametric loss tuning and black-box adversarial attacks validate our approach.



Paperid:3662
Authors:Emil Biju, Shayan Talaei, Zhemin Huang, Mohammadreza Pourreza, Azalia Mirhoseini, Amin Saberi
Abstract:
Large reasoning models (LRMs) excel at complex reasoning tasks but typically generate lengthy sequential chains-of-thought, resulting in long inference times before arriving at the final answer. To address this challenge, we introduce SPRINT, a novel post-training and inference-time framework designed to enable LRMs to dynamically identify and exploit opportunities for parallelization during their reasoning process. SPRINT incorporates an innovative data curation pipeline that reorganizes natural language reasoning trajectories into structured rounds of long-horizon planning and parallel execution. By fine-tuning LRMs on a small amount of such curated data, the models learn to dynamically identify independent subtasks within extended reasoning processes and effectively execute them in parallel. Through extensive evaluations, we show that the models fine-tuned with the SPRINT framework match the performance of reasoning models on complex domains such as mathematics while generating up to 39% fewer sequential tokens on problems requiring more than 8000 output tokens. Finally, we observe consistent results transferred to two out-of-distribution tasks of GPQA and Countdown with up to 45% and 65% reduction in average sequential tokens for longer reasoning trajectories, while achieving the performance of the fine-tuned reasoning model.



Paperid:3663
Authors:Gerard Ben Arous, Murat Erdogdu, Nuri Mert Vural, Denny Wu
Abstract:
Abstract:We study the optimization and sample complexity of gradient-based training of a two-layer neural network with quadratic activation function in the high-dimensional regime, where the data is generated as $y \propto \sum_{j=1}^{r}\lambda_j \sigma\left(\langle \boldsymbol{\theta_j}, \boldsymbol{x}\rangle\right), \boldsymbol{x} \sim \mathcal{N}(0,\boldsymbol{I}_d)$, where $\sigma$ is the 2nd Hermite polynomial, and $\lbrace \boldsymbol{\theta}_j \rbrace _{j=1}^{r} \subset \mathbb{R}^d$ are orthonormal signal directions. We consider the extensive-width regime $r \asymp d^\beta$ for $\beta \in (0, 1)$, and assume a power-law decay on the (non-negative) second-layer coefficients $\lambda_j\asymp j^{-\alpha}$ for $\alpha \geq 0$. We provide a sharp analysis of the SGD dynamics in the feature learning regime, for both the population limit and the finite-sample (online) discretization, and derive scaling laws for the prediction risk that highlight the power-law dependencies on the optimization time, the sample size, and the model width. Our analysis combines a precise characterization of the associated matrix Riccati differential equation with novel matrix monotonicity arguments to establish convergence guarantees for the infinite-dimensional effective dynamics.



Paperid:3664
Authors:Max Weltevrede, Moritz Zanger, Matthijs Spaan, Wendelin Boehmer
Abstract:
In the zero-shot policy transfer setting in reinforcement learning, the goal is to train an agent on a fixed set of training environments so that it can generalise to similar, but unseen, testing environments. Previous work has shown that policy distillation after training can sometimes produce a policy that outperforms the original in the testing environments. However, it is not yet entirely clear why that is, or what data should be used to distil the policy. In this paper, we prove, under certain assumptions, a generalisation bound for policy distillation after training. The theory provides two practical insights: for improved generalisation, you should 1) train an ensemble of distilled policies, and 2) distil it on as much data from the training environments as possible. We empirically verify that these insights hold in more general settings, when the assumptions required for the theory no longer hold. Finally, we demonstrate that an ensemble of policies distilled on a diverse dataset can generalise significantly better than the original agent.



Paperid:3665
Authors:Yuyang Li, Wenxin Du, Chang Yu, Puhao Li, Zihang Zhao, Tengyu Liu, Chenfanfu Jiang, Yixin Zhu, Siyuan Huang
Abstract:
Tactile sensing is crucial for achieving human-level robotic capabilities in manipulation tasks. VBTSs have emerged as a promising solution, offering high spatial resolution and cost-effectiveness by sensing contact through camera-captured deformation patterns of elastic gel pads. However, these sensors' complex physical characteristics and visual signal processing requirements present unique challenges for robotic applications. The lack of efficient and accurate simulation tools for VBTS has significantly limited the scale and scope of tactile robotics research. Here we present Taccel, a high-performance simulation platform that integrates IPC and ABD to model robots, tactile sensors, and objects with both accuracy and unprecedented speed, achieving an 18-fold acceleration over real-time across thousands of parallel environments. Unlike previous simulators that operate at sub-real-time speeds with limited parallelization, Taccel provides precise physics simulation and realistic tactile signals while supporting flexible robot-sensor configurations through user-friendly APIs. Through extensive validation in object recognition, robotic grasping, and articulated object manipulation, we demonstrate precise simulation and successful sim-to-real transfer. These capabilities position Taccel as a powerful tool for scaling up tactile robotics research and development. By enabling large-scale simulation and experimentation with tactile sensing, Taccel accelerates the development of more capable robotic systems, potentially transforming how robots interact with and understand their physical environment.



Paperid:3666
Authors:Xu Yang, Chen Liu, Ying Wei
Abstract:
Selecting a compact set of visual instruction–following data has emerged as an effective way to align large multimodal models with human intentions while avoiding the cost of full-dataset training. Yet we observe that both full-data and state-of-the-art-method selected-data instruction tuning can inherit underlying dataset biases, leading to biased model behavior due to position bias and spurious correlations. We introduce ARDS, a robustness-aware targeted visual instruction-selection framework that explicitly targets these weaknesses, sidestepping the need for access to downstream data or time-consuming gradient computation. Specifically, we first generate the worst-case evaluation subgroups through visual and textual task-specific perturbations. The robust training mixture is then constructed by prioritizing the samples that are semantically closer to these subgroups in the rich multimodal embedding space. Extensive experiments show that ARDS substantially boosts both robustness and data efficiency for visual instruction tuning. We also showcase that the robust mixtures produced with a smaller model transfer effectively to larger architectures.



Paperid:3667
Authors:Rohit Jayanti, Swayam Agrawal, Vansh Garg, Siddharth Tourani, Muhammad Haris Khan, Sourav Garg, Madhava Krishna
Abstract:
Abstract:Segment matching is an important intermediate task in computer vision that establishes correspondences between semantically or geometrically coherent regions across images. Unlike keypoint matching, which focuses on localized features, segment matching captures structured regions, offering greater robustness to occlusions, lighting variations, and viewpoint changes. In this paper, we leverage the spatial understanding of 3D foundation models to tackle wide-baseline segment matching, a challenging setting involving extreme viewpoint shifts. We propose an architecture that uses the inductive bias of these 3D foundation models to match segments across image pairs with up to $180^\circ$ rotation. Extensive experiments show that our approach outperforms state-of-the-art methods, including the SAM2 video propagator and local feature matching methods, by upto 30\% on the AUPRC metric, on ScanNet++ and Replica datasets. We further demonstrate benefits of the proposed model on relevant downstream tasks, including 3D instance segmentation and image-goal navigation.



Paperid:3668
Authors:Zhihao Li, Yufei Wang, Heliang Zheng, Yihao Luo, Bihan Wen
Abstract:
Abstract:High-fidelity 3D object synthesis remains significantly more challenging than 2D image generation due to the unstructured nature of mesh data and the cubic complexity of dense volumetric grids. Existing two-stage pipelines—compressing meshes with a VAE (using either 2D or 3D supervision), followed by latent diffusion sampling—often suffer from severe detail loss caused by inefficient representations and modality mismatches introduced in VAE. We introduce **SparC**, a unified framework that combines a sparse deformable marching cubes representation **SparseCubes** with a novel encoder **SparConv-VAE**. SparseCubes converts raw meshes into high-resolution ($1024^3$) surfaces with arbitrary topology by scattering signed distance and deformation fields onto a sparse cube, allowing differentiable optimization. SparConv-VAE is the first modality-consistent variational autoencoder built entirely upon sparse convolutional networks, enabling efficient and near-lossless 3D reconstruction suitable for high-resolution generative modeling through latent diffusion. SparC achieves state-of-the-art reconstruction fidelity on challenging inputs, including open surfaces, disconnected components, and intricate geometry. It preserves fine-grained shape details, reduces training and inference cost, and integrates naturally with latent diffusion models for scalable, high-resolution 3D generation.



Paperid:3669
Authors:Zeqian Li, Shangzhe Di, Zhonghua Zhai, Weilin Huang, Yanfeng Wang, Weidi Xie
Abstract:
This paper presents a computational model for universal video temporal grounding, which accurately localizes temporal moments in videos based on natural language queries (e.g., questions or descriptions). Unlike existing methods that are often limited to specific video domains or durations, we proposeUniTime, a robust and universal video grounding model leveraging the strong vision-language understanding capabilities of generative Multi-modal Large Language Models (MLLMs).Our model effectively handles videos of diverse views, genres, and lengths while comprehending complex language queries.The key contributions include:(i) We consider steering strong MLLMs for temporal grounding in videos. To enable precise timestamp outputs, we incorporate temporal information by interleaving timestamp tokens with video tokens.(ii) By training the model to handle videos with different input granularities through adaptive frame scaling, our approach achieves robust temporal grounding for both short and long videos.(iii) Comprehensive experiments show that UniTime outperforms state-of-the-art approaches in both zero-shot and dataset-specific finetuned settings across five public temporal grounding benchmarks.(iv) When employed as a preliminary moment retriever for long-form video question-answering (VideoQA), UniTime significantly improves VideoQA accuracy, highlighting its value for complex video understanding tasks.



Paperid:3670
Authors:Yichen Li, Yijing Shan, YI LIU, Haozhao Wang, Cheng Wang, wangshi.ww, Yi Wang, Ruixuan Li
Abstract:
The current Federated Recommendation System (FedRS) focuses on personalized recommendation services and assumes clients are personalized IoT devices (e.g., Mobile phones). In this paper, we deeply dive into new but practical FedRS applications within the joint venture ecosystem. Subsidiaries engage as participants with their users and items. However, in such a situation, merely exchanging item embedding is insufficient, as user bases always exhibit both overlaps and exclusive segments, demonstrating the complexity of user information. Meanwhile, directly uploading user information is a violation of privacy and unacceptable. To tackle the above challenges, we propose an efficient and privacy-enhanced federated recommendation for the joint venture ecosystem (FR-JVE) that each client transfers more common knowledge from other clients with a distilled user's \textit{rating preference} from the local dataset. More specifically, we first transform the local data into a new format and apply model inversion techniques to distill the rating preference with frozen user gradients before the federated training. Then, a bridge function is employed on each client side to align the local rating preference and aggregated global preference in a privacy-friendly manner. Finally, each client matches similar users to make a better prediction for overlapped users. From a theoretical perspective, we analyze how effectively FR-JVE can guarantee user privacy. Empirically, we show that FR-JVE achieves superior performance compared to state-of-the-art methods.



Paperid:3671
Authors:Wenjun Huang, Ziteng Cui, Yinqiang Zheng, Yirui He, Tatsuya Harada, Mohsen Imani
Abstract:
We introduce Dr. RAW, a unified and tuning-efficient framework for high-level computer vision tasks directly operating on camera RAW data. Unlike previous approaches that optimize image signal processing (ISP) pipelines and fully fine-tune networks for each task, Dr. RAW achieves state-of-the-art performance with minimal parameter updates. At the input stage, we apply lightweight pre-processing modules, sensor and illumination mapping, followed by re-mosaicing, to mitigate data inconsistencies stemming from sensor variation and lighting. At the network level, we introduce task-specific adaptation through two modules: Sensor Prior Prompts (SPP) and Low-Rank Adaptation (LoRA). SPP injects sensor-aware conditioning into the network via learnable prompts derived from imaging priors, while LoRA enables efficient task-specific tuning by updating only low-rank matrices in key backbone layers. Despite minimal tuning, our method delivers superior results across four RAW-based tasks (object detection, semantic segmentation, instance segmentation, and pose estimation) on nine datasets encompassing low-light and over-exposed conditions. By harnessing the intrinsic physical cues of RAW data alongside parameter-efficient techniques, our method advances RAW-based vision systems, achieving both high accuracy and computational economy. We will release our source code.



Paperid:3672
Authors:Xiaoyi Zhang, Zhaoyang Jia, Zongyu Guo, Jiahao Li, Bin Li, Houqiang Li, Yan Lu
Abstract:
Abstract:Long-form video understanding presents significant challenges due to extensive temporal-spatial complexity and the difficulty of question answering under such extended contexts. While Large Language Models (LLMs) have demonstrated considerable advancements in video analysis capabilities and long context handling, they continue to exhibit limitations when processing information-dense hour-long videos. To overcome such limitations, we propose the $\textbf{D}eep \ \textbf{V}ideo \ \textbf{D}iscovery \ (\textbf{DVD})$ agent to leverage an $\textit{agentic search}$ strategy over segmented video clips. Different from previous video agents manually designing a rigid workflow, our approach emphasizes the autonomous nature of agents.By providing a set of search-centric tools on multi-granular video database,our DVD agent leverages the advanced reasoning capability of LLM to plan on its current observation state, strategically selects tools, formulates appropriate parameters for actions, and iteratively refines its internal reasoning in light of the gathered information.We perform comprehensive evaluation on multiple long video understanding benchmarks that demonstrates the advantage of the entire system design. Our DVD agent achieves SOTA performance, significantly surpassing prior works by a large margin on the challenging LVBench dataset. Comprehensive ablation studies and in-depth tool analyses are also provided, yielding insights to further advance intelligent agents tailored for long-form video understanding tasks. The code will be released later.



Paperid:3673
Authors:Hanlin Zhu, Shibo Hao, Zhiting Hu, Jiantao Jiao, Stuart J Russell, Yuandong Tian
Abstract:
Abstract:Large Language Models (LLMs) have demonstrated remarkable performance in many applications, including challenging reasoning problems via chain-of-thoughts (CoTs) techniques that generate ``thinking tokens'' before answering the questions. While existing theoretical works demonstrate that CoTs with discrete tokens boost the capability of LLMs, recent work on continuous CoTs lacks a theoretical understanding of why it outperforms discrete counterparts in various reasoning tasks such as directed graph reachability, a fundamental graph reasoning problem that includes many practical domain applications as special cases. In this paper, we prove that a two-layer transformer with $D$ steps of continuous CoTs can solve the directed graph reachability problem, where $D$ is the diameter of the graph, while the best known result of constant-depth transformers with discrete CoTs requires $O(n^2)$ decoding steps where $n$ is the number of vertices ($D



Paperid:3674
Authors:Juan Chen, Honglin liu, Yingying Ao, Ting Zhang, Yan Huang, Xudong Liu, Biao Li, Jintao Fang
Abstract:
Vision-language models heavily rely on visual representations, yet ensuring its efficiency remains a critical challenge. Most existing approaches focus on reducing visual tokens either at the visual encoder phase or during the LLM decoder stage. Inspired by human visual cognition, where an initial global glance precedes focused attention on semantically salient regions, we introduce Glance2Gaze, a cognitively inspired framework that mimics the human two-stage attention process. The framework consists of two key components: the Glance Fusion module, which integrates multi-layer vision transformer features with text-aware attention to generate a semantically enriched global representation, and the Gaze Compression module, which utilizes a novel query-guided mechanism to selectively compress visual tokens based on their semantic relevance. Experimental results on widely adopted benchmarks demonstrate that Glance2Gaze outperforms existing methods, achieving superior performance with equal or lower computational cost. Furthermore, it generalizes well to high-resolution and video scenarios, showcasing robust and scalable efficiency improvements in VLMs.



Paperid:3675
Authors:Seunghee Ryu, Donghoon Kwon, Seongjin Choi, Aryan Deshwal, Seungmo Kang, Carolina Osorio
Abstract:
We introduce BO4Mob, a new benchmark framework for high-dimensional Bayesian Optimization (BO), driven by the challenge of origin-destination (OD) travel demand estimation in large urban road networks. Estimating OD travel demand from limited traffic sensor data is a difficult inverse optimization problem, particularly in real-world, large-scale transportation networks. This problem involves optimizing over high-dimensional continuous spaces where each objective evaluation is computationally expensive, stochastic, and non-differentiable. BO4Mob comprises five scenarios based on real-world San Jose, CA road networks, with input dimensions scaling up to 10,100. These scenarios utilize high-resolution, open-source traffic simulations that incorporate realistic nonlinear and stochastic dynamics. We demonstrate the benchmark's utility by evaluating four optimization methods: three state-of-the-art BO algorithms and one non-BO baseline. This benchmark is designed to support both the development of scalable optimization algorithms and their application for the design of data-driven urban mobility models, including high-resolution digital twins of metropolitan road networks. Code and documentation are available at https://github.com/UMN-Choi-Lab/BO4Mob



Paperid:3676
Authors:Hao Zhong, Muzhi Zhu, Zongze Du, Zheng Huang, Canyu Zhao, Mingyu Liu, Wen Wang, Hao Chen, Chunhua Shen
Abstract:
Enabling intelligent systems to simultaneously process and reason over information from multiple modalities—such as text, video, and audio—in complex real-world scenarios, while performing planning and precise decision-making, has long been regarded as the ultimate goal in the field of artificial intelligence.Recently, with the rapid advancement of multimodal pretraining and supervised fine-tuning (SFT) techniques, a series of Omni-modal models have emerged, bringing us closer to this vision.However, current Omni-modal models still exhibit significant shortcomings in understanding and reasoning over long video and audio sequences, as well as in fine-grained, pixel-level comprehension tasks.On the other hand, while reinforcement learning (RL) has achieved remarkable success in enhancing the reasoning capabilities of MLLMs in pure-text domains, its application within Omni-modal models remains at an early stage.This is due to two major challenges: first, the lack of effective reasoning datasets and task formulations for training Omni-modal models; second, the effectiveness of existing RL techniques in Omni-modal settings has not yet been thoroughly validated.To address these issues, we focus on two of the most challenging tasks: Referring Audio-Visual Segmentation (RefAVS) and Referring Video Object Segmentation (REVOS).We decouple the complex video understanding task into two key subtasks: (1) identifying long-range keyframes in videos, and (2) generating task-specific re-captions.We propose a fully RL-based framework that leverages large-scale existing datasets to jointly optimize both capabilities of Omni-modal models.Experimental results demonstrate that our model achieves state-of-the-art performance on both the RefAVS and REVOS benchmarks.Furthermore, we show that our RL approach brings significant improvements on other general-purpose understanding tasks as well.



Paperid:3677
Authors:Yunwei Ren, Jason Lee
Abstract:
Abstract:The information exponent (Ben Arous et al. [2021]) and its extensions --- which are equivalent to the lowest degree in the Hermite expansion of the link function (after a potential label transform) for Gaussian single-index models --- have played an important role in predicting the sample complexity of online stochastic gradient descent (SGD) in various learning tasks. In this work, we demonstrate that, for multi-index models, focusing solely on the lowest degree can miss key structural details of the model and result in suboptimal rates. Specifically, we consider the task of learning target functions of form $f_*(x) = \sum_{k=1}^{P} \phi(v_k^* \cdot x)$, where $P \le d$, the ground-truth directions $\\{ v_k^* \\}_{k=1}^P$ are orthonormal, and the information exponent of $\phi$ is $L$. Based on the theory of information exponent, when $L = 2$, only the relevant subspace (not the exact directions) can be recovered due to the rotational invariance of the second-order terms, and when $L > 2$, recovering the directions using online SGD require $\tilde{O}(P d^{L-1})$ samples. In this work, we show that by considering both second- and higher-order terms, we can first learn the relevant space using the second-order terms, and then the exact directions using the higher-order terms, and the overall sample and complexity of online SGD is $\tilde{O}( P d P^{L-2} )$.



Paperid:3678
Authors:Zhen Zhang, Javen Qinfeng Shi, Wee Sun Lee
Abstract:
The Asymmetric Traveling Salesman Problem (ATSP) ranks among the most fundamental and notoriously difficult problems in combinatorial optimization. We propose a novel continuous relaxation framework for the Asymmetric Traveling Salesman Problem (ATSP) by leveraging differentiable constraints that encourage acyclic structures and valid permutations. Our approach integrates a differentiable trace-based Directed Acyclic Graph (DAG) constraint with a doubly stochastic matrix relaxation of the assignment problem, enabling gradient-based optimization over soft permutations. We develop a projected exponentiated gradient method with adaptive step size to minimize tour cost while satisfying the relaxed constraints. To recover high-quality discrete tours, we introduce a greedy post-processing procedure that iteratively corrects subtours using cost-aware cycle merging. Our method achieves state-of-the-art performance on standard asymmetric TSP benchmarks and demonstrates competitive scalability and accuracy, particularly on large or asymmetric instances where heuristic solvers such as LKH-3 struggle.



Paperid:3679
Authors:Phuc Tran, Van Vu, Nisheeth K. Vishnoi
Abstract:
Abstract:A central challenge in machine learning is to understand how noise or measurement errors affect low-rank approximations, particularly in the spectral norm. This question is especially important in differentially private low-rank approximation, where one aims to preserve the top-$p$ structure of a data-derived matrix while ensuring privacy. Prior work often analyzes Frobenius norm error or changes in reconstruction quality, but these metrics can over- or under-estimate true subspace distortion. The spectral norm, by contrast, captures worst-case directional error and provides the strongest utility guarantees. We establish new high-probability spectral-norm perturbation bounds for symmetric matrices that refine the classical Eckart--Young--Mirsky theorem and explicitly capture interactions between a matrix $A \in \mathbb{R}^{n \times n}$ and an arbitrary symmetric perturbation $E$. Under mild eigengap and norm conditions, our bounds yield sharp estimates for $\| (A + E)_p - A_p \|$, where $A_p$ is the best rank-$p$ approximation of $A$, with improvements of up to a factor of $\sqrt{n}$. As an application, we derive improved utility guarantees for differentially private PCA, resolving an open problem in the literature. Our analysis relies on a novel contour bootstrapping method from complex analysis and extends it to a broad class of spectral functionals, including polynomials and matrix exponentials. Empirical results on real-world datasets confirm that our bounds closely track the actual spectral error under diverse perturbation regimes.



Paperid:3680
Authors:Harshita Chopra, Chirag Shah
Abstract:
Effective decision-making and problem-solving in conversational systems require the ability to identify and acquire missing information through targeted questioning. A key challenge lies in efficiently narrowing down a large space of possible outcomes by posing questions that minimize uncertainty. To address this, we introduce a novel framework that leverages Large Language Models (LLMs) to generate information-seeking questions, with Monte Carlo Tree Search (MCTS) to strategically select questions that maximize information gain, as a part of inference-time planning. Our primary contribution includes a hierarchical feedback mechanism that exploits past interaction patterns to guide future strategy. Specifically, each new problem is mapped to a cluster based on semantic similarity, and our UCT (Upper Confidence bound for Trees) formulation employs a cluster-specific bonus reward to prioritize successful question trajectories that have proven effective for similar problems in the past. Extensive empirical evaluation across medical diagnosis and technical troubleshooting domains shows that our method achieves an average of 12\% improvement in success rates and about 10x reduction in the number of LLM calls made for planning per conversation, compared to the state of the art. An additional 8\% gain in success rate is observed on average when we start with a constrained set of possibilities. Our results underscore the efficacy of feedback-aware MCTS in enhancing information-seeking in goal-oriented dialogues.



Paperid:3681
Authors:Sai Sumedh R. Hindupur, Ekdeep S Lubana, Thomas Fel, Demba Ba
Abstract:
Sparse Autoencoders (SAEs) are widely used to interpret neural networks by identifying meaningful concepts from their representations. However, do SAEs truly uncover all concepts a model relies on, or are they inherently biased toward certain kinds of concepts? We introduce a unified framework that recasts SAEs as solutions to a bilevel optimization problem, revealing a fundamental challenge: each SAE imposes structural assumptions about how concepts are encoded in model representations, which in turn shapes what it can and cannot detect. This means different SAEs are not interchangeable -- switching architectures can expose entirely new concepts or obscure existing ones. To systematically probe this effect, we evaluate SAEs across a spectrum of settings: from controlled toy models that isolate key variables, to semi-synthetic experiments on real model activations and finally to large-scale, naturalistic datasets. Across this progression, we examine two fundamental properties that real-world concepts often exhibit: heterogeneity in intrinsic dimensionality (some concepts are inherently low-dimensional, others are not) and nonlinear separability. We show that SAEs fail to recover concepts when these properties are ignored, and we design a new SAE that explicitly incorporates both, enabling the discovery of previously hidden concepts and reinforcing our theoretical insights. Our findings challenge the idea of a universal SAE and underscores the need for architecture-specific choices in model interpretability. Overall, we argue an SAE does not just reveal concepts -- it determines what can be seen at all.



Paperid:3682
Authors:Nelson Higuera, Thomas Eiter, Sota Moriyama, Katsumi Inoue
Abstract:
Integrating logical constraints into object detection models for autonomous driving (AD) is a promising way to enhance their compliance with rules and thereby increase the safety of the system. T-norms have been utilized to calculate the constrained loss, i.e., the violations of logical constraints as losses. While prior works have statically selected a few t-norms, we conduct an extensive experimental study to identify the most effective choices, as suboptimal t-norms can lead to undesired model behavior.To this end, we present MOD-ECL, a neurosymbolic framework that implements a wide range of t-norms and applies them in an adaptive manner. It includes an algorithm that selects well-performing t-norms during training and a scheduler that regulates the impact of the constrained loss. We evaluate its effectiveness on the ROAD-R and ROAD-Waymo-R datasets for object detection in AD, using attached common-sense constraints.Our results show that careful selection of parameters is crucial for effective constrained loss behavior. Moreover, our framework not only reduces constraint violations but also, in some cases, improves detection performance. Additionally, our methods offer fine-grained control over the trade-off between accuracy and constraint violation.



Paperid:3683
Authors:Yanchen Luo, ZHIYUAN LIU, Yi Zhao, Sihang Li, Hengxing Cai, Kenji Kawaguchi, Tat-Seng Chua, Yang Zhang, Xiang Wang
Abstract:
3D molecule generation is crucial for drug discovery and material science, requiring models to process complex multi-modalities, including atom types, chemical bonds, and 3D coordinates. A key challenge is integrating these modalities of different shapes while maintaining SE(3) equivariance for 3D coordinates. To achieve this, existing approaches typically maintain separate latent spaces for invariant and equivariant modalities, reducing efficiency in both training and sampling.In this work, we proposeUnified VariationalAuto-Encoder for3DMolecular Latent Diffusion Modeling (UAE-3D), a multi-modal VAE that compresses 3D molecules into latent sequences from a unified latent space, while maintaining near-zero reconstruction error. This unified latent space eliminates the complexities of handling multi-modality and equivariance when performing latent diffusion modeling. We demonstrate this by employing the Diffusion Transformer--a general-purpose diffusion model without any molecular inductive bias--for latent generation. Extensive experiments on GEOM-Drugs and QM9 datasets demonstrate that our method significantly establishes new benchmarks in bothde novoand conditional 3D molecule generation, achieving leading efficiency and quality.



Paperid:3684
Authors:Shen Zhang, Siyuan Liang, Yaning Tan, Zhaowei Chen, Linze Li, Ge Wu, Yuhao Chen, Shuheng Li, Zhenyu Zhao, Caihua Chen, Jiajun Liang, Yao Tang
Abstract:
Abstract:Diffusion transformers (DiTs) struggle to generate images at resolutions higher than their training resolutions. The primary obstacle is that the explicit positional encodings (PE), such as RoPE, need extrapolating to unseen positions which degrades performance when the inference resolution differs from training. In this paper, We propose a Length-Extrapolatable Diffusion Transformer (LEDiT) to overcome this limitation. LEDiT needs no explicit PEs, thereby avoiding PE extrapolation. The key innovation of LEDiT lies in the use of causal attention. We demonstrate that causal attention can implicitly encode global positional information and show that such information facilitates extrapolation. We further introduce a locality enhancement module, which captures fine-grained local information to complement the global coarse-grained position information encoded by causal attention. Experimental results on both conditional and text-to-image generation tasks demonstrate that LEDiT supports up to 4× resolution scaling (e.g., from 256$\times$256 to 512$\times$512), achieving better image quality compared to the state-of-the-art length extrapolation methods. We believe that LEDiT marks a departure from the standard RoPE-based methods and offers a promising insight into length extrapolation.



Paperid:3685
Authors:Jisung Hwang, Jaihoon Kim, Minhyuk Sung
Abstract:
We propose a novel regularization loss that enforces standard Gaussianity, encouraging samples to align with a standard Gaussian distribution. This facilitates a range of downstream tasks involving optimization in the latent space of text-to-image models.We treat elements of a high-dimensional sample as one-dimensional standard Gaussian variables and define a composite loss that combines moment-based regularization in the spatial domain with power spectrum-based regularization in the spectral domain. Since the expected values of moments and power spectrum distributions are analytically known, the loss promotes conformity to these properties. To ensure permutation invariance, the losses are applied to randomly permuted inputs. Notably, existing Gaussianity-based regularizations fall within our unified framework: some correspond to moment losses of specific orders, while the previous covariance-matching loss is equivalent to our spectral loss but incurs higher time complexity due to its spatial-domain computation. We showcase the application of our regularization in generative modeling for test-time reward alignment with a text-to-image model, specifically to enhance aesthetics and text alignment. Our regularization outperforms previous Gaussianity regularization, effectively prevents reward hacking and accelerates convergence.



Paperid:3686
Authors:Yichen Li, Xiuying Wang, Wenchao Xu, Haozhao Wang, Yining Qi, Jiahua Dong, Ruixuan Li
Abstract:
Model-Heterogeneous Federated Learning (Hetero-FL) has attracted growing attention for its ability to aggregate knowledge from heterogeneous models while keeping private data locally. To better aggregate knowledge from clients, ensemble distillation, as a widely used and effective technique, is often employed after global aggregation to enhance the performance of the global model. However, simply combining Hetero-FL and ensemble distillation does not always yield promising results and can make the training process unstable. The reason is that existing methods primarily focus on logit distillation, which, while being model-agnostic with softmax predictions, fails to compensate for the knowledge bias arising from heterogeneous models.To tackle this challenge, we propose a stable and efficient Feature Distillation for model-heterogeneous Federated learning, dubbed FedFD, that can incorporate aligned feature information via orthogonal projection to integrate knowledge from heterogeneous models better. Specifically, a new feature-based ensemble federated knowledge distillation paradigm is proposed. The global model on the server needs to maintain a projection layer for each client-side model architecture to align the features separately. Orthogonal techniques are employed to re-parameterize the projection layer to mitigate knowledge bias from heterogeneous models and thus maximize the distilled knowledge. Extensive experiments show that FedFD achieves superior performance compared to state-of-the-art methods.



Paperid:3687
Authors:Anupam Pani, Yanchao Yang
Abstract:
Abstract:Eye gaze offers valuable cues about attention, short-term intent, and future actions, making it a powerful signal for modeling egocentric behavior. In this work, we propose a gaze-regularized framework that enhances VLMs for two key egocentric understanding tasks: fine-grained future event prediction and current activity understanding.Unlike prior approaches that rely solely on visual inputs or use gaze as an auxiliary input signal , our method uses gaze only during training. We introduce a gaze-regularized attention mechanism that aligns model focus with human visual gaze. This design is flexible and modular, allowing it to generalize across multiple VLM architectures that utilize attention.Experimental results show that our approach improves semantic prediction scores by up to 11 $\%$ for future event prediction and around 7 $\%$ for current activity understanding, compared to the corresponding baseline models trained without gaze regularization. These results highlight the value of gaze-guided training in improving the accuracy and robustness of egocentric VLMs. Overall, this work establishes a foundation for using human gaze to enhance the predictive capabilities of VLMs in real-world scenarios like assistive robots and human-machine collaboration.



Paperid:3688
Authors:Yongxiang Li, Yanglin Feng, Yuan Sun, Dezhong Peng, Xi Peng, Peng Hu
Abstract:
In this paper, we investigate source-free domain adaptation (SFDA) for Visible-Infrared Person Re-Identification (VI-ReID), aiming to adapt a pre-trained source model to an unlabeled target domain without access to source domain data or target domain labels. To tackle this, we propose a novel learning paradigm, Source-Free Domain Adaptation for Visible-Infrared Person Re-Identification (SVIP), which fully exploits the prior knowledge embedded in the source model to guide target domain adaptation. Our method consists of three core mechanisms designed specifically for the source-free setting: (1) the Source Guided Contrastive Learning (SGCL) mechanism, which leverages the discriminative feature space of the fixed source model as a reference to perform contrastive learning on unlabeled target data, thereby preserving discriminative capability without requiring source samples; (2) the Residual Transfer Learning (RTL) mechanism, which learns a residual mapping to adapt the target model’s representations towards the target domain while maintaining the source model’s prior knowledge fixed; and (3) the Structural Consistency Guided Cross-modal Alignment (SCCA) mechanism, which enforces reciprocal structural consistency constraints between visible and infrared modalities in the target domain to identify reliable cross-modal pairs and facilitate robust modality alignment without source supervision. Extensive experiments on benchmark datasets demonstrate that SVIP significantly improves target domain performance and outperforms existing unsupervised VI-ReID approaches under source-free adaptation settings.



Paperid:3689
Authors:Huajie Tan, Yuheng Ji, Xiaoshuai Hao, Xiansheng Chen, Pengwei Wang, Zhongyuan Wang, Shanghang Zhang
Abstract:
Visual reasoning abilities play a crucial role in understanding complex multimodal data, advancing both domain-specific applications and artificial general intelligence (AGI). Existing methods enhance Vision-Language Models (VLMs) through Chain-of-Thought (CoT) supervised fine-tuning using meticulously annotated data. However, this approach may lead to overfitting and cognitive rigidity, limiting the model’s generalization ability under domain shifts and reducing real-world applicability. To overcome these limitations, we propose Reason-RFT, a two-stage reinforcement fine-tuning framework for visual reasoning. First, Supervised Fine-Tuning (SFT) with curated CoT data activates the reasoning potential of VLMs. This is followed by reinforcement learning based on Group Relative Policy Optimization (GRPO), which generates multiple reasoning-response pairs to enhance adaptability to domain shifts. To evaluate Reason-RFT, we reconstructed a comprehensive dataset covering visual counting, structural perception, and spatial transformation, serving as a benchmark for systematic assessment across three key dimensions. Experimental results highlight three advantages: (1) performance enhancement, with Reason-RFT achieving state-of-the-art results and outperforming both open-source and proprietary models; (2) generalization superiority, maintaining robust performance under domain shifts across various tasks; and (3) data efficiency, excelling in few-shot learning scenarios and surpassing full-dataset SFT baselines. Reason-RFT introduces a novel training paradigm for visual reasoning and marks a significant step forward in multimodal research.



Paperid:3690
Authors:Hossein Askari, Yadan Luo, Hongfu Sun, Fred Roosta
Abstract:
Recent advances ininverse problemsolving have increasingly adopted flowpriorsover diffusion models due to their ability to construct straight probability paths from noise to data, thereby enhancing efficiency in both training and inference. However, current flow-based inverse solvers face two primary limitations: (i) they operate directly in pixel space, which demands heavy computational resources for training and restricts scalability to high-resolution images, and (ii) they employ guidance strategies withprior-agnostic posterior covariances, which can weaken alignment with the generative trajectory and degrade posterior coverage. In this paper, we proposeLFlow(Latent Refinement viaFlows), atraining-freeframework for solving linear inverse problems via pretrained latent flow priors. LFlow leverages the efficiency of flow matching to perform ODE sampling in latent space along an optimal path. Furthermore, we introduce a theoretically grounded, time-dependent variance for the latent identity posterior covariance, enabling more effective flow guidance. Experimental results demonstrate that our proposed method outperforms state-of-the-art latent diffusion solvers in reconstruction quality across most tasks.



Paperid:3691
Authors:Hubert Baniecki, Maximilian Muschalik, Fabian Fumagalli, Barbara Hammer, Eyke Hüllermeier, Przemyslaw Biecek
Abstract:
Language-image pre-training (LIP) enables the development of vision-language models capable of zero-shot classification, localization, multimodal retrieval, and semantic understanding. Various explanation methods have been proposed to visualize the importance of input image-text pairs on the model's similarity outputs. However, popular saliency maps are limited by capturing only first-order attributions, overlooking the complex cross-modal interactions intrinsic to such encoders. We introduce faithful interaction explanations of LIP models (FIxLIP) as a unified approach to decomposing the similarity in vision-language encoders. FIxLIP is rooted in game theory, where we analyze how using the weighted Banzhaf interaction index offers greater flexibility and improves computational efficiency over the Shapley interaction quantification framework. From a practical perspective, we propose how to naturally extend explanation evaluation metrics, like the pointing game and area between the insertion/deletion curves, to second-order interaction explanations. Experiments on MS COCO and ImageNet-1k benchmarks validate that second-order methods like FIxLIP outperform first-order attribution methods. Beyond delivering high-quality explanations, we demonstrate the utility of FIxLIP in comparing different models like CLIP vs. SigLIP-2 and ViT-B/32 vs. ViT-L/16.



Paperid:3692
Authors:Sungmin Cha, Kyunghyun Cho
Abstract:
Knowledge distillation (KD) is a core component in the training and deployment of modern generative models, particularly large language models (LLMs). While its empirical benefits are well documented—enabling smaller student models to emulate the performance of much larger teachers—the underlying mechanisms by which KD improves generative quality remain poorly understood.In this work, we present a minimal working explanation of KD in generative modeling. Using a controlled simulation with mixtures of Gaussians, we demonstrate that distillation induces a trade-off between precision and recall in the student model. As the teacher distribution becomes more selective, the student concentrates more probability mass on high-likelihood regions at the expense of coverage—a behavior modulated by a single entropy-controlling parameter.We then validate this effect in a large-scale language modeling setup using the SmolLM2 family of models. Empirical results reveal the same precision–recall dynamics observed in simulation, where precision corresponds to sample quality and recall to distributional coverage.This precision–recall trade-off proves especially beneficial in scenarios where sample quality outweighs diversity, such as instruction tuning or downstream generation. Our analysis provides a simple and general explanation for the effectiveness of KD in generative modeling.



Paperid:3693
Authors:Kang Yang, Gaofeng Dong, Sijie Ji, Wan Du, Mani Srivastava
Abstract:
Synthesizing Radio-Frequency (RF) signal data is essential for wireless sensing applications but remains challenging due to complex interactions, including reflection, diffraction, scattering, and phase-dependent interference. State-of-the-art Neural Radiance Field (NeRF)-based methods achieve high fidelity in RF data synthesis but suffer from long training times and high inference latency. We introduce GSRF, a framework that extends 3D Gaussian Splatting (3DGS) to the RF domain, enabling efficient, high-fidelity RF signal data synthesis. Unlike nanoscale visible light, RF signals with centimeter-scale wavelengths require phase-aware modeling to capture complex propagation effects. GSRF addresses these challenges with three innovations: first, complex-valued 3D Gaussians with a hybrid Fourier-Legendre basis to model directional and phase-dependent radiance; second, orthographic splatting for efficient ray-Gaussian intersection identification; and third, a complex-valued ray tracing algorithm executed on RF-customized CUDA kernels, grounded in wavefront propagation principles, to synthesize RF signal data in real time. Evaluated on data synthesis across various RF technologies, GSRF demonstrates significant improvements in training efficiency, training time, and inference latency.



Paperid:3694
Authors:Yan Wu, Esther Wershof, Sebastian Schmon, Marcel Nassar, Błażej Osiński, Ridvan Eksi, Zichao Yan, Rory Stark, Kun Zhang, Thore Graepel
Abstract:
We introduce a comprehensive framework for perturbation response modeling in single cells, aimed at standardizing benchmarking in this rapidly evolving field. Our approach includes a modular and user-friendly model development and evaluation platform, a collection of diverse perturbational datasets, and a set of metrics designed to fairly compare models and dissect their performance nuances. Through extensive evaluation of both published and baseline models across diverse datasets, we highlight the limitations of widely used models, such as mode collapse. We also demonstrate the importance of rank metrics which complement traditional model fit measures, such as RMSE, for validating model effectiveness. Notably, our results show that while no single model architecture clearly outperforms others, simpler architectures are generally competitive and scale well with larger datasets. Overall, this benchmarking exercise sets new standards for model evaluation, supports robust model development, and advances the potential of these models to use high-throughput genetic and chemical screens for disease target discovery.



Paperid:3695
Authors:David Burt, Renato Berlinghieri, Stephen Bates, Tamara Broderick
Abstract:
Estimating associations between spatial covariates and responses — rather than merely predicting responses — is central to environmental science, epidemiology, and economics. For instance, public health officials might be interested in whether air pollution has a strictly positive association with a health outcome, and the magnitude of any effect. Standard machine learning methods often provide accurate predictions but offer limited insight into covariate-response relationships. And we show that existing methods for constructing confidence (or credible) intervals for associations fail to provide nominal coverage in the face of model misspecification and distribution shift — despite both being essentially always present in spatial problems. We introduce a method that constructs valid frequentist confidence intervals for associations in spatial settings. Our method requires minimal assumptions beyond a form of spatial smoothness. In particular, we do not require model correctness or covariate overlap between training and target locations. Our approach is the first to guarantee nominal coverage in this setting and outperforms existing techniques in both real and simulated experiments.



Paperid:3696
Authors:Daoyuan Chen, Yilun Huang, Xuchen Pan, Jiang Nana, Haibin Wang, Yilei Zhang, Ce Ge, Yushuo Chen, Wenhao Zhang, Zhijian Ma, Jun Huang, Wei Lin, Yaliang Li, Bolin Ding, Jingren Zhou
Abstract:
The burgeoning field of foundation models necessitates advanced data processing mechanisms capable of harnessing vast and valuable data with various types used by these models. Nevertheless, the current landscape presents unique challenges that traditional data processing frameworks struggle to handle effectively, particularly in handling the complexity of multimodal data. In response, we present Data-Juicer 2.0, a data processing system backed by 100+ data processing operators spanning text, image, video, and audio modalities, supporting more critical tasks including data analysis, synthesis, annotation, and foundation model post-training. With seamless compatibility and dedicated optimization for popular dataset hubs like Hugging Face and computing engines like Ray, it improves upon its predecessor in terms of usability, efficiency, and programmability. It features an easily accessible user interface layer that supports decoupled Python interactions, RESTful APIs, and conversational commands. It contains a new runtime layer optimized for adaptive execution and management across varying dataset scales, processing demands, and computational environments, while hiding unnecessary system details. Extensive empirical evaluations demonstrate Data-Juicer 2.0's remarkable performance and scalability, highlighting its capability to efficiently process TB-level data with 10k+ CPU cores. The system is publicly available and has been widely adopted in diverse research fields and real-world products such as Alibaba Cloud PAI. We actively maintain it and share insights from practical feedback, with the goal of facilitating research and application of next-generation foundation models.



Paperid:3697
Authors:Dong Bok Lee, Aoxuan Zhang, Byungjoo Kim, Junhyeon Park, Steven Adriaensen, Juho Lee, Sung Ju Hwang, Hae Beom Lee
Abstract:
In this paper, we address the problem of cost-sensitive hyperparameter optimization (HPO) built upon freeze-thaw Bayesian optimization (BO). Specifically, we assume a scenario where users want to early-stop the HPO process when the expected performance improvement is not satisfactory with respect to the additional computational cost. Motivated by this scenario, we introduce \emph{utility} in the freeze-thaw framework, a function describing the trade-off between the cost and performance that can be estimated from the user's preference data. This utility function, combined with our novel acquisition function and stopping criterion, allows us to dynamically continue training the configuration that we expect to maximally improve the utility in the future, and also automatically stop the HPO process around the maximum utility. Further, we improve the sample efficiency of existing freeze-thaw methods with transfer learning to develop a specialized surrogate model for the cost-sensitive HPO problem. We validate our algorithm on established multi-fidelity HPO benchmarks and show that it outperforms all the previous freeze-thaw BO and transfer-BO baselines we consider, while achieving a significantly better trade-off between the cost and performance.



Paperid:3698
Authors:Panwang Pan, Tingting Shen, Chenxin Li, Yunlong Lin, Kairun Wen, Jingjing Zhao, Yixuan Yuan
Abstract:
Abstract:Recent advances in generative models have achieved high-fidelity in 3D human reconstruction, yet their utility for downstream applications (e.g., human 3D segmentation) remains constrained. We propose $\textbf{Human3R}$, a unified framework that enables simultaneous modeling of appearance and human-part semantics in a $\textbf{generalizable}$ manner. Specifically, we integrate human geometric priors in the reconstruction stage and self-supervised semantic priors in the segmentation stage. To address labeled 3D human datasets scarcity, we further develop an interactive annotation procedure for generating high-quality data-label pairs. Our pixel-aligned aggregation enables cross-task synergy, while the multi-task objective simultaneously optimizes texture modeling fidelity and semantic consistency. Extensive experiments demonstrate that Human3R surpasses existing state-of-the-art methods in both 3D human-part segmentation tasks and 3D human reconstruction. The ablation studies validate the efficacy of critical model designs. The entire pipeline (code and constructed dataset) will be publicly available.



Paperid:3699
Authors:Xiaohao Liu, Xiaobo Xia, Weixiang Zhao, Manyi Zhang, Xianzhi Yu, Xiu Su, Shuo Yang, See-Kiong Ng, Tat-Seng Chua
Abstract:
Large language models (LLMs) have achieved notable progress. Despite their success, next-token prediction (NTP), the dominant method for LLM training and inference, is constrained in both contextual coverage and inference efficiency due to its inherently sequential process. To overcome these challenges, we propose leap multi-token prediction~(L-MTP), an innovative token prediction method that extends the capabilities of multi-token prediction (MTP) by introducing a leap-based mechanism. Unlike conventional MTP, which generates multiple tokens at adjacent positions, L-MTP strategically skips over intermediate tokens, predicting non-sequential ones in a single forward pass. This structured leap not only enhances the model's ability to capture long-range dependencies but also enables a decoding strategy specially optimized for non-sequential leap token generation, effectively accelerating inference. We theoretically demonstrate the benefit of L-MTP in improving inference efficiency. Experiments across diverse benchmarks validate its merit in boosting both LLM performance and inference speed. The source code will be publicly available.



Paperid:3700
Authors:Zichen Wen, Shaobo Wang, Yufa Zhou, Junyuan Zhang, Qintong Zhang, Yifeng Gao, Zhaorun Chen, Bin Wang, Weijia Li, Conghui He, Linfeng Zhang
Abstract:
Visual tokens consume substantial computational resources in multi-modal large models (MLLMs), significantly compromising their efficiency. Recent works have attempted to improve efficiency by compressing visual tokens during training, either through modifications to model components or by introducing additional parameters. However, they often overlook the increased learning difficulty caused by such compression, as the model’s parameter space struggles to quickly adapt to the substantial perturbations in the feature space induced by token compression. In this work, we propose to develop Efficient MLLMs via Progressive Consistency Distillation (EPIC), a progressive learning framework. Specifically, by decomposing the feature space perturbations introduced by token compression along the token-wise and layer-wise dimensions, we introduce token consistency distillation and layer consistency distillation, respectively, aiming to reduce the training difficulty by leveraging guidance from a teacher model and following a progressive learning trajectory. Extensive experiments demonstrate the superior effectiveness, robustness, and generalization capabilities of our proposed framework.



Paperid:3701
Authors:Yan Zhuang, Minhao Liu, Wei Bai, Yanru Zhang, Wei Li, Jiawen Deng, Fuji Ren
Abstract:
Multimodal Sentiment Analysis (MSA) aims to infer human emotions by integrating complementary signals from diverse modalities. However, in real-world scenarios, missing modalities are common due to data corruption, sensor failure, or privacy concerns, which can significantly degrade model performance. To tackle this challenge, we propose Hyper-Modality Enhancement (HME), a novel framework that avoids explicit modality reconstruction by enriching each observed modality with semantically relevant cues retrieved from other samples. This cross-sample enhancement reduces reliance on fully observed data during training, making the method better suited to scenarios with inherently incomplete inputs. In addition, we introduce an uncertainty-aware fusion mechanism that adaptively balances original and enriched representations to improve robustness. Extensive experiments on three public benchmarks show that HME consistently outperforms state-of-the-art methods under various missing modality conditions, demonstrating its practicality in real-world MSA applications.



Paperid:3702
Authors:Jieyuan (Eric) Zhang, Xiaolong Zhou, Wenjie Wei, Hanwen Liu, Qian Sun, Shuai Wang, Malu Zhang, Yang Yang, Haizhou Li
Abstract:
Spiking Neural Networks (SNNs) offer promising biological plausibility, energy efficiency, and event-driven processing for neuromorphic computing. Despite architectural and training advances, their complex spatiotemporal dynamics complicate understanding of intrinsic learning behaviors. We introduce Spatio-Temporal Effective Receptive Field (ST-ERF), extending traditional receptive field analysis to accommodate SNN temporal dynamics. Our analysis reveals two key ST-ERF properties: Gaussian-like spatial distributions and power-law temporal decay patterns. Furthermore, examination of Transformer-based SNN architectures reveals significant receptive field variations, with several prominent models (SDT-V1, Meta-SDT, QKFormer) showing limited global receptive fields that impair visual long-sequence modeling. Identifying the channel-mixer component as the critical factor, we propose two novel designs: MLPixer and Splash-and-reconstruct Block (SR Block). These innovations enhance global spatial receptive fields in early network stages, significantly improving performance on challenging visual tasks. Experimental results validate our theoretical findings and demonstrate the effectiveness of our proposed architectures, advancing both theoretical understanding and practical applications of SNNs in complex visual processing.



Paperid:3703
Authors:Kailai Li, Jiawei Sun, Honghong Zeng, Jiong Lou, Chentao Wu, Yuan Luo, Wen Li, Wei Zhao, shouguo du, Jie LI
Abstract:
Abstract:Graph Neural Networks (GNNs) achieve strong performance in node classification tasks but exhibit substantial performance degradation under label noise. Despite recent advances in noise-robust learning, a principled approach that exploits the node-neighbor interdependencies inherent in graph data for label noise detection remains underexplored. To address this gap, we propose GD$^2$, a noise-aware \underline{G}raph learning framework that detects label noise by leveraging \underline{D}ual-view prediction \underline{D}iscrepancies. The framework contrasts the \textit{ego-view}, constructed from node-specific features, with the \textit{structure-view}, derived through the aggregation of neighboring representations. The resulting discrepancy captures disruptions in semantic coherence between individual node representations and the structural context, enabling effective identification of mislabeled nodes. Building upon this insight, we further introduce a view-specific training strategy that enhances noise detection by amplifying prediction divergence through differentiated view-specific supervision. Extensive experiments on multiple datasets and noise settings demonstrate that \name~achieves superior performance over state-of-the-art baselines.



Paperid:3704
Authors:Wanxin Tian, Shijie Zhang, Yulin Luo, Xiaowei Chi, Junyu Lu, Chun-Kai Fan, Kevin Zhang, d, Ning Liu, Qiang Zhou, Siyu Lin, Zhiyuan Qin, Xiaozhu Ju, Shanghang Zhang, Jian Tang
Abstract:
Self-evolution, the ability of agents to autonomously improve their reasoning and behavior, is essential for the embodied domain with long-horizon, real-world tasks. Despite current advancements in reinforcement fine-tuning (RFT) showing strong performance in enhancing reasoning in LLMs, its potential to enable self-evolving embodied intelligence with multi-modal interactions remains largely unexplored. Specifically, reinforcement fine-tuning faces two fundamental obstacles in embodied settings: (i) the lack of accessible intermediate rewards in multi-step reasoning tasks limits effective learning signals, and (ii) reliance on hand-crafted reward functions restricts generalization to novel tasks and environments. To address these challenges, we present \textit{\contour{black}{\textbf{S}}elf-\contour{black}{\textbf{E}}volving \contour{black}{\textbf{E}}mbodied \contour{black}{\textbf{A}}gents-R1}, \textbf{SEEA-R1}, the first RFT framework designed for enabling the self-evolving capabilities of embodied agents. Specifically, to convert sparse delayed rewards into denser intermediate signals that improve multi-step reasoning, we propose Tree-based group relative policy optimization (\textbf{Tree-GRPO}) integrates Monte Carlo Tree Search into GRPO. To generalize reward estimation across tasks and scenes, supporting autonomous adaptation and reward-driven self-evolution, we further introduce Multi-modal Generative Reward Model (\textbf{MGRM}). To holistically evaluate the effectiveness of SEEA-R1, we evaluate on the ALFWorld benchmark, surpassing state-of-the-art methods with scores of 85.07\% (textual) and 36.19\% (multi-modal), outperforming prior models including GPT-4o. SEEA-R1 also achieves scores of 80.3\% without environmental reward, surpassing all open-source baselines and highlighting its scalability as a self-evolving embodied agent. Additional experiments and qualitative analysis further support the potential of SEEA-R1 for future research in scalable embodied intelligence.



Paperid:3705
Authors:Bufang Yang, Lilin Xu, Liekang Zeng, Kaiwei Liu, Siyang Jiang, Wenrui Lu, Hongkai Chen, Xiaofan Jiang, Guoliang Xing, Zhenyu Yan
Abstract:
Recent advances in Large Language Models (LLMs) have propelled intelligent agents from reactive responses to proactive support. While promising, existing proactive agents either rely exclusively on observations from enclosed environments (e.g., desktop UIs) with direct LLM inference or employ rule-based proactive notifications, leading to suboptimal user intent understanding and limited functionality for proactive service. In this paper, we introduce ContextAgent, the first context-aware proactive agent that incorporates extensive sensory contexts to enhance the proactive capabilities of LLM agents. ContextAgent first extracts multi-dimensional contexts from massive sensory perceptions on wearables (e.g., video and audio) to understand user intentions. ContextAgent then leverages the sensory contexts and the persona contexts from historical data to predict the necessity for proactive services. When proactive assistance is needed, ContextAgent further automatically calls the necessary tools to assist users unobtrusively. To evaluate this new task, we curate ContextAgentBench, the first benchmark for evaluating context-aware proactive LLM agents, covering 1,000 samples across nine daily scenarios and twenty tools. Experiments on ContextAgentBench show that ContextAgent outperforms baselines by achieving up to 8.5% and 6.0% higher accuracy in proactive predictions and tool calling, respectively. We hope our research can inspire the development of more advanced, human-centric, proactive AI assistants.



Paperid:3706
Authors:Xiaoqiang Wang, Suyuchen Wang, Yun Zhu, Bang Liu
Abstract:
Chain-of-thought (CoT) reasoning enables large language models (LLMs) to move beyond fast System-1 responses and engage in deliberative System-2 reasoning. However, this comes at the cost of significant inefficiency due to verbose intermediate output. Recent latent-space reasoning methods improve efficiency by operating on hidden states without decoding into language, yet they treat all steps uniformly, failing to distinguish critical deductions from auxiliary steps and resulting in suboptimal use of computational resources. In this paper, we propose System-1.5 Reasoning, an adaptive reasoning framework that dynamically allocates computation across reasoning steps through shortcut paths in latent space.Specifically, System-1.5 Reasoning introduces two types of dynamic shortcuts. The model depth shortcut (DS) adaptively reasons along the vertical depth by early exiting non-critical tokens through lightweight adapter branches, while allowing critical tokens to continue through deeper Transformer layers. The step shortcut (SS) reuses hidden states across the decoding steps to skip trivial steps and reason horizontally in latent space. Training System-1.5 Reasoning involves a two-stage self-distillation process: first distilling natural language CoT into latent-space continuous thought, and then distilling full-path System-2 latent reasoning into adaptive shortcut paths (System-1.5 Reasoning).Experiments on reasoning tasks demonstrate the superior performance of our method.For example, on GSM8K, System-1.5 Reasoning achieves reasoning performance comparable to traditional CoT fine-tuning methods while accelerating inference by over 20× and reducing token generation by 92.31\% on average.



Paperid:3707
Authors:Yikang Li, Yeqing Qiu, Yuxuan Chen, Lingshen He, Lexiang Hu, Zhouchen Lin
Abstract:
Equivariant networks enhance model efficiency and generalization by embedding symmetry priors into their architectures.However, most existing methods, primarily based on group convolutions and steerable convolutions, face significant limitations when dealing with complex transformation groups, particularly the projective group, which plays a crucial role in vision.In this work, we tackle the challenge by constructing projective equivariant networks based on differential invariants.Using the moving frame method with a carefully selected cross section tailored for multi-dimensional functions, we derive a complete and concise set of second-order fundamental differential invariants of the projective group.We provide a rigorous analysis of the properties and transformation relationships of their underlying components, yielding a further simplified and unified set of fundamental differential invariants, which facilitates both theoretical analysis and practical applications.Building on this foundation, we develop the first deep projective equivariant networks, PDINet, which achieve full projective equivariance without discretizing or sampling the group.Empirical results on projectively transformed STL-10 and Imagenette datasets show that PDINet achieves improvements of 11.39\% and 5.66\% in accuracy over baseline results, respectively, demonstrating strong generalization to complex geometric transformations under out-of-distribution settings.



Paperid:3708
Authors:Zixun Wang, Ben Dai
Abstract:
Abstract:Semantic segmentation labels each pixel in an image with its corresponding class, and is typically evaluated using the Intersection over Union (IoU) and Dice metrics to quantify the overlap between predicted and ground-truth segmentation masks. In the literature, most existing methods estimate pixel-wise class probabilities, then apply argmax or thresholding to obtain the final prediction. These methods have been shown to generally lead to inconsistent or suboptimal results, as they do not directly maximize segmentation metrics. To address this issue, a novel consistent segmentation framework, RankSEG, has been proposed, which includes RankDice and RankIoU specifically designed to optimize the Dice and IoU metrics, respectively. Although RankSEG almost guarantees improved performance, it suffers from two major drawbacks. First, it is its computational expense—RankDice has a complexity of $\mathcal{O}(d \log d)$ with a substantial constant factor (where $d$ represents the number of pixels), while RankIoU exhibits even higher complexity $\mathcal{O}(d^2)$, thus limiting its practical application. For instance, in LiTS, prediction with RankSEG takes 16.33 seconds compared to just 0.01 seconds with the argmax rule. Second, RankSEG is only applicable to overlapping segmentation settings, where multiple classes can occupy the same pixel, which contrasts with standard benchmarks that typically assume non-overlapping segmentation. In this paper, we overcome these two drawbacks via a \textit{reciprocal moment approximation} (RMA) of RankSEG with the following contributions: (i) we improve RankSEG using RMA, namely RankSEG-RMA, reduces the complexity of both algorithms to $\mathcal{O}(d)$ while maintaining comparable performance; (ii) inspired by RMA, we develop a pixel-wise score function that allows efficient implementation for non-overlapping segmentation settings. We illustrate the effectiveness of our method across various datasets and state-of-the-art models.



Paperid:3709
Authors:Jiarui Zhang, Xiangyu Liu, Yong Hu, Chaoyue Niu, Fan Wu, Guihai Chen
Abstract:
Retrieval-Augmented Generation (RAG) significantly improves the performance of Large Language Models (LLMs) on knowledge-intensive tasks. However, varying response quality across LLMs under RAG necessitates intelligent routing mechanisms, which select the most suitable model for each query from multiple retrieval-augmented LLMs via a dedicated router model. We observe that external documents dynamically affect LLMs' ability to answer queries, while existing routing methods, which rely on static parametric knowledge representations, exhibit suboptimal performance in RAG scenarios. To address this, we formally define the new retrieval-augmented LLM routing problem, incorporating the influence of retrieved documents into the routing framework. We propose RAGRouter, a RAG-aware routing design, which leverages document embeddings and RAG capability embeddings with contrastive learning to capture knowledge representation shifts and enable informed routing decisions. Extensive experiments on diverse knowledge-intensive tasks and retrieval settings show that RAGRouter outperforms the best individual LLM by 3.61\% on average and existing routing methods by 3.29\%–9.33\%. With an extended score-threshold-based mechanism, it also achieves strong performance-efficiency trade-offs under low-latency constraints.



Paperid:3710
Authors:Vaggos Chatziafratis, Ishani Karmarkar, Yingxi Li, Ellen Vitercik
Abstract:
Data-driven algorithm selection is a powerful approach for choosing effective heuristics for computational problems. It operates by evaluating a set of candidate algorithms on a collection of representative training instances and selecting the one with the best empirical performance. However, running each algorithm on every training instance is computationally expensive, making scalability a central challenge. In practice, a common workaround is to evaluate algorithms on smaller proxy instances derived from the original inputs. However, this practice has remained largely ad hoc and lacked theoretical grounding. We provide the first theoretical foundations for this practice by formalizing the notion of size generalization: predicting an algorithm's performance on a large instance by evaluating it on a smaller, representative instance, subsampled from the original instance. We provide size generalization guarantees for three widely used clustering algorithms (single-linkage, k-means++, and Gonzalez's k-centers heuristic) and two canonical max-cut algorithms (Goemans-Williamson and Greedy). We characterize the subsample size sufficient to ensure that performance on the subsample reflects performance on the full instance, and our experiments support these findings.



Paperid:3711
Authors:Yash Patel, Eduardo Ochoa Rivera, Ambuj Tewari
Abstract:
Distribution-free uncertainty estimation for ensemble methods is increasingly desirable due to the widening deployment of multi-modal black-box predictive models. Conformal prediction is one approach that avoids such distributional assumptions. Methods for conformal aggregation have in turn been proposed for ensembled prediction, where the prediction regions of individual models are merged as to retain coverage guarantees while minimizing conservatism. Merging the prediction regions directly, however, sacrifices structures present in the conformal scores that can further reduce conservatism. We, therefore, propose a novel framework that extends the standard scalar formulation of a score function to a multivariate score that produces more efficient prediction regions. We then demonstrate that such a framework can be efficiently leveraged in both classification and predict-then-optimize regression settings downstream and empirically show the advantage over alternate conformal aggregation methods.



Paperid:3712
Authors:Yixiong Fang, Ziran Yang, Zhaorun Chen, Zhuokai Zhao, Jiawei Zhou
Abstract:
Large vision-language models (LVLMs) excel at multimodal tasks but are prone to misinterpreting visual inputs, often resulting in hallucinations and unreliable outputs. We present Dropout Decoding, a novel inference-time approach that quantifies the uncertainty of visual tokens and selectively masks uncertain tokens to improve decoding. Our method measures the uncertainty of each visual token by projecting it onto the text space and decomposing it into aleatoric and epistemic components. Specifically, we focus on epistemic uncertainty, which captures perception-related errors more effectively. Inspired by dropout regularization, we introduce uncertainty-guided token dropout, which applies the dropout principle to input visual tokens instead of model parameters, and during inference rather than training. By aggregating predictions from an ensemble of masked decoding contexts, we can robustly mitigate errors arising from visual token misinterpretations. Evaluations on benchmarks including CHAIR, THRONE, and MMBench demonstrate that Dropout Decoding significantly reduces object hallucinations (OH) and enhances both reliability and quality of LVLM outputs across diverse visual contexts.



Paperid:3713
Authors:Shuai Liu, Quanmin Liang, Zefeng Li, Boyang Li, Kai Huang
Abstract:
Multi-sensor fusion is crucial for improving the performance and robustness of end-to-end autonomous driving systems. Existing methods predominantly adopt either attention-based flatten fusion or bird’s eye view fusion through geometric transformations. However, these approaches often suffer from limited interpretability or dense computational overhead. In this paper, we introduce GaussianFusion, a Gaussian-based multi-sensor fusion framework for end-to-end autonomous driving. Our method employs intuitive and compact Gaussian representations as intermediate carriers to aggregate information from diverse sensors. Specifically, we initialize a set of 2D Gaussians uniformly across the driving scene, where each Gaussian is parameterized by physical attributes and equipped with explicit and implicit features. These Gaussians are progressively refined by integrating multi-modal features. The explicit features capture rich semantic and spatial information about the traffic scene, while the implicit features provide complementary cues beneficial for trajectory planning. To fully exploit rich spatial and semantic information in Gaussians, we design a cascade planning head that iteratively refines trajectory predictions through interactions with Gaussians. Extensive experiments on the NAVSIM and Bench2Drive benchmarks demonstrate the effectiveness and robustness of the proposed GaussianFusion framework. The source code is included in the supplementary material and will be released publicly.



Paperid:3714
Authors:Jiawen Wei, jiang lan, Pengbo Wei, Ziwen Ye, Teng Song, Chen Chen, Guangrui Ma
Abstract:
Time series data is ubiquitous, with forecasting applications spanning from finance to healthcare. Beyond popular deterministic methods, generative models are gaining attention due to advancements in areas like image synthesis and video generation, as well as their inherent ability to provide probabilistic predictions. However, existing generative approaches mostly involve recurrent generative operations or repeated denoising steps, making the prediction laborious, particularly for long-term forecasting. Most of them only conduct experiments for relatively short-term forecasting, with limited comparison to deterministic methods in long-term forecasting, leaving their practical advantages unclear. This paper presents TARFVAE, a novel generative framework that combines the Transformer-based autoregressive flow (TARFLOW) and variational autoencoder (VAE) for efficient one-step generative time series forecasting. Inspired by the rethinking that complex architectures for extracting time series representations might not be necessary, we add a flow module, TARFLOW, to VAE to promote spontaneous learning of latent variables that benefit predictions. TARFLOW enhances VAE's posterior estimation by breaking the Gaussian assumption, thereby enabling a more informative latent space. TARFVAE uses only the forward process of TARFLOW, avoiding autoregressive inverse operations and thus ensuring fast generation. During generation, it samples from the prior latent space and directly generates full-horizon forecasts via the VAE decoder. With simple MLP modules, TARFVAE achieves superior performance over state-of-the-art deterministic and generative models across different forecast horizons on benchmark datasets while maintaining efficient prediction speed, demonstrating its effectiveness as an efficient and powerful solution for generative time series forecasting. Our code is available at https://anonymous.4open.science/r/-TARFVAE.



Paperid:3715
Authors:Jifeng Hu, Sili Huang, Li Shen, Zhejian Yang, Shengchao Hu, Shisong Tang, Hechang Chen, Lichao Sun, Yi Chang, Dacheng Tao
Abstract:
Continual offline reinforcement learning (CORL) has shown impressive ability in diffusion-based continual learning systems by modeling the joint distributions of trajectories. However, most research only focuses on limited continual task settings where the tasks have the same observation and action space, which deviates from the realistic demands of training agents in various environments. In view of this, we propose Vector-Quantized Continual Diffuser, named VQ-CD, to break the barrier of different spaces between various tasks. Specifically, our method contains two complementary sections, where the quantization spaces alignment provides a unified basis for the selective weights activation. In the quantized spaces alignment, we leverage vector quantization to align the different state and action spaces of various tasks, facilitating continual training in the same space. Then, we propose to leverage a unified diffusion model attached by the inverse dynamic model to master all tasks by selectively activating different weights according to the task-related sparse masks. Finally, we conduct extensive experiments on 15 continual learning (CL) tasks, including conventional CL task settings (identical state and action spaces) and general CL task settings (various state and action spaces). Compared with 17 baselines, our method reaches the SOTA performance.



Paperid:3716
Authors:Ye Liu, Zongyang Ma, Junfu Pu, Zhongang Qi, Yang Wu, Ying Shan, Chang Chen
Abstract:
Recent advances in Large Multi-modal Models (LMMs) have demonstrated their remarkable success as general-purpose multi-modal assistants, with particular focuses on holistic image- and video-language understanding. Conversely, less attention has been given to scaling fine-grained pixel-level understanding capabilities, where the models are expected to realize pixel-level alignment between visual signals and language semantics. Some previous studies have applied LMMs to related tasks such as region-level captioning and referring expression segmentation, however, these models are limited to performing either referring or segmentation tasks independently and fail to integrate these fine-grained perception capabilities into visual reasoning. To bridge this gap, we propose UniPixel, a large multi-modal model capable of flexibly comprehending visual prompt inputs and generating mask-grounded responses. Our model distinguishes itself by seamlessly integrating pixel-level perception with general visual understanding capabilities. Specifically, UniPixel processes visual prompts and generates relevant masks on demand, and performs subsequent reasoning conditioning on these intermediate pointers during inference, thereby enabling fine-grained pixel-level reasoning. The effectiveness of the proposed method has been verified on 10 benchmarks across a diverse set of tasks, including pixel-level referring/segmentation and object-centric understanding in images/videos. A novel PixelQA task is also introduced to verify the significance of our method. Code and models will be publicly available.



Paperid:3717
Authors:Tom Waknine, Shay Moran, Bogdan Chornomaz, Yonatan Koren
Abstract:
Abstract:We study the problem of learning in the presence of an adversary that can corrupt an $\eta$ fraction of the training examples with the goal of causing failure on a specific test point. In the realizable setting, prior work established that the optimal error under such instance-targeted poisoning attacks scales as $\Theta(d\eta)$, where $d$ is the VC dimension of the hypothesis class [Hanneke, Karbasi, Mahmoody, Mehalel, and Moran (NeurIPS 2022)]. In this work, we resolve the corresponding question in the agnostic setting. We show that the optimal excess error is $\widetilde\Theta(\sqrt{d\eta})$, answering one of the main open problems left by Hanneke et al. To achieve this rate, it is necessary to use randomized learners: Hanneke et al.\ showed that deterministic learners can be forced to suffer error close to $1$ even under small amounts of poisoning. Perhaps surprisingly, our upper bound remains valid even when the learner’s random bits are fully visible to the adversary. In the other direction, our lower bound is stronger than standard PAC-style bounds: instead of tailoring a hard distribution separately for each sample size, we exhibit a single fixed distribution under which the adversary can enforce an excess error of $\Omega(\sqrt{d\eta})$ infinitely often.



Paperid:3718
Authors:Riccardo De Santi, Marin Vlastelica, Ya-Ping Hsieh, Zebang Shen, Niao He, Andreas Krause
Abstract:
Adapting large-scale foundational flow and diffusion generative models to optimize task-specific objectives while preserving prior information is crucial for real-world applications such as molecular design, protein docking, and creative image generation. Existing principled fine-tuning methods aim to maximize the expected reward of generated samples, while retaining knowledge from the pre-trained model via KL-divergence regularization. In this work, we tackle the significantly more general problem of optimizing general utilities beyond average rewards, including risk-averse and novelty-seeking reward maximization, diversity measures for exploration, and experiment design objectives among others. Likewise, we consider more general ways to preserve prior information beyond KL-divergence, such as optimal transport distances and Rényi divergences. To this end, we introduce Flow Density Control (FDC), a simple algorithm that reduces this complex problem to a specific sequence of simpler fine-tuning tasks, each solvable via scalable established methods. We derive convergence guarantees for the proposed scheme under realistic assumptions by leveraging recent understanding of mirror flows. Finally, we validate our method on illustrative settings, text-to-image, and molecular design tasks, showing that it can steer pre-trained generative models to optimize objectives and solve practically relevant tasks beyond the reach of current fine-tuning schemes.



Paperid:3719
Authors:Hua Ye, Siyuan Chen, Haoliang Zhang, Weihao Luo, Yanbin Li, Xuan Zhang
Abstract:
Large language models (LLMs) demonstrate impressive generalization abilities, yet adapting them effectively across multiple heterogeneous domains remains challenging due to inter-domain interference. To overcome this challenge, we propose a partition-based multi-stage fine-tuning framework designed to exploit inter-domain synergies while minimizing negative transfer. Our approach strategically partitions domains into subsets (stages) by balancing domain discrepancy, synergy, and model capacity constraints. We theoretically analyze the proposed framework and derive novel generalization bounds that justify our partitioning strategy. Extensive empirical evaluations on various language understanding tasks show that our method consistently outperforms state-of-the-art baselines.



Paperid:3720
Authors:Jaekyun Park, Hye Won Chung
Abstract:
In the era of large-scale foundation models, fully fine-tuning pretrained networks for each downstream task is often prohibitively resource-intensive. Prompt tuning offers a lightweight alternative by introducing tunable prompts while keeping the backbone frozen. However, existing visual prompt tuning methods often fail to specialize the prompts or enrich the representation space--especially when applied to self-supervised backbones. We show that these limitations become especially pronounced in challenging tasks and data-scarce settings, where effective adaptation is most critical. In this work, we introduce VIPAMIN, a visual prompt initialization strategy that enhances adaptation of self-supervised models by (1) aligning prompts with semantically informative regions in the embedding space, and (2) injecting novel representational directions beyond the pretrained subspace. Despite its simplicity--requiring only a single forward pass and lightweight operations--VIPAMIN consistently improves performance across diverse tasks and dataset sizes, setting a new state of the art in visual prompt tuning.



Paperid:3721
Authors:Xiaoyang Liu, Kangjie Bao, Jiashuo Zhang, Yunqi Liu, Yuntian Liu, Yu Chen, Yang Jiao, Tao Luo
Abstract:
Abstract:Autoformalization, the automatic translation of mathematical content from natural language into machine-verifiable formal languages, has seen significant progress driven by advances in large language models (LLMs). Nonetheless, a primary barrier to further improvements is the limited availability of parallel corpora that map informal mathematical text to its formal counterpart. To address this limitation, we propose ATLAS (Autoformalizing Theorems through Lifting, Augmentation, and Synthesis of Data), a novel data generation framework designed to produce large-scale, high-quality parallel corpora of theorem statements. Distinct from prior approaches, ATLAS begins with a concept repository, accelerates the improvement of student model through expert iteration combined with knowledge distillation, and introduces two novel augmentation strategies that exploit the structural characteristics of formal languages. With the proposed ATLAS running for 10 iterations, we construct an undergraduate-level dataset comprising 117k theorem statements and develop ATLAS Translator, which demonstrates statistically significant improvements over both the HERALD Translator and the Kimina-Autoformalizer across all benchmarks ($p<0.05$, two-sided t-test), achieving a new state of the art. The datasets, model, and code will be released to the public soon.



Paperid:3722
Authors:Mohammad Pedramfar, Christopher Quinn, Vaneet Aggarwal
Abstract:
Abstract:This paper presents novel contributions to the field of online optimization, particularly focusing on the adaptation of algorithms from concave optimization to more challenging classes of functions.Key contributions include the introduction of uniform wrappers, a class of meta-algorithms that could be used for algorithmic conversions such as converting algorithms for convex optimization into those for quadratizable optimization.Moreover, we propose a guideline that, given a base algorithm $\mathcal{A}$ for concave optimization and a uniform wrapper $\mathcal{W}$, describes how to convert a proof of the regret bound of $\mathcal{A}$ in the concave setting into a proof of the regret bound of $\mathcal{W}(\mathcal{A})$ for quadratizable setting.Through this framework, the paper demonstrates improved regret guarantees for various classes of DR-submodular functions under zeroth-order feedback. Furthermore, the paper extends zeroth-order online algorithms to bandit feedback and offline counterparts, achieving notable improvements in regret/sample complexity compared to existing approaches.



Paperid:3723
Authors:Marianne Arriola, Yair Schiff, Hao Phung, Aaron Gokaslan, Volodymyr Kuleshov
Abstract:
Discrete diffusion models have steadily improved in quality, but still face a performance gap relative to autoregressive approaches. Partitioning sequences into blocks and performing diffusion within each block improves performance, but doubles computational costs in standard decoder-only architectures. Here, we propose an encoder-decoder architecture for block discrete diffusion, which relies on a lightweight decoder to iteratively refine each block of generated data. This greatly improves training and inference efficiency and allows training larger and higher quality block diffusion models for the same compute budget. We introduce encoder-decoder discrete diffusion (E2D2), a block diffusion model with specialized training and sampling algorithms, and we show that E2D2 achieves superior trade-offs between generation quality and inference throughput on summarization, translation, and mathematical reasoning tasks.



Paperid:3724
Authors:Haonan Yuan, Qingyun Sun, Junhua Shi, Xingcheng Fu, Bryan Hooi, Jianxin Li, Philip S Yu
Abstract:
Inspired by the remarkable success of foundation models in language and vision, Graph Foundation Models (GFMs) hold significant promise for broad applicability across diverse graph tasks and domains. However, existing GFMs struggle with unstable few-shot fine-tuning, where both performance and adaptation efficiency exhibit significant fluctuations caused by the randomness in the support sample selection and structural discrepancies between the pre-trained and target graphs. How to fine-tune GFMs robustly and efficiently to enable trustworthy knowledge transfer across domains and tasks is the major challenge. In this paper, we proposeGRAVER, a novelGenerative gRAphVocabulariEs forRobust GFM fine-tuning framework that tackles the aforementioned instability via generative augmentations. Specifically, to identify transferable units, we analyze and extract key class-specific subgraph patterns by ego-graph disentanglement and validate their transferability both theoretically and empirically. To enable effective pre-training across diverse domains, we leverage a universal task template based on ego-graph similarity and construct graph vocabularies via graphon-based generative experts. To facilitate robust and efficient prompt fine-tuning, we grave the support samples with in-context vocabularies, where the lightweight MoE-CoE network attentively routes knowledge from source domains. Extensive experiments demonstrate the superiority of GRAVER overeffectiveness,robustness, andefficiencyon downstream few-shot node and graph classification tasks compared with15state-of-the-art baselines.



Paperid:3725
Authors:Chao Huang, Benfeng Wang, Jie Wen, Chengliang Liu, Wei Wang, Li Shen, Xiaochun Cao
Abstract:
Recent advancements in reasoning capability of Multimodal Large Language Models (MLLMs) demonstrate its effectiveness in tackling complex visual tasks. However, existing MLLM-based Video Anomaly Detection (VAD) methods remain limited to shallow anomaly descriptions without deep reasoning. In this paper, we propose a new task named Video Anomaly Reasoning (VAR), which aims to enable deep analysis and understanding of anomalies in the video by requiring MLLMs to think explicitly before answering. To this end, we propose Vad-R1, an end-to-end MLLM-based framework for VAR. Specifically, we design a Perception-to-Cognition Chain-of-Thought (P2C-CoT) that simulates the human process of recognizing anomalies, guiding the MLLM to reason anomaly step-by-step. Based on the structured P2C-CoT, we construct Vad-Reasoning, a dedicated dataset for VAR. Furthermore, we propose an improved reinforcement learning algorithm AVA-GRPO, which explicitly incentivizes the anomaly reasoning capability of MLLMs through a self-verification mechanism with limited annotations. Experimental results demonstrate that Vad-R1 achieves superior performance, outperforming both open-source and proprietary models on VAD and VAR tasks.



Paperid:3726
Authors:Guojingfeng, Jian Liu, Jinnan Chen, Shiwei Mao, Changrong Hu, Puhua Jiang, Junlin Yu, Jing Xu, Qi Liu, LiXin Xu, Zhuo Chen, Chunchao Guo
Abstract:
Abstract:We introduce Auto-Connect, a novel approach for automatic rigging that explicitly preserves skeletal connectivity through a connectivity-preserving tokenization scheme. Unlike previous methods that predict bone positions represented as two joints or first predict points before determining connectivity, our method employs special tokens to define endpoints for each joint's children and for each hierarchical layer, effectively automating connectivity relationships. This approach significantly enhances topological accuracy by integrating connectivity information directly into the prediction framework.To further guarantee high-quality topology, we implement a topology-aware reward function that quantifies topological correctness, which is then utilized in a post-training phase through reward-guided Direct Preference Optimization. Additionally, we incorporate implicit geodesic features for latent top-$k$ bone selection, which substantially improves skinning quality. By leveraging geodesic distance information within the model's latent space, our approach intelligently determines the most influential bones for each vertex, effectively mitigating common skinning artifacts.This combination of connectivity-preserving tokenization, reward-guided fine-tuning, and geodesic-aware bone selection enables our model to consistently generate more anatomically plausible skeletal structures with superior deformation properties.



Paperid:3727
Authors:Weiyi Xue, Fan Lu, Yunwei Zhu, Zehan Zheng, Haiyun Wei, Sanqing Qu, Jiangtong Li, Ya Wu, Guang Chen
Abstract:
Pose-free Neural Radiance Field (NeRF) aims at novel view synthesis (NVS) without relying on accurate poses, exhibiting significant practical value. Image and LiDAR point cloud are two pivotal modalities in autonomous driving scenarios. While demonstrating impressive performance, single-modality pose-free NeRFs often suffer from local optima due to the limited geometric information provided by dense image textures or the sparse, textureless nature of point clouds. Although prior methods have explored the complementary strengths of both modalities, they have only leveraged inherently sparse point clouds for discrete, non-pixel-wise depth supervision, and are limited to NVS of images. As a result, a Multimodal Unified Pose-free framework remains notably absent. In light of this, we propose MUP, a pose-free framework for LiDAR-Camera joint NVS in large-scale scenes. This unified framework enables continuous depth supervision for image reconstruction using LiDAR-Fields rather than discrete point clouds. By leveraging multimodal inputs, pose optimization receives gradients from the rendering loss of point cloud geometry and image texture, thereby alleviating the issue of local optima commonly encountered in single-modality pose-free tasks. Moreover, to further guide pose optimization of NeRF, we propose a multimodal geometric optimizer that leverages geometric relations from point clouds and photometric regularization from adjacent image frames. Besides, to alleviate the domain gap between modalities, we propose a multimodal-specific coarse-to-fine training approach for unified, compact reconstruction. Extensive experiments on KITTI-360 and NuScenes datasets demonstrate MUP's superiority in accomplishing geometry-aware, modality-consistent, and pose-free 3D reconstruction.



Paperid:3728
Authors:Yimeng Chen, Piotr Piękos, Mateusz Ostaszewski, Firas Laakom, Jürgen Schmidhuber
Abstract:
Evaluating the scientific discovery capabilities of large language model based agents, particularly how they cope with varying environmental complexity and utilize prior knowledge, requires specialized benchmarks currently lacking in the landscape. To address this gap, we introduce \textsc{Physicist}, a novel benchmark suite and simulation platform for rigorously assessing LLM-based scientific reasoning in interactive physics environments. \textsc{Physicist}'s primary contribution lies in its sophisticated control over the level of prior knowledge provided to the agent. This allows researchers to dissect agent performance along axes including the complexity of the problem and the prior knowledge levels. The benchmark comprises a suite of interactive simulations, where agents must actively probe environments, gather data sequentially under constraints and formulate hypotheses about underlying physical laws. \textsc{Physicist} provides standardized evaluation protocols and metrics for assessing hypothesis accuracy and model fidelity. We demonstrate the benchmark's utility by presenting results from baseline LLMs, showcasing its ability to differentiate capabilities based on varying priors and task complexity.



Paperid:3729
Authors:Xiaobin Liu, Jianing Li, Baiwei Guo, WenbinZhu, Jing Yuan
Abstract:
Recent works show that contrastive learning based on memory banks is an effective framework for unsupervised person Re-IDentification (ReID). In existing methods, memory banks are typically initialized with cluster centroids and rewritten with positive samples via the momentum mechanism along with the model training. However, this mechanism solely focuses on the intra-class compactness by pulling memory banks close to positive samples, neglecting the inter-class separability among different memory banks. Rewriting memory banks with partial constraint limits their discrimination capacities, and hence hinders learning discriminative features based on those memory banks. In this paper, we claim that memory banks should be rewritten with both intra-class and inter-class constraints, and therefore propose a unified memory bank rewriting mechanism, Bidirectionally Memory bank reWriting (BMW), to chase enhanced discrimination capacity. Specifically, BMW formulates the memory bank rewriting as the gradient descent update with two objectives, i.e., reducing intra-class diversity and enhancing inter-class separability. To effectively enhance the separability of memory banks with limited number of rewriting steps, we further design a novel objective formulation for the inter-class constraint, which is more effective for one step update. BMW enhances both representation and discrimination capacities of memory banks, thus leads to an effective ReID feature optimization. BMW is simple yet effective and can serve as a new paradigm for person ReID methods based on memory banks. Extensive experiments on standard benchmarks demonstrate the effectiveness of our BMW method in unsupervised ReID model training. Specially, BMW even outperforms previous methods that use stronger backbones. Code will be available.



Paperid:3730
Authors:Jumin Han, Jun-Hui Kim, Seong-Whan Lee
Abstract:
3D Human Pose Estimation (HPE) is a fundamental task in the computer vision. Generalization in 3D HPE task is crucial due to the need for robustness across diverse environments and datasets. Existing methods often focus on learning relationships between joints to enhance the generalization capability, but the role of the loss landscape, which is closely tied to generalization, remains underexplored. In this paper, we empirically visualize the loss landscape of the 3D HPE task, revealing its complexity and the challenges it poses for optimization. To address this, we first introduce a simple adaptive scaling mechanism that smooths the loss landscape. We further observe that different solutions on this smoothed loss landscape exhibit varying generalization behaviors. Based on this insight, we propose an efficient ensemble approach that combines diverse solutions on the smooth loss landscape induced by our adaptive scaling mechanism. Extensive experimental results demonstrate that our approach improves the generalization capability of 3D HPE models, and can be easily applied, regardless of model architecture, with consistent performance gains.



Paperid:3731
Authors:Shaohang Wei, Wei Li, Feifan Song, Wen Luo, Tianyi Zhuang, Haochen Tan, Zhijiang Guo, Houfeng Wang
Abstract:
Temporal reasoning is pivotal for Large Language Models (LLMs) to comprehend the real world. However, existing works neglect the real-world challenges for temporal reasoning: (1) intensive temporal information, (2) fast-changing event dynamics, and (3) complex temporal dependencies in social interactions. To bridge this gap, we propose a multi-level benchmark TimE, designed for temporal reasoning in real-world scenarios. TimE consists of 38,522 QA pairs, covering 3 levels with 11 fine-grained sub-tasks. This benchmark encompasses 3 sub-datasets reflecting different real-world challenges: TimE-Wiki, TimE-News, and TimE-Dial. We conduct extensive experiments on reasoning models and non-reasoning models. And we conducted an in-depth analysis of temporal reasoning performance across diverse real-world scenarios and tasks, and summarized the impact of test-time scaling on temporal reasoning capabilities. Additionally, we release TimE-Lite, a human-annotated subset to foster future research and standardized evaluation in temporal reasoning.



Paperid:3732
Authors:Marcos Negre Saura, Richard Allmendinger, Wei Pan, Theodore Papamarkou
Abstract:
This paper explores the integration of ring attractors, a mathematical model inspired by neural circuit dynamics, into the Reinforcement Learning (RL) action selection process. Serving as specialized brain-inspired structures that encode spatial information and uncertainty, ring attractors offer a biologically plausible mechanism to improve learning speed and accuracy in RL. They do so by explicitly encoding the action space, facilitating the organization of neural activity, and enabling the distribution of spatial representations across the neural network in the context of Deep Reinforcement Learning (DRL). For example, preserving the continuity between rotation angles in robotic control or adjacency between tactical moves in game-like environments. The application of ring attractors in the action selection process involves mapping actions to specific locations on the ring and decoding the selected action based on neural activity. We investigate the application of ring attractors by both building an exogenous model and integrating them as part of DRL agents. Our approach significantly improves state-of-the-art performance on the Atari 100k benchmark, achieving a 53% increase in performance across selected state-of-the-art baselines.



Paperid:3733
Authors:Hua Ye, Hang Ding, Siyuan Chen, Yiyang Jiang, changyuan zhang, Xuan Zhang
Abstract:
Abstract:Most multimodal models treat every negative pair alike, ignoring the ambiguous negatives that differ from the positive by only a small detail. We propose Boundary-A ware Curriculum with Local Attention(BACL), a lightweight add-on that turns these borderline cases into a curriculum signal. A Boundary-aware Negative Sampler gradually raises difficulty, while a Contrastive Local Attention loss highlights where the mismatch occurs. The two modules are fully differentiable and work with any off-the-shelf dual encoder. Theory predicts a fast $\tilde{\mathcal{O}}(1/n)$ error rate; practice shows up to +32 \% R@1 over CLIP and new SOTA on four large-scale benchmarks, all without extra labels.



Paperid:3734
Authors:Achleshwar Luthra, Tianbao Yang, Tomer Galanti
Abstract:
Abstract:Despite its empirical success, the theoretical foundations of self-supervised contrastive learning (CL) are not yet fully established. In this work, we address this gap by showing that standard CL objectives implicitly approximate a supervised variant we call the negatives-only supervised contrastive loss (NSCL), which excludes same-class contrasts. We prove that the gap between the CL and NSCL losses vanishes as the number of semantic classes increases, under a bound that is both label-agnostic and architecture-independent.We characterize the geometric structure of the global minimizers of the NSCL loss: the learned representations exhibit augmentation collapse, within-class collapse, and class centers that form a simplex equiangular tight frame. We further introduce a new bound on the few-shot error of linear-probing. This bound depends on two measures of feature variability—within-class dispersion and variation along the line between class centers. We show that directional variation dominates the bound and that the within-class dispersion's effect diminishes as the number of labeled samples increases. These properties enable CL and NSCL-trained representations to support accurate few-shot label recovery using simple linear probes.Finally, we empirically validate our theoretical findings: the gap between CL and NSCL losses decays at a rate of $\mathcal{O}(\frac{1}{\#\text{classes}})$; the two losses are highly correlated; minimizing the CL loss implicitly brings the NSCL loss close to the value achieved by direct minimization; and the proposed few-shot error bound provides a tight estimate of probing performance in practice.



Paperid:3735
Authors:Junyu Huang, Qilong Feng, Zhen Zhang, Beirong Cui, Jianxin Wang
Abstract:
Abstract:Local search is a powerful clustering technique that provides high-quality solutions with theoretical guarantees. With distance-based sampling strategies, local search methods can achieve constant approximations for clustering with linear running time in data size. Despite their effectiveness, existing algorithms still face scalability issues as they require scanning the entire dataset for iterative center swaps. This typically leads to an $O(ndk)$ running time, where $n$ is the data size, $d$ is the dimension, $k$ is the number of clusters. To further improve the efficiency of local search algorithms, we propose adaptive sampling and bandit-based methods. Specifically, the adaptive sampling method can well approximate the distance-based sampling distribution without maintaining pairwise distances between data points and the centers, enabling fast and accurate sampling in sublinear time after an $\tilde{O}(nd)$ time preprocessing step. The bandit-based method models the best swap pair selection as a bandit problem, where a grouping strategy is proposed for fast identification of the optimal swap pair. With these techniques, our proposed algorithm can achieve constant approximation in expected running time $\tilde{O}(nd + k^4 d)$ under mild assumptions about optimal cluster and swap pair distributions. Our approach also extends naturally to the $k$-median problem, achieving constant approximation in expected running time $\tilde{O}(nd + \sqrt{n}k^3d)$ without distributional assumptions. Empirical results demonstrate that our algorithm achieves up to 1000× speedup over existing local search methods on datasets with 100 million points, while delivering comparable clustering quality. Compared to coreset-based approaches, it provides up to 80× speedup and consistently yields better clustering results.



Paperid:3736
Authors:Ido Cohen, Ronen Talmon
Abstract:
We propose a novel functorial graph coarsening method that preserves inner products between node features, a property often overlooked by existing approaches focusing primarily on structural fidelity.By treating node features as functions on the graph and preserving their inner products, our method retains both structural and feature relationships, facilitating substantial benefits for downstream tasks. To formalize this, we introduce the Inner Product Error (IPE), which quantifies how the inner products between node features are preserved. Leveraging the underlying geometry of the problem on the Grassmann manifold, we formulate an optimization objective that minimizes the IPE, also for unseen smooth functions. We show that minimizing the IPE improves standard coarsening metrics, and illustrate our method’s properties through visual examples that highlight its clustering ability. Empirical results on benchmarks for graph coarsening and node classification show that our approach outperforms existing state-of-the-art methods.



Paperid:3737
Authors:Chenyang Le, Bing Han, Jinshun Li, Songyong Chen, Yanmin Qian
Abstract:
Simultaneous Speech Translation (SimulST) enables real-time cross-lingual communication by jointly optimizing speech recognition and machine translation under strict latency constraints. Existing systems struggle to balance translation quality, latency, and semantic coherence, particularly in multilingual many-to-many scenarios where divergent read/write policies hinder unified strategy learning. In this paper, we present SimulMEGA(Simultaneous Generation by Mixture-of-Experts GAting), an unsupervised policy learning framework that combines prefix-based training with a Mixture-of-Experts refiner to learn effective read/write decisions in an implicit manner, without adding inference-time overhead. Our design requires only minimal modifications to standard transformer architectures and generalizes across both speech-to-text and text-to-speech streaming tasks. Through comprehensive evaluation on six language pairs, our 500 M-parameter speech-to-text model outperforms the Seamless baseline, achieving under 7% BLEU degradation at 1.5 s average lag and under 3% at 3 s. We further demonstrate SimulMEGA’s versatility by extending it to streaming TTS via a unidirectional backbone, yielding superior latency–quality trade-offs.



Paperid:3738
Authors:Manyi Yao, Abhishek Aich, Bingbing Zhuang, Sparsh Garg, Amit Roy-Chowdhury, Christian Shelton, Manmohan Chandraker
Abstract:
Grounding large language models (LLMs) in domain-specific tasks like post-hoc dash-cam driving video analysis is challenging due to their general-purpose training and lack of structured inductive biases. As vision is often the sole modality available for such analysis (i.e., no LiDAR, GPS, etc.), existing video-based vision-language models (V-VLMs) struggle with spatial reasoning, causal inference, and explainability of events in the input video. To this end, we introduce iFinder, a structured semantic grounding framework that decouples perception from reasoning by translating dash-cam videos into a hierarchical, interpretable data structure for LLMs. iFinder operates as a modular, training-free pipeline that employs pretrained vision models to extract critical cues—object pose, lane positions, and object trajectories—which are hierarchically organized into frame- and video-level structures. Combined with a three-block prompting strategy, it enables step-wise, grounded reasoning for the LLM to refine a peer V-VLM's outputs and provide accurate reasoning.Evaluations on four public dash-cam video benchmarks show that iFinder's proposed grounding with domain-specific cues—especially object orientation and global context—significantly outperforms end-to-end V-VLMs on four zero-shot driving benchmarks, with up to 39% gains in accident reasoning accuracy. By grounding LLMs with driving domain-specific representations, iFinder offers a zero-shot, interpretable, and reliable alternative to end-to-end V-VLMs for post-hoc driving video understanding.



Paperid:3739
Authors:junyan ye, DONGZHI JIANG, Jun He, Baichuan Zhou, Zilong Huang, Zhiyuan Yan, Hongsheng Li, Conghui He, Weijia Li
Abstract:
Recently, Multimodal Large Language Models (MLLMs) have made rapid progress, particularly in enhancing their reasoning capabilities. However, existing reasoning benchmarks still primarily assess language-based reasoning, often treating visual input as replaceable context. To address this gap, we introduce BLINK-Twice, a vision-centric reasoning benchmark grounded in challenging perceptual tasks. Instead of relying on external knowledge, our tasks require models to reason from visual content alone, shifting the focus from language-based to image-grounded reasoning. Compared to prior perception benchmarks, it moves beyond shallow perception ("see") and requires fine-grained observation and analytical reasoning ("observe"). BLINK-Twice integrates three core components: seven types of visual challenges for testing visual reasoning, natural adversarial image pairs that enforce reliance on visual content, and annotated reasoning chains for fine-grained evaluation of the reasoning process rather than final answers alone. We evaluate 20 leading MLLMs, including 12 foundation models and 8 reasoning-enhanced models. BLINK-Twice poses a significant challenge to current models. While existing reasoning strategies in the language space—such as chain-of-thought or self-criticism can improve performance, they often result in unstable and redundant reasoning. We observe that repeated image observation improves performance across models, and active visual interaction, as demonstrated by models like o3, highlights the need for a new paradigm for vision reasoning. The dataset is publicly available at https://huggingface.co/datasets/PicoTrex/BLINK-Twice.



Paperid:3740
Authors:Yangshi Ge, Yiwei Bao, Feng Lu
Abstract:
In recent years, face-based deep-learning gaze estimation methods have achieved significant advancements. However, while face images provide supplementary information beneficial for gaze inference, the substantial extraneous information they contain also increases the risk of overfitting during model training and compromises generalization capability. To alleviate this problem, we propose the 3DPE-Gaze framework, explicitly modeling 3D facial priors for feature decoupling and generalized gaze estimation. The 3DPE-Gaze framework consists of two core modules: the 3D Geometric Prior module (3DGP) incorporating the FLAME model to parameterize facial structures and gaze-irrelevant facial appearances while extracting gaze features; the Semantic Concept Adversarial Module (SCAM) separates gaze-related and unrelated concepts through CLIP-guided contrastive learning. Finally, the 3DPE-Gaze framework combines 3D facial landmark as prior for generalized gaze estimation. Experimental results show that 3DPE-Gaze outperforms existing state-of-the-art methods on four major cross-domain tasks, with particularly outstanding performance in challenging scenarios such as lighting variations, extreme head poses, and glasses occlusion.



Paperid:3741
Authors:Korneel Van den Berghe, Stein Stroobants, Vijay Janapa Reddi, Guido de Croon
Abstract:
Neuromorphic computing systems are set to revolutionize energy-constrained robotics by achieving orders-of-magnitude efficiency gains while enabling native processing of temporal information. Spiking Neural Networks (SNNs) represent a promising algorithmic approach for these systems, yet their application to complex control tasks faces two critical challenges: (1) the non-differentiable nature of spiking neurons necessitates surrogate gradients with unclear optimization properties, and (2) the stateful dynamics of SNNs require training on sequences, which in reinforcement learning (RL) is hindered by limited sequence lengths during early training, preventing the network from bridging its warm-up period.We address these challenges by first providing a systematic analysis of surrogate gradient slope settings, demonstrating that shallower surrogate gradient slope settings increase gradient magnitude in deeper layers while reducing alignment with true gradients. In supervised learning, we find a slight preference of using steep slopes or scheduled slope settings, but the effect is much more pronounced in RL settings, where shallower slopes or scheduled slopes lead to a 2.1x improvement in both training performance and final deployed performance. Next, we propose a novel training approach that leverages a privileged guiding policy to bootstrap the learning process, while still exploiting online environment interactions. Combining our method with an adaptive slope schedule for a drone position control task, we achieve an average return of 400 points, substantially outperforming prior techniques, including behavioral cloning and TD3+BC, which achieve at most –200 points under the same conditions. This work advances both the theoretical understanding of surrogate gradients in SNNs and practical training methodologies for neuromorphic controllers demonstrated in real-world robotic systems.



Paperid:3742
Authors:Ibrahim Ethem Hamamci, Sezgin Er, Suprosanna Shit, Hadrien Reynaud, Dong Yang, Pengfei Guo, Marc Edgar, Daguang Xu, Bernhard Kainz, bjoern menze
Abstract:
Abstract:Recent progress in vision-language modeling for 3D medical imaging has been fueled by large-scale computed tomography (CT) corpora with paired free-text reports, stronger architectures, and powerful pretrained models. This has enabled applications such as automated report generation and text-conditioned 3D image synthesis. Yet, current approaches struggle with high-resolution, long-sequence volumes: contrastive pretraining often yields vision encoders that are misaligned with clinical language, and slice-wise tokenization blurs fine anatomy, reducing diagnostic performance on downstream tasks. We introduce BTB3D (Better Tokens for Better 3D), a causal convolutional encoder-decoder that unifies 2D and 3D training and inference while producing compact, frequency-aware volumetric tokens. A three-stage training curriculum enables (i) local reconstruction, (ii) overlapping-window tiling, and (iii) long-context decoder refinement, during which the model learns from short slice excerpts yet generalizes to scans exceeding $300$ slices without additional memory overhead. BTB3D sets a new state-of-the-art on two key tasks: it improves BLEU scores and increases clinical F1 by 40\% over CT2Rep, CT-CHAT, and Merlin for report generation; and it reduces FID by 75\% and halves FVD compared to GenerateCT and MedSyn for text-to-CT synthesis, producing anatomically consistent $512\times512\times241$ volumes. These results confirm that precise three-dimensional tokenization, rather than larger language backbones alone, is essential for scalable vision-language modeling in 3D medical imaging.



Paperid:3743
Authors:Junpyo Seo, HanbinKoo, jieun yook, Byung-Ro Moon
Abstract:
We propose a novel diffusion-based framework for automatic colorization of Anime-style facial sketches, which preserves the structural fidelity of the input sketch while effectively transferring stylistic attributes from a reference image. Our approach builds upon recent continuous-time diffusion models, but departs from traditional methods that rely on predefined noise schedules, which often fail to maintain perceptual consistency across the generative trajectory. To address this, we introduce SSIMBaD (Sigma Scaling with SSIM-Guided Balanced Diffusion), a sigma-space transformation that ensures linear alignment of perceptual degradation, as measured by structural similarity. This perceptual scaling enforces uniform visual difficulty across timesteps, enabling more balanced and faithful reconstructions. Experiments on a large-scale Anime face dataset show that our method significantly outperforms state-of-the-art (SOTA) models in terms of both pixel-level accuracy and perceptual quality, while generalizing robustly to diverse styles and structural variations.



Paperid:3744
Authors:Aiden Chang, Celso de Melo, Stephanie Lukin
Abstract:
Real-time understanding of continuous video streams is essential for intelligent agents operating in high-stakes environments, including autonomous vehicles, surveillance drones, and disaster response robots. Yet, most existing video understanding and highlight detection methods assume access to the entire video during inference, making them unsuitable for online or streaming scenarios. In particular, current models optimize for offline summarization, failing to support step-by-step reasoning needed for real-time decision-making. We introduce Aha!, an autoregressive highlight detection framework that predicts the relevance of each video frame against a task described in natural language. Without accessing future video frames, Aha! utilizes a multimodal language-vision model and lightweight, decoupled heads trained on a large, curated dataset of human-centric video labels. To enable scalability, we adopt a fixed-size SinkCache mechanism that achieves constant memory usage across infinite-length streams without degrading performance on standard benchmarks. This encourages the hidden representation to capture high-level task objectives, enabling effective frame-level rankings for informativeness, relevance, and uncertainty with respect to the natural language task. Aha! achieves state-of-the-art performance on highlight detection benchmarks, surpassing prior full-context and video-language models by +5.5\% on TVSum and +8.3\% on Mr. HiSum in mAP. We explore Aha!’s potential for real-world robotics applications given a task-oriented natural language input and a continuous, robot-centric video. Both experiments demonstrate Aha!'s potential effectiveness as a real-time reasoning module for downstream planning and long-horizon understanding.



Paperid:3745
Authors:David Chanin, James Wilken-Smith, Tomáš Dulka, Hardik Bhatnagar, Satvik Golechha, Joseph Bloom
Abstract:
Sparse Autoencoders (SAEs) aim to decompose the activation space of large language models (LLMs) into human-interpretable latent directions or features. As we increase the number of features in the SAE, hierarchical features tend to split into finer features (“math” may split into “algebra”, “geometry”, etc.), a phenomenon referred to as feature splitting. However, we show that sparse decomposition and splitting of hierarchical features is not robust. Specifically, we show that seemingly monosemantic features fail to fire where they should, and instead get “absorbed” into their children features. We coin this phenomenon feature absorption, and show that it is caused by optimizing for sparsity in SAEs whenever the underlying features form a hierarchy. We introduce a metric to detect absorption in SAEs, and validate our findings empirically on hundreds of LLM SAEs. Our investigation suggests that varying SAE sizes or sparsity is insufficient to solve this issue. We discuss the implications of feature absorption in SAEs and some potential approaches to solve the fundamental theoretical issues before SAEs can be used for interpreting LLMs robustly and at scale.



Paperid:3746
Authors:Shuhong Liu, Lin Gu, Ziteng Cui, Xuangeng Chu, Tatsuya Harada
Abstract:
Participating in efforts to endow generative AI with the 3D physical world perception, we propose I2-NeRF, a novel neural radiance field framework that enhances isometric and isotropic metric perception under media degradation. While existing NeRF models predominantly rely on object-centric sampling, I2-NeRF introduces a reverse-stratified upsampling strategy to achieve near-uniform sampling across 3D space, thereby preserving isometry. We further present a general radiative formulation for media degradation that unifies emission, absorption, and scattering into a particle model governed by the Beer–Lambert attenuation law. By matting direct and media-induced in-scatter radiance, this formulation extends naturally to complex media environments such as underwater, haze, and even low-light scenes. By treating light propagation uniformly in both vertical and horizontal directions, I2-NeRF enables isotropic metric perception and can even estimate medium properties such as water depth. Experiments on real-world datasets demonstrate that our method significantly improves both reconstruction fidelity and physical plausibility compared to existing approaches. The source code will be released.



Paperid:3747
Authors:Chenlong Deng, Zhisong Zhang, Kelong Mao, Shuaiyi Li, Tianqing Fang, Hongming Zhang, Haitao Mi, Dong Yu, Zhicheng Dou
Abstract:
Large language models are increasingly capable of handling long-context inputs, but the memory overhead of KV cache remains a major bottleneck for general-purpose deployment. While many compression strategies have been explored, sequence-level compression is particularly challenging due to its tendency to lose important details. We present UniGist, a gist token-based long context compression framework that removes the need for chunk-wise training, enabling the model to learn how to compress and utilize long-range context during training. To fully exploit the sparsity, we introduce a gist shift trick that transforms the attention layout into a right-aligned block structure and develop a block-table-free sparse attention kernel based on it. UniGist further supports one-pass training and flexible chunk sizes during inference, allowing efficient and adaptive context processing. Experiments across multiple long-context tasks show that UniGist significantly improves compression quality, with especially strong performance in recalling details and long-range dependency modeling.



Paperid:3748
Authors:Wei Zhu, Zhiwen Tang, Kun Yue
Abstract:
Recent advancements have increasingly focused on leveraging large language models (LLMs) to construct autonomous agents for complex problem-solving tasks. However, existing approaches predominantly employ a single-agent framework to generate search branches and estimate rewards during Monte Carlo Tree Search (MCTS) planning. This single-agent paradigm inherently limits exploration capabilities, often resulting in insufficient diversity among generated branches and suboptimal planning performance.To overcome these limitations, we propose \textbf{SYMPHONY}, \textbf{SY}nergistic \textbf{M}ulti-agent \textbf{P}lanning with \textbf{H}eter\textbf{O}geneous la\textbf{N}gauge model assembl\textbf{Y}, a novel multi-agent planning framework that integrates a pool of heterogeneous language model-based agents. By leveraging diverse reasoning patterns across agents, SYMPHONY enhances rollout diversity and facilitates more effective exploration.Empirical results across multiple benchmark tasks show that SYMPHONY achieves strong performance even when instantiated with open-source LLMs deployable on consumer-grade hardware. When enhanced with cloud-based LLMs accessible via API, SYMPHONY demonstrates further improvements, outperforming existing state-of-the-art baselines and underscoring the effectiveness of heterogeneous multi-agent coordination in planning tasks.



Paperid:3749
Authors:Congyu Qiao, Ning Xu, Yihao Hu, Xin Geng
Abstract:
Instance-dependent Partial Label Learning (ID-PLL) aims to learn a multi-class predictive model given training instances annotated with candidate labels related to features, among which correct labels are hidden fixed but unknown. The previous works involve leveraging the identification capability of the training model itself to iteratively refine supervision information. However, these methods overlook a critical aspect of ID-PLL: within the original label space, the model may fail to distinguish some incorrect candidate labels that are strongly correlated with features from correct labels. This leads to poor-quality supervision signals and creates a bottleneck in the training process. In this paper, we propose to leverage reduction-based pseudo-labels to alleviate the influence of incorrect candidate labels and train our predictive model to overcome this bottleneck. Specifically, reduction-based pseudo-labels are generated by performing weighted aggregation on the outputs of a multi-branch auxiliary model, with each branch trained in a label subspace that excludes certain labels. This approach ensures that each branch explicitly avoids the disturbance of the excluded labels, allowing the pseudo-labels provided for instances troubled by these excluded labels to benefit from the unaffected branches. Theoretically, we demonstrate that reduction-based pseudo-labels exhibit greater consistency with the Bayes optimal classifier compared to pseudo-labels directly generated from the training predictive model.



Paperid:3750
Authors:Junsheng Zhou, XingYu Shi, Haichuan Song, Yi Fang, Yu-Shen Liu, Zhizhong Han
Abstract:
Point clouds captured by scanning sensors are often perturbed by noise, which have a highly negative impact on downstream tasks (e.g. surface reconstruction and shape understanding). Previous works mostly focus on training neural networks with noisy-clean point cloud pairs for learning denoising priors, which requires extensively manual efforts. In this work, we introduce U-CAN, an Unsupervised framework for point cloud denoising with Consistency-Aware Noise2Noise matching. Specifically, we leverage a neural network to infer a multi-step denoising path for each point of a shape or scene with a noise to noise matching schema. We achieve this by a novel loss which enables statistical reasoning on multiple noisy point cloud observations. We further introduce a novel constraint on the denoised geometry consistency for learning consistency-aware denoising patterns. We justify that the proposed constraint is a general term which is not limited to 3D domain and can also contribute to the area of 2D image denoising. Our evaluations under the widely used benchmarks in point cloud denoising, upsampling and image denoising show significant improvement over the state-of-the-art unsupervised methods, where U-CAN also produces comparable results with the supervised methods.



Paperid:3751
Authors:Xiangyu Wang, Donglin Yang, Yue Liao, Wenhao Zheng, wenjun wu, Bin Dai, Hongsheng Li, Si Liu
Abstract:
Unmanned Aerial Vehicles (UAVs) are evolving into language-interactive platforms, enabling more intuitive forms of human-drone interaction. While prior works have primarily focused on high-level planning and long-horizon navigation, we shift attention to language-guided fine-grained trajectory control, where UAVs execute short-range, reactive flight behaviors in response to language instructions. We formalize this problem as the Flying-on-a-Word (Flow) task and introduce UAV imitation learning as an effective approach. In this framework, UAVs learn fine-grained control policies by mimicking expert pilot trajectories paired with atomic language instructions. To support this paradigm, we present UAV-Flow, the first real-world benchmark for language-conditioned, fine-grained UAV control. It includes a task formulation, a large-scale dataset collected in diverse environments, a deployable control framework, and a simulation suite for systematic evaluation. Our design enables UAVs to closely imitate the precise, expert-level flight trajectories of human pilots and supports direct deployment without sim-to-real gap. We conduct extensive experiments on UAV-Flow, benchmarking VLN and VLA paradigms. Results show that VLA models are superior to VLN baselines and highlight the critical role of spatial grounding in the fine-grained Flow setting.



Paperid:3752
Authors:Yuhan Helena Liu, Victor Geadah, Jonathan Pillow
Abstract:
Understanding how animals learn is a central challenge in neuroscience, with growing relevance to the development of animal- or human-aligned artificial intelligence. However, most existing approaches to this problem assume specific parametric forms for the learning rule (e.g., Q-learning, policy gradient) or are limited to simplified settings like bandit tasks, which do not involve learning a new input–output mapping from scratch. In contrast, animals must often learn to perform new behaviors de novo, which presents a rich challenge for learning rule inference. Here, we propose a flexible approach for inferring learning rules directly from animal decision-making data during de novo task learning. Specifically, we use a deep neural network (DNN) to parameterize a learning rule that governs the update of an animal's policy weights after each trial. We validate our approach by showing that this model can recover specific ground-truth learning rules in simulated datasets. We then consider more complex learning rules by incorporating a recurrent neural network (RNN), which allows per-trial weight updates to depend on the history of stimuli, decisions, and rewards over multiple trials. We applied our methods to a large behavioral dataset of mice learning a sensory decision-making task over multiple weeks and observed improved predictions on held-out data. The inferred learning rules revealed that mice make substantially different weight updates after correct and incorrect trials, and that learning is non-Markovian, with strong dependencies on history over multiple trials. Overall, these results provide a flexible new framework for inferring biological learning rules from behavioral data and yield powerful insights into the features of de novo task learning.



Paperid:3753
Authors:Liang Zhang, Justin Lieffers, Adarsh Pyarelal
Abstract:
In this paper, we explore semantic clustering properties of deep reinforcement learning (DRL) to improve its interpretability and deepen our understanding of its internal semantic organization. In this context, semantic clustering refers to the ability of neural networks to cluster inputs based on their semantic similarity in the internal space. We propose a DRL architecture that incorporates a novel semantic clustering module that combines feature dimensionality reduction with online clustering. This module integrates seamlessly into the DRL training pipeline, addressing the instability of t-SNE and eliminating the need for extensive manual annotation inherent to prior semantic analysis methods. We experimentally validate the effectiveness of the proposed module and demonstrate its ability to reveal semantic clustering properties within DRL. Furthermore, we introduce new analytical methods based on these properties to provide insights into the hierarchical structure of policies and semantic organization within the feature space.



Paperid:3754
Authors:jingnan zheng, Xiangtian Ji, Yijun Lu, Chenhang Cui, Weixiang Zhao, Gelei Deng, Zhenkai Liang, An Zhang, Tat-Seng Chua
Abstract:
Large Language Models (LLMs) continue to exhibit vulnerabilities despite deliberate safety alignment efforts, posing significant risks to users and society. To safeguard against the risk of policy-violating content, system-level moderation via external guard models—designed to monitor LLM inputs and outputs and block potentially harmful content—has emerged as a prevalent mitigation strategy. Existing approaches of training guard models rely heavily on extensive human curated datasets and struggle with out-of-distribution threats, such as emerging harmful categories or jailbreak attacks. To address these limitations, we propose RSafe, an adaptive reasoning-based safeguard that conducts guided safety reasoning to provide robust protection within the scope of specified safety policies. RSafe operatesin two stages: (1) guided reasoning, where it analyzes safety risks of input content through policy-guided step-by-step reasoning, and (2) reinforced alignment, where rule-based RL optimizes its reasoning paths to align with accurate safety prediction. This two-stage training paradigm enables RSafe to internalize safety principles to generalize safety protection capability over unseen or adversarial safety violationscenarios. During inference, RSafe accepts user-specified safety policies to provide enhanced safeguards tailored to specific safety requirements. Experiments demonstrate that RSafe matches state-of-the-art guard models using limited amount of public data in both prompt- and response-level harmfulness detection, while achieving superior out-of-distribution generalization on both emerging harmful category and jailbreak attacks. Furthermore, RSafe provides human-readable explanations for its safety judgments for better interpretability. RSafe offers a robust, adaptive, and interpretable solution for LLM safety moderation, advancing the development of reliable safeguards in dynamic real-world environments. Our code is available at https://anonymous.4open.science/r/RSafe-996D.



Paperid:3755
Authors:Wanying Dai, Beibei Li, Naipeng Dong, Guangdong Bai, Jin Song Dong
Abstract:
Face recognition is essential for identity authentication, but the rich visual clues in facial images pose significant privacy risks, highlighting the critical importance of privacy-preserving solutions. For instance, numerous studies have shown that generative models are capable of effectively performing reconstruction attacks that result in the restoration of original visual clues. To mitigate this threat, we introduce FracFace, a fractal-based privacy-preserving face recognition framework. This approach effectively weakens the visual clues that can be exploited by reconstruction attacks by disrupting the spatial structure in frequency domain features, while retaining the vital visual clues required for identity recognition. To achieve this, we craft a Frequency Channels Refining module that reduces sparsity in the frequency domain. It suppresses visual clues that could be exploited by reconstruction attacks, while preserving features indispensable for recognition, thus making these attacks more challenging. More significantly, we design a Frequency Fractal Mapping module that obfuscates deep representations by remapping refined frequency channels into a fractal-based privacy structure. By leveraging the self-similarity of fractals, this module preserves identity relevant features while enhancing defense capabilities, thereby improving the overall robustness of the protection scheme. Experiments conducted on multiple public face recognition benchmarks demonstrate that the proposed FracFace significantly reduces the visual recoverability of facial features, while maintaining high recognition accuracy, as well as the superiorities over state-of-the-art privacy protection approaches.



Paperid:3756
Authors:Sebastian Bruch, Aditya Krishnan, Franco Maria Nardini
Abstract:
Abstract:Clustering-based nearest neighbor search algorithms partition points into shards to form an index, and search only a subset of shards to process a query. Even though search efficacy is heavily influenced by the algorithm that identifies the shards to probe, it has received little attention in the literature. We study routing in clustering-based maximum inner product search, which includes cosine similarity search. We unpack existing routers and notice the surprising role of optimism. We then take a page from the sequential decision making literature and formalize that insight following the principle of ``optimism in the face of uncertainty.'' In particular, we present a framework that incorporates the moments of the distribution of inner products within each shard to estimate the maximum inner product. We then develop a practical instance of our algorithm that uses only the first two moments to reach the same accuracy as state-of-the-art routers by probing up to $50\%$ fewer points on benchmark datasets without compromising efficiency. Our algorithm is also space-efficient: we design a sketch of the second moment whose size is independent of the number of points and requires $\mathcal{O}(1)$ vectors per shard.



Paperid:3757
Authors:Xusheng Cao, Haori Lu, Linlan Huang, Fei Yang, Xialei Liu, Ming-Ming Cheng
Abstract:
Continual learning in computer vision faces the critical challenge of catastrophic forgetting, where models struggle to retain prior knowledge while adapting to new tasks.Although recent studies have attempted to leverage the generalization capabilities of pre-trained models to mitigate overfitting on current tasks, models still tend to forget details of previously learned categories as tasks progress, leading to misclassification. To address these limitations, we introduce a novel Knowledge Graph Enhanced Generative Multi-modal model (KG-GMM) that builds an evolving knowledge graph throughout the learning process. Our approach utilizes relationships within the knowledge graph to augment the class labels and assigns different relations to similar categories to enhance model differentiation. During testing, we propose a Knowledge Graph Augmented Inference method that locates specific categories by analyzing relationships within the generated text, thereby reducing the loss of detailed information about old classes when learning new knowledge and alleviating forgetting. Experiments demonstrate that our method effectively leverages relational information to help the model correct mispredictions, achieving state-of-the-art results in both conventional CIL and few-shot CIL settings, confirming the efficacy of knowledge graphs at preserving knowledge in the continual learning scenarios.



Paperid:3758
Authors:Jingyi Xu, Shengnan Wang, Weidong Yang, Keyi Liu, Yeqi Luo, Ben Fei, LEI BAI
Abstract:
Arctic sea ice performs a vital role in global climate and has paramount impacts on both polar ecosystems and coastal communities. In the last few years, multiple deep learning based pan-Arctic sea ice concentration (SIC) forecasting methods have emerged and showcased superior performance over physics-based dynamical models. However, previous methods forecast SIC at a fixed temporal granularity, e.g. sub-seasonal or seasonal, thus only leveraging inter-granularity information and overlooking the plentiful inter-granularity correlations. SIC at various temporal granularities exhibits cumulative effects and are naturally consistent, with short-term fluctuations potentially impacting long-term trends and long-term trends provides effective hints for facilitating short-term forecasts in Arctic sea ice. Therefore, in this study, we propose to cultivate temporal multi-granularity that naturally derived from Arctic sea ice reanalysis data and provide a unified perspective for modeling SIC via our Sea Ice Fusion framework. SIFusion is delicately designed to leverage both intra-granularity and inter-granularity information for capturing granularity-consistent representations that promote forecasting skills. Our extensive experiments show that SIFusion outperforms off-the-shelf deep learning models for their specific temporal granularity.



Paperid:3759
Authors:Yuxiang Wei, Olivier Duchenne, Jade Copet, Quentin Carbonneaux, LINGMING ZHANG, Daniel Fried, Gabriel Synnaeve, Rishabh Singh, Sida Wang
Abstract:
The recent DeepSeek-R1 release has demonstrated the immense potential of reinforcement learning (RL) in enhancing the general reasoning capabilities of large language models (LLMs). While DeepSeek-R1 and other follow-up work primarily focus on applying RL to competitive coding and math problems, this paper introduces SWE-RL, the first approach to scale RL-based LLM reasoning for real-world software engineering. Leveraging a lightweight rule-based reward (e.g., the similarity score between ground-truth and LLM-generated solutions), SWE-RL enables LLMs to autonomously recover a developer's reasoning processes and solutions by learning from extensive open-source software evolution data -- the record of a software's entire lifecycle, including its code snapshots, code changes, and events such as issues and pull requests. Trained on top of Llama 3, our resulting reasoning model, Llama3-SWE-RL-70B, achieves a 41.0% solve rate on SWE-bench Verified -- a human-verified collection of real-world GitHub issues. To our knowledge, this is the best performance reported for medium-sized (<100B) LLMs to date, even comparable to leading proprietary LLMs like GPT-4o. Surprisingly, despite performing RL solely on software evolution data, Llama3-SWE-RL has even emerged with generalized reasoning skills. For example, it shows improved results on five out-of-domain tasks, namely, function coding, library use, code reasoning, mathematics, and general language understanding, whereas a supervised-finetuning baseline even leads to performance degradation on average. Overall, SWE-RL opens up a new direction to improve the reasoning capabilities of LLMs through reinforcement learning on massive software engineering data.



Paperid:3760
Authors:Philip Schroeder, Ondrej Biza, Thomas Weng, Hongyin Luo, Jim Glass
Abstract:
Vision-language models (VLMs) have exhibited impressive capabilities across diverse image understanding tasks, but still struggle in settings that require reasoning over extended sequences of camera frames from a video. This limits their utility in embodied settings, which require reasoning over long frame sequences from a continuous stream of visual input at each moment of a task attempt. To address this limitation, we propose ROVER (Reasoning Over VidEo Recursively), a framework that enables the model to recursively decompose long-horizon video trajectories into segments corresponding to shorter subtasks within the trajectory. In doing so, ROVER facilitates more focused and accurate reasoning over temporally localized frame sequences without losing global context. We evaluate ROVER, implemented using an in-context learning approach, on diverse OpenX Embodiment videos and on a new dataset derived from RoboCasa that consists of 543 videos showing both expert and perturbed non-expert trajectories across 27 manipulation tasks. ROVER outperforms strong baselines across three video reasoning tasks: task progress estimation, frame-level natural language reasoning, and video question answering. We observe that, by reducing the number of frames the model reasons over at each timestep, ROVER mitigates model hallucinations, especially during unexpected or non-optimal moments of a trajectory. In addition, by enabling the implementation of a subtask-specific sliding context window, ROVER's time complexity scales linearly with video length, an asymptotic improvement over baselines.



Paperid:3761
Authors:Shayan Mohajer Hamidi, Ben Liang, EN-HUI YANG
Abstract:
Abstract:Inverse problems, where the goal is to recover an unknown signal from noisy or incomplete measurements, are central to applications in medical imaging, remote sensing, and computational biology. Diffusion models have recently emerged as powerful priors for solving such problems. However, existing methods either rely on projection-based techniques that enforce measurement consistency through heuristic updates, or they approximate the likelihood $p(\boldsymbol{y} \mid \boldsymbol{x})$, often resulting in artifacts and instability under complex or high-noise conditions.To address these limitations, we propose a novel framework called coupled data and measurement space diffusion posterior sampling (C-DPS), which eliminates the need for constraint tuning or likelihood approximation. C-DPS introduces a forward stochastic process in the measurement space $\{\boldsymbol{y}_t\}$, evolving in parallel with the data-space diffusion $\{\boldsymbol{x}_t\}$, which enables the derivation of a closed-form posterior $p(\boldsymbol{x}_{t-1} \mid \boldsymbol{x}_t, \boldsymbol{y}_{t-1})$. This coupling allows for accurate and recursive sampling based on a well-defined posterior distribution. Empirical results demonstrate that C-DPS consistently outperforms existing baselines, both qualitatively and quantitatively, across multiple inverse problem benchmarks. Code is available at \url{https://anonymous.4open.science/r/C-DPS-CF88}.



Paperid:3762
Authors:Feiyang Pan, Shenghe Zheng, Chunyan Yin, Guangbin Dou
Abstract:
MEMS gyroscopes play a critical role in inertial navigation and motion control applications but typically suffer from a fundamental trade-off between measurement range and noise performance. Existing hardware-based solutions aimed at mitigating this issue introduce additional complexity, cost, and scalability challenges. Deep-learning methods primarily focus on noise reduction and typically require precisely aligned ground-truth signals, making them difficult to deploy in practical scenarios and leaving the fundamental trade-off unresolved. To address these challenges, we introduce Mixture of Experts for MEMS Gyroscopes (MoE-Gyro), a novel self-supervised framework specifically designed for simultaneous over-range signal reconstruction and noise suppression. MoE-Gyro employs two experts: an Over‑Range Reconstruction Expert (ORE), featuring a Gaussian-Decay Attention mechanism for reconstructing saturated segments; and a Denoise Expert (DE), utilizing dual-branch complementary masking combined with FFT-guided augmentation for robust noise reduction. A lightweight gating module dynamically routes input segments to the appropriate expert. Furthermore, existing evaluation lack a comprehensive standard for assessing multi-dimensional signal enhancement. To bridge this gap, we introduce IMU Signal Enhancement Benchmark (ISEBench), an open-source benchmarking platform comprising the GyroPeak-100 dataset and a unified evaluation of IMU signal enhancement methods. We evaluate MoE-Gyro using our proposed ISEBench, demonstrating that our framework significantly extends the measurable range from ±450°/s to ±1500°/s, reduces Bias Instability by 98.4%, and achieves state-of-the-art performance, effectively addressing the long-standing trade-off in inertial sensing.



Paperid:3763
Authors:Ting Han, Linara Adilova, Henning Petzka, Jens Kleesiek, Michael Kamp
Abstract:
Neural collapse, i.e., the emergence of highly symmetric, class-wise clustered representations, is frequently observed in deep networks and is often assumed to reflect or enable generalization. In parallel, flatness of the loss landscape has been theoretically and empirically linked to generalization. Yet, the causal role of either phenomenon remains unclear: Are they prerequisites for generalization, or merely by-products of training dynamics? We disentangle these questions using grokking, a training regime in which memorization precedes generalization, allowing us to temporally separate generalization from training dynamics and we find that while both neural collapse and relative flatness emerge near the onset of generalization, only flatness consistently predicts it. Models encouraged to collapse or prevented from collapsing generalize equally well, whereas models regularized away from flat solutions exhibit delayed generalization, resembling grokking---even in architectures and datasets where it does not typically occur. Furthermore, we show theoretically that neural collapse implies relative flatness under classical assumptions, explaining their empirical co-occurrence. Our results support the view that relative flatness is a potentially necessary and more fundamental property for generalization, and demonstrate how grokking can serve as a powerful probe for isolating its geometric underpinnings.



Paperid:3764
Authors:Yana Wei, Liang Zhao, Jianjian Sun, Kangheng Lin, En Yu, jisheng yin, Jingcheng Hu, Yinmin Zhang, Qi Han, Haoran Lv, Zejia Weng, Jia Wang, Zheng Ge, Xiangyu Zhang, Daxin Jiang, Vishal Patel
Abstract:
Cognitive behaviors such as subgoal decomposition, reflection, and verification have emerged as key drivers of reasoning performance in large language models (LLMs) trained with reinforcement learning (RL). While recent studies have formalized these patterns and demonstrated their impact on generalization, it remains unclear whether such behaviors can transfer from language to vision in multimodal settings.In this work, we present a simple yet revealing observation: a vision-language model trained with language-only cold start exhibits surprising gains on visual reasoning tasks, along with emergent visual reflection behaviors. Motivated by DeepSeek-R1 and Open-Reasoner-Zero, we propose a three-stage training pipeline to systematically study cross-modal cognitive behavior transfer: (1) supervised fine-tuning on language-only reasoning data, (2) language-only RL to reinforce cognitive traits, and (3) multimodal RL for cross-modal adaptation.Our 7B model, \textbf{Open-Vision-Reasoner}, achieves strong performance across both language and vision benchmarks, including 94.1% on MATH500, 50.0% on MathVision, and 52.9% on MathVerse. Our visual cognitive behavior analysis further reveals the mechanisms behind these gains and provides a promising path toward general-purpose multimodal reasoning.



Paperid:3765
Authors:Mingyao Zhou, Hao Sun, Wei Xie, Ming Dong, Chengji Wang, Mang Ye
Abstract:
With the exponential growth of video content, aiming at localizing relevant video moments based on natural language queries, video moment retrieval (VMR) has gained significant attention. Existing weakly supervised VMR methods focus on designing various feature modeling and modal interaction modules to alleviate the reliance on precise temporal annotations. However, these methods have poor generalization capabilities on compositional queries with novel syntactic structures or vocabulary in real-world scenarios. To this end, we propose a new task: weakly supervised compositional moment retrieval (WSCMR). This task trains models using only video-query pairs without precise temporal annotations, while enabling generalization to complex compositional queries. Furthermore, a proposal-centric network (PC-Net) is proposed to tackle this challenging task. First, video and query features are extracted through frozen feature extractors, followed by modality interaction to obtain multimodal features. Second, to handle compositional queries with explicit temporal associations, a dual-granularity proposal generator decodes multimodal global and frame-level features to obtain query-relevant proposal boundaries with fine-grained temporal perception. Third, to improve the discrimination of proposal features, a proposal feature aggregator is constructed to conduct semantic alignment of frames and queries, and employ a learnable peak-aware Gaussian distributor to fit the frame weights within the proposals to derive proposal features from the frame features. Finally, the proposal quality is assessed based on the results of reconstructing the masked query using the obtained proposal features. To further enhance the model's ability to capture semantic associations between proposals and queries, a quality margin regularizer is constructed to dynamically stratify proposals into high and low query-relevance subsets and enhance the association between queries and common elements within proposals, and suppress spurious correlations via inter-subset contrastive learning. Notably, PC-Net achieves superior performance with 54\% fewer parameters than prior works by parameter-efficient design. Experiments on Charades-CG and ActivityNet-CG demonstrate PC-Net’s ability to generalize across diverse compositional queries.



Paperid:3766
Authors:Minseok Kang, Sangyoun Lee, Minhyeok Lee, Donghyeong Kim, Minjung Kim
Abstract:
Video Temporal Grounding (VTG) aims to localize temporal segments in long, untrimmed videos that align with a given natural language query. This task typically comprises two subtasks: \textit{Moment Retrieval (MR)} and \textit{Highlight Detection (HD)}. While recent advances have been progressed by powerful pretrained vision-language models such as CLIP and InternVideo2, existing approaches commonly treat all text tokens uniformly during cross-modal attention, disregarding their distinct semantic roles. To validate the limitations of this approach, we conduct controlled experiments demonstrating that VTG models overly rely on [EOS]-driven global semantics while failing to effectively utilize word-level signals, which limits their ability to achieve fine-grained temporal alignment. Motivated by this limitation, we propose DualGround, a dual-branch architecture that explicitly separates global and local semantics by routing the [EOS] token through a sentence-level path and clustering word tokens into phrase-level units for localized grounding. Our method introduces (1) token-role-aware cross modal interaction strategies that align video features with sentence-level and phrase-level semantics in a structurally disentangled manner, and (2) a joint modeling framework that not only improves global sentence-level alignment but also enhances fine-grained temporal grounding by leveraging structured phrase-aware context. This design allows the model to capture both coarse and localized semantics, enabling more expressive and context-aware video grounding. DualGround achieves state-of-the-art performance on both Moment Retrieval and Highlight Detection tasks across QVHighlights and Charades-STA benchmarks, demonstrating the effectiveness of disentangled semantic modeling in video-language alignment.



Paperid:3767
Authors:Amirmohammad Farzaneh, Osvaldo Simeone
Abstract:
The selection of hyperparameters, such as prompt templates in large language models (LLMs), must often strike a balance between reliability and cost. In many cases, structural relationships between the expected reliability levels of the hyperparameters can be inferred from prior information and held-out data -- e.g., longer prompt templates may be more detailed and thus more reliable. However, existing hyperparameter selection methods either do not provide formal reliability guarantees or are unable to incorporate structured knowledge in the hyperparameter space. This paper introduces reliability graph-based Pareto testing (RG-PT), a novel multi-objective hyperparameter selection framework that maintains formal reliability guarantees in terms of false discovery rate (FDR), while accounting for known relationships among hyperparameters via a directed acyclic graph. Edges in the graph reflect expected reliability and cost trade-offs among hyperparameters, which are inferred via the Bradley-Terry (BT) ranking model from prior information and held-out data. Experimental evaluations demonstrate that RG-PT significantly outperforms existing methods such as learn-then-test (LTT) and Pareto testing (PT) through a more efficient exploration of the hyperparameter space.



Paperid:3768
Authors:Yanqiao Zhu, Yidan Shi, Yuanzhou Chen, Fang Sun, Yizhou Sun, Wei Wang
Abstract:
Molecular representation learning has emerged as a promising approach for modeling molecules with deep learning in chemistry and beyond. While 3D geometric models effectively capture molecular structure, they typically process single static conformers, overlooking the inherent flexibility and dynamics of molecules. In reality, many molecular properties depend on distributions of thermodynamically accessible conformations rather than single structures. Recent works show that learning from conformer ensembles can improve molecular representations, but existing approaches either produce unphysical structures through averaging or require restrictive molecular alignment. In this paper, we propose SymmetryPreserving Conformer Ensemble networks (SPiCE), which introduces two key innovations: (1) geometric mixture-of-experts for selective processing of scalar and vector features, and (2) hierarchical ensemble encoding that combines ensemblelevel representation with cross-conformer integration. Crucially, SPiCE ensures physically meaningful representations by maintaining joint equivariance to geometric transformations of individual conformers and conformer permutations. Extensive experiments demonstrate that SPiCE consistently outperforms existing conformer ensemble methods and state-of-the-art structural aggregation models across quantum mechanical and biological property prediction tasks.



Paperid:3769
Authors:Zhanyi Sun, Shuran Song
Abstract:
Visuomotor policies trained via behavior cloning are vulnerable to covariate shift, where small deviations from expert trajectories can compound into failure. Common strategies to mitigate this issue involve expanding the training distribution through human-in-the-loop corrections or synthetic data augmentation. However, these approaches are often labor-intensive, rely on strong task assumptions, or compromise the quality of imitation. We introduce Latent Policy Barrier, a framework for robust visuomotor policy learning. Inspired by Control Barrier Functions, LPB treats the latent embeddings of expert demonstrations as an implicit barrier separating safe, in-distribution states from unsafe, out-of-distribution (OOD) ones. Our approach decouples the role of precise expert imitation and OOD recovery into two separate modules: a base diffusion policy solely on expert data, and a dynamics model trained on both expert and suboptimal policy rollout data. At inference time, the dynamics model predicts future latent states and optimizes them to stay within the expert distribution. Both simulated and real-world experiments show that LPB improves both policy robustness and data efficiency, enabling reliable manipulation from limited expert data and without additional human correction or annotation. More details are on our anonymous project website https://latentpolicybarrier.github.io.



Paperid:3770
Authors:YINYI WEI, Xiao LI
Abstract:
We present Text2MBL, a text-to-code generation framework that generates executable Building Information Modeling (BIM) code directly from textual descriptions of modular building layout (MBL) design. Unlike conventional layout generation approaches that operate in 2D space, Text2MBL produces fully parametric, semantically rich BIM layouts through on‑the‑fly code instantiation. To address MBLs' unique challenges due to their hierarchical three-tier structure: modules (physical building blocks), units (self-contained dwellings), and rooms (functional spaces), we developed an object-oriented code architecture and fine-tuned large language models to output structured action sequences in code format. To train and evaluate the framework, we curated a dataset of paired descriptions and ground truth layouts drawn from real‑world modular housing projects. Performance were assessed using metrics for executable validity, semantic fidelity, and geometric consistency. By tightly unifying natural language understanding with BIM code generation, Text2MBL establishes a scalable pipeline from high-level conceptual design to automation-ready modular construction workflows. Our implementation is available at https://anonymous.4open.science/r/Text2MBL-EE38/README.md.



Paperid:3771
Authors:Futoshi Futami, Masahiro Fujisawa
Abstract:
Latent variables (LVs) play a crucial role in encoder-decoder models by enabling effective data compression, prediction, and generation. Although their theoretical properties, such as generalization, have been extensively studied in supervised learning, similar analyses for unsupervised models such as variational autoencoders (VAEs) remain insufficiently underexplored. In this work, we extend information-theoretic generalization analysis to vector-quantized (VQ) VAEs with discrete latent spaces, introducing a novel data-dependent prior to rigorously analyze the relationship among LVs, generalization, and data generation. We derive a novel generalization error bound of the reconstruction loss of VQ-VAEs, which depends solely on the complexity of LVs and the encoder, independent of the decoder. Additionally, we provide the upper bound of the 2-Wasserstein distance between the distributions of the true data and the generated data, explaining how the regularization of the LVs contributes to the data generation performance.



Paperid:3772
Authors:Chantal Shaib, Vinith Suriyakumar, Byron Wallace, Marzyeh Ghassemi
Abstract:
Abstract:For an LLM to correctly respond to an instruction it must understand both the semantics and the domain (i.e., subject area) of a given task-instruction pair. However, syntax can also convey implicit information. Recent work shows that syntactic templates---frequent sequences of Part-of-Speech (PoS) tags---are prevalent in training data and often appear in model outputs. In this work we characterize syntactic templates, domain, and semantics in task-instruction pairs. We identify cases of spurious correlations between syntax and domain, where models learn to associate a domain with syntax during training; this can sometimes override prompt semantics. Using a synthetic training dataset, we find that the syntactic-domain correlation can lower performance (mean 0.51$\pm0.06$) on entity knowledge tasks in OLMo-2 models (1B-13B). We introduce an evaluation framework to detect this phenomenon in trained models, and show that it occurs on a subset of the FlanV2 dataset in open (OLMo-2-7B; Llama-4-Maverick), and closed (GPT-4o) models. Finally, we present a case study on the implications for safety finetuning, showing that unintended syntactic-domain correlations can be used to bypass refusals in OLMo-2-7B Instruct and GPT-4o. Our findings highlight two needs: (1) to explicitly test for syntactic-domain correlations, and (2) to ensure syntactic diversity in training data, specifically within domains, to prevent such spurious correlations.



Paperid:3773
Authors:Ben Cohen, Emaad Khwaja, Youssef Doubli, Salahidine Lemaachi, Chris Lettieri, Charles Masson, Hugo Miccinilli, Elise Ramé, Qiqi Ren, Afshin Rostamizadeh, Jean du Terrail, Anna-Monica Toon, Kan Wang, Stephan Xie, Zongzhe Xu, Viktoriya Zhukova, David Asker, Ameet Talwalkar, Othmane Abou-Amal
Abstract:
Abstract:We introduce Toto, a time series forecasting foundation model with 151 million parameters. Toto uses a modern decoder-only architecture coupled with architectural innovations designed to account for specific challenges found in multivariate observability time series data. Toto's pre-training corpus is a mixture of observability data, open datasets, and synthetic data, and is 4-10$\times$ larger than those of leading time series foundation models. Additionally, we introduce BOOM, a large-scale benchmark consisting of 350 million observations across 2,807 real-world time series. For both Toto and BOOM, we source observability data exclusively from our own telemetry and internal observability metrics. Extensive evaluations demonstrate that Toto achieves state-of-the-art performance on both BOOM and on established general purpose time series forecasting benchmarks. Toto's model weights, inference code, and evaluation scripts, as well as BOOM's data and evaluation code, are all available as open source under the Apache 2.0 License.



Paperid:3774
Authors:Tianle Li, Chenyang Zhang, Xingwu Chen, Yuan Cao, Difan Zou
Abstract:
Transformer-based Large Language Models (LLMs) have demonstrated powerful in-context learning capabilities. However, their predictions can be disrupted by factually correct context, a phenomenon known as context hijacking, revealing a significant robustness issue. To understand this phenomenon theoretically, we explore an in-context linear classification problem based on recent advances in linear transformers. In our setup, context tokens are designed as factually correct query-answer pairs, where the queries are similar to the final query but have opposite labels. Then, we develop a general theoretical analysis on the robustness of the linear transformers, which is formulated as a function of the model depth, training context lengths, and number of hijacking context tokens. A key finding is that a well-trained deeper transformer can achieve higher robustness, which aligns with empirical observations. We show that this improvement arises because deeper layers enable more fine-grained optimization steps, effectively mitigating interference from context hijacking. This is also well supported by our numerical experiments. Our findings provide theoretical insights into the benefits of deeper architectures and contribute to enhancing the understanding of transformer architectures.



Paperid:3775
Authors:Linye Li, Yufei Chen, Xiaodong Yue
Abstract:
Uncertainty estimation is crucial for ensuring the reliability of machine learning models in safety-critical applications. Evidential Deep Learning (EDL) offers a principled framework by modeling predictive uncertainty through Dirichlet distributions over class probabilities. However, existing EDL methods predominantly rely on level-0 hard labels, which supervised a uncertainty-aware model with full certainty. We argue that hard labels not only fail to capture epistemic uncertainty but also obscure the aleatoric uncertainty arising from inherent data noise and label ambiguity. As a result, EDL models often produce degenerate Dirichlet distributions that collapse to near-deterministic outputs. To overcome these limitations, we propose a vicinal risk minimization paradigm for EDL by incorporating level-1 supervision in the form of vicinally smoothed conditional label distributions. This richer supervision exposes the model to local label uncertainty, enhancing aleatoric uncertainty quantification, while also mitigating the degeneration of the Dirichlet distribution into a Dirac delta function, thereby improving epistemic uncertainty modeling.Extensive experiments show that our approach consistently outperforms standard EDL baselines across synthetic datasets, covariate-shifted out-of-distribution generalization tasks, and out-of-distribution detection benchmarks, providing more reliable uncertainty estimates.



Paperid:3776
Authors:Yang Shi, Huanqian Wang, Xie, Huanyao Zhang, Lijie Zhao, yifan zhang, Xinfeng Li, Chaoyou Fu, Zhuoer Wen, Wenting Liu, Zhuoran Zhang, Xinlong Chen, Bohan Zeng, Sihan Yang, Yushuo Guan, Zhang Zhang, Liang Wang, Haoxuan Li, Zhouchen Lin, Yuanxing Zhang, Pengfei Wan, Haotian Wang, Wenjing Yang
Abstract:
Multimodal Large Language Models (MLLMs) have achieved considerable accuracy in Optical Character Recognition (OCR) from static images. However, their efficacy in video OCR is significantly diminished due to factors such as motion blur, temporal variations, and visual effects inherent in video content. To provide clearer guidance for training practical MLLMs, we introduce MME-VideoOCR, which encompasses a comprehensive range of video OCR application scenarios. MME-VideoOCR features 10 task categories comprising 25 individual tasks and spans 44 diverse scenarios, including not only text recognition tasks but also those requiring deeper comprehension and reasoning regarding the textual content within videos. The benchmark consists of 1,464 videos with varying resolutions, aspect ratios, and durations, along with 2,000 meticulously curated, manually annotated question-answer pairs. We evaluate 15 state-of-the-art MLLMs on MME-VideoOCR, revealing that even the best-performing model (Gemini-2.5 Pro) scores below 1,500 out of 2,000. Fine-grained analysis indicates that while existing MLLMs demonstrate strong performance on tasks where relevant texts are contained within a single or few frames, they exhibit limited capability in generating effective responses for tasks demanding holistic video comprehension.



Paperid:3777
Authors:Hwiwon Lee, Ziqi Zhang, Hanxiao Lu, LINGMING ZHANG
Abstract:
Rigorous security-focused evaluation of large language model (LLM) agents is imperative for establishing trust in their safe deployment throughout the software development lifecycle. However, existing benchmarks largely rely on synthetic challenges or simplified vulnerability datasets that fail to capture the complexity and ambiguity encountered by security engineers in practice. We introduce SEC-bench, the first fully automated benchmarking framework for evaluating LLM agents on authentic security engineering tasks. SEC-bench employs a novel multi-agent scaffold that automatically constructs code repositories with harnesses, reproduces vulnerabilities in isolated environments, and generates gold patches for reliable evaluation. Our framework automatically creates high-quality software vulnerability datasets with reproducible artifacts at a cost of only $0.87 per instance. Using SEC-bench, we implement two critical software security tasks to rigorously evaluate LLM agents’ capabilities: proof-of-concept (PoC) generation and vulnerability patching. A comprehensive evaluation of state-of-the-art LLM code agents reveals significant performance gaps, achieving at most 12.5% success in PoC generation and 31.5% in vulnerability patching on our complete dataset. These results highlight the crucial steps needed toward developing LLM agents that are more practical, intelligent, and autonomous for security engineering.



Paperid:3778
Authors:Lin Guo, Xiaoqing Luo, Wei Xie, Zhancheng Zhang, Hui Li, Rui Wang, Zhenhua Feng, Xiaoning Song
Abstract:
Existing infrared and visible image fusion methods often face the dilemma of balancing modal information. Generative fusion methods reconstruct fused images by learning from data distributions, but their generative capabilities remain limited. Moreover, the lack of interpretability in modal information selection further affects the reliability and consistency of fusion results in complex scenarios. This manuscript revisits the essence of generative image fusion from the perspective of human cognitive laws and proposes a novel infrared and visible image fusion method, termed HCLFuse. First, HCLFuse investigates the quantification theory of information mapping in unsupervised fusion networks, which leads to the design of a multi-scale mask-regulated variational bottleneck encoder. This encoder applies posterior probability modeling and information decomposition to extract accurate and concise low-level modal information, thereby supporting the generation of high-fidelity structural details. Furthermore, the probabilistic generative capability of the diffusion model is integrated with physical laws, forming a time-varying physical guidance mechanism that adaptively regulates the generation process at different stages, thereby enhancing the ability of the model to perceive the intrinsic structure of data and reducing dependence on data quality. Experimental results show that the proposed method achieves state-of-the-art fusion performance in qualitative and quantitative evaluations across multiple datasets and significantly improves semantic segmentation metrics. This fully demonstrates the advantages of this generative image fusion method, rooted in human cognitive laws, in enhancing structural consistency and detail quality.



Paperid:3779
Authors:Juan Formanek, Omayma Mahjoub, Louay Nessir, Sasha Abramowitz, Ruan John de Kock, Wiem Khlifi, Daniel Rajaonarivonivelomanantsoa, Simon Du Toit, Arnol Fokam, Siddarth Singh, Ulrich Armel Mbou Sob, Felix Chalumeau, Arnu Pretorius
Abstract:
A key challenge in offline multi-agent reinforcement learning (MARL) is achieving effective many-agent multi-step coordination in complex environments. In this work, we propose Oryx, a novel algorithm for offline cooperative MARL to directly address this challenge. Oryx adapts the recently proposed retention-based architecture Sable and combines it with a sequential form of implicit constraint Q-learning (ICQ), to develop a novel offline auto-regressive policy update scheme. This allows Oryx to solve complex coordination challenges while maintaining temporal coherence over lengthy trajectories. We evaluate Oryx across a diverse set of benchmarks from prior works—SMAC, RWARE, and Multi-Agent MuJoCo—covering tasks of both discrete and continuous control, varying in scale and difficulty. Oryx achieves state-of-the-art performance on more than 80% of the 65 tested datasets, outperforming prior offline MARL methods and demonstrating robust generalisation across domains with many agents and long horizons. Finally, we introduce new datasets to push the limits of many-agent coordination in offline MARL, and demonstrate Oryx's surperior ability to scale effectively in such settings. We make all of our datasets, experimental data, and code available.



Paperid:3780
Authors:Xianzhe Fan, Xuhui Zhou, Chuanyang Jin, Kolby T Nottingham, Hao Zhu, Maarten Sap
Abstract:
Humans continuously infer the states, goals, and behaviors of others by perceiving their surroundings in dynamic, real-world social interactions. However, most Theory of Mind (ToM) benchmarks only evaluate static, text-based scenarios, which have a significant gap compared to real interactions. We propose the SoMi-ToM benchmark, designed to evaluate multi-perspective ToM in embodied multi-agent complex social interactions. This benchmark is based on rich multimodal interaction data generated by the interaction environment SoMi, covering diverse crafting goals and social relationships. Our framework supports multi-level evaluation: (1) first-person evaluation provides multimodal (visual, dialogue, action, etc.) input from a first-person perspective during a task for real-time state inference, (2) third-person evaluation provides complete third-person perspective video and text records after a task for goal and behavior inference. This evaluation method allows for a more comprehensive examination of a model's ToM capabilities from both the subjective immediate experience and the objective global observation. We constructed a challenging dataset containing 35 third-person perspective videos, 363 first-person perspective images, and 1225 expert-annotated multiple-choice questions (three options). On this dataset, we systematically evaluated the performance of human subjects and several state-of-the-art large vision-language models (LVLMs). The results show that LVLMs perform significantly worse than humans on SoMi-ToM: the average accuracy gap between humans and models is 40.1% in first-person evaluation and 26.4% in third-person evaluation. This indicates that future LVLMs need to further improve their ToM capabilities in embodied, complex social interactions.



Paperid:3781
Authors:Anlin Zheng, Xin Wen, Xuanyang Zhang, Chuofan Ma, Tiancai Wang, Gang Yu, Xiangyu Zhang, Xiaojuan Qi
Abstract:
Leveraging the powerful representations of pre-trained vision foundation models-- traditionally used for visual comprehension---we explore a novel direction: building an image tokenizer directly atop such models, a largely underexplored area. Specifically, we employ a frozen vision foundation model as the encoder of our tokenizer. To enhance its effectiveness, we introduce two key components: (1) a region-adaptive quantization framework that reduces redundancy in the pre-trained features on regular 2D grids, and (2) a semantic reconstruction objective that aligns the tokenizer’s outputs with the foundation model’s representations to preserve semantic fidelity. Based on these designs, our proposed image tokenizer, \textbf{VFMTok}, achieves substantial improvements in image reconstruction and generation quality, while also enhancing token efficiency. It further boosts autoregressive (AR) generation---achieving a gFID of \textbf{2.07} on ImageNet benchmarks, while accelerating model convergence by \textbf{three times}, and enabling high-fidelity class-conditional synthesis without the need for classifier-free guidance (CFG). The code will be released publicly to benefit the community.



Paperid:3782
Authors:Yue Yang, Jinhao Li, Hao Wang
Abstract:
Machine unlearning has become a focal point in recent research, yet the specific area of feature unlearning has not been thoroughly explored. Feature unlearning involves the elimination of specific features' effects from an already trained model, presenting distinct challenges that are still not comprehensively addressed. This paper presents a novel and straightforward approach to feature unlearning that employs a tactical shuffling of the features designated for removal. By redistributing the values of the features targeted for unlearning throughout the original training dataset and subsequently fine-tuning the model with this shuffled data, our proposed method provides a theoretical guarantee for effective feature unlearning. Under mild assumptions, our method can effectively disrupt the established correlations between unlearned features and the target outcomes, while preserving the relationships between the remaining features and the predicted outcomes. Our empirical studies across various datasets,validate that our approach not only successfully removes the effects of specified features but also maintains the informational integrity of the remaining features while achieving a faster convergence rate.



Paperid:3783
Authors:Zhuo Cao, Xuan Zhao, Lena Krieger, Hanno Scharr, Ira Assent
Abstract:
The growing integration of machine learning (ML) and artificial intelligence (AI) models into high-stakes domains such as healthcare and scientific research calls for models that are not only accurate but also interpretable. Among the existing explainable methods, counterfactual explanations offer interpretability by identifying minimal changes to inputs that would alter a model’s prediction, thus providing deeper insights. However, current counterfactual generation methods suffer from critical limitations, including gradient vanishing, discontinuous latent spaces, and an overreliance on the alignment between learned and true decision boundaries. To overcome these limitations, we propose LeapFactual, a novel counterfactual explanation algorithm based on conditional flow matching. LeapFactual generates robust and informative counterfactuals, even when true and learned decision boundaries diverge. LeapFactual is not limited to models with differentiable loss functions. It can even handle human-in-the-loop systems, expanding the scope of counterfactual explanations to domains that require the participation of human annotators, such as citizen science. We provide extensive experiments on benchmark and real-world datasets highlighting that LeapFactual generates accurate and in-distribution counterfactual explanations that offer actionable insights. We observe, for instance, that our robust counterfactual samples with labels aligning to ground truth can be beneficially used as new training data to enhance the model. The proposed method is diversely applicable and enhances scientific knowledge discovery as well as non-expert interpretability.



Paperid:3784
Authors:Kavana Venkatesh, Connor Dunlop, Pinar Yanardag Delul
Abstract:
Creativity in AI imagery remains a fundamental challenge, requiring not only the generation of visually compelling content but also the capacity to add novel, expressive, and artistically rich transformations to images. Unlike conventional editing tasks that rely on direct prompt-based modifications, creative image editing demands an autonomous, iterative approach that balances originality, coherence, and artistic intent. To address this, we introduce CREA, a novel multi-agent collaborative framework that mimics the human creative process. Our framework leverages a team of specialized AI agents who dynamically collaborate to conceptualize, generate, critique, and enhance images. Through extensive qualitative and quantitative evaluations, we demonstrate that CREA significantly outperforms state-of-the-art methods in diversity, semantic alignment, and creative transformation. To the best of our knowledge, this is the first work to introduce the task of creative editing.



Paperid:3785
Authors:Tonglong Wei, Yan Lin, Zeyu Zhou, Haomin Wen, Jilin Hu, Shengnan Guo, Youfang Lin, Gao Cong, Huaiyu Wan
Abstract:
Abstract:Vehicle GPS trajectories provide valuable movement information that supports various downstream tasks and applications. A desirable trajectory learning model should be able to transfer across regions and tasks without retraining, avoiding the need to maintain multiple specialized models and subpar performance with limited training data. However, each region has its unique spatial features and contexts, which are reflected in vehicle movement patterns and difficult to generalize. Additionally, transferring across different tasks faces technical challenges due to the varying input-output structures required for each task. Existing efforts towards transferability primarily involve learning embedding vectors for trajectories, which perform poorly in region transfer and require retraining of prediction modules for task transfer.To address these challenges, we propose $\textit{TransferTraj}$, a vehicle GPS trajectory learning model that excels in both region and task transferability. For region transferability, we introduce RTTE as the main learnable module within TransferTraj. It integrates spatial, temporal, POI, and road network modalities of trajectories to effectively manage variations in spatial context distribution across regions. It also introduces a TRIE module for incorporating relative information of spatial features and a spatial context MoE module for handling movement patterns in diverse contexts. For task transferability, we propose a task-transferable input-output scheme that unifies the input-output structure of different tasks into the masking and recovery of modalities and trajectory points. This approach allows TransferTraj to be pre-trained once and transferred to different tasks without retraining. We conduct extensive experiments on three real-world vehicle trajectory datasets under various transfer settings, including task transfer, zero-shot region transfer, and few-shot region transfer. Experimental results demonstrate that TransferTraj significantly outperforms state-of-the-art baselines in different scenarios, validating its effectiveness in region and task transfer. Code is available at https://anonymous.4open.science/r/TransferTraj-2D60.



Paperid:3786
Authors:Zhaolan Huang, Emmanuel Baccelli
Abstract:
AI spans from large language models to tiny models running on microcontrollers (MCUs). Extremely memory-efficient model architectures are decisive to fit within an MCU's tiny memory budget e.g., 128kB of RAM. However, inference latency must remain small to fit real-time constraints. An approach to tackle this ispatch-based fusion, which aims to optimize data flows across neural network layers. In this paper, we introducemsf-CNN, a novel technique that efficiently finds optimal fusion settings for convolutional neural networks (CNNs) by walking through the fusion solution space represented as a directed acyclic graph. Compared to previous work on CNN fusion for MCUs, msf-CNN identifies a wider set of solutions. We published an implementation of msf-CNN running on various microcontrollers (ARM Cortex-M, RISC-V, ESP32). We show that msf-CNN can achieve inference using 50% less RAM compared to the prior art (MCUNetV2 and StreamNet). We thus demonstrate how msf-CNN offers additional flexibility for system designers.



Paperid:3787
Authors:Xinran Wang, Songyu Xu, Shan Xiangxuan, Yuxuan Zhang, Muxi Diao, Xueyan Duan, Yanhua huang, Kongming Liang, Zhanyu Ma
Abstract:
Cinematography is a cornerstone of film production and appreciation, shaping mood, emotion, and narrative through visual elements such as camera movement, shot composition, and lighting. Despite recent progress in multimodal large language models (MLLMs) and video generation models, the capacity of current models to grasp and reproduce cinematographic techniques remains largely uncharted, hindered by the scarcity of expert-annotated data. To bridge this gap, we present CineTechBench, a pioneering benchmark founded on precise, manual annotation by seasoned cinematography experts across key cinematography dimensions. Our benchmark covers seven essential aspects—shot scale, shot angle, composition, camera movement, lighting, color, and focal length—and includes over 600 annotated movie images and 120 movie clips with clear cinematographic techniques. For the understanding task, we design question–answer pairs and annotated descriptions to assess MLLMs’ ability to interpret and explain cinematographic techniques. For the generation task, we assess advanced video generation models on their capacity to reconstruct cinema-quality camera movements given conditions such as textual prompts or keyframes. We conduct a large-scale evaluation on 15+ MLLMs and 5+ video generation models. Our results offer insights into the limitations of current models and future directions for cinematography understanding and generation in automatically film production and appreciation. The code and benchmark can be accessed at \url{https://github.com/PRIS-CV/CineTechBench}.



Paperid:3788
Authors:Mohammad Reza Taesiri, Abhijay Ghildyal, Saman Zadtootaghaj, Nabajeet Barman, Cor-Paul Bezemer
Abstract:
With video games leading in entertainment revenues, optimizing game development workflows is critical to the industry’s long-term success. Recent advances in vision-language models (VLMs) hold significant potential to automate and enhance various aspects of game development—particularly video game quality assurance (QA), which remains one of the most labor-intensive processes with limited automation. To effectively measure VLM performance in video game QA tasks and evaluate their ability to handle real-world scenarios, there is a clear need for standardized benchmarks, as current ones fall short in addressing this domain. To bridge this gap, we introduce VideoGameQA-Bench - a comprehensive benchmark designed to encompass a wide range of game QA activities, including visual unit testing, visual regression testing, needle-in-a-haystack, glitch detection, and bug report generation for both images and videos.



Paperid:3789
Authors:Guy Davidson, Todd Gureckis, Brenden Lake, Adina Williams
Abstract:
Demonstrations and instructions are two primary approaches for prompting language models to perform in-context learning (ICL) tasks. Do identical tasks elicited in different ways result in similar representations of the task? An improved understanding of task representation mechanisms would offer interpretability insights and may aid in steering models. We study this throughfunction vectors, recently proposed as a mechanism to extract few-shot ICL task representations. We generalize function vectors to alternative task presentations, focusing on short textual instruction prompts, and successfully extract instruction function vectors that promote zero-shot task accuracy. We find evidence that demonstration- and instruction-based function vectors leverage different model components, and offer several controls to dissociate their contributions to task performance. Our results suggest that different task presentations do not induce a common task representation but elicit different, partly overlapping mechanisms. Our findings offer principled support to the practice of combining textual instructions and task demonstrations, imply challenges in universally monitoring task inference across presentation forms, and encourage further examinations of LLM task inference mechanisms.



Paperid:3790
Authors:Hongyu Cheng, Amitabh Basu
Abstract:
Mixed-integer programming (MIP) provides a powerful framework for optimization problems, with Branch-and-Cut (B&C) being the predominant algorithm in state-of-the-art solvers. The efficiency of B&C critically depends on heuristic policies for making sequential decisions, including node selection, cut selection, and branching variable selection. While traditional solvers often employ heuristics with manually tuned parameters, recent approaches increasingly leverage machine learning, especially neural networks, to learn these policies directly from data. A key challenge is to understand the theoretical underpinnings of these learned policies, particularly their generalization performance from finite data. This paper establishes rigorous sample complexity bounds for learning B&C policies where the scoring functions guiding each decision step (node, cut, branch) have a certain piecewise polynomial structure. This structure generalizes the linear models that form the most commonly deployed policies in practice and investigated recently in a foundational series of theoretical works by Balcan et al. Such piecewise polynomial policies also cover the neural network architectures (e.g., using ReLU activations) that have been the focal point of contemporary practical studies. Consequently, our theoretical framework closely reflects the models utilized by practitioners investigating machine learning within B&C, offering a unifying perspective relevant to both established theory and modern empirical research in this area. Furthermore, our theory applies to quite general sequential decision making problems beyond B&C.



Paperid:3791
Authors:Eilon Laufer, Boaz Nadler
Abstract:
Abstract:Recovering a low rank matrix from a subset of its entries, some of which may be corrupted, is known as the robust matrix completion (RMC) problem.Existing RMC methods have several limitations: they require a relatively large number of observed entries; they may fail under overparametrization, when their assumed rank is higher than the correct one;and many of them fail to recover even mildly ill-conditioned matrices; In this paper we propose a novel RMC method, denoted $\texttt{RGNMR}$, which overcomes these limitations. $\texttt{RGNMR}$ is a simple factorization-based iterative algorithm, which combines a Gauss–Newton linearization with removal of entries suspected to be outliers.On the theoretical front, we prove that under suitable assumptions, $\texttt{RGNMR}$ is guaranteed exact recovery of the underlying low rank matrix.Our theoretical results improve upon the best currently known for factorization-based methods. On the empirical front, we show via several simulationsthe advantages of $\texttt{RGNMR}$ over existing RMC methods, and in particular its ability to handle a small number of observed entries, overparameterization of the rank and ill-conditioned matrices.



Paperid:3792
Authors:Shenzhi Wang, Le Yu, Chang Gao, Chujie Zheng, Shixuan Liu, Kai Dang, Xiong-Hui Chen, Jianxin Yang, Zhenru Zhang, Yuqiong Liu, An Yang, Rui Lu, Andrew Zhao, Yang Yue, Shiji Song, Bowen Yu, Gao Huang, Junyang Lin
Abstract:
Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a powerful approach to enhancing the reasoning capabilities of Large Language Models (LLMs), yet its underlying mechanisms remain insufficiently understood. In this work, we undertake a pioneering exploration of RLVR through the novel perspective of token entropy patterns, comprehensively analyzing how different tokens influence reasoning performance. By examining token entropy patterns in Chain-of-Thought (CoT) reasoning, we observe that only a small fraction (approximately 20\%) of tokens exhibit high entropy, and these tokens semantically act as critical forks that steer the model toward diverse reasoning pathways. We further demonstrate that moderately increasing the entropy of these high-entropy tokens via decoding temperature adjustments leads to improved performance, quantitatively confirming their role as decision points in reasoning. We ultimately refine RLVR by restricting policy gradient updates to these forking tokens. Despite utilizing only 20\% of tokens, our approach achieves comparable performance to full-gradient updates on the Qwen3-8B base model. Moreover, it demonstrates remarkable improvements on the larger Qwen3-32B base model, boosting AIME'25 scores by 11.04 and AIME'24 scores by 7.71. In contrast, training exclusively on the 80\% lowest-entropy tokens leads to a marked decline in performance. These findings indicate that the efficacy of RLVR primarily arises from optimizing the high-entropy tokens that dictate key reasoning directions. Collectively, our results suggest promising avenues for optimizing RLVR algorithms by strategically leveraging the potential of these high-entropy minority tokens to further enhance the reasoning abilities of LLMs.



Paperid:3793
Authors:Anuja Negi, SUBBAREDDY OOTA, Anwar Nunez-Elizalde, Manish Gupta, Fatma Deniz
Abstract:
Recent studies have demonstrated that fine-tuning language models with brain data can improve their semantic understanding, although these findings have so far been limited to English. Interestingly, similar to the shared multilingual embedding space of pretrained multilingual language models, human studies provide strong evidence for a shared semantic system in bilingual individuals. Here, we investigate whether fine-tuning language models with bilingual brain data changes model representations in a way that improves them across multiple languages. To test this, we fine-tune monolingual and multilingual language models using brain activity recorded while bilingual participants read stories in English and Chinese. We then evaluate how well these representations generalize to the bilingual participants’ first language, their second language, and several other languages that the participant is not fluent in. We assess the fine-tuned language models on brain encoding performance and downstream NLP tasks. Our results show that bilingual brain-informed fine-tuned language models outperform their vanilla (pretrained) counterparts in both brain encoding performance and most downstream NLP tasks across multiple languages. These findings suggest that brain-informed fine-tuning improves multilingual understanding in language models, offering a bridge between cognitive neuroscience and NLP research.



Paperid:3794
Authors:Yanna Ding, Songtao Lu, Yingdong Lu, Tomasz Nowicki, Jianxi Gao
Abstract:
Transformer architectures can solve unseen tasks based on input-output pairs in a given prompt due to in-context learning (ICL). Existing theoretical studies on ICL have mainly focused on linear regression tasks, often with i.i.d. inputs. To understand how transformers express in-context learning when modeling dynamics-driven functions, we investigate Markovian function learning through a structured ICL setup, where we characterize the loss landscape to reveal underlying optimization behaviors. Specifically, we (1) provide the closed-form expression of the global minimizer (in an enlarged parameter space) for a single-layer linear self-attention (LSA) model; (2) prove that recovering transformer parameters that realize the optimal solution is NP-hard in general, revealing a fundamental limitation of one-layer LSA in representing structured dynamical functions; and (3) supply a novel interpretation of a multilayer LSA as performing preconditioned gradient descent to optimize multiple objectives beyond the square loss. These theoretical results are numerically validated using simplified transformers.



Paperid:3795
Authors:Zhenyi Zhang, Zihan Wang, Yuhao Sun, Tiejun Li, Peijie Zhou
Abstract:
Modeling the dynamics from sparsely time-resolved snapshot data is crucial for understanding complex cellular processes and behavior. Existing methods leverage optimal transport, Schrödinger bridge theory, or their variants to simultaneously infer stochastic, unbalanced dynamics from snapshot data. However, these approaches remain limited in their ability to account for cell-cell interactions. This integration is essential in real-world scenarios since intercellular communications are fundamental life processes and can influence cell state-transition dynamics. To address this challenge, we formulate the Unbalanced Mean-Field Schrödinger Bridge (UMFSB) framework to model unbalanced stochastic interaction dynamics from snapshot data. Inspired by this framework, we further proposeCytoBridge, a deep learning algorithm designed to approximate the UMFSB problem. By explicitly modeling cellular growth, death, and interactions through neural networks, CytoBridge offers the flexibility to learn these processes directly from data. The effectiveness of our method has been extensively validated using both synthetic gene regulatory data and real scRNA-seq datasets. Compared to existing methods, CytoBridge identifies growth, transition, and interaction patterns, eliminates false transitions, and reconstructs the developmental landscape with greater accuracy.



Paperid:3796
Authors:Rob Romijnders, Antti Koskela
Abstract:
The hidden state threat model of differential privacy (DP) assumes that the adversary has access only to the final trained machine learning (ML) model, without seeing intermediate states during training. However, the current privacy analyses under this model are restricted to convex optimization problems, reducing their applicability to multi-layer neural networks, which are essential in modern deep learning applications. Notably, the most successful applications of the hidden state privacy analyses in classification tasks have only been for logistic regression models. We demonstrate that it is possible to privately train convex problems with privacy-utility trade-offs comparable to those of 2-layer ReLU networks trained with DP stochastic gradient descent (DP-SGD). This is achieved through a stochastic approximation of a dual formulation of the ReLU minimization problem, resulting in a strongly convex problem. This enables the use of existing hidden state privacy analyses and provides accurate privacy bounds also for the noisy cyclic mini-batch gradient descent (NoisyCGD) method with fixed disjoint mini-batches. Empirical results on benchmark classification tasks demonstrate that NoisyCGD can achieve privacy-utility trade-offs on par with DP-SGD applied to 2-layer ReLU networks.



Paperid:3797
Authors:Jing Zuo, Jiaqi Wang, Yonggang Qi, Yi-Zhe Song
Abstract:
Abstract:Recent work shows that features from Stable Diffusion (SD) and contrastively pretrained models like DINO can be directly used for zero-shot semantic correspondence via naive feature concatenation. In this paper, we explore the stronger potential of Stable Diffusion 3 (SD3), a rectified flow-based model with a multimodal transformer backbone (MM-DiT). We show that semantic signals in SD3 are scattered across multiple timesteps and transformer layers, and propose a multi-level fusion scheme to extract discriminative features. Moreover, we identify that naive fusion across models suffers from inconsistent distributions, thus leading to suboptimal performance. To address this, we propose a simple yet effective confidence-aware feature fusion strategy that re-weights each model’s contribution based on prediction confidence scores derived from their matching uncertainties. Notably, this fusion approach is not only training-free but also enables per-pixel adaptive integration of heterogeneous features. The resulting representation, E-SD$^3$, significantly outperforms strong baselines on SPair-71k, PF-Pascal, and PSC6K, validating the benefit of combining SD3, SD, and DINO through our proposed confidence-aware feature fusion.



Paperid:3798
Authors:Piyush Nitin Bagad, Andrew Zisserman
Abstract:
Our objective is to develop compact video representations that are sensitive to visual change over time. To measure such time-sensitivity, we introduce a new task: chiral action recognition, where one needs to distinguish between a pair of temporally opposite actions, such as “opening vs. closing a door", “approaching vs. moving away from something", “folding vs. unfolding paper", etc. Such actions (i) occur frequently in everyday life, (ii) require understanding of simple visual change over time (in object state, size, spatial position, count . . . ), and (iii) are known to be poorly represented by many video embeddings. Our goal is to build time aware video representations which offer linear separability between these chiral pairs. To that end, we propose a self-supervised adaptation recipe to inject time-sensitivity into a sequence of frozen image features. Our model is based on an auto-encoder with a latent space with inductive bias inspired by perceptual straightening. We show that this results in a compact but time-sensitive video representation for the proposed task across three datasets: Something-Something, EPIC-Kitchens, and Charade. Our method (i) outperforms much larger video models pre-trained on large-scale video datasets, and (ii) leads to an improvement in classification performance on standard benchmarks when combined with these existing models.



Paperid:3799
Authors:Zenghui Yang, Xingquan Zuo, Hai Huang, Gang Chen, Xinchao Zhao, Tianle Zhang
Abstract:
Deep neural networks have become foundational in various applications but remain vulnerable to adversarial patch attacks. Crafting effective adversarial patches is inherently challenging due to the combinatorial complexity involved in jointly optimizing critical factors such as patch shape, location, number, and content. Existing approaches often simplify this optimization by addressing each factor independently, which limits their effectiveness. To tackle this significant challenge, we introduce a novel and flexible adversarial attack framework termed IMPACT (Irregular Multi-Patch Adversarial Composition based on Two phase optimization). IMPACT uniquely enables comprehensive optimization of all essential patch factors using gradient-free methods. Specifically, we propose a novel dimensionality reduction encoding scheme that substantially lowers computational complexity while preserving expressive power. Leveraging this encoding, we further develop a two-phase optimization framework: phase 1 employs differential evolution for joint optimization of patch mask and content, while phase 2 refines patch content using an evolutionary strategy for enhanced precision. Additionally, we introduce a new aggregation algorithm explicitly designed to produce contiguous, irregular patches by merging localized regions, ensuring physical applicability. Extensive experiments demonstrate that our method significantly outperforms several state-of-the-art approaches, highlighting the critical benefit of jointly optimizing all patch factors in adversarial patch attacks.



Paperid:3800
Authors:Jiehua Chen, Jiong Guo, Yinghui Wen
Abstract:
We investigate congested assignment problems where agents have preferences over both resources and their associated congestion levels. These agents are \emph{averse} towards congestion, i.e., consistently preferring lower congestion for identical resources. Such scenarios are ubiquitous across domains including traffic management and school choice, where fair resource allocation is essential. We focus on the concept of \emph{competitiveness}, recently introduced by Bogomolnaia and Moulin [6], and contribute a polynomial-time algorithm that determines competitiveness, resolving their open question. Additionally, we explore two optimization variants of congested assignments by examining the problem of finding envy-free or maximally competitive assignments that guarantee a certain amount of social welfare for every agent, termed \emph{top-guarantees} [6]. While we prove that both problems are NP-hard, we develop parameterized algorithms with respect to the number of agents or resources.



Paperid:3801
Authors:Jinwoo Lim, Suhyun Kim, Soo-Mook Moon
Abstract:
One of the chronic problems of deep-learning models is shortcut learning. In a case where the majority of training data are dominated by a certain feature, neural networks prefer to learn such a feature even if the feature is not generalizable outside the training set. Based on the framework of Neural Tangent Kernel (NTK), we analyzed the case of linear neural networks to derive some important properties of shortcut learning. We defined a “feature” of a neural network as an eigenfunction of NTK. Then, we found that shortcut features correspond to features with larger eigenvalues when the shortcuts stem from the imbalanced number of samples in the clustered distribution. We also showed that the features with larger eigenvalues still have a large influence on the neural network output even after training, due to data variances in the clusters. Such a preference for certain features remains even when a margin of a neural network output is controlled, which shows that the max-margin bias is not the only major reason for shortcut learning. These properties of linear neural networks are empirically extended for more complex neural networks as a two-layer ReLU FC network and a ResNet-18.



Paperid:3802
Authors:Pingyi Chen, Yujing Lou, Shen Cao, Jinhui Guo, Lubin Fan, Yue Wu, Lin Yang, Lizhuang Ma, Jieping Ye
Abstract:
While vision language models (VLMs) excel in 2D semantic visual understanding, their ability to quantitatively reason about 3D spatial relationships remains under-explored, due to the deficiency of 2D images' spatial representation ability. In this paper, we analyze the problem hindering VLMs' spatial understanding abilities and propose SD-VLM, a novel framework that significantly enhances fundamental spatial perception abilities of VLMs through two key contributions: (1) propose Massive Spatial Measuring and Understanding (MSMU) dataset with precise spatial annotations, and (2) introduce a simple depth positional encoding method strengthening VLMs' spatial awareness. MSMU dataset covers massive quantitative spatial tasks with 700K QA pairs, 2.5M physical numerical annotations, and 10K chain-of-thought augmented samples. We have trained SD-VLM, a strong generalist VLM which shows superior quantitative spatial measuring and understanding capability. SD-VLM not only achieves state-of-the-art performance on our proposed MSMU-Bench, but also shows spatial generalization abilities on other spatial understanding benchmarks including Q-Spatial and SpatialRGPT-Bench. Extensive experiments demonstrate that SD-VLM outperforms GPT-4o and Intern-VL3-78B by 26.91% and 25.56% respectively on MSMU-Bench. We will release MSMU dataset and SD-VLM to facilitate future research in quantitative spatial measuring and understanding.



Paperid:3803
Authors:Yanglin Feng, Hongyuan Zhu, Dezhong Peng, Xi Peng, Xiaomin Song, Peng Hu
Abstract:
Grounding target objects in 3D environments via natural language is a fundamental capability for autonomous agents to successfully fulfill user requests. Almost all existing works typically assume that the target object lies within a known scene and focus solely on in-scene localization. In practice, however, agents often encounter unknown or previously visited environments and need to search across a large archive of scenes to ground the described object, thereby invalidating this assumption. To address this, we reveal a novel task called Cross-Scene Spatial Reasoning and Grounding (CSSRG), which aims to locate a described object anywhere across an entire collection of 3D scenes rather than predetermined scenes. Due to the difference from existing 3D visual grounding, CSSRG poses two challenges: the prohibitive cost of exhaustively traversing all scenes and more complex cross-modal spatial alignment. To address the challenges, we propose a Cross-Scene 3D Object Reasoning Framework (CoRe), which adopts a matching-then-grounding pipeline to reduce computational overhead. Specifically, CoRe consists of i) a Robust Text-Scene Aligning (RTSA) module that learns global scene representations for robust alignment between object descriptions and the corresponding 3D scenes, enabling efficient retrieval of candidate scenes; and ii) a Tailored Word-Object Associating (TWOA) module that establishes fine-grained alignment between words and target objects to filter out redundant context, supporting precise object-level reasoning and alignment. Additionally, to benchmark CSSRG, we construct a new CrossScene-RETR dataset and evaluation protocol tailored for cross-scene grounding. Extensive experiments across four multimodal datasets demonstrate that CoRe dramatically reduces computational overhead while showing superiority in both scene retrieval and object grounding.



Paperid:3804
Authors:Fei Wang, Li Shen, Liang Ding, Chao Xue, Ye Liu, Changxing Ding
Abstract:
Large Language Models (LLMs) excel at natural language processing tasks, but their massive size leads to high computational and storage demands.Recent works have sought to reduce their model size through layer-wise structured pruning.However, they tend to ignore retaining the capabilities in the pruned part. In this work, we re-examine structured pruning paradigms and uncover several key limitations: 1) notable performance degradation due to direct layer removal, 2) incompetent linear weighted layer aggregation, and 3) the lack of effective post-training recovery mechanisms.To address these limitations, we propose CoMe, including a progressive layer pruning framework with a Concatenation-based Merging technology and a hierarchical distillation post-training process. Specifically, we introduce a channel sensitivity metric that utilizes activation intensity and weight norms for fine-grained channel selection. Subsequently, we employ a concatenation-based layer merging method to fuse the most critical channels in the adjacent layers, enabling a progressive model size reduction. Finally, we propose a hierarchical distillation protocol, which leverages the correspondences between the original and pruned model layers established during pruning, enabling efficient knowledge transfer.Experiments on seven benchmarks show that CoMe achieves state-of-the-art performance; when pruning 30% of LLaMA-2-7b's parameters, the pruned model retains 83% of its original average accuracy.



Paperid:3805
Authors:xin zhang, Ziruo Zhang, JIAWEI DU, Zuozhu Liu, Joey Tianyi Zhou
Abstract:
Abstract:Multimodal Dataset Distillation (MDD) seeks to condense large-scale image-text datasets into compact surrogates while retaining their effectiveness for cross-modal learning. Despite recent progress, existing MDD approaches often suffer from ***Modality Collapse***, characterized by over-concentrated intra-modal representations and enlarged distributional gap across modalities. In this paper, at the first time, we identify this issue as stemming from a fundamental conflict between the over-compression behavior inherent in dataset distillation and the cross-modal supervision imposed by contrastive objectives. To alleviate modality collapse, we introduce **RepBlend**, a novel MDD framework that weakens overdominant cross-modal supervision via representation blending, thereby significantly enhancing intra-modal diversity. Additionally, we observe that current MDD methods impose asymmetric supervision across modalities, resulting in biased optimization. To address this, we propose symmetric projection trajectory matching, which synchronizes the optimization dynamics using modality-specific projection heads, thereby promoting balanced supervision and enhancing cross-modal alignment.Experiments on Flickr-30K and MS-COCO show that RepBlend consistently outperforms prior state-of-the-art MDD methods, achieving significant gains in retrieval performance (e.g., +9.4 IR@10, +6.3 TR@10 under the 100-pair setting) and offering up to 6.7$\times$ distillation speedup.



Paperid:3806
Authors:Timothy Truong Jr, Tristan Bepler
Abstract:
Protein language models (PLMs) learn probability distributions over natural protein sequences. By learning from hundreds of millions of natural protein sequences, protein understanding and design capabilities emerge. Recent works have shown that scaling these models improves structure prediction, but does not seem to improve mutation understanding and representation quality for protein function prediction. We introduce PoET-2, a multimodal, retrieval-augmented protein foundation model that incorporates in-context learning of family-specific evolutionary constraints with optional structure conditioning to learn generative distributions over protein sequences. PoET-2 uses a hierarchical transformer encoder that is equivariant to sequence context ordering and a dual decoder architecture with both causal and masked language modeling objectives, allowing PoET-2 to operate in both fully generative and bidirectional representation learning modes. PoET-2 achieves state-of-the-art performance on zero-shot variant effect prediction, excelling at scoring variants with multiple mutations and challenging indel mutations. In supervised settings, PoET-2 embeddings outperform previous methods for learning sequence-function relationships, especially with small datasets. This work highlights the benefits of combining retrieval augmentation with multimodal, family-centric modeling for advancing protein foundation models.



Paperid:3807
Authors:Shayan Karimi, Xiaoqi Tan
Abstract:
Abstract:This paper investigates the computational complexity of reinforcement learning within a novel linear function approximation regime, termed partial $q^{\pi}$-realizability. In this framework, the objective is to learn an $\epsilon$-optimal policy with respect to a predefined policy set $\Pi$, under the assumption that all value functions corresponding to policies in $\Pi$ are linearly realizable. This framework adopts assumptions that are weaker than those in the $q^{\pi}$-realizability setting yet stronger than those in the $q^*$-realizability setup. As a result, it provides a more practical model for reinforcement learning scenarios where function approximation naturally arise. We prove that learning an $\epsilon$-optimal policy in this newly defined setting is computationally hard. More specifically, we establish NP-hardness under a parameterized greedy policy set (i.e., argmax) and, further, show that—unless NP = RP—an exponential lower bound (exponential in feature vector dimension) holds when the policy set contains softmax policies, under the Randomized Exponential Time Hypothesis. Our hardness results mirror those obtained in the $q^*$-realizability settings, and suggest that computational difficulty persists even when the policy class $ \Pi $ is expanded beyond the optimal policy, reinforcing the unbreakable nature of the computational hardness result regarding partial $ q^{\pi} $-realizability under two important policy sets. To establish our negative result, our primary technical contribution is a reduction from two complexity problems, $\delta$-Max-3SAT and $\delta$-Max-3SAT($b$) (a variant of MAX-3SAT), to instances of our problem settings: GLinear-$\kappa$-RL (under the greedy policy set) and SLinear-$\kappa$-RL (under the softmax policy set), respectively. Our findings indicate that positive computational results are generally unattainable in the context of partial $ q^{\pi} $-realizability, in sharp contrast to the $ q^{\pi} $-realizability setting under a generative access model.



Paperid:3808
Authors:Jianhui Chen, Xiaozhi Wang, Zijun Yao, Yushi Bai, Lei Hou, Juanzi Li
Abstract:
Large language models (LLMs) excel in various capabilities but pose safety risks such as generating harmful content and misinformation, even after safety alignment. In this paper, we explore the inner mechanisms of safety alignment through the lens of mechanistic interpretability, focusing on identifying and analyzing safety neurons within LLMs that are responsible for safety behaviors. We propose inference-time activation contrasting to locate these neurons and dynamic activation patching to evaluate their causal effects on model safety. Experiments on multiple prevalent LLMs demonstrate that we can consistently identify about 5% safety neurons, and by only patching their activations we can restore over 90% of the safety performance across various red-teaming benchmarks without influencing general ability. The finding of safety neurons also helps explain the ''alignment tax'' phenomenon by revealing that the key neurons for model safety and helpfulness significantly overlap, yet they require different activation patterns for the same neurons. Furthermore, we demonstrate an application of our findings in safeguarding LLMs by detecting unsafe outputs before generation.



Paperid:3809
Authors:Adriel Marco, John D. Kirwan, Alexia Toumpa, Simos Gerasimou
Abstract:
Quantifying uncertainty in deep regression models is important both for understanding the confidence of the model and for safe decision-making in high-risk domains. Existing approaches that yield prediction intervals overlook distributional information, neglecting the effect of multimodal or asymmetric distributions on decision-making. Similarly, full or approximated Bayesian methods, albeit yielding the predictive posterior density, demand major modifications to the model architecture and retraining. We introduce MCNF, a novel post hoc uncertainty quantification method that produces both prediction intervals and the full conditioned predictive distribution. MCNF operates on top of the underlying trained predictive model; thus, no predictive model retraining is needed. We provide experimental evidence that the MCNF-based uncertainty estimate is well calibrated, is competitive with state-of-the-art uncertainty quantification methods, and provides richer information for downstream decision-making tasks.



Paperid:3810
Authors:Dehao Zhang, Malu Zhang, Shuai Wang, Jingya Wang, Wenjie Wei, Zeyu Ma, Guoqing Wang, Yang Yang, Haizhou Li
Abstract:
The explosive growth in sequence length has intensified the demand for effective and efficient long sequence modeling. Benefiting from intrinsic oscillatory membrane dynamics, Resonate-and-Fire (RF) neurons can efficiently extract frequency components from input signals and encode them into spatiotemporal spike trains, making them well-suited for long sequence modeling. However, RF neurons exhibit limited effective memory capacity and a trade-off between energy efficiency and training speed on complex temporal tasks. Inspired by the dendritic structure of biological neurons, we propose a Dendritic Resonate-and-Fire (D-RF) model, which explicitly incorporates a multi-dendritic and soma architecture. Each dendritic branch encodes specific frequency bands by utilizing the intrinsic oscillatory dynamics of RF neurons, thereby collectively achieving comprehensive frequency representation. Furthermore, we introduce an adaptive threshold mechanism into the soma structure that adjusts the threshold based on historical spiking activity, reducing redundant spikes while maintaining training efficiency in long sequence tasks. Extensive experiments demonstrate that our method maintains competitive accuracy while substantially ensuring sparse spikes without compromising computational efficiency during training. These results underscore its potential as an effective and efficient solution for long sequence modeling on edge platforms.



Paperid:3811
Authors:Zhihao Zhan, Jianan Zhao, Zhaocheng Zhu, Jian Tang
Abstract:
Long-context modeling is a fundamental challenge in natural language processing (NLP) with numerous practical applications. State-space models (SSMs) have gained attention as a promising alternative to Transformers, offering sub-quadratic efficiency for long-context sequence modeling. Despite this advantage, both empirical evidence and theoretical studies have shown that SSMs are substantially less effective than Transformers in capturing long-range dependencies. While recent approaches such as Transformer-SSM hybrid architectures have yielded better performance, they suffer from poor scalability with quadratic time complexity. In this work, we take an initial step towards identifying design principles for sub-quadratic hybrid architectures. We motivate the study of hybrid sparse attention via a novel synthetic task, joint recall. Through theoretical analysis and empirical evaluation, we demonstrate that hybrid sparse attention significantly improves the long-context modeling capacity of SSMs. Moreover, we highlight the importance of incorporating input-dependent sparse attention patterns. Our findings are further validated on real-world long-context NLP tasks by continual pre-training Mamba augmented with hybrid sparse attention.



Paperid:3812
Authors:Jon Saad-Falcon, Estefany Kelly Buchanan, Mayee Chen, Tzu-Heng Huang, Brendan McLaughlin, Tanvir Bhathal, Zhu, Ben Athiwaratkun, Frederic Sala, Scott Linderman, Azalia Mirhoseini, Christopher Ré
Abstract:
Verifiers can improve language model (LM) capabilities by providing feedback or selecting the best response from a pool of generated candidates. Currently, high-quality verifiers are either unscalable (e.g., humans) or limited in utility (e.g., tools like Lean for formal proofs). While LM judges and reward models have become broadly useful as general-purpose verifiers, a significant performance gap remains between them and oracle verifiers. To help close this gap, we introduce Weaver, a framework for designing a strong verifier by combining multiple weak, imperfect verifiers. First we find that weighted ensembles of verifiers, which typically require learning from labeled data, significantly outperform unweighted combinations due to differences in the verifiers. To reduce the dependency on labeled data, Weaver leverages weak supervision to estimate each verifier’s accuracy and combines their outputs into a unified score that better reflects true response quality. However, directly applying weak supervision algorithms poses several challenges, including inconsistent verifier output formats and handling low-quality verifiers. Weaver addresses these challenges by using dataset statistics to normalize outputs and filter specific verifiers. We study the effectiveness of Weaver in repeated sampling settings, where a model generates multiple candidate responses at test time and a verifier is used to select the correct one. Our evaluations demonstrate that Weaver significantly improves the pass@1 performance across several reasoning and math tasks, achieving o3-mini level accuracy with Llama 3.3 70B Instruct (a much cheaper non-reasoning model) as the generator, and an ensemble of smaller judge and reward models as the verifiers (86.2% average). This gain mirrors the jump achieved between GPT-4o and o3-mini (69.0% vs. 86.7%), which required extensive finetuning and post-training interventions. To make Weaver more efficient, we train a compact 400M cross-encoder using Weaver's combined output scores. This distilled model retains 98.7% of Weaver's full accuracy while reducing verification compute by up to 99.97%.



Paperid:3813
Authors:Panagiotis Giannoulis, Yorgos Pantis, Christos Tzamos
Abstract:
Abstract:Despite their proficiency in various language tasks, Large Language Models (LLMs) struggle with combinatorial problems like Satisfiability, Traveling Salesman Problem, or even basic arithmetic. We address this gap through a novel approach for solving problems in the class NP. We focus on the paradigmatic task of Sudoku and achieve state-of-the-art accuracy (99\%) compared to prior neuro-symbolic approaches. Unlike prior work that used custom architectures, our method employs a vanilla decoder-only Transformer (GPT-2) without external tools or function calling.Our method integrates imitation learning of simpleSudoku rules with an explicit Depth-First Search (DFS) exploration strategyinvolving informed guessing and backtracking.Moving beyond imitation learning, we seek to minimize the number of guesses until reaching a solution. We provide a rigorous analysis of this setup by formalizing its connection to a contextual variant of $\textit{Min-Sum Set Cover}$, a well-studied problem in algorithms and stochastic optimization.



Paperid:3814
Authors:Yi Liu, Dianqing Liu, Mingye Zhu, Junbo Guo, Yongdong Zhang, Zhendong Mao
Abstract:
The widespread adoption of large language models (LLMs) across industries has increased the demand for high-quality and customizable outputs. However, traditional alignment methods often require retraining large pretrained models, making it difficult to quickly adapt and optimize LLMs for diverse applications. To address this limitation, we propose a novel \textit{Residual Alignment Model} (\textit{RAM}) that formalizes the alignment process as a type of importance sampling. In this framework, the unaligned upstream model serves as the proposal distribution, while the alignment process is framed as secondary sampling based on an autoregressive alignment module that acts as an estimator of the importance weights. This design enables a natural detachment of the alignment module from the target aligned model, improving flexibility and scalability. Based on this model, we derive an efficient sequence-level training strategy for the alignment module, which operates independently of the proposal module. Additionally, we develop a resampling algorithm with iterative token-level decoding to address the common first-token latency issue in comparable methods. Experimental evaluations on two leading open-source LLMs across diverse tasks, including instruction following, domain adaptation, and preference optimization, demonstrate that our approach consistently outperforms baseline models.



Paperid:3815
Authors:Jiyu Guo, Shuo Yang, Yiming Huang, Yancheng Long, Xiaobo Xia, Xiu Su, Bo Zhao, Zeke Xie, Liqiang Nie
Abstract:
Data augmentation using generative models has emerged as a powerful paradigm for enhancing performance in computer vision tasks. However, most existing augmentation approaches primarily focus on optimizing intrinsic data attributes -- such as fidelity and diversity -- to generate visually high-quality synthetic data, while often neglecting task-specific requirements. Yet, it is essential for data generators to account for the needs of downstream tasks, as training data requirements can vary significantly across different tasks and network architectures.To address these limitations, we propose UtilGen, a novel utility-centric data augmentation framework that adaptively optimizes the data generation process to produce task-specific, high-utility training data via downstream task feedback. Specifically, we first introduce a weight allocation network to evaluate the task-specific utility of each synthetic sample. Guided by these evaluations, UtilGen iteratively refines the data generation process using a dual-level optimization strategy to maximize the synthetic data utility: (1) model-level optimization tailors the generative model to the downstream task, and (2) instance-level optimization adjusts generation policies -- such as prompt embeddings and initial noise -- at each generation round.Extensive experiments on eight benchmark datasets of varying complexity and granularity demonstrate that UtilGen consistently achieves superior performance, with an average accuracy improvement of 3.49\% over previous SOTA. Further analysis of data influence and distribution reveals that UtilGen produces more impactful and task-relevant synthetic data, validating the effectiveness of the paradigm shift from visual characteristics-centric to task utility-centric data augmentation.



Paperid:3816
Authors:Ke Wang, Yiming QIN, Nikolaos Dimitriadis, Alessandro Favero, Pascal Frossard
Abstract:
Language models deployed in real-world systems often require post-hoc updates to incorporate new or corrected knowledge. However, editing such models efficiently and reliably—without retraining or forgetting previous information—remains a major challenge. Existing methods for lifelong model editing either compromise generalization, interfere with past edits, or fail to scale to long editing sequences. We propose MEMOIR, a novel scalable framework that injects knowledge through a residual memory, i.e., a dedicated parameter module, while preserving the core capabilities of the pre-trained model. By sparsifying input activations through data-dependent masks, MEMOIR confines each edit to a distinct subset of the memory parameters, minimizing interference among edits. At inference, it identifies relevant edits by comparing the sparse activation patterns of new queries to those stored during editing. This enables generalization to rephrased queries by activating only the relevant knowledge while suppressing unnecessary memory activation for unrelated prompts. Experiments on question answering, hallucination correction, and out-of-distribution generalization benchmarks across LLaMA-3 and Mistral demonstrate that MEMOIR achieves state-of-the-art performance across reliability, generalization, and locality metrics, scaling to thousands of sequential edits with minimal forgetting.



Paperid:3817
Authors:Xuhui Chen, Fei Hou, Wencheng Wang, Hong Qin, Ying He
Abstract:
Abstract:Unsigned distance fields (UDFs) are widely used in 3D deep learning due to their ability to represent shapes with arbitrary topology. While prior work has largely focused on learning UDFs from point clouds or multi-view images, extracting meshes from UDFs remains challenging, as the learned fields rarely attain exact zero distances. A common workaround is to reconstruct signed distance fields (SDFs) locally from UDFs to enable surface extraction via Marching Cubes. However, this often introduces topological artifacts such as holes or spurious components. Moreover, local SDFs are inherently incapable of representing non-manifold geometry, leading to complete failure in such cases. To address this gap, we propose MIND ($\underline{M}aterial$ $\underline{I}nterface$ $from$ $\underline{N}on$-$manifold$ $\underline{D}istance$ $fields$), a novel algorithm for generating material interfaces directly from UDFs, enabling non-manifold mesh extraction from a global perspective. The core of our method lies in deriving a meaningful spatial partitioning from the UDF, where the target surface emerges as the interface between distinct regions. We begin by computing a two-signed local field to distinguish the two sides of manifold patches, and then extend this to a multi-labeled global field capable of separating all sides of a non-manifold structure. By combining this multi-labeled field with the input UDF, we construct material interfaces that support non-manifold mesh extraction via a multi-labeled Marching Cubes algorithm. Extensive experiments on UDFs generated from diverse data sources, including point cloud reconstruction, multi-view reconstruction, and medial axis transforms, demonstrate that our approach robustly handles complex non-manifold surfaces and significantly outperforms existing methods.



Paperid:3818
Authors:Daniel Kunin, Giovanni Luca Marchetti, Feng Chen, Dhruva Karkada, James Simon, Michael Deweese, Surya Ganguli, Nina Miolane
Abstract:
What features neural networks learn, and how, remains an open question. In this paper, we introduce Alternating Gradient Flows (AGF), an algorithmic framework that describes the dynamics of feature learning in two-layer networks trained from small initialization. Prior works have shown that gradient flow in this regime exhibits a staircase-like loss curve, alternating between plateaus where neurons slowly align to useful directions and sharp drops where neurons rapidly grow in norm. AGF approximates this behavior as an alternating two-step process: maximizing a utility function over dormant neurons and minimizing a cost function over active ones. AGF begins with all neurons dormant. At each round, a dormant neuron activates, triggering the acquisition of a feature, and a drop in the loss. AGF quantifies the order, timing, and magnitude of these drops, matching experiments across architectures. We show that AGF unifies and extends existing saddle-to-saddle analyses in fully connected linear networks and attention-only linear transformers, where the learned features are singular modes and principal components, respectively. In diagonal linear networks, we prove AGF converges to gradient flow in the limit of vanishing initialization. Applying AGF to quadratic networks trained to perform modular addition, we give the first complete characterization of the training dynamics, revealing that networks learn Fourier features in decreasing order of coefficient magnitude. Altogether, AGF offers a promising step towards understanding feature learning in neural networks.



Paperid:3819
Authors:Chenghao Xu, guangtao lyu, Jiexi Yan, Muli Yang, Cheng Deng
Abstract:
In dance performances, choreographers define the visual expression of movement, while cinematographers shape its final presentation through camera work. Consequently, the synthesis of camera movements informed by both music and dance has garnered increasing research interest. While recent advancements have led to notable progress in this area, existing methods predominantly operate in an offline manner—that is, they require access to the entire dance sequence before generating corresponding camera motions. This constraint renders them impractical for real-time applications, particularly in live stage performances, where immediate responsiveness is essential. To address this limitation, we introduce a more practical yet challenging task: online camera movement synthesis, in which camera trajectories must be generated using only the current and preceding segments of dance and music. In this paper, we propose TemMEGA (Temporal Masked Generative Modeling), a unified framework capable of handling both online and offline camera movement generation. TemMEGA consists of three key components. First, a discrete camera tokenizer encodes camera motions as discrete tokens via a discrete quantization scheme. Second, a consecutive memory encoder captures historical context by jointly modeling long- and short-term temporal dependencies across dance and music sequences. Finally, a temporal conditional masked transformer is employed to predict future camera motions by leveraging masked token prediction. Extensive experimental evaluations demonstrate the effectiveness of our TemMEGA, highlighting its superiority in both online and offline camera movement synthesis.



Paperid:3820
Authors:Paul Rosu, Lawrence Carin, Xiang Cheng
Abstract:
Transformers have emerged as powerful meta-learners, with growing evidence that they implement learning algorithms within their forward pass. We study this phenomenon in the context of denoising, presenting a unified framework that shows Transformers can implement (a) manifold denoising via Laplacian flows, (b) score-based denoising from diffusion models, and (c) a generalized form of anisotropic diffusion denoising. Our theory establishes exact equivalence between Transformer attention updates and these algorithms. Empirically, we validate these findings on image denoising tasks, showing that even simple Transformers can perform robust denoising both with and without context. These results illustrate the Transformer’s flexibility as a denoising meta-learner.



Paperid:3821
Authors:Xuelin Shen, Haifeng Jiao, Yitong Wang, Yulin HE, Wenhan Yang
Abstract:
Abstract:Although RAW images offer advantages over sRGB by avoiding ISP-induced distortion and preserving more information in low-light conditions, their widespread use is limited due to high storage costs, transmission burdens, and the need for significant architectural changes for downstream tasks. To address the issues, this paper explores a new raw-based machine vision paradigm, termed Compact RAW Metadata-guided Image Refinement (CRM-IR). In particular, we propose a Machine Vision-oriented Image Refinement (MV-IR) module that refines sRGB images to better suit machine vision preferences, guided by learned raw metadata. Such a design allows the CRM-IR to focus on extracting the most essential metadata from raw images to support downstream machine vision tasks, while remaining plug-and-play and fully compatible with existing imaging pipelines, without any changes to model architectures or ISP modules. We implement our CRM-IR scheme on various object detection networks, and extensive experiments under low-light conditions demonstrate that it can significantly improve performance with an additional bitrate cost of less than $10^{-3}$ bits per pixel.



Paperid:3822
Authors:Xiaoshan Wu, Yifei Yu, Bo Wang, Yihua Huang, Xiaoyang Lyu, Baoheng Zhang, Zhongrui Wang, Xiaojuan Qi
Abstract:
Robust 3D geometry estimation from videos is critical for applications such as autonomous navigation, SLAM, and 3D scene reconstruction. Recent methods like DUSt3R demonstrate that regressing dense pointmaps from image pairs enables accurate and efficient pose-free reconstruction. However, existing RGB-only approaches struggle under real-world conditions involving dynamic objects and extreme illumination, due to the inherent limitations of conventional cameras. In this paper, we propose EAG3R, a novel geometry estimation framework that augments pointmap-based reconstruction with asynchronous event streams. Built upon the MonST3R backbone, EAG3R introduces two key innovations: (1) a retinex-inspired image enhancement module and a lightweight event adapter with SNR-aware fusion mechanism that adaptively combines RGB and event features based on local reliability; and (2) a novel event-based photometric consistency loss that reinforces spatiotemporal coherence during global optimization. Our method enables robust geometry estimation in challenging dynamic low-light scenes without requiring retraining on night-time data. Extensive experiments demonstrate that EAG3R significantly outperforms state-of-the-art RGB-only baselines across monocular depth estimation, camera pose tracking, and dynamic reconstruction tasks.



Paperid:3823
Authors:Yuchen Zhang, Hanyue Du, Chun Cao, Jingwei Xu
Abstract:
Abstract:Low-Rank Adaptation (LoRA) has emerged as a widely adopted parameter-efficient fine-tuning (PEFT) technique to enhance the performance of large language models (LLMs) for downstream tasks. While prior work has explored integration strategies for LLM training and serving, there remains a gap in unifying fine-tuning and inference processes for LoRA-based models. In this paper, we present Loquetier, a virtualized multi-LoRA framework that enables unified LoRA fine-tuning and serving. Specifically, Loquetier consists of two main components: (1) a Virtualized Module that isolates model modifications to support multiple PEFT methods within a shared base model architecture,and (2) an optimized computation flow and kernel design that merges fine-tuning and inference paths in forward propagation, enabling efficient batching and minimizing kernel invocation overhead. Extensive experiments across three task settings demonstrate that Loquetier consistently outperforms existing baselines in performance and flexibility, achieving 3.0$\times$ the performance of the state-of-the-art co-serving system for the inference-only tasks, and 46.4$\times$ higher SLO attainment than PEFT in the unified tasks. The implementation of Loquetier is available at https://github.com/s3co3wjy5tr2bdfj/Loquetier.



Paperid:3824
Authors:Fengxiang Wang, Mingshuo Chen, Yueying Li, Di Wang, Haotian Wang, Zonghao Guo, Zefan Wang, Shan Boqi, Long Lan, Yulin Wang, Hongzhen Wang, Wenjing Yang, Bo Du, Jing Zhang
Abstract:
Abstract:Ultra-high-resolution (UHR) remote sensing (RS) imagery offers valuable data for Earth observation but pose challenges for existing multimodal foundation models due to two key bottlenecks: (1) limited availability of UHR training data, and (2) token explosion caused by the large image size. To address data scarcity, we introduce **SuperRS-VQA** (avg. 8,376$\times$8,376) and **HighRS-VQA** (avg. 2,000$\times$1,912), the highest-resolution vision-language datasets in RS to date, covering 22 real-world dialogue tasks. To mitigate token explosion, our pilot studies reveal significant redundancy in RS images: crucial information is concentrated in a small subset of object-centric tokens, while pruning background tokens (e.g., ocean or forest) can even improve performance.Motivated by these findings, we propose two strategies: *Background Token Pruning* and *Anchored Token Selection*, to reduce the memory footprint while preserving key semantics.Integrating these techniques, we introduce **GeoLLaVA-8K**, the first RS-focused multimodal large language model capable of handling inputs up to 8K$\times$8K resolution, built on the LLaVA framework. Trained on SuperRS-VQA and HighRS-VQA, GeoLLaVA-8K sets a new state-of-the-art on the XLRS-Bench. Datasets and code will be released.



Paperid:3825
Authors:Zixuan Xie, Xinyu Liu, Rohan Chandra, Shangtong Zhang
Abstract:
Abstract:Linear TD($\lambda$) is one of the most fundamental reinforcement learning algorithms for policy evaluation. Previously, convergence rates are typically established under the assumption of linearly independent features, which does not hold in many practical scenarios. This paper instead establishes the first $L^2$ convergence rates for linear TD($\lambda$) operating under arbitrary features, without making any algorithmic modification or additional assumptions. Our results apply to both the discounted and average-reward settings. To address the potential non-uniqueness of solutions resulting from arbitrary features, we develop a novel stochastic approximation result featuring convergence rates to the solution set instead of a single point.



Paperid:3826
Authors:James Chapman, Kedar Karhadkar, Guido Montufar
Abstract:
Deep reinforcement learning (DRL) has achieved remarkable success across multiple domains, including competitive games, natural language processing, and robotics. Despite these advancements, policies trained via DRL often struggle to generalize to evaluation environments with different parameters. This challenge is typically addressed by training with multiple contexts and/or by leveraging additional structure in the problem. However, obtaining sufficient training data across diverse contexts can be impractical in real-world applications. In this work, we consider contextual Markov decision processes (CMDPs) with transition and reward functions that exhibit regularity in context parameters. We introduce the context-enhanced Bellman equation (CEBE) to improve generalization when training on a single context. We prove both analytically and empirically that the CEBE yields a first-order approximation to the Q function trained across multiple contexts. We then derive context sample enhancement (CSE) as an efficient data augmentation method for approximating the CEBE in deterministic control environments. We numerically validate the performance of CSE in simulation environments, showcasing its potential to improve generalization in DRL.



Paperid:3827
Authors:Hiroshi Kera, Nico Pelleriti, Yuki Ishihara, Max Zimmer, Sebastian Pokutta
Abstract:
Abstract:Solving systems of polynomial equations, particularly those with finitely many solutions, is a crucial challenge across many scientific fields. Traditional methods like Gröbner and Border bases are fundamental but suffer from high computational costs, which have motivated recent Deep Learning approaches to improve efficiency, albeit at the expense of output correctness. In this work, we introduce the Oracle Border Basis Algorithm, the first Deep Learning approach that accelerates Border basis computation while maintaining output guarantees. To this end, we design and train a Transformer-based oracle that identifies and eliminates computationally expensive reduction steps, which we find to dominate the algorithm's runtime. By selectively invoking this oracle during critical phases of computation, we achieve substantial speedup factors of up to 3.5x compared to the base algorithm, without compromising the correctness of results. To generate the training data, we develop a sampling method and provide the first sampling theorem for border bases. We construct a tokenization and embedding scheme tailored to monomial-centered algebraic computations, resulting in a compact and expressive input representation, which reduces the number of tokens to encode an $n$-variate polynomial by a factor of $O(n)$. Our learning approach is data efficient, stable, and a practical enhancement to traditional computer algebra algorithms and symbolic computation.



Paperid:3828
Authors:Maryam Aliakbarpour, Zhan Shi, Ria Stevens, Vincent Wang
Abstract:
Abstract:Estimating the density of a distribution from its samples is a fundamental problem in statistics. \emph{Hypothesis selection} addresses the setting where, in addition to a sample set, we are given $n$ candidate distributions---referred to as \emph{hypotheses}---and the goal is to determine which one best describes the underlying data distribution. This problem is known to be solvable very efficiently, requiring roughly $O(\log n)$ samples and running in $\tilde{O}(n)$ time. The quality of the output is measured via the total variation distance to the unknown distribution, and the approximation factor of the algorithm determines how large this distance is compared to the optimal distance achieved by the best candidate hypothesis. It is known that $\alpha = 3$ is the optimal approximation factor for this problem.We study hypothesis selection under the constraint of \emph{differential privacy}. We propose a differentially private algorithm in the central model that runs in nearly linear time with respect to the number of hypotheses, achieves the optimal approximation factor, and incurs only a modest increase in sample complexity, which remains polylogarithmic in $n$. This resolves an open question posed by [Bun, Kamath, Steinke, Wu, NeurIPS 2019]. Prior to our work, existing upper bounds required quadratic time.



Paperid:3829
Authors:Ning Li, Xiangmou Qu, Jiamu Zhou, Jun Wang, Muning Wen, Kounianhua Du, Xingyu Lou, Qiuying Peng, Jun Wang, Weinan Zhang
Abstract:
Recent advances in Multimodal Large Language Models (MLLMs) have enabled the development of mobile agents that can understand visual inputs and follow user instructions, unlocking new possibilities for automating complex tasks on mobile devices.However, applying these models to real-world mobile scenarios remains a significant challenge due to the long-horizon task execution, difficulty in error recovery, and the cold-start problem in unfamiliar environments. To address these challenges, we propose MobileUse, a GUI agent designed for robust and adaptive mobile task execution. To improve resilience in long-horizon tasks and dynamic environments, we introduce a hierarchical reflection architecture that enables the agent to self-monitor, detect, and recover from errors across multiple temporal scales—ranging from individual actions to overall task completion—while maintaining efficiency through a reflection-on-demand strategy. To tackle cold-start issues, we further introduce a proactive exploration module, which enriches the agent’s understanding of the environment through self-planned exploration. Evaluations on the AndroidWorld and AndroidLab benchmarks demonstrate that MobileUse establishes new state-of-the-art performance, achieving success rates of 62.9% and 44.2%, respectively. To facilitate real-world applications, we release an out-of-the-box toolkit for automated task execution on physical mobile devices.



Paperid:3830
Authors:Yanlong Chen, Mattia Orlandi, Pierangelo Rapa, Simone Benatti, Luca Benini, Yawei Li
Abstract:
Physiological signals are often corrupted by motion artifacts, baseline drift, and other low-SNR disturbances, posing significant challenges for analysis. Additionally, these signals exhibit strong non-stationarity, with sharp peaks and abrupt changes that evolve continuously, making them difficult to represent using traditional time-domain or filtering methods. To address these issues, a novel wavelet-based approach for physiological signal analysis is presented, aimed at capturing multi-scale time-frequency features across various physiological signals. Leveraging this technique, two large-scale pretrained models specific to EMG and ECG are introduced for the first time, achieving superior performance and setting new baselines in downstream tasks. Additionally, a unified multi-modal framework is constructed by integrating a pretrained EEG model, where each modality is guided through its dedicated branch and fused via learnable weighted fusion. This design effectively addresses challenges such as low signal-to-noise ratio, high inter-subject variability, and device mismatch, outperforming existing methods on multi-modal tasks. The proposed wavelet-based architecture lays a solid foundation for the analysis of diverse physiological signals, while the multi-modal design points to next-generation physiological signal processing with potential impacts on wearable health monitoring, clinical diagnostics, and broader biomedical applications.



Paperid:3831
Authors:Landon Butler, Abhineet Agarwal, Justin Kang, Yigit Efe Erginbas, Bin Yu, Kannan Ramchandran
Abstract:
Abstract:Large Language Models (LLMs) have achieved remarkable performance by capturing complex interactions between input features. To identify these interactions, most existing approaches require enumerating all possible combinations of features up to a given order, causing them to scale poorly with the number of inputs $n$. Recently, Kang et al. (2025) proposed SPEX, an information-theoretic approach that uses interaction sparsity to scale to $n \approx 10^3$ features. SPEX greatly improves upon prior methods but requires tens of thousands of model inferences, which can be prohibitive for large models. In this paper, we observe that LLM feature interactions are often *hierarchical*—higher-order interactions are accompanied by their lower-order subsets—which enables more efficient discovery.To exploit this hierarchy, we propose ProxySPEX, an interaction attribution algorithm that first fits gradient boosted trees to masked LLM outputs and then extracts the important interactions. Experiments across four challenging high-dimensional datasets show that ProxySPEX more faithfully reconstructs LLM outputs by 20\% over marginal attribution approaches while using *$10\times$ fewer inferences* than SPEX. By accounting for interactions, ProxySPEX identifies features that influence model output over 20\% more than those selected by marginal approaches. Further, we apply ProxySPEX to two interpretability tasks. *Data attribution*, where we identify interactions among CIFAR-10 training samples that influence test predictions, and *mechanistic interpretability*, where we uncover interactions between attention heads, both within and across layers, on a question-answering task.ProxySPEX identifies interactions that enable more aggressive pruning of heads than marginal approaches.



Paperid:3832
Authors:Jingpu Cheng, Qianxiao Li, Ting Lin, Zuowei Shen
Abstract:
We investigate the universal approximation property (UAP) of transformer-type architectures, providing a unified theoretical framework that extends prior results on residual networks to models incorporating attention mechanisms. Our work identifies token distinguishability as a fundamental requirement for UAP and introduces a general sufficient condition that applies to a broad class of architectures. Leveraging an analyticity assumption on the attention layer, we can significantly simplify the verification of this condition, providing a non-constructive approach in establishing UAP for such architectures. We demonstrate the applicability of our framework by proving UAP for transformers with various attention mechanisms, including kernel-based and sparse ones. The corollaries of our results either generalize prior works or establish UAP for architectures not previously covered. Furthermore, our framework offers a principled foundation for designing novel transformer architectures with inherent UAP guarantees, including those with specific functional symmetries. We propose examples to illustrate these insights.



Paperid:3833
Authors:Yinxu Tang, Stylianos Loukas Vasileiou, Vincent Derkinderen, William Yeoh
Abstract:
In human-AI interaction, effective communication relies on aligning the AI agent’s model with the human user’s mental model -- a process known as model reconciliation. However, existing model reconciliation approaches predominantly assume deterministic models, overlooking the fact that human knowledge is often uncertain or probabilistic. To bridge this gap, we present a probabilistic model reconciliation framework that resolves inconsistencies in MPE outcome probabilities between an agent’s and a user’s models.Our approach is built on probabilistic logic programming (PLP) using ProbLog, where explanations are generated as cost-optimal model updates that reconcile these probabilistic differences. We develop two search algorithms -- a generic and an optimized version, with the latter guided by theoretical insights for scalability. Our approach is validated through a user study on how explanation types impact user understanding and computational experiments showing the optimized search consistently outperforms the generic.



Paperid:3834
Authors:Anay Majee, Amitesh Gangrade, Rishabh Iyer
Abstract:
Abstract:Open-World Object Detection (OWOD) enriches traditional object detectors by enabling continual discovery and integration of unknown objects via human guidance. However, existing OWOD approaches frequently suffer from semantic confusion between known and unknown classes, alongside catastrophic forgetting, leading to diminished unknown recall and degraded known-class accuracy. To overcome these challenges, we propose **C**ombinato**r**ial **O**pen-**W**orld **D**etection (**CROWD**), a unified framework reformulating unknown object discovery and adaptation as an interwoven combinatorial (set-based) data-discovery (CROWD-Discover) and representation learning (CROWD-Learn) task. CROWD-Discover strategically mines unknown instances by maximizing Submodular Conditional Gain (SCG) functions, selecting representative examples distinctly dissimilar from known objects. Subsequently, CROWD-Learn employs novel combinatorial objectives that jointly disentangle known and unknown representations while maintaining discriminative coherence among known classes, thus mitigating confusion and forgetting. Extensive evaluations on OWOD benchmarks illustrate that CROWD achieves improvements of 2.83% and 2.05% in known-class accuracy on M-OWODB and S-OWODB, respectively, and nearly 2.4$\times$ unknown recall compared to leading baselines.



Paperid:3835
Authors:Guanchen Li, Yixing Xu, Zeping Li, Ji Liu, Xuanwu Yin, Dong Li, Emad Barsoum
Abstract:
Structural pruning enhances hardware-agnostic inference efficiency for large language models (LLMs) but often struggles to maintain performance. Local pruning performs efficient layer-by-layer compression but ignores global topology. Although global pruning aims to identify an optimal sparse model, existing methods tend to adopt a two-stage paradigm—first evaluating the saliency of individual substructures and then applying pruning globally, which ignores inter-structure dependencies and fails to achieve end-to-end optimization.To address these limitations, we proposeTýr-the-Pruner, an efficient end-to-end search-based global structural pruning framework. This framework constructs a supernet by repeatedly applying local pruning across a range of sparsity ratios to each layer in an LLM, with the core goal of determining the optimal sparsity distribution under a target overall sparsity ratio. Concretely, we introduce an effective local pruning and an expectation error accumulation approach to improve supernet construction. Furthermore, we employ an iterative prune-and-search strategy with coarse-to-fine sparsity granularity to ensure efficient search convergence. Experimental results show that Týr-the-Pruner achieves state-of-the-art structural pruning, retaining97%of the dense model's performance while removing a challenging50%of Llama-3.1-70B's parameters.



Paperid:3836
Authors:Changshuo Wang, Shuting He, Xiang Fang, Zhijian Hu, JIA-HONG HUANG, Yixian Shen, Prayag Tiwari
Abstract:
Point Cloud Few-Shot Semantic Segmentation (PC-FSS) aims to segment unknown categories in query samples using only a small number of annotated support samples. However, scene complexity and insufficient representation of local geometric structures pose significant challenges to PC-FSS. To address these issues, we propose a novel pre-training-free Visual Introspective Prototype Segmentation network (VIP-Seg). Specifically, inspired by Chain-of-Thought reasoning, we design a Visual Introspective Prototype (VIP) module to tackle intra-class diversity and domain gaps between support and query sets. The VIP module consists of a Prototype Enhancement Module (PEM) and a Prototype Difference Module (PDM), which work alternately to form a multi-step reasoning process. The PEM enhances prototype discriminability and reduces intra-class diversity, while the PDM learns common representations from the differences between query and support features, effectively eliminating semantic inconsistencies caused by domain gaps. To further reduce intra-class diversity and enhance point discriminative ability, we propose a Dynamic Taylor Convolution (DyTaylorConv), inspired by the Taylor series, which effectively captures local geometric structures and detailed features of point clouds. Extensive experiments on S3DIS and ScanNet demonstrate that our proposed VIP-Seg significantly outperforms current state-of-the-art methods, proving the its effectiveness in PC-FSS tasks.



Paperid:3837
Authors:Jiajun Fan, Tong Wei, Chaoran Cheng, Yuxin Chen, Ge Liu
Abstract:
Balancing exploration and exploitation during reinforcement learning fine-tuning of generative models presents a critical challenge, as existing approaches rely on fixed divergence regularization that creates an inherent dilemma: strong regularization preserves model capabilities but limits reward optimization, while weak regularization enables greater alignment but risks instability or reward hacking. We introduce Adaptive Divergence Regularized Policy Optimization (ADRPO), which automatically adjusts regularization strength based on advantage estimates—reducing regularization for high-value samples while applying stronger regularization to poor samples, enabling policies to navigate between exploration and aggressive exploitation according to data quality. Our implementation with Wasserstein-2 regularization for flow matching generative models achieves remarkable results on text-to-image generation, achieving better semantics alignment and diversity than offline methods like DPO and online methods with fixed regularization like ORW-CFM-W2. ADRPO also enables 2B parameter SD3 model to surpass much larger models with 4.8B and 12B parameters in attribute binding, semantic consistency, artistic style transfer, and compositional control while maintaining generation diversity. ADRPO can also generalize to KL-regularized LLM fine-tuning, enhancing existing online RL methods like GRPO while requiring no additional networks or complex architectural changes. In LLM fine-tuning tasks, we observe that ADRPO even demonstrates an emergent ability to escape local optima by actively increasing exploration to discover superior policies, thus offering an effective, plug-and-play solution to the exploration-exploitation challenge in generative model fine-tuning.



Paperid:3838
Authors:Hadi Askari, Shivanshu Gupta, Fei Wang, Anshuman Chhabra, Muhao Chen
Abstract:
Pretrained Large Language Models (LLMs) achieve strong performance across a wide range of tasks, yet exhibit substantial variability in the various layers' training quality with respect to specific downstream applications, limiting their downstream performance. It is therefore critical to estimate layer-wise training quality in a manner that accounts for both model architecture and training data. However, existing approaches predominantly rely on model-centric heuristics (such as spectral statistics, outlier detection, or uniform allocation) while overlooking the influence of data. To address these limitations, we proposeLayerIF, a data-driven framework that leveragesInfluence Functionsto quantify the training quality of individual layers in a principled and task-sensitive manner. By isolating each layer's gradients and measuring the sensitivity of the validation loss to training examples by computing layer-wise influences, we derive data-driven estimates of layer importance. Notably, our method producestask-specificlayer importance estimates for thesameLLM, revealing how layers specialize for different test-time evaluation tasks. We demonstrate the utility of our scores by leveraging them for two downstream applications: (a) expert allocation in LoRA-MoE architectures and (b) layer-wise sparsity distribution for LLM pruning. Experiments across multiple LLM architectures demonstrate that our model-agnostic, influence-guided allocation leads to consistent gains in task performance.



Paperid:3839
Authors:Sonia Mazelet, Rémi Flamary, Bertrand Thirion
Abstract:
Optimal transport between graphs, based on Gromov-Wasserstein and other extensions, is a powerful tool for comparing and aligning graph structures. However, solving the associated non-convex optimization problems is computationally expensive, which limits the scalability of these methods to large graphs. In this work, we present Unbalanced Learning of Optimal Transport (ULOT), a deep learning method that predicts optimal transport plans between two graphs. Our method is trained by minimizing the fused unbalanced Gromov-Wasserstein (FUGW) loss. We propose a novel neural architecture with cross-attention that is conditioned on the FUGW tradeoff hyperparameters. We evaluate ULOT on synthetic stochastic block model (SBM) graphs and on real cortical surface data obtained from fMRI. ULOT predicts transport plans with competitive loss up to two orders of magnitude faster than classical solvers. Furthermore, the predicted plan can be used as a warm start for classical solvers to accelerate their convergence. Finally, the predicted transport plan is fully differentiable with respect to the graph inputs and FUGW hyperparameters, enabling the optimization of functionals of the ULOT plan.



Paperid:3840
Authors:Songhao Han, Boxiang Qiu, Yue Liao, Siyuan Huang, Chen Gao, Shuicheng Yan, Si Liu
Abstract:
Recent advances in vision-language models (VLMs) have enabled instruction-conditioned robotic systems with improved generalization. However, most existing work focuses on reactive System 1 policies, underutilizing VLMs’ strengths in semantic reasoning and long-horizon planning. These System 2 capabilities—characterized by deliberative, goal-directed thinking—remain underexplored due to the limited temporal scale and structural complexity of current benchmarks. To address this gap, we introduce RoboCerebra, a benchmark for evaluating high-level reasoning in long-horizon robotic manipulation. RoboCerebra includes: (1) a large-scale simulation dataset with extended task horizons and diverse subtask sequences in household environments; (2) a hierarchical framework combining a high-level VLM planner with a low-level vision-language-action (VLA) controller; and (3) an evaluation protocol targeting planning, reflection, and memory through structured System 1–System 2 interaction. The dataset is constructed via a top-down pipeline, where GPT generates task instructions and decomposes them into subtask sequences. Human operators execute the subtasks in simulation, yielding high-quality trajectories with dynamic object variations. Compared to prior benchmarks, RoboCerebra features significantly longer action sequences and denser annotations. We further benchmark state-of-the-art VLMs as System 2 modules and analyze their performance across key cognitive dimensions, advancing the development of more capable and generalizable robotic planners.



Paperid:3841
Authors:kaiyuan Li, Xiaoyue Chen, Chen Gao, Yong Li, Xinlei Chen
Abstract:
Large Vision-Language Models (LVLMs) have shown impressive performance across multi-modal tasks by encoding images into thousands of tokens. However, the large number of image tokens results in significant computational overhead, and the use of dynamic high-resolution inputs further increases this burden. Previous approaches have attempted to reduce the number of image tokens through token pruning, typically by selecting tokens based on attention scores or image token diversity. Through empirical studies, we observe that existing methods often overlook the joint impact of pruning on both the current layer’s output (local) and the outputs of subsequent layers (global), leading to suboptimal pruning decisions. To address this challenge, we propose Balanced Token Pruning (BTP), a plug-and-play method for pruning vision tokens. Specifically, our method utilizes a small calibration set to divide the pruning process into multiple stages. In the early stages, token pruning emphasizes their impact on downstream layers, whereas in the deeper stages, the focus shifts to their influence on outputs within the current layer. Extensive experiments across various LVLMs demonstrate the broad effectiveness of our approach on multiple benchmarks. Our source code is publicly available at https://anonymous.4open.science/r/BTP-EE00TY89U/.



Paperid:3842
Authors:Jiahao Ma, Lei Wang, miaomiao Liu, David Ahmedt-Aristizabal, Chuong Nguyen
Abstract:
Multi-view 3D reconstruction remains a core challenge in computer vision. Recent methods, such as DUSt3R and its successors, directly regress pointmaps from image pairs without relying on known scene geometry or camera parameters. However, the performance of these models is constrained by the diversity and scale of available training data. In this work, we introduce Puzzles, a data augmentation strategy that synthesizes an unbounded volume of high-quality, posed video-depth data from just a single image or video clip. By simulating diverse camera trajectories and realistic scene geometry through targeted image transformations, Puzzles significantly enhances data variety. Extensive experiments show that integrating Puzzles into existing video‑based 3D reconstruction pipelines consistently boosts performance, all without modifying the underlying network architecture. Notably, models trained on only 10% of the original data, augmented with Puzzles, achieve accuracy comparable to those trained on the full dataset.



Paperid:3843
Authors:Léo Dana, Loucas Pillaud-Vivien, Francis Bach
Abstract:
Abstract:We analyse the convergence of one-hidden-layer ReLU networks trained by gradient flow on $n$ data points. Our main contribution leverages the high dimensionality of the ambient space, which implies low correlation of the input samples, to demonstrate that a network with width of order $\log(n)$ neurons suffices for global convergence with high probability. Our analysis uses a Polyak–Łojasiewicz viewpoint along the gradient-flow trajectory, which provides an exponential rate of convergence of $\frac{1}{n}$. When the data are exactly orthogonal, we give further refined characterizations of the convergence speed, proving its asymptotic behavior lies between the orders $\frac{1}{n}$ and $\frac{1}{\sqrt{n}}$, and exhibiting a phase-transition phenomenon in the convergence rate, during which it evolves from the lower bound to the upper, and in a relative time of order $\frac{1}{\log(n)}$.



Paperid:3844
Authors:Dylan Zapzalka, Trenton Chang, Lindsay Warrenburg, Sae-Hwan Park, Daniel Shenfeld, Ravi Parikh, Jenna Wiens, Maggie Makar
Abstract:
In settings where ML models are used to inform the allocation of resources, agents affected by the allocation decisions might have an incentive to strategically change their features to secure better outcomes. They may do so in two ways: through genuine modification – where agents genuinely alter their behavior leading to changes in their true feature values – or misreporting, where agents report false features without changing underlying behaviors. E.g., unemployed loan applicants may genuinely modify their behavior by getting a job or misreport their employment status as employed to get a more favorable risk prediction. While prior work has studied strategic responses broadly, disentangling misreporting from genuine modification remains a fundamental challenge. In this paper, we propose a causally-motivated approach to identify and quantify how much an agent misreports on average by distinguishing deceptive changes in their features from genuine modification. Our key insight is that, unlike genuine modification, misreported features do not causally affect downstream variables (i.e., causal descendants). We exploit this asymmetry by comparing the causal effect of misreported features on their causal descendants as derived from manipulated datasets against those from unmanipulated datasets. We formally prove identifiability of the misreporting rate and characterize the variance of our estimator. We empirically validate our theoretical results using a semi-synthetic and real Medicare dataset with misreported data, demonstrating that our approach can be employed to identify misreporting in real-world scenarios.



Paperid:3845
Authors:Kunjun Li, Zigeng Chen, Cheng-Yen Yang, Jenq-Neng Hwang
Abstract:
Visual Autoregressive (VAR) modeling has garnered significant attention for its innovative next-scale prediction approach, which yields substantial improvements in efficiency, scalability, and zero-shot generalization. Nevertheless, the coarse-to-fine methodology inherent in VAR results in exponential growth of the KV cache during inference, causing considerable memory consumption and computational redundancy. To address these bottlenecks, we introduce ScaleKV, a novel KV cache compression framework tailored for VAR architectures. ScaleKV leverages two critical observations: varying cache demands across transformer layers and distinct attention patterns at different scales. Based on these insights, ScaleKV categorizes transformer layers into two functional groups: drafters and refiners. Drafters exhibit dispersed attention across multiple scales, thereby requiring greater cache capacity. Conversely, refiners focus attention on the current token map to process local details, consequently necessitating substantially reduced cache capacity. ScaleKV optimizes the multi-scale inference pipeline by identifying scale-specific drafters and refiners, facilitating differentiated cache management tailored to each scale. Evaluation on the state-of-the-art text-to-image VAR model family, Infinity, demonstrates that our approach effectively reduces the required KV cache memory to 10% while preserving pixel-level fidelity.



Paperid:3846
Authors:Pengwei Liu, Hangjie Yuan, Bo Dong, Jiazheng Xing, Jinwang Wang, Rui Zhao, Weihua Chen, Fan Wang
Abstract:
Image and video relighting remains a challenging task, where controlling illumination under ensures physical consistency, and efficiency under diverse lighting conditions.However, existing relighting methods lack explicit physical grounding, often resulting in perceptually implausible illumination under diverse lighting conditions, such as overexposed highlights or inconsistent shadow placement.In addition, similar to general-purpose generative models, their reliance on multi-iterative sampling leads to slow inference. This is particularly problematic considering the strong conditioning in relighting tasks, which should intuitively enable much faster generation, especially for real-time applications.To address these challenges, we propose UniLumos, a unified framework for fast and physically plausible relighting of both images and videos. First, we introduce LumosData, a lightweight pipeline for constructing diverse relighting pairs from the illumination augmentations of real-world data. Leveraging the strong conditioning provided by LumosData, UniLumos extends a flow-matching backbone with path consistency learning to enable fast inference without multi-step sampling.Second, inspired by the inherent geometric consistency in dense estimation, we introduce a physics-plausible feedback mechanism that supervises the latent diffusion process through depth and normal alignment with the labeled references, encouraging physically realistic illumination.Experiments demonstrate that UniLumos achieves high-quality and physically plausible relighting with significantly faster inference for both image and video relighting.



Paperid:3847
Authors:Zelin Peng, Zhengqin Xu, Qingyang Liu, Xiaokang Yang, Wei Shen
Abstract:
Multi-modal large language models (MLLMs) have emerged as a transformative approach for aligning visual and textual understanding. They typically require extremely high computational resources (e.g., thousands of GPUs) for training to achieve cross-modal alignment at multi-granularity levels. We argue that a key source of this inefficiency lies in the vision encoders they widely equip with, e.g., CLIP and SAM, which lack the alignment with language at multi-granularity levels. To address this issue, in this paper, we leverage hyperbolic space, which inherently models hierarchical levels and thus provides a principled framework for bridging the granularity gap between visual and textual modalities at an arbitrary granularity level. Concretely, we propose an efficient training paradigm for MLLMs, dubbed as \blg, which can optimize visual representations to align with their textual counterparts at an arbitrary granularity level through dynamic hyperbolic radius adjustment in hyperbolic space. \alg employs learnable matrices with M\"{o}bius multiplication operations, implemented via three effective configurations: diagonal scaling matrices, block-diagonal matrices, and banded matrices, providing a flexible yet efficient parametrization strategy. Comprehensive experiments across multiple MLLM benchmarks demonstrate that \alg consistently improves both existing pre-training and fine-tuning MLLMs by large margins with less than 1\% additional parameters.



Paperid:3848
Authors:Hoonhee Cho, Jae-Young Kang, Giwon Lee, Hyemin Yang, Heejun Park, Seokwoo Jung, Kuk-Jin Yoon
Abstract:
End-to-end autonomous driving (E2E-AD) has emerged as a promising paradigm that unifies perception, prediction, and planning into a holistic, data-driven framework. However, achieving robustness to varying camera viewpoints, a common real-world challenge due to diverse vehicle configurations, remains an open problem. In this work, we propose VR-Drive, a novel E2E-AD framework that addresses viewpoint generalization by jointly learning 3D scene reconstruction as an auxiliary task to enable planning-aware view synthesis. Unlike prior scene-specific synthesis approaches, VR-Drive adopts a feed-forward inference strategy that supports online training-time augmentation from sparse views without additional annotations. To further improve viewpoint consistency, we introduce a viewpoint-mixed memory bank that facilitates temporal interaction across multiple viewpoints and a viewpoint-consistent distillation strategy that transfers knowledge from original to synthesized views. Trained in a fully end-to-end manner, VR-Drive effectively mitigates synthesis-induced noise and improves planning under viewpoint shifts. In addition, we release a new benchmark dataset to evaluate E2E-AD performance under novel camera viewpoints, enabling comprehensive analysis. Our results demonstrate that VR-Drive is a scalable and robust solution for the real-world deployment of end-to-end autonomous driving systems. Our code and datasets will be made publicly available.



Paperid:3849
Authors:Yinfang Chen, Jiaqi Pan, Jackson Clark, Yiming Su, Noah Zheutlin, Bhavya Bhavya, Rohan R. Arora, Yu Deng, Saurabh Jha, Tianyin Xu
Abstract:
In cloud-scale systems, failures are the norm. A distributed computing cluster exhibits hundreds of machine failures and thousands of disk failures; software bugs and misconfigurations are reported to be more frequent. The demand of autonomous, AI-driven reliability engineering continues to grow, as existing human-in-the-loop practices can hardly keep up with the scale of modern clouds. This paper presents STRATUS, an LLM-based multi-agent system for realizing autonomous Site Reliability Engineering (SRE) of cloud services. STRATUS consists of multiple specialized agents (e.g., for failure detection, diagnosis, mitigation), organized in a state machine to assist system-level safety reasoning and enforcement. We formalize a key safety specification of agentic SRE systems like STRATUS, termed Transactional No-Regression (TNR), which enables safe exploration and iteration. We show that TNR can effectively improve autonomous failure mitigation. STRATUS significantly outperforms state-of-the-art SRE agents in terms of success rate of failure mitigation problems in AIOpsLab and ITBench (two SRE benchmark suites), by at least 1.5 times across various models. STRATUS shows a promising path toward practical deployment of agentic systems for cloud reliability.



Paperid:3850
Authors:Jonas Kulhanek, Torsten Sattler
Abstract:
Novel view synthesis is an important problem with many applications, including AR/VR, gaming, and robotic simulations. With the recent rapid development of Neural Radiance Fields (NeRFs) and 3D Gaussian Splatting (3DGS) methods, it is becoming difficult to keep track of the current state of the art (SoTA) due to methods using different evaluation protocols, codebases being difficult to install and use, and methods not generalizing well to novel 3D scenes. In our experiments, we show that even tiny differences in the evaluation protocols of various methods can artificially boost the performance of these methods. This raises questions about the validity of quantitative comparisons performed in the literature. To address these questions, we propose NerfBaselines, an evaluation framework which provides consistent benchmarking tools, ensures reproducibility, and simplifies the installation and use of various methods. We validate our implementation experimentally by reproducing the numbers reported in the original papers. For improved accessibility, we release a web platform that compares commonly used methods on standard benchmarks. We strongly believe NerfBaselines is a valuable contribution to the community as it ensures that quantitative results are comparable and thus truly measure progress in the field of novel view synthesis.



Paperid:3851
Authors:Jian-Ting Guo, Yu-Cheng Chen, Ping-Chun Hsieh, Kuo-Hao Ho, Po-Wei Huang, Ti-Rong Wu, I-Chen Wu
Abstract:
Human-like agents have long been one of the goals in pursuing artificial intelligence.Although reinforcement learning (RL) has achieved superhuman performance in many domains, relatively little attention has been focused on designing human-like RL agents.As a result, many reward-driven RL agents often exhibit unnatural behaviors compared to humans, raising concerns for both interpretability and trustworthiness.To achieve human-like behavior in RL, this paper first formulates human-likeness as trajectory optimization, where the objective is to find an action sequence that closely aligns with human behavior while also maximizing rewards, and adapts the classic receding-horizon control to human-like learning as a tractable and efficient implementation.To achieve this, we introduce Macro Action Quantization (MAQ), a human-like RL framework that distills human demonstrations into macro actions via Vector-Quantized VAE.Experiments on D4RL Adroit benchmarks show that MAQ significantly improves human-likeness, increasing trajectory similarity scores, and achieving the highest human-likeness rankings among all RL agents in the human evaluation study.Our results also demonstrate that MAQ can be easily integrated into various off-the-shelf RL algorithms, opening a promising direction for learning human-like RL agents.



Paperid:3852
Authors:Yunyang Li, Lin Huang, Zhihao Ding, Xinran Wei, Chu Wang, Han Yang, Zun Wang, Chang Liu, Yu Shi, Peiran Jin, Jia Zhang, Mark Gerstein, Tao Qin
Abstract:
Abstract:Equivariant Graph Neural Networks (EGNNs) have demonstrated significant success in modeling microscale systems, including those in chemistry, biology and materials science. However, EGNNs face substantial computational challenges due to the high cost of constructing edge features via spherical tensor products, making them almost impractical for large-scale systems. To address this limitation, we introduce E2Former, an equivariant and efficient transformer architecture that incorporates a Wigner $6j$ convolution (Wigner $6j$ Conv). By shifting the computational burden from edges to nodes, Wigner $6j$ Conv reduces the complexity from $O(| \mathcal{E}|)$ to $O(| \mathcal{V}|)$ while preserving both the model's expressive power and rotational equivariance.We show that this approach achieves a 7x–30x speedup compared to conventional $\mathrm{SO}(3)$ convolutions. Furthermore, our empirical results demonstrate that the derived E2Former mitigates the computational challenges of existing approaches without compromising the ability to capture detailed geometric information. This development could suggest a promising direction for scalable molecular modeling.



Paperid:3853
Authors:Qing Li, Huifang Feng, Xun Gong, Yu-Shen Liu
Abstract:
3D Gaussian Splatting has recently emerged as an efficient solution for high-quality and real-time novel view synthesis. However, its capability for accurate surface reconstruction remains underexplored. Due to the discrete and unstructured nature of Gaussians, supervision based solely on image rendering loss often leads to inaccurate geometry and inconsistent multi-view alignment. In this work, we propose a novel method that enhances the geometric representation of 3D Gaussians through view alignment (VA). Specifically, we incorporate edge-aware image cues into the rendering loss to improve surface boundary delineation. To enforce geometric consistency across views, we introduce a visibility-aware photometric alignment loss that models occlusions and encourages accurate spatial relationships among Gaussians. To further mitigate ambiguities caused by lighting variations, we incorporate normal-based constraints to refine the spatial orientation of Gaussians and improve local surface estimation. Additionally, we leverage deep image feature embeddings to enforce cross-view consistency, enhancing the robustness of the learned geometry under varying viewpoints and illumination. Extensive experiments on standard benchmarks demonstrate that our method achieves state-of-the-art performance in both surface reconstruction and novel view synthesis. We will release our code to support future research.



Paperid:3854
Authors:Natalie Maus, Kyurae Kim, Yimeng Zeng, Haydn Jones, Fangping Wan, Marcelo Der Torossian Torres, Cesar de la Fuente-Nunez, Jacob Gardner
Abstract:
Abstract:In multi-objective black-box optimization, the goal is typically to find solutions that optimize a set of $T$ black-box objective functions, $f_1, \ldots f_T$, simultaneously. Traditional approaches often seek a single Pareto-optimal set that balances trade-offs among all objectives. In this work, we consider a problem setting that departs from this paradigm: finding a small set of $K < T$ solutions, that collectively "covers" the $T$ objectives. A set of solutions is defined as "covering" if, for each objective $f_1, \ldots f_T$, there is at least one good solution. A motivating example for this problem setting occurs in drug design. For example, we may have $T$ pathogens and aim to identify a set of $K < T$ antibiotics such that at least one antibiotic can be used to treat each pathogen. To address this problem, we propose Multi-Objective Coverage Bayesian Optimization (MOCOBO), a principled algorithm designed to efficiently find a covering set. We validate our approach through experiments on challenging high-dimensional tasks, including applications in peptide and molecular design, where MOCOBO is shown to find high-performing covering sets of solutions.The results show that the coverage of the $K < T$ solutions found by MOCOBO matches or nearly matches the coverage of $T$ solutions obtained by optimizing each objective individually.Furthermore, in *in vitro* experiments, the peptides found by MOCOBO exhibited high potency against drug-resistant pathogens, further demonstrating the potential of MOCOBO for drug discovery.



Paperid:3855
Authors:Han-Jia Ye, Si-Yang Liu, Wei-Lun (Harry) Chao
Abstract:
Tabular datasets are inherently heterogeneous, presenting significant challenges for developing pre-trained foundation models. The recently introduced transformer-based Tabular Prior-data Fitted Network v2 (TabPFN v2) achieves unprecedentedin-context learningperformance across diverse downstream datasets, marking a pivotal advancement in tabular foundation models. In this paper, we take a closer look at TabPFN v2 to examine how it effectively handles heterogeneity and achieves high predictive accuracy, and to explore how its limitations in high-dimensional, many-category, and large-scale tasks can be mitigated. We find that TabPFN v2 can infer attribute relationships even when provided with randomized attribute token inputs, eliminating the need to explicitly learn dataset-specific attribute embeddings to address heterogeneity. We further show that TabPFN v2 can be transformed into a feature extractor, revealing its ability to construct a highly separable feature space for accurate predictions. Lastly, we demonstrate that TabPFN v2's limitations can be addressed through a test-time divide-and-conquer strategy, enabling scalable inference without requiring re-training. By uncovering the mechanisms behind TabPFN v2's success and introducing strategies to extend its applicability, this study offers key insights into the design of future tabular foundation models.



Paperid:3856
Authors:Sunqi Fan, Jiashuo Cui, Meng-Hao Guo, Shuojin Yang
Abstract:
Video Question Answering (VideoQA) task serves as a critical playground for evaluating whether foundation models can effectively perceive, understand, and reason about dynamic real-world scenarios. However, existing Multimodal Large Language Models (MLLMs) struggle with simultaneously ensuring the ability to model spatial relationships between video frames and to understand the causal dynamics of temporal evolution on complex and reasoning-intensive VideoQA. In this work, we equip MLLM with a comprehensive and extensible Video Toolkit, to enhance MLLM’s spatiotemporal reasoning capabilities as well as guarantee the harmony between the quantity and diversity of tools. To better control the tool invocation sequence and avoid toolchain shortcut issues, we propose a Spatiotemporal Reasoning Framework (STAR) that strategically schedules temporal and spatial tools, thereby progressively localizing the key area in the video. Our STAR framework enhances GPT-4o using lightweight tools, achieving an 8.2% gain on VideoMME and 4.6% on LongVideoBench. We believe that our proposed Video Toolkit and STAR framework make an important step towards building autonomous and intelligent video analysis assistants. Code will be publicly available.



Paperid:3857
Authors:Lam Ngo, Huong Ha, Jeffrey Chan, Hongyu Zhang
Abstract:
Bayesian Optimization (BO) is a powerful tool for optimizing expensive black-box objective functions. Much research has been conducted on the single-objective optimization problem, while the multi-objective optimization problem remains challenging. In this paper, we propose NBI-MOBO, a multi-objective Bayesian Optimization algorithm inspired by the Normal Boundary Intersection framework. NBI-MOBO decomposes the multi-objective problem into a series of single-objective subproblems, referred to as NBI-MOBO subproblems, each constrained along a predefined search direction in the objective space. Solving these subproblems yields a set of Pareto-optimal candidate solutions. NBI-MOBO employs a well-distributed set of search directions across the objective space, which enhances solution diversity, hence improving hypervolume performance of selected solutions. Furthermore, to improve coverage of the Pareto-optimal candidate solutions without requiring a large number of subproblems, we leverage a Pareto Front Estimation technique that locally explores the neighborhood of each obtained candidate, generating additional Pareto-optimal candidates. Additionally, NBI-MOBO can scale to more than three objectives while maintaining good hypervolume performance. Our experiments and analysis on various synthetic and real-world benchmark functions with various objective functions, ranging from 2 to 6 objectives, show that NBI-MOBO can outperform the state of the arts.



Paperid:3858
Authors:Luca Arnaboldi, Bruno Loureiro, Ludovic Stephan, Florent Krzakala, Lenka Zdeborová
Abstract:
We study the dynamics of stochastic gradient descent (SGD) for a class of sequence models termed Sequence Single-Index (SSI) models, where the target depends on a single direction in input space applied to a sequence of tokens. This setting generalizes classical single-index models to the sequential domain, encompassing simplified one-layer attention architectures. We derive a closed-form expression for the population loss in terms of a pair of sufficient statistics capturing semantic and positional alignment, and characterize the induced high-dimensional SGD dynamics for these coordinates. Our analysis reveals two distinct training phases: escape from uninformative initialization and alignment with the target subspace, and demonstrates how the sequence length and positional encoding influence convergence speed and learning trajectories. These results provide a rigorous and interpretable foundation for understanding how sequential structure in data can be beneficial for learning with attention-based models.



Paperid:3859
Authors:Xiao Li, Zekai Zhang, Xiang Li, Siyi Chen, Zhihui Zhu, Peng Wang, Qing Qu
Abstract:
Diffusion models, though originally designed for generative tasks, have demonstrated impressive self-supervised representation learning capabilities. A particularly intriguing phenomenon in these models is the emergence of unimodal representation dynamics, where the quality of learned features peaks at an intermediate noise level. In this work, we conduct a comprehensive theoretical and empirical investigation of this phenomenon. Leveraging the inherent low-dimensionality structure of image data, we theoretically demonstrate that the unimodal dynamic emerges when the diffusion model successfully captures the underlying data distribution. The unimodality arises from an interplay between denoising strength and class confidence across noise scales. Empirically, we further show that, in classification tasks, the presence of unimodal dynamics reliably indicates generalization: it emerges when the model generalizes and gradually transitions to a monotonically decreasing curve as the model begins to memorize the training data.



Paperid:3860
Authors:Yunheng Li, Jing Cheng, Shaoyong Jia, Hangyi Kuang, Shaohui Jiao, Qibin Hou, Ming-Ming Cheng
Abstract:
This paper introduces TempSamp-R1, a new reinforcement fine-tuning framework designed to improve the effectiveness of adapting multimodal large language models (MLLMs) to video temporal grounding tasks. We reveal that existing reinforcement learning methods, such as Group Relative Policy Optimization (GRPO), rely on on-policy sampling for policy updates. However, in tasks with large temporal search spaces, this strategy becomes both inefficient and limited in performance, as it often fails to identify temporally accurate solutions. To address this limitation, TempSamp-R1 leverages ground-truth annotations as off-policy supervision to provide temporally precise guidance, effectively compensating for the sparsity and misalignment in on-policy solutions. To further stabilize training and reduce variance in reward-based updates, TempSamp-R1 provides a non-linear soft advantage computation method that dynamically reshapes the reward feedback via an asymmetric transformation. By employing a hybrid Chain-of-Thought (CoT) training paradigm, TempSamp-R1 optimizes a single unified model to support both CoT and non-CoT inference modes, enabling efficient handling of queries with varying reasoning complexity. Experimental results demonstrate that TempSamp-R1 outperforms GRPO-based baselines, establishing new state-of-the-art performance on benchmark datasets:Charades-STA (R1\@0.7: 52.9\%, +2.7\%), ActivityNet Captions (R1\@0.5: 56.0\%, +5.3\%), and QVHighlights (mAP: 30.0\%, +3.0\%). Moreover, TempSamp-R1 shows robust few-shot generalization capabilities under limited data. The code will be released publicly.



Paperid:3861
Authors:Qingsong Wang, Zhengchao Wan, Misha Belkin, Yusu Wang
Abstract:
Diffusion models transform random noise into images of remarkable fidelity, yet the structure of this noise-to-image map remains largely unexplored. We investigate this relationship using patch-wise Principal Component Analysis (PCA) and empirically demonstrate that low-frequency components of the initial noise predominantly influence the compositional structure of generated images. Our analyses reveal that noise seeds inherently contain universal compositional cues, evident when identical seeds produce images with similar structural attributes across different datasets and model architectures. Leveraging these insights, we develop and theoretically justify a simple yet effective Patch PCA denoiser that extracts underlying structure from noise using only generic natural image statistics. The robustness of these structural cues is observed to persist across both pixel-space models and latent diffusion models, highlighting their fundamental nature. Finally, we introduce a zero-shot editing method that enables injecting compositional control over generated images, providing an intuitive approach to guided generation without requiring model fine-tuning or additional training.



Paperid:3862
Authors:Kaihang Pan, Yang Wu, Wendong Bu, Shen Kai, Juncheng Li, Yingting Wang, Yunfei Li, Siliang Tang, Jun Xiao, Fei Wu, ZhaoHang, Yueting Zhuang
Abstract:
Recent endeavors in Multimodal Large Language Models (MLLMs) aim to unify visual comprehension and generation. However, these two capabilities remain largely independent, as if they are two separate functions encapsulated within the same model. Consequently, visual comprehension does not enhance visual generation, and the reasoning mechanisms of LLMs have not been fully integrated to revolutionize image generation. In this paper, we propose to enable the collaborative co-evolution of visual comprehension and generation, advancing image generation into an iterative introspective process. We introduce a two-stage training approach: supervised fine-tuning teaches the MLLM with the foundational ability to generate genuine CoT for visual generation, while reinforcement learning activates its full potential via an exploration-exploitation trade-off. Ultimately, we unlock the Aha moment in visual generation, advancing MLLMs from text-to-image tasks to unified image generation. Extensive experiments demonstrate the superior performance of our model in both text-to-image generation and image editing. Anonymous Project: https://anonymous.4open.science/r/RefineDraw_Anonym-8714.



Paperid:3863
Authors:Zailong Tian, Zhuoheng Han, Houfeng Wang, Lizi Liao
Abstract:
Spreadsheets are widely used for data analysis and reporting, yet their complex structure and formula logic pose significant challenges for AI systems. We introduce Sheetpedia, a large-scale corpus of over 295,000 diverse spreadsheets (from 324,000+ workbooks) compiled from enterprise email archives and online forums. We detail a rigorous collection and preprocessing pipeline (integrating the Enron email spreadsheet archive and the Fuse web corpus, plus a new crawl of Excel forums) to standardize formats, filter languages, and remove duplicates. Sheetpedia provides extensive coverage of real formulas and annotations – addressing a gap left by prior table datasets (e.g. web tables used in TURL or Text-to-SQL in Spider) which often lack formula semantics. We present comprehensive corpus statistics, highlighting rich formula diversity and a majority (76\%+) of English content. To demonstrate the corpus’s utility, we fine-tune large language models on Sheetpedia for two novel spreadsheet understanding tasks: Natural Language to Semantic Range (NL2SR) and Natural Language to Formula (NL2Formula). Using a rejection-sampling data generation strategy, our fine-tuned models achieve up to 97.5\% accuracy on NL2SR and 71.7\% on NL2Formula – substantially outperforming baseline approaches. Sheetpedia (to be released publicly) fills a crucial need for a large, high-quality spreadsheet benchmark, enabling more effective spreadsheet intelligence and natural language interfaces for spreadsheet tools.



Paperid:3864
Authors:Zaifei Yang, Hong Chang, RuiBing Hou, Shiguang Shan, Xilin Chen
Abstract:
The molecular large language models have garnered widespread attention due to their promising potential on molecular applications. However, current molecular large language models face significant limitations in understanding molecules due to inadequate textual descriptions and suboptimal molecular representation strategies during pretraining. To address these challenges, we introduce KnowMol-100K, a large-scale dataset with 100K fine-grained molecular annotations across multiple levels, bridging the gap between molecules and textual descriptions. Additionally, we propose chemically-informative molecular representation, effectively addressing limitations in existing molecular representation strategies. Building upon these innovations, we develop KnowMol, a state-of-the-art multi-modal molecular large language model. Extensive experiments demonstrate that KnowMol achieves superior performance across molecular understanding and generation tasks.



Paperid:3865
Authors:Yan Gong, Yiren Song, Yicheng Li, Chenglin Li, Yin Zhang
Abstract:
Inspired by the in-context learning mechanism of large language models (LLMs), a new paradigm of generalizable visual prompt-based image editing is emerging. Existing single-reference methods typically focus on style or appearance adjustments and struggle with non-rigid transformations. To address these limitations, we propose leveraging source-target image pairs to extract and transfer content-aware editing intent to novel query images. To this end, we introduce RelationAdapter, a lightweight module that enables Diffusion Transformer (DiT) based models to effectively capture and apply visual transformations from minimal examples. We also introduce Relation252K, a comprehensive dataset comprising 218 diverse editing tasks, to evaluate model generalization and adaptability in visual prompt-driven scenarios. Experiments on Relation252K show that RelationAdapter significantly improves the model’s ability to understand and transfer editing intent, leading to notable gains in generation quality and overall editing performance.



Paperid:3866
Authors:Tao Yin, Shaochen Fu, Zhibin Zhang, Li Huang, Xiaohong Zhang, Yiyuan Yang, Kaixiang Yang, Meng Yan
Abstract:
One main challenge in time series anomaly detection for industrial IoT lies in the complex spatio-temporal couplings within multivariate data. However, as traditional anomaly detection methods focus on modeling spatial or temporal dependencies independently, resulting in suboptimal representation learning and limited sensitivity to anomalous dispersion in high-dimensional spaces. In this work, we conduct an empirical analysis showing that both normal and anomalous samples tend to scatter in high-dimensional space, especially anomalous samples are markedly more dispersed. We formalize this dispersion phenomenon as scattering, quantified by the mean pairwise distance among sample representations, and leverage it as an inductive signal to enhance spatio-temporal anomaly detection. Technically, we propose ScatterAD to model representation scattering across temporal and topological dimensions. ScatterAD incorporates a topological encoder for capturing graph-structured scattering and a temporal encoder for constraining over-scattering through mean squared error minimization between neighboring time steps. We introduce a contrastive fusion mechanism to ensure the complementarity of the learned temporal and topological representations. Additionally, we theoretically show that maximizing the conditional mutual information between temporal and topological views improves cross-view consistency and enhances more discriminative representations. Extensive experiments on multiple public benchmarks show that ScatterAD achieves state-of-the-art performance on multivariate time series anomaly detection.



Paperid:3867
Authors:William Zhi, Ziyong Ma, Tianyi Zhang, Matthew Johnson-Roberson
Abstract:
Autonomous robots typically need to construct representations of their surroundings and adapt their motions to the geometry of their environment. Here, we tackle the problem of constructing a policy model for collision-free motion generation, consistent with the environment, from a single input RGB image. Extracting 3D structures from a single image often involves monocular depth estimation. Developments in depth estimation have given rise to large pre-trained models such as \emph{DepthAnything}. However, using outputs of these models for downstream motion generation is challenging due to frustum-shaped errors that arise. Instead, we propose a framework known as Video-Generation Environment Representation (VGER), which leverages the advances of large-scale video generation models to generate a moving camera video conditioned on the input image. Frames of this video, which form a multiview dataset, are then input into a pre-trained 3D foundation model to produce a dense point cloud. We then introduce a multi-scale noise approach to train an implicit representation of the environment structure and build a motion generation model that complies with the geometry of the representation. We extensively evaluate VGER over a diverse set of indoor and outdoor environments. We demonstrate its ability to produce smooth motions that account for the captured geometry of a scene, all from a single RGB input image.



Paperid:3868
Authors:Xiao Han, Zhao Zimo, Wanyu Wang, Maolin Wang, Zitao Liu, Yi Chang, Xiangyu Zhao
Abstract:
Recent advancements in Large Language Models (LLMs) have emphasized the critical role of fine-tuning (FT) techniques in adapting LLMs to specific tasks, especially in privacy-sensitive scenarios.Users can leverage this technology by incorporating their personal data to tailor a unique LLM that perfectly aligns with their needs.However, conventional FT approaches often suffer from catastrophic forgetting and suboptimal data efficiency, limiting their real-world applicability. To address these challenges, this paper proposes DEAL, a novel framework that integrates Low-Rank Adaptation (LoRA) with a continuous fine-tuning strategy. By incorporating knowledge retention and adaptive parameter update modules, the framework mitigates the limitations of existing FT methods while maintaining efficiency in privacy-preserving settings.Experiments on \textbf{15} diverse datasets show that DEAL consistently outperforms baseline methods, yielding substantial gains in task accuracy and resource efficiency. These results underscore the potential of the proposed approach to improve model adaptability while safeguarding data privacy. The source code is publicly available at https://anonymous.4open.science/r/DEAL-Continuous-Low-Rank-Fine-Tuning-058A



Paperid:3869
Authors:Yuyang Hu, Kangfu Mei, Mojtaba Ardakani, Ulugbek Kamilov, Peyman Milanfar, Mauricio Delbracio
Abstract:
Abstract:Diffusion models show promise for image restoration, but existing methods often struggle with inconsistent fidelity and undesirable artifacts. To address this, we introduce Kernel Density Steering (KDS), a novel inference-time framework promoting robust, high-fidelity outputs through explicit local mode-seeking. KDS employs an $N$-particle ensemble of diffusion samples, computing patch-wise kernel density estimation gradients from their collective outputs. These gradients steer patches in each particle towards shared, higher-density regions identified within the ensemble. This collective local mode-seeking mechanism, acting as "collective wisdom", steers samples away from spurious modes prone to artifacts, arising from independent sampling or model imperfections, and towards more robust, high-fidelity structures. This allows us to obtain better quality samples at the expense of higher compute by simultaneously sampling multiple particles. As a plug-and-play framework, KDS requires no retraining or external verifiers, seamlessly integrating with various diffusion samplers. Extensive numerical validations demonstrate KDS substantially improves both quantitative and qualitative performance on challenging real-world super-resolution and image inpainting tasks.



Paperid:3870
Authors:Zhongyi Yu, Jianqiu Wu, Zhenghao Wu, Shuhan Zhong, Weifeng Su, Chul-Ho Lee, Weipeng Zhuo
Abstract:
Temporal graph link prediction aims to predict future interactions between nodes in a graph based on their historical interactions, which are encoded in node embeddings. We observe that heterogeneity naturally appears in temporal interactions, e.g., a few node pairs can make most interaction events, and interaction events happen at varying intervals. This leads to the problems of ineffective temporal information encoding and forgetting of past interactions for a pair of nodes that interact intermittently for their link prediction. Existing methods, however, do not consider such heterogeneity in their learning process, and thus their learned temporal node embeddings are less effective, especially when predicting the links for infrequently interacting node pairs. To cope with the heterogeneity, we propose a novel framework called TAMI, which contains two effective components, namely log time encoding function (LTE) and link history aggregation (LHA). LTE better encodes the temporal information through transforming interaction intervals into more balanced ones, and LHA prevents the historical interactions for each target node pair from being forgotten. State-of-the-art temporal graph neural networks can be seamlessly and readily integrated into TAMI to improve their effectiveness. Experiment results on 13 classic datasets and three newest temporal graph benchmark (TGB) datasets show that TAMI consistently improves the link prediction performance of the underlying models in both transductive and inductive settings.



Paperid:3871
Authors:Federico Vasile, Ri-Zhao Qiu, Lorenzo Natale, Xiaolong Wang
Abstract:
System identification involving the geometry, appearance, and physical properties from video observations is a challenging task with applications in robotics and graphics. Recent approaches have relied on fully differentiable Material Point Method (MPM) and rendering for simultaneous optimization of these properties. However, they are limited to simplified object-environment interactions with planar colliders and fail in more challenging scenarios where objects collide with non-planar surfaces. We propose AS-DiffMPM, a differentiable MPM framework that enables physical property estimation with arbitrarily shaped colliders. Our approach extends existing methods by incorporating a differentiable collision handling mechanism, allowing the target object to interact with complex rigid bodies while maintaining end-to-end optimization. We show AS-DiffMPM can be easily interfaced with various novel view synthesis methods as a framework for system identification from visual observations. The code will be open-sourced.



Paperid:3872
Authors:Zijie Geng, Jie Wang, Ziqi Liu, Feng Ju, Yiming Li, Xing Li, Mingxuan Yuan, Jianye Hao, Defu Lian, Enhong Chen, Feng Wu
Abstract:
Abstract:Key-Value (KV) cache eviction---which retains the KV pairs of the most important tokens while discarding less important ones---is a critical technique for optimizing both memory usage and inference latency in large language models (LLMs).However, existing approaches often rely on simple heuristics---such as attention weights---to measure token importance, overlooking the spatial relationships between token value states in the vector space.This often leads to suboptimal token selections and thus performance degradation.To tackle this problem, we propose a novel method, namely **AnDPro** (**An**chor **D**irection **Pro**jection), which introduces a projection-based scoring function to more accurately measure token importance.Specifically, AnDPro operates in the space of value vectors and leverages the projections of these vectors onto an *``Anchor Direction''*---the direction of the pre-eviction output---to measure token importance and guide more accurate token selection.Experiments on $16$ datasets from the LongBench benchmark demonstrate that \ours{} can maintain $96.07\\%$ of the full cache accuracy using only $3.44\\%$ KV cache budget, reducing KV cache budget size by $46.0\\%$ without compromising quality compared to previous state-of-the-arts.



Paperid:3873
Authors:Xinbin Yuan, Jian Zhang, Kaixin Li, Zhuoxuan Cai, Jie Chen, Lujian Yao, Enguang Wang, Qibin Hou, Jinwei Chen, Peng-Tao Jiang, Bo Li
Abstract:
Graphical User Interface (GUI) agents have made substantial strides in understanding and executing user instructions across diverse platforms. Yet, grounding these instructions to precise interface elements remains challenging—especially in complex, high-resolution, professional environments. Traditional supervised fine-tuning (SFT) methods often require large volumes of diverse data and exhibit weak generalization. To overcome these limitations, we introduce a reinforcement learning (RL)-based framework that incorporates three core strategies: (1) seed data curation to ensure high-quality training samples, (2) a dense policy gradient that provides continuous feedback based on prediction accuracy, and (3) a self-evolutionary reinforcement finetuning mechanism that iteratively refines the model using attention maps. With only 3k training samples, our 7B-parameter model achieves state-of-the-art results among similarly sized models on three grounding benchmarks. Notably, it attains 47.3\% accuracy on the ScreenSpot-Pro dataset—outperforming much larger models, such as UI-TARS-72B, by a margin of 24.2\%. These findings underscore the effectiveness of RL-based approaches in enhancing GUI agent performance, particularly in high-resolution, complex environments.



Paperid:3874
Authors:Shiyue Wang, Ziao Guo, Changhong Lu, Junchi Yan
Abstract:
Abstract:Langevin Dynamics (LD) and its discrete proposal have been widely applied in the field of Combinatorial Optimization (CO). Both sampling-based and data-driven approaches have benefited significantly from these methods. However, LD's reliance on Gaussian noise limits its ability to escape narrow local optima, requires costly parallel chains, and performs poorly in rugged landscapes or with non-strict constraints. These challenges have impeded the development of more advanced approaches. To address these issues, we introduce Fractional Langevin Dynamics (FLD) for CO, replacing Gaussian noise with $\alpha$-stable L\'evy noise. FLD can escape from local optima more readily via L\'evy flights, and in multiple-peak CO problems with high potential barriers it exhibits a polynomial escape time that outperforms the exponential escape time of LD. Moreover, FLD coincides with LD when $\alpha = 2$, and by tuning $\alpha$ it can be adapted to a wider range of complex scenarios in the CO fields. We provide theoretical proof that our method offers enhanced exploration capabilities and improved convergence. Experimental results on the Maximum Independent Set, Maximum Clique, and Maximum Cut problems demonstrate that incorporating FLD advances both sampling-based and data-driven approaches, achieving state-of-the-art (SOTA) performance in most of the experiments.



Paperid:3875
Authors:Haoyu Wang, Peihao Wang, Mufei Li, Shikun Liu, Siqi Miao, Zhangyang "Atlas" Wang, Pan Li
Abstract:
Modern large language models (LLMs) are inherently auto-regressive, requiring input to be serialized into flat sequences regardless of their structural dependencies. This serialization hinders the model’s ability to leverage structural inductive biases, especially in tasks such as retrieval-augmented generation (RAG) and reasoning on data with native graph structures, where inter-segment dependencies are crucial. We introduce Graph-KV with the potential to overcome this limitation. Graph-KV leverages the KV-cache of text segments as condensed representations and governs their interaction through structural inductive biases. In this framework, ''target'' segments selectively attend only to the KV-caches of their designated ''source'' segments, rather than all preceding segments in a serialized sequence. This approach induces a graph-structured block mask, sparsifying attention and enabling a message-passing-like step within the LLM. Furthermore, strategically allocated positional encodings for source and target segments reduce positional bias and context window consumption. We evaluate Graph-KV across three scenarios: (1) seven RAG benchmarks spanning direct inference, multi-hop reasoning, and long-document understanding; (2) Arxiv-QA, a novel academic paper QA task with full-text scientific papers structured as citation ego-graphs; and (3) paper topic classification within a citation network.By effectively reducing positional bias and harnessing structural inductive biases, Graph-KV substantially outperforms baselines, including standard costly sequential encoding, across various settings.



Paperid:3876
Authors:Hoang Nguyen, Priya Donti
Abstract:
Efficiently solving constrained optimization problems is crucial for numerous real-world applications, yet traditional solvers are often computationally prohibitive for real-time use. Machine learning-based approaches have emerged as a promising alternative to provide approximate solutions at faster speeds, but they struggle to strictly enforce constraints, leading to infeasible solutions in practice. To address this, we propose the Feasibility-Seeking-Integrated Neural Network (FSNet), which integrates a feasibility-seeking step directly into its solution procedure to ensure constraint satisfaction. This feasibility-seeking step solves an unconstrained optimization problem that minimizes constraint violations in a differentiable manner, enabling end-to-end training and providing guarantees on feasibility and convergence. Our experiments across a range of different optimization problems, including both smooth/nonsmooth and convex/nonconvex problems, demonstrate that FSNet can provide feasible solutions with solution quality comparable to (or in some cases better than) traditional solvers, at significantly faster speeds.



Paperid:3877
Authors:Siting Li, Xiang Gao, Simon Du
Abstract:
While an image is worth more than a thousand words, only a few provide crucial information for a given task and thus should be focused on. In light of this, ideal text-to-image (T2I) retrievers should prioritize specific visual attributes relevant to queries. To evaluate current retrievers on handling attribute-focused queries, we build COCO-Facet, a COCO-based benchmark with 9,112 queries about diverse attributes of interest. We find that CLIP-like retrievers, which are widely adopted due to their efficiency and zero-shot ability, have poor and imbalanced performance, possibly because their image embeddings focus on global semantics and subjects while leaving out other details. Notably, we reveal that even recent Multimodal Large Language Model (MLLM)-based, stronger retrievers with a larger output dimension struggle with this limitation. Hence, we hypothesize that retrieving withgeneralimage embeddings is suboptimal for performing such queries. As a solution, we propose to usepromptableimage embeddings enabled by these multimodal retrievers, which boost performance by highlighting required attributes. Our pipeline for deriving such embeddings generalizes across query types, image pools, and base retriever architectures. To enhance real-world applicability, we offer two acceleration strategies: Pre-processing promptable embeddings and using linear approximations. We show that the former yields a 15% improvement in Recall@5 when prompts are predefined, while the latter achieves an 8% improvement when prompts are only available during inference.



Paperid:3878
Authors:Shangbin Feng, Zifeng Wang, Palash Goyal, Yike Wang, Weijia Shi, Huang Xia, Hamid Palangi, Luke Zettlemoyer, Yulia Tsvetkov, Chen-Yu Lee, Tomas Pfister
Abstract:
We propose Heterogeneous Swarms, an algorithm to design multi-LLM systems by jointly optimizing model roles and weights. We represent multi-LLM systems as directed acyclic graphs (DAGs) of LLMs with topological message passing for collaborative generation. Given a pool of LLM experts and a utility function, Heterogeneous Swarms employs two iterative steps: role-step and weight-step. For role-step, we interpret model roles as learning a DAG that specifies the flow of inputs and outputs between LLMs. Starting from a swarm of random continuous adjacency matrices, we decode them into discrete DAGs, call the LLMs in topological order, evaluate on the utility function (e.g. accuracy on a task), and optimize the adjacency matrices with particle swarm optimization based on the utility score. For weight-step, we assess the contribution of individual LLMs in the multi-LLM systems and optimize model weights with swarm intelligence. We propose JFK-score to quantify the individual contribution of each LLM in the best-found DAG of the role-step, then optimize model weights with particle swarm optimization based on the JFK-score. Experiments demonstrate that Heterogeneous Swarms outperforms 17 role- and/or weight-based baselines by 18.5% on average across 12 tasks. Further analysis reveals that Heterogeneous Swarms discovers multi-LLM systems with heterogeneous model roles and substantial collaborative gains, and benefits from the diversity of language models.



Paperid:3879
Authors:Wenzhen Yue, Yong Liu, Haoxuan Li, Hao Wang, Xianghua Ying, Ruohao Guo, Bowei Xing, Ji Shi
Abstract:
Abstract:This paper presents $\mathbf{OLinear}$, a $\mathbf{linear}$-based multivariate time series forecasting model that operates in an $\mathbf{o}$rthogonally transformed domain. Recent forecasting models typically adopt the temporal forecast (TF) paradigm, which directly encode and decode time series in the time domain. However, the entangled step-wise dependencies in series data can hinder the performance of TF. To address this, some forecasters conduct encoding and decoding in the transformed domain using fixed, dataset-independent bases (e.g., sine and cosine signals in the Fourier transform). In contrast, we propose $\mathbf{OrthoTrans}$, a data-adaptive transformation based on an orthogonal matrix that diagonalizes the series' temporal Pearson correlation matrix. This approach enables more effective encoding and decoding in the decorrelated feature domain and can serve as a plug-in module to enhance existing forecasters. To enhance the representation learning for multivariate time series, we introduce a customized linear layer, $\mathbf{NormLin}$, which employs a normalized weight matrix to capture multivariate dependencies. Empirically, the NormLin module shows a surprising performance advantage over multi-head self-attention, while requiring nearly half the FLOPs. Extensive experiments on 24 benchmarks and 140 forecasting tasks demonstrate that OLinear consistently achieves state-of-the-art performance with high efficiency. Notably, as a plug-in replacement for self-attention, the NormLin module consistently enhances Transformer-based forecasters. The code and datasets are available at https://anonymous.4open.science/r/OLinear.



Paperid:3880
Authors:Peng Chen, Hailiang Zhao, Jiaji Zhang, Xueyan Tang, Yixuan Wang, Shuiguang Deng
Abstract:
Abstract:The online caching problem aims to minimize cache misses when serving a sequence of requests under a limited cache size. While naive learning-augmented caching algorithms achieve ideal $1$-consistency, they lack robustness guarantees. Existing robustification methods either sacrifice $1$-consistency or introduce significant computational overhead. In this paper, we introduce Guard, a lightweight robustification framework that enhances the robustness of a broad class of learning-augmented caching algorithms to $2H_k + 2$, while preserving their $1$-consistency. Guard achieves the current best-known trade-off between consistency and robustness, with only $\mathcal{O}(1)$ additional per-request overhead, thereby maintaining the original time complexity of the base algorithm. Extensive experiments across multiple real-world datasets and prediction models validate the effectiveness of Guard in practice.



Paperid:3881
Authors:Xingjian Ran, Yixuan Li, Linning Xu, Mulin Yu, Bo Dai
Abstract:
Realistic 3D indoor scene synthesis is crucial for Embodied AI and digital content creation. However, achieving high fidelity, strong generalization and precise controllability remains challenging due to complex semantic and physical constraints. Existing methods follow two paradigms: (1) Training models on layout datasets to directly generate numerical 3D layouts, which often generalize poorly to unseen room types; (2) Using LLMs/VLMs to produce open-vocabulary intermediate representations (e.g., scene graphs) followed by constraint-based optimization, improving plausibility but sacrificing flexibility due to predefined rules. Both approaches struggle to adapt to fine-grained user requirements. We introduce DirectLayout, a framework that directly generates numerical 3D layouts from text descriptions, without relying on intermediate representations and constrained optimization. DirectLayout decomposes the generation into three stages: producing a Bird's-Eye View (BEV) layout, lifting it into 3D space, and refining object placements for plausibility. To enable explicit spatial reasoning and help the model grasp basic principles of object placement, we employ Chain-of-Thought (CoT) activation based on the 3D-Front dataset. Additionally, we design CoT-Grounded Generative Layout Reward to enhance generalization and spatial planning. During inference, DirectLayout addresses asset-layout mismatches via Iterative Asset-Layout Alignment through in-context learning. Extensive experiments demonstrate that DirectLayout achieves impressive semantic consistency, generalization and physical plausibility.



Paperid:3882
Authors:Xuyang Zhong, Haochen Luo, Chen Liu
Abstract:
Existing machine unlearning (MU) approaches exhibit significant sensitivity to hyperparameters, requiring meticulous tuning that limits practical deployment. In this work, we first empirically demonstrate the instability and suboptimal performance of existing popular MU methods when deployed in different scenarios. To address this issue, we propose Dual Optimizer (DualOptim), which incorporates adaptive learning rate and decoupled momentum factors. Empirical and theoretical evidence demonstrates that DualOptim contributes to effective and stable unlearning. Through extensive experiments, we show that DualOptim can significantly boost MU efficacy and stability across diverse tasks, including image classification, image generation, and large language models, making it a versatile approach to empower existing MU algorithms.



Paperid:3883
Authors:Han Lin, Jaemin Cho, Amir Zadeh, Chuan Li, Mohit Bansal
Abstract:
There is growing interest in integrating high-fidelity visual synthesis capabilities into large language models (LLMs) without compromising their strong reasoning capabilities. Existing methods that directly train LLMs or bridge LLMs and diffusion models usually suffer from costly training since the backbone LLMs have not seen image representations during pretraining. We present Bifrost-1, a unified framework that bridges pretrained multimodal LLMs (MLLMs) and diffusion models using patch-level CLIP image embeddings as latent variables, which are natively aligned with the MLLM’s CLIP visual encoder. These patch-level image embeddings are integrated into the diffusion model with a lightweight adaptation of its ControlNet. To retain the original multimodal reasoning capabilities of MLLMs, we equip the MLLM with a visual generation branch initialized from the original MLLM parameters when predicting the patch-level image embeddings. By seamlessly integrating pretrained MLLMs and diffusion models with patch-level CLIP latents, our framework enables high-fidelity controllable image generation with significant training efficiency. Our experiments demonstrate that Bifrost-1 achieves comparable or better performance than previous methods in terms of visual fidelity and multimodal understanding, with substantially lower compute during training. We also provide comprehensive ablation studies showing the effectiveness of our design choices. Code, technical details and additional experiment results are included in the supplementary materials.



Paperid:3884
Authors:Yiqi Jiang, Kaiwen Sheng, Yujia Gao, Estefany Kelly Buchanan, Yu Shikano, Seung Je Woo, Yixiu Zhao, Tony Hyun Kim, Fatih Dinc, Scott Linderman, Mark Schnitzer
Abstract:
Abstract:Recent work indicates that low-dimensional dynamics of neural and behavioral data are often preserved across days and subjects. However, extracting these preserved dynamics remains challenging: high-dimensional neural population activity and the recorded neuron populations vary across recording sessions. While existing modeling tools can improve alignment between neural and behavioral data, they often operate on a per-subject basis or discretize behavior into categories, disrupting its natural continuity and failing to capture the underlying dynamics. We introduce $\underline{\text{C}}$ontrastive $\underline{\text{A}}$ligned $\underline{\text{N}}$eural embe$\underline{\text{D}}$ding and d$\underline{\text{Y}}$namics (CANDY), an end‑to‑end framework that aligns neural and behavioral data using rank-based contrastive learning, adapted for continuous behavioral variables, to project neural activity from different sessions onto a shared low-dimensional embedding space. CANDY fits a shared linear dynamical system to the aligned embeddings, enabling an interpretable model of the conserved temporal structure in the latent space. We validate CANDY on several datasets, spanning multiple species, behaviors and recording modalities. Our results show that CANDY is able to learn aligned latent embeddings and preserved dynamics across sessions and subjects, and it achieves improved cross-session behavior decoding performance. We further show that the latent linear dynamical system generalizes to new sessions and subjects, achieving behavior decoding performance that can match and even outperform models trained from scratch on the new datasets. These advances enable robust cross‑session behavioral decoding and offer a path towards identifying shared neural dynamics that underlie behavior across individuals and recording conditions.



Paperid:3885
Authors:Weijian Mai, Jiamin Wu, Yu Zhu, Zhouheng Yao, Dongzhan Zhou, Andrew Luo, Qihao Zheng, Wanli Ouyang, Chunfeng Song
Abstract:
Deciphering how visual stimuli are transformed into cortical responses is a fundamental challenge in computational neuroscience. This visual-to-neural mapping is inherently a one-to-many relationship, as identical visual inputs reliably evoke variable hemodynamic responses across trials, contexts, and subjects. However, existing deterministic methods struggle to simultaneously model this biological variability while capturing the underlying functional consistency that encodes stimulus information. To address these limitations, we propose SynBrain, a generative framework that simulates the transformation from visual semantics to neural responses in a probabilistic and biologically interpretable manner. SynBrain introduces two key components: (i) BrainVAE models neural representations as continuous probability distributions via probabilistic learning while maintaining functional consistency through visual semantic constraints; (ii) A Sematic-to-Neural Mapper acts as a semantic transmission pathway, projecting visual semantics into the neural response manifold to facilitate high-fidelity fMRI synthesis. Experimental results demonstrate that SynBrain surpasses state-of-the-art methods in subject-specific visual-to-fMRI encoding performance. Furthermore, SynBrain adapts efficiently to new subjects with few-shot data and synthesizes high-quality fMRI signals that are effective in improving data-limited fMRI-to-image decoding performance. Beyond that, SynBrain reveals functional consistency across trials and subjects, with synthesized signals capturing interpretable patterns shaped by biological neural variability. The code will be made publicly available.



Paperid:3886
Authors:Yuping He, Yifei Huang, Guo Chen, Baoqi Pei, Jilan Xu, Jiangmiao Pang, Tong Lu
Abstract:
Transferring and integrating knowledge across first-person (egocentric) and third-person (exocentric) viewpoints is intrinsic to human intelligence, enabling humans to learn from others and convey insights from their own experiences. Despite rapid progress in multimodal large language models (MLLMs), their ability to perform such cross-view reasoning remains unexplored. To address this, we introduce EgoExoBench, the first benchmark for egocentric exocentric video understanding and reasoning. Built from publicly available datasets, EgoExoBench comprises over 7300 question–answer pairs spanning eleven sub-tasks organized into three core challenges: semantic alignment, viewpoint association, and temporal reasoning. We evaluate 13 state-of-the-art MLLMs and find that while these models excel on single-view tasks, they struggle to align semantics across perspectives, accurately associate views, and infer temporal dynamics in the ego-exo context. We hope EgoExoBench can serve as a valuable resource for research on embodied agents and intelligent assistants seeking human-like cross-view intelligence.



Paperid:3887
Authors:Chenghu Du, Shengwu Xiong, junyin Wang, Yi Rong, Shili Xiong
Abstract:
This paper investigates the occlusion problems in virtual try-on (VTON) tasks. According to how they affect the try-on results, the occlusion issues of existing VTON methods can be grouped into two categories: (1) Inherent Occlusions, which are the ghosts of the clothing from reference input images that exist in the try-on results. (2) Acquired Occlusions, where the spatial structures of the generated human body parts are disrupted and appear unreasonable. To this end, we analyze the causes of these two types of occlusions, and propose a novel mask-free VTON framework based on our analysis to deal with these occlusions effectively. In this framework, we develop two simple-yet-powerful operations: (1) The background pre-replacement operation prevents the model from confusing the target clothing information with the human body or image background, thereby mitigating inherent occlusions. (2) The covering-and-eliminating operation enhances the model's ability of understanding and modeling human semantic structures, leading to more realistic human body generation and thus reducing acquired occlusions. Moreover, our method is highly generalizable, which can be applied in in-the-wild scenarios, and our proposed operations can also be easily integrated into different generative network architectures (e.g., GANs and diffusion models) in a plug-and-play manner. Extensive experiments on three VTON datasets validate the effectiveness and generalization ability of our method. Both qualitative and quantitative results demonstrate that our method outperforms recently proposed VTON benchmarks.



Paperid:3888
Authors:Yibo Zhao, Yang Zhao, Hongru Du, Hao Frank Yang
Abstract:
Individual decision-making, particularly in high-stakes scenarios like vaccine uptake, often diverges from predictions by population-optimal. This gap arises from the uniqueness of the individual decision-making process, shaped by quantifiable attributes (e.g., cost, time) and unstructured influences (e.g., personal preferences and constraints).Building upon Utility Theory and leveraging the textual-reasoning capabilities of Large Language Models (LLMs), this paper proposes an Adaptive Textual-symbolic Human-centric Reasoning framework (ATHENA) to address this challenge.ATHENA uniquely integrates two stages: First, it discovers robust, group-level symbolic utility functions via LLM-augmented symbolic discovery; Second, it implements individual-level semantic adaptation, creating personalized semantic templates guided by the optimal utility to model personalized choices. Validated on real-world travel mode and vaccine choice tasks, ATHENA consistently outperforms utility-based, machine learning, and other LLM-based models, lifting F1 score by at least 6.5% over the strongest cutting-edge models. Further, ablation studies confirm that both stages of ATHENA are critical and complementary, as removing either significantly degrades overall predictive performance. By organically integrating symbolic utility modeling and semantic adaptation, ATHENA provides a new scheme for modeling human-centric decisions.



Paperid:3889
Authors:Rui Peng, Yuchen Lu, Qichen Sun, Yuxing Lu, Chi Zhang, Ziru Liu, Jinzhuo Wang
Abstract:
Gene regulatory network (GRN) inference serves as a cornerstone for deciphering cellular decision-making processes. Early approaches rely exclusively on gene expression data, thus their predictive power remain fundamentally constrained by the vast combinatorial space of potential gene-gene interactions. Subsequent methods integrate prior knowledge to mitigate this challenge by restricting the solution space to biologically plausible interactions. However, we argue that the effectiveness of these approaches is contingent upon the precision of prior information and the reduction in the search space will circumscribe the models' potential for novel biological discoveries. To address these limitations, we introduce KINDLE, a three-stage framework that decouples GRN inference from prior knowledge dependencies. KINDLE trains a teacher model that integrates prior knowledge with temporal gene expression dynamics and subsequently distills this encoded knowledge to a student model, enabling accurate GRN inference solely from expression data without access to any prior. KINDLE achieves state-of-the-art performance across four benchmark datasets. Notably, it successfully identifies key transcription factors governing mouse embryonic development and precisely characterizes their functional roles. In mouse hematopoietic stem cell data, KINDLE accurately predicts fate transition outcomes following knockout of two critical regulators (Gata1 and Spi1). These biological validations demonstrate our framework's dual capability in maintaining topological inference precision while preserving discovery potential for novel biological mechanisms.



Paperid:3890
Authors:Dongyang Zeng, Yaping Liu, Wei Zhang, Shuo Zhang, Xinwang Liu, Binxing Fang
Abstract:
Abstract:With the increasing adoption of Retrieval-Augmented Generation (RAG) systems for knowledge-intensive tasks, ensuring the adequacy of retrieved documents has become critically important for generation quality. Traditional reranking approaches face three significant challenges: substantial computational overhead that scales with document length, dependency on plain text that limits application in sensitive scenarios, and insufficient assessment of document value beyond simple relevance metrics. We propose EAReranker, an efficient embedding-based adequacy assessment framework that evaluates document utility for RAG systems without requiring access to original text content. The framework quantifies document adequacy through a comprehensive scoring methodology considering verifiability, coverage, completeness and structural aspects, providing interpretable adequacy classifications for downstream applications. EAReranker employs a Decoder-Only Transformer architecture that introduces embedding dimension expansion method and bin-aware weighted loss, designed specifically to predict adequacy directly from embedding vectors. Our comprehensive evaluation across four public benchmarks demonstrates that EAReranker achieves competitive performance with state-of-the-art plaintext rerankers while maintaining constant memory usage ($\sim$550MB) regardless of input length and processing 2-3x faster than traditional approaches. The semantic bin adequacy prediction accuracy of 92.85\% LACC@10 and 86.12\% LACC@25 demonstrates its capability to effectively filter out inadequate documents that could potentially mislead or adversely impact RAG system performance, thereby ensuring only high-utility information serves as generation context. These results establish EAReranker as an efficient and practical solution for enhancing RAG system performance through improved context selection while addressing the computational and privacy challenges of existing methods.



Paperid:3891
Authors:Inguk Choi, Woo-Jin Shin, Sang-Hyun Cho, Hyun-Jung Kim
Abstract:
Reinforcement Learning (RL) has shown promising results in solving Job Scheduling Problems (JSPs), automatically deriving powerful dispatching rules from data without relying on expert knowledge. However, most RL-based methods rely on single policy strategies, which significantly limit their ability to explore diverse scheduling behaviors. Moreover, designing tailored reward functions for each JSP variant remains a challenging and labor-intensive task. To address these issues, we introduce a generic learning framework that optimizes multiple policies sharing a common objective and a single neural network, yet each policy can learn specialized and complementary problem-solving strategies. This optimization process is completely derived from autonomously generated self-teaching labels, eliminating the need for manually crafted reward functions. In addition, we develop a training scheme that adaptively controls the imitation degree to reflect the quality of self-labels and enhances sample efficiency. Experimental results show that our method successfully addresses the aforementioned challenges, significantly outperforming previous state-of-the-art methods across five JSP variants. Furthermore, our approach demonstrates remarkable effectiveness on other combinatorial optimization problems, highlighting its broad applicability and generality beyond JSPs.



Paperid:3892
Authors:Yuhao QING, Boyu Zhu, Mingzhe Du, Zhijiang Guo, Terry Yue Zhuo, Qianru Zhang, Jie Zhang, Heming Cui, Siu Ming Yiu, Dong HUANG, See-Kiong Ng, Anh Tuan Luu
Abstract:
Existing code generation benchmarks primarily evaluate functional correctness, with limited focus on code efficiency and often restricted to a single language like Python. To address this gap, we introduce EffiBench-X, the first multi-language benchmark designed to measure the efficiency of LLM-generated code. EffiBench-X supports Python, C++, Java, JavaScript, Ruby, and Golang. It comprises competitive programming tasks with human-expert solutions as efficiency baselines. Evaluating state-of-the-art LLMs on EffiBench-X reveals that while models generate functionally correct code, they consistently underperform human experts in efficiency. Even the most efficient LLM-generated solutions (Qwen3-32B) achieve only around 62\% of human efficiency on average, with significant language-specific variations. LLMs show better efficiency in Python, Ruby, and JavaScript than in Java, C++, and Golang. For instance, DeepSeek-R1's Python code is significantly more efficient than its Java code. These results highlight the critical need for research into LLM optimization techniques to improve code efficiency across diverse languages. The dataset and evaluation infrastructure are submitted and available at \url{https://github.com/EffiBench/EffiBench-X.git} and \url{https://huggingface.co/datasets/EffiBench/effibench-x}.



Paperid:3893
Authors:Yue Tan, Xiaoqian Hu, Hao Xue, Celso de Melo, Flora Salim
Abstract:
Frontier vision-language models (VLMs) have made remarkable improvements in video understanding tasks. However, real-world videos typically exist as continuously evolving data streams (e.g., dynamic scenes captured by wearable glasses), necessitating models to continually adapt to shifting data distributions and novel scenarios. Considering the prohibitive computational costs of fine-tuning models on new tasks, usually, a small subset of parameters is updated while the bulk of the model remains frozen. This poses new challenges to existing continual learning frameworks in the context of large multimodal foundation models, i.e., catastrophic forgetting and update conflict. While the foundation models struggle with parameter-efficient continual learning, the hippocampus in the human brain has evolved highly efficient mechanisms for memory formation and consolidation. Inspired by the rapidBinding and patternseparation mechanisms in the hippocampus, in this work, we proposeBiseclefor video-languagecontinuallearning, where a multi-directional supervision module is used to capture more cross-modal relationships and a contrastive prompt learning scheme is designed to isolate task-specific knowledge to facilitate efficient memory storage. Binding and separation processes further strengthen the ability of VLMs to retain complex experiences, enabling robust and efficient continual learning in video understanding tasks. We perform a thorough evaluation of the proposed Bisecle, demonstrating its ability to mitigate forgetting and enhance cross-task generalization on several VideoQA benchmarks.



Paperid:3894
Authors:Wenkai Fang, Shunyu Liu, Yang Zhou, Kongcheng Zhang, Tongya Zheng, Kaixuan Chen, Mingli Song, Dacheng Tao
Abstract:
Recent advances have demonstrated the effectiveness of Reinforcement Learning (RL) in improving the reasoning capabilities of Large Language Models (LLMs). However, existing works inevitably rely on high-quality instructions and verifiable rewards for effective training, both of which are often difficult to obtain in specialized domains. In this paper, we propose Self-play Reinforcement Learning (SeRL) to bootstrap LLM training with limited initial data. Specifically, SeRL comprises two complementary modules: self-instruction and self-rewarding. The former module generates additional instructions based on the available data at each training step, employing comprehensive online filtering strategies to ensure instruction quality, diversity, and difficulty. The latter module introduces a simple yet effective majority-voting mechanism to estimate response rewards for additional instructions, eliminating the need for external annotations. Finally, SeRL performs conventional RL based on the generated data, facilitating iterative self-play learning.Extensive experiments on various reasoning benchmarks and across different LLM backbones demonstrate that the proposed SeRL yields results superior to the counterparts, and achieves performance on par with those obtained by high-quality data with verifiable rewards. Our code will be made publicly available.



Paperid:3895
Authors:Ming Hu, Zhengdi Yu, feilong tang, Kaiwen Chen, Yulong Li, Imran Razzak, Junjun He, Tolga Birdal, Kai-Jing Zhou, Zongyuan Ge
Abstract:
Accurate 3D reconstruction of hands and instruments is critical for vision-based analysis of ophthalmic microsurgery, yet progress has been hampered by the lack of realistic, large-scale datasets and reliable annotation tools. In this work, we introduce OphNet-3D, the first extensive RGB-D dynamic 3D reconstruction dataset for ophthalmic surgery, comprising 41 sequences from 40 surgeons and totaling 7.1 million frames, with fine-grained annotations of 12 surgical phases, 10 instrument categories, dense MANO hand meshes, and full 6-DoF instrument poses. To scalably produce high-fidelity labels, we design a multi-stage automatic annotation pipeline that integrates multi-view data observation, data-driven motion prior with cross-view geometric consistency and biomechanical constraints, along with a combination of collision-aware interaction constraints for instrument interactions. Building upon OphNet-3D, we establish two challenging benchmarks—bimanual hand pose estimation and hand–instrument interaction reconstruction—and propose two dedicated architectures: H-Net for dual-hand mesh recovery and OH-Net for joint reconstruction of two-hand–two-instrument interactions. These models leverage a novel spatial reasoning module with weak-perspective camera modeling and collision-aware center-based representation. Both architectures outperform existing methods by substantial margins, achieving improvements of over 2mm in Mean Per Joint Position Error (MPJPE) and up to 23\% in ADD-S metrics for hand and instrument reconstruction, respectively.



Paperid:3896
Authors:Zhijian Zhuo, Yutao Zeng, Ya Wang, Sijun Zhang, Xiaoqing Li, Jian Yang, zhou Xun, Jinwen Ma
Abstract:
Transformers have become the de facto architecture for a wide range of machine learning tasks, particularly in large language models (LLMs). Despite their remarkable performance, challenges remain in training deep transformer networks, especially regarding the position of layer normalization. While Pre-Norm structures facilitate more stable training owing to their stronger identity path, they often lead to suboptimal performance compared to Post-Norm. In this paper, we proposeHybridNorm, a simple yet effective hybrid normalization strategy that integrates the advantages of both Pre-Norm and Post-Norm. Specifically, HybridNorm employs QKV normalization within the attention mechanism and Post-Norm in the feed-forward network (FFN) of each transformer block. We provide both theoretical insights and empirical evidence demonstrating that HybridNorm improves gradient flow and model robustness. Extensive experiments on large-scale transformer models, including both dense and sparse variants, show that HybridNorm consistently outperforms both Pre-Norm and Post-Norm approaches across multiple benchmarks. These findings highlight the potential of HybridNorm as a more stable and effective technique for improving the training and performance of deep transformer models.



Paperid:3897
Authors:Yue Gong, Raul Fernandez
Abstract:
As hypothesis generation becomes increasingly automated, a new bottleneck has emerged: hypothesis assessment. Modern systems can surface thousands of statistical relationships—correlations, trends, causal links—but offer little guidance on which ones are novel, non-trivial, or worthy of expert attention. In this work, we study the complementary problem to hypothesis generation: automatic hypothesis assessment. Specifically, we ask—given a large set of statistical relationships, can we automatically assess which ones are novel and worth further exploration? We focus on correlations as they are a common entry point in exploratory data analysis that often serve as the basis for forming deeper scientific or causal hypotheses.To support automatic assessment, we propose to leverage the vast knowledge encoded in LLMs' weights to derive a prior distribution over the correlation value of a variable pair. If an LLM's prior expects the correlation value observed, then such correlation is not surprising, and vice versa. We propose the Logit-based Calibrated Prior, an LLM-elicited correlation prior that transforms the model’s raw output logits into a calibrated, continuous predictive distribution over correlation values. We evaluate the prior on a benchmark of 2,096 real-world variable pairs and it achieves a sign accuracy of 78.8%, a mean absolute error of 0.26, and 95% credible interval coverage of 89.2% in predicting Pearson correlation coefficient. It also outperforms a fine-tuned RoBERTa classifier in binary correlation prediction and achieves higher precision@K in hypothesis ranking. We further show that the prior generalizes to correlations not seen during LLM pretraining, reflecting context-sensitive reasoning rather than memorization.



Paperid:3898
Authors:Jian Xiao, Zijie Song, Jialong Hu, Hao Cheng, Jia Li, Zhenzhen Hu, Richang Hong
Abstract:
Abstract:Recent advances in text-video retrieval have been largely driven by contrastive learning frameworks. However, existing methods overlook a key source of optimization tension: the separation between text and video distributions in the representation space—referred to as the modality gap—and the prevalence of false negatives in batch sampling. These factors lead to conflicting gradients under the InfoNCE loss, impeding stable alignment. To mitigate this, we propose GARE—a Gap-Aware Retrieval framework that introduces a learnable, pair-specific increment $\Delta_{ij}$ between text $t_i$ and video $v_j$ to offload the tension from the global anchor representation. We first derive the ideal form of $\Delta_{ij}$ via a coupling multivariate first-order Taylor approximation of the InfoNCE loss under a trust-region constraint, revealing it as a key mechanism for resolving gradient conflicts by guiding updates along a locally optimal descent direction in the coupled optimization landscape. Due to the expensive cost of directly approximate $\Delta_{ij}$, we introduce a lightweight neural module conditioned on the semantic gap between each video-text pair, enabling structure-aware correction guided by gradient supervision. To further stabilize learning and promote interpretability, we regularize $\Delta$ via three components: a trust-region constraint regularization to prevent oscillations, a directional diversity term to expand the semantic difference space, and an information bottleneck over $\Delta$ to restrict redundant information. Experiments across four retrieval benchmarks show that GARE consistently improves alignment accuracy and robustness to noisy supervision, confirming the effectiveness of gap-aware tension unloading.



Paperid:3899
Authors:Yuhan Li, Zhiyu Jin, Yifan Tong, Wenxiang Shang, Benlei Cui, Xuanhong Chen, Ran Lin, Bingbing Ni
Abstract:
Virtual footwear try-on (VFTON), a critical yet underexplored area in virtual try-on (VTON), aims to synthesize faithful try-on results given diverse footwear and model images while maintaining 3D consistency and texture authenticity. Unlike conventional garment-focused VTON methods, VFTON presents unique challenges due to (1) Data Scarcity, which arises from the difficulty of perfectly matching product shoes with models wearing the identical ones, (2) Viewpoint Misalignment, where the target foot pose and source shoe views are always misaligned, leading to incomplete texture information and detail distortion, and (3) Background-induced Color Distortion, where complex material of footwear interacts with environmental lighting, causing unintended color contamination. To address these challenges, we introduce MVShoes, a multi-view shoe try-on dataset consisting of 7305 well-annotated image triplets, covering diverse footwear categories and challenging try-on scenarios. Furthermore, we propose a dual-stream DiT architecture, ShoeFit, designed to mitigate viewpoint misalignment through Multi-View Conditioning with 3D Rotary Position Embedding, and alleviate background-induced distortion using the LayeredRefAttention which leverages background features to modulate footwear latents. The proposed framework effectively decouples shoe appearance from environmental interferences while preserving high-quality texture detail through decoupled denoising and conditioning branches. Extensive quantitative and qualitative experiments demonstrate that our method substantially improves rendering fidelity and robustness under challenging real-world product shoes, establishing a new benchmark in high-fidelity footwear try-on synthesis. The dataset and benchmark will be publicly available upon acceptance of the paper.



Paperid:3900
Authors:Anindya Sarkar, Binglin Ji, Yevgeniy Vorobeychik
Abstract:
In various scientific and engineering domains, where data acquisition is costly—such as in medical imaging, environmental monitoring, or remote sensing—strategic sampling from unobserved regions, guided by prior observations, is essential to maximize target discovery within a limited sampling budget. In this work, we introduce Diffusion-guided Active Target Discovery (DiffATD), a novel method that leverages diffusion dynamics for active target discovery. DiffATD maintains a belief distribution over each unobserved state in the environment, using this distribution to dynamically balance exploration-exploitation. Exploration reduces uncertainty by sampling regions with the highest expected entropy, while exploitation targets areas with the highest likelihood of discovering the target, indicated by the belief distribution and an incrementally trained reward model designed to learn the characteristics of the target. DiffATD enables efficient target discovery in a partially observable environment within a fixed sampling budget, all without relying on any prior supervised training. Furthermore, DiffATD offers interpretability, unlike existing black-box policies that require extensive supervised training. Through extensive experiments and ablation studies across diverse domains, including medical imaging, species discovery and remote sensing, we show that DiffATD performs significantly better than baselines and competitively with supervised methods that operate under full environmental observability.



Paperid:3901
Authors:Jongyeong Lee, Junya Honda, Shinji Ito, Min-hwan Oh
Abstract:
Abstract:Follow-the-Regularized-Leader (FTRL) policies have achieved Best-of-Both-Worlds (BOBW) results in various settings through hybrid regularizers, whereas analogous results for Follow-the-Perturbed-Leader (FTPL) remain limited due to inherent analytical challenges. To advance the analytical foundations of FTPL, we revisit classical FTRL-FTPL duality for unbounded perturbations and establish BOBW results for FTPL under a broad family of asymmetric unbounded Fréchet-type perturbations, including hybrid perturbations combining Gumbel-type and Fréchet-type tails. These results not only extend the BOBW results of FTPL but also offer new insights into designing alternative FTPL policies competitive with hybrid regularization approaches. Motivated by earlier observations in two-armed bandits, we further investigate the connection between the $1/2$-Tsallis entropy and a Fréchet-type perturbation. Our numerical observations suggest that it corresponds to a symmetric Fréchet-type perturbation, and based on this, we establish the first BOBW guarantee for symmetric unbounded perturbations in the two-armed setting. In contrast, in general multi-armed bandits, we find an instance in which symmetric Fréchet-type perturbations violate the standard condition for BOBW analysis, which is a problem not observed with asymmetric or nonnegative Fréchet-type perturbations. Although this example does not rule out alternative analyses achieving BOBW results, it suggests the limitations of directly applying the relationship observed in two-armed cases to the general case and thus emphasizes the need for further investigation to fully understand the behavior of FTPL in broader settings.



Paperid:3902
Authors:Jiaqi Zhang, Juntuo Wang, Zhixin Sun, John Zou, Randall Balestriero
Abstract:
Large-scale vision foundation models such as DINOv2 boast impressive performances by leveraging massive architectures and training datasets. The expense of large-scale pre-training puts such research out of reach for many, hence limiting scientific advancements. We thus propose a novel pretraining strategy for DINOv2 that simultaneously accelerates convergence–and strengthens robustness to common corruptions as a by-product. Our approach involves a frequency filtering curriculum–low-frequency being seen first–and the Gaussian noise patching augmentation. Applied to a ViT-B/16 backbone trained on ImageNet-1K, while pre-training time is reduced by 1.6×–from 16.64 to 10.32 l40s days–and FLOPs by 2.25×, our method still achieves matching robustness in corruption benchmarks (ImageNet-C) and maintains competitive linear probing performance compared with the DINOv2 baseline. This dual benefit of efficiency and robustness makes large-scale self-supervised foundation modeling more attainable, while opening the door to novel exploration around data curriculum and augmentation as a means to improve self-supervised learning models robustness.



Paperid:3903
Authors:Jingyuan Zhou, Hao Qian, Shikui Tu, Lei Xu
Abstract:
Structure-based drug design (SBDD), aiming to generate 3D molecules with high binding affinity toward target proteins, is a vital approach in novel drug discovery. Although recent generative models have shown great potential, they suffer from unstable probability dynamics and mismatch between generated molecule size and the protein pockets geometry, resulting in inconsistent quality and off-target effects. We propose PAFlow, a novel target-aware molecular generation model featuring prior interaction guidance and a learnable atom number predictor. PAFlow adopts the efficient flow matching framework to model the generation process and constructs a new form of conditional flow matching for discrete atom types. A protein–ligand interaction predictor is incorporated to guide the vector field toward higher-affinity regions during generation, while an atom number predictor based on protein pocket information is designed to better align generated molecule size with target geometry. Extensive experiments on the CrossDocked2020 benchmark show that PAFlow achieves a new state-of-the-art in binding affinity (up to -8.31 Avg. Vina Score), simultaneously maintains favorable molecular properties.



Paperid:3904
Authors:Belinda Li, Been Kim, Zi Wang
Abstract:
Recently, large language models (LLMs) have shown impressive performance on reasoning benchmarks like math and logic. However, past work has largely assumed well-defined tasks, while real-world queries are often underspecified and only solvable by acquiring missing information. We formalize this as a constraint satisfaction problem (CSP) with missing variable assignments. Using a special case where only one necessary variable assignment is missing, we can rigorously evaluate an LLM’s ability to identify the minimal necessary question to ask. We present QUESTBENCH, a set of underspecified reasoning tasks solvable by asking at most one question, which includes: (1) Logic-Q: logical reasoning tasks with one missing proposition, (2) Planning-Q: PDDL planning problems with partially-observed initial states, (3) GSM-Q: human-annotated grade school math problems with one missing variable, and (4) GSME-Q: equation-based version of GSM-Q. The LLM must select the correct clarification question from multiple options. While current models excel at GSM-Q and GSME-Q, they achieve only 40–50% accuracy on Logic-Q and Planning-Q. Analysis shows that the ability to solve well-specified reasoning problems is not sufficient for success on our benchmark: models struggle to identify the right question even when they can solve the fully specified version. This highlights the need for specifically optimizing models’ information acquisition capabilities.



Paperid:3905
Authors:Yifei Liu, Li Lyna Zhang, Yi Zhu, Bingcheng Dong, Xudong Zhou, Ning Shang, Fan Yang, Cheng Li, Mao Yang
Abstract:
Advancing code reasoning in large language models (LLMs) is fundamentally limited by the scarcity of high-difficulty datasets, especially those with verifiable input-output test cases necessary for rigorous solution validation at scale. We introduce rStar-Coder, which significantly improves LLM code reasoning capabilities by constructing a large-scale, verified dataset of 418K competition-level code problems, 580K long-reasoning solutions along with rich test cases of varying difficulty. This is achieved through three core contributions: (1) we curate competitive programming code problems and oracle solutions to synthesize new, solvable problems; (2) we introduce a reliable input-output test case synthesis pipeline that decouples the generation into a three-step input generation method and a mutual verification mechanism for effective output labeling; (3) we augment problems with high-quality, test-case-verified long-reasoning solutions. Extensive experiments on Qwen models (1.5B-14B) across various code reasoning benchmarks demonstrate the superiority of rStar-Coder dataset, achieving leading performance comparable to frontier reasoning LLMs with much smaller model sizes. On LiveCodeBench, rStar-Coder improves Qwen2.5-7B from 17.4% to an impressive 57.3%, and Qwen2.5-14B from 23.3% to 62.5%, surpassing o3-mini (low) by 3.1%. On the more challenging USA Computing Olympiad, our 7B model achieves an average pass@1 accuracy of 16.15%, outperforming the frontier-level QWQ-32B. We will release rStar-Coder dataset to advance open research in code reasoning.



Paperid:3906
Authors:Ioannis Anagnostides, Ioannis Panageas, Tuomas Sandholm, Jingming Yan
Abstract:
Abstract:We consider the problem of computing stationary points in min-max optimization, with a focus on the special case of Nash equilibria in (two-)team zero-sum games. We first show that computing $\epsilon$-Nash equilibria in $3$-player $\text{\emph{adversarial}}$ team games---wherein a team of $2$ players competes against a $\text{\emph{single}}$ adversary---is $\textsf{CLS}$-complete, resolving the complexity of Nash equilibria in such settings. Our proof proceeds by reducing from $\text{\emph{symmetric}}$ $\epsilon$-Nash equilibria in $\text{\emph{symmetric}}$, identical-payoff, two-player games, by suitably leveraging the adversarial player so as to enforce symmetry---without disturbing the structure of the game. In particular, the class of instances we construct comprises solely polymatrix games, thereby also settling a question left open by Hollender, Maystre, and Nagarajan (2024). Moreover, we establish that computing $\text{\emph{symmetric}}$ (first-order) equilibria in $\text{\emph{symmetric}}$ min-max optimization is $\textsf{PPAD}$-complete, even for quadratic functions. Building on this reduction, we show that computing symmetric $\epsilon$-Nash equilibria in symmetric, $6$-player ($3$ vs. $3$) team zero-sum games is also $\textsf{PPAD}$-complete, even for $\epsilon = \text{poly}(1/n)$. As a corollary, this precludes the existence of symmetric dynamics---which includes many of the algorithms considered in the literature---converging to stationary points. Finally, we prove that computing a $\text{\emph{non-symmetric}}$ $\text{poly}(1/n)$-equilibrium in symmetric min-max optimization is $\textsf{FNP}$-hard.



Paperid:3907
Authors:Haoran Li, Yingjie Qin, Baoyuan Ou, Lai Xu, Ruiwen Xu
Abstract:
Vision-Language Models (VLMs) have made significant progress in multimodal tasks. However, their performance often deteriorates in long-context scenarios, particularly long videos. While Rotary Position Embedding (RoPE) has been widely adopted for length generalization in Large Language Models (LLMs), extending vanilla RoPE to capture the intricate spatial-temporal dependencies in videos remains an unsolved challenge. Existing methods typically allocate different frequencies within RoPE to encode 3D positional information. However, these allocation strategies mainly rely on heuristics, lacking in-depth theoretical analysis. In this paper, we first study how different allocation strategies impact the long-context capabilities of VLMs. Our analysis reveals that current multimodal RoPEs fail to reliably capture semantic similarities over extended contexts. To address this issue, we propose HoPE, a Hybrid of Position Embedding designed to improve the long-context capabilities of VLMs. HoPE introduces a hybrid frequency allocation strategy for reliable semantic modeling over arbitrarily long context, and a dynamic temporal scaling mechanism to facilitate robust learning and flexible inference across diverse context lengths. Extensive experiments across four video benchmarks on long video understanding and retrieval tasks demonstrate that HoPE consistently outperforms existing methods, confirming its effectiveness.



Paperid:3908
Authors:Michael Green, Matan Levy, Issar Tzachor, Dvir Samuel, Nir Darshan, Rami Ben-Ari
Abstract:
We address the challenge of Small Object Image Retrieval (SoIR), where the goal is to retrieve images containing a specific small object, in a cluttered scene. The key challenge in this setting is constructing a single image descriptor, for scalable and efficient search, that effectively represents all objects in the image. In this paper, we first analyze the limitations of existing methods on this challenging task and then introduce new benchmarks to support SoIR evaluation. Next, we introduce Multi-object Attention Optimization (MaO), a novel retrieval framework which incorporates a dedicated multi-object pre-training phase. This is followed by a refinement process that leverages attention-based feature extraction with object masks, integrating them into a single unified image descriptor. Our MaO approach significantly outperforms existing retrieval methods and strong baselines, achieving notable improvements in both zero-shot and lightweight multi-object fine-tuning. We hope this work will lay the groundwork and inspire further research to enhance retrieval performance for this highly practical task.



Paperid:3909
Authors:Honghao Chen, Xingzhou Lou, Xiaokun Feng, Kaiqi Huang, Xinlong Wang
Abstract:
Chain of thought reasoning has demonstrated remarkable success in large language models, yet its adaptation to vision-language reasoning remains an open challenge with unclear best practices. Existing attempts typically employ reasoning chains at a coarse-grained level, which struggles to perform fine-grained structured reasoning and, more importantly, are difficult to evaluate the reward and quality of intermediate reasoning. In this work, we delve into chain of step reasoning for vision-language models, enabling assessing reasoning step quality accurately and leading to effective reinforcement learning and inference-time scaling with fine-grained rewards. We present a simple, effective, and fully transparent framework, including the step-level reasoning data, process reward model (PRM), and reinforcement learning training. With the proposed approaches, our models set strong baselines with consistent improvements on challenging vision-language benchmarks. More importantly, we conduct a thorough empirical analysis and ablation study, unveiling the impact of each component and several intriguing properties of inference-time scaling. We believe this paper serves as a baseline for vision-language models and offers insights into more complex multimodal reasoning. Our dataset, PRM, and code will be made publicly available.



Paperid:3910
Authors:Dong Li, Xujiang Zhao, Linlin Yu, Yanchi Liu, Wei Cheng, Zhengzhang Chen, Zhong Chen, Feng Chen, Chen Zhao, Haifeng Chen
Abstract:
Large Language Models (LLMs) offer promising capabilities for tackling complex reasoning tasks, including optimization problems. However, existing methods either rely on prompt engineering—leading to poor generalization across problem types—or require costly supervised training. We introduce SolverLLM, a training-free framework that leverages test-time scaling to solve diverse optimization problems. Rather than solving directly, SolverLLM generates mathematical formulations and translates them into solver-ready code, guided by a novel Monte Carlo Tree Search (MCTS) strategy. To enhance the search process, we modify classical MCTS with (1) dynamic expansion for adaptive formulation generation, (2) prompt backpropagation to guide exploration via outcome-driven feedback, and (3) uncertainty backpropagation to incorporate reward reliability into decision-making. Experiments on six standard benchmark datasets demonstrate that \sysname{} outperforms both prompt-based and learning-based baselines, achieving strong generalization without additional training. Code available: https://anonymous.4open.science/r/LLM_Solver-BD93.



Paperid:3911
Authors:Tianpai Luo, Xinyuan Fan, Weichi Wu
Abstract:
This work presents the first theoretically justified simultaneous inference framework for off-policy evaluation (OPE). In contrast to existing methods that focus on point estimates or pointwise confidence intervals (CIs), the new framework quantifies global uncertainty across an infinite or continuous initial state space, offering valid inference over the entire state space. Our method leverages sieve-based Q-function estimation and (high-dimensional) Gaussian approximation techniques over convex regions, which further motivates a new multiplier bootstrap algorithm for constructing asymptotically correct simultaneous confidence regions (SCRs). The widths of the SCRs exceed those of the pointwise CIs by only a logarithmic factor, indicating that our procedure is nearly optimal in terms of efficiency. The effectiveness of the proposed approach is demonstrated through simulations and analysis of the OhioT1DM dataset.



Paperid:3912
Authors:Sizhe Tang, Jiayu Chen, Tian Lan
Abstract:
Abstract:Monte Carlo Tree Search (MCTS), which leverages Upper Confidence Bound for Trees (UCTs) to balance exploration and exploitation through randomized sampling, is instrumental to solving complex planning problems. However, for multi-agent planning, MCTS is confronted with a large combinatorial action space that often grows exponentially with the number of agents. As a result, the branching factor of MCTS during tree expansion also increases exponentially, making it very difficult to efficiently explore and exploit during tree search. To this end, we propose MALinZero, a new approach to leverage low-dimensional representational structures on joint-action returns and enable efficient MCTS in complex multi-agent planning. Our solution can be viewed as projecting the joint-action returns into the low-dimensional space representable using a contextual linear bandit problem formulation. We solve the contextual linear bandit problem with convex and $\mu$-smooth loss functions -- in order to place more importance on better joint actions and mitigate potential representational limitations -- and derive a linear Upper Confidence Bound applied to trees (LinUCT) to enable novel multi-agent exploration and exploitation in the low-dimensional space. We analyze the regret of MALinZero for low-dimensional reward functions and propose an $(1-\tfrac1e)$-approximation algorithm for the joint action selection by maximizing a sub-modular objective. MALinZero demonstrates state-of-the-art performance on multi-agent benchmarks such as matrix games, SMAC, and SMACv2, outperforming both model-based and model-free multi-agent reinforcement learning baselines with faster learning speed and better performance.



Paperid:3913
Authors:Wangkai Li, Rui Sun, Zhaoyang Li, Tianzhu Zhang
Abstract:
Pseudo-label learning is widely used in semantic segmentation, particularly in label-scarce scenarios such as unsupervised domain adaptation (UDA) and semi-supervised learning (SSL). Despite its success, this paradigm can generate erroneous pseudo-labels, which are further amplified during training due to utilization of one-hot encoding. To address this issue, we propose ECOCSeg, a novel perspective for segmentation models that utilizes error-correcting output codes (ECOC) to create a fine-grained encoding for each class. ECOCSeg offers several advantages. First, an ECOC-based classifier is introduced, enabling model to disentangle classes into attributes and handle partial inaccurate bits, improving stability and generalization in pseudo-label learning. Second, a bit-level label denoising mechanism is developed to generate higher-quality pseudo-labels, providing adequate and robust supervision for unlabeled images. ECOCSeg can be easily integrated with existing methods and consistently demonstrates significant improvements on multiple UDA and SSL benchmarks across different segmentation architectures.



Paperid:3914
Authors:Qingmei Li, Yang Zhang, Zurong Mai, Yuhang Chen, Loushuohong, Henglian Huang, Jiarui Zhang, Zhiwei Zhang, Yibin Wen, Weijia Li, Haohuan Fu, Huang Jianxi, Juepeng Zheng
Abstract:
Large Multimodal Models (LMMs) has demonstrated capabilities across various domains, but comprehensive benchmarks for agricultural remote sensing (RS) remain scarce. Existing benchmarks designed for agricultural RS scenarios exhibit notable limitations, primarily in terms of insufficient scene diversity in the dataset and oversimplified task design. To bridge this gap, we introduce AgroMind, a comprehensive agricultural remote sensing benchmark covering four task dimensions: spatial perception, object understanding, scene understanding, and scene reasoning, with a total of 13 task types, ranging from crop identification and health monitoring to environmental analysis. We curate a high-quality evaluation set by integrating eight public datasets and one private farmland plot dataset, containing 25,026 QA pairs and 15,556 images. The pipeline begins with multi-source data preprocessing, including collection, format standardization, and annotation refinement. We then generate a diverse set of agriculturally relevant questions through the systematic definition of tasks. Finally, we employ LMMs for inference, generating responses, and performing detailed examinations. We evaluated 18 open-source LMMs and 3 closed-source models on AgroMind. Experiments reveal significant performance gaps, particularly in spatial reasoning and fine-grained recognition, it is notable that human performance lags behind several leading LMMs. By establishing a standardized evaluation framework for agricultural RS, AgroMind reveals the limitations of LMMs in domain knowledge and highlights critical challenges for future work. Data and code can be accessed at https://rssysu.github.io/AgroMind/.



Paperid:3915
Authors:Ankur Sinha, Shobhit Arora, Dhaval Pujara
Abstract:
This study presents AutoOpt-11k, a unique image dataset of 11000 handwritten and printed mathematical optimization models corresponding to single-objective, multi-objective, bi-level, and stochastic optimization problems exhibiting various types of complexities such as non-linearity, non-convexity, non-differentiability, discontinuity, and high-dimensionality. The labels consist of the LaTeX representation for all the images and modeling language representation for a subset of images. The dataset is created by 25 experts following ethical data creation guidelines and verified in two-phases to avoid errors. Further, we develop AutoOpt framework, a machine learning based automated approach for solving optimization problems, where the user just needs to provide an image of the formulation and AutoOpt solves it efficiently without any further human intervention. AutoOpt framework consists of three Modules: (i) M1 (ImagetoText)- a deep learning model performs the Mathematical Expression Recognition (MER) task to generate the LaTeX code corresponding to the optimization formulation in image; (ii) M2 (TexttoText)- a small-scale fine-tuned LLM generates the PYOMO script (optimization modeling language) from LaTeX code; (iii) M3 (Optimization)- a Bilevel Optimization based Decomposition (BOBD) method solves the optimization formulation described in the PYOMO script. We use AutoOpt-11k dataset for training and testing of deep learning models employed in AutoOpt. The deep learning model for MER task (M1) outperforms ChatGPT, Gemini, and Nougat on BLEU score metric. BOBD method (M3), which is a hybrid approach, yields better results on complex test problems compared to common approaches, like interior-point algorithm and genetic algorithm. The dataset and the framework are being released publicly on GitHub.



Paperid:3916
Authors:Xianghua Zeng, Hao Peng, Yicheng Pan, Angsheng Li, Guanlin Wu
Abstract:
Diffusion-based generative methods have shown promising potential for modeling trajectories from offline reinforcement learning (RL) datasets, and hierarchical diffusion has been introduced to mitigate variance accumulation and computational challenges in long-horizon planning tasks.However, existing approaches typically assume a fixed two-layer diffusion hierarchy with a single predefined temporal scale, which limits adaptability to diverse downstream tasks and reduces flexibility in decision making.In this work, we propose SIHD, a novel Structural Information-based Hierarchical Diffusion framework for effective and stable offline policy learning in long-horizon environments with sparse rewards.Specifically, we analyze structural information embedded in offline trajectories to construct the diffusion hierarchy adaptively, enabling flexible trajectory modeling across multiple temporal scales.Rather than relying on reward predictions from localized sub-trajectories, we quantify the structural information gain of each state community and use it as a conditioning signal within the corresponding diffusion layer.To reduce overreliance on offline datasets, we introduce a structural entropy regularizer that encourages exploration of underrepresented states while avoiding extrapolation errors from distributional shifts.Extensive evaluations on challenging offline RL tasks show that SIHD significantly outperforms state-of-the-art baselines in decision-making performance and demonstrates superior generalization across diverse scenarios.



Paperid:3917
Authors:Qingyi Pan, Yicheng Li
Abstract:
Abstract:Real-world time series analysis, such as healthcare, autonomous driving, and solar energy, faces unique challenges arising from the scarcity of labeled data, highlighting the need for effective semi-supervised learning methods. While the Virtual Adversarial Training (VAT) method has shown promising performance in leveraging unlabeled data for smoother predictive distributions, straightforward extensions of VAT often fall short on time series tasks as they neglect the temporal structure of the data in the adversarial perturbation. In this paper, we propose the framework of functional Virtual Adversarial Training (f-VAT) that can incorporate the functional structure of the data into perturbations. By theoretically establishing a duality between the perturbation norm and the functional model sensitivity, we propose to use an appropriate Sobolev ($H^{-s}$) norm to generate structured functional adversarial perturbations for semi-supervised time series classification. Our proposed f-VAT method outperforms recent methods and achieves superior performance in extensive semi-supervised time series classification tasks (e.g., up to $ \approx 9$ % performance improvement). We also provide additional visualization studies to offer further insights into the superiority of f-VAT.



Paperid:3918
Authors:Marton Havasi, Brian Karrer, Itai Gat, Ricky T. Q. Chen
Abstract:
Autoregressive generative models naturally generate variable-length sequences, while non-autoregressive models struggle, often imposing rigid, token-wise structures. We propose Edit Flows, a non-autoregressive model that overcomes these limitations by defining a discrete flow over sequences through edit operations—insertions, deletions, and substitutions. By modeling these operations within a Continuous-time Markov Chain over the sequence space, Edit Flows enable flexible, position-relative generation that aligns more closely with the structure of sequence data. Our training method sidesteps intractable computations by leveraging an expanded state space with auxiliary variables, making the learning process efficient and tractable. Empirical results show that Edit Flows outperform baselines on image captioning tasks and significantly improve upon prior non-autoregressive models in code and text generation.



Paperid:3919
Authors:Kuicai Dong, CHANG YUJING, Shijie Huang, Yasheng Wang, Ruiming Tang, Yong Liu
Abstract:
Document Visual Question Answering (DocVQA) faces dual challenges in processing lengthy multimodal documents (text, images, tables) and performing cross-modal reasoning. Current document retrieval-augmented generation (DocRAG) methods remain limited by their text-centric approaches, frequently missing critical visual information. The field also lacks robust benchmarks for assessing multimodal evidence selection and integration. We introduce MMDocRAG, a comprehensive benchmark featuring 4,055 expert-annotated QA pairs with multi-page, cross-modal evidence chains. Our framework introduces innovative metrics for evaluating multimodal quote selection and enables answers that interleave text with relevant visual elements. Through large-scale experiments with 60 VLM/LLM models and 14 retrieval systems, we identify persistent challenges in multimodal evidence retrieval, selection, and integration. Key findings reveal that advanced proprietary LVMs show superior performance than open-sourced alternatives. Also, they show moderate advantages using multimodal inputs over text-only inputs, while open-source alternatives show significant performance degradation. Notably, fine-tuned LLMs achieve substantial improvements when using detailed image descriptions. MMDocRAG establishes a rigorous testing ground and provides actionable insights for developing more robust multimodal DocVQA systems. Our benchmark and code are available at https://mmdocrag.github.io/MMDocRAG.



Paperid:3920
Authors:Jindong Yang, Han Fang, Kejiang Chen, Weiming Zhang, Nenghai Yu
Abstract:
Diffusion models have advanced rapidly in recent years, producing high-fidelity images while raising concerns about intellectual property protection and the misuse of generative AI. Image watermarking for diffusion models, particularly Noise-as-Watermark (NaW) methods, embeds watermarks into latent representations drawn from a standard Gaussian distribution to preserve image quality. For detection, the generation process is inverted to recover the initial noise vector containing the watermark before extraction. However, existing NaW methods struggle to balance watermark robustness with generation diversity. Some methods achieve strong robustness by heavily constraining the initial noise sampling, which degrades user experience, while others preserve diversity but prove too fragile for real-world deployment. To address this issue, we propose T2SMark, a training-free watermarking scheme based on two-stage tail-truncated sampling. Unlike prior methods that simply map bits to positive or negative values, tail-truncated sampling excludes the easily flipped regions of the latent distribution, enhancing robustness with even less redundancy. Our two-stage framework then compensates sampling diversity by incorporating a random key into both encryption pipelines—first as the payload and then as the encryption key. We evaluate T2SMark on diffusion models with both U-Net and DiT backbones. Extensive experiments show that it achieves an optimal balance between robustness and diversity.



Paperid:3921
Authors:Yaohui Zhang, Haijing ZHANG, Wenlong Ji, Tianyu Hua, Nick Haber, Hancheng Cao, Weixin Liang
Abstract:
The advent of large language models (LLMs) offers unprecedented opportunities to reimagine peer review beyond the constraints of traditional workflows.Despite these opportunities, prior efforts have largely focused on replicating traditional review workflows with LLMs serving as direct substitutes for human reviewers, while limited attention has been given to exploring new paradigms that fundamentally rethink how LLMs can participate in the academic review process.In this paper, we introduce and explore a novel mechanism that employs LLM agents to perform pairwise comparisons among manuscripts instead of individual scoring. By aggregating outcomes from substantial pairwise evaluations, this approach enables a more accurate and robust measure of relative manuscript quality.Our experiments demonstrate that this comparative approach significantly outperforms traditional rating-based methods in identifying high-impact papers. However, our analysis also reveals emergent biases in the selection process, notably a reduced novelty in research topics and an increased institutional imbalance. These findings highlight both the transformative potential of rethinking peer review with LLMs and critical challenges that future systems must address to ensure equity and diversity.



Paperid:3922
Authors:Wenhao Tang, Rong Qin, Heng Fang, Fengtao Zhou, Hao CHEN, Xiang Li, Ming-Ming Cheng
Abstract:
Abstract:Pre-trained encoders for offline feature extraction followed by multiple instance learning (MIL) aggregators have become the dominant paradigm in computational pathology (CPath), benefiting cancer diagnosis and prognosis.However, performance limitations arise from the absence of encoder fine-tuning for downstream tasks and disjoint optimization with MIL. While slide-level supervised end-to-end (E2E) learning is an intuitive solution to this issue, it faces challenges such as high computational demands and suboptimal results.These limitations motivate us to revisit E2E learning. We argue that prior work neglects inherent E2E optimization challenges, leading to performance disparities compared to traditional two-stage methods.In this paper, we pioneer the elucidation of optimization challenge caused by sparse-attention MIL and propose a novel MIL called ABMILX. ABMILX mitigates this problem through global correlation-based attention refinement and multi-head mechanisms.With the efficient multi-scale random patch sampling strategy, an E2E trained ResNet with ABMILX surpasses SOTA foundation models under the two-stage paradigm across multiple challenging benchmarks,while remaining computationally efficient ($<$ 10 RTX3090 GPU hours).We demonstrate the potential of E2E learning in CPath and calls for greater research focus in this area.The code is~\href{https://anonymous.4open.science/r/ABMILX-E480}{here}.



Paperid:3923
Authors:Yuhan Liu, LingHui Fu, Zhen Yang, Hao Chen, Youfu Li, Yongjian Deng
Abstract:
Abstract:Event cameras, with their capacity to provide high temporal resolution information between frames, are increasingly utilized for video frame interpolation (VFI) in challenging scenarios characterized by high-speed motion and significant occlusion. However, prevalent issues of blur and distortion within the keyframes and ground truth data used for training and inference in these demanding conditions are frequently overlooked. This oversight impedes the perceptual realism and multi-scene generalization capabilities of existing event-based VFI (E-VFI) methods when generating interpolated frames. Motivated by the observation that semantic-perceptual discrepancies between degraded and pristine images are considerably smaller than their image-level differences, we introduce EPA. This novel E-VFI framework diverges from approaches reliant on direct image-level supervision by constructing multilevel, degradation-insensitive semantic perceptual supervisory signals to enhance the perceptual realism and multi-scene generalization of the model's predictions. Specifically, EPA operates in two phases: it first employs a DINO-based perceptual extractor, a customized style adapter, and a reconstruction generator to derive multi-layered, degradation-insensitive semantic-perceptual features ($\mathcal{S}$). Second, a novel Bidirectional Event-Guided Alignment (BEGA) module utilizes deformable convolutions to align perceptual features from keyframes to ground truth with inter-frame temporal guidance extracted from event signals. By decoupling the learning process from direct image-level supervision, EPA enhances model robustness against degraded keyframes and unreliable ground truth information. Extensive experiments demonstrate that this approach yields interpolated frames more consistent with human perceptual preferences. *The code will be released upon acceptance.*



Paperid:3924
Authors:QiXu, Dongxu Wei, Lingzhe Zhao, Wenpu Li, Zhangchi Huang, Shunping Ji, Peidong Liu
Abstract:
Simultaneous understanding and 3D reconstruction plays an important role in developing end-to-end embodied intelligent systems.To achieve this, recent approaches resort to 2D-to-3D feature alignment paradigm, which leads to limited 3D understanding capability and potential semantic information loss.In light of this, we propose SIU3R, the first alignment-free framework for generalizable simultaneous understanding and 3D reconstruction from unposed images.Specifically, SIU3R bridges reconstruction and understanding tasks via pixel-aligned 3D representation, and unifies multiple understanding tasks into a set of unified learnable queries, enabling native 3D understanding without the need of alignment with 2D models.To encourage collaboration between the two tasks with shared representation, we further conduct in-depth analyses of their mutual benefits, and propose two lightweight modules to facilitate their interaction.Extensive experiments demonstrate that our method achieves state-of-the-art performance not only on the individual tasks of 3D reconstruction and understanding, but also on the task of simultaneous understanding and 3D reconstruction, highlighting the advantages of our alignment-free framework and the effectiveness of the mutual benefit designs.



Paperid:3925
Authors:Kyurae Kim, Yian Ma, Trevor Campbell, Jacob Gardner
Abstract:
Abstract:We prove that, given a mean-field location-scale variational family, black-box variational inference (BBVI) with the reparametrization gradient converges at an almost dimension-independent rate. Specifically, for strongly log-concave and log-smooth targets, the number of iterations for BBVI with a sub-Gaussian family to achieve an objective $\epsilon$-close to the global optimum is $\mathrm{O}(\log d)$, which improves over the $\mathrm{O}(d)$ dependence of full-rank location-scale families. For heavy-tailed families, we provide a weaker $\mathrm{O}(d^{2/k})$ dimension dependence, where $k$ is the number of finite moments. Additionally, if the Hessian of the target log-density is constant, the complexity is free of any explicit dimension dependence. We also prove that our bound on the gradient variance, which is key to our result, cannot be improved using only spectral bounds on the Hessian of the target log-density.



Paperid:3926
Authors:Vivek Myers, Bill Zheng, Benjamin Eysenbach, Sergey Levine
Abstract:
Learned successor features provide a powerful framework for learning goal-reaching policies. These representations are constructed such that similarity in the representation space predicts future outcomes, allowing goal-reaching policies to be extracted. Representations learned for forward inference have some practical limitations - stitching of behaviors does not arise naturally with forward objectives like contrastive classification, and additional regularization is required to enable valid policy extraction. In this work, we propose a new representation learning objective that enables extraction of goal-reaching policies. Our key insight is that rather than learning representations of the future, we should really learn representations that can associate outcomes with preceding states. We show that when combined with existing quasimetric network parameterization and the right invariances, these representations let us learn optimal goal-reaching policies from offline data. On existing offline GCRL benchmarks, the hindsight classification objective improves performance with a simpler algorithm and fewer independent networks/parameters to learn relative to past methods.



Paperid:3927
Authors:Zekun Qi, Wenyao Zhang, Yufei Ding, Runpei Dong, XinQiang Yu, Jingwen Li, Lingyun Xu, Baoyu Li, Xialin He, Guofan Fan, Jiazhao Zhang, Jiawei He, Jiayuan Gu, Xin Jin, Kaisheng Ma, Zhizheng Zhang, He Wang, Li Yi
Abstract:
While spatial reasoning has made progress in object localization relationships, it often overlooks object orientation— a key factor in 6-DoF fine-grained manipulation. Traditional pose representations rely on pre-defined frames or templates, limiting generalization and semantic grounding. In this paper, we introduce the concept of semantic orientation, which defines object orientations using natural language in a reference-frame-free manner (e.g., the "plug-in" direction of a USB or the "handle" direction of a cup). To support this, we construct OrienText300K, a large-scale dataset of 3D objects annotated with semantic orientations, and develop PointSO, a general model for zero-shot semantic orientation prediction. By integrating semantic orientation into VLM agents, our SoFAR framework enables 6-DoF spatial reasoning and generates robotic actions. Extensive experiments demonstrated the effectiveness and generalization of our SOFAR, e.g., zero-shot 48.7% successful rate on OpenDOR and 58.3% successful rate on SIMPLER Widox-X setting.



Paperid:3928
Authors:R. Teal Witter, Yurong Liu, Christopher Musco
Abstract:
Abstract:With origins in game-theory, probabilistic values like Shapley values, Banzhaf values, and semi-values have emerged as a central tool in explainable AI. They are used for feature attribution, data attribution, data valuation, and more. Since all of these values require exponential time to compute exactly, research has focused on efficient approximation methods using two techniques: Monte Carlo sampling and linear regression formulations. In this work, we present a new way of combining both of these techniques. Our approach is more flexible than prior algorithms, allowing for linear regression to be replaced with any function family whose probabilistic values can be computed efficiently. This allows us to harness the accuracy of tree-based models like XGBoost, while still producing unbiased estimates. From experiments across eight datasets, we find that our methods give state-of-the-art performance for estimating probabilistic values. For Shapley values, the error of our methods is up to $6\times$ lower than Permutation SHAP (the most popular Monte Carlo method), $2.75\times$ lower than Kernel SHAP (the most popular linear regression method), and $1.75\times$ lower than Leverage SHAP (the prior state-of-the-art Shapley value estimator). For more general probabilistic values, we can obtain error up to $60\times$ lower than prior work.



Paperid:3929
Authors:Minsoo Kim, Kyuhong Shim, Jungwook Choi, Simyung Chang
Abstract:
Modern multimodal large language models (MLLMs) can reason over hour-long video, yet their key–value (KV) cache grows linearly with time—quickly exceeding the fixed memory of phones, AR glasses, and edge robots. Prior compression schemes either assume the whole video and user query are available offline or must first build the full cache, so memory still scales with stream length. InfiniPot-V is the first training-free, query-agnostic framework that enforces a hard, length-independent memory cap for streaming video understanding. During video encoding it monitors the cache and, once a user-set threshold is reached, runs a lightweight compression pass that (i) removes temporally redundant tokens via Temporal-axis Redundancy (TaR) metric and (ii) keeps semantically significant tokens via Value-Norm (VaN) ranking. Across four open-source MLLMs and four long-video and two streaming-video benchmarks, InfiniPot-V cuts peak GPU memory by up to 94\%, sustains real-time generation, and matches or surpasses full-cache accuracy—even in multi-turn dialogues. By dissolving the KV-cache bottleneck without retraining or query knowledge, InfiniPot-V closes the gap for on-device streaming video assistants.



Paperid:3930
Authors:Markus Gross, Aya Fahmy, Danit Niwattananan, Dominik Muhle, Rui Song, Daniel Cremers, Henri Meeß
Abstract:
Abstract:Semantic Scene Completion (SSC) has emerged as a pivotal approach for jointly learning scene geometry and semantics, enabling downstream applications such as navigation in mobile robotics. The recent generalization to Panoptic Scene Completion (PSC) advances the SSC domain by integrating instance-level information, thereby enhancing object-level sensitivity in scene understanding. While PSC was introduced using LiDAR modality, methods based on camera images remain largely unexplored. Moreover, recent Transformer-based SSC approaches utilize a fixed set of learned queries to reconstruct objects within the scene volume. Although these queries are typically updated with image context during training, they remain static at test time, limiting their ability to dynamically adapt specifically to the observed scene. To address these limitations, we propose IPFormer, the first approach that leverages context-adaptive instance proposals at train and test time to solve vision-based 3D Panoptic Scene Completion. Specifically, IPFormer dynamically initializes these queries as panoptic instance proposals derived from image context and further refines them through attention-based encoding and decoding to reason about semantic instance-voxel relationships. Experimental results show that our approach surpasses state-of-the-art methods in overall panoptic metrics PQ$^\dagger$ and PQ-All, matches performance in individual metrics, and achieves a runtime reduction exceeding 14$\times$. Furthermore, our ablation studies reveal that dynamically deriving instance proposals from image context, as opposed to random initialization, leads to a 3.62% increase in PQ-All and a remarkable average improvement of 18.65% in combined Thing-metrics. These results underscore the effectiveness of IPFormer and highlight its introduction of context-adaptive instance proposals as a pioneering effort in addressing vision-based 3D Panoptic Scene Completion.



Paperid:3931
Authors:Yun Hua, Haosheng Chen, Shiqin Wang, Wenhao Li, Xiangfeng Wang, Jun Luo
Abstract:
Large Language Models (LLMs) are increasingly deployed as autonomous agents in multi-agent systems, and promising coordination has been demonstrated in handling complex tasks under predefined roles and scripted workflows.However, significant challenges remain in open-ended environments, where agents are inherently self-interested and explicit coordination guidelines are absent. In such scenarios, misaligned incentives frequently lead to social dilemmas and inefficient collective outcomes.Inspired by how human societies tackle similar coordination challenges—through temporary collaborations like employment or subcontracting—a cooperative workflow \textbf{Shapley-Coop} is proposed. This workflow enables self-interested Large Language Model (LLM) agents to engage in emergent collaboration by using a fair credit allocation mechanism to ensure each agent’s contributions are appropriately recognized and rewarded.Shapley-Coop introduces structured negotiation protocols and Shapley-inspired reasoning to estimate agents’ marginal contributions, thereby enabling effective task-time coordination and equitable post-task outcome redistribution. This results in effective coordination that fosters collaboration while preserving agent autonomy, through a rational pricing mechanism that encourages cooperative behavior.Evaluated in two multi-agent games and a software engineering simulation, Shapley-Coop consistently enhances LLM agent collaboration and facilitates equitable outcome redistribution, accurately reflecting individual contributions during the task execution process.



Paperid:3932
Authors:Zhangdie Yuan, Zifeng Ding, Andreas Vlachos
Abstract:
Forecasting is an important task in many domains. However, existing forecasting benchmarks lack comprehensive confidence assessment, focusing on limited question types, and often consist of artificial questions that do not reflect real-world needs. To address these gaps, we introduce FOReCAst (Future Outcome Reasoning and Confidence Assessment), a benchmark that evaluates models' ability to make predictions and their confidence in them. FOReCAst spans diverse forecasting scenarios involving Boolean questions, timeframe prediction, and quantity estimation, enabling a comprehensive evaluation of both prediction accuracy and confidence calibration for real-world applications.



Paperid:3933
Authors:Hsin Yi Hsieh, Shang-Wei Liu, Chang Meng, Chien-Hua Chen, Shuo-Yueh Lin, Hung-Ju Lin, Hen-Hsen Huang, I-Chen Wu
Abstract:
Vision-language models (VLMs) struggle with culturally specific content, an issue often overlooked as existing benchmarks focus on dominant languages and globalized datasets. We introduce TaiwanVQA, a VQA benchmark designed for Taiwanese culture to evaluate recognition and reasoning in regional contexts. TaiwanVQA contains 2,736 images and 5,472 manually designed questions on topics like traditional food, public signs, festivals, and landmarks. The official benchmark set includes 1,000 images and 2,000 questions for systematic assessment, with the remainder of the data used as training material. Tests on state-of-the-art VLMs show strong visual recognition but weaknesses in cultural reasoning. We propose a data augmentation strategy combining human-annotated and synthesized dialogues to improve cultural understanding. Fine-tuning yields significant gains on TaiwanVQA while maintaining stable performance on other multimodal tasks. TaiwanVQA offers a scalable framework for building culturally grounded AI models in low-resource cultures, promoting diversity and fairness in multimodal AI.Our dataset and code are publicly available onHugging FaceandGitHub.



Paperid:3934
Authors:Yi Li, Zhichun Guo, Guanpeng Li, Bingzhe Li
Abstract:
Abstract:Training Graph Neural Networks (GNNs) often relies on repeated, irregular, and expensive message-passing operations over all nodes (e.g., $N$), leading to high computational overhead. To alleviate this inefficiency, we revisit the GNNs training from a spectral perspective. In many real-world graphs, node features and embeddings exhibit sparse representation in the Graph Fourier domain. This inherent spectral sparsity aligns well with the principles of Compressed Sensing, which posits that signals sparse in one transform domain can be accurately reconstructed from a significantly reduced number of measurements. This observation motivates the design of a more efficient GNNs that operates predominantly in compressed spectral subspace. Thus, we propose You Only Spectralize Once (YOSO), a GNN training scheme that performs single Graph Fourier Transformation to project features onto a learnable orthonormal Fourier basis, retaining only $M$ spectral coefficients ($M \ll N$). The entire GNN computation is then carried out in reduced spectral domain. Final full-graph embeddings are recovered only at output layer by solving a bounded $\ell_{2,1}$-regularized optimization problem. Theoretically, drawing upon Compressed Sensing theory, we prove stable recovery throughout training by showing that the projection onto our learnable Fourier basis can satisfy the Restricted Isometry Property when $M=\mathcal{O}(k \log N)$ for $k$-row-sparse spectra, acting as the measurement process. Empirically, YOSO achieves an average 74\% reduction in training time across five benchmark datasets compared to state-of-the-art methods, while maintaining competitive accuracy.



Paperid:3935
Authors:HongQiong Zhong, Qingyang Teng, Baolin Zheng, Guanlin Chen, Yingshui Tan, Zhendong Liu, Jiaheng Liu, Wenbo Su, Xiaoyong Zhu, Bo Zheng, Kaifu Zhang
Abstract:
As large Visual-Language Models (VLMs) continue to evolve, they have demonstrated increasingly sophisticated logical reasoning capabilities and multimodal thought generation, opening doors to widespread applications. However, this advancement raises serious concerns about content security, particularly when these models process complex multimodal inputs requiring intricate reasoning. When faced with these safety challenges, the critical competition between logical reasoning and safety objectives of VLMs is often overlooked in previous works. In this paper, we introduce Visualization-of-Thought Attack (VoTA), a novel and automated attack framework that strategically constructs chains of images with risky visual thoughts to challenge victim models. Our attack provokes the inherent conflict between the model's logical processing and safety protocols, ultimately leading to the generation of unsafe content. Through comprehensive experiments, VoTA achieves remarkable effectiveness, improving the average attack success rate (ASR) by 26.71\% (from 63.70\% to 90.41\%) on 9 open-source VLMs and 6 commercial VLMs, compared to the state-of-the-art methods. These results expose a critical vulnerability: current VLMs struggle to maintain safety guarantees when processing insecure multimodal visualization-of-thought inputs, highlighting the urgency and necessity of enhancing safety alignment.



Paperid:3936
Authors:Zhiheng Zhang, Zichen Wang
Abstract:
Network interference has attracted significant attention in the field of causal inference, encapsulating various sociological behaviors in which the treatment assigned to one individual within a networkmay affect the outcomes of others, such as their neighbors. A key challenge in this setting is thatstandard causal inference methods often assume independent treatment effects among individuals,which may not hold in networked environments. To estimate interference-aware causal effects, atraditional approach is to inherit the independent settings, where practitioners randomly assign experimental participants to different groups and compare their outcomes. Although effective in offlinesettings, this strategy becomes problematic in sequential experiments, where suboptimal decisions persist, leading to substantial regret. To address this issue, we introduce a unified interference-aware framework for online experimental design. Compared to existing studies, we extend thedefinition of arm space using the statistical concept of exposure mapping, which allows fora more flexible and context-aware representation of treatment effects in network settings. Crucially, we establish a Pareto-optimal trade-off between estimation accuracy and regret under thenetwork concerning both time period and arm space, which remains superior to baseline modelseven without network interference. Furthermore, we propose an algorithmic implementation anddiscuss its generalization in different learning settings and network topology.



Paperid:3937
Authors:Yatin Dandi, Luca Pesce, Lenka Zdeborová, Florent Krzakala
Abstract:
Understanding the advantages of deep neural networks trained by gradient descent (GD) compared to shallow models remains an open theoretical challenge. In this paper, we introduce a class of target functions (single and multi-index Gaussian hierarchical targets) that incorporate a hierarchy of latent subspace dimensionalities. This framework enables us to analytically study the learning dynamics and generalization performance of deep networks compared to shallow ones in the high-dimensional limit. Specifically, our main theorem shows that feature learning with GD successively reduces the effective dimensionality, transforming a high-dimensional problem into a sequence of lower-dimensional ones. This enables learning the target function with drastically less samples than with shallow networks. While the results are proven in a controlled training setting, we also discuss more common training procedures and argue that they learn through the same mechanisms. These findings open the way to further quantitative studies of the crucial role of depth in learning hierarchical structures with deep networks.



Paperid:3938
Authors:Yang Zhao, Pu Wang, Hao Frank Yang
Abstract:
Task-adaptive prompting and reasoning for Large Language Models (LLMs) are necessity in real-world applications. Determining how to integrate domain knowledge, enhance reasoning efficiency, and even provide domain experts with refined knowledge integration hints are particularly crucial yet unresolved tasks. In this research, we propose Evolutionary Graph Optimization for Prompting (EGO-Prompt), an automated framework to designing optimal prompts, efficient reasoning processes and providing enhanced casual process. EGO-Prompt begins with an general prompt and fault-tolerant initial Semantic Causal Graph (SCG) descriptions, constructed by human experts, which is then automatically refined and optimized to guide LLM reasoning. Recognizing that expert-defined SCGs may be partial or imperfect and that their optimal integration varies across LLMs, EGO-Prompt integrates a novel casual-guided textual gradient process in two steps: first, generating nearly deterministic reasoning guidance from the SCG for each instance, and second, adapting the LLM to effectively utilize the guidance alongside the original input. The iterative optimization algorithm further refines both the SCG and the reasoning mechanism using textual gradients with ground-truth. We tested the framework on real-world public health, transportation and human behavior tasks and EGO-Prompt demonstrates clear improvements, achieving increases of 8% to 12% in F1 score compared to state-of-the-art methods, together with a refined domain SCG.



Paperid:3939
Authors:Marc Brooks, Gabriel Durham, Kihyuk Hong, Ambuj Tewari
Abstract:
Recent advances in generative artificial intelligence (GenAI) models have enabled the generation of personalized content that adapts to up-to-date user context. While personalized decision systems are often modeled using bandit formulations, the integration of GenAI introduces new structure into otherwise classical sequential learning problems. In GenAI-powered interventions, the agent selects a query, but the environment experiences a stochastic response drawn from the generative model. Standard bandit methods do not explicitly account for this structure, where actions influence rewards only through stochastic, observed treatments. We introduce generator-mediated bandit-Thompson sampling (GAMBITTS), a bandit approach designed for this action/treatment split, using mobile health interventions with large language model-generated text as a motivating case study. GAMBITTS explicitly models both the treatment and reward generation processes, using information in the delivered treatment to accelerate policy learning relative to standard methods. We establish regret bounds for GAMBITTS by decomposing sources of uncertainty in treatment and reward, identifying conditions where it achieves stronger guarantees than standard bandit approaches. In simulation studies, GAMBITTS consistently outperforms conventional algorithms by leveraging observed treatments to more accurately estimate expected rewards.



Paperid:3940
Authors:KunHo Heo, GiHyun Kim, SuYeon Kim, MyeongAh Cho
Abstract:
3D Semantic Scene Graph Prediction aims to detect objects and their semantic relationships in 3D scenes, and has emerged as a crucial technology for robotics and AR/VR applications. While previous research has addressed dataset limitations and explored various approaches including Open-Vocabulary settings, they frequently fail to optimize the representational capacity of object and relationship features, showing excessive reliance on Graph Neural Networks despite insufficient discriminative capability. In this work, we demonstrate through extensive analysis that the quality of object features plays a critical role in determining overall scene graph accuracy. To address this challenge, we design a highly discriminative object feature encoder and employ a contrastive pretraining strategy that decouples object representation learning from the scene graph prediction. This design not only enhances object classification accuracy but also yields direct improvements in relationship prediction. Notably, when plugging in our pretrained encoder into existing frameworks, we observe substantial performance improvements across all evaluation metrics. Additionally, whereas existing approaches have not fully exploited the integration of relationship information, we effectively combine both geometric and semantic features to achieve superior relationship prediction. Comprehensive experiments on the 3DSSG dataset demonstrate that our approach significantly outperforms previous state-of-the-art methods.



Paperid:3941
Authors:Hanqing Zeng, Yinglong Xia, Zhuokai Zhao, Chuan Jiang, Qiang Zhang, Jiayi Liu, Lizhu Zhang, Qunshu Zhang, Xiangjun Fan, Benyu Zhang
Abstract:
Fine-tuning pre-trained large language models (LLMs) presents a dual challenge of balancing parameter efficiency and model capacity. Existing methods like low-rank adaptations (LoRA) are efficient but lack flexibility, while Mixture-of-Experts (MoE) architectures enhance model capacity at the cost of more & under-utilized parameters. To address these limitations, we propose Structural Mixture of Residual Experts (S'MoRE), a novel framework that seamlessly integrates the efficiency of LoRA with the flexibility of MoE. Specifically, S'MoRE employs hierarchical low-rank decomposition of expert weights, yielding residuals of varying orders interconnected in a multi-layer structure. By routing input tokens through sub-trees of residuals, S'MoRE emulates the capacity of many experts by instantiating and assembling just a few low-rank matrices. We craft the inter-layer propagation of S'MoRE's residuals as a special type of Graph Neural Network (GNN), and prove that under similar parameter budget, S'MoRE improves "structural flexibility" of traditional MoE (or Mixture-of-LoRA) by exponential order. Comprehensive theoretical analysis and empirical results demonstrate that S'MoRE achieves superior fine-tuning performance, offering a transformative approach for efficient LLM adaptation.



Paperid:3942
Authors:Jiahao Wang, Weiye Xu, Aijun Yang, Wengang Zhou, Lewei Lu, Houqiang Li, Xiaohua Wang, Jinguo Zhu
Abstract:
Outcome‑reward reinforcement learning (RL) is a common—and increasingly significant—way to refine the step‑by‑step reasoning of multimodal large language models (MLLMs). In the multiple‑choice setting—a dominant format for multimodal reasoning benchmarks—the paradigm faces a significant yet often overlooked obstacle: unfaithful trajectories that guess the correct option after a faulty chain of thought receive the same reward as genuine reasoning, which is a flaw that cannot be ignored. We propose Self‑Consistency Sampling (SCS) to correct this issue. For each question, SCS (i) introduces small visual perturbations and (ii) performs repeated truncation‑and‑resampling of a reference trajectory; agreement among the resulting trajectories yields a differentiable consistency score that down‑weights unreliable traces during policy updates. Plugging SCS into RLOO, GRPO, REINFORCE++ series improves accuracy by up to 7.7 percentage points on six multimodal benchmarks with negligible extra computation, offering a simple, general remedy for outcome‑reward RL in MLLMs.



Paperid:3943
Authors:Runzhe Zhan, Zhihong Huang, Xinyi Yang, Lidia Chao, Min Yang, Derek Wong
Abstract:
Recent advancements in large reasoning models (LRMs) have introduced the "slow thinking" paradigm, which leverages their inherent strengths to enhance reasoning capabilities for complex downstream tasks. However, the potential of LRMs as evaluators for machine translation (MT) quality remains underexplored. We provides the first systematic analysis of LRM-as-a-judge in MT evaluation. We identify key challenges, revealing LRMs require tailored evaluation materials, tend to ``overthink'' simpler instances and have issues with scoring mechanisms leading to overestimation. To address these, we propose to calibrate LRM thinking by training them on synthetic, human-like thinking trajectories. Our experiments on WMT24 Metrics benchmarks demonstrate that this approach largely reduces thinking budgets by ~35x while concurrently improving evaluation performance across different LRM scales from 7B to 32B (e.g., R1-Distill-Qwen-7B achieves a +8.7 correlation point improvement). These findings highlight the substantial potential of efficiently calibrated LRMs to advance human-centric MT evaluation.



Paperid:3944
Authors:Ziyan Wang, Sizhe Wei, Xiaoming Huo, Hao Wang
Abstract:
Diffusion models have made significant advancements in recent years. However, their performance often deteriorates when trained or fine-tuned on imbalanced datasets. This degradation is largely due to the disproportionate representation of majority and minority data in image-text pairs. In this paper, we propose a general fine-tuning approach, dubbed PoGDiff, to address this challenge. Rather than directly minimizing the KL divergence between the predicted and ground-truth distributions, PoGDiff replaces the ground-truth distribution with a Product of Gaussians (PoG), which is constructed by combining the original ground-truth targets with the predicted distribution conditioned on a neighboring text embedding. Experiments on real-world datasets demonstrate that our method effectively addresses the imbalance problem in diffusion models, improving both generation accuracy and quality.



Paperid:3945
Authors:Zhaoze Wang, Genela Morris, Dori Derdikman, Pratik Chaudhari, Vijay Balasubramanian
Abstract:
Grid cells in the medial entorhinal cortex (MEC) are believed to path integrate speed and direction signals to activate at triangular grids of locations in an environment, thus implementing a population code for position. In parallel, place cells in the hippocampus (HC) fire at spatially confined locations, with selectivity tuned not only to allocentric position but also to environmental contexts, such as sensory cues. Although grid and place cells both encode spatial information and support memory for multiple locations, why animals maintain two such representations remains unclear. Noting that place representations seem to have other functional roles in intrinsically motivated tasks such as recalling locations from sensory cues, we propose that animals maintain grid and place representations together to support planning. Specifically, we posit that place cells auto-associate not only sensory information relayed from the MEC but also grid cell patterns, enabling recall of goal location grid patterns from sensory and motivational cues, permitting subsequent planning with only grid representations. We extend a previous theoretical framework for grid-cell-based planning and show that local transition rules can generalize to long-distance path forecasting. We further show that a planning network can sequentially update grid cell states toward the goal. During this process, intermediate grid activity can trigger place cell pattern completion, reconstructing experiences along the planned path. We demonstrate all these effects using a single-layer RNN that simultaneously models the HC-MEC loop and the planning subnetwork. We show that such recurrent mechanisms for grid cell-based planning, with goal recall driven by the place system, make several characteristic, testable predictions.



Paperid:3946
Authors:Qi Wang, Tianfei Zhou, Yanrui Yu, Ye Yuan, Rui Mao
Abstract:
Reinforcement fine-tuning (RFT) has shown great promise in achieving human-level reasoning capabilities of Large Language Models (LLMs), and has recently been extended to MLLMs. Nevertheless, reasoning about videos, which is a fundamental aspect of human intelligence, remains a persistent challenge due to the complex logic, temporal and causal structures inherent in video data. To fill this gap, we propose VideoRFT, a novel approach that extends the RFT paradigm to cultivate human-like video reasoning capabilities in MLLMs. VideoRFT follows the standard two-stage scheme in RFT: supervised fine-tuning (SFT) with chain-of-thought (CoT) annotations, followed by online reinforcement learning (RL) to improve generalization. A central challenge to achieve this in the video domain lies in the scarcity of large-scale, high-quality video CoT datasets. We address this by building a fully automatic CoT curation pipeline. First, we devise a cognition-inspired prompting strategy to elicit a reasoning LLM to generate preliminary CoTs based solely on rich, structured, and literal representations of video content. Subsequently, these CoTs are revised by a visual-language model conditioned on the actual video, ensuring visual consistency and reducing visual hallucinations. This pipeline results in two new datasets -- VideoRFT-CoT-102K for SFT and VideoRFT-RL-310K for RL. To further strength the RL phase, we introduce a novel semantic-consistency reward that explicitly promotes the alignment between textual reasoning with visual evidence. This reward encourages the model to produce coherent, context-aware reasoning outputs grounded in visual input. Extensive experiments show that VideoRFT achieves state-of-the-art performance on six video reasoning benchmarks. Code, model, and data will be released.



Paperid:3947
Authors:Wen-Bo Du, Tian Qin, Tian-Zuo Wang, Zhi-Hua Zhou
Abstract:
Abstract:In machine learning, a critical class of decision-making problems involves *Avoiding Undesired Future* (AUF): given a predicted undesired outcome, how can one make decisions to prevent it? Recently, the \textit{rehearsal learning} framework has been proposed to address the AUF problem. While existing methods offer reliable decisions for single-round success, they fall short in long-term settings that involve coordinating multiple future outcomes, which better reflect real-world requirements. To address this challenge, we generalize the AUF objective to characterize a long-term decision target that incorporates cross-temporal relations among variables. As directly optimizing the *AUF probability* $P_{AUF}$ over this objective remains challenging, we derive an explicit expression for the objective and further propose a quadratic programming (QP) reformulation that transforms the intractable probabilistic AUF optimization into a tractable one. Under mild assumptions, we prove that QP solutions equals solutions to the AUF optimization, upon which we develop two novel rehearsal learning methods for long-term decision-making: (i) a *greedy* method that maximizes the single-round $P_{AUF}$ at each step, and (ii) a *far-sighted* method that accounts for future consequences in each decision, yielding a higher overall $P_{AUF}$ through a $T/(T+1)$ variance reduction in the AUF objective. We further establish an $\mathcal{O}(1/\sqrt{N})$ excess risk bound for decisions based on estimated parameters, ensuring reliable practical applicability with finite data. Experiments validate both theoretical results and empirical effectiveness of our approach.



Paperid:3948
Authors:zhentao he, Can Zhang, Ziheng Wu, Zhenghao Chen, Yufei Zhan, Yifan Li, Zhao Zhang, XIAN WANG, Minghui Qiu
Abstract:
Recent advancements in multimodal large language models (MLLMs) have enhanced document understanding by integrating textual and visual information. However, existing models exhibit incompleteness within their paradigm in real-world scenarios, particularly under visual degradation (e.g., blur, occlusion, low contrast). In such conditions, the current response paradigm often fails to adequately perceive visual degradation and ambiguity, leading to overreliance on linguistic priors or misaligned visual-textual reasoning. This difficulty in recognizing uncertainty frequently results in the generation of hallucinatory content, especially when a precise answer is not feasible. To better demonstrate and analyze this phenomenon and problem, we propose KIE-HVQA, the first benchmark dedicated to evaluating OCR hallucination in degraded document understanding. This dataset includes test samples spanning identity cards, invoices, and prescriptions, with simulated real-world degradations and pixel-level annotations for OCR reliability. This setup allows for evaluating models' capacity, under degraded input, to distinguish reliable visual information and answer accordingly, thereby highlighting the challenge of avoiding hallucination on uncertain data. To achieve vision-faithful reasoning and thereby avoid the aforementioned issues, we further introduce a Group Relative Policy Optimization (GRPO)-based framework featuring a novel reward mechanism. By incorporating a self-awareness of visual uncertainty and an analysis method that initiates refusal to answer to increase task difficulty within our supervised fine-tuning and reinforcement learning framework, we successfully mitigated hallucinations in ambiguous regions. Experiments on Qwen2.5-VL demonstrate that our 7B-parameter model achieves a ~28% absolute improvement in hallucination-free accuracy over GPT-4o on KIE-HVQA and there is no significant performance drop in standard tasks, highlighting both effectiveness and robustness. This work advances the development of reliable MLLMs for real-world document analysis by addressing critical challenges in visual-linguistic alignment under degradation.



Paperid:3949
Authors:Mojtaba Kolahdouzi, Hatice Gunes, Ali Etemad
Abstract:
We study whether and how the choice of optimization algorithm can impact group fairness in deep neural networks. Through stochastic differential equation analysis of optimization dynamics in an analytically tractable setup, we demonstrate that the choice of optimization algorithm indeed influences fairness outcomes, particularly under severe imbalance. Furthermore, we show that when comparing two categories of optimizers, adaptive methods and stochastic methods, RMSProp (from the adaptive category) has a higher likelihood of converging to fairer minima than SGD (from the stochastic category). Building on this insight, we derive two new theoretical guarantees showing that, under appropriate conditions, RMSProp exhibits fairer parameter updates and improved fairness in a single optimization step compared to SGD. We then validate these findings through extensive experiments on three publicly available datasets, namely CelebA, FairFace, and MS-COCO, across different tasks as facial expression recognition, gender classification, and multi-label classification, using various backbones. Considering multiple fairness definitions including equalized odds, equal opportunity, and demographic parity, adaptive optimizers like RMSProp and Adam consistently outperform SGD in terms of group fairness, while maintaining comparable predictive accuracy. Our results highlight the role of adaptive updates as a crucial yet overlooked mechanism for promoting fair outcomes. We release the source code at: https://anonymous.4open.science/r/Some-Optimizers-Are-More-Equal.



Paperid:3950
Authors:Yunuo Chen, Junli Cao, Vidit Goel, Sergei Korolev, Chenfanfu Jiang, Jian Ren, Sergey Tulyakov, Anil Kag
Abstract:
We present a novel video generation framework that integrates 3-dimensional geometry and dynamic awareness. To achieve this, we augment 2D videos with 3D point trajectories and align them in pixel space. The resulting 3D-aware video dataset, PointVid, is then used to fine-tune a latent diffusion model, enabling it to track 2D objects with 3D Cartesian coordinates. Building on this, we regularize the shape and motion of objects in the video to eliminate undesired artifacts, e.g., non-physical deformation. Consequently, we enhance the quality of generated RGB videos and alleviate common issues like object morphing, which are prevalent in current video models due to a lack of shape awareness. With our 3D augmentation and regularization, our model is capable of handling contact-rich scenarios such as task-oriented videos, where 3D information is essential for perceiving shape and motion of interacting solids. Our method can be seamlessly integrated into existing video diffusion models to improve their visual plausibility.



Paperid:3951
Authors:Gang Li, Ming Lin, Tomer Galanti, Zhengzhong Tu, Tianbao Yang
Abstract:
The recent success and openness of DeepSeek-R1 have brought widespread attention to Group Relative Policy Optimization (GRPO) as a reinforcement learning method for large reasoning models (LRMs). In this work, we analyze the GRPO objective under a binary reward setting and reveal an inherent limitation of question-level difficulty bias arising from its group relative advantage function. We also identify a connection between GRPO and traditional discriminative methods in supervised learning. Motivated by these insights, we introduce a newDiscriminative Constrained Optimization (DisCO)framework for reinforcing LRMs, grounded in the principle of discriminative learning: increasing the scores of positive answers while decreasing those of negative ones. The main differences between DisCO and GRPO and its recent variants are: (1) it replaces the group relative objective with a discriminative objective defined by a scoring function; (2) it abandons clipping-based surrogates in favor of non-clipping RL surrogate objectives used as scoring functions; (3) it employs a simple yet effective constrained optimization approach to enforce the KL divergence constraint, ensuring stable training. As a result, DisCO offers notable advantages over GRPO and its variants: (i) it completely eliminates difficulty bias by adopting discriminative objectives; (ii) it addresses the entropy instability in GRPO and its variants through the use of non-clipping scoring functions and a constrained optimization approach; (iii) it allows the incorporation of advanced discriminative learning techniques to address data imbalance, where a significant number of questions have more negative than positive generated answers during training. Our experiments on enhancing the mathematical reasoning capabilities of SFT-finetuned models show that DisCO significantly outperforms GRPO and its improved variants such as DAPO, achieving average gains of 7\% over GRPO and 6\% over DAPO across six benchmark tasks for an 1.5B model.



Paperid:3952
Authors:Pu Yang, Yunzhen Feng, Ziyuan Chen, Yuhang Wu, Zhuoyuan Li
Abstract:
Modern foundation models often undergo iterative ``bootstrapping'' in their post-training phase: a model generates synthetic data, an external verifier filters out low-quality samples, and the high-quality subset is used for further fine-tuning. Over multiple iterations, the model's performance improves—raising a crucial question: how should the total budget on generation and training be allocated across iterations to maximize final performance? In this work, we develop a theoretical framework to analyze budget allocation strategies. Specifically, we show that constant policies fail to converge with high probability, while increasing policies—particularly exponential growth policies—exhibit significant theoretical advantages. Experiments on image denoising with diffusion probabilistic models and math reasoning with large language models show that both exponential and polynomial growth policies consistently outperform constant policies, with exponential policies often providing more stable performance.



Paperid:3953
Authors:Nan Sun, LuYu Yuan, Han Fang, Yuxing Lu, Hefei Ling, Sijing Xie, Chengxin Zhao
Abstract:
Recent developments in DNN-based image watermarking techniques have achieved impressive results in protecting digital content. However, most existing methods are constrained to low-resolution images as they need to encode the entire image, leading to prohibitive memory and computational costs when applied to high-resolution images. Moreover, they lack robustness to distortions prevalent in large-image transmission, such as extreme scaling and random cropping. To address these issues, we propose a novel watermarking method based on implicit neural representations (INRs). Leveraging the properties of INRs, our method employs resolution-independent coordinate sampling mechanism to generate watermarks pixel-wise, achieving ultra-high resolution watermark generation with fixed and limited memory and computational resources. This design ensures strong robustness in watermark extraction, even under extreme cropping and scaling distortions. Additionally, we introduce a hierarchical multi-scale coordinate embedding and a low-rank watermark injection strategy to ensure high-quality watermark generation and robust decoding. Experimental results demonstrate that our method significantly outperforms existing schemes in terms of both robustness and computational efficiency while preserving high image quality. Our approach achieves an accuracy greater than 98\% in watermark extraction with only 0. 4\% of the image area in 2K images. These results highlight the effectiveness of our method, making it a promising solution for large-scale and high-resolution image watermarking applications.



Paperid:3954
Authors:Shuteng Wang, Christian Theobalt, Vladislav Golyanik
Abstract:
Abstract:Implicit representations with parametrized quantum circuits for gate-based quantum systems emerged as a promising paradigm for efficient learning, but still with many challenges. However, an optimal strategy for incorporating such circuits remains unclear, owing to limited understanding of different ways to leverage quantum-mechanical properties in representation learning. Existing works face problems such as: 1) reliance on heavy post-processing that overshadows quantum-mechanical effects, and 2) worse performance compared to classical modules. We introduce QVFs ($\textbf{Q}$uantum $\textbf{V}$isual $\textbf{F}$ields), a complete learnable quantum circuit without classical post-processing inspired by gate-based quantum system evolution. QVF integrates Hamiltonian-driven trainable neural operators with unitary quantum dynamics, regulating information flow through quantum conservation laws while employing projective measurement for feature extraction. Notably, our framework naturally inherits quantum-mechanical properties, including: (1) Fourier-spectral interpretability and (2) stable, reversible unitary evolution that ensures numerical stability—properties rigorously established in literature as critical for robust, bias-free learning. We evaluate QVF's properties, including expressiveness and robustness, with up to six qubits against existing methods on representing 2D images and 3D geometries. A benchmark study is performed against established quantum circuit architectures. Our results demonstrate that QVF, as a complete quantum circuit, can (1) efficiently learn high-fidelity multi-dimensional visual fields and (2) consistently outperform existing widely-used baselines across metrics such as learning speed, even at scale, therefore highlighting its practical potential.



Paperid:3955
Authors:Pengxiang Li, Zhi Gao, Bofei Zhang, Yapeng Mi, Xiaojian (Shawn) Ma, Chenrui Shi, Tao Yuan, Yuwei Wu, Yunde Jia, Song-Chun Zhu, Qing Li
Abstract:
Multimodal agents, which integrate a controller (e.g., a vision language model) with external tools, have demonstrated remarkable capabilities in tackling complex multimodal tasks.Existing approaches for training these agents, both supervised fine-tuning and reinforcement learning, depend on extensive human-annotated task-answer pairs and tool trajectories.However, for complex multimodal tasks, such annotations are prohibitively expensive or impractical to obtain.In this paper, we propose an iterative tool usage exploration method for multimodal agents without any pre-collected data, namely SPORT, via step-wise preference optimization to refine the trajectories of tool usage. Our method enables multimodal agents to autonomously discover effective tool usage strategies through self-exploration and optimization, eliminating the bottleneck of human annotation.SPORT has four iterative components: task synthesis, step sampling, step verification, and preference tuning.We first synthesize multimodal tasks using language models. Then, we introduce a novel trajectory exploration scheme, where step sampling and step verification are executed alternately to solve synthesized tasks.In step sampling, the agent tries different tools and obtains corresponding results. In step verification, we employ a verifier to provide AI feedback to construct step-wise preference data. The data is subsequently used to update the controller for tool usage through preference tuning, producing a SPORT agent.By interacting with real environments, the SPORT agent gradually evolves into a more refined and capable system.Evaluation in the GTA and GAIA benchmarks shows that the SPORT agent achieves 6.41% and 3.64% improvements, underscoring the generalization and effectiveness introduced by our method.



Paperid:3956
Authors:Gaojian Xiong, Yu Sun, Jianhua Liu, Jian Cui, Jianwei Liu
Abstract:
Abstract:While Large Language Models (LLMs) have gained remarkable success, they are consistently at risk of being stolen when deployed on untrusted edge devices. As a solution, TEE-based secure inference has been proposed to protect valuable model property. However, we identify a statistical vulnerability in existing protection methods, and furtherly compromise their security guarantees by proposed Model Stealing Attacks with Prior. To eliminate this vulnerability, LoRO is presented in this paper, which leverages dense mask to completely obfuscate parameters. LoRO includes two innovations: (1) Low Rank Mask, which uses low-rank factors to generate dense masks efficiently. The computing complexity in TEE is hence reduced by an exponential amount to achieve inference speed up, while providing robust model confidentiality. (2) Factors Multiplexing, which reuses several cornerstone factors to generate masks for all layers. Compared to one-mask-per-layer, the secure memory requirement is reduced from GB-level to tens of MB, hence avoiding the hundred-fold latency introduced by secure memory paging. Experimental results indicate that LoRO achieve a $0.94\times$ Model Stealing (MS) accuracy, while SOTA methods presents $3.37\times$ at least. The averaged inference latency of LoRO is only $1.49\times$, compared to the best $112\times$ of SOTA. Moreover, LoRO results no accuracy loss, and requires no re-training and structure modification. LoRO can solve the concerns regarding model thefts on edge devices in an efficient and secure manner, facilitating the wide edge application of LLMs.



Paperid:3957
Authors:Hadi Hosseini, Samarth Khanna
Abstract:
The growing interest in employing large language models (LLMs) for decision-making in social and economic contexts has raised questions about their potential to function as agents in these domains. A significant number of societal problems involve the distribution of resources, where fairness, along with economic efficiency, play a critical role in the desirability of outcomes. In this paper, we examine whether LLM responses adhere to fundamental fairness concepts such as equitability, envy-freeness, and Rawlsian maximin, and investigate their alignment with human preferences. We evaluate the performance of several LLMs, providing a comparative benchmark of their ability to reflect these measures. Our results demonstrate a lack of alignment between current LLM responses and human distributional preferences. Moreover, LLMs are unable to utilize money as a transferable resource to mitigate inequality. Nonetheless, we demonstrate a stark contrast when (some) LLMs are tasked with selecting from a predefined menu of options rather than generating one. In addition, we analyze the robustness of LLM responses to variations in semantic factors (e.g. intentions or personas) or non-semantic prompting changes (e.g. templates or orderings). Finally, we highlight potential strategies aimed at enhancing the alignment of LLM behavior with well-established fairness concepts.



Paperid:3958
Authors:Moritz Gögl, Yu Liu, Christopher Yau, Peter Watkinson, Tingting Zhu
Abstract:
Predicting heterogeneous treatment effects (HTEs) of continuous-valued interventions on survival, that is, time-to-event (TTE) outcomes, is crucial in various fields, notably in clinical decision making and in driving the advancement of next-generation clinical trials. However, while HTE estimation for continuous-valued (e.g, dosage-dependent) interventions and for TTE outcomes have been separately explored, their combined application remains largely overlooked in recent machine learning literature. We propose DoseSurv, a varying coefficient network designed for estimating the HTE of (multiple) continuous-valued as well as binary treatment options from TTE data. DoseSurv uses radial basis functions to model the continuity in dose-response relationships, and uses balancing representations to address covariate shifts arising in HTE estimation from observational TTE data. We present experiments in various treatment scenarios on both simulated and real-world data, demonstrating DoseSurv's superior performance over existing baseline models.



Paperid:3959
Authors:Jiawei Huang, Minming Li, Hu Ding
Abstract:
Deep learning models often struggle with distribution shifts between training and deployment environments. Distributionally Robust Optimization (DRO) offers a promising framework by optimizing worst-case performance over a set of candidate distributions, which is called as the \emph{uncertainty set}. However, the efficacy of DRO heavily depends on the design of uncertainty set, and existing methods often perform suboptimally due to inappropriate and inflexible uncertainty sets. In this work, we first propose a novel perspective that casts entropy-regularized Wasserstein DRO as a dynamic process of distributional exploration and semantic alignment, both driven by optimal transport (OT). This unified viewpoint yields two key new techniques: \emph{semantic calibration}, which bootstraps semantically meaningful transport costs via inverse OT, and \emph{adaptive refinement}, which adjusts uncertainty set using OT-driven feedback. Together, these components form an exploration-and-feedback system, where the transport costs and uncertainty set evolve jointly during training, enabling the model to better adapt to potential distribution shifts. Moreover, we provide an in-depth analysis on this adaptive process and prove the theoretical convergence guarantee. Finally, we present our experimental results across diverse distribution shift scenarios, which demonstrate that our approach significantly outperforms existing methods, achieving state-of-the-art robustness.



Paperid:3960
Authors:Omar Bennouna, Amine Bennouna, Saurabh Amin, Asuman Ozdaglar
Abstract:
We study the fundamental question of how informative a dataset is for solving a given decision-making task. In our setting, the dataset provides partial information about unknown parameters that influence task outcomes. Focusing on linear programs, we characterize when a dataset is sufficient to recover an optimal decision, given an uncertainty set on the cost vector. Our main contribution is a sharp geometric characterization that identifies the directions of the cost vector that matter for optimality, relative to the task constraints and uncertainty set.We further develop a practical algorithm that, for a given task, constructs a minimal or least-costly sufficient dataset. Our results reveal that small, well-chosen datasets can often fully determine optimal decisions---offering a principled foundation for task-aware data selection.



Paperid:3961
Authors:Wenzhe Ouyang, Qi Ye, Jinghua Wang, Zenglin Xu, Jiming Chen
Abstract:
Abstract:We introduce RFMPose, a novel generative framework for category-level 6D object pose estimation that learns deterministic pose trajectories through Riemannian Flow Matching (RFM). Existing discriminative approaches struggle with multi-hypothesis predictions (e.g., symmetry ambiguities) and often require specialized network architectures. RFMPose advances this paradigm through three key innovations:(1) Ensuring geometric consistency via geodesic interpolation on Riemannian manifolds combined with bi-invariant metric constraints;(2) Alleviating symmetry-induced ambiguities through Riemannian Optimal Transport for probability mass redistribution without ad-hoc design;(3) Enabling end-to-end likelihood estimation through Hutchinson trace approximation, thereby eliminating auxiliary model dependencies.Extensive experiments on the Omni6DPose demonstrate state-of-the-art performance of the proposed method, with significant improvements of $\textbf{+4.1}$ in $\mathrm{\textbf{IoU}_{25}}$ and $\textbf{+2.4}$ in $\textbf{5°2cm}$ metrics compared to prior generative approaches. Furthermore, the proposed RFM framework exhibits robust sim-to-real transfer capabilities and facilitates pose tracking extensions with minimal architectural adaptation.



Paperid:3962
Authors:Ran Xu, Yuchen Zhuang, Zihan Dong, Ruiyu Wang, Yue Yu, Joyce Ho, Linjun Zhang, Haoyu Wang, Wenqi Shi, Carl Yang
Abstract:
Retrieval-augmented generation (RAG) systems often struggle with complex reasoning tasks due to ineffective multi-hop retrieval and limited reasoning ability. We propose AceRAG, a cooperative self-play framework that trains a single large language model (LLM) to alternate between two roles: a decomposer that breaks down complex queries and a solver that integrates retrieved contexts for answer generation. AceRAG couples supervised fine-tuning on a diverse mixture of retrieval, reasoning, and decomposition tasks with reinforcement fine-tuning optimized for final answer accuracy, eliminating the need for intermediate annotations. Extensive experiments on three reasoning-intensive tasks across 10 datasets show that AceRAG outperforms state-of-the-art baselines, achieving an average exact match improvement of 7.6%. Remarkably, on document-level reasoning tasks, AceRAG-32B matches the performance of the giant DeepSeek-V3 model using less than 5% of its parameters. Even at smaller scales (1.5B and 8B), AceRAG often surpasses existing RAG models with up to 9x more parameters, highlighting its exceptional efficiency and effectiveness in tackling complex reasoning tasks.



Paperid:3963
Authors:Haoyu Zhang, Yinan Zhang, Chaolong Ying, Xiaoying Tang, Tianshu Yu
Abstract:
Multimodal Sentiment Analysis (MSA) aims to predict sentiment from diverse data types, such as video, audio, and language. Recent progress in Multimodal Large Language Models (MLLMs) have demonstrated impressive performance across various tasks. However, in MSA, the increase in computational costs does not always correspond to a significant improvement in performance, raising concerns about the cost-effectiveness of applying MLLMs to MSA. This paper introduces the MLLM-Guided Multimodal Sentiment Learning Framework (MMSLF). It improves the performance of task-specific MSA models by leveraging the generalized knowledge of MLLMs through a teacher-student framework, rather than directly using MLLMs for sentiment prediction. The proposed teacher, equipped with a powerful MLLM (e.g., GPT-4o-mini), guides the student model by generating context-aware prompts, facilitating better alignment and learning of multimodal sentiment features. Extensive experiments on the SIMS, MOSI, and MOSEI datasets demonstrate that our framework enables task-specific models to achieve state-of-the-art performance across most metrics. This also provides new insights into the application of general MLLMs for improving MSA.



Paperid:3964
Authors:Weisen Jiang, Sinno Pan
Abstract:
This paper introduces MetaDefense, a novel framework for defending against finetuning-based jailbreak attacks in large language models (LLMs). We observe that existing defense mechanisms fail to generalize to harmful queries disguised by unseen attack templates, despite LLMs being capable of distinguishing disguised harmful queries in the embedding space. Based on these insights, we propose a two-stage defense approach: (i) pre-generation defense that detects harmful queries before response generation begins, and (ii) mid-generation defense that monitors partial responses during generation to prevent outputting more harmful content. Our MetaDefense trains the LLM to predict the harmfulness of both queries and partial responses using specialized prompts, enabling early termination of potentially harmful interactions. Extensive experiments across multiple LLM architectures (LLaMA-2-7B, Qwen-2.5-3B-Instruct, and LLaMA-3.2-3B-Instruct) demonstrate that MetaDefense significantly outperforms existing defense mechanisms, achieving robust defense against harmful queries with seen and unseen attack templates while maintaining competitive performance on benign tasks.



Paperid:3965
Authors:Xiaowei Zhu, Yubing Ren, Fang Fang, Qingfeng Tan, Shi Wang, Yanan Cao
Abstract:
The rapid advancement of large language models (LLMs) has blurred the line between AI-generated and human-written text. This progress brings societal risks such as misinformation, authorship ambiguity, and intellectual property concerns, highlighting the urgent need for reliable AI-generated text detection methods. However, recent advances in generative language modeling have resulted in significant overlap between the feature distributions of human-written and AI-generated text, blurring classification boundaries and making accurate detection increasingly challenging. To address the above challenges, we propose a DNA-inspired perspective, leveraging a repair-based process to directly and interpretably capture the intrinsic differences between human-written and AI-generated text. Building on this perspective, we introduceDNA-DetectLLM, a zero-shot detection method for distinguishing AI-generated and human-written text. The method constructs an ideal AI-generated sequence for each input, iteratively repairs non-optimal tokens, and quantifies the cumulative repair effort as an interpretable detection signal. Empirical evaluations demonstrate that our method achieves state-of-the-art detection performance and exhibits strong robustness against various adversarial attacks and input lengths. Specifically, DNA-DetectLLM achieves relative improvements of5.55\%in AUROC and2.08\%in F1 score across multiple public benchmark datasets.



Paperid:3966
Authors:Aobo Li, Jinjian Wu, Yongxu Liu, Leida Li, Weisheng Dong
Abstract:
Blind Image Quality Assessment (BIQA) has advanced significantly through deep learning, but the scarcity of large-scale labeled datasets remains a challenge. While synthetic data offers a promising solution, models trained on existing synthetic datasets often show limited generalization ability. In this work, we make a key observation that representations learned from synthetic datasets often exhibit a discrete and clustered pattern that hinders regression performance: features of high-quality images cluster around reference images, while those of low-quality images cluster based on distortion types. Our analysis reveals that this issue stems from the distribution of synthetic data rather than model architecture. Consequently, we introduce a novel framework SynDR-IQA, which reshapes synthetic data distribution to enhance BIQA generalization. Based on theoretical derivations of sample diversity and redundancy's impact on generalization error, SynDR-IQA employs two strategies: distribution-aware diverse content upsampling, which enhances visual diversity while preserving content distribution, and density-aware redundant cluster downsampling, which balances samples by reducing the density of densely clustered areas. Extensive experiments across three cross-dataset settings (synthetic-to-authentic, synthetic-to-algorithmic, and synthetic-to-synthetic) demonstrate the effectiveness of our method. The source code will be publicly available.



Paperid:3967
Authors:Zhao Jin, Rong-Cheng Tu, Jingyi Liao, Wenhao Sun, Xiao Luo, Shunyu Liu, Dacheng Tao
Abstract:
Abstract:3D Visual Grounding (3DVG) aims to localize target objects within a 3D scene based on natural language queries. To alleviate the reliance on costly 3D training data, recent studies have explored zero-shot 3DVG by leveraging the extensive knowledge and powerful reasoning capabilities of pre-trained LLMs and VLMs. However, existing paradigms tend to emphasize either spatial (3D-based) or semantic (2D-based) understanding, limiting their effectiveness in complex real-world applications. In this work, we introduce SPAZER — a VLM-driven agent that combines both modalities in a progressive reasoning framework. It first holistically analyzes the scene and produces a 3D rendering from the optimal viewpoint. Based on this, anchor-guided candidate screening is conducted to perform a coarse-level localization of potential objects. Furthermore, leveraging retrieved relevant 2D camera images, 3D-2D joint decision-making is efficiently performed to determine the best-matching object. By bridging spatial and semantic reasoning neural streams, SPAZER achieves robust zero-shot grounding without training on 3D-labeled data. Extensive experiments on ScanRefer and Nr3D benchmarks demonstrate that SPAZER significantly outperforms previous state-of-the-art zero-shot methods, achieving notable gains of $\mathbf{9.0\}$% and $\mathbf{10.9\}$% in accuracy.



Paperid:3968
Authors:Yi-Lun Wu, Bo-Kai Ruan, Chiang Tseng, Hong-Han Shuai
Abstract:
Direct preference optimization (DPO) methods have shown strong potential in aligning text-to-image diffusion models with human preferences by training on paired comparisons. These methods improve training stability by avoiding the REINFORCE algorithm but still struggle with challenges such as accurately estimating image probabilities due to the non-linear nature of the sigmoid function and the limited diversity of offline datasets. In this paper, we introduce Diffusion Denoising Ranking Optimization (Diffusion-DRO), a new preference learning framework grounded in inverse reinforcement learning. Diffusion-DRO removes the dependency on a reward model by casting preference learning as a ranking problem, thereby simplifying the training objective into a denoising formulation and overcoming the non-linear estimation issues found in prior methods. Moreover, Diffusion-DRO uniquely integrates offline expert demonstrations with online policy-generated negative samples, enabling it to effectively capture human preferences while addressing the limitations of offline data. Comprehensive experiments show that Diffusion-DRO delivers improved generation quality across a range of challenging and unseen prompts, outperforming state-of-the-art baselines in both both quantitative metrics and user studies.



Paperid:3969
Authors:Hancheng Ye, Zhengqi Gao, Mingyuan Ma, Qinsi Wang, Yuzhe Fu, Ming-Yu Chung, Yueqian Lin, Zhijian Liu, Jianyi Zhang, Danyang Zhuo, Yiran Chen
Abstract:
Abstract:Multi-agent large language model (LLM) systems are increasingly adopted for complex language processing tasks that require communication and coordination among agents. However, these systems often suffer substantial overhead from repeated reprocessing of overlapping contexts across agents. In typical pipelines, once an agent receives a message from its predecessor, the full context—including prior turns—must be reprocessed from scratch, leading to inefficient processing. While key-value (KV) caching is an effective solution for avoiding redundant computation in single-agent settings where prefixes remain unchanged, it cannot be directly reused in multi-agent scenarios due to diverging prefixes introduced by agent-specific context extensions. We identify that the core challenge lies in the *offset variance* of KV-caches across agents. To address this, we propose **KVComm**, a training-free framework that enables efficient prefilling in multi-agent inference by reusing KV-caches and aligning cache offsets of overlapping contexts under diverse prefix contexts. KVComm estimates and adjusts KV-caches for shared content by referencing a pool of cached examples—termed *anchors*—that store observed cache deviations under varying prefixes. The anchor pool is maintained and updated online, allowing dynamic adaptation to distinct user requests and context structures. KVComm achieves an average 3.7$\times$ speedup in prefill latency and over 70\% reuse rate across diverse multi-agent workloads, including retrieval-augmented generation, math reasoning, and collaborative coding tasks, all without degradation in output quality. Particularly, when each fully-connected agent receives 1K input tokens with 512 prefix tokens and 512 output tokens under a five-agent setting, KVComm achieves up to 13.7$\times$ speedup compared to the standard prefill pipeline, reducing TTFT from $\sim$450ms to $\sim$33ms.



Paperid:3970
Authors:Zhenyuan Qin, Xincheng Shuai, Henghui Ding
Abstract:
Controllable image generation has attracted increasing attention in recent years, enabling users to manipulate visual content such as identity and style. However, achieving simultaneous control over the 9D poses (location, size, and orientation) of multiple objects remains an open challenge. Despite recent progress, existing methods often suffer from limited controllability and degraded quality, falling short of comprehensive multi-object 9D pose control. To address these limitations, we proposeSceneDesigner, a method for accurate and flexible multi-object 9-DoF pose manipulation.SceneDesignerincorporates a branched network to the pre-trained base model and leverages a new representation,CNOCS map, which encodes 9D pose information from the camera view. This representation exhibits strong geometric interpretation properties, leading to more efficient and stable training. To support training, we construct a new dataset,ObjectPose9D, which aggregates images from diverse sources along with 9D pose annotations. To further address data imbalance issues, particularly performance degradation on low-frequency poses, we introduce a two-stage training strategy with reinforcement learning, where the second stage fine-tunes the model using a reward-based objective on rebalanced data. At inference time, we proposeDisentangled Object Sampling, a technique that mitigates insufficient object generation and concept confusion in complex multi-object scenes. Moreover, by integrating user-specific personalization weights,SceneDesignerenables customized pose control for reference subjects. Extensive qualitative and quantitative experiments demonstrate thatSceneDesignersignificantly outperforms existing approaches in both controllability and quality.



Paperid:3971
Authors:Zican Dong, Han Peng, Peiyu Liu, Xin Zhao, Dong Wu, Feng Xiao, Zhifeng Wang
Abstract:
Abstract:Mixture-of-Experts (MoE) models achieve a favorable trade-off between performance and inference efficiency by activating only a subset of experts. However, the memory overhead of storing all experts remains a major limitation, especially in large-scale MoE models such as DeepSeek-R1 (671B). In this study, we investigate domain specialization and expert redundancy in large-scale MoE models and uncover a consistent behavior we term *few-shot expert localization*, with only a few in-domain demonstrations, the model consistently activates a sparse and stable subset of experts on tasks within the same domain.Building on this observation, we propose a simple yet effective pruning framework, **EASY-EP**, that leverages a few domain-specific demonstrations to identify and retain only the most relevant experts.EASY-EP comprises two key components: **output-aware expert importance assessment** and **expert-level token contribution estimation**. The former evaluates the importance of each expert for the current token by considering the gating scores and L2 norm of the outputs of activated experts, while the latter assesses the contribution of tokens based on representation similarities before and after routed experts.Experiments on DeepSeek-R1 and DeepSeek-V3-0324 show that our method can achieve comparable performances and $2.99\times$ throughput under the same memory budget with full model with only half the experts.



Paperid:3972
Authors:Yassine ABBAHADDOU, Amine Aboussalah
Abstract:
Generative models such as Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and diffusion models often fail to capture the full diversity of their training data, leading to mode collapse. While this issue is well-explored in image generation, it remains underinvestigated for time series data. We introduce a new definition of mode collapse specific to time series and propose a new geometry-aware metric, DMD-GEN, to quantify its severity. Our metric leverages Dynamic Mode Decomposition (DMD), a data-driven technique for identifying coherent spatiotemporal patterns through spectral analysis, and employs optimal transport between DMD eigenvectors to assess discrepancies in the underlying dynamics of the original and generated data. The geometry-aware nature of our method comes from modeling DMD modes as points on a Grassmann manifold and comparing them using Wasserstein distances computed via principal angles. When using mini-batch evaluation, DMD-GEN compares the subspaces spanned by the dominant modes of original and generated time series through optimal transport of their corresponding subspaces, enabling a principled geometric comparison. DMD-GEN is efficient in practice: it is used only during evaluation, supports mini-batch approximations, and is highly parallelizable. It not only quantifies the preservation of essential dynamic characteristics but also provides interpretability by highlighting which modes are poorly captured in the generated data. We validate DMD-GEN on both synthetic and real-world datasets using TimeGAN, TimeVAE, and DiffusionTS. The results show that DMD-GEN aligns well with traditional metrics while, for the first time, offering a principled definition of mode collapse for time series.



Paperid:3973
Authors:Shuo Wang, Yongcai Wang, Wanting Li, Xudong Cai, Yucheng Wang, Maiyue Chen, kaihui.wang, Zhizhong Su, Deying Li, Zhaoxin Fan
Abstract:
Vision-Language Navigation is a critical task for developing embodied agents that can follow natural language instructions to navigate in complex real-world environments. Recent advances by finetuning large pretrained models have significantly improved generalization and instruction grounding compared to traditional approaches. However, the role of reasoning strategies in navigation—an action-centric, long-horizon task—remains underexplored, despite Chain-of-Thought reasoning's demonstrated success in static tasks like question answering and visual reasoning. To address this gap, we conduct the first systematic evaluation of reasoning strategies for VLN, including No-Think (direct action prediction), Pre-Think (reason before action), and Post-Think (reason after action). Surprisingly, our findings reveal the Inference-time Reasoning Collaps issue, where inference-time reasoning degrades navigation accuracy, highlighting the challenges of integrating reasoning into VLN. Based on this insight, we propose Aux-Think, a framework that trains models to internalize structured reasoning patterns through CoT supervision during training, while preserving No-Think inference for efficient action prediction. To support this framework, we release R2R-CoT-320k, a large-scale Chain-of-Thought annotated dataset. Empirically, Aux-Think significantly reduces training effort without compromising performance.



Paperid:3974
Authors:Md Mobasshir Arshed Naved, Wenbo Xie, Wojciech Szpankowski, Ananth Grama
Abstract:
We propose a novel gradient-based framework for learning parameterized quantum circuits (PQCs) in the presence of Pauli noise in gate operation. The key innovation in our framework is the simultaneous optimization of model parameters and learning of an inverse noise channel, specifically designed to mitigate Pauli noise. Our parametrized inverse noise model utilizes the Pauli-Lindblad equation and relies on the principle underlying the Probabilistic Error Cancellation (PEC) protocol to learn an effective and scalable mechanism for noise mitigation. In contrast to conventional approaches that apply predetermined inverse noise models during execution, our method systematically mitigates Pauli noise by dynamically updating the inverse noise parameters in conjunction with the model parameters, facilitating task-specific noise adaptation throughout the learning process. We employ proximal stochastic gradient descent (proximal SGD) to ensure that updates are bounded within a feasible range to ensure stability. This approach allows the model to converge efficiently to a stationary point, balancing the trade-off between noise mitigation and computational overhead, resulting in a highly adaptable quantum model that performs robustly in noisy quantum environments. Our framework is well-suited to near-term quantum devices in the noisy intermediate-scale quantum (NISQ) era, where noise is a significant challenge.



Paperid:3975
Authors:Runa Eschenhagen, Aaron Defazio, Tsung-Hsien Lee, Richard Turner, Hao-Jun Shi
Abstract:
The recent success of Shampoo in the AlgoPerf contest has sparked renewed interest in non-diagonal Kronecker-factorization-based optimization algorithms for training neural networks. Despite its success, Shampoo relies heavily on several heuristics such as learning rate grafting and stale preconditioning to achieve performance at-scale. These heuristics increase algorithmic complexity, necessitate further hyperparameter tuning, and lack theoretical justification. This paper investigates these heuristics from the angle of Frobenius norm approximation to full-matrix Adam and decouples the preconditioner's eigenvalues and eigenbasis updates. We show that grafting from Adam mitigates the staleness and mis-scaling of the preconditioner'seigenvaluesand how correcting the eigenvalues directly can eliminate the need for learning rate grafting. To manage the error induced by infrequenteigenbasiscomputations, we propose an adaptive criterion for determining the eigenbasis computation frequency motivated by terminating a warm-started QR algorithm, translating to improved efficiency. These practical techniques offer a principled angle towards removing Shampoo's heuristics and developing improved Kronecker-factorization-based training algorithms.



Paperid:3976
Authors:Alan J.X. Guo, Sihan Sun, Xiang Wei, Mengyi Wei, Xin Chen
Abstract:
With the emergence of new storage and communication methods, the insertion, deletion, and substitution (IDS) channel has attracted considerable attention. However, many topics on the IDS channel and the associated Levenshtein distance remain open, making the invention of a novel IDS-correcting code a hard task. Furthermore, current studies on single-IDS-correcting code misalign with the requirements of applications which necessitates the correcting of multiple errors. Compromise solutions have involved shortening codewords to reduce the chance of multiple errors. However, the code rates of existing codes are poor at short lengths, diminishing the overall storage density. In this study, a novel method is introduced for designing high-code-rate single-IDS-correcting codewords through deep Levenshtein distance embedding. A deep learning model is utilized to project the sequences into embedding vectors that preserve the Levenshtein distances between the original sequences. This embedding space serves as a proxy for the complex Levenshtein domain, within which algorithms for codeword search and segment correcting is developed. While the concept underpinning this approach is straightforward, it bypasses the mathematical challenges typically encountered in code design. The proposed method results in a code rate that outperforms existing combinatorial solutions, particularly for designing short-length codewords.



Paperid:3977
Authors:Xiaoang Xu, Shuo Wang, Xu Han, Zhenghao Liu, Huijia Wu, Peipei Li, Zhiyuan Liu, Maosong Sun, Zhaofeng He
Abstract:
Abstract:Large Reasoning Models (LRMs) achieve superior performance by extending the thought length. However, a lengthy thinking trajectory leads to reduced efficiency. Most of the existing methods are stuck in the assumption of overthinking and attempt to reason efficiently by compressing the Chain-of-Thought, but this often leads to performance degradation. To address this problem, we introduce A*-Thought, an efficient tree search-based unified framework designed to identify and isolate the most essential thoughts from the extensive reasoning chains produced by these models. It formulates the reasoning process of LRMs as a search tree, where each node represents a reasoning span in the giant reasoning space. By combining the A* search algorithm with a cost function specific to the reasoning path, it can efficiently compress the chain of thought and determine a reasoning path with high information density and low cost. In addition, we also propose a bidirectional importance estimation mechanism, which further refines this search process and enhances its efficiency beyond uniform sampling. Extensive experiments on several advanced math tasks show that A*-Thought effectively balances performance and efficiency over a huge search space. Specifically, A*-Thought can improve the performance of QwQ-32B by 2.39$\times$ with low-budget and reduce the length of the output token by nearly 50\% with high-budget. The proposed method is also compatible with several other LRMs, demonstrating its generalization capability.



Paperid:3978
Authors:Ruizhe Zheng, Lingyan Mao, DINGDING HAN, Tian Luo, Yi Wang, Jing Ding, Yuguo Yu
Abstract:
Abstract:Precise, generalizable subject-agnostic seizure prediction (SASP) remains a fundamental challenge due to the intrinsic complexity and significant spectral variability of electrophysiologial signals across individuals and recording modalities. We propose \model{FAPEX}, a novel architecture that introduces a learnable \emph{fractional neural frame operator} (FrNFO) for adaptive time–frequency decomposition. Unlike conventional models that exhibit spectral bias toward low frequencies, our FrNFO employs fractional-order convolutions to capture both high and low-frequency dynamics, achieving approximately $10\%$ improvement in F1-score and sensitivity over state-of-the-art baselines. The FrNFO enables the extraction of \emph{instantaneous phase and amplitude representations} that are particularly informative for preictal biomarker discovery and enhance out-of-distribution generalization. \model{FAPEX} further integrates structural state-space modeling and channelwise attention, allowing it to handle heterogeneous electrode montages. Evaluated across 12 benchmarks spanning species (human, rat, dog, macaque) and modalities (Scalp‑EEG, SEEG, ECoG, LFP), \model{FAPEX} consistently outperforms 23 supervised and 10 self-supervised baselines under nested cross-validation, with gains of up to $15\%$ in sensitivity on complex cross-domain scenarios. It further demonstrates superior performance in several external validation cohorts. To our knowledge, these establish \model{FAPEX} as the first epilepsy model to show consistent superiority in SASP, offering a promising solution for discovering epileptic biomarker evidence supporting the existence of a distinct and identifiable preictal state for and clinical translation.



Paperid:3979
Authors:Wei Shen, Guanlin Liu, Yu Yue, Ruofei Zhu, Qingping Yang, Chao Xin, Lin Yan
Abstract:
Reinforcement Learning from Human Feedback (RLHF) is essential for aligning large language models (LLMs) with human preferences and values. While recent research has primarily focused on algorithmic advancements—such as reducing computational overhead or strengthening reward models to mitigate reward hacking—the critical role of prompt-data construction and its scalability has received comparatively less attention. In this paper, we address this gap by systematically exploring data-driven bottlenecks that currently hinder RLHF performance scaling, focusing specifically on the challenges posed by reward hacking and decreasing response diversity.To mitigate reward hacking, we introduce a hybrid reward system combining reasoning task verifiers (RTV) and a generative reward model (GenRM). This approach not only exhibits enhanced resistance to reward hacking, but also enables accurate assessment of responses against clearly defined ground-truth solutions. Additionally, in order to ensure response diversity and enhance learning effectiveness, we propose a novel prompt-selection method named \textbf{Pre-PPO}, explicitly identifying training prompts that are inherently challenging and thus less prone to reward hacking. Furthermore, we find that \textbf{prioritizing mathematical and coding tasks during the early phases of RLHF training} significantly boosts performance, given that these tasks naturally encode fine-grained response distinctions and possess clearly defined ground truths.Through comprehensive experiments conducted across two model sizes, we validate the effectiveness and scalability of our proposed methods. Results show that RTV exhibits the strongest resistance to reward hacking, followed by GenRM with ground truth, and finally GenRM relying on SFT Best-of-N responses. Moreover, our proposed strategies enable the model to rapidly capture subtle task-specific distinctions, leading to substantial improvements in overall RLHF performance. This work underscores the importance of careful data construction and provides practical methodologies to overcome critical performance barriers in RLHF.



Paperid:3980
Authors:Seonghoon Yu, Dongjun Nam, Dina Katabi, Jeany Son
Abstract:
Abstract:Knowledge Distillation (KD) aims to train a lightweight student model by transferring knowledge from a large, high-capacity teacher.Recent studies have shown that leveraging diverse teacher perspectives can significantly improve distillation performance; however, achieving such diversity typically requires multiple teacher networks, leading to high computational costs. In this work, we propose a novel cost-efficient knowledge augmentation method for KD that generates diverse multi-views by attaching multiple branches to a single teacher. To ensure meaningful semantic variation across multi-views, we introduce two angular diversity objectives: 1) $\textit{constrained inter-angle diversify loss}$, which maximizes angles between augmented views while preserving proximity to the original teacher output, and 2) $\textit{intra-angle diversify loss}$, which encourages an even distribution of views around the original output. The ensembled knowledge from these angularly diverse views, along with the original teacher, is distilled into the student. We further theoretically demonstrate that our objectives increase the diversity among ensemble members and thereby reduce the upper bound of the ensemble's expected loss, leading to more effective distillation. Experimental results show that our method surpasses an existing knowledge augmentation method across diverse configurations. Moreover, the proposed method is compatible with other KD frameworks in a plug-and-play fashion, providing consistent improvements in generalization performance.



Paperid:3981
Authors:Chenbei Lu, Zaiwei Chen, Tongxin Li, Chenye Wu, Adam Wierman
Abstract:
Traditional reinforcement learning (RL) assumes the agents make decisions based on Markov decision processes (MDPs) with one-step transition models. In many real-world applications, such as energy management and stock investment, agents can access multi-step predictions of future states, which provide additional advantages for decision making. However, multi-step predictions are inherently high-dimensional: naively embedding these predictions into an MDP leads to an exponential blow-up in state space and the curse of dimensionality. Moreover, existing RL theory provides few tools to analyze prediction-augmented MDPs, as it typically works on one-step transition kernels and cannot accommodate multi-step predictions with errors or partial action-coverage. We address these challenges with three key innovations: First, we propose the \emph{Bayesian value function} to characterize the optimal prediction-aware policy tractably. Second, we develop a novel \emph{Bellman–Jensen Gap} analysis on the Bayesian value function, which enables characterizing the value of imperfect predictions. Third, we introduce BOLA (Bayesian Offline Learning with Online Adaptation), a two-stage model-based RL algorithm that separates offline Bayesian value learning from lightweight online adaptation to real-time predictions. We prove that BOLA remains sample-efficient even under imperfect predictions. We validate our theory and algorithm on synthetic MDPs and a real-world wind energy storage control problem.



Paperid:3982
Authors:Xudong Li, Mengdan Zhang, Peixian Chen, Xiawu Zheng, Yan Zhang, Jingyuan Zheng, Yunhang Shen, Ke Li, Chaoyou Fu, Xing Sun, Rongrong Ji
Abstract:
Multi-modal Large Language Models (MLLMs) excel at single-image tasks but struggle with multi-image understanding due to cross-modal misalignment, leading to hallucinations (context omission, conflation, and misinterpretation). Existing methods using Direct Preference Optimization (DPO) constrain optimization to a solitary image reference within the input sequence, neglecting holistic context modeling. We propose \textbf{C}ontext-to-\textbf{C}ue \textbf{D}irect \textbf{P}reference \textbf{O}ptimization~\textbf{(CcDPO)}, a multi-level preference optimization framework that enhances per-image perception in multi-image settings by zooming into visual clues—from sequential context to local details. It features: (i) \textit{Context-Level Optimization} : Re-evaluates cognitive biases underlying MLLMs' multi-image context comprehension and integrates a spectrum of low-cost global sequence preferences for bias mitigation. (ii) \textit{Needle-Level Optimization} : Directs attention to fine-grained visual details through region-targeted visual prompts and multimodal preference supervision. To support scalable optimization, we also construct \textbf{MultiScope-42k}, an automatically generated dataset with high-quality multi-level preference pairs. Experiments show that CcDPO significantly reduces hallucinations and yields consistent performance gains across general multi-image vision-language tasks.



Paperid:3983
Authors:Haifan Gong, Xuanye Zhang, Ruifei Zhang, Yun Su, Zhuo Li, Yuhao Du, Xiang Wan, Anningzhe Gao, Haofeng Li
Abstract:
Recent advances in artificial intelligence have significantly impacted image retrieval tasks, yet Patent-Product Image Retrieval (PPIR) has received limited attention. PPIR, which retrieves patent images based on product images to identify potential infringements, presents unique challenges: (1) both product and patent images often contain numerous categories of artificial objects, but models pre-trained on standard datasets exhibit limited discriminative power to recognize some of those unseen objects; and (2) the significant domain gap between binary patent line drawings and colorful RGB product images further complicates similarity comparisons for product-patent pairs. To address these challenges, we formulate it as an open-set image retrieval task and introduce a comprehensive Patent-Product Image Retrieval Dataset (PPIRD) including a test set with 439 product-patent pairs, a retrieval pool of 727,921 patents, and an unlabeled pre-training set of 3,799,695 images. We further propose a novel Intermediate Domain Alignment and Morphology Analogy (IDAMA) strategy. IDAMA maps both image types to an intermediate sketch domain using edge detection to minimize the domain discrepancy, and employs a Morphology Analogy Filter to select discriminative patent images based on visual features via analogical reasoning. Extensive experiments on PPIRD demonstrate that IDAMA significantly outperforms baseline methods (+7.58 mAR) and offers valuable insights into domain mapping and representation learning for PPIR.



Paperid:3984
Authors:Junfei Zhou, Penglin Dai, Quanmin Wei, Bingyi Liu, Xiao Wu, Jianping Wang
Abstract:
Multi-agent collaboration enhances the perception capabilities of individual agents through information sharing. However, in real-world applications, differences in sensors and models across heterogeneous agents inevitably lead to domain gaps during collaboration. Existing approaches based on adaptation and reconstruction fail to supportpragmatic heterogeneous collaborationdue to two key limitations: (1) retraining the encoder or core modules of agents disrupts the consistency of the semantic space of each agent; and (2) accommodating new agents incurs high computational costs, limiting scalability. To address these challenges, we present a novelGenerativeCommunication mechanism (GenComm) that facilitates seamless perception across heterogeneous multi-agent systems through feature generation, without altering the original network, and employs lightweight numerical alignment of spatial information to efficiently integrate new agents at minimal cost. Specifically, a tailored Deformable Message Extractor is used to extract spatial information for each collaborator, which is then transmitted in place of intermediate features. The Spatial-Aware Feature Generator, utilizing a conditional diffusion model, generates features aligned with the ego agent’s semantic space while preserving the spatial information of the collaborators. These features are further refined by a Channel Enhancer before fusion. Experiments conducted on the OPV2V-H and DAIR-V2X datasets demonstrate that GenComm outperforms existing state-of-the-art methods, achieving an 81\% reduction in both computational cost and parameter count when incorporating new agents.



Paperid:3985
Authors:Zhisheng Zhang, Derui Wang, Yifan Mi, Zhiyong Wu, Jie Gao, Yuxin Cao, Kai Ye, Minhui Xue, Jie Hao
Abstract:
Recent advancements in speech synthesis technology have significantly enriched our daily lives, with high-quality and human-like audio widely adopted across real-world applications. However, malicious exploitation like voice-cloning fraud poses severe security risks. Existing defense techniques struggle to address the production large language model (LLM)-based speech synthesis. While previous studies have considered the protection for fine-tuning synthesizers, they assume manually annotated transcripts. Given the labor intensity of manual annotation, end-to-end systems leveraging automatic speech recognition (ASR) to generate transcripts are becoming increasingly prevalent, e.g., voice cloning via commercial APIs. Therefore, this end-to-end speech synthesis also requires new security mechanisms. To tackle these challenges, we propose E2E-VGuard, a proactive defense framework for two emerging threats: (1) production LLM-based speech synthesis, and (2) the novel attack arising from ASR-driven end-to-end fine-tuning. Specifically, we employ the encoder ensemble with a feature extractor to protect timbre through feature loss optimization, while ASR-targeted adversarial examples disrupt pronunciation. Moreover, we incorporate the psychoacoustic model to ensure perturbative imperceptibility. For a comprehensive evaluation, we test 13 open-source synthesizers and 3 commercial APIs across Chinese and English datasets, confirming E2E-VGuard's effectiveness in timbre preservation and pronunciation protection. Real-world deployment validation is also conducted. Our anonymous demo and code are available at https://e2e-vguard.github.io/.



Paperid:3986
Authors:Haozhen Zhang, Tao Feng, Jiaxuan You
Abstract:
The rapid emergence of diverse large language models (LLMs) has spurred the development of LLM routers that assign user queries to the most suitable model. However, existing LLM routers typically perform a single-round, one-to-one mapping (\textit{i.e.}, assigning each query to a single model in isolation), which limits their capability to tackle complex tasks that demand the complementary strengths of multiple LLMs. In this paper, we present \textbf{Router-R1}, an reinforcement learning (RL)-based framework that formulates multi-LLM routing and aggregation as a sequential decision process. Router-R1 instantiates the router itself as a capable LLM, leveraging its reasoning ability to interleave "think" actions (internal deliberation) with "route" actions (dynamic model invocation), and integrates each response into its evolving context. To guide learning, we employ a lightweight rule-based reward comprising format rewards, final outcome rewards, and a novel cost reward for performance and cost trade-off optimization, opening a pathway toward optimizing performance-cost tradeoffs via RL. Router-R1 also conditions only on simple model descriptors such as pricing, latency, and example performance, enabling strong generalization to unseen model selection. Experiments on seven general and multi-hop QA benchmarks show that Router-R1 outperforms over several strong baselines, achieving superior performance while maintaining robust generalization and cost management.



Paperid:3987
Authors:Yining Hong, Rui Sun, Bingxuan Li, Xingcheng Yao, Maxine Wu, Alexander Chien, Da Yin, Ying Nian Wu, Zhecan Wang, Kai-Wei Chang
Abstract:
AI agents today are mostly siloed — they either retrieve and reason over vast amount of digital information and knowledge obtained online; or interact with the physical world through embodied perception, planning and action — but rarely both. This separation limits their ability to solve tasks that require integrated physical and digital intelligence, such as cooking from online recipes, navigating with dynamic map data, or interpreting real-world landmarks using web knowledge. We introduce \textsc{Embodied Web Agents}, a novel paradigm for AI agents that fluidly bridge embodiment and web-scale reasoning. To operationalize this concept, we first develop the \textsc{Embodied Web Agents} task environments, a unified simulation platform that integrates realistic 3D indoor and outdoor environments with functional web interfaces. Building upon this platform, we construct and release the \textsc{Embodied Web Agents} Benchmark, which encompasses a diverse suite of tasks including cooking, navigation, shopping, tourism, and geolocation — all requiring coordinated reasoning across physical and digital realms for systematic assessment of cross-domain intelligence. Experimental results reveal significant performance gaps between state-of-the-art AI systems and human capabilities, establishing both challenges and opportunities at the intersection of embodied cognition and web-scale knowledge access.



Paperid:3988
Authors:Ziyang Ma, Yinghao Ma, Yanqiao Zhu, Chen Yang, Yi-Wen Chao, Ruiyang Xu, Wenxi Chen, Yuanzhe Chen, Zhuo Chen, Jian Cong, Kai Li, Keliang Li, Siyou Li, Xinfeng Li, Xiquan Li, Zheng Lian, Yuzhe Liang, Minghao Liu, Zhikang Niu, Tianrui Wang, Wang Yuping, Yuxuan Wang, Yihao Wu, Guanrou Yang, Jianwei Yu, Ruibin Yuan, Zhisheng Zheng, Ziya Zhou, Haina Zhu, Wei Xue, Emmanouil Benetos, Kai Yu, Eng-Siong Chng, Xie Chen
Abstract:
We introduce MMAR, a new benchmark designed to evaluate the deep reasoning capabilities of Audio-Language Models (ALMs) across massive multi-disciplinary tasks. MMAR comprises 1,000 meticulously curated audio-question-answer triplets, collected from real-world internet videos and refined through iterative error corrections and quality checks to ensure high quality. Unlike existing benchmarks that are limited to specific domains of sound, music, or speech, MMAR extends them to a broad spectrum of real-world audio scenarios, including mixed-modality combinations of sound, music, and speech. Each question in MMAR is hierarchically categorized across four reasoning layers: Signal, Perception, Semantic, and Cultural, with additional sub-categories within each layer to reflect task diversity and complexity. To further foster research in this area, we annotate every question with a Chain-of-Thought (CoT) rationale to promote future advancements in audio reasoning. Each item in the benchmark demands multi-step deep reasoning beyond surface-level understanding. Moreover, a part of the questions requires graduate-level perceptual and domain-specific knowledge, elevating the benchmark's difficulty and depth. We evaluate MMAR using a broad set of models, including Large Audio-Language Models (LALMs), Large Audio Reasoning Models (LARMs), Omni Language Models (OLMs), Large Language Models (LLMs), and Large Reasoning Models (LRMs), with audio caption inputs. The performance of these models on MMAR highlights the benchmark's challenging nature, and our analysis further reveals critical limitations of understanding and reasoning capabilities among current models. These findings underscore the urgent need for greater research attention in audio-language reasoning, including both data and algorithm innovation. We hope MMAR will serve as a catalyst for future advances in this important but little-explored area.



Paperid:3989
Authors:rongchao zhang, Yongzhi Cao, Hanpin Wang, Yu Huang
Abstract:
Molecular chemical space projecting (MCSP) is a critical technique in de novo molecular design, which aims to rectify molecules without synthesizability guarantee by converting them into synthetic postfix notations. However, the vast synthesizable chemical space and the discrete data modalities involved pose significant challenges to postfix notation conversion benchmarking, and existing MCSP frameworks suffering from error accumulation. In this paper, we exploit conditional probability transitions in discrete state space and introduce MoleBridge, a deep generative model built on the Markov bridge approach for designing postfix notations of molecular synthesis pathways. MoleBridge consists of two iterative optimizations: i) semi-autoregressive extending of reaction tokens from molecular graphs, and ii) generation of discrete reaction postfix notations from the entire approximate token sequence while allowing the latter to backward jump for error rectification. For the challenging second iteration, which demands sensitivity to incorrect generative probability paths within intricate chemical spaces, we employ a thinking and denoising separation approach to denoise when an error occurs at a certain step. Endowed by this, our uniquely modulated generative perspective allows the model to retain scalable search capability for the progressively increasing chemical space and prevents the accumulation of error. Empirically, we find that MoleBridge is capable of accurately predicting synthesis pathways while exhibiting excellent performance in a variety of application scenarios.



Paperid:3990
Authors:Lin Zhang, Wenshuo Dong, Zhuoran Zhang, Shu Yang, Lijie Hu, Ninghao Liu, Pan Zhou, Di Wang
Abstract:
Understanding the internal mechanisms of transformer-based language models remains challenging. Mechanistic interpretability based on circuit discovery aims to reverse engineer neural networks by analyzing their internal processes at the level of computational subgraphs. In this paper, we revisit existing gradient-based circuit identification methods and find that their performance is either affected by the zero-gradient problem or saturation effects, where edge attribution scores become insensitive to input changes, resulting in noisy and unreliable attribution evaluations for circuit components. To address the saturation effect, we propose Edge Attribution Patching with GradPath (EAP-GP), EAP-GP introduces an integration path, starting from the input and adaptively following the direction of the difference between the gradients of corrupted and clean inputs to avoid the saturated region. This approach enhances attribution reliability and improves the faithfulness of circuit identification. We evaluate EAP-GP on 6 datasets using GPT-2 Small, GPT-2 Medium, and GPT-2 XL. Experimental results demonstrate that EAP-GP outperforms existing methods in circuit faithfulness, achieving improvements up to 17.7\%. Comparisons with manually annotated ground-truth circuits demonstrate that EAP-GP achieves precision and recall comparable to or better than previous approaches, highlighting its effectiveness in identifying accurate circuits.



Paperid:3991
Authors:Keyon Vafa, Sarah Bentley, Jon Kleinberg, Sendhil Mullainathan
Abstract:
How should we evaluate the quality of generative models? Many existing metrics focus on a model's producibility, i.e. the quality and breadth of outputs it can generate. However, the actual value from using a generative model stems not just from what it can produce but whether a user with a specific goal can produce an output that satisfies that goal. We refer to this property as steerability. In this paper, we first introduce a mathematical framework for quantifying steerability independently from producibility. Steerability is more challenging to evaluate than producibility because it requires knowing a user's goals. We address this issue by creating a benchmark task that relies on one key idea: sample an output from a generative model and ask users to reproduce it. We implement this benchmark in a large-scale user study of text-to-image models and large language models. Despite the ability of these models to produce high-quality outputs, they all perform poorly on steerability. We also find that steerability can be predicted using machine learning methods, providing insights into model differences. These results suggest that we need to focus on improving the steerability of generative models. We show such improvements are indeed possible: simple image-based steering mechanisms achieve more than 2x improvement on this benchmark.



Paperid:3992
Authors:Yifei Wang, Weimin Bai, colin zhang, Debing Zhang, Weijian Luo, He Sun
Abstract:
Abstract:In this paper, we unify more than 10 existing one-step diffusion distillation approaches, such as Diff-Instruct, DMD, SIM, SiD, $f$-distill, etc, inside a theory-driven framework which we name the \textbf{\emph{Uni-Instruct}}. Uni-Instruct is motivated by our proposed diffusion expansion theory of the $f$-divergence family. Then we introduce key theories that overcome the intractability issue of the original expanded $f$-divergence, resulting in an equivalent yet tractable loss that effectively trains one-step diffusion models by minimizing the expanded $f$-divergence family. The novel unification introduced by Uni-Instruct not only offers new theoretical contributions that help understand existing approaches from a high-level perspective but also leads to state-of-the-art one-step diffusion generation performances. On the CIFAR10 generation benchmark, Uni-Instruct achieves record-breaking Frechet Inception Distance (FID) values of \textbf{\emph{1.46}} for unconditional generation and \textbf{\emph{1.38}} for conditional generation. On the ImageNet-$64\times 64$ generation benchmark, Uni-Instruct achieves a new SoTA one-step diffusion FID value of \textbf{\emph{1.06}}, which outperforms its 79-step teacher diffusion with a significant improvement margin of 1.29 (1.06 vs 2.35). We also apply Uni-Instruct on broader tasks like text-to-3D generation. For text-to-3D generation, Uni-Instruct gives decent results, which slightly outperforms previous methods, such as SDS and VSD, in terms of both generation quality and diversity. Both the solid theoretical and empirical contributions of Uni-Instruct will potentially help future studies on one-step diffusion distillation and knowledge transferring of diffusion models.



Paperid:3993
Authors:Siyu Zhou, Tianyi Zhou, Yijun Yang, Guodong Long, Deheng Ye, Jing Jiang, Chengqi Zhang
Abstract:
Can we build accurate world models out of large language models (LLMs)? How can world models benefit LLM agents? The gap between the prior knowledge of LLMs and the specified environment's dynamics usually bottlenecks LLMs' performance as world models. To bridge the gap, we propose a training-free "world alignment" that learns an environment's symbolic knowledge complementary to LLMs. The symbolic knowledge covers action rules, knowledge graphs, and scene graphs, which are extracted by LLMs from exploration trajectories and encoded into executable codes to regulate LLM agents' policies. We further propose an RL-free, model-based agent "WALL-E" through the model-predictive control (MPC) framework. Unlike classical MPC requiring costly optimization on the fly, we adopt an LLM agent as an efficient look-ahead optimizer of future steps' actions by interacting with the neurosymbolic world model. While the LLM agent's strong heuristics make it an efficient planner in MPC, the quality of its planned actions is also secured by the accurate predictions of the aligned world model. They together considerably improve learning efficiency in a new environment. On open-world challenges in Mars (Minecraft like) and ALFWorld (embodied indoor environments), WALL-E significantly outperforms existing methods, e.g., surpassing baselines in Mars by 16.1%–51.6% of success rate and by at least 61.7% in score. In ALFWorld, it achieves a new record 98% success rate after only 4 iterations.



Paperid:3994
Authors:Yibo Wang, Hai-Long Sun, Guangda Huzhang, Qingguo Chen, Zhao Xu, Weihua Luo, Kaifu Zhang, Lijun Zhang
Abstract:
Abstract:Recently, self-play fine-tuning (SPIN) has been proposed to adapt large language models to downstream applications with scarce expert-annotated data, by iteratively generating synthetic responses from the model itself. However, SPIN is designed to optimize the current reward advantages of annotated responses over synthetic responses at hand, which may gradually vanish during iterations, leading to \textit{unstable optimization}. Moreover, the utilization of reference policy induces a \textit{misalignment} issue between the reward formulation for training and the metric for generation. To address these limitations, we propose a novel \textbf{T}riplet-based \textbf{S}elf-\textbf{P}lay f\textbf{I}ne-tu\textbf{N}ing (TSPIN) method that integrates two key designs. First, beyond current advantages, TSPIN additionally incorporates historical advantages between iteratively generated responses and proto-synthetic responses produced by the initial policy. Even if the current advantages diminish, historical advantages remain effective, stabilizing the overall optimization. Second, TSPIN introduces the entropy constraint into the self-play framework, which is theoretically justified to support reference-free fine-tuning, eliminating the training-generation discrepancy. Empirical results on various tasks demonstrate not only the superior performance of TSPIN over SPIN, but also its stable evolution during iterations. Remarkably, compared to supervised fine-tuning, TSPIN achieves comparable or even better performance with only $25\\%$ samples, highlighting its effectiveness when faced with scarce annotated data.



Paperid:3995
Authors:Felix Koehler, Nils Thuerey
Abstract:
Neural operators or emulators for PDEs trained on data from numerical solversare conventionally assumed to be limited by their training data's fidelity. Wechallenge this assumption by identifying "emulator superiority," where neuralnetworks trained purely on low-fidelity solver data can achieve higher accuracythan those solvers when evaluated against a higher-fidelity reference. Ourtheoretical analysis reveals how the interplay between emulator inductivebiases, training objectives, and numerical error characteristics enablessuperior performance during multi-step rollouts. We empirically validate thisfinding across different PDEs using standard neural architectures, demonstratingthat emulators can implicitly learn dynamics that are more regularized orexhibit more favorable error accumulation properties than their training data,potentially surpassing training data limitations and mitigating numericalartifacts. This work prompts a re-evaluation of emulator benchmarking,suggesting neural emulators might achieve greater physical fidelity than theirtraining source within specific operational regimes.



Paperid:3996
Authors:Ege Özsoy, Arda Mamur, Felix Tristram, Chantal Pellegrini, Magdalena Wysocki, Benjamin Busam, Nassir Navab
Abstract:
Operating rooms (ORs) demand precise coordination among surgeons, nurses, and equipment in a fast-paced, occlusion-heavy environment, necessitating advanced perception models to enhance safety and efficiency. Existing datasets either provide partial egocentric views or sparse exocentric multi-view context, but do not explore the comprehensive combination of both. We introduce EgoExOR, the first OR dataset and accompanying benchmark to fuse first-person and third-person perspectives. Spanning 94 minutes (84,553 frames at 15 FPS) of two emulated spine procedures, Ultrasound-Guided Needle Insertion and Minimally Invasive Spine Surgery, EgoExOR integrates egocentric data (RGB, gaze, hand tracking, audio) from wearable glasses, exocentric RGB and depth from RGB-D cameras, and ultrasound imagery. Its detailed scene graph annotations, covering 36 entities and 22 relations (568,235 triplets), enable robust modeling of clinical interactions, supporting tasks like action recognition and human-centric perception. We evaluate the surgical scene graph generation performance of two adapted state-of-the-art models and offer a new baseline that explicitly leverages EgoExOR’s multimodal and multi-perspective signals. This new dataset and benchmark set a new foundation for OR perception, offering a rich, multimodal resource for next-generation clinical perception. Our code and data are available at https://github.com/ardamamur/EgoExOR.



Paperid:3997
Authors:Wei Shen, Weiqi Liu, Mingde Chen, Wenke Huang, Mang Ye
Abstract:
Vertical Federated Learning (VFL) has emerged as a critical privacy-preserving learning paradigm, enabling collaborative model training by leveraging distributed features across clients. However, due to privacy concerns, there are few publicly available real-world datasets for evaluating VFL methods, which poses significant challenges to related research. To bridge this gap, we propose MARS-VFL, a unified benchmark for realistic VFL evaluation. It integrates data from practical applications involving collaboration across different features, maintaining compatibility with the VFL setting. Based on this, we standardize the evaluation of VFL methods from the mainstream aspects of efficiency, robustness, and security. We conduct comprehensive experiments to assess different VFL approaches, providing references for unified evaluation. Furthermore, we are the first to unify the evaluation of robustness challenges in VFL and introduce a new method for addressing robustness challenges, establishing standard baselines for future research.



Paperid:3998
Authors:Haoxiang Wang, Zinan Lin, Da Yu, Huishuai Zhang
Abstract:
Abstract:Generating high-fidelity, differentially private (DP) synthetic images offers a promising route to share and analyze sensitive visual data without compromising individual privacy. However, existing DP image synthesis methods struggle to produce high-resolution outputs that faithfully capture the structure of the original data. In this paper, we introduce a novel method, referred to as Synthesis via Private Textual Intermediaries (SPTI), that can generate high-resolution images with strong privacy guarantees and easy adoptions. The key idea is to shift the challenge of DP image synthesis from the image domain to the text domain by leveraging state-of-the-art DP text generation methods. SPTI first summarizes each private image into a concise textual description using image-to-text models, then applies a modified Private Evolution algorithm to generate DP text, and finally reconstructs images using text-to-image models. Notably, SPTI requires no model training, only inferences with off-the-shelf models. Given a private dataset, SPTI produces synthetic images of substantially higher quality than prior DP approaches. On the LSUN Bedroom dataset, SPTI attains an FID $\le$ 26.71 under $\epsilon=1.0$, improving over Private Evolution’s FID of 40.36. Similarly, on MM-CelebA-HQ, SPTI achieves an FID $\le$ 33.27 at $\epsilon=1.0$, compared to 57.01 from DP fine-tuning baselines. Overall, our results demonstrate that Synthesis via Private Textual Intermediaries provides a resource-efficient and proprietary-model-compatible framework for generating high-resolution DP synthetic images, greatly expanding access to private visual datasets.



Paperid:3999
Authors:Rey Koki, Michael McCabe, Dhruv Kedar, Josh Myers-Dean, Annabel Wade, Jebb Stewart, Christina Kumler-Bonfanti, Jed Brown
Abstract:
The global rise in wildfire frequency and intensity over the past decade underscores the need for improved fire monitoring techniques. To advance deep learning research on wildfire detection and its associated human health impacts, we introduceSmokeViz, a large-scale machine learning dataset of smoke plumes in satellite imagery. The dataset is derived from expert annotations created by smoke analysts at the National Oceanic and Atmospheric Administration, which provide coarse temporal and spatial approximations of smoke presence. To enhance annotation precision, we proposepseudo-label dimension reduction (PLDR), a generalizable method that applies pseudo-labeling to refine datasets with mismatching temporal and/or spatial resolutions. Unlike typical pseudo-labeling applications that aim to increase the number of labeled samples, PLDR maintains the original labels but increases the dataset quality by solving for intermediary pseudo-labels (IPLs) that align each annotation to the most representative input data. For SmokeViz, a parent model produces IPLs to identify the single satellite image within each annotations time window that best corresponds with the smoke plume. This refinement process produces a succinct and relevant deep learning dataset consisting of over 180,000 manual annotations. The SmokeViz dataset is expected to be a valuable resource to develop further wildfire-related machine learning models and is publicly available at https://noaa-gsl-experimental-pds.s3.amazonaws.com/index.html#SmokeViz/.



Paperid:4000
Authors:Yan Gao, Massimo R. Scamarcia, Javier Fernandez-Marques, Mohammad Naseri, Chong Ng, Dimitris Stripelis, Zexi Li, Tao Shen, Jiamu Bai, Daoyuan Chen, Zikai Zhang, Rui Hu, InSeo Song, KangYoon Lee, Hong Jia, Ting Dang, Junyan Wang, Zheyuan Liu, Daniel J. Beutel, Lingjuan Lyu, Nicholas Lane
Abstract:
Large Language Models (LLMs) have achieved state-of-the-art results across diverse domains, yet their development remains reliant on vast amounts of publicly available data, raising concerns about data scarcity and the lack of access to domain-specific, sensitive information. Federated Learning (FL) presents a compelling framework to address these challenges by enabling decentralized fine-tuning on pre-trained LLMs without sharing raw data. However, the compatibility and performance of pre-trained LLMs in FL settings remain largely under explored. We introduce the FlowerTune LLM Leaderboard, a first-of-its-kind benchmarking suite designed to evaluate federated fine-tuning of LLMs across four diverse domains: general NLP, finance, medical, and coding. Each domain includes federated instruction-tuning datasets and domain-specific evaluation metrics. Our results, obtained through a collaborative, open-source and community-driven approach, provide the first comprehensive comparison across 26 pre-trained LLMs with different aggregation and fine-tuning strategies under federated settings, offering actionable insights into model performance, resource constraints, and domain adaptation. This work lays the foundation for developing privacy-preserving, domain-specialized LLMs for real-world applications.



Paperid:4001
Authors:Wancong Zhang, Uladzislau Sobal, Kyunghyun Cho, Randall Balestriero, Tim G. J. Rudner, Yann LeCun
Abstract:
A long-standing goal in AI is to develop agents capable of solving diverse tasks across a range of environments, including those never seen during training. Two dominant paradigms address this challenge: (i) reinforcement learning (RL), which learns policies via trial and error, and (ii) optimal control, which plans actions using a known or learned dynamics model. However, their comparative strengths in the offline setting—where agents must learn from reward-free trajectories—remain underexplored. In this work, we systematically evaluate RL and control-based methods on a suite of navigation tasks, using offline datasets of varying quality. On the RL side, we consider goal-conditioned and zero-shot approaches. On the control side, we train a latent dynamics model using the Joint Embedding Predictive Architecture (JEPA) and employ it for planning. We investigate how factors such as data diversity, trajectory quality, and environment variability influence the performance of these approaches. Our findings reveal that model-free RL excels when abundant, high-quality data is available, while model-based planning demonstrates superior generalization to novel layouts, better trajectory stitching, and greater data efficiency. Notably, planning with a latent dynamics model proves to be a strong approach for handling suboptimal offline data and adapting to diverse environment configurations.



Paperid:4002
Authors:Joel Valdivia Ortega, Lorenz Lamm, Franziska Eckardt, Benedikt Schworm, Marion Jasnin, Tingying Peng
Abstract:
Vision Transformers (ViTs), such as DINOv2, achieve strong performance across domains but often repurpose low-informative patch tokens in ways that reduce the interpretability of attention and feature maps. This challenge is especially evident in medical imaging, where domain shifts can degrade both performance and transparency. In this paper, we introduce Randomized-MLP (RMLP) regularization, a contrastive learning-based method that encourages more semantically aligned representations. We apply RMLP when fine-tuning DINOv2 to both medical and natural image modalities, showing that it improves or maintains downstream performance while producing more interpretable attention maps. We also provide a mathematical analysis of RMLPs, offering insights into its role in enhancing ViT-based models and advancing our understanding of contrastive learning in this context.



Paperid:4003
Authors:Sungho Koh, SeungJu Cha, Hyunwoo Oh, Kwanyoung Lee, Dong-Jin Kim
Abstract:
Text-to-image diffusion models often exhibit degraded performance when generating images beyond their training resolution.Recent training-free methods can mitigate this limitation, but they are primarily designed for U-Net-based models and often yield suboptimal results when applied to recent Diffusion Transformer models.While patch-based methods are architecture-agnostic, they incur substantial computational due to overlapping patches.In this paper, we propose ScaleDiff, a model-agnostic and highly efficient framework for extending the resolution of pretrained diffusion models without any additional training.A core component of our framework is Neighborhood Patch Attention (NPA), an efficient mechanism that reduces computational redundancy in self-attention layer with non-overlapping patches.We integrate NPA into an upsample-diffuse-denoise pipeline and introduce Latent Frequency Mixing (LFM) to better generate fine details.Furthermore, we apply Structure Guidance to enhance global structure during the denoising process.Experimental results demonstrate that ScaleDiff achieves state-of-the-art performance among training-free methods in terms of both image quality and inference speed on both U-Net and Diffusion Transformer architectures.



Paperid:4004
Authors:Yichen Zhu, Weiyu Chen, James Kwok, Zhou Zhao
Abstract:
Autoregressive models (ARMs) show strong capabilities in many domains but face challenges with planning and complex reasoning due to their sequential generation. Masked diffusion models (MDMs) address these issues by enabling controllable, any-order, and parallel generation but encounter training difficulties as token prediction complexity varies with unmasked token positions. This work identifies two key characteristics of efficient MDM sampling paths: prioritizing tokens near unmasked ones and generating subsequence earlier in reasoning. We propose the Simple Path Mask Diffusion Model (SPMDM), which partitions sequences into fixed-length, non-overlapping subsequences and applies varying noise scales to learn token-level and cross-subsequence dependencies. Experiments on synthetic data and tasks like Countdown and Sudoku show SPMDM captures structural rules effectively, significantly outperforming existing MDMs and ARMs, with competitive results on broader reasoning benchmarks.



Paperid:4005
Authors:Fivos Kalogiannis, Gabriele Farina
Abstract:
Abstract:Multi-agent reinforcement learning (MARL) has long been seen as inseparable from Markov games (Littman 1994). Yet, the most remarkable achievements of practical MARL have arguably been in extensive-form games (EFGs)---spanning games like Poker, Stratego, and Hanabi. At the same time, little is known about provable equilibrium convergence for MARL algorithms applied to EFGs as they stumble upon the inherent nonconvexity of the optimization landscape and the failure of the value-iteration subroutine in EFGs. To this goal, we utilize contemporary advances in nonconvex optimization theory to prove that regularized alternating policy gradient with (i) *direct policy parametrization*, (ii) *softmax policy parametrization*, and (iii) *softmax policy parametrization with natural policy gradient* updates converge to an approximate Nash equilibrium (NE) in the *last-iterate* in imperfect-information perfect-recall zero-sum EFGs. Namely, we observe that since the individual utilities are concave with respect to the sequence-form strategy, they satisfy gradient dominance w.r.t. the behavioral strategy---or, \textit{policy}, in reinforcement learning terms. We exploit this structure to further prove that the regularized utility satisfies the much stronger Polyak- Łojasiewicz condition. In turn, we show that the different flavors of alternating policy gradient methods converge to an $\epsilon$-approximate NE with a number of iterations and trajectory samples that are polynomial in $1/\epsilon$ and the natural parameters of the game. Our work is a preliminary---yet principled---attempt in bridging the conceptual gap between the theory of Markov and imperfect-information EFGs while it aspires to stimulate a deeper dialogue between them.



Paperid:4006
Authors:shu chen, HongJun Xu, Ruichi Zhang, Mengke Li, Yonggang Zhang, Yang Lu, Bo Han, Yiu-ming Cheung, Hanzi Wang
Abstract:
In real world, the observed label distribution of a dataset often mismatches its true distribution due to noisy labels. In this situation, noisy labels learning (NLL) methods directly integrated with long-tail learning (LTL) methods tend to fail due to a dilemma: NLL methods normally rely on unbiased model predictions to recover true distribution by selecting and correcting noisy labels; while LTL methods like logit adjustment depends on true distributions to adjust biased predictions, leading to a deadlock of mutual dependency defined in this paper. To address this, we propose \texttt{Unlocker}, a bilevel optimization framework that integrates NLL methods and LTL methods to iteratively disentangle this deadlock. The inner optimization leverages NLL to train the model, incorporating LTL methods to fairly select and correct noisy labels. The outer optimization adaptively determines an adjustment strength, mitigating model bias from over- or under-adjustment. We also theoretically prove that this bilevel optimization problem is convergent by transferring the outer optimization target to an equivalent problem with a closed-form solution. Extensive experiments on synthetic and real-world datasets demonstrate the effectiveness of our method in alleviating model bias and handling long-tailed noisy label data. Code is available at \url{https://anonymous.4open.science/r/neurips-2025-anonymous-1015/}.



Paperid:4007
Authors:Wenjun Wu, Lingling Zhang, Bo Zhao, Muye Huang, QianYing Wang, Jun Liu
Abstract:
The task of geometry problem solving has been a long-standing focus in the automated mathematics community and draws growing attention due to its complexity for both symbolic and neural models. Although prior studies have explored various effective approaches for enhancing problem solving performances, two fundamental challenges remain unaddressed, which are essential to the application in practical scenarios. First, the multi-step reasoning gap between the initial geometric conditions and ultimate problem goal leads to a great search space for solution exploration. Second, obtaining multiple interpretable and shorter solutions remains an open problem. In this work, we introduce the Causal-Reasoning Geometry Problem Solver to overcome these challenges. Specifically, the Causal Graph Reasoning theory is proposed to perform symbolic reasoning before problem solving. Several causal graphs are constructed according to predefined rule base, where each graph is composed of primitive nodes, causal edges and prerequisite edges. By applying causal graph deduction from initial conditions, the reachability status of nodes are iteratively conveyed by causal edges until reaching the target nodes, representing feasible causal deduction paths. In this way, the search space of solutions is compressed from the beginning, the end and intermediate reasoning paths, while ensuring the interpretability and variety of solutions. To achieve this, we further propose Forward Matrix Deduction which transforms the causal graphs into matrices and vectors, and applies matrix operations to update the status value of reachable nodes in iterations. Finally, multiple solutions can be generated by tracing back from the target nodes after validation. Experiments demonstrate the effectiveness of our method to obtain multiple shorter and interpretable solutions. Code is available after acceptance.



Paperid:4008
Authors:Jiaqi Wang, Kevin Qinghong Lin, James Cheng, Mike Zheng Shou
Abstract:
Reinforcement Learning (RL) has proven to be an effective post-training strategy for enhancing reasoning in vision–language models (VLMs). Group Relative Policy Optimization (GRPO) is a recent prominent method that encourages models to generate complete reasoning traces before answering, leading to increased token usage and computational cost. Inspired by the human-like thinking process—where people skip reasoning for easy questions but think carefully when needed—we explore how to enable VLMs to first decidewhen reasoning is necessary.To realize this, we propose \ours, a two-stage training strategy:(i)a supervised fine-tuning (SFT) stage with a simple yet effective “thought dropout” operation, where reasoning traces are randomly replaced with empty thoughts. This introduces a think-or-not format that serves as a cold start for selective reasoning;(ii)a GRPO stage that enables the model to freely explore when to think or not, while maximizing task-aware outcome rewards.Experimental results show that \ours canreduce the completion length by up to90%compared to vanilla GRPO, without sacrificing performance or even improving it. Further evaluations across diverse vision-language tasks—covering a range of reasoning difficulties under both 3B and 7B models—consistently reveal that themodel progressively learns to bypass unnecessary reasoning steps as training advances. These findings shed light on the path toward human-like reasoning patterns in reinforcement learning approaches.



Paperid:4009
Authors:Wei Chen, Yuxuan Liang
Abstract:
Spatio-temporal forecasting is crucial in many domains, such as transportation, meteorology, and energy. However, real-world scenarios frequently present challenges such as signal anomalies, noise, and distributional shifts. Existing solutions primarily enhance robustness by modifying network architectures or training procedures. Nevertheless, these approaches are computationally intensive and resource-demanding, especially for large-scale applications. In this paper, we explore anoveltest-timecomputing paradigm, namely learning with calibration,ST-TTC, forspatio-temporal forecasting. Through learning with calibration, we aim to capture periodic structural biases arising from non-stationarity during the testing phase and perform real-time bias correction on predictions to improve accuracy. Specifically, we first introduce a spectral-domain calibrator with phase-amplitude modulation to mitigate periodic shift and then propose a flash updating mechanism with a streaming memory queue for efficient test-time computation.ST-TTCeffectively bypasses complex training-stage techniques, offering an efficient and generalizable paradigm. Extensive experiments on real-world datasets demonstrate the effectiveness, universality, flexibility and efficiency of our proposed method.



Paperid:4010
Authors:Seong-Hyeon Hwang, Soyoung Choi, Steven Whang
Abstract:
Multimodal models often over-rely on dominant modalities, failing to achieve optimal performance. While prior work focuses on modifying training objectives or optimization procedures, data-centric solutions remain underexplored. We propose MIDAS, a novel data augmentation strategy that generates misaligned samples with semantically inconsistent cross-modal information, labeled using unimodal confidence scores to compel learning from contradictory signals. However, this confidence-based labeling can still favor the more confident modality. To address this within our misaligned samples, we introduce weak-modality weighting, which dynamically increases the loss weight of the least confident modality, thereby helping the model fully utilize weaker modality. Furthermore, when misaligned features exhibit greater similarity to the aligned features, these misaligned samples pose a greater challenge, thereby enabling the model to better distinguish between classes. To leverage this, we propose hard-sample weighting, which prioritizes such semantically ambiguous misaligned samples. Experiments on multiple multimodal classification benchmarks demonstrate that MIDAS significantly outperforms related baselines in addressing modality imbalance.



Paperid:4011
Authors:Jiaye Qian, Ge Zheng, Yuchen Zhu, Sibei Yang
Abstract:
Despite their impressive performance across a wide range of tasks, Large Vision-Language Models (LVLMs) remain prone to hallucination. In this study, we propose a comprehensive intervention framework aligned with the transformer’s causal architecture in LVLMs, integrating the effects of different intervention paths on hallucination. We find that hallucinations in LVLMs do not arise from a single causal path, but rather from the interplay among image-to-input-text, image-to-output-text, and text-to-text pathways. For the first time, we also find that LVLMs rely on different pathways depending on the question–answer alignment format. Building on these insights, we propose simple yet effective methods to identify and intervene on critical hallucination heads within each pathway, tailored to discriminative and generative formats. Experiments across multiple benchmarks demonstrate that our approach consistently reduces hallucinations across diverse alignment types. Code will be made available.



Paperid:4012
Authors:Jae-Won Chung, Jiachen Liu, Jeff J., Ruofan Wu, Oh Kweon, Yuxuan Xia, Zhiyu Wu, Mosharaf Chowdhury
Abstract:
As the adoption of Generative AI in real-world services grow explosively, energy has emerged as a critical bottleneck resource.However, energy remains a metric that is often overlooked, under-explored, or poorly understood in the context of building ML systems.We present the ML.ENERGY Benchmark, a benchmark suite and tool for measuring inference energy consumption under realistic service environments, and the corresponding ML.ENERGY Leaderboard, which have served as a valuable resource for those hoping to understand and optimize the energy consumption of their generative AI services.In this paper, we explain four key design principles for benchmarking ML energy we have acquired over time, and then describe how they are implemented in the ML.ENERGY Benchmark.We then highlight results from the latest iteration of the benchmark, including energy measurements of 40 widely used model architectures across 6 different tasks, case studies of how ML design choices impact energy consumption, and how automated optimization recommendations can lead to significant (sometimes more than 40%) energy savings without changing what is being computed by the model.The ML.ENERGY Benchmark is open-source and can be easily extended to various customized models and application scenarios.



Paperid:4013
Authors:Sheng Wang, Pengan CHEN, Jingqi Zhou, Qintong Li, Jingwei Dong, Jiahui Gao, Boyang XUE, Jiyue Jiang, Lingpeng Kong, Chuan Wu
Abstract:
Model customization necessitates high-quality and diverse datasets, but acquiring such data remains time-consuming and labor-intensive. Despite the great potential of large language models (LLMs) for data synthesis, current approaches are constrained by limited seed data, model biases and low-variation prompts, resulting in limited diversity and biased distribution with the increase of data scales.To tackle this challenge, we introduce TreeSynth, a tree-guided subspace-based data synthesis approach inspired by decision trees. It constructs a spatial partitioning tree to recursively divide a task-specific full data space (i.e., root node) into numerous atomic subspaces (i.e., leaf nodes) with mutually exclusive and exhaustive attributes to ensure both distinctiveness and comprehensiveness, before synthesizing samples within each atomic subspace. This globally divide-and-synthesize method finally collects subspace samples into a comprehensive dataset, effectively circumventing repetition and space collapse to ensure the diversity of large-scale data synthesis.Furthermore, the spatial partitioning tree enables sample allocation into atomic subspaces, allowing the re-balancing of existing datasets for more balanced and comprehensive distributions.Empirically, extensive experiments across diverse benchmarks consistently validates the superior data diversity, model performance, and robust scalability of TreeSynth compared to both human-crafted datasets and peer data synthesis methods, with the average performance gain reaching 10%.Besides, the consistent improvements of TreeSynth-balanced datasets highlight its efficacious application to redistribute existing datasets for more comprehensive coverage and the induced performance enhancement. The code is available in the anonymous repository via https://anonymous.4open.science/r/TreeSynth-EF04.



Paperid:4014
Authors:Xiaoan Zhu, Yue Zhao, Tianyang Hu, Jiaming Guo, Yulan Zeng, Renjing Pei, Fenglong Song, Huajun Feng
Abstract:
Dual-camera smartphones suffer from geometric and photometric inconsistencies during zoom transitions, primarily due to disparities in intrinsic/extrinsic parameters and divergent image processing pipelines between the two cameras. Existing interpolation methods struggle to effectively address this issue, constrained by the lack of ground-truth datasets and motion ambiguity in dynamic scenarios.To overcome these challenges, we propose OmniZoom, a universal plug-and-play paradigm for cross-device smooth zoom interpolation. Specifically, we present a novel cross-device virtual data generation method utilizing 3D Gaussian Splatting. This method tackles data scarcity by decoupling geometric features via spatial transition modeling and correcting photometric variations with dynamic color adaptation. It is further enhanced by cross-domain consistency learning for device-agnostic semantic alignment. Additionally, we introduce a plug-and-play 3D-TPR (3D Trajectory Progress Ratio Mapping) framework that surmounts 2D spatial limitations. As components of our framework, a texture-focus strategy is introduced for high-frequency detail preservation, incorporating mask penalty constraints to suppress interpolation artifacts. Our pipeline exhibits broad compatibility with diverse interpolation methods and achieves good performance across multiple public benchmarks. Real-world evaluations on various smartphone platforms also reveal significant quality improvements after fine-tuning on our synthetic data, which underscores the robustness and practical effectiveness of our approach for cross-device zoom applications.



Paperid:4015
Authors:Ana Rivera, Anvita Bhagavathula, Alvaro Carbonero, Priya Donti
Abstract:
Large-scale integration of renewable energy and climate-induced extreme weather events increase uncertainty in power system operations, which calls for tools that can accurately and efficiently simulate a wide range of scenarios and operating conditions. Machine learning approaches offer potential speedup over traditional solvers, but their performance has not been systematically assessed on benchmarks that capture real-world variability. This paper introduces PF∆, a benchmark dataset designed to evaluate power flow methods under variations in load, generation, and topology. PF∆ contains 693,000 solved power flow instances spanning six different bus system sizes,, capturing three types of contingency scenarios (N, N-1, and N-2), and including close-to-infeasible cases near steady-state voltage stability limits. We evaluate traditional solvers and GNN-based methods, highlighting key areas where existing approaches struggle, and identifying open problems for future research. Our dataset is available at https://huggingface.co/datasets/pfdelta/pfdelta/tree/main and our code with data generation scripts and model implementations at https://anonymous.4open.science/r/pfdelta-A769.



Paperid:4016
Authors:chen yang, Hui Wang, Shiyao Wang, Junyang Chen, Jiabei He, Jiaming Zhou, Xi Yang, Yequan Wang, Yonghua Lin, Yong Qin
Abstract:
While voice technologies increasingly serve aging populations, current systems exhibit significant performance gaps due to inadequate training data capturing elderly-specific vocal characteristics like presbyphonia and dialectal variations. The limited data available on super-aged individuals in existing elderly speech datasets, coupled with overly simple recording styles and annotation dimensions, exacerbates this issue. To address the critical scarcity of speech data from individuals aged 75 and above, we introduce SeniorTalk, a carefully annotated Chinese spoken dialogue dataset. This dataset contains 55.53 hours of speech from 101 natural conversations involving 202 participants, ensuring a strategic balance across gender, region, and age. Through detailed annotation across multiple dimensions, it can support a wide range of speech tasks. We perform extensive experiments on speaker verification, speaker diarization, speech recognition, and speech editing tasks, offering crucial insights for the development of speech technologies targeting this age group. Code is available at https://github.com/flageval-baai/SeniorTalk and data at https://huggingface.co/datasets/BAAI/SeniorTalk.



Paperid:4017
Authors:Julia Nakhleh, Robert Nowak
Abstract:
Abstract:Overparameterized neural networks can interpolate a given dataset in many different ways, prompting the fundamental question: which among these solutions should we prefer, and what explicit regularization strategies will provably yield these solutions? This paper addresses the challenge of finding the sparsest interpolating ReLU network—i.e., the network with the fewest nonzero parameters or neurons—a goal with wide-ranging implications for efficiency, generalization, interpretability, theory, and model compression. Unlike post hoc pruning approaches, we propose a continuous, almost-everywhere differentiable training objective whose global minima are guaranteed to correspond to the sparsest single-hidden-layer ReLU networks that fit the data. This result marks a conceptual advance: it recasts the combinatorial problem of sparse interpolation as a smooth optimization task, potentially enabling the use of gradient-based training methods. Our objective is based on minimizing $\ell^p$ quasinorms of the weights for $0 < p < 1$, a classical sparsity-promoting strategy in finite-dimensional settings. However, applying these ideas to neural networks presents new challenges: the function class is infinite-dimensional, and the weights are learned using a highly nonconvex objective. We prove that, under our formulation, global minimizers correspond exactly to sparsest solutions. Our work lays a foundation for understanding when and how continuous sparsity-inducing objectives can be leveraged to recover sparse networks through training.



Paperid:4018
Authors:Xinyan Chen, Renrui Zhang, DONGZHI JIANG, Aojun Zhou, Shilin Yan, Weifeng Lin, Hongsheng Li
Abstract:
Chain-of-Thought (CoT) has widely enhanced mathematical reasoning in Large Language Models (LLMs), but it still remains challenging for extending it to multimodal domains. Existing works either adopt a similar textual reasoning for image input, or seek to interleave visual signals into mathematical CoT. However, they face three key limitations for math problem-solving:reliance on coarse-grained box-shaped image regions, limited perception of vision encoders on math content, and dependence on external capabilities for visual modification. In this paper, we proposeMINT-CoT, introducingMathematicalINterleavedTokens forChain-of-Thought visual reasoning. MINT-CoT adaptively interleaves relevant visual tokens into textual reasoning steps via an Interleave Token, which dynamically selects visual regions of any shapes within math figures. To empower this capability, we construct the MINT-CoT dataset, containing 54K mathematical problems aligning each reasoning step with visual regions at the token level, accompanied by a rigorous data generation pipeline. We further present a three-stage MINT-CoT training strategy, progressively combining text-only CoT SFT, interleaved CoT SFT, and interleaved CoT RL, which derives our MINT-CoT-7B model. Extensive experiments demonstrate the effectiveness of our method for effective visual interleaved reasoning in mathematical domains, where MINT-CoT-7B outperforms the baseline model by +34.08% on MathVista and +28.78% on GeoQA, respectively.



Paperid:4019
Authors:Linus Aronsson, Morteza Haghir Chehreghani
Abstract:
Abstract:Signed networks, where edges are labeled as positive or negative to represent friendly or antagonistic interactions, offer a natural framework for analyzing polarization, trust, and conflict in social systems. Detecting meaningful group structures in such networks is crucial for understanding online discourse, political divisions, and trust dynamics. A key challenge is to identify communities that are internally cohesive and externally antagonistic, while allowing for neutral or unaligned vertices. In this paper, we propose a method for identifying $k$ polarized communities that addresses a major limitation of prior methods: their tendency to produce highly size-imbalanced solutions. We introduce a novel optimization objective that avoids such imbalance. In addition, it is well known that approximation algorithms based on *local search* are highly effective for clustering signed networks when neutral vertices are not allowed. We build on this idea and design the first local search algorithm that extends to the setting with neutral vertices while scaling to large networks. By connecting our approach to block-coordinate Frank-Wolfe optimization, we prove a linear convergence rate, enabled by the structure of our objective. Experiments on real-world and synthetic datasets demonstrate that our method consistently outperforms state-of-the-art baselines in solution quality, while remaining competitive in computational efficiency.



Paperid:4020
Authors:Junchi Yu, Yujie Liu, Jindong Gu, Philip Torr, Dongzhan Zhou
Abstract:
Retrieval-Augmented Generation (RAG) based on knowledge graphs (KGs) enhances large language models (LLMs) by providing structured and interpretable external knowledge.However, existing KG-based RAG methods struggle to retrieve accurate and diverse information from text-rich KGs for complex real-world queries.Process Reward Models (PRMs) offer a way to align the retrieval process of KG-based RAG with query-specific knowledge requirements, but they heavily rely on process-level supervision signals that are expensive and hard to obtain on KGs.To address this challenge, we propose GraphFlow, a framework that efficiently retrieves accurate and diverse knowledge required for real-world queries from text-rich KGs.GraphFlow employs a transition-based flow matching objective to jointly optimize a retrieval policy and a flow estimator.The flow estimator factorizes the reward of the retrieval outcome into the intermediate retrieval states.Such reward factorization guides the retrieval policy to retrieve candidates from KGs in proportion to their reward.This allows GraphFlow to explore high-quality regions of KGs that yield diverse and relevant results. We evaluate GraphFlow on the STaRK benchmark, which includes real-world queries from multiple domains over text-rich KGs. GraphFlow outperforms strong KG-RAG baselines, including GPT-4o, by 10\% on average in hit rate and recall. It also shows strong generalization to unseen KGs, demonstrating its effectiveness and robustness.



Paperid:4021
Authors:Anna van Elst, Igor Colin, Stephan Clémençon
Abstract:
Abstract:This paper addresses the problem of robust estimation in gossip algorithms over arbitrary communication graphs. Gossip algorithms are fully decentralized, relying only on local neighbor-to-neighbor communication, making them well-suited for situations where communication is constrained. A fundamental challenge in existing mean-based gossip algorithms is their vulnerability to malicious or corrupted nodes. In this paper, we show that an outlier-robust mean can be computed by globally estimating a robust statistic. More specifically, we propose a novel gossip algorithm for rank estimation, referred to as \textsc{GoRank}, and leverage it to design a gossip procedure dedicated to trimmed mean estimation, coined \textsc{GoTrim}. In addition to a detailed description of the proposed methods, a key contribution of our work is a precise convergence analysis: we establish an $ \mathcal{O}(1/t) $ rate for rank estimation and an $ \mathcal{O}(\log(t)/t) $ rate for trimmed mean estimation, where by $ t $ is meant the number of iterations. Moreover, we provide a breakdown point analysis of \textsc{GoTrim}. We empirically validate our theoretical results through experiments on diverse network topologies, data distributions, and contamination schemes.



Paperid:4022
Authors:Haiyun He, Yepeng Liu, Ziqiao Wang, Yongyi Mao, Yuheng Bu
Abstract:
Abstract:Watermarking has emerged as a crucial method to distinguish AI-generated text from human-created text. Current watermarking approaches often lack formal optimality guarantees or address the scheme and detector design separately. In this paper, we introduce a novel, unified theoretical framework for watermarking Large Language Models (LLMs) that jointly optimizes both the watermarking scheme and detector. Our approach aims to maximize detection performance while maintaining control over the worst-case false positive rate (FPR) and distortion on text quality. We derive closed-form optimal solutions for this joint design and characterize the fundamental trade-off between watermark detectability and distortion. Notably, we reveal that the optimal watermarking schemes should be adaptive to the LLM’s generative distribution. Building on our theoretical insights, we propose a distortion-free, distribution-adaptive watermarking algorithm (DAWA) that leverages a surrogate model for model-agnosticism and efficiency. Experiments on Llama2-13B and Mistral-8$\times$7B models confirm the effectiveness of our approach, particularly at ultra-low FPRs.



Paperid:4023
Authors:Jun Wang, Shuai Wang, Liang Ding, Hongyu Li, Yong Luo, Huangxuan Zhao, Han Hu, Bo Du
Abstract:
Abstract:Continual learning for large language models (LLMs) demands a precise balance between $\textbf{plasticity}$ - the ability to absorb new tasks - and $\textbf{stability}$ - the preservation of previously learned knowledge. Conventional rehearsal methods, which replay stored examples, are limited by long-term data inaccessibility; earlier pseudo-rehearsal methods require additional generation modules, while self-synthesis approaches often generate samples that poorly align with real tasks, suffer from unstable outputs, and ignore task relationships. We present $\textbf{\textit{Self-Evolving Pseudo-Rehearsal for Catastrophic Forgetting with Task Similarity}}(\textbf{SERS})$, a lightweight framework that 1) decouples pseudo-input synthesis from label creation, using semantic masking and template guidance to produce diverse, task-relevant prompts without extra modules; 2) applies label self-evolution, blending base-model priors with fine-tuned outputs to prevent over-specialization; and 3) introduces a dynamic regularizer driven by the Wasserstein distance between task distributions, automatically relaxing or strengthening constraints in proportion to task similarity. Experiments across diverse tasks on different LLMs show that our SERS reduces forgetting by over 2\% points against strong pseudo-rehearsal baselines, by ensuring efficient data utilization and wisely transferring knowledge. The code will be released at https://anonymous.4open.science/r/LLM_CL_SERS-0ECD/.



Paperid:4024
Authors:Chenwang Wu, Yiu-ming Cheung, Bo Han, Defu Lian
Abstract:
Existing machine-generated text (MGT) detection methods implicitly assume labels as the "golden standard". However, we uncover boundary ambiguity in MGT detection, implying that traditional training paradigms are inexact. The limitations of human cognition and the super intelligence of detectors make inexact learning widespread and inevitable. To this end, we propose an easy-to-hard enhancement framework to provide reliable supervision under such inexact conditions. Distinct from knowledge distillation, which transfers expertise from strong to weak models, our framework employs an easy supervisor targeting relatively simple longer-text detection tasks (despite weaker capabilities), to enhance the more challenging target detector. Firstly, longer texts targeted by supervisors theoretically alleviate the impact of inexact labels, laying the foundation for reliable supervision. Secondly, by structurally incorporating the detector into the supervisor, we theoretically model the supervisor as a lower performance bound for the detector. Thus, optimizing the supervisor indirectly optimizes the detector, ultimately approximating the underlying "golden" labels. Extensive experiments across diverse practical scenarios, including cross-LLM, cross-domain, mixed text, and paraphrase attacks, demonstrate the framework's significant effectiveness in improving detection capabilities. The code is available at: \url{https://anonymous.4open.science/r/Easy2Hard}.



Paperid:4025
Authors:David Dai, Peilin Chen, Chanakya Ekbote, Paul Pu Liang
Abstract:
Clinical decision‑making routinely demands reasoning over heterogeneous data, yet existing multimodal language models (MLLMs) remain largely vision‑centric and fail to generalize across clinical specialties. To bridge this gap, we introduce QoQ-Med-7B/32B, the first open generalist clinical foundation model that jointly reasons across medical images, time‑series signals, and text reports. QoQ-Med is trained with Domain‑aware Relative Policy Optimization (DRPO), a novel reinforcement‑learning objective that hierarchically scales normalized rewards according to domain rarity and modality difficulty, mitigating performance imbalance caused by skewed clinical data distributions. Trained on 2.61 million instruction tuning pairs spanning 9 clinical domains, we show that DRPO training boosts diagnostic performance by 43% in macro‑F1 on average across all visual domains as compared to other critic-free training methods like GRPO. Furthermore, with QoQ-Med trained on intensive segmentation data, it is able to highlight salient regions related to the diagnosis, with an IoU 10x higher than open models while reaching the performance of OpenAI o4-mini. To foster reproducibility and downstream research, we release (i) the full model weights, (ii) the modular training pipeline, and (iii) all intermediate reasoning traces.



Paperid:4026
Authors:Haiming Zhang, Yiyao Zhu, Wending Zhou, Xu Yan, Yingjie CAI, Bingbing Liu, Shuguang Cui, Zhen Li
Abstract:
Sparse Perception Models (SPMs) adopt a query-driven paradigm that forgoes explicit dense BEV or volumetric construction, enabling highly efficient computation and accelerated inference. In this paper, we introduce SQS, a novel query-based splatting pre-training specifically designed to advance SPMs in autonomous driving. SQS introduces a plug-in module that predicts 3D Gaussian representations from sparse queries during pre-training, leveraging self-supervised splatting to learn fine-grained contextual features through the reconstruction of multi-view images and depth maps. During fine-tuning, the pre-trained Gaussian queries are seamlessly integrated into downstream networks via query interaction mechanisms that explicitly connect pre-trained queries with task-specific queries, effectively accommodating the diverse requirements of occupancy prediction and 3D object detection. Extensive experiments on autonomous driving benchmarks demonstrate that SQS delivers considerable performance gains across multiple query-based 3D perception tasks, notably in occupancy prediction and 3D object detection, outperforming prior state-of-the-art pre-training approaches by a significant margin (i.e., +1.3 mIoU on occupancy prediction and +1.0 NDS on 3D detection).



Paperid:4027
Authors:Siyuan Xu, Minghui Zhu
Abstract:
This paper studies the safe meta-reinforcement learning (safe meta-RL) problem where anytime safety is ensured during the meta-test. We develop a safe meta-RL framework that consists of two modules, safe policy adaptation and safe meta-policy training, and propose efficient algorithms for the two modules. Beyond existing safe meta-RL analyses, we prove the anytime safety guarantee of policy adaptation and provide a lower bound of the expected total reward of the adapted policies compared with the optimal policies, which shows that the adapted policies are nearly optimal. Our experiments demonstrate three key advantages over existing safe meta-RL methods: (i) superior optimality, (ii) anytime safety guarantee, and (iii) high computational efficiency.



Paperid:4028
Authors:Weihao Xuan, Junjue Wang, Heli Qi, Zihang Chen, Zhuo Zheng, Yanfei Zhong, Junshi Xia, Naoto YOKOYA
Abstract:
Multimodal large language models have demonstrated remarkable capabilities in visual understanding, but their application to long-term Earth observation analysis remains limited, primarily focusing on single-temporal or bi-temporal imagery. To address this gap, we introduceDVL-Suite, a comprehensive framework for analyzing long-term urban dynamics through remote sensing imagery. Our suite comprises 15,063 high-resolution (1.0m) multi-temporal images spanning 42 megacities in the U.S. from 2005 to 2023, organized into two components:DVL-BenchandDVL-Instruct. The DVL-Bench includes seven urban understanding tasks, from fundamental change detection (pixel-level) to quantitative analyses (regional-level) and comprehensive urban narratives (scene-level), capturing diverse urban dynamics including expansion/transformation patterns, disaster assessment, and environmental challenges. We evaluate 17 state-of-the-art multimodal large language models and reveal their limitations in long-term temporal understanding and quantitative analysis. These challenges motivate the creation ofDVL-Instruct, a specialized instruction-tuning dataset designed to enhance models' capabilities in multi-temporal Earth observation. Building upon this dataset, we developDVLChat, a baseline model capable of both image-level question-answering and pixel-level segmentation, facilitating a comprehensive understanding of city dynamics through language interactions.



Paperid:4029
Authors:Ziqi Yuan, Haoyang Zhang, Yirui Zhou, Apoorve Mohan, I-Hsin Chung, Seetharami Seelam, Jian Huang
Abstract:
We present the design and implementation of a new lifetime-aware tensor offloadingframework for GPU memory expansion using low-cost PCIe-based solid-statedrives (SSDs). Our framework, TERAIO, is developed explicitly for large languagemodel (LLM) training with multiple GPUs and multiple SSDs. Its design is drivenby our observation that the active tensors take only a small fraction (1.7% onaverage) of allocated GPU memory in each LLM training iteration, the inactivetensors are usually large and will not be used for a long period of time, creatingample opportunities for offloading/prefetching tensors to/from slow SSDs withoutstalling the GPU training process. TERAIO accurately estimates the lifetime (activeperiod of time in GPU memory) of each tensor with the profiling of the first fewiterations in the training process. With the tensor lifetime analysis, TERAIO willgenerate an optimized tensor offloading/prefetching plan and integrate it into thecompiled LLM program via PyTorch. TERAIO has a runtime tensor migrationengine to execute the offloading/prefetching plan via GPUDirect storage, whichallows direct tensor migration between GPUs and SSDs for alleviating the CPUbottleneck and maximizing the SSD bandwidth utilization. In comparison withstate-of-the-art studies such as ZeRO-Offload and ZeRO-Infinity, we show thatTERAIO improves the training performance of various LLMs by 1.47× on average,and achieves 80.7% of the ideal performance assuming unlimited GPU memory.



Paperid:4030
Authors:Yixiao Huang, Hanlin Zhu, Tianyu Guo, Jiantao Jiao, Somayeh Sojoudi, Michael Jordan, Stuart J Russell, Song Mei
Abstract:
Abstract:Recent work found that large language models (LLMs) are able to deduce implications from learned facts during training. For example, when seeing individuals who both live in Paris and speak French in the training data, LLM will automatically deduce that a new person speaks French if the model learns that the same person also lives in Paris. In this paper, we theoretically study how transformers learn to deduce such implications from facts. Surprisingly, we find that a one-layer attention-only transformer is enough to learn to deduce implications from learned facts -- a task typically framed as a two-hop problem that would seemingly require at least a two-layer transformer. Our theoretical analysis on the output-value matrix parameters $(\boldsymbol{W}\_{\mathsf O},\boldsymbol{W}\_{\mathsf V})$ shows that the generalization capability can be attributed to the implicit bias of gradient descent. Specifically, gradient descent directionally converges to an SVM solution minimizing the nuclear norm of the combined parameter $\boldsymbol{W}\_{\mathsf{OV}}:= \boldsymbol{W}\_{\mathsf O}\boldsymbol{W}\_{\mathsf V}^\top$. Further analysis of this SVM solution reveals that one-layer attention-only transformers can associate facts and implications with only a few training samples, no matter whether the fact and implication are causally correlated. Finally, we empirically verify our finding in both a one-layer attention model and several real-world LLMs (e.g., Gemma-2-9B). While this sample-efficient reasoning pattern equips LLMs with strong generalization capability, it also causes LLMs to hallucinate easily by deducing false implications when two events are not causally related but only co-occur several times, which provides a possible explanation of hallucination in LLMs.



Paperid:4031
Authors:Dongki Kim, Wonbin Lee, Sung Ju Hwang
Abstract:
Understanding molecules is key to understanding organisms and driving advances in drug discovery, requiring interdisciplinary knowledge across chemistry and biology. Although large molecular language models have achieved notable success in task transfer, they often struggle to accurately analyze molecular features due to limited knowledge and reasoning capabilities. To address this issue, we present Mol-LLaMA, a large molecular language model that grasps the general knowledge centered on molecules and exhibits explainability and reasoning ability. To this end, we design key data types that encompass the fundamental molecular features, taking into account the essential abilities for molecular reasoning. Further, to improve molecular understanding, we propose a module that integrates complementary information from different molecular encoders, leveraging the distinct advantages of molecular representations. Our experimental results demonstrate that Mol-LLaMA is capable of comprehending the general features of molecules and providing informative responses, implying its potential as a general-purpose assistant for molecular analysis.



Paperid:4032
Authors:Alex Lewandowski, Aditya Ramesh, Edan Meyer, Dale Schuurmans, Marlos C. Machado
Abstract:
Continual learning is often motivated by the idea, known as the big world hypothesis, that the "world is bigger" than the agent. Recent problem formulations capture this idea by explicitly constraining an agent relative to the environment. These constraints lead to solutions in which the agent continually adapts to best use its limited capacity, rather than converging to a fixed solution. However, explicit constraints can be ad hoc, difficult to incorporate, and limiting to the effectiveness of scaling up the agent's capacity. In this paper, we characterize a general problem setting in which an agent of any capacity is implicitly constrained. In particular, we consider the implicit constraint faced by an agent embedded in an environment. We introduce a universal-local environment to embed such an agent using computational universality and transition dynamics that depend on a local neighbourhood of the state-space. The embedded agent is implicitly constrained relative to its environment, represented as a partially observable Markov decision process. We then propose interactivity as a measure of an embedded agent's ability to adapt its future behaviour, conditioned on its past behaviour, using Kolmogorov complexity. We show that maximizing interactivity is a continual reinforcement learning problem, meaning that any agent that stops learning is suboptimal. To support experimentation on continual adaptation, we develop a synthetic benchmark in which an interactivity-maximizing agent constructs its own nonstationary stream of experience from which it must continually learn.



Paperid:4033
Authors:Bo Li, Wei Wang, Peng Ye
Abstract:
Abstract:We revisit the problem of private online learning, in which a learner receives a sequence of $T$ data points and has to respond at each time-step a hypothesis. It is required that the entire stream of output hypotheses should satisfy differential privacy. Prior work of Golowich and Livni [2021] established that every concept class $\mathcal{H}$ with finite Littlestone dimension $d$ is privately online learnable in the realizable setting. In particular, they proposed an algorithm that achieves an $O_{d}(\log T)$ mistake bound against an oblivious adversary. However, their approach yields a suboptimal $\tilde{O}\_{d}(\sqrt{T})$ bound against an adaptive adversary. In this work, we present a new algorithm with a mistake bound of $O_{d}(\log T)$ against an adaptive adversary, closing this gap. We further investigate the problem in the agnostic setting, which is more general than the realizable setting as it does not impose any assumptions on the data. We give an algorithm that obtains a sublinear regret of $\tilde{O}_d(\sqrt{T})$ for generic Littlestone classes, demonstrating that they are also privately online learnable in the agnostic setting.



Paperid:4034
Authors:Yunkee Chae, Kyogu Lee
Abstract:
We present MGE-LDM, a unified latent diffusion framework for simultaneous music generation, source imputation, and query-driven source separation. Unlike prior approaches constrained to fixed instrument classes, MGE-LDM learns a joint distribution over full mixtures, submixtures, and individual stems within a single compact latent diffusion model. At inference, MGE-LDM enables (1) complete mixture generation, (2) partial generation (i.e., source imputation), and (3) text-conditioned extraction of arbitrary sources. By formulating both separation and imputation as conditional inpainting tasks in the latent space, our approach supports flexible, class-agnostic manipulation of arbitrary instrument sources. Notably, MGE-LDM can be trained jointly across heterogeneous multi-track datasets (e.g., Slakh2100, MUSDB18, MoisesDB) without relying on predefined instrument categories.



Paperid:4035
Authors:Shufan Shen, Junshu Sun, Qingming Huang, Shuhui Wang
Abstract:
The alignment of vision-language representations endows current Vision-Language Models (VLMs) with strong multi-modal reasoning capabilities. However, the interpretability of the alignment component remains uninvestigated due to the difficulty in mapping the semantics of multi-modal representations into a unified concept set. To address this problem, we propose VL-SAE, a sparse autoencoder that encodes vision-language representations into its hidden activations. Each neuron in the hidden layer correlates to a concept represented by semantically similar images and texts, thereby interpreting these representations with a unified concept set. To establish the neuron-concept correlation, we encourage semantically similar representations to exhibit consistent neuron activations during self-supervised training. First, to measure the semantic similarity of multi-modal representations, we perform their alignment in an explicit form based on cosine similarity. Second, we construct the VL-SAE with a distance-based encoder and two modality-specific decoders to ensure the activation consistency of semantically similar representations. Experiments across multiple VLMs (e.g., CLIP, LLaVA) demonstrate the superior capability of VL-SAE in interpreting and enhancing the vision-language alignment. For interpretation, the alignment between vision and language representations can be understood by comparing their semantics with concepts. For enhancement, the alignment can be strengthened by aligning vision-language representations at the concept level, contributing to performance improvements in downstream tasks, including zero-shot image classification and hallucination elimination. Codes are provided in the supplementary and will be released to GitHub.



Paperid:4036
Authors:Yumeng He, Yunbo Wang
Abstract:
Out-of-distribution (OOD) 3D relighting requires novel view synthesis under unseen lighting conditions that differ significantly from the observed images. Existing relighting methods, which assume consistent light source distributions between training and testing, often degrade in OOD scenarios. We introduceMetaGSto tackle this challenge from two perspectives. First, we propose a meta-learning approach to train 3D Gaussian splatting, which explicitly promotes learning generalizable Gaussian geometries and appearance attributes across diverse lighting conditions, even with biased training data. Second, we embed fundamental physical priors from theBlinn-Phongreflection model into Gaussian splatting, which enhances the decoupling of shading components and leads to more accurate 3D scene reconstruction. Results on both synthetic and real-world datasets demonstrate the effectiveness of MetaGS in challenging OOD relighting tasks, supporting efficient point-light relighting and generalizing well to unseen environment lighting maps.



Paperid:4037
Authors:Yuxuan Zhang, Zhenbo Shi, han ye, Shuchang Wang, Zhidong Yu, Shaowei Wang, Wei Yang
Abstract:
Out-of-distribution (OOD) fine-tuning has emerged as a promising approach for anomaly segmentation. Current OOD fine-tuning strategies typically employ global-level objectives, aiming to guide segmentation models to accurately predict a large number of anomaly pixels. However, these strategies often perform poorly on small anomalies. To address this issue, we propose an instance-level OOD fine-tuning framework, dubbed LNOIB (Leaving No OOD Instance Behind). We start by theoretically analyzing why global-level objectives fail to segment small anomalies. Building on this analysis, we introduce a simple yet effective instance-level objective. Moreover, we propose a feature separation objective to explicitly constrain the representations of anomalies, which are prone to be smoothed by their in-distribution (ID) surroundings. LNOIB integrates these objectives to enhance the segmentation of small anomalies and serves as a paradigm adaptable to existing OOD fine-tuning strategies, without introducing additional inference cost. Experimental results show that integrating LNOIB into various OOD fine-tuning strategies yields significant improvements, particularly in component-level results, highlighting its strength in comprehensive anomaly segmentation.



Paperid:4038
Authors:Yandan Yang, Baoxiong Jia, Shujie Zhang, Siyuan Huang
Abstract:
Indoor scene synthesis has become increasingly important with the rise of Embodied AI, which requires 3D environments that are not only visually realistic but also physically plausible and functionally diverse. While recent approaches have advanced visual fidelity, they often remain constrained to fixed scene categories, lack sufficient object-level detail and physical consistency, and struggle to align with complex user instructions. In this work, we present SceneWeaver, a reflective agentic framework that unifies diverse scene synthesis paradigms through tool-based iterative refinement. At its core, SceneWeaver employs a language model-based planner to select from a suite of extensible scene generation tools, ranging from data-driven generative models to visual- and LLM-based methods, guided by self-evaluation of physical plausibility, visual realism, and semantic alignment with user input. This closed-loop reason-act-reflect design enables the agent to identify semantic inconsistencies, invoke targeted tools, and update the environment over successive iterations. Extensive experiments on both common and open-vocabulary room types demonstrate that \model not only outperforms prior methods on physical, visual, and semantic metrics, but also generalizes effectively to complex scenes with diverse instructions, marking a step toward general-purpose 3D environment generation.



Paperid:4039
Authors:Szymon Plotka, Gizem Mert, Maciej Chrabaszcz, Ewa Szczurek, Arkadiusz Sitek
Abstract:
In recent years, artificial intelligence has significantly advanced medical image segmentation. However, challenges remain, including efficient 3D medical image processing across diverse modalities and handling data variability. In this work, we introduce Hierarchical Soft Mixture-of-Experts (HoME), a two-level token-routing layer for efficient long-context modeling, specifically designed for 3D medical image segmentation. Built on the Mamba state-space model (SSM) backbone, HoME enhances sequential modeling through sparse, adaptive expert routing. The first stage employs a Soft Mixture-of-Experts (SMoE) layer to partition input sequences into local groups, routing tokens to specialized per-group experts for localized feature extraction. The second stage aggregates these outputs via a global SMoE layer, enabling cross-group information fusion and global context refinement. This hierarchical design, combining local expert routing with global expert refinement improves generalizability and segmentation performance, surpassing state-of-the-art results across datasets from the three most commonly used 3D medical imaging modalities and data quality.



Paperid:4040
Authors:Emile Anand, Sarah Liaw
Abstract:
Thompson Sampling (TS) is widely used to address the exploration/exploitation tradeoff in contextual bandits, yet recent theory shows that it does not explore aggressively enough in high-dimensional problems. Feel-Good Thompson Sampling (FG-TS) addresses this by adding an optimism bonus that favors high-reward models, and it achieves the minimax-optimal regret in the linear setting when posteriors are exact. However, its performance withapproximateposteriors--common in large-scale or neural problems--has not been benchmarked. We provide the first systematic study of FG-TS and its smoothed variant (SFG-TS) across eight real-world and synthetic benchmarks. We compare regimes with exact posteriors (linear and logistic bandits) to approximate regimes produced by fast but coarse stochastic-gradient samplers. Ablations over preconditioning, bonus scale, and prior strength reveal a trade-off: larger bonuses help when posterior samples are accurate, but hurt when sampling noise dominates. FG-TS generally outperforms vanilla TS in linear and logistic bandits, but tends to be weaker in neural bandits. Since FG-TS and its variants are competitive and easy-to-use, we recommend them as baselines in modern contextual-bandit benchmarks.



Paperid:4041
Authors:Yurun Yuan, Fan Chen, Zeyu Jia, Alexander Rakhlin, Tengyang Xie
Abstract:
Abstract:Policy-based methods currently dominate reinforcement learning (RL) pipelines for large language model (LLM) reasoning, leaving value-based approaches largely unexplored. We revisit the paradigm of Bellman Residual Minimization---a decades-old concept---and introduce Trajectory Bellman Residual Minimization (TBRM), an algorithm that naturally adapts this classical idea to LLMs, yielding a simple yet effective off-policy algorithm that optimizes a single trajectory-level Bellman objective using the model's own logits as $Q$-values. TBRM removes the need for critics, importance-sampling ratios, or clipping, and functions with only one rollout per prompt. We prove convergence to the near-optimal KL-regularized policy from arbitrary off-policy data via an improved change-of-trajectory-measure analysis. Experiments on standard mathematical-reasoning benchmarks show that TBRM consistently outperforms policy-based baselines, like PPO and GRPO, with comparable or lower computational and memory overhead. Our results indicate that value-based RL might be a principled and efficient alternative for enhancing reasoning capabilities in LLMs.



Paperid:4042
Authors:Wenjing Tang, Xinyu He, Yongxi Huang, Yunxiao Xiao, Cewu Lu, Panpan Cai
Abstract:
Task planning under uncertainty is essential for home-service robots operating in the real world. Tasks involve ambiguous human instructions, hidden or unknown object locations, and open-vocabulary object types, leading to significant open-ended uncertainty and a boundlessly large planning space. To address these challenges, we propose Tru-POMDP, a planner that combines structured belief generation using Large Language Models (LLMs) with principled POMDP planning. Tru-POMDP introduces a hierarchical Tree of Hypotheses (TOH), which systematically queries an LLM to construct high-quality particle beliefs over possible world states and human goals. We further formulate an open-ended POMDP model that enables rigorous Bayesian belief tracking and efficient belief-space planning over these LLM-generated hypotheses. Experiments on complex object rearrangement tasks across diverse kitchen environments show that Tru-POMDP significantly outperforms state-of-the-art LLM-based and LLM-tree-search hybrid planners, achieving higher success rates with significantly better plans, stronger robustness to ambiguity and occlusion, and greater planning efficiency.



Paperid:4043
Authors:Sibo Tian, Xin Wang, Zeyang Zhang, Haibo Chen, Wenwu Zhu
Abstract:
Graph anomaly detection (GAD) is widely prevalent in scenarios such as financial fraud detection, anti-money laundering, and social bot detection. However, structural distribution shifts are commonly observed in real-world GAD data due to selection bias, resulting in reduced homophily. Existing GAD methods tend to rely on homophilic shortcuts when trained on high-homophily structures, limiting their ability to generalize well to data with low homophily under structural distribution shifts. In this study, we propose to handle structural distribution shifts by generating novel environments characterized by diverse homophilic structures and utilizing invariant patterns, i.e., features and structures with the capability of stable prediction across structural distribution shifts, which face two challenges: (1) How to discover invariant patterns from entangled features and structures, as structures are sensitive to varying homophilic distributions. (2) How to systematically construct new environments with diverse homophilic structures. To address these challenges, we propose the Ego-Neighborhood Disentangled Encoder with Homophily-aware Environment Mixup (HEM), which effectively handles structural distribution shifts in GAD by discovering invariant patterns. Specifically, we first propose an ego-neighborhood disentangled encoder to decouple the learning of feature embeddings and structural embeddings, which facilitates subsequent improvements in the invariance of structural embeddings for prediction. Next, we introduce a homophily-aware environment mixup that dynamically adjusts edge weights through adversarial learning, effectively generating environments with diverse structural distributions. Finally, we iteratively train the classifier and environment mixup via adversarial training, simultaneously improving the diversity of constructed environments and discovering invariant patterns under structural distribution shifts. Extensive experiments on real-world datasets demonstrate that our method outperforms existing baselines and achieves state-of-the-art performance under structural distribution shift conditions.



Paperid:4044
Authors:Luting Wang, Yinghao Xiang, Hongliang Huang, Dongjun Li, Chen Gao, Si Liu
Abstract:
Abstract:Agile Earth Observation Satellites (AEOSs) constellations offer unprecedented flexibility for monitoring the Earth’s surface, but their scheduling remains challenging under large-scale scenarios, dynamic environments, and stringent constraints. Existing methods often simplify these complexities, limiting their real-world performance. We address this gap with a unified framework integrating a standardized benchmark suite and a novel scheduling model. Our benchmark suite, AEOS-Bench, contains $3,907$ finely tuned satellite assets and $16,410$ scenarios. Each scenario features $1$ to $50$ satellites and $50$ to $300$ imaging tasks. These scenarios are generated via a high-fidelity simulation platform, ensuring realistic satellite behavior such as orbital dynamics and resource constraints. Ground truth scheduling annotations are provided for each scenario. To our knowledge, AEOS-Bench is the first large-scale benchmark suite tailored for realistic constellation scheduling. Building upon this benchmark, we introduce AEOS-Former, a Transformer-based scheduling model that incorporates a constraint-aware attention mechanism. A dedicated internal constraint module explicitly models the physical and operational limits of each satellite. Through simulation-based iterative learning, AEOS-Former adapts to diverse scenarios, offering a robust solution for AEOS constellation scheduling. Experimental results demonstrate that AEOS-Former outperforms baseline models in task completion and energy efficiency, with ablation studies highlighting the contribution of each model component. Code and sample data are provided in the supplemental materials.



Paperid:4045
Authors:Zhong Li, Qi Huang, Yuxuan Zhu, Lincen Yang, Mohammad Mohammadi Amiri, Niki van Stein, Matthijs van Leeuwen
Abstract:
We introduce Time-Conditioned Contraction Matching (TCCM), a novel method for semi-supervised anomaly detection in tabular data. TCCM is inspired by flow matching, a recent generative modeling framework that learns velocity fields between probability distributions and has shown strong performance compared to diffusion models and generative adversarial networks. Instead of directly applying flow matching as originally formulated, TCCM builds on its core idea—learning velocity fields between distributions—but simplifies the framework by predicting a time-conditioned contraction vector toward a fixed target (the origin) at each sampled time step. This design offers three key advantages: (1) a lightweight and scalable training objective that removes the need for solving ordinary differential equations during training and inference; (2) an efficient scoring strategy called one time-step deviation, which quantifies deviation from expected contraction behavior in a single forward pass, addressing the inference bottleneck of existing continuous-time models such as DTE (a diffusion-based model with leading anomaly detection accuracy but heavy inference cost);and (3) explainability and provable robustness, as the learned velocity field operates directly in input space, making the anomaly score inherently feature-wise attributable; moreover, the score function is Lipschitz-continuous with respect to the input, providing theoretical guarantees under small perturbations. Extensive experiments on the ADBench benchmark show that TCCM strikes a favorable balance between detection accuracy and inference cost, outperforming state-of-the-art methods—especially on high-dimensional and large-scale datasets. The source code is provided at https://anonymous.4open.science/r/TCCM-502C



Paperid:4046
Authors:Yousef Al-Jazzazi, Haya Diwan, Jinrui Gou, Cameron Musco, Christopher Musco, Torsten Suel
Abstract:
Nearest neighbor search is central in machine learning, information retrieval, and databases.For high-dimensional datasets, graph-based methods such as HNSW, DiskANN, and NSG have become popular thanks to their empirical accuracy and efficiency. These methods construct a directed graph over the dataset and perform beam search on the graph to find nodes close to a given query. While significant work has focused on practical refinements and theoretical understanding of graph-based methods, many questions remain. We propose a new distance-based termination condition for beam search to replace the commonly used condition based on beam width. We prove that, as long as the search graph is \emph{navigable}, our resulting {Adaptive Beam Search} method is guaranteed to approximately solve the nearest-neighbor problem, establishing a connection between navigability and the performance of graph-based search. We also provide extensive experiments on our new termination condition for both navigable graphs and approximately navigable graphs used in practice, such as HNSW and Vamana graphs. We find that Adaptive Beam Search outperforms standard beam search over a range of recall values, data sets, graph constructions, and target number of near neighbors. It thus provides a simple and practical way to improve the performance of popular methods.



Paperid:4047
Authors:Zhiwen Chen, Zhuoren Li, Bo Leng, Hanming Deng, Guizhe Jin, Ran Yu, Huanxi Wen
Abstract:
Integrating the understanding and reasoning capabilities of Large Language Models (LLM) with the self-learning capabilities of Reinforcement Learning (RL) enables more reliable driving performance under complex driving conditions. There has been a lot of work exploring LLM-Dominated RL methods in the field of autonomous driving motion planning. These methods, which utilize LLM to directly generate policies or provide decisive instructions during policy learning of RL agent, are centrally characterized by an over-reliance on LLM outputs. However, LLM outputs are susceptible to hallucinations. Evaluations show that state-of-the-art LLM indicates a non-hallucination rate of only approximately 57.95\% when assessed on essential driving-related tasks. Thus, in these methods, hallucinations from the LLM can directly jeopardize the performance of driving policies. This paper argues that maintaining relative independence between the LLM and the RL is vital for solving the hallucinations problem. Consequently, this paper is devoted to propose a novel LLM-Hinted RL paradigm. The LLM is used to generate semantic hints for state augmentation and policy optimization to assist RL agent in motion planning, while the RL agent counteracts potential erroneous semantic indications through policy learning to achieve excellent driving performance. Based on this paradigm, we propose the HCRMP (LLM-Hinted Contextual Reinforcement Learning Motion Planner) architecture, which is designed that includes ①Augmented Semantic Representation Module to extend state space. ②Contextual Stability Anchor Module enhances the reliability of multi-critic weight hints by utilizing information from the knowledge base. ③Semantic Cache Module is employed to seamlessly integrate LLM low-frequency guidance with RL high-frequency control. Extensive experiments in CARLA validate HCRMP's strong overall driving performance. HCRMP achieves a task success rate of up to 80.3\% under diverse driving conditions with different traffic densities. Under safety-critical driving conditions, HCRMP significantly reduces the collision rate by 11.4\%, which effectively improves the driving performance in complex scenarios.



Paperid:4048
Authors:Konstantinos Tsiolis, Alireza Mousavi-Hosseini, Murat Erdogdu
Abstract:
To understand feature learning dynamics in neural networks, recent theoretical works have focused on gradient-based learning of Gaussian single-index models, where the label is a nonlinear function of a latent one-dimensional projection of the input. While the sample complexity of online SGD is determined by the information exponent of the nonlinear link, recent works improved this by reusing samples or modifying the loss function --- transformations which introduce non-correlational updates --- and instead are limited by the (potentially much smaller) generative exponent. However, this picture is only valid if the learning rate is sufficiently large. In this paper, we characterize the relationship between learning rate and sample complexity for a broad class of gradient-based algorithms that encapsulates both correlational and non-correlational updates, and demonstrate a phase transition from an "information exponent regime" with small learning rate to a "generative exponent regime" with large learning rate. Our framework covers prior analyses of one-pass SGD and SGD with batch reuse, while also introducing a new layer-wise training algorithm that leverages a two-timescales approach to go beyond correlational queries without reusing samples or modifying the loss from squared error. Our theoretical study demonstrates that the choice of learning rate is as important as the design of the algorithm in achieving statistical and computational efficiency.



Paperid:4049
Authors:Changmin Lee, Jihyun Lee, Tae-Kyun Kim
Abstract:
While there has been significant progress in the field of 3D avatar creation from visual observations, modeling physically plausible dynamics of humans with loose garments remains a challenging problem. Although a few existing works address this problem by leveraging physical simulation, they suffer from limited accuracy or robustness to novel animation inputs. In this work, we present MPMAvatar, a framework for creating 3D human avatars from multi-view videos that supports highly realistic, robust animation, as well as photorealistic rendering from free viewpoints. For accurate and robust dynamics modeling, our key idea is to use a Material Point Method-based simulator, which we carefully tailor to model garments with complex deformations and contact with the underlying body by incorporating an anisotropic constitutive model and a novel collision handling algorithm. We combine this dynamics modeling scheme with our canonical avatar that can be rendered using 3D Gaussian Splatting with quasi-shadowing, enabling high-fidelity rendering for physically realistic animations. In our experiments, we demonstrate that MPMAvatar significantly outperforms the existing state-of-the-art physics-based avatar in terms of (1) dynamics modeling accuracy, (2) rendering accuracy, and (3) robustness and efficiency. Additionally, we present a novel application in which our avatar generalizes to unseen interactions in a zero-shot manner—which was not achievable with previous learning-based methods due to their limited simulation generalizability. Our code will be publicly available.



Paperid:4050
Authors:Kerui Ren, Jiayang Bai, Linning Xu, Lihan Jiang, Jiangmiao Pang, Mulin Yu, Bo Dai
Abstract:
Object compositing offers significant promise for augmented reality (AR) and embodied intelligence applications. Existing approaches predominantly focus on single-image scenarios or intrinsic decomposition techniques, facing challenges with multi-view consistency, complex scenes, and diverse lighting conditions. Recent inverse rendering advancements, such as 3D Gaussian and diffusion-based methods, have enhanced consistency but are limited by scalability, heavy data requirements, or prolonged reconstruction time per scene. To broaden its applicability, we introduce MV-CoLight, a two-stage framework for illumination-consistent object compositing in both 2D images and 3D scenes. Our novel feed-forward architecture models lighting and shadows directly, avoiding the iterative biases of diffusion-based methods. We employ a Hilbert curve–based mapping to align 2D image inputs with 3D Gaussian scene representations seamlessly. To facilitate training and evaluation, we further introduce a large-scale 3D compositing dataset. Experiments demonstrate state-of-the-art harmonized results across standard benchmarks and our dataset, as well as casually captured real-world scenes demonstrate the framework’s robustness and wide generalization.



Paperid:4051
Authors:Saibo Geng, Nathan Ranchin, Yunzhen Yao, Maxime Peyrard, Chris Wendler, Michael Gastpar, Robert West
Abstract:
Tokenization efficiency plays a critical role in the performance and cost of large language models (LLMs), yet most models rely on static tokenizers optimized for general-purpose corpora. These tokenizers' fixed vocabularies often fail to adapt to domain- or language-specific inputs, leading to longer token sequences and higher computational costs. We introduce zip2zip, a framework that enables LLMs to dynamically adjust token vocabulary at inference time, allowing for fewer generated tokens and thus faster inference. zip2zip consists of three key components: (1) a tokenizer based on Lempel-Ziv-Welch (LZW) compression that incrementally compresses tokens into reusable “hypertokens” on the fly; (2) an embedding layer that computes embeddings for newly formed hypertokens at runtime ; and (3) a causal language modeling variant that trains the model to operate on hypertokenized, compressed sequences. We show that an existing LLM can be zip2zip-fied in 10 GPU-hours via parameter-efficient finetuning. The resulting zip2zip LLMs effectively learn to use hypertokens at inference time, reducing input and output sequence length by 20–60%, with significant improvements in inference latency.



Paperid:4052
Authors:Ming Cheng, Hao Chen, Zhiqing Li, Jia Wang, Senzhang Wang
Abstract:
Emergency traffic control presents an increasingly critical challenge, requiring seamless coordination among emergency vehicles, regular vehicles, and traffic lights to ensure efficient passage for all vehicles. Existing models primarily only focus on traffic light control, leaving emergency and regular vehicles prone to delay due to the lack of navigation strategies. To address this issue, we propose theRole-awareMulti-agentTrafficControl (RMTC)framework, which dynamically assigns appropriate roles to traffic components for better cooperation by considering their relations with emergency vehicles and adaptively adjusting their policies. Specifically, RMTC introduces aHeterogeneous Temporal Traffic Graph (HTTG)to model the spatial and temporal relationships among all traffic components (traffic lights, regular and emergency vehicles) at each time step. Furthermore, we develop aDynamic Role Learningmodel to infer the evolving roles of traffic lights and regular vehicles based on HTTG. Finally, we present aRole-aware Multi-agent Reinforcement Learningapproach that learns traffic policies conditioned on the dynamically roles. Extensive experiments across four public traffic scenarios show that RMTC outperforms existing traffic light control methods by significantly reducing emergency vehicle travel time, while effectively preserving traffic efficiency for regular vehicles. The code is released athttps://anonymous.4open.science/r/RMTC-5E28.



Paperid:4053
Authors:Yiping Wang, Qing Yang, Zhiyuan Zeng, Liliang Ren, Liyuan Liu, Baolin Peng, Hao Cheng, Xuehai He, Kuan Wang, Jianfeng Gao, Weizhu Chen, Shuohang Wang, Simon Du, yelong shen
Abstract:
We show that reinforcement learning with verifiable reward using one training example (1-shot RLVR) is effective in incentivizing the math reasoning capabilities of large language models (LLMs). Applying RLVR to the base model Qwen2.5-Math-1.5B, we identify a single example that elevates model performance on MATH500 from 36.0% to 73.6%, and improves the average performance across six common mathematical reasoning benchmarks from 17.6% to 35.7%. This result matches the performance obtained using the 1.2k DeepScaleR subset (MATH500: 73.6%, average: 35.9%), which includes the aforementioned example. Similar substantial improvements are observed across various models (Qwen2.5-Math-7B, Llama3.2-3B-Instruct, DeepSeek-R1-Distill-Qwen-1.5B), RL algorithms (GRPO and PPO), and different math examples (many of which yield approximately 30% or greater improvement on MATH500 when employed as a single training example). In addition, we identify some interesting phenomena during 1-shot RLVR, including cross-domain generalization, increased frequency of self-reflection, and sustained test performance improvement even after the training accuracy has saturated, a phenomenon we term post-saturation generalization. Moreover, we verify that the effectiveness of 1-shot RLVR primarily arises from the policy gradient loss, distinguishing it from the "grokking" phenomenon. We also show the critical role of promoting exploration (e.g., by adding entropy loss with an appropriate coefficient) in 1-shot RLVR training. As a bonus, we observe that applying entropy loss alone, without any outcome reward, significantly enhances Qwen2.5-Math-1.5B's performance on MATH500 by 27.4%. These findings can inspire future work on RLVR data efficiency and encourage a re-examination of both recent progress and the underlying mechanisms in RLVR.



Paperid:4054
Authors:Marcel Kollovieh, Nils Fleischmann, Filippo Guerranti, Bertrand Charpentier, Stephan Günnemann
Abstract:
In this work, we introduce TreeGen, a novel generative framework modeling distributions over hierarchies. We extend Bayesian Flow Networks (BFNs) to enable transitions between probabilistic and discrete hierarchies parametrized via categorical distributions. Our proposed scheduler provides smooth and consistent entropy decay across varying numbers of categories. We empirically evaluate TreeGen on the jet-clustering task in high-energy physics, demonstrating that it consistently generates valid trees that adhere to physical constraints and closely align with ground-truth log-likelihoods. Finally, by comparing TreeGen’s samples to the exact posterior distribution and performing likelihood maximization via rejection sampling, we demonstrate that TreeGen outperforms various baselines.



Paperid:4055
Authors:Kongcheng Zhang, QI YAO, Shunyu Liu, Yingjie Wang, Baisheng Lai, Jieping Ye, Mingli Song, Dacheng Tao
Abstract:
In this work, we propose a novel self-rewarding reinforcement learning framework to enhance Large Language Model (LLM) reasoning by leveraging the consistency of intermediate reasoning states across different response trajectories. Our key insight is that correct responses often exhibit consistent trajectory patterns in terms of model likelihood: their intermediate reasoning states tend to converge toward their own final answers (high consistency) with minimal deviation toward other candidates (low volatility). Inspired by this observation, we introduce CoVo, an intrinsic reward mechanism that integrates Consistency and Volatility via a robust vector-space aggregation strategy, complemented by a curiosity bonus to promote diverse exploration. CoVo enables LLMs to perform reinforcement learning in a self-rewarding manner, offering a scalable pathway for learning to reason without external supervision. Moreover, our theoretical analysis reveals that CoVo in fact treats reasoning paths as latent variables within a variational inference objective, thereby enabling the joint optimization of both reasoning paths and final answers. Extensive experiments on diverse reasoning benchmarks show that CoVo achieves performance comparable to or even surpassing supervised RL, boosting Qwen2.5-3B-Instruct from 63.6% to 68.2% on MATH dataset. Our code is available at https://anonymous.4open.science/r/CoVo.



Paperid:4056
Authors:Jie Yan, Jing Liu, Zhong-Yuan Zhang
Abstract:
Federated clustering (FC) aims to discover global cluster structures across decentralized clients without sharing raw data, making privacy preservation a fundamental requirement. There are two critical challenges: (1) privacy leakage during collaboration, and (2) robustness degradation due to aggregation of proxy information from non-independent and identically distributed (Non-IID) local data, leading to inaccurate or inconsistent global clustering. Existing solutions typically rely on model-specific local proxies, which are sensitive to data heterogeneity and inherit inductive biases from their centralized counterparts, thus limiting robustness and generality. We propose Omni Federated Clustering (OmniFC), a unified and model-agnostic framework. Leveraging Lagrange coded computing, our method enables clients to share only encoded data, allowing exact reconstruction of the global distance matrix—a fundamental representation of sample relationships—without leaking private information, even under client collusion. This construction is naturally resilient to Non-IID data distributions. This approach decouples FC from model-specific proxies, providing a unified extension mechanism applicable to diverse centralized clustering methods. Theoretical analysis confirms both reconstruction fidelity and privacy guarantees, while comprehensive experiments demonstrate OmniFC's superior robustness, effectiveness, and generality across various benchmarks compared to state-of-the-art methods. Code will be released.



Paperid:4057
Authors:Tyler Farghly, Peter Potaptchik, Samuel Howard, George Deligiannidis, Jakiw Pidstrigach
Abstract:
Diffusion models have achieved state-of-the-art performance, demonstrating remarkable generalisation capabilities across diverse domains. However, the mechanisms underpinning these strong capabilities remain only partially understood. A leading conjecture, based on the manifold hypothesis, attributes this success to their ability to adapt to low-dimensional geometric structure within the data. This work provides evidence for this conjecture, focusing on how such phenomena could result from the formulation of the learning problem through score matching. We inspect the role of implicit regularisation by investigating the effect of smoothing minimisers of the empirical score matching objective. Our theoretical and empirical results confirm that smoothing the score function—or equivalently, smoothing in the log-density domain—produces smoothing tangential to the data manifold. In addition, we show that the manifold along which the diffusion model generalises can be controlled by choosing an appropriate smoothing.



Paperid:4058
Authors:Jincheng Cao, Ruichen Jiang, Erfan Yazdandoost Hamedani, Aryan Mokhtari
Abstract:
Abstract:In this paper, we study the problem of solving a simple bilevel optimization problem, where the upper-level objective is minimized over the solution set of the lower-level problem. We focus on the general setting in which both the upper- and lower-level objectives are smooth but potentially nonconvex. Due to the absence of additional structural assumptions for the lower-level objective—such as convexity or the Polyak–Łojasiewicz (PL) condition—guaranteeing global optimality is generally intractable. Instead, we introduce a suitable notion of stationarity for this class of problems and aim to design a first-order algorithm that finds such stationary points in polynomial time. Intuitively, stationarity in this setting means the upper-level objective cannot be substantially improved locally without causing a larger deterioration in the lower-level objective. To this end, we show that a simple and implementable variant of the dynamic barrier gradient descent (DBGD) framework can effectively solve the considered nonconvex simple bilevel problems up to stationarity. Specifically, to reach an $(\epsilon_f, \epsilon_g)$-stationary point—where $\epsilon_f$ and $\epsilon_g$ denote the target stationarity accuracies for the upper- and lower-level objectives, respectively—the considered method achieves a complexity of $\mathcal{O}(\max(\epsilon_f^{-\frac{3+p}{1+p}}, \epsilon_g^{-\frac{3+p}{2}}))$, where $p \geq 0$ is an arbitrary constant balancing the terms. To the best of our knowledge, this is the first complexity result for a discrete-time algorithm that guarantees joint stationarity for both levels in general nonconvex simple bilevel problems.



Paperid:4059
Authors:Zaoming Yan, Yaomin Huang, Qizhou Chen, Pengcheng Lei, Guixu Zhang, Faming Fang
Abstract:
Motion in the real world takes place in 3D space.Existing Frame Interpolation methods often estimate global receptive fields in 2D frame space. Due to the limitations of 2D space, these global receptive fields are limited, which makes it difficult to match object correspondences between frames, resulting in sub-optimal performance when handling large-motion scenarios.In this paper, we introduce a novel pipeline for exploring object correspondences based on differential surface theory.The differential surface coordinate system provides a better representation of the real world, enabling effective exploration of object correspondences.Specifically, the pipeline first transforms an input pair of video frames from the image coordinate system to the differential surface coordinate system.Subsequently, within this coordinate system, object correspondences are explored based on surface geometric properties and the surface uniqueness theorem.Experimental findings showcase that our method attains state-of-the-art performance across large motion benchmarks. Our method demonstrates the state-of-the-art performance on these VFI subsets with large motion.



Paperid:4060
Authors:Weiji Xie, Jinrui Han, Jiakun Zheng, Huanyu Li, Xinzhe Liu, Jiyuan Shi, Weinan Zhang, Chenjia Bai, Xuelong Li
Abstract:
Humanoid robots are promising to acquire various skills by imitating human behaviors. However, existing algorithms are only capable of tracking smooth, low-speed human motions, even with delicate reward and curriculum design. This paper presents a physics-based humanoid control framework, aiming to master highly-dynamic human behaviors such as Kungfu and dancing through multi-steps motion processing and adaptive motion tracking. For motion processing, we design a pipeline to extract, filter out, correct, and retarget motions, while ensuring compliance with physical constraints to the maximum extent. For motion imitation, we formulate a bi-level optimization problem to dynamically adjust the tracking accuracy tolerance based on the current tracking error, creating an adaptive curriculum mechanism. We further construct an asymmetric actor-critic framework for policy training. In experiments, we train whole-body control policies to imitate a set of highly dynamic motions. Our method achieves significantly lower tracking errors than existing approaches and is successfully deployed on the Unitree G1 robot, demonstrating stable and expressive behaviors. The project page is https://kungfubot.github.io.



Paperid:4061
Authors:Zipeng Wang, Dan Xu
Abstract:
Recently, 3D Gaussian Splatting (3DGS) has emerged as a powerful alternative to NeRF-based approaches, enabling real-time, high-quality novel view synthesis through explicit, optimizable 3D Gaussians. However, 3DGS suffers from significant memory overhead due to its reliance on per-Gaussian parameters to model view-dependent effects and anisotropic shapes. While recent works propose compressing 3DGS with neural fields, these methods struggle to capture high-frequency spatial variations in Gaussian properties, leading to degraded reconstruction of fine details. We present Hybrid Radiance Fields (HyRF), a novel scene representation that combines the strengths of explicit Gaussians and neural fields. HyRF decomposes the scene into (1) a compact set of explicit Gaussians storing only critical high-frequency parameters and (2) grid-based neural fields that predict remaining properties. To enhance representational capacity, we introduce a decoupled neural field architecture, separately modeling geometry (scale, opacity, rotation) and view-dependent color. Additionally, we propose a hybrid rendering scheme that composites Gaussian splatting with a neural field-predicted background, addressing limitations in distant scene representation.Experiments demonstrate that HyRF achieves state-of-the-art rendering quality while reducing model size by over 20× compared to 3DGS and maintaining real-time performance.



Paperid:4062
Authors:Jack Goffinet, Youngjo Min, Carlo Tomasi, David Carlson
Abstract:
Accurate and scalable quantification of animal pose and appearance is crucial for studying behavior. Current 3D pose estimation techniques, such as keypoint- and mesh-based techniques, often face challenges including limited representational detail, labor-intensive annotation requirements, and expensive per-frame optimization. These limitations hinder the study of subtle movements and can make large-scale analyses impractical. We propose "Pose Splatter", a novel framework leveraging shape carving and 3D Gaussian splatting to model the complete pose and appearance of laboratory animals without prior knowledge of animal geometry, per-frame optimization, or manual annotations. We also propose a novel rotation-invariant visual embedding technique for encoding pose and appearance, designed to be a plug-in replacement for 3D keypoint data in downstream behavioral analyses. Experiments on datasets of mice, rats, and zebra finches show Pose Splatter learns accurate 3D animal geometries. Notably, Pose Splatter represents subtle variations in pose, provides better low-dimensional pose embeddings over state-of-the-art as evaluated by humans, and generalizes to unseen data. By eliminating annotation and per-frame optimization bottlenecks, Pose Splatter enables analysis of large-scale, longitudinal behavior needed to map genotype, neural activity, and micro-behavior at unprecedented resolution.



Paperid:4063
Authors:Sepehr Dehdashtian, Mashrur Mahmud Morshed, Jacob Seidman, Gaurav Bharaj, Vishnu Boddeti
Abstract:
Synthetic image detectors (SIDs) are a key defense against the risks posed by the growing realism of images from text-to-image (T2I) models. Red teaming improves SID’s effectiveness by identifying and exploiting their failure modes via misclassified synthetic images. However, existing red-teaming solutions (i) require white-box access to SIDs, which is infeasible for proprietary state-of-the-art detectors, and (ii) generate image-specific attacks through expensive online optimization. To address these limitations, we propose PolyJuice, the first black-box, image-agnostic red-teaming method for SIDs, based on an observed distribution shift in the T2I latent space between samples correctly and incorrectly classified by the SID. PolyJuice generates attacks by (i) identifying the direction of this shift through a lightweight offline process that only requires black-box access to the SID, and (ii) exploiting this direction by universally steering all generated images towards the SID’s failure modes. PolyJuice-steered T2I models are significantly more effective at deceiving SIDs (up to 84%) compared to their unsteered counterparts. We also show that the steering directions can be estimated efficiently at lower resolutions and transferred to higher resolutions using simple interpolation, reducing computational overhead. Finally, tuning SID models on PolyJuice-augmented datasets notably enhances the performance of the detectors (up to 30%).



Paperid:4064
Authors:Andrew Li, Toryn Klassen, Andrew Wang, Parand A. Alamdari, Sheila McIlraith
Abstract:
Large-scale foundation models have gained broad capabilities through training on large, diverse datasets. But is diverse data necessary for broad capabilities? In this work, we explore a data-efficient approach to building up the capabilities of agents called Ground-Compose-Reinforce. We first ground a predefined set of symbols in the environment by training on a labelled trajectory dataset, which is harnessed by an agent endowed with the compositional semantics of a formal language. This agent can then be tasked through this formal language and can learn to solve any such task through a self-generated learning signal. Controlled experiments on both an image-based gridworld and a MuJoCo robotics domain show that our approach is data-efficient to train from trajectories while supporting diverse, out-of-distribution capabilities by explicitly leveraging compositionality.



Paperid:4065
Authors:Qin Xu, Qihang Wu, Lu Hongtao, Xiaoxia Cheng, Bo Jiang
Abstract:
Scene understanding in adverse conditions, such as fog, snow, and night, is challenging due to the visual appearance degeneration. In this context, we propose a Cross-modal Semantic Compensation Adaptation method (CroPe) for scene understanding. Distinct from the existing methods, which only use the visual information to learn the domain-invariant features, CroPe establishes a visual-textual paradigm which provides textual semantic compensation for visual features, enabling the model to learn more consistent representations. We propose the Complementary Perceptual Text Generation (CPTG) module which generates a set of multi-level complementary-perceptive text embeddings incorporating both generalization and domain awareness. To achieve cross-modal semantic compensation, the Reverse Chain Text-Visual Fusion (RCTVF) module is developed. By the unified attention and reverse decoding chain, compensation information is successively fused to the visual features from the deep (semantic dense) to shallow (semantic sparse) features, maximizing compensation gain. CroPe yields competitive results under all adverse conditions and significantly improves the state-of-the-art performance by 6.5 mIoU for ACDC-Night dataset and 1.2 mIoU for ACDC-All dataset, respectively.



Paperid:4066
Authors:Gerardo Flores, Alyssa H. Smith, Julia Fukuyama, Ashia Wilson
Abstract:
Machine learning-based decision support systems are increasingly deployed in clinical settings, where probabilistic scoring functions are used to inform and prioritize patient management decisions.However, widely used scoring rules, such as accuracy and AUC-ROC, fail to adequately reflect key clinical priorities, including calibration, robustness to distributional shifts, and sensitivity to asymmetric error costs.In this work, we propose a principled yet practical evaluation framework for selecting calibrated thresholded classifiers that explicitly accounts for uncertainty in class prevalences and domain-specific cost asymmetries.Building on the theory of proper scoring rules, particularly the Schervish representation, we derive an adjusted variant of cross-entropy (log score) that averages cost-weighted performance over clinically relevant ranges of class balance.The resulting evaluation is simple to apply, sensitive to clinical deployment conditions, and designed to prioritize models that are both calibrated and robust to real-world variations.



Paperid:4067
Authors:Florian Barthel, Wieland Morgenstern, Paul Hinzer, Anna Hilsmann, Peter Eisert
Abstract:
Abstract:Recently, 3D GANs based on 3D Gaussian splatting have been proposed for high quality synthesis of human heads. However, existing methods stabilize training and enhance rendering quality from steep viewpoints by conditioning the random latent vector on the current camera position. This compromises 3D consistency, as we observe significant identity changes when re-synthesizing the 3D head with each camera shift. Conversely, fixing the camera to a single viewpoint yields high-quality renderings for that perspective but results in poor performance for novel views. Removing view-conditioning typically destabilizes GAN training, often causing the training to collapse. In response to these challenges, we introduce CGS-GAN, a novel 3D Gaussian Splatting GAN framework that enables stable training and high-quality 3D-consistent synthesis of human heads without relying on view-conditioning. To ensure training stability, we introduce a multi-view regularization technique that enhances generator convergence with minimal computational overhead. Additionally, we adapt the conditional loss used in existing 3D Gaussian splatting GANs and propose a generator architecture designed to not only stabilize training but also facilitate efficient rendering and straightforward scaling, enabling output resolutions up to $2048^2$. To evaluate the capabilities of CGS-GAN, we curate a new dataset derived from FFHQ. This dataset enables very high resolutions, focuses on larger portions of the human head, reduces view-dependent artifacts for improved 3D consistency, and excludes images where subjects are obscured by hands or other objects. As a result, our approach achieves very high rendering quality, supported by competitive FID scores, while ensuring consistent 3D scene generation.



Paperid:4068
Authors:Andrés Guzmán-Cordero, Felix Dangel, Gil Goldshlager, Marius Zeinhofer
Abstract:
Abstract:Natural gradient methods significantly accelerate the training of Physics-Informed Neural Networks (PINNs), but are often prohibitively costly. We introduce a suite of techniques to improve the accuracy and efficiency of energy natural gradient descent (ENGD) for PINNs. First, we leverage the Woodbury formula to dramatically reduce the computational complexity of ENGD. Second, we adapt the Subsampled Projected-Increment Natural Gradient Descent algorithm from the variational Monte Carlo literature to accelerate the convergence. Third, we explore the use of randomized algorithms to further reduce the computational cost in the case of large batch sizes. We find that randomization accelerates progress in the early stages of training for low-dimensional problems, and we identify key barriers to attaining acceleration in other scenarios. Our numerical experiments demonstrate that our methods outperform previous approaches, achieving the same $L^2$ error as the original ENGD up to $75\times$ faster.



Paperid:4069
Authors:Sitong Wu, Haoru Tan, Jingyao Li, Shaofeng Zhang, Xiaojuan Qi, Bei Yu, Jiaya Jia
Abstract:
Long chain-of-thought (LongCoT) has emerged as a powerful reasoning paradigm for enabling large language models (LLMs) to solve complex tasks through a systematic and thorough thinking phase.Although supervised fine-tuning (SFT) on high-quality LongCoT traces has proven effective to activate LongCoT abilities, we find that models trained in this way tend to overfit problem-specific knowledge and heuristics, leading to degraded out-of-distribution performance.To address this issue, we propose a Decoupled LongCoT Fine-Tuning (DLoFT) algorithm, which enables the model to learn generalizable LongCoT reasoning abilities while preventing overfitting to the reasoning content with problem-specific information.The key idea is to decouple the gradient into two orthogonal components: 1) a paradigm-relevant gradient corresponding to the general LongCoT paradigm and 2) a content-relevant gradient reflecting the problem-specific information, where only the former gradient is used to update model parameters.Specifically, by leveraging the unique two-phase composition (thinking and solution) of the LongCoT response, our gradient decoupling mechanism isolates the content-relevant gradient via a projection operation and separates the paradigm-relevant gradient through orthogonalization.Our DLoFT ensures the model concentrate on internalizing the LongCoT paradigm rather than memorizing problem-specific knowledge and heuristics.Extensive experiments demonstrate that our DLoFT significantly improves the generalization behavior of LongCoT abilities compared to SFT while maintaining strong in-distribution performance.



Paperid:4070
Authors:Panqi Chen, Yifan Sun, Lei Cheng, YANG YANG, Weichang Li, Yang Liu, Weiqing Liu, Jiang Bian, Shikai Fang
Abstract:
Modeling and reconstructing multidimensional physical dynamics from sparse and off-grid observations presents a fundamental challenge in scientific research. Recently, diffusion-based generative modeling shows promising potential for physical simulation. However, current approaches typically operate on on-grid data with preset spatiotemporal resolution, but struggle with the sparsely observed and continuous nature of real-world physical dynamics. To fill the gaps, we present SDIFT, Sequential DIffusion in Functional Tucker space, a novel framework that generates full-field evolution of physical dynamics from irregular sparse observations. SDIFT leverages the functional Tucker model as the latent space representer with proven universal approximation property, and represents sparse observations as latent functions and Tucker core sequences. We then construct a sequential diffusion model with temporally augmented UNet in the functional Tucker space, denoising noise drawn from a Gaussian process to generate the sequence of core tensors. At the posterior sampling stage, we propose a Message-Passing Posterior Sampling mechanism, enabling conditional generation of the entire sequence guided by observations at limited time steps. We validate SDIFT on three physical systems spanning astronomical (supernova explosions, light-year scale), environmental (ocean sound speed fields, kilometer scale), and molecular (organic liquid, millimeter scale) domains, demonstrating significant improvements in both reconstruction accuracy and computational efficiency compared to state-of-the-art approaches.



Paperid:4071
Authors:Xilin Xia, Jie Wang, Wanbo Zhang, Zhihai Wang, Mingxuan Yuan, Jianye Hao, Feng Wu
Abstract:
Abstract:Arithmetic circuit optimization remains a fundamental challenge in modern integrated circuit design. Recent advances have cast this problem within the Learning to Optimize (L2O) paradigm, where intelligent agents autonomously explore high-performance design spaces with encouraging results. However, existing approaches predominantly target coarse-grained architectural configurations, while the crucial interconnect optimization stage is often relegated to oversimplified proxy models or a heuristic approach. This disconnect undermines design quality, leading to suboptimal solutions in the circuit topology search space. To bridge this gap, we present **Arith-DAS**, a **D**ifferentiable **A**rchitecture **S**earch framework for **Arith**metic circuits. To the best of our knowledge, **Arith-DAS** is the first to formulate interconnect optimization within arithmetic circuits as a differentiable edge prediction problem over a multi-relational directed acyclic graph, enabling fine-grained, proxy-free optimization at the interconnection level. We evaluate **Arith-DAS** on a suite of representative arithmetic circuits, including multipliers and multiply-accumulate units. Experiments show substantial improvements over state-of-the-art L2O and conventional methods, achieving up to $\textbf{27.05}$% gain in hypervolume of area-delay Pareto front, a standard metric for evaluating multi-objective optimization performance. Moreover, integrating our optimized arithmetic units into large-scale AI accelerators yields up to $\textbf{6.59}$% delay reduction, demonstrating both scalability and real-world applicability.



Paperid:4072
Authors:Zhongben Gong, Xiaoqun Wu, Mingyang Zhou
Abstract:
We propose PreBoot-SI, a general bootstrapped framework for structural inference in interacting dynamical systems. It integrates a pretrained structural estimator and a joint inference module based on the Variational Autoencoder (VAE); these components are alternately updated to progressively refine the inferred structures. Initially, the structural estimator is trained on labels from either a meta-dataset or a baseline model to extract features and generate structural priors, which provide multi-level guidance for training the joint inference module. In subsequent iterations, pseudo-labels from the joint module replace the initial labels. PreBoot-SI is compatible with various VAE-based models. Experiments on synthetic datasets of physical systems demonstrate that PreBoot-SI significantly enhances the performance of structural inference models such as Neural Relational Inference (NRI). Ablation studies reveal that feature and structural prior inputs to the joint module offer complementary improvements from representational and generative perspectives.



Paperid:4073
Authors:Tianyi Bai, Yuxuan Fan, Qiu Jiantao, Fupeng Sun, Jiayi Song, Junlin Han, Zichen Liu, Conghui He, Wentao Zhang, Binhang Yuan
Abstract:
Multimodal large language models (MLLMs) have achieved strong performance on vision-language tasks but still struggle with fine-grained visual differences, leading to hallucinations or missed semantic shifts. We attribute this to limitations in both training data and learning objectives. To address these issues, we propose a controlled data generation pipeline that produces minimally edited image pairs with semantically aligned captions. Using this pipeline, we construct the Micro Edit Dataset (MED), containing over 50K image-text pairs spanning 11 fine-grained edit categories, including attribute, count, position, and object presence changes.Building on MED Dataset, we introduce a supervised fine-tuning (SFT) framework with a feature-level consistency loss that promotes stable visual embeddings under small edits. We evaluate our approach on the Micro Edit Detection benchmark, which includes carefully balanced evaluation pairs designed to test sensitivity to subtle visual variations across the same edit categories.Our method improves difference detection accuracy and reduces hallucinations compared to strong baselines, including GPT-4o. Moreover, it yields consistent gains on standard vision-language tasks such as image captioning and visual question answering. These results demonstrate the effectiveness of combining targeted data and alignment objectives for enhancing fine-grained visual reasoning in MLLMs.



Paperid:4074
Authors:Yitian Chen, Jingfan Xia, Siyu Shao, DongDong Ge, Yinyu Ye
Abstract:
Optimization modeling is fundamental to decision-making across diverse domains. Despite progress in automating optimization formulation from natural language descriptions, Large Language Models (LLMs) often struggle to generate formally correct and usable models due to hallucinations, posing a challenge for reliable automation. Inspired by the success of Reinforcement Learning (RL) in enhancing Large Reasoning Models, we present Solver-Informed Reinforcement Learning (SIRL). This novel framework leverages external optimization solvers as verifiable reward mechanisms to significantly improve the authenticity of LLMs for optimization modeling. Acting as precise verifiers, these solvers automatically assess the executable code and the instance-level mathematical model represented by the associated LP file, yielding precise and comprehensive feedback signals---including syntax, feasibility, and solution quality that directly inform the RL process. This automated verification process, powered by classic optimization solvers, also underpins our instance-enhanced self-consistency method to synthesize high-quality training data. Extensive experiments on diverse public benchmarks demonstrate that SIRL achieves state-of-the-art performance, substantially outperforming existing methods in generating accurate and executable optimization models.



Paperid:4075
Authors:Abel Jansma
Abstract:
We introduce a novel framework for decomposing interventional causal effects into synergistic, redundant, and unique components, building on the intuition of Partial Information Decomposition (PID) and the principle of Möbius inversion. While recent work has explored a similar decomposition of an observational measure, we argue that a proper causal decomposition must be interventional in nature. We develop a mathematical approach that systematically quantifies how causal power is distributed among variables in a system, using a recently derived closed-form expression for the Möbius function of the redundancy lattice. The formalism is then illustrated by decomposing the causal power in logic gates, cellular automata, and chemical reaction networks. Our results reveal how the distribution of causal power can be context- and parameter-dependent. The decomposition provides new insights into complex systems by revealing how causal influences are shared and combined among multiple variables, with potential applications ranging from attribution of responsibility in legal or AI systems, to the analysis of biological networks or climate models.



Paperid:4076
Authors:Qiannan Zhang, Manqi Zhou, Zilong Bai, Chang Su, Fei Wang
Abstract:
Clinical risk prediction plays a crucial role in early disease detection and personalized intervention. While recent models increasingly incorporate multimodal data, their development typically assumes access to large-scale, multimodal datasets and substantial computational resources. In practice, however, most clinical sites operate under resource constraints, with access limited to EHR data alone and insufficient capacity to train complicated models. This gap highlights the urgent need to democratize clinical risk prediction by enabling effective deployment in data- and resource-limited local clinical settings. In this work, we propose a cross-cohort cross-modal knowledge transfer framework that leverages the multimodal model trained on a nationwide cohort and adapts it to local cohorts with only EHR data. We focus on EHR and genetic data as representative multimodal inputs and address two key challenges. First, to mitigate the influence of noisy or less informative biological signals, we propose a novel mixture-of-aggregations design to enhance the modeling of informative and relevant genetic features. Second, to support rapid model adaptation in low-resource sites, we develop a lightweight graph-guided fine-tuning method that adapts pretrained phenotypical EHR representations to target cohorts using limited patient data. Extensive experiments on real-world clinical data validate the effectiveness of our proposed model.



Paperid:4077
Authors:Xiyu Zhang, Chong Bao, YiPeng Chen, Hongjia Zhai, Yitong Dong, Hujun Bao, Zhaopeng Cui, Guofeng Zhang
Abstract:
3D reconstruction of indoor and urban environments is a prominent research topic with various downstream applications. However, current geometry prior to address low-texture regions in indoor and urban is not globally consistent.Besides, Gaussian Splatting (GS) or implicit SDF field either suffer from discontinuities or are computationally inefficient with fewer details.To address these issues, we propose an Atlanta-world guided implicit-structured Gaussian Splatting that achieves smooth indoor and urban scene reconstruction while preserving high-frequency details and rendering efficiency.By leveraging the Atlanta-world model, we ensure the accurate surface reconstruction for low-texture regions, while proposed novel implicit-structured GS representations provide smoothness without sacrificing efficiency and high-frequency details.Specifically, we propose a semantic GS representation to predict probability of all semantic regions and deploy a structure plane regularization with learnable plane indicators to global accurate surface reconstruction.Extensive experiments demonstrate that our method outperforms state-of-the-art approaches in both indoor and urban scenes, delivering superior surface reconstruction quality.



Paperid:4078
Authors:Hanru Bai, Weiyang Ding, Difan Zou
Abstract:
Diffusion models have achieved impressive success in high-fidelity image generation but suffer from slow sampling due to their inherently iterative denoising process. While recent one-step methods accelerate inference by learning direct noise-to-image mappings, they sacrifice the interpretability and fine-grained control intrinsic to diffusion dynamics, key advantages that enable applications like editable generation. To resolve this dichotomy, we introduceHierarchical Koopman Diffusion, a novel framework that achieves both one-step sampling and interpretable generative trajectories. Grounded in Koopman operator theory, our method lifts the nonlinear diffusion dynamics into a latent space where evolution is governed by globally linear operators, enabling closed-form trajectory solutions. This formulation not only eliminates iterative sampling but also provides full access to intermediate states, allowing manual intervention during generation. To model the multi-scale nature of images, we design a hierarchical architecture that disentangles generative dynamics across spatial resolutions via scale-specific Koopman subspaces, capturing coarse-to-fine details systematically. We empirically show that the Hierarchical Koopman Diffusion not only achieves competitive one-step generation performance but also provides a principled mechanism for interpreting and manipulating the generative process through spectral analysis. Our framework bridges the gap between fast sampling and interpretability in diffusion models, paving the way for explainable image synthesis in generative modeling.



Paperid:4079
Authors:Huming Qiu, Zhaoxiang Wang, Mi Zhang, Xiaohan Zhang, Xiaoyu You, Min Yang
Abstract:
Image watermarking embeds imperceptible signals into AI-generated images for deepfake detection and provenance verification. Although recent semantic-level watermarking methods demonstrate strong resistance against conventional pixel-level removal attacks, their robustness against more advanced removal strategies remains underexplored, raising concerns about their reliability in practical scenarios. Existing removal attacks primarily operate in the pixel domain without altering image semantics, which limits their effectiveness against semantic-level watermarks.In this paper, we propose Next Frame Prediction Attack (NFPA), the first semantic-level removal attack. Unlike pixel-level attacks, NFPA formulates watermark removal as a video generation task: it treats the watermarked image as the initial frame and aims to subtly manipulate the image semantics to generate the next-frame image, i.e., the unwatermarked image.We conduct a comprehensive evaluation on eight state-of-the-art semantic watermarking schemes, demonstrating that NFPA consistently outperforms nine removal attack baselines in both watermark removal and image quality preservation. Our results reveal the vulnerabilities of current image watermarking methods and highlight the urgent need for more robust watermarks.Code is available at https://anonymous.4open.science/r/NFPA/.



Paperid:4080
Authors:Daniele Baieri, Filippo Maggioli, Emanuele Rodolà, Simone Melzi, Zorah Lähner
Abstract:
Neural fields have emerged as a powerful representation for 3D geometry, enabling compact and continuous modeling of complex shapes. Despite their expressive power, manipulating neural fields in a controlled and accurate manner -- particularly under spatial constraints -- remains an open challenge, as existing approaches struggle to balance surface quality, robustness, and efficiency. We address this by introducing a novel method for handle-guided neural field deformation, which leverages discrete local surface representations to optimize the As-Rigid-As-Possible deformation energy. To this end, we propose the local patch mesh representation, which discretizes level sets of a neural signed distance field by projecting and deforming flat mesh patches guided solely by the SDF and its gradient. We conduct a comprehensive evaluation showing that our method consistently outperforms baselines in deformation quality, robustness, and computational efficiency. We also present experiments that motivate our choice of discretization over marching cubes. By bridging classical geometry processing and neural representations through local patch meshing, our work enables scalable, high-quality deformation of neural fields and paves the way for extending other geometric tasks to neural domains.



Paperid:4081
Authors:Yingying Feng, Jie Li, Jie Hu, Yukang Zhang, Lei Tan, Jiayi Ji
Abstract:
The challenge of inconsistent modalities in real-world applications presents significant obstacles to effective object re-identification (ReID). However, most existing approaches assume modality-matched conditions, significantly limiting their effectiveness in modality-mismatched scenarios. To overcome this limitation and achieve a more flexible ReID, we introduce MDReID to allow any-to-any image-level ReID systems. MDReID is inspired by the widely recognized perspective that modality information comprises both modality-shared features, predictable across modalities, and unpredictable modality-specific features, which are inherently modality-dependent and consist of two key components: the Modality Decoupling Module (MDM) and Modality-aware Metric Learning (MML). Specifically, MDM explicitly decomposes modality features into modality-shared and modality-specific representations, enabling effective retrieval in both modality-aligned and mismatched scenarios. MML, a tailored metric learning strategy, further enhances feature discrimination and decoupling by exploiting distributional relationships between shared and specific modality features. Extensive experiments conducted on three challenging multi-modality ReID benchmarks (RGBNT201, RGBNT100, MSVR310) consistently demonstrate the superiority of MDL. MDReID achieves significant mAP improvements of 9.8\%, 3.0\%, and 11.5\% in modality-matched scenarios, and average gains of 3.4\%, 11.8\%, and 10.9\% in modality-mismatched scenarios, respectively.



Paperid:4082
Authors:Zhichao Zhu, YANG QI, Hengyuan Ma, Wenlian Lu, Jianfeng Feng
Abstract:
The Forward-Forward (FF) algorithm provides a bottom-up alternative to backpropagation (BP) for training neural networks, relying on a layer-wise "goodness" function to guide learning. Existing goodness functions, inspired by energy-based learning (EBL), are typically defined as the sum of squared post-synaptic activations, neglecting the correlations between neurons. In this work, we propose a novel goodness function termed dimensionality compression that uses the effective dimensionality (ED) of fluctuating neural responses to incorporate second-order statistical structure. Our objective minimizes ED for clamped inputs when noise is considered while maximizing it across the sample distribution, promoting structured representations without the need to prepare negative samples.We demonstrate that this formulation achieves competitive performance compared to other non-BP methods. Moreover, we show that noise plays a constructive role that can enhance generalization and improve inference when predictions are derived from the mean of squared outputs, which is equivalent to making predictions based on the energy term. Our findings contribute to the development of more biologically plausible learning algorithms and suggest a natural fit for neuromorphic computing, where stochasticity is a computational resource rather than a nuisance.



Paperid:4083
Authors:Kotaro Ikeda, Masanori Koyama, Jinzhe Zhang, Kohei Hayashi, Kenji Fukumizu
Abstract:
In this paper, we propose a flow-based method for learning all-to-all transfer maps among conditional distributions that approximates pairwise optimal transport. The proposed method addresses the challenge of handling the case of continuous conditions, which often involve a large set of conditions with sparse empirical observations per condition. We introduce a novel cost function that enables simultaneous learning of optimal transports for all pairs of conditional distributions. Our method is supported by a theoretical guarantee that, in the limit, it converges to the pairwise optimal transports among infinite pairs of conditional distributions. The learned transport maps are subsequently used to couple data points in conditional flow matching. We demonstrate the effectiveness of this method on synthetic and benchmark datasets, as well as on chemical datasets in which continuous physical properties are defined as conditions.



Paperid:4084
Authors:Shiyi Xu, Hu Yiwen, Yingqian Min, Zhipeng Chen, Xin Zhao, Ji-Rong Wen
Abstract:
With the significant progress of large reasoning models in complex coding and reasoning tasks, existing benchmarks, like LiveCodeBench and CodeElo, are insufficient to evaluate the coding capabilities of large language models (LLMs) in real competition environments. Moreover, current evaluation metrics such as Pass@K fail to capture the reflective abilities of reasoning models. To address these challenges, we propose ICPC-Eval, a top-level competitive coding benchmark designed to probing the frontiers of LLM reasoning. ICPC-Eval includes 118 carefully curated problems from 11 recent ICPC contests held in various regions of the world, offering three key contributions: 1) A challenging realistic ICPC competition scenario, featuring a problem type and difficulty distribution consistent with actual contests. 2) A robust test case generation method and a corresponding local evaluation toolkit, enabling efficient and accurate local evaluation. 3) An effective test-time scaling evaluation metric, Refine@K, which allows iterative repair of solutions based on execution feedback. The results underscore the significant challenge in evaluating complex reasoning abilities: top-tier reasoning models like DeepSeek-R1 often rely on multi-turn code feedback to fully unlock their in-context reasoning potential when compared to non-reasoning counterparts. Furthermore, despite recent advancements in code generation, these models still lag behind top-performing human teams. We release the benchmark at: https://github.com/RUCAIBox/ICPC-Eval



Paperid:4085
Authors:Guancheng Wan, Jiaru Qian, Wenke Huang, Qilin Xu, Xianda Guo, Boheng Li, Guibin Zhang, Bo Du, Mang Ye
Abstract:
Federated Graph Learning (FGL) offers a promising framework for collaboratively training Graph Neural Networks (GNNs) while preserving data privacy. In resource-constrained environments, One-shot Federated Learning (OFL) emerges as an effective solution by limiting communication to a single round. Current OFL approaches employing generative models have attracted considerable attention; however, they face unresolved challenges: these methods are primarily designed for traditional image data and fail to capture the fine-grained structural information of local graph data. Consequently, they struggle to integrate the intricate correlations necessary and transfer subtle structural insights from each client to the global model. To address these issues, we introduceOASIS, an innovative one-shot FGL framework. In OASIS, we propose a Synergy Graph Synthesizer designed to generate informative synthetic graphs and introduce a Topological Codebook to construct a structural latent space. Moreover, we propose the Wasserstein-Enhanced Semantic Affinity Distillation (WESAD) to incorporate rich inter-class relationships and the Wasserstein-Driven Structural Relation Distillation (WDSRD) to facilitate the effective transfer of structural knowledge from the Topological Codebook. Extensive experiments on real-world tasks demonstrate the superior performance and generalization capability of OASIS. The code is available for anonymous access at https://anonymous.4open.science/r/OASIS-NeurIPS25.



Paperid:4086
Authors:Jiahui Zhang, Yurui Chen, Yueming Xu, Ze Huang, Yanpeng Zhou, Yu-Jie Yuan, Xinyue Cai, Guowei Huang, Xingyue Quan, Hang Xu, Li Zhang
Abstract:
Leveraging diverse robotic data for pretraining remains a critical challenge. Existing methods typically model the dataset’s action distribution using simple observations as inputs. However, these inputs are often incomplete, resulting in a dispersed conditional action distribution—an issue we refer to as coordinate system chaos and state chaos. This inconsistency significantly hampers pretraining efficiency. To address this, we propose 4D-VLA, a novel approach that effectively integrates 4D information into the input to mitigate these sources of chaos. Our model introduces depth and temporal information into visual features with sequential RGB-D inputs, aligning the coordinate systems of the robot and the scene. This alignment endows the model with strong spatiotemporal reasoning capabilities while minimizing training overhead. Additionally, we introduce Memory bank sampling, a frame sampling strategy designed to extract informative frames from historical images, further improving effectiveness and efficiency. Experimental results demonstrate that our pretraining method and architectural components substantially enhance model performance. In both simulated and real-world experiments, our model achieves a significant increase in success rate over OpenVLA.To further assess spatial perception and generalization to novel views, we introduce MV-Bench, a multi-view simulation benchmark. Our model consistently outperforms existing methods, demonstrating stronger spatial understanding and adaptability.



Paperid:4087
Authors:Sebastian Joseph, Syed M. Husain, Stella Offner, Stéphanie Juneau, Paul Torrey, Adam Bolton, Juan Farias, Niall Gaffney, Greg Durrett, Junyi Jessy Li
Abstract:
Large Language Models (LLMs) are being explored for applications in scientific research, including their capabilities to synthesize literature, answer research questions, generate research ideas, and even conduct computational experiments.Ultimately, our goal is for these to help scientists derive novel scientific insights. In many areas of science, such insights often arise from processing and visualizing data to understand its patterns. However, evaluating whether an LLM-mediated scientific workflow produces outputs conveying the correct scientific insights is challenging to evaluate and has not been addressed in past work.We introduce AstroVisBench, the first benchmark for both scientific computing and visualization in the astronomy domain.AstroVisBench judges a language model’s ability to both (1) create astronomy-specific workflows to process and analyze data and (2) visualize the results of these workflows through complex plots.Our evaluation of visualizations uses a novel LLM-as-a-judge workflow, which is validated against annotation by five professional astronomers.Using AstroVisBench we present an evaluation of state-of-the-art language models, showing a significant gap in their ability to engage in astronomy research as useful assistants.This evaluation provides a strong end-to-end evaluation for AI scientists that offers a path forward for the development of visualization-based workflows, which are central to a broad range of domains from physics to biology.



Paperid:4088
Authors:Jie Long Lee, Gim Hee Lee
Abstract:
Event cameras are neuromorphic vision sensors that asynchronously capture pixel-level intensity changes with high temporal resolution and dynamic range. These make them well suited for monocular depth estimation under challenging lighting conditions. However, progress in event-based monocular depth estimation remains constrained by the quality of supervision: LiDAR-based depth labels are inherently sparse, spatially incomplete, and prone to artifacts. Consequently, these signals are suboptimal for learning dense depth from sparse events. To address this problem, we propose Distil-E2D, a framework that distills depth priors from the image domain into the event domain by generating dense synthetic pseudolabels from co-recorded APS or RGB frames using foundational depth models. These pseudolabels complement sparse LiDAR depths with dense semantically rich supervision informed by large-scale image-depth datasets. To reconcile discrepancies between synthetic and real depths, we introduce a Confidence-Guided Calibrated Depth Loss that learns nonlinear depth alignment and adaptively weights supervision by alignment confidence. Additionally, our architecture integrates past predictions via a Context Transformer and employs a Dual-Decoder Training scheme that enhances encoder representations by jointly learning metric and relative depth abstractions. Experiments on benchmark datasets show that Distil-E2D achieves state-of-the-art performance in event-based MDE across both event-only and event+APS settings. Our code will be released upon paper acceptance.



Paperid:4089
Authors:Michal Balcerak, Tamaz Amiranashvili, Antonio Terpin, Suprosanna Shit, Lea Bogensperger, Sebastian Kaltenbach, Petros Koumoutsakos, bjoern menze
Abstract:
The most widely used generative models map noise and data distributions by matching flows or scores. However, they struggle to incorporate partial observations and additional priors—something energy-based models (EBMs) handle elegantly by simply adding corresponding scalar energy terms. We address this issue by proposing Energy Matching, a framework that endows flow-based approaches with the flexibility of EBMs. Far from the data manifold, samples move along curl-free, optimal transport paths from noise to data. As they approach the data manifold, an entropic energy term guides the system into a Boltzmann equilibrium distribution, explicitly capturing the underlying likelihood structure of the data. We parameterize this dynamic with a single time-independent scalar field, which serves as both a powerful generator and a flexible prior for effective regularization of inverse problems. Our method substantially outperforms existing EBMs on CIFAR-10 and ImageNet generation in terms of fidelity, while retaining simulation-free training of transport-based approaches away from the data manifold. Furthermore, we leverage the method’s flexibility to introduce an interaction energy that supports diverse mode exploration, which we demonstrate in a controlled protein‐generation setting. Our approach focuses on learning a scalar potential energy—without time-conditioning, auxiliary generators, or additional networks—which marks a significant departure from recent EBM methods. We believe that this simplified framework significantly advances EBMs capabilities and paves the way for their wider adoption in generative modeling across diverse domains.



Paperid:4090
Authors:Reda Marzouk, Shahaf Bassan, Guy Katz
Abstract:
Although Shapley additive explanations (SHAP) can be computed in polynomial time for simple models like decision trees, they unfortunately become NP-hard to compute for more expressive black-box models like neural networks - where generating explanations is often most critical. In this work, we analyze the problem of computing SHAP explanations forTensor Networks (TNs), a broader and more expressive class of models than those for which current exact SHAP algorithms are known to hold, and which is widely used for neural network abstraction and compression. First, we introduce a general framework for computing provably exact SHAP explanations for general TNs with arbitrary structures. Interestingly, we show that, when TNs are restricted to aTensor Train (TT)structure, SHAP computation can be performed inpoly-logarithmictime usingparallelcomputation. Thanks to the expressiveness power of TTs, this complexity result can be generalized to many other popular ML models such as decision trees, tree ensembles, linear models, and linear RNNs, therefore tightening previously reported complexity results for these families of models. Finally, by leveraging reductions of binarized neural networks to Tensor Network representations, we demonstrate that SHAP computation can becomeefficiently tractablewhen the network’swidthis fixed, while it remains computationally hard even with constantdepth. This highlights an important insight: for this class of models, width - rather than depth - emerges as the primary computational bottleneck in SHAP computation.



Paperid:4091
Authors:Zafir Stojanovski, Oliver Stanley, Joe Sharratt, Richard Jones, Abdulhakeem Adefioye, Jean Kaddour, Andreas Köpf
Abstract:
We introduce Reasoning Gym, a library of reasoning environments for reinforcement learning with verifiable rewards (RLVR). It provides over 100 tasks spanning multiple domains including algebra, arithmetic, computation, cognition, geometry, graph theory, logic, and various common games. Its key innovation is the ability to generate virtually infinite training data with adjustable complexity, unlike most previous reasoning datasets, which are typically fixed. This procedural generation approach allows for continuous evaluation across varying difficulty levels and task configurations. Our experimental results demonstrate the efficacy of Reasoning Gym in both evaluating and reinforcement learning of reasoning models.



Paperid:4092
Authors:Zhitao Zeng, Guojian Yuan, Junyuan Mao, Yuxuan Wang, Xiaoshuang Jia, Yueming Jin
Abstract:
Accurate temporal prediction is the bridge between comprehensive scene understanding and embodied artificial intelligence. However, predicting multiple fine-grained states of scene at multiple temporal scales is difficult for vision-language models.We formalize the Multi‐Scale Temporal Prediction (MSTP) task in general and surgical scene by decomposing multi‐scale into two orthogonal dimensions: the temporal scale, forecasting states of human and surgery at varying look‐ahead intervals, and the state scale, modeling a hierarchy of states in general and surgical scene. For instance in general scene, states of contacting relationship are finer-grained than states of spatial relationship. For instance in surgical scene, medium‐level steps are finer‐grained than high‐level phases yet remain constrained by their encompassing phase. To support this unified task, we introduce the first MSTP Benchmark, featuring synchronized annotations across multiple state scales and temporal scales. We further propose a novel method, Incremental Generation and Multi‐agent Collaboration (IG-MC), which integrates two key innovations. Firstly, we propose an plug-and-play incremental generation to keep high-quality temporal prediction that continuously synthesizes up-to-date visual previews at expanding temporal scales to inform multiple decision-making agents, ensuring decision content and generated visuals remain synchronized and preventing performance degradation as look‐ahead intervals lengthen.Secondly, we propose a decision‐driven multi‐agent collaboration framework for multiple states prediction, comprising generation, initiation, and multi‐state assessment agents that dynamically triggers and evaluates prediction cycles to balance global coherence and local fidelity. Extensive experiments on the MSTP Benchmark in general and surgical scene show that IG‐MC is a generalizable plug-and-play method for MSTP, demonstrating the effectiveness of incremental generation and the stability of decision‐driven multi‐agent collaboration.



Paperid:4093
Authors:Hongzhi Wang, Xiubo Liang, Jinxing Han, Weidong Geng
Abstract:
Low-light image enhancement (LLIE) aims to improve the visibility and quality of images captured under poor illumination. However, existing deep enhancement methods often neglect computational efficiency, resulting in high energy and memory costs. We propose \textbf{Spike-RetinexFormer}, a novel LLIE architecture that synergistically combines Retinex theory with spiking neural networks(SNNs) and Transformer-based design. By leveraging sparse spike-based computations, Spike-RetinexFormer drastically reduces energy consumption and memory usage compared to conventional approaches. Extensive evaluations across thirteen diverse LLIE benchmarks demonstrate that Spike-RetinexFormer achieves state-of-the-art performance, often improving PSNR by 0.5–1.5 dB over previous best methods (attaining 25.50 dB on LOL-v1 and 30.37 dB on SDSD-out) while using significantly less power. Our work pioneers the synergistic integration of SNNs within Transformer architectures for LLIE, establishing a compelling pathway towards powerful, energy-efficient low-level vision on resource-constrained platforms.



Paperid:4094
Authors:Moonsoo Jeong, Dongbeen Kim, Minseong Kim, Sungkil Lee
Abstract:
We present a Directional Consistency (DC)-driven Adaptive Density Control (ADC) for 3D Gaussian Splatting (DC4GS). Whereas the conventional ADC bases its primitive splitting on the magnitudes of positional gradients, we further incorporate the DC of the gradients into ADC, and realize it through the angular coherence of the gradients. Our DC better captures local structural complexities in ADC, avoiding redundant splitting. When splitting is required, we again utilize the DC to define optimal split positions so that sub-primitives best align with the local structures than the conventional random placement. As a consequence, our DC4GS greatly reduces the number of primitives (up to 30\% in our experiments) than the existing ADC, and also enhances reconstruction fidelity greatly.



Paperid:4095
Authors:Shanshan Li, Da Huang, Yu He, Yanwei Fu, Yu-Gang Jiang, Xiangyang Xue
Abstract:
In daily life, people often move through spaces to find objects that meet their needs, a key challenge in embodied AI. Traditional Demand-Driven Navigation (DDN) handles one need at a time but doesnot reflect the complexity of real-world tasks involving multiple needs and personal choices. To bridge this gap, we introduce Task-Preferenced Multi-Demand-Driven Navigation (TP-MDDN), a new benchmark for long-horizon navigation involving multiple sub-demands with explicit task preferences. To solve TP-MDDN, we propose AWMSystem, an autonomous decision-making system composed of three key modules: BreakLLM (instruction decomposition), LocateLLM (goal selection), and StatusLLM (task monitoring). For spatial memory, we design MASMap, which combines 3D point cloud accumulation with 2D semantic mapping for accurate and efficient environmental understanding. Our Dual-Tempo action generation framework integrates zero-shot planning with policy-based fine control, and is further supported by an Adaptive Error Corrector that handles failure cases in real time. Experiments demonstrate that our approach outperforms state-of-the-art baselines in both perception accuracy and navigation robustness. Code and supplementary materials will be released to support future research.



Paperid:4096
Authors:Shoaib Siddiqui, Adrian Weller, David Krueger, Gintare Karolina Dziugaite, Michael Mozer, Eleni Triantafillou
Abstract:
Abstract:Recent unlearning methods for LLMs are vulnerable to relearning attacks: knowledge believed-to-be-unlearned re-emerges by fine-tuning on a small set of (even seemingly-unrelated) examples. We study this phenomenon in a controlled setting for example-level unlearning in vision classifiers. We make the surprising discovery that forget-set accuracy can recover from around 50\% post-unlearning to nearly 100\% with fine-tuning on just the *retain* set---i.e., zero examples of the forget set. We observe this effect across a wide variety of unlearning methods, whereas for a model retrained from scratch excluding the forget set (gold standard), the accuracy remains at 50\%. We observe that resistance to relearning attacks can be predicted by weight-space properties, specifically, $L_2$-distance and linear mode connectivity between the original and the unlearned model. Leveraging this insight, we propose a new class of methods that achieve state-of-the-art resistance to relearning attacks.



Paperid:4097
Authors:Xinya Qin, Lu Bai, Lixin Cui, Ming Li, Hangyuan Du, Edwin Hancock
Abstract:
The problem of over-smoothing has emerged as a fundamental issue for Graph Convolutional Networks (GCNs). While existing efforts primarily focus on enhancing the discriminability of node representations for node classification, they tend to overlook the over-smoothing at the graph level, significantly influencing the performance of graph classification. In this paper, we provide an explanation of the graph-level over-smoothing phenomenon, and propose a novel Adaptive Multi-Viewed Subgraph Convolutional Network (MultiNet) to address this challenge. Specifically, the MultiNet introduces a local subgraph convolution module that adaptively divides each input graph into multiple subgraph views. Then a number of subgraph-based view-specific convolution operations are applied to constrain the extent of node information propagation over the original global graph structure, not only mitigating the over-smoothing issue but also generating more discriminative local node representations. Moreover, we develop an alignment-based readout that establishes correspondences between nodes over different graphs, thereby effectively preserving the local node-level structure information and improving the discriminative ability of the resulting graph-level representations. Theoretical analysis and empirical studies show that the MultiNet mitigates the graph-level over-smoothing and achieves excellent performance for graph classification.



Paperid:4098
Authors:Mehrdad Noori, David OSOWIECHI, Gustavo Vargas Hakim, Ali Bahri, Moslem Yazdanpanah, Sahar Dastani, Farzad Beizaee, Ismail Ayed, Christian Desrosiers
Abstract:
Recently, test-time adaptation has attracted wide interest in the context of vision-language models for image classification. However, to the best of our knowledge, the problem is completely overlooked in dense prediction tasks such as Open-Vocabulary Semantic Segmentation (OVSS). In response, we propose a novel TTA method tailored to adapting VLMs for segmentation during test time. Unlike TTA methods for image classification, our Multi-Level and Multi-Prompt (MLMP) entropy minimization integrates features from intermediate vision-encoder layers and is performed with different text-prompt templates at both the global CLS token and local pixel-wise levels. Our approach could be used as plug-and-play for any segmentation network, does not require additional training data or labels, and remains effective even with a single test sample. Furthermore, we introduce a comprehensive OVSS TTA benchmark suite, which integrates a rigorous evaluation protocol, seven segmentation datasets, and 15 common corruptions, with a total of 82 distinct test scenarios, establishing a standardized and comprehensive testbed for future TTA research in open-vocabulary segmentation. Our experiments on this suite demonstrate that our segmentation-tailored method consistently delivers significant gains over direct adoption of TTA classification baselines. Our code and data can be found at https://anonymous.4open.science/r/MTL_TTA-AE6C.



Paperid:4099
Authors:Kang Zhang, Trung X. Pham, Suyeon Lee, Axi Niu, Arda Senocak, Joon Son Chung
Abstract:
We present MGAudio, a novel flow-based framework for open-domain video-to-audio generation, which introduces model-guided dual-role alignment as a central design principle. Unlike prior approaches that rely on classifier-based or classifier-free guidance, MGAudio enables the generative model to guide itself through a dedicated training objective designed for video-conditioned audio generation. The framework integrates three main components: (1) a scalable flow-based Transformer denoiser, (2) a dual-role alignment mechanism where the audio-visual encoder serves both as a conditioning module and as a feature aligner to improve generation quality, and (3) a model-guided objective that enhances cross-modal coherence and audio realism.MGAudio achieves state-of-the-art performance on VGGSound, reducing FAD to 0.40, substantially surpassing the best classifier-free guidance baselines, and consistently outperforms existing methods across FD, IS, and alignment metrics. It also generalizes well to the challenging UnAV-100 benchmark. These results highlight model-guided dual-role alignment as a powerful and scalable paradigm for conditional video-to-audio generation.



Paperid:4100
Authors:Andrew Chin, Jake Callahan, Jason Pacheco, Tommie Catanach
Abstract:
Expected information gain (EIG) is a crucial quantity in Bayesian optimal experimental design (BOED), quantifying how useful an experiment is by the amount we expect the posterior to differ from the prior.However, evaluating the EIG can be computationally expensive since it generally requires estimating the posterior normalizing constant.In this work, we leverage two idiosyncrasies of BOED to improve efficiency of EIG estimation via sequential Monte Carlo (SMC).First, in BOED we simulate the data and thus know the true underlying parameters.Second, we ultimately care about the EIG, not the individual normalizing constants.Often we observe that the Monte Carlo variance of standard SMC estimators for the normalizing constant of a single dataset are significantly lower than the variance of the normalizing constants across datasets; the latter thus contributes the majority of the variance for EIG estimates.This suggests the potential to slightly increase variance while drastically decreasing computation time by reducing the SMC population size, which leads us to an EIG-specific SMC estimator that starts with a only a single sample from the posterior and tempers \textit{backwards} towards the prior. Using this single-sample estimator, which we call reverse-annealed SMC (RA-SMC), we show that it is possible to estimate EIG with orders of magnitude fewer likelihood evaluations in three models: a four-dimensional spring-mass, a six-dimensional Johnson-Cook model and a four-dimensional source-finding problem.



Paperid:4101
Authors:Toby Boyne, Juan Campos, Rebecca Langdon, Jixiang Qing, Yilin Xie, Shiqiang Zhang, Calvin Tsay, Ruth Misener, Daniel Davies, Kim Jelfs, Sarah Boyall, Thomas Dixon, Linden Schrecker, Jose Pablo Folch
Abstract:
Machine learning has promised to change the landscape of laboratory chemistry, with impressive results in molecular property prediction and reaction retro-synthesis. However, chemical datasets are often inaccessible to the machine learning community as they tend to require cleaning, thorough understanding of the chemistry, or are simply not available. In this paper, we introduce a novel dataset for yield prediction, providing the first-ever transient flow dataset for machine learning benchmarking, covering over 1200 process conditions. While previous datasets focus on discrete parameters, our experimental set-up allow us to sample a large number of continuous process conditions, generating new challenges for machine learning models. We focus on solvent selection, a task that is particularly difficult to model theoretically and therefore ripe for machine learning applications. We showcase benchmarking for regression algorithms, transfer-learning approaches, feature engineering, and active learning, with important applications towards solvent replacement and sustainable manufacturing.



Paperid:4102
Authors:Xiaoyu Yue, ZiDong Wang, Yuqing Wang, zhangwenlong, Xihui Liu, Wanli Ouyang, LEI BAI, Luping Zhou
Abstract:
Recent studies have demonstrated the importance of high-quality visual representations in image generation and have highlighted the limitations of generative models in image understanding. As a generative paradigm originally designed for natural language, autoregressive models face similar challenges. In this work, we present the first systematic investigation into the mechanisms of applying the next-token prediction paradigm to the visual domain. We identify three key properties that hinder the learning of high-level visual semantics: local and conditional dependence, inter-step semantic inconsistency, and spatial invariance deficiency. We show that these issues can be effectively addressed by introducing self-supervised objectives during training, leading to a novel training framework, Self-guided Training for AutoRegressive models (ST-AR). Without relying on pre-trained representation models, ST-AR significantly enhances the image understanding ability of autoregressive models and leads to improved generation quality. Specifically, ST-AR brings approximately 42% FID improvement for LlamaGen-L and 49% FID improvement for LlamaGen-XL, while maintaining the same sampling strategy.



Paperid:4103
Authors:Yimin Jiang, Huibing Wang, Jinjia peng
Abstract:
Person Search aims to locate query persons in gallery scene images, but faces severe performance degradation under domain shifts. Unsupervised domain adaptation transfers knowledge from the labeled source domain to the unlabeled target domain and iteratively rectifies the pseudo-labels. However, the pseudo-labels are inevitably contaminated by the source-biased model, which misleads the training process. This, in turn, reduces the quality of the pseudo-labels themselves and ultimately affects the search performance. In this paper, we propose a Spatiotemporal Consensus with Scene Prior (STCSP) framework that effectively eliminates the interference of noise on pseudo-labels, establishes positive feedback, and thus gradually bridging the domain gap. Firstly, STCSP uses a Spatiotemporal Consensus pipeline to suppress the noise from being mixed into the pseudo-labels. Secondly, leveraging the scene prior, STCSP employs our designed Iterative Bilateral Extremum Matching method to prevent the occurrence of some incorrect pseudo-labels. Thirdly, we propose a Scene Prior Contrastive Learning module, which encourages the model to directly acquire the scene prior knowledge from the target domain, thereby mitigating the generation of noise. By suppressing noise contamination, avoiding noise occurrence and mitigating noise generation, our framework achieves state-of-the-art performance on two benchmark datasets, PRW with 50.2% mAP and CUHK-SYSU with 87.0% mAP.



Paperid:4104
Authors:Haidong Xu, Guangwei Xu, Zhedong Zheng, Xiatian Zhu, Wei Ji, Xiangtai Li, Ruijie Guo, Meishan Zhang, Min Zhang, Hao Fei
Abstract:
This paper introducesVimoRAG, a novel video-based retrieval-augmented motion generation framework for motion large language models (LLMs). As motion LLMs face severe out-of-domain/out-of-vocabulary issues due to limited annotated data,VimoRAGleverages large-scale in-the-wild video databases to enhance 3D motion generation by retrieving relevant 2D human motion signals. While video-based motion RAG is nontrivial, we address two key bottlenecks: (1) developing an effective motion-centered video retrieval model that distinguishes human poses and actions, and (2) mitigating the issue of error propagation caused by suboptimal retrieval results.We design the Gemini Motion Video Retriever mechanism and the Motion-centric Dual-alignment DPO Trainer, enabling effective retrieval and generation processes. Experimental results show thatVimoRAGsignificantly boosts the performance of motion LLMs constrained to text-only input.



Paperid:4105
Authors:Yinlin Zhu, Xunkai Li, Jishuo Jia, Miao Hu, Di Wu, Meikang Qiu
Abstract:
Recent advances in graph machine learning have shifted to data-centric paradigms, driven by two emerging research fields: (1) Federated graph learning (FGL) facilitates multi-client collaboration but struggles with data and task heterogeneity, resulting in limited practicality; (2) Graph foundation model (GFM) enables desirable domain generalization but is typically confined to single-machine training, neglecting the potential of cross-silo data and computational resources. It is evident that these two paradigms are complementary, and their integration offers substantial advantages. Motivated by this, we present a pioneering study about the federated graph foundation model (FedGFM), a novel decentralized GFM training paradigm. Despite the promising vision of FedGFM, knowledge entanglement has emerged as a critical challenge, where multi-domain knowledge is encoded into indistinguishable representations, thereby limiting downstream adaptation. To this end, we propose FedGFM+, an effective FedGFM framework with two key modules to mitigate knowledge entanglement in a dual-pronged manner. (1) AncDAI: From a global perspective, we introduce a novel anchor-based domain-aware initialization strategy. Before pre-training, each client encodes its local graph into a domain-specific prototypes, which serve as semantic anchors in the representation space. Around each anchor, we construct synthetic embeddings to initialize the global model. We theoretically show that these prototypes are distinguishable across domains, and the initialization provides a strong inductive bias that facilitates disentanglement of domain-specific knowledge. (2) AdaDPP: From a local perspective, during pre-training, each client independently learns a lightweight graph prompt that imbues the GFM with its own domain semantic preferences. During fine-tuning, prompts from all clients are aggregated into an adaptive domain-sensitive prompt pool, from which the GFM selects relevant prompts to augment the target graph’s attributes, thereby improving the downstream adaptation. FedGFM+ is extensively evaluated on 8 diverse benchmarks spanning multiple domains and tasks, outperforming 20 baselines from isolated supervised learning, FGL, and federated variants of centralized GFM paradigms.



Paperid:4106
Authors:Tom Kouwenhoven, Kiana Shahrasbi, Tessa Verhoef
Abstract:
Recent advances in multimodal models have raised questions about whether vision-and-language models (VLMs) integrate cross-modal information in ways that reflect human cognition. One well-studied test case in this domain is the bouba-kiki effect, where humans reliably associate pseudowords like "bouba" with round shapes and "kiki" with jagged ones. Given the mixed evidence found in prior studies for this effect in VLMs, we present a comprehensive re-evaluation focused on two variants of CLIP, ResNet and Vision Transformer (ViT), given their centrality in many state-of-the-art VLMs. We apply two complementary methods closely modelled after human experiments: a prompt-based evaluation that uses probabilities as model preference, and we use Grad-CAM as a novel way to interpret visual attention in shape-word matching tasks. Our findings show that these models do not consistently exhibit the bouba-kiki effect. While ResNet shows a preference for round shapes, overall performance across both models lacks the expected associations. Moreover, direct comparison with prior human data on the same task shows that the models' responses fall markedly short of the robust, modality-integrated behaviour characteristic of human cognition. These results contribute to the ongoing debate about the extent to which VLMs truly understand cross-modal concepts, highlighting limitations in their internal representations and alignment with human intuitions.



Paperid:4107
Authors:Zhaowei Wang, Wenhao Yu, Xiyu REN, Jipeng Zhang, Yu Zhao, Rohit Saxena, Liang Cheng, Ginny Wong, Simon See, Pasquale Minervini, Yangqiu Song, Mark Steedman
Abstract:
The rapid extension of context windows in large vision-language models has given rise to long-context vision-language models (LCVLMs), which are capable of handling hundreds of images with interleaved text tokens in a single forward pass. In this work, we introduce MMLongBench, the first benchmark covering a diverse set of long-context vision-language tasks, to evaluate LCVLMs effectively and thoroughly. MMLongBench is composed of 13,331 examples spanning five different categories of downstream tasks, such as Visual RAG and Many-Shot ICL. It also provides broad coverage of image types, including various natural and synthetic images. To assess the robustness of the models to different input lengths, all examples are delivered at five standardized input lengths (8K-128K tokens) via a cross-modal tokenization scheme that combines vision patches and text tokens. Through a thorough benchmarking of 46 closed-source and open-source LCVLMs, we provide a comprehensive analysis of the current models' vision-language long-context ability. Our results show that: i) performance on a single task is a weak proxy for overall long-context capability; ii) both closed-source and open-source models face challenges in long-context vision-language tasks, indicating substantial room for future improvement; iii) models with stronger reasoning ability tend to exhibit better long-context performance. By offering wide task coverage, various image types, and rigorous length control, MMLongBench provides the missing foundation for diagnosing and advancing the next generation of LCVLMs.



Paperid:4108
Authors:Ruiqi Wu, Xinjie wang, Liu.Liu, Chun-Le Guo, Jiaxiong Qiu, Chongyi Li, Lichao Huang, Zhizhong Su, Ming-Ming Cheng
Abstract:
We presentDIPO, a novel framework for the controllable generation of articulated 3D objects from a pair of images: one depicting the object in a resting state and the other in an articulated state.Compared to the single-image approach, our dual-image input imposes only a modest overhead for data collection, but at the same time provides important motion information, which is a reliable guide for predicting kinematic relationships between parts.Specifically, we propose a dual-image diffusion model that captures relationships between the image pair to generate part layouts and joint parameters. In addition, we introduce a Chain-of-Thought (CoT) basedgraph reasonerthat explicitly infers part connectivity relationships.To further improve robustness and generalization on complex articulated objects, we develop a fully automated dataset expansion pipeline, nameLEGO-Art, that enriches the diversity and complexity of PartNet-Mobility dataset. We proposePM-X, a large-scale dataset of complex articulated 3D objects, accompanied by rendered images, URDF annotations, and textual descriptions.Extensive experiments demonstrate that DIPO significantly outperforms existing baselines in both the resting state and the articulated state, while the proposed PM-X dataset further enhances generalization to diverse and structurally complex articulated objects.Our code and dataset will be released to the community upon publication.



Paperid:4109
Authors:Ferdinand Genans, Antoine Godichon-Baggioni, Francois-Xavier Vialard, Olivier Wintenberger
Abstract:
Abstract:Adding entropic regularization to Optimal Transport (OT) problems has become a standard approach for designing efficient and scalable solvers. However, regularization introduces a bias from the true solution. To mitigate this bias while still benefiting from the acceleration provided by regularization, a natural solver would adaptively decrease the regularization as it approaches the solution. Although some algorithms heuristically implement this idea, their theoretical guarantees and the extent of their acceleration compared to using a fixed regularization remain largely open. In the setting of semi-discrete OT, where the source measure is continuous and the target is discrete, we prove that decreasing the regularization can indeed accelerate convergence. To this end, we introduce DRAG: Decreasing (entropic) Regularization Averaged Gradient, a stochastic gradient descent algorithm where the regularization decreases with the number of optimization steps. We provide a theoretical analysis showing that DRAG benefits from decreasing regularization compared to a fixed scheme, achieving an unbiased $\mathcal{O}(1/t)$ sample and iteration complexity for both the OT cost and the potential estimation, and a $\mathcal{O}(1/\sqrt{t})$ rate for the OT map. Our theoretical findings are supported by numerical experiments that validate the effectiveness of DRAG and highlight its practical advantages.



Paperid:4110
Authors:Yuichi Inoue, Kou Misaki, Yuki Imajuku, So Kuroki, Taishi Nakamura, Takuya Akiba
Abstract:
Recent advances demonstrate that increasing inference-time computation can significantly boost the reasoning capabilities of large language models (LLMs). Although repeated sampling (i.e., generating multiple candidate outputs) is a highly effective strategy, it does not leverage external feedback signals for refinement, which are often available in tasks like coding. In this work, we propose Adaptive Branching Monte Carlo Tree Search (AB-MCTS), a novel inference-time framework that generalizes repeated sampling with principled multi-turn exploration and exploitation. At each node in the search tree, AB-MCTS dynamically decides whether to ''go wider'' by expanding new candidate responses or ''go deeper'' by revisiting existing ones based on external feedback signals. We evaluate our method on complex coding and engineering tasks using frontier models. Empirical results show that AB-MCTS outperforms both repeated sampling and standard MCTS, underscoring the importance of combining the response diversity of LLMs with multi-turn solution refinement for effective inference-time scaling.



Paperid:4111
Authors:Wujian Peng, Lingchen Meng, Yitong Chen, Yiweng Xie, Yang Liu, Tao Gui, Hang Xu, Xipeng Qiu, Zuxuan Wu, Yu-Gang Jiang
Abstract:
Large Multimodal Models (LMMs) have made significant breakthroughs with the advancement of instruction tuning. However, while existing models can understand images and videos at a holistic level, they still struggle with instance-level understanding that requires a more fine-grained comprehension and alignment. Instance-level understanding is crucial for LMMs, as it focuses on the specific elements that we are most interested in. Excitingly, existing works find that the state-of-the-art LMMs exhibit strong instance understanding capabilities when provided with explicit visual cues. Motivated by this, we proposed Inst-IT, a solution to enhance LMMs in Instance understanding via explicit visual prompt Instruction Tuning for instance guidance. Inst-IT consists of a benchmark to diagnose multimodal instance-level understanding, a large-scale instruction-tuning dataset, and a continuous instruction-tuning training paradigm to effectively enhance spatial-temporal instance understanding capabilities of existing LMMs. Experimental results show that, enhanced by Inst-IT, our models not only achieve outstanding performance on Inst-IT-Bench and other instance understanding benchmarks, but also demonstrate significant improvements across various generic image and video understanding benchmarks. This highlights that our method not only boosts instance-level understanding but also strengthens the overall capabilities of generic image and video comprehension.



Paperid:4112
Authors:Haoyu Wang, Wei Dai, Jiawei Zhang, Jialun Ma, Mingyi Huang, Yuguo Yu
Abstract:
Loss of plasticity (LoP) is the primary cause of cognitive decline in normal aging brains next to cell loss.Recent works show that similar LoP also plagues neural networks during deep continual learning (DCL). While it has been shown that random perturbations of learned weights can alleviate LoP, its underlying mechanisms remain insufficiently understood.Here we offer a unique view of LoP and dissect its mechanisms through the lenses of an innovative framework combining the theory of neural collapse and finite-time Lyapunov exponents (FTLE) analysis.We show that LoP actually consists of two contrasting types: (i) type-1 LoP is characterized by highly negative FTLEs, where the network is prevented from learning due to the collapse of representations; (ii) while type-2 LoP is characterized by excessively positive FTLEs, where the network can train well but the growingly chaotic behaviors reduce its test accuracy.Based on these understandings, we introduce Generalized Mixup, designed to relax the representation space for prolonged DCL and demonstrate its superior efficacy vs. existing methods.



Paperid:4113
Authors:Huan Li, Yiming Dong, Zhouchen Lin
Abstract:
Abstract:As the default optimizer for training large language models, AdamW has achieved remarkable success in deep learning. However, its convergence behavior is not theoretically well-understood. This paper establishes the convergence rate $\frac{1}{K}\sum_{k=1}^K E[||\nabla f(x^k)||_1]\leq O(\frac{\sqrt{d}C}{K^{1/4}})$ for AdamW measured by $\ell_1$ norm, where K represents the iteration number, d denotes the model dimension, and C matches the constant in the optimal convergence rate of SGD. Theoretically, we have $E[||\nabla f(x)||_1]\geq\sqrt{\frac{2d}{\pi}}E[||\nabla f(x)||_2]$ when each element of $\nabla f(x)$ is generated from Gaussian distribution $\mathcal N(0,1)$. Empirically, our experimental results on real-world deep learning tasks reveal $||\nabla f(x)||_1=\varTheta(\sqrt{d})||\nabla f(x)||_2$. Both support that our convergence rate can be considered to be analogous to the optimal convergence rate of SGD.



Paperid:4114
Authors:Ziang Yan, Yinan He, Xinhao Li, Zhengrong Yue, Xiangyu Zeng, Yali Wang, Yu Qiao, Limin Wang, Yi Wang
Abstract:
Inducing reasoning in multimodal large language models (MLLMs) is critical for achieving human-level perception and understanding. Existing methods mainly leverage LLM reasoning to analyze parsed visuals, often limited by static perception stages. This paper introduces Visual Test-Time Scaling (VTTS), a novel approach to enhance MLLMs' reasoning via iterative perception during inference. VTTS mimics humans' hierarchical attention by progressively refining focus on high-confidence spatio-temporal regions, guided by updated textual predictions. Specifically, VTTS employs an Iterative Perception (ITP) mechanism, incorporating reinforcement learning with spatio-temporal supervision to optimize reasoning. To support this paradigm, we also present VTTS-80K, a dataset tailored for iterative perception.These designs allows a MLLM to enhance its performance by increasing its perceptual compute. Extensive experiments validate VTTS's effectiveness and generalization across diverse tasks and benchmarks. It exhibits substantial improvements (by more than 5\% on average) over strong baselines (Qwen2.5VL-3B and -7B) on more than 20 benchmarks covering video conversation, image reasoning, and spatio-temporal perception.



Paperid:4115
Authors:Haijing Liu, Zhiyuan Song, Hefeng Wu, Tao Pu, Keze Wang, Liang Lin
Abstract:
Egocentric Referring Video Object Segmentation (Ego-RVOS) aims to segment the specific object actively involved in a human action, as described by a language query, within first-person videos. This task is critical for understanding egocentric human behavior. However, achieving such segmentation robustly is challenging due to ambiguities inherent in egocentric videos and biases present in training data. Consequently, existing methods often struggle, learning spurious correlations from skewed object-action pairings in datasets and fundamental visual confounding factors of the egocentric perspective, such as rapid motion and frequent occlusions. To address these limitations, we introduce a novel framework, Causal Ego-REferring Segmentation (CERES), that is grounded in causal inference to enhance models. CERES implements dual modal causal intervention: applying backdoor adjustment principles to counteract language representation biases learned from dataset statistics, and leveraging front-door adjustment concepts to address visual confounding by intelligently integrating semantic visual features with geometric depth information guided by causal principles, creating representations more robust to egocentric distortions. Extensive experiments demonstrate that CERES achieves state-of-the-art performance on Ego-RVOS benchmarks, highlighting the potential of applying causal reasoning to build more reliable models for broader egocentric video understanding.



Paperid:4116
Authors:Chao Wang, Yixin Song, Jinhui Ye, Chuan Qin, Dazhong Shen, Lingfeng Liu, Xiang Wang, Yanyong Zhang
Abstract:
Recently, large language models (LLMs) have been explored for integration with collaborative filtering (CF)-based recommendation systems, which are crucial for personalizing user experiences. However, a key challenge is that LLMs struggle to interpret the latent, non-semantic embeddings produced by CF approaches, limiting recommendation effectiveness. To address this, we propose FACE, a general interpretable framework that aligns CF embeddings with pre-trained LLM tokens. Specifically, we introduce a disentangled projection module to decompose CF embeddings into concept-specific vectors, followed by a quantized autoencoder to convert continuous embeddings into LLM tokens (descriptors). Then we design a contrastive alignment objective to ensure that the tokens align with corresponding textual signals. Hence, the model-agnostic FACE framework achieves semantic alignment without fine-tuning LLMs and enhances recommendation performance by leveraging their pre-trained capabilities. Empirical results on three real-world recommendation datasets demonstrate performance improvements in benchmark models, with interpretability studies confirming the interpretability of the descriptors. Code is available in an anonymous GitHub repository \url{https://anonymous.4open.science/r/VQ-RAF-CA54}.



Paperid:4117
Authors:Muye Huang, Lingling Zhang, Jie Ma, Han Lai, Fangzhi Xu, Yifei Li, Wenjun Wu, Yaqiang Wu, Jun Liu
Abstract:
Charts are high-density visualization carriers for complex data, serving as a crucial medium for information extraction and analysis. Automated chart understanding poses significant challenges to existing multimodal large language models (MLLMs) due to the need for precise and complex visual reasoning. Current step-by-step reasoning models primarily focus on text-based logical reasoning for chart understanding. However, they struggle to refine or correct their reasoning when errors stem from flawed visual understanding, as they lack the ability to leverage multimodal interaction for deeper comprehension. Inspired by human cognitive behavior, we propose ChartSketcher, a multimodal feedback-driven step-by-step reasoning method designed to address these limitations. ChartSketcher is a chart understanding model that employs Sketch-CoT, enabling MLLMs to annotate intermediate reasoning steps directly onto charts using a programmatic sketching library, iteratively feeding these visual annotations back into the reasoning process. This mechanism enables the model to visually ground its reasoning and refine its understanding over multiple steps. We employ a two-stage training strategy: a cold start phase to learn sketch-based reasoning patterns, followed by off-policy reinforcement learning to enhance reflection and generalization. Experiments demonstrate that ChartSketcher achieves promising performance on chart understanding benchmarks and general vision tasks, providing an interactive and interpretable approach to chart comprehension.



Paperid:4118
Authors:Geeho Kim, Jinu Lee, Bohyung Han
Abstract:
We propose a novel federated learning framework to effectively address the challenges of generalized category discovery (GCD) in heterogeneous federated learning environments.Our algorithm does not require unrealistic assumptions often made in existing GCD work, such as the need for prior knowledge of novel class counts or the assumption of uniform class distributions.Within our federated pipeline, each client estimates its unique category distribution and the number of local classes by constructing a local, graph-based data representation. Leveraging the category structures identified via clustering, we propose a novel self-distillation strategy, which regularizes the model to align its batch-wise predictions for unlabeled examples directly with these discovered local category structures.This synergy of local structure grounding and dynamic prior adaptation enables robust generalized category discovery under severe data heterogeneity. Our framework significantly outperforms existing federated generalized category discovery approaches on fine-grained and standard benchmarks, as demonstrated by extensive experimental results.



Paperid:4119
Authors:Alan Lahoud, Erik Schaffernicht, Johannes Andreas Stork
Abstract:
Learning representations for solutions of constrained optimization problems (COPs) with unknown cost functions is challenging, as models like (Variational) Autoencoders struggle to enforce constraints when decoding structured outputs. We propose an Inverse Optimization Latent Variable Model (IO-LVM) that learns a latent space of COP cost functions from observed solutions and reconstructs feasible outputs by solving a COP with a solver in the loop. Our approach leverages estimated gradients of a Fenchel-Young loss through a non-differentiable deterministic solver to shape the latent space. Unlike standard Inverse Optimization or Inverse Reinforcement Learning methods, which typically recover a single or context-specific cost function, IO-LVM captures a distribution over cost functions, enabling the identification of diverse solution behaviors arising from different agents or conditions not available during the training process. We validate our method on real-world datasets of ship and taxi routes, as well as paths in synthetic graphs, demonstrating its ability to reconstruct paths and cycles, predict their distributions, and yield interpretable latent representations.



Paperid:4120
Authors:Jaebyeong Jeon, Hyeonseo Jang, Jy-yong Sohn, Kibok Lee
Abstract:
Equivariant representation learning aims to capture variations induced by input transformations in the representation space, whereas invariant representation learning focuses on encoding semantic information by disregarding such transformations. Recent studies have shown that jointly learning both types of representations is often beneficial for downstream tasks, typically achieved by employing separate projection heads. However, we argue that there exists essential information shared between invariant and equivariant learning, such that using independent projection heads leads to representational redundancy and suboptimal utilization of model capacity. To address this, we propose a task-aware routing mechanism that enables the model to dynamically extract task-relevant information from multiple projection heads. This encourages the projection heads to specialize as either shared or task-specific experts, resulting in reduced representational redundancy, as evidenced by lower canonical correlations between invariant and equivariant representations. Experimental results demonstrate the effectiveness of our approach across various transfer learning tasks. The code will be released.



Paperid:4121
Authors:Yuchao Lin, Cong Fu, Zachary Krueger, Haiyang Yu, Maho Nakata, Jianwen Xie, Emine Kucukbenli, Xiaofeng Qian, Shuiwang Ji
Abstract:
Abstract:$\rm{SO}(3)$-equivariant networks are the dominant models for machine learning force fields (MLFF). The key operation of such networks is the Clebsch-Gordan (CG) tensor product, which is computationally expensive. To accelerate the computation, we develop tensor decomposition networks (TDNs) as a class of approximately equivariant networks whose CG tensor products are replaced by low-rank tensor decompositions, such as the CANDECOMP/PARAFAC (CP) decomposition. With the CP decomposition, we prove (i) a uniform bound on the induced error of $\rm{SO}(3)$-equivariance, and (ii) the universality of approximating any equivariant bilinear map. To further reduce the number of parameters, we propose path-weight sharing that ties all multiplicity-space weights across the $\mathcal{O}(L^3)$ CG paths into a single path without compromising equivariance, where $L$ is the maximum angular degree. The resulting layer acts as a plug-and-play replacement for tensor products in existing networks, and the computational complexity of tensor products is reduced from $\mathcal{O}(L^6)$ to $\mathcal{O}(L^4)$. We evaluate TDNs on a newly curated, DFT-calculated molecular force field relaxation dataset with 105 million snapshots. We also use existing datasets, including OC20, and OC22. Results show that TDNs achieve competitive performance with dramatic speedup in computations.



Paperid:4122
Authors:Leonardo Zini, Elia Frigieri, Sebastiano Aloscari, Lorenzo Baraldi
Abstract:
We introduce HeisenVec, a large-scale dataset designed to advance research in vector graphics generation from natural language descriptions. Unlike conventional image generation datasets that focus on raster images, HeisenVec targets the structured and symbolic domain of Scalable Vector Graphics (SVG), where images are represented as sequences of drawing commands and style attributes. The dataset comprises 2.2 million SVGs collected from different online sources, each paired with four complementary textual descriptions generated by multi-modal models. To ensure structural consistency and efficiency for autoregressive modeling, all SVGs are standardized through a pre-processing pipeline that unifies geometric primitives as paths, applies affine transformations, and compresses syntax via custom tokens set. HeisenVec exhibits broad coverage among visual styles and sequence lengths, with a substantial portion of samples exceeding 8,000 tokens, making it particularly well-suited for benchmarking long-context language models. Our benchmark enables rigorous evaluation of text-conditioned SVG generation, encourages progress on sequence modeling with symbolic outputs, and bridges the gap between vision, graphics, and language. We release the dataset, tokenization tools, and evaluation pipeline to foster further research in this emerging domain.



Paperid:4123
Authors:Yunqi Hong, Sohyun An, Andrew Bai, Neil Lin, Cho-Jui Hsieh
Abstract:
Despite Multimodal Large Language Models (MLLMs) showing promising results on general zero-shot image classification tasks, fine-grained image classification remains challenging. It demands precise attention to subtle visual details to distinguish between visually similar subcategories—details that MLLMs may easily overlook without explicit guidance. To address this, we introduce AutoSEP, an iterative self-supervised prompt learning framework designed to enhance MLLM fine-grained classification capabilities in a fully unsupervised manner. Our core idea is to leverage unlabeled data to learn a description prompt that guides MLLMs in identifying crucial discriminative features within an image, and boost classification accuracy. We developed an automatic self-enhancing prompt learning framework called AutoSEP to iteratively improve the description prompt using unlabeled data, based on instance-level classification scoring function. AutoSEP only requires black-box access to MLLMs, eliminating the need for any training or fine-tuning. We evaluate our approach on multiple fine-grained classification datasets. It consistently outperforms other unsupervised baselines, demonstrating the effectiveness of our self-supervised optimization framework. Notably, AutoSEP in average improves 13\% over standard zero-shot classification and 5\% over the best-performing baselines.



Paperid:4124
Authors:Mohammad Jalali, Haoyu Lei, Amin Gohari, Farzan Farnia
Abstract:
Abstract:Diffusion models have demonstrated remarkable success in high-fidelity image synthesis and prompt-guided generative modeling. However, ensuring adequate diversity in generated samples of prompt-guided diffusion models remains a challenge, particularly when the prompts span a broad semantic spectrum and the diversity of generated data needs to be evaluated in a prompt-aware fashion across semantically similar prompts. Recent methods have introduced guidance via diversity measures to encourage more varied generations. In this work, we extend the diversity measure-based approaches by proposing the Scalable Prompt-Aware Latent Entropy-based Diversity Guidance (SPALE) method for prompt-aware diversity guidance. SPALE utilizes conditional entropy for diversity guidance, which dynamically conditions diversity measurement on similar prompts and enables prompt-aware diversity control. While the entropy-based guidance approach enhances prompt-aware diversity, its reliance on the matrix-based entropy scores poses computational challenges in large-scale generation settings. To address this, we focus on the special case of \textit{Conditional latent RKE Score Guidance}, reducing entropy computation and gradient-based optimization complexity from the $\mathcal{O}(n^3)$ of general entropy measures to $\mathcal{O}(n)$. The reduced computational complexity allows for diversity-guided sampling over potentially thousands of generation rounds on different prompts. We numerically test the SPALE method on several text-to-image diffusion models, demonstrating that the proposed method improves the prompt-aware diversity of the generated data without incurring significant computational costs.



Paperid:4125
Authors:Kunlun Xu, Yibo Feng, Jiangmeng Li, Yongsheng Qi, Jiahuan Zhou
Abstract:
Abstract:Federated continual learning (FCL) tackles scenarios of learning from continuously emerging task data across distributed clients, where the key challenge lies in addressing both temporal forgetting over time and spatial forgetting simultaneously. Recently, prompt-based FCL methods have shown advanced performance through task-wise prompt communication.In this study, we underscore that the existing prompt-based FCL methods are prone to class-wise knowledge coherence between prompts across clients. The class-wise knowledge coherence includes two aspects: (1) intra-class distribution gap across clients, which degrades the learned semantics across prompts, (2) inter-prompt class-wise relevance, which highlights cross-class knowledge confusion. During prompt communication, insufficient class-wise coherence exacerbates knowledge conflicts among new prompts and induces interference with old prompts, intensifying both spatial and temporal forgetting. To address these issues, we propose a novel Class-aware Client Knowledge Interaction (C${}^2$Prompt) method that explicitly enhances class-wise knowledge coherence during prompt communication. Specifically, a local class distribution compensation mechanism (LCDC) is introduced to reduce intra-class distribution disparities across clients, thereby reinforcing intra-class knowledge consistency. Additionally, a class-aware prompt aggregation scheme (CPA) is designed to alleviate inter-class knowledge confusion by selectively strengthening class-relevant knowledge aggregation. Extensive experiments on multiple FCL benchmarks demonstrate that C${}^2$Prompt achieves state-of-the-art performance. Our code will be released.



Paperid:4126
Authors:Songjun Tu, Jiahao Lin, Qichao Zhang, Xiangyu Tian, Linjing Li, Xiangyuan Lan, Dongbin Zhao
Abstract:
Large reasoning models (LRMs) are proficient at generating explicit, step-by-step reasoning sequences before producing final answers. However, such detailed reasoning can introduce substantial computational overhead and latency, particularly for simple problems. To address this over-thinking problem, we explore how to equip LRMs with adaptive thinking capabilities—enabling them to dynamically decide whether or not to engage in explicit reasoning based on problem complexity.Building on R1-style distilled models, we observe that inserting a simple ellipsis ("...") into the prompt can stochastically trigger either a thinking or no-thinking mode, revealing a latent controllability in the reasoning behavior. Leveraging this property, we propose AutoThink, a multi-stage reinforcement learning (RL) framework that progressively optimizes reasoning policies via stage-wise reward shaping.AutoThink learns to invoke explicit reasoning only when necessary, while defaulting to succinct responses for simpler tasks.Experiments on five mainstream mathematical benchmarks demonstrate that AutoThink achieves favorable accuracy–efficiency trade-offs compared to recent prompting and RL-based pruning methods. It can be seamlessly integrated into any R1-style model, including both distilled and further fine-tuned variants. Notably, AutoThink improves relative accuracy by 6.4\% while reducing token usage by 52\% on DeepSeek-R1-Distill-Qwen-1.5B, establishing a scalable and adaptive reasoning paradigm for LRMs.Project Page: https://anonymous.4open.science/r/AutoThink-BF7C/.



Paperid:4127
Authors:Stephen Zhao, Aidan Li, Rob Brekelmans, Roger Grosse
Abstract:
To avoid bad language model (LM) outputs, there are many alignment approaches (e.g., RLHF, DPO). Ideally, we would like our LM to have zero probability of undesirable outputs. Standard reinforcement learning (RL) would achieve this at optimality (if unregularized). However, in practice, there may be a tradeoff between methods focusing on the expected reward (standard RL) and methods explicitly focused on reducing the probability of undesired outputs. Our goal is to improve this tradeoff, reducing the probability of bad outputs as much as possible, while maintaining performance on expected reward. To do this, we introduce RePULSe, a new training method that augments the standard RL loss with an additional loss that uses learned proposals to guide sampling low-reward outputs, and then reduces those outputs' probability. We run experiments to test whether our method provides better reduction of the probability of bad outputs and adversarial robustness, at minimal cost to expected reward, compared to standard RL alignment approaches and other alternatives.



Paperid:4128
Authors:Li Hao, He CAO, Bin Feng, Daniel Shao, Robert Tang, Zhiyuan Yan, Li Yuan, Yonghong Tian, Yu Li
Abstract:
While large language models (LLMs) with Chain-of-Thought (CoT) reasoning excel in mathematics and coding, their potential for systematic reasoning in chemistry, a domain demanding rigorous structural analysis for real-world tasks like drug design and reaction engineering, remains untapped. Current benchmarks focus on simple knowledge retrieval, neglecting step-by-step reasoning required for complex tasks such as molecular optimization and reaction prediction. To address this, we introduce ChemCoTBench, a reasoning framework that bridges molecular structure understanding with arithmetic-inspired operations, including addition, deletion, and substitution, to formalize chemical problem-solving into transparent, step-by-step workflows. By treating molecular transformations as modular "chemical operations", the framework enables slow-thinking reasoning, mirroring the logic of mathematical proofs while grounding solutions in real-world chemical constraints. We evaluate models on two high-impact tasks: Molecular Property Optimization and Chemical Reaction Prediction. These tasks mirror real-world challenges while providing structured evaluability. By providing annotated datasets, a reasoning taxonomy, and baseline evaluations, ChemCoTBench bridges the gap between abstract reasoning methods and practical chemical discovery, establishing a foundation for advancing LLMs as tools for AI-driven scientific innovation.



Paperid:4129
Authors:Xuan Tang, Han Zhang, Yuan Cao, Difan Zou
Abstract:
Abstract:Adam is a popular and widely used adaptive gradient method in deep learning, which has also received tremendous focus in theoretical research. However, most existing theoretical work primarily analyzes its full-batch version, which differs fundamentally from the stochastic variant used in practice. Unlike SGD, stochastic Adam does not converge to its full-batch counterpart even with infinitesimal learning rates. We present the first theoretical characterization of how batch size affects Adam's generalization, analyzing two-layer over-parameterized CNNs on image data. Our results reveal that while both Adam and AdamW with proper weight decay $\lambda$ converge to poor test error solutions, their mini-batch variants can achieve near-zero test error. We further prove Adam has a strictly smaller effective weight decay bound than AdamW, theoretically explaining why Adam requires more sensitive $\lambda$ tuning. Extensive experiments validate our findings, demonstrating the critical role of batch size and weight decay in Adam's generalization performance.



Paperid:4130
Authors:Haoxin Sun, Yubo Sun, Xiaotian Zhou, Zhongzhi Zhang
Abstract:
In this paper, we address the problem of fast computation and optimization of opinion-based quantities in the Friedkin–Johnsen (FJ) model. We first introduce the concept of partial rooted forests and present an efficient algorithm for computing these quantities using this method. Furthermore, we study two optimization problems in the FJ model: the Opinion Minimization Problem and the Polarization and Disagreement Minimization Problem. For both problems, we propose fast algorithms based on partial rooted forest sampling. Our methods reduce the time complexity from linear to sublinear. Extensive experiments on real-world networks demonstrate that our algorithms are both accurate and efficient, outperforming state-of-the-art methods and scaling effectively to large-scale networks.



Paperid:4131
Authors:Arie Soeteman, Balder ten Cate
Abstract:
We propose and study Hierarchical Ego Graph Neural Networks (HE-GNNs), an expressive extension of GNNs with hierarchical node individualization, inspired by the Individualization-Refinement paradigm for graph isomorphism testing. HE-GNNs generalize subgraph-GNNs and form a hierarchy of increasingly expressive models that, in the limit, can distinguish graphs up to isomorphism.We provide a logical characterization of HE-GNN node classifiers, with and without subgraph restrictions, using graded hybrid logic. This characterization enables us to relate the separating power of HE-GNNs to that of higher-order GNNs, GNNs enriched with local homomorphism count features, and color refinement algorithms based on Individualization-Refinement.Our experimental results confirm the practical feasibility of HE-GNNs and show benefits in comparison with traditional GNN architectures, both with and without local homomorphism count features.



Paperid:4132
Authors:Shuqing Luo, Ye Han, Pingzhi Li, Jiayin Qin, Jie Peng, Yang Zhao, Yu Cao, Tianlong Chen
Abstract:
Abstract:Mixture-of-Experts (MoE) architecture offers enhanced efficiency for Large Language Models (LLMs) with modularized computation, yet its inherent sparsity poses significant hardware deployment challenges, including memory locality issues, communication overhead, and inefficient computing resource utilization. Inspired by the modular organization of the human brain, we propose $\texttt{Mozart}$, a novel algorithm-hardware co-design framework tailored for efficient training of MoE-based LLMs on 3.5D wafer-scale chiplet architectures. On the algorithm side, $\texttt{Mozart}$ exploits the inherent modularity of chiplets and introduces: ($1$) an expert allocation strategy that enables efficient on-package all-to-all communication, and ($2$) a fine-grained scheduling mechanism that improves communication-computation overlap through streaming tokens and experts. On the architecture side, $\texttt{Mozart}$ adaptively co-locates heterogeneous modules on specialized chiplets with a 2.5D NoP-Tree topology and hierarchical memory structure.Evaluation across three popular MoE models demonstrates significant efficiency gains, enabling more effective parallelization and resource utilization for large-scale modularized MoE-LLMs.



Paperid:4133
Authors:Yuan Yao, Shifan Jiang, Yangqing Hou, Chengxu Zuo, Xinrui Chen, Shihui Guo, Yipeng Qin
Abstract:
Sparse inertial measurement units (IMUs) provide a portable, low-cost solution for human motion tracking but struggle with error accumulation from drift and sensor noise when estimating joint position through time-based linear acceleration integration (i.e., indirect measurement). To address this, we propose ToF-IP, a novel 3D full-body pose estimation system that integrates Time-of-Flight (ToF) sensors with sparse IMUs. The distinct advantage of our approach is that ToF sensors provide direct distance measurements, effectively mitigating error accumulation without relying on indirect time-based integration. From a hardware perspective, we maintain the portability of existing solutions by attaching ToF sensors to selected IMUs with a negligible volume increase of just 3\%.On the software side, we introduce two novel techniques to enhance multi-sensor integration: (i) a Node-Centric Data Integration strategy that leverages a Transformer encoder to explicitly model both intra-node and inter-node data integration by treating each sensing node as a token; and (ii) a Dynamic Spatial Positional Encoding scheme that encodes the continuously changing spatial positions of wearable nodes as motion-conditioned functions, enabling the model to better capture human body dynamics in the embedding space.Additionally, we contribute a 208-minute human motion dataset from 10 participants, including synchronized IMU-ToF measurements and ground-truth from optical tracking.Extensive experiments demonstrate that our method outperforms state-of-the-art approaches such as PNP, achieving superior accuracy in tracking complex and slow motions like Tai Chi, which remains challenging for inertial-only methods.We will release our dataset and code upon acceptance.



Paperid:4134
Authors:Sid Bharthulwar, Stone Tao, Hao Su
Abstract:
Massively parallel GPU simulation environments have accelerated reinforcement learning (RL) research by enabling fast data collection for on-policy RL algorithms like Proximal Policy Optimization (PPO). To maximize throughput, it is common to use short rollouts per policy update, increasing the update-to-data (UTD) ratio. However, we find that, in this setting, standard synchronous resets introduce harmful nonstationarity, skewing the learning signal and destabilizing training. We introduce staggered resets, a simple yet effective technique where environments are initialized and reset at varied points within the task horizon. This yields training batches with more uniform state visitation distributions, reducing the nonstationarity induced by synchronized rollouts. We characterize dimensions along which RL environments can benefit significantly from staggered resets through toy environments. We then apply this technique to challenging high-dimensional robotics environments, achieving significantly higher sample efficiency, faster wall-clock convergence, and stronger final performance. Finally, this technique scales better with more parallel environments compared to naive synchronized rollouts, yielding more optimal utilization of computational resources.



Paperid:4135
Authors:Liangliang Zhang, Zhuorui Jiang, Hongliang Chi, Haoyang Chen, Mohammed ElKoumy, Fali Wang, Qiong Wu, Zhengyi Zhou, Shirui Pan, Suhang Wang, Yao Ma
Abstract:
Knowledge Graph Question Answering (KGQA) systems rely on high-quality benchmarks to evaluate complex multi-hop reasoning. However, despite their widespread use, popular datasets such as WebQSP and CWQ suffer from critical quality issues, including inaccurate or incomplete ground-truth annotations, poorly constructed questions that are ambiguous, trivial, or unanswerable, and outdated or inconsistent knowledge. Through a manual audit of 16 popular KGQA datasets—including WebQSP and CWQ—we find that the average factual correctness rate is only 57%. To address these issues, we introduce KGQAGen, an LLM-in-the-loop framework that systematically resolves these pitfalls. KGQAGen combines structured knowledge grounding, LLM-guided generation, and symbolic verification to produce challenging and verifiable QA instances. Using KGQAGen, we construct KGQAGen-10k, a 10K-scale benchmark grounded in Wikidata, and evaluate a diverse set of KG-RAG models. Experimental results demonstrate that even state-of-the-art systems struggle on this benchmark, highlighting its ability to expose limitations of existing models. Our findings advocate for more rigorous benchmark construction and position KGQAGen as a scalable framework for advancing KGQA evaluation.



Paperid:4136
Authors:Jiayu Wang, Yifei Ming, Zixuan Ke, Caiming Xiong, Shafiq Joty, Aws Albarghouthi, Frederic Sala
Abstract:
Reinforcement learning (RL) has become the dominant paradigm for endowing language models with advanced reasoning capabilities. Despite the substantial empirical gains demonstrated by RL-based training methods like GRPO, a granular understanding of their advantages is still lacking. To address this gap, we introduce a fine-grained analytic framework to dissect the impact of RL on reasoning. Our framework specifically investigates key elements that have been hypothesized to benefit from RL training: (1) plan-following and execution, (2) problem decomposition, and (3) improved reasoning and knowledge utilization. Using this framework, we gain insights beyond mere accuracy. For instance, providing models with explicit step-by-step plans surprisingly degrades performance on the most challenging benchmarks, yet RL-tuned models exhibit greater robustness, experiencing markedly smaller performance drops than their base counterparts. This suggests that RL may not primarily enhance the execution of external plans but rather empower models to formulate and follow internal strategies better suited to their reasoning processes. Conversely, we observe that RL enhances the model's capacity to integrate provided knowledge into its reasoning process, leading to performance improvements across diverse tasks. We also study difficulty, showing improved training by developing new ways to exploit hard problems. Our findings lay a foundation for more principled training and evaluation of reasoning models.



Paperid:4137
Authors:Ming Nie, Chunwei Wang, Jianhua Han, Hang Xu, Li Zhang
Abstract:
Unified vision-language models have made significant progress in multimodal understanding and generation, yet they largely fall short in producing multimodal interleaved outputs, which is a crucial capability for tasks like visual storytelling and step-by-step visual reasoning.In this work, we propose a reinforcement learning-based post-training strategy to unlock this capability in existing unified models, without relying on large-scale multimodal interleaved datasets.We begin with a warm-up stage using a hybrid dataset comprising curated interleaved sequences and limited data for multimodal understanding and text-to-image generation, which exposes the model to interleaved generation patterns while preserving its pretrained capabilities.To further refine interleaved generation, we propose a unified policy optimization framework that extends Group Relative Policy Optimization (GRPO) to the multimodal setting.Our approach jointly models text and image generation within a single decoding trajectory and optimizes it with our novel hybrid rewards covering textual relevance, visual-text alignment, and structural fidelity.Additionally, we incorporate process-level rewards to provide step-wise guidance, enhancing training efficiency in complex multimodal tasks.Experiments on MMIE and InterleavedBench demonstrate that our approach significantly enhances the quality and coherence of multimodal interleaved generation.



Paperid:4138
Authors:woohyeon Byeon, Giseung Park, Jongseong Chae, Amir Leshem, Youngchul Sung
Abstract:
In this paper, we propose a provably convergent and practical framework for multi-objective reinforcement learning with max-min criterion. From a game-theoretic perspective, we reformulate max-min multi-objective reinforcement learning as a two-player zero-sum regularized continuous game and introduce an efficient algorithm based on mirror descent. Our approach simplifies the policy update while ensuring global last-iterate convergence. We provide a comprehensive theoretical analysis on our algorithm, including iteration complexity under both exact and approximate policy evaluations, as well as sample complexity bounds. To further enhance performance, we modify the proposed algorithm with adaptive regularization.Our experiments demonstrate the convergence behavior of the proposed algorithm in tabular settings, and our implementation for deep reinforcement learning significantly outperforms previous baselines in many MORL environments.



Paperid:4139
Authors:Zhiding Liu, Mingyue Cheng, Guanhao Zhao, Jiqian Yang, Qi Liu, Enhong Chen
Abstract:
Time series forecasting plays a pivotal role in various real-world applications and has attracted significant attention in recent decades. While recent methods have achieved remarkable accuracy by incorporating advanced inductive biases and training strategies, we observe that instance-level variations remain a significant challenge. These variations—stemming from distribution shifts, missing data, and long-tail patterns—often lead to suboptimal forecasts for specific instances, even when overall performance appears strong. To address this issue, we propose a model-agnostic framework, PIR, designed to enhance forecasting performance through Post-forecasting Identification and Revision. Specifically, PIR first identifies biased forecast instances by estimating their predictive accuracy. Based on this, the framework revises the forecasts using contextual information, including covariates and historical time series, from both local and global perspectives in a post-processing fashion. Extensive experiments on real-world datasets with mainstream forecasting models demonstrate that PIR effectively mitigates instance-level errors and significantly improves forecasting reliability.



Paperid:4140
Authors:Peng Xie, Xingyuan Liu, Tsz Wai Chan, Yequan Bie, Yangqiu Song, Yang Wang, Hao CHEN, Kani Chen
Abstract:
Code-switching (CS) is the alternating use of two or more languages within a conversation or utterance, often influenced by social context and speaker identity. This linguistic phenomenon poses challenges for Automatic Speech Recognition (ASR) systems, which are typically designed for a single language and struggle to handle multilingual inputs. The growing global demand for multilingual applications, including Code-Switching ASR (CSASR), Text-to-Speech (CSTTS), and Cross-Lingual Information Retrieval (CLIR), highlights the inadequacy of existing monolingual datasets. Although some code-switching datasets exist, most are limited to bilingual mixing within homogeneous ethnic groups, leaving a critical need for a large-scale, diverse benchmark akin to ImageNet in computer vision. To bridge this gap, we introduce \textbf{LinguaMaster}, a multi-agent collaboration framework specifically designed for efficient and scalable multilingual data synthesis. Leveraging this framework, we curate \textbf{SwitchLingua}, the first large-scale multilingual and multi-ethnic code-switching dataset, including: (1) 420K CS textual samples across 12 languages, and (2) over 80 hours of audio recordings from 174 speakers representing 18 countries/regions and 63 racial/ethnic backgrounds, based on the textual data. This dataset captures rich linguistic and cultural diversity, offering a foundational resource for advancing multilingual and multicultural research. Furthermore, to address the issue that existing ASR evaluation metrics lack sensitivity to code-switching scenarios, we propose the \textbf{Semantic-Aware Error Rate (SAER)}, a novel evaluation metric that incorporates semantic information, providing a more accurate and context-aware assessment of system performance. Benchmark experiments on SwitchLingua with state-of-the-art ASR models reveal substantial performance gaps, underscoring the dataset’s utility as a rigorous benchmark for CS capability evaluation. In addition, SwitchLingua aims to encourage further research to promote cultural inclusivity and linguistic diversity in speech technology, fostering equitable progress in the ASR field. LinguaMaster (Code): github.com/Shelton1013/SwitchLingua, SwitchLingua (Data): https://huggingface.co/datasets/Shelton1013/SwitchLinguatext, https://huggingface.co/datasets/Shelton1013/SwitchLinguaaudio



Paperid:4141
Authors:Dipam Goswami, Simone Magistri, KAI WANG, Bartłomiej Twardowski, Andrew Bagdanov, Joost van de Weijer
Abstract:
Using pre-trained models has been found to reduce the effect of data heterogeneity and speed up federated learning algorithms. Recent works have explored training-free methods using first- and second-order statistics to aggregate local client data distributions at the server and achieve high performance without any training. In this work, we propose a training-free method based on an unbiased estimator of class covariance matrices which only uses first-order statistics in the form of class means communicated by clients to the server. We show how these estimated class covariances can be used to initialize the global classifier, thus exploiting the covariances without actually sharing them. We also show that using only within-class covariances results in a better classifier initialization. Our approach improves performance in the range of 4-26\% with exactly the same communication cost when compared to methods sharing only class means and achieves performance competitive or superior to methods sharing second-order statistics with dramatically less communication overhead. The proposed method is much more communication-efficient than federated prompt-tuning methods and still outperforms them. Finally, using our method to initialize classifiers and then performing federated fine-tuning or linear probing again yields better performance.



Paperid:4142
Authors:Adel Nabli, Louis Fournier, Pierre ERBACHER, Louis Serrano, Eugene Belilovsky, Edouard Oyallon
Abstract:
Abstract:Training Large Language Models (LLMs) relies heavily on distributed implementations, employing multiple GPUs to compute stochastic gradients on model replicas in parallel. However, synchronizing gradients in data parallel settings induces a communication overhead increasing with the number of distributed workers, which can impede the efficiency gains of parallelization. To address this challenge, optimization algorithms reducing inter-worker communication have emerged, such as local optimization methods used in Federated Learning. While effective in minimizing communication overhead, these methods incur significant memory costs, hindering scalability: in addition to extra momentum variables, if communications are only allowed between multiple local optimization steps, then the optimizer's states cannot be sharded among workers. In response, we propose **AC**cumulate while **CO**mmunicate ($\texttt{ACCO}$), a memory-efficient optimization algorithm tailored for distributed training of LLMs. $\texttt{ACCO}$ allows to shard optimizer states across workers, overlaps gradient computations and communications to conceal communication costs, and accommodates heterogeneous hardware. Our method relies on a novel technique to mitigate the one-step delay inherent in parallel execution of gradient computations and communications, eliminating the need for warmup steps and aligning with the training dynamics of standard distributed optimization while converging faster in terms of wall-clock time. We demonstrate the effectiveness of $\texttt{ACCO}$ on several LLMs training and fine-tuning tasks.



Paperid:4143
Authors:Zimeng Huang, Jinxin Ke, Xiaoxuan Fan, Yufeng Yang, Yang Liu, Liu Zhonghan, Zedi Wang, Junteng Dai, Haoyi Jiang, Yuyu Zhou, Keze Wang, Ziliang Chen
Abstract:
Large Vision-Language Models (LVLMs) have exhibited remarkable progress. However, deficiencies remain compared to human intelligence, such as hallucination and shallow pattern matching. In this work, we aim to evaluate a fundamental yet underexplored intelligence: association, a cornerstone of human cognition for creative thinking and knowledge integration. Current benchmarks, often limited to closed-ended tasks, fail to capture the complexity of open-ended association reasoning vital for real-world applications. To address this, we present MM-OPERA, a systematic benchmark with 11,497 instances across two open-ended tasks: Remote-Item Association (RIA) and In-Context Association (ICA), aligning association intelligence evaluation with human psychometric principles. It challenges LVLMs to resemble the spirit of divergent thinking and convergent associative reasoning through free-form responses and explicit reasoning paths. We deploy tailored LLM-as-a-Judge strategies to evaluate open-ended outputs, applying process-reward-informed judgment to dissect reasoning with precision. Extensive empirical studies on state-of-the-art LVLMs, including sensitivity analysis of task instances, validity analysis of LLM-as-a-Judge strategies, and diversity analysis across abilities, domains, languages, cultures, etc., provide a comprehensive and nuanced understanding of the limitations of current LVLMs in associative reasoning, paving the way for more human-like and general-purpose AI.



Paperid:4144
Authors:Owen Queen, Harrison Zhang, James Zou
Abstract:
Variant and gene interpretation are vital prerequisites for clinical genetics and personalized medicine as they guide the diagnosis and management of rare or common diseases with strong genetic etiologies. However, traditional approaches for this task are manual and labor-intensive.Generative language models (LMs) provide an opportunity to facilitate the review of genes and variants, thereby accelerating the translation of genetic sequencing data into clinically-actionable insights. While existing benchmarks have attempted to quantify the capabilities of LMs for interpreting scientific data, these benchmarks often focus on narrow tasks that do not translate to real-world research.Harnessing the ClinGen database, a resource of expert-curated literature interpretations of publications in clinical genetics, we built CGBench, a robust benchmark that tests complex reasoning capabilities of LMs on scientific publications.The tasks in CGBench measure the ability to 1) extract relevant experimental results following precise protocols and guidelines,2) judge the strength of presented evidence, and 3) categorize and describe assays and experiments and their outcomes as is relevant to specific genes and variants.We test 8 different LMs on these tasks and find that while models show promise, substantial gaps still exist in abilities to correctly interpret literature, especially fine-grained instructions.Our experiments reveal that while reasoning models are best-performing across fine-grained instructions, non-reasoning models are still better at answering high-level questions about scientific data. Additionally, in-context learning of demonstrations can significantly boost performance.Our evaluation of explanations shows that models often hallucinate or misinterpret results even when correctly classifying evidence.CGBench introduces a rigorous, challenging benchmark that precisely measures scientific literature interpretation, revealing strengths and points of weakness for development of LMs and agentic systems.



Paperid:4145
Authors:Sanghyun Ahn, Wonje Choi, Junyong Lee, Jinwoo Park, Honguk Woo
Abstract:
Recent advances in large language models (LLMs) have enabled the automatic generation of executable code for task planning and control in embodied agents such as robots, demonstrating the potential of LLM-based embodied intelligence. However, these LLM-based code-as-policies approaches often suffer from limited environmental grounding, particularly in dynamic or partially observable settings, leading to suboptimal task success rates due to incorrect or incomplete code generation. In this work, we propose a neuro-symbolic embodied task planning framework that incorporates explicit symbolic verification and interactive validation processes during code generation. In the validation phase, the framework generates exploratory code that actively interacts with the environment to acquire missing observations while preserving task-relevant states. This integrated process enhances the grounding of generated code, resulting in improved task reliability and success rates in complex environments. We evaluate our framework on RLBench and in real-world settings across dynamic, partially observable scenarios. Experimental results demonstrate that our framework improves task success rates by 46.2\% over Code as Policies baselines and attains over 86.8\% executability of task-relevant actions, thereby enhancing the reliability of task planning in dynamic environments.



Paperid:4146
Authors:Yuheng Zhao, Yu-Hu Yan, Kfir Y. Levy, Peng Zhao
Abstract:
Smoothness is known to be crucial for acceleration in offline optimization, and for problem-dependent regret that scale with gradient variations in online learning. Interestingly, these two problems are actually closely connected -- accelerated optimization can be understood from the lens of gradient-variation online learning. In this paper, we systematically investigate online learning with \emph{Hölder smooth} functions, a class encompassing both smooth and non-smooth (Lipschitz) functions, and further explore its implications for offline optimization. First, we propose an online algorithm with optimal gradient-variation regret for convex functions, which implies an optimal accelerated method for stochastic convex optimization under Hölder smoothness. Then, we extend the results in three aspects: \rom{1} strongly convex functions, \rom{2} non-stationary online learning where the comparator sequence changes over time; \rom{3} universal online learning where the curvature of functions is unknown. In all three cases, we have achieved the first gradient-variation regret that can well interpolate the results between the smooth and Lipschitz regimes, and recover the optimal results in each case. Notably, our proposed algorithms do not require prior knowledge of the Hölder smoothness parameter, greatly improving the adaptivity of existing methods that depend on this parameter, even when designed specifically for smooth functions. Finally, we demonstrate several implications for offline optimization through carefully tailored online-to-batch conversions.



Paperid:4147
Authors:Melody Li, Kumar Krishna Agrawal, Arna Ghosh, Komal Teru, Adam Santoro, Guillaume Lajoie, Blake Richards
Abstract:
The geometry of representations in a neural network can significantly impact downstream generalization. It is unknown how representation geometry changes in large language models (LLMs) over pretraining and post-training. Here, we characterize the evolving geometry of LLM representations using spectral methods (effective rank and eigenspectrum decay). With the OLMo and Pythia model families we uncover a consistent non-monotonic sequence of three distinct geometric phases in pretraining. An initial \warmup phase sees rapid representational compression. This is followed by an "entropy-seeking" phase, characterized by expansion of the representation manifold's effective dimensionality, which correlates with an increase in memorization. Subsequently, a "compression seeking" phase imposes anisotropic consolidation, selectively preserving variance along dominant eigendirections while contracting others, correlating with improved downstream task performance. We link the emergence of these phases to the fundamental interplay of cross-entropy optimization, information bottleneck, and skewed data distribution. Additionally, we find that in post-training the representation geometry is further transformed: Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO) correlate with another "entropy-seeking" dynamic to integrate specific instructional or preferential data, reducing out-of-distribution robustness. Conversely, Reinforcement Learning with Verifiable Rewards (RLVR) often exhibits a "compression seeking" dynamic, consolidating reward-aligned behaviors and reducing the entropy in its output distribution. This work establishes the utility of spectral measures of representation geometry for understanding the multiphase learning dynamics within LLMs.



Paperid:4148
Authors:Sangmin Bae, Yujin Kim, Reza Bayat, Sungnyun Kim, Tal Schuster, Adam Fisch, Hrayr Harutyunyan, Ziwei Ji, Jiyoun Ha, Aaron Courville, Se-Young Yun
Abstract:
Scaling Transformer language models unlocks impressive capabilities, but the accompanying computational and memory demands make both training and deployment expensive. Existing efficiency efforts typically target either parameter sharing or adaptive computation, leaving open the question of how to attain both simultaneously. We introduce Mixture-of-Recursions (MoR), a unified framework that combines the two axes of efficiency inside a single Recursive Transformer. MoR reuses a shared stack of layers across recursion steps to achieve parameter efficiency, while a lightweight token-level router dynamically assigns recursion depth to tokens, thereby focusing quadratic attention computation only where it is most useful. Further enhancing its efficiency, MoR incorporates a recursion-wise key-value caching mechanism that eliminates redundant memory access across recursion steps by selectively storing only the key-value caches for designated tokens. Across pretraining runs at model scales ranging from 135M to 1.7B parameters, MoR forms a new Pareto frontier: at equal training FLOPs, it significantly lowers validation perplexity and improves few-shot accuracy, while delivering up to 2.18x higher inference throughput compared with vanilla and existing recursive baselines. These gains demonstrate that MoR is an effective path towards large-model quality without incurring large-model cost.



Paperid:4149
Authors:Xiaobao Wang, Ruoxiao Sun, Yujun Zhang, Bingdao Feng, Dongxiao He, Luzhi Wang, Di Jin
Abstract:
Graph Neural Networks (GNNs) have demonstrated strong performance across tasks such as node classification, link prediction, and graph classification, but remain vulnerable to backdoor attacks that implant imperceptible triggers during training to control predictions. While node-level attacks exploit local message passing, graph-level attacks face the harder challenge of manipulating global representations while maintaining stealth. We identify two main sources of anomaly in existing graph classification backdoor methods: structural deviation from rare subgraph triggers and semantic deviation caused by label flipping, both of which make poisoned graphs easily detectable by anomaly detection models. To address this, we propose DPSBA, a clean-label backdoor framework that learns in-distribution triggers via adversarial training guided by anomaly-aware discriminators. DPSBA effectively suppresses both structural and semantic anomalies, achieving high attack success while significantly improving stealth. Extensive experiments on real-world datasets validate that DPSBA achieves a superior balance between effectiveness and detectability compared to state-of-the-art baselines.



Paperid:4150
Authors:Alper KALLE, Théo Rudkiewicz, Mohamed Ouerfelli, Mohamed Tamaazousti
Abstract:
Abstract:Neural networks are widely used for image–related tasks but typically demand considerable computing power. Once a network has been trained, however, its memory‑ and compute‑footprint can be reduced by compression. In this work, we focus on compression through tensorization and low‑rank representations. Whereas classical approaches search for a low‑rank approximation by minimizing an isotropic norm such as the Frobenius norm in weight‑space, we use data‑informed norms that measure the error in function space. Concretely, we minimize the change in the layer’s output distribution, which can be expressed as $\lVert (W - \widetilde{W}) \Sigma^{1/2}\rVert_F$ where $\Sigma^{1/2}$ is the square root of the covariance matrix of the layer’s input and $W$, $\widetilde{W}$ are the original and compressed weights. We propose new alternating least square algorithms for the two most common tensor decompositions (Tucker‑2 and CPD) that directly optimize the new norm. Unlike conventional compression pipelines, which almost always require post‑compression fine‑tuning, our data‑informed approach often achieves competitive accuracy without any fine‑tuning. We further show that the same covariance‑based norm can be transferred from one dataset to another with only a minor accuracy drop, enabling compression even when the original training dataset is unavailable.Experiments on several CNN architectures (ResNet‑18/50, and GoogLeNet) and datasets (ImageNet, FGVC‑Aircraft, Cifar10, and Cifar100) confirm the advantages of the proposed method.



Paperid:4151
Authors:Gijs Wijngaard, Elia Formisano, Michele Esposito, Michel Dumontier
Abstract:
Audio-language models have shown promising results in various sound understanding tasks, yet they remain limited in their ability to reason over the fine-grained semantics of sound. In this paper, we present AudSemThinker, a model whose reasoning is structured around a framework of auditory semantics inspired by human cognition. To support this, we introduce AudSem, a novel dataset specifically curated for semantic descriptor reasoning in audio-language models. AudSem addresses the persistent challenge of data contamination in zero-shot evaluations by providing a carefully filtered collection of audio samples paired with captions generated through a robust multi-stage pipeline. Our experiments demonstrate that AudSemThinker outperforms state-of-the-art models across multiple training settings, highlighting its strength in semantic audio reasoning. Both AudSemThinker and the AudSem dataset will be publicly released upon acceptance.



Paperid:4152
Authors:Riccardo Cadei, Ilker Demirel, Piersilvio De Bartolomeis, Lukas Lindorfer, Sylvia Cremer, Cordelia Schmid, Francesco Locatello
Abstract:
In many scientific domains, the cost of data annotation limits the scale and pace of experimentation. Yet, modern machine learning systems offer a promising alternative—provided their predictions yield correct conclusions. We focus on Prediction-Powered Causal Inferences (PPCI), i.e., estimating the treatment effect in a target experiment with unlabeled factual outcomes, retrievable zero-shot from a pre-trained model. We first identify the conditional calibration property to guarantee valid PPCI at population level. Then, we introduce a new necessary ``causal lifting'' constraint transferring validity across experiments, which we propose to enforce in practice in Deconfounded Empirical Risk Minimization, our new model-agnostic training objective. We validate our method on synthetic and real-world scientific data, offering solutions to instances not solvable by vanilla Empirical Risk Minimization and invariant training. In particular, we solve zero-shot PPCI on the ISTAnt dataset for the first time, fine-tuning a foundational model on our replica dataset of their ecological experiment with a different recording platform and treatment.



Paperid:4153
Authors:Yan Xu, Yixing Wang, Stella X. Yu
Abstract:
We reinterpret sparse-input novel view synthesis as completing a natural video---recovering missing frames between wide-baseline inputs. This perspective allows us to leverage strong generative priors from pretrained video diffusion models to synthesize plausible intermediate views, providing constraints for under-observed regions during 3D-GS training.We propose a \textit{zero-shot, generation-guided reconstruction} framework where pseudo views are generated at specified camera poses via a video diffusion model, guided by an \textit{uncertainty-aware modulation mechanism} for spatial coherence. These generated views provide dense supervision for training 3D-GS and serve as a bridge to \textit{densify Gaussian primitives} in under-observed regions, helping the 3D-GS model reconstruct the scene appearance and geometry. We further refine pseudo views and 3D-GS iteratively, allowing geometry-aware feedback to improve synthesis quality. Our approach enhances the 3D representation in sparse-view settings and significantly reduces rendering artifacts. Experiments on DL3DV, LLFF, and DTU demonstrate that our approach consistently outperforms baseline 3D-GS with substantial improvements.



Paperid:4154
Authors:Ran Li, Hao Wang, Chengzhi Mao
Abstract:
Efficient red-teaming method to uncover vulnerabilities in Large Language Models (LLMs) is crucial. While recent attacks often use LLMs as optimizers, the discrete language space make gradient-based methods struggle. We introduce LARGO (Latent Adversarial Reflection through Gradient Optimization), a novel latent self-reflection attack that reasserts the power of gradient-based optimization for generating fluent jailbreaking prompts. By operating within the LLM's continuous latent space, LARGO first optimizes an adversarial latent vector and then recursively call the same LLM to decode the latent into natural language. This methodology yields a fast, effective, and transferable attack that produces fluent and stealthy prompts. On standard benchmarks like AdvBench and JailbreakBench, LARGO surpasses leading jailbreaking techniques, including AutoDAN, by 44 points in attack success rate. Our findings demonstrate a potent alternative to agentic LLM prompting, highlighting the efficacy of interpreting and attacking LLM internals through gradient optimization.



Paperid:4155
Authors:Joshua Tian Jin Tee, Hee Suk Yoon, Abu Hanif Muhammad Syarubany, Eunseop Yoon, Chang Yoo
Abstract:
Direct Preference Optimization (DPO) is a key framework for aligning text-to-image models with human preferences, extended by Stepwise Preference Optimization (SPO) to leverage intermediate steps for preference learning, generating more aesthetically pleasing images with significantly less computational cost. While effective, SPO's underlying mechanisms remain underexplored. In light of this, We critically re-examine SPO by formalizing its mechanism as gradient guidance. This novel lens reveals SPO’s biased temporal weighting—underweighting later generative steps—and, uniquely compared to likelihood-centric views, highlights the presence of significant noise in these gradient estimates. Leveraging these insights, our GradSPO algorithm introduces a simplified loss and a targeted, variance-informed noise reduction strategy, enhancing training stability. Evaluations on SD 1.5 and SDXL show GradSPO substantially outperforms leading baselines in human preference, yielding images with markedly improved aesthetics and semantic faithfulness, leading to more robust alignment.



Paperid:4156
Authors:Haibo Tong, Zhaoyang Wang, Zhaorun Chen, Haonian Ji, Shi Qiu, Siwei Han, Kexin Geng, Zhongkai Xue, Yiyang Zhou, Peng Xia, Mingyu Ding, Rafael Rafailov, Chelsea Finn, Huaxiu Yao
Abstract:
Recent advancements in video generation have significantly improved the ability to synthesize videos from text instructions. However, existing models still struggle with key challenges such as instruction misalignment, content hallucination, safety concerns, and generation bias. To address these limitations, we introduce MJ-BENCH-VIDEO, a large-scale video preference benchmark designed to evaluate video generation across five critical aspects: Alignment, Safety, Fineness, Coherence & Consistency, and Bias & Fairness. This benchmark further incorporates 28 fine-grained criteria to provide a comprehensive evaluation of video preference. Building upon this dataset, we propose MJ-VIDEO, a Mixture-of-Experts (MoE)-based video reward model designed to deliver fine-grained reward. MJ-VIDEO can dynamically select relevant experts to accurately judge the preference based on the input text-video pair. This architecture enables more precise and adaptable preference judgments. Through extensive benchmarking on MJ-BENCH-VIDEO, we analyze the limitations of existing video reward models and demonstrate the superior performance of MJ-VIDEO in video preference assessment, achieving 17.58% and 15.87% improvements in overall and fine-grained preference judgments, respectively. Additionally, MJ-VIDEO is able to improve the alignment performance in video generation via preference fine-tuning.



Paperid:4157
Authors:Dingling Yao, Shimeng Huang, Riccardo Cadei, Kun Zhang, Francesco Locatello
Abstract:
Causal reasoning and discovery, two fundamental tasks of causal analysis, often face challenges in applications due to the complexity, noisiness, and high-dimensionality of real-world data.Despite recent progress in identifying latent causal structures using causal representation learning (CRL), what makes learned representations useful for causal downstream tasks and how to evaluate them are still not well understood.In this paper, we reinterpret CRL using a measurement model framework, where the learned representations are viewed as proxy measurements of the latent causal variables. Our approach clarifies the conditions under which learned representations support downstream causal reasoning and provides a principled basis for quantitatively assessing the quality of representations using a new Test-based Measurement EXclusivity (T-MEX) score. We validate T-MEX across diverse causal inference scenarios, including numerical simulations and real-world ecological video analysis, demonstrating that the proposed framework and corresponding score effectively assess the identification of learned representations and their usefulness for causal downstream tasks.



Paperid:4158
Authors:Xiaoxuan Gong, Jie Ma
Abstract:
Existing research in low-level vision has shifted its focus from "one-by-one" task-specific methods to "all-in-one" multi-task unified architectures. However, current all-in-one image restoration approaches primarily aim to improve overall performance across a limited number of tasks. In contrast, how to incrementally add new image restoration capabilities on top of an existing model — that is, task-incremental learning — has been largely unexplored. To fill this research gap, we propose a minimalistic and universal paradigm for task-incremental learning called MINI. It addresses the problem of parameter interference across different tasks through a simple yet effective mechanism, enabling nearly forgetting-free task-incremental learning. Specifically, we design a special meta-convolution called MINI-Conv, which generates parameters solely through lightweight embeddings instead of complex convolutional networks or MLPs. This not only significantly reduces the number of parameters and computational overhead but also achieves complete parameter isolation across different tasks. Moreover, MINI-Conv can be seamlessly integrated as a plug-and-play replacement for any convolutional layer within existing backbone networks, endowing them with incremental learning capabilities. Therefore, our method is highly generalizable. Finally, we demonstrate that our method achieves state-of-the-art performance compared to existing incremental learning approaches across five common image restoration tasks. Moreover, the near forgetting-free nature of our method makes it highly competitive even against all-in-one image restoration methods trained in a full-supervised manner. Our code is available at https://github.com.



Paperid:4159
Authors:Jintao Tong, Wenwei Jin, Pengda Qin, Anqi Li, Yixiong Zou, Yuhong Li, Yuhua Li, Ruixuan Li
Abstract:
Abstract:Large vision-language models (LVLMs) excel at multimodal understanding but suffer from high computational costs due to redundant vision tokens. Existing pruning methods typically rely on single-layer attention scores to rank and prune redundant visual tokens to solve this inefficiency. However, as the interaction between tokens and layers is complicated, this raises a basic question: Is such a simple single-layer criterion sufficient to identify redundancy? To answer this question, we rethink the emergence of redundant visual tokens from a fundamental perspective: information flow, which models the interaction between tokens and layers by capturing how information moves between tokens across layers. We find (1) the CLS token acts as an information relay, which can simplify the complicated flow analysis; (2) the redundancy emerges progressively and dynamically via layer-wise attention concentration; and (3) relying solely on attention scores from single layers can lead to contradictory redundancy identification. Based on this, we propose FlowCut, an information-flow-aware pruning framework, mitigating the insufficiency of the current criterion for identifying redundant tokens and better aligning with the model's inherent behaviors. Extensive experiments show FlowCut achieves superior results, outperforming SoTA by 1.6% on LLaVA-1.5-7B with 88.9% token reduction, and by 4.3% on LLaVA-NeXT-7B with 94.4% reduction, delivering 3.2$\times$ speed-up in the prefilling stage. Our codes will be released.



Paperid:4160
Authors:Yunan Lu, Xixi Zhang, Yaojin Lin, Weiwei Li, Lei Yang, Xiuyi Jia
Abstract:
Label distribution in recent years has been applied in a diverse array of complex decision-making tasks. To address the availability of label distributions, label enhancement has been established as an effective learning paradigm that aims to automatically infer label distributions from readily available multi-label data, e.g., logical labels. Recently, numerous works have demonstrated that the label ranking is significantly beneficial to label enhancement. However, these works still exhibit deficiencies in representing the probabilistic relationships between label distribution and tie-allowed label rankings. Therefore, we propose PROM, a pairwise ranking model with orderliness and monotonicity, to explain the probabilistic relationship between label distributions and label rankings. Specifically, we propose the monotonicity and orderliness assumptions for the probabilities of different ranking relationships and derive the mass functions for PROM, which are theoretically ensured to preserve the monotonicity and orderliness. Further, we propose a generative label enhancement algorithm based on PROM, which directly learns a label distribution predictor from the readily available multi-label data. Finally, extensive experiments demonstrate the efficacy of our proposed PROM.



Paperid:4161
Authors:Jinyong Jeong, Hyungu Kahng, Seoung Bum Kim
Abstract:
Abstract:Distribution shifts between training and test data undermine the reliability of deep neural networks, challenging real-world applications across domains and subpopulations. While distributionally robust optimization (DRO) methods like GroupDRO aim to improve robustness by optimizing worst-case performance over predefined groups, their use of a single global classifier can be restrictive when facing substantial inter-environment variability. We propose Multi-Expert Distributionally Robust Optimization (MEDRO), a novel extension of GroupDRO designed to address such complex shifts. MEDRO employs a shared feature extractor with $m$ environment-specific expert classifier heads, and introduces a min-max objective over all $m^{2}$ expert-environment pairings, explicitly modeling cross-environment risks. This expanded uncertainty set captures fine-grained distributional variations that a single classifier might overlook. Empirical evaluations on a range of standard distribution shift benchmarks demonstrate that MEDRO often achieves robust predictive performance compared to existing methods. Furthermore, MEDRO offers practical inference strategies, such as ensembling or gating mechanisms, for typical scenarios where environment labels are unavailable at test time. Our findings suggest MEDRO as a promising step toward resilient and generalizable machine learning under real-world distribution shifts.



Paperid:4162
Authors:Xiaoxiao Ma, Feng Zhao, Pengyang Ling, Haibo Qiu, Zhixiang Wei, Hu Yu, Jie Huang, Zhixiong Zeng, Lin Ma
Abstract:
In this work, we first revisit the sampling issues in current autoregressive (AR) image generation models and identify that image tokens, unlike text tokens, exhibit lower information density and non-uniform spatial distribution. Accordingly, we present an entropy-informed decoding strategy that facilitates higher autoregressive generation quality with faster synthesis speed. Specifically, the proposed method introduces two main innovations: 1) dynamic temperature control guided by spatial entropy of token distributions, enhancing the balance between content diversity, alignment accuracy, and structural coherence in both mask-based and scale-wise models, without extra computational overhead, and 2) entropy-aware acceptance rules in speculative decoding, achieving near-lossless generation at about 85% of the inference cost of conventional acceleration methods. Extensive experiments across multiple benchmarks using diverse AR image generation models demonstrate the effectiveness and generalizability of our approach in enhancing both generation quality and sampling speed.



Paperid:4163
Authors:Chuhao Zhou, Jianfei Yang
Abstract:
Embodied agents operating in smart homes must understand human behavior through diverse sensory inputs and communicate via natural language. While Vision-Language Models (VLMs) have enabled impressive language-grounded perception, their reliance on visual data limits robustness in real-world scenarios with occlusions, poor lighting, or privacy constraints. In this paper, we introduce HoloLLM, a Multimodal Large Language Model (MLLM) that integrates uncommon but powerful sensing modalities, such as LiDAR, infrared, mmWave radar, and WiFi, to enable seamless human perception and reasoning across heterogeneous environments. We address two key challenges: (1) the scarcity of aligned modality-text data for rare sensors, and (2) the heterogeneity of their physical signal representations. To overcome these, we design a Universal Modality-Injection Projector (UMIP) that enhances pre-aligned modality embeddings with fine-grained, text-aligned features from tailored encoders via coarse-to-fine cross-attention without introducing significant alignment overhead. We further introduce a human-VLM collaborative data curation pipeline to generate paired textual annotations for sensing datasets. Extensive experiments on two newly constructed benchmarks show that HoloLLM significantly outperforms existing MLLMs, improving language-grounded human sensing accuracy by up to 30%. This work establishes a new foundation for real-world, language-informed multisensory embodied intelligence.



Paperid:4164
Authors:Yicong Chen, Jiahua Rao, Jiancong Xie, Dahao Xu, Zhen Wang, Yuedong Yang
Abstract:
Virtual screening (VS) is vital for drug discovery but struggles with low hit rates and high computational costs. While active learning (AL) has shown promise in improving the efficiency of VS, traditional methods rely on inflexible and handcrafted heuristics, limiting adaptability in complex chemical spaces, particularly in balancing molecular diversity and selection accuracy. To overcome these challenges, we propose GLARE, a reinforced active learning framework that reformulates VS as a Markov Decision Process (MDP). Using Group Relative Policy Optimization (GRPO), GLARE dynamically balances chemical diversity, biological relevance, and computational constraints, eliminating the need for inflexible heuristics.Experiments show GLARE outperforms state-of-the-art AL methods, with a 64.8% average improvement in enrichment factors (EF). Additionally, GLARE enhances the performance of VS foundation models like DrugCLIP, achieving up to an 8-fold improvement in EF-0.5% with as few as 15 active molecules. These results highlight the transformative potential of GLARE for adaptive and efficient drug discovery.



Paperid:4165
Authors:Rongyuan Wu, Lingchen Sun, Zhengqiang ZHANG, Shihao Wang, Tianhe Wu, Qiaosi Yi, Shuai Li, Lei Zhang
Abstract:
Real-world image super-resolution (Real-ISR) benefits from pre-trained text-to-image diffusion models, which can generate rich details. However, these models exhibit inherent stochasticity, where different noise inputs yield results of varying perceptual quality. This stochasticity actually widens the perceptual qualityrangeof outputs, which motivates us to developDirectPerceptualPreferenceOptimization for Real-ISR(DP²O-SR). We first define a Combined Perceptual Score (CPS) to quantify image perceptual quality preference, then employ Direct Preference Optimization (DPO) to fine-tune the pre-trained Real-ISR model towards generating outputs with higher CPS scores. While simply increasing the number of noise samples could improve DPO performance, this is computationally expensive. We therefore propose to extract more training signals from a limited group of candidate outputs. Instead of forming only one preference pair (best and worst samples) per input, we generate multiple pairs by selecting higher- and lower-scoring samples from the same group, significantly increasing the DPO training efficacy. In addition, considering the unequal importance of these pairs, we introduce Hierarchical Significance Weighting (HSW) to weight the DPO loss based on intra-pair significance (CPS difference within the pair) and inter-group learning potential (CPS variance of the source candidate group), prioritizing more informative pairs and maximizing DPO learning from available samples. Experiments show thatDP²O-SRsignificantly enhances the perceptual quality over the diffusion-based Real-ISR models, validated by quantitative metrics and visualization comparison. The code, model, and data will be made publicly available.



Paperid:4166
Authors:Moritz Willig, Tim Woydt, Devendra Singh Dhami, Kristian Kersting
Abstract:
Many causal inference frameworks rely on a staticity assumption, where repeated interventions are expected to yield consistent outcomes, often summarized by metrics like the Average Treatment Effect (ATE). This assumption, however, frequently fails in dynamic environments where interventions can alter the system's underlying causal structure, rendering traditional `static' ATE insufficient or misleading. Recent works on meta-causal models (MCM) offer a promising avenue by enabling qualitative reasoning over evolving relationships. In this work, we propose a specific class of MCM with desirable properties for explicitly modeling and predicting intervention outcomes under meta-causal dynamics. Through expository examples in high-impact domains of medical treatment and judicial decision-making, we highlight the severe consequences that arise when system dynamics are neglected and demonstrate the successful application of meta-causal strategies to navigate these challenges.



Paperid:4167
Authors:Alexander Lappe, Martin Giese
Abstract:
Recent work has shown that the attention maps of the widely popular DINOv2 model exhibit artifacts, which hurt both model interpretability and performance on dense image tasks. These artifacts emerge due to the model repurposing patch tokens with redundant local information for the storage of global image information. To address this problem, additional register tokens have been incorporated in which the model can store such information instead. We carefully examine the influence of these register tokens on the relationship between global and local image features, showing that while register tokens yield cleaner attention maps, these maps do not accurately reflect the integration of local image information in large models. Instead, global information is dominated by information extracted from register tokens, leading to a disconnect between local and global features. Inspired by these findings, we show that the [CLS] token itself, which can be interpreted as a register, leads to a very similar phenomenon in models without explicit register tokens. Our work shows that care must be taken when interpreting attention maps of large ViTs. Further, by clearly attributing the faulty behaviour to register and [CLS] tokens, we show a path towards more interpretable vision models.



Paperid:4168
Authors:Junyan Liu, Ziyun Chen, Kun Wang, Haipeng Luo, Lillian Ratliff
Abstract:
Abstract:We study the Pandora’s Box problem in an online learning setting with semi-bandit feedback. In each round, the learner sequentially pays to open up to $n$ boxes with unknown reward distributions, observes rewards upon opening, and decides when to stop. The utility of the learner is the maximum observed reward minus the cumulative cost of opened boxes, and the goal is to minimize regret defined as the gap between the cumulative expected utility and that of the optimal policy. We propose a new algorithm that achieves $\widetilde{O}(\sqrt{nT})$ regret after $T$ rounds, which improves the $\widetilde{O}(n\sqrt{T})$ bound of Agarwal et al. [2024] and matches the known lower bound up to logarithmic factors. To better capture real-life applications, we then extend our results to a natural but challenging contextual linear setting, where each box's expected reward is linear in some known but time-varying $d$-dimensional context and the noise distribution is fixed over time.We design an algorithm that learns both the linear function and the noise distributions, achieving $\widetilde{O}(nd\sqrt{T})$ regret. Finally, we show that our techniques also apply to the online Prophet Inequality problem, where the learner must decide immediately whether or not to accept a revealed reward. In both non-contextual and contextual settings, our approach achieves similar improvements and regret bounds.



Paperid:4169
Authors:Zitao Yang, Amin Ullah, Shuai Li, Fuxin Li, Jun Li
Abstract:
This paper introduces the Convex Potential Mirror Langevin Algorithm (CPMLA), a novel method to improve sampling efficiency for Energy-Based Models (EBMs). CPMLA uses mirror Langevin dynamics with a convex potential flow as a dynamic mirror map for EBM sampling. This dynamic mirror map enables targeted geometric exploration on the data manifold, accelerating convergence to the target distribution. Theoretical analysis proves that CPMLA achieves exponential convergence with vanishing bias under relaxed log-concave conditions, supporting its efficiency in adapting to complex data distributions. Experiments on benchmarks like CIFAR-10, SVHN, and CelebA demonstrate CPMLA's improved sampling quality and inference efficiency over existing techniques.



Paperid:4170
Authors:Dongjie Yang, Chengqiang Lu, Qimeng Wang, Xinbei Ma, Yan Gao, Yao Hu, Hai Zhao
Abstract:
Abstract:Travel planning is a complex task requiring the integration of diverse real-world information and user preferences. While LLMs show promise, existing methods with long-horizon thinking struggle with handling multifaceted constraints and preferences in the context, leading to suboptimal itineraries. We formulate this as an $L^3$ planning problem, emphasizing long context, long instruction, and long output. To tackle this, we introduce Multiple Aspects of Planning (MAoP), enabling LLMs to conduct wide-horizon thinking to solve complex planning problems. Instead of direct planning, MAoP leverages the strategist to conduct pre-planning from various aspects and provide the planning blueprint for planning models, enabling strong inference-time scalability for better performance. In addition, current benchmarks overlook travel's dynamic nature, where past events impact subsequent journeys, failing to reflect real-world feasibility. To address this, we propose Travel-Sim, an agent-based benchmark assessing plans via real-world travel simulation. This work advances LLM capabilities in complex planning and offers novel insights for evaluating sophisticated scenarios through agent-based simulation.



Paperid:4171
Authors:Jing Wang, Wonho Bae, Jiahong Chen, Wenxu Wang, Junhyug Noh
Abstract:
Recent work on latent diffusion models (LDMs) has focused almost exclusively on generative tasks, leaving their potential for discriminative transfer largely unexplored. We introduce Discriminative Vicinity Diffusion (DVD), a novel LDM-based framework for a more practical variant of source-free domain adaptation (SFDA): the source provider may share not only a pre-trained classifier but also an auxiliary latent diffusion module, trained once on the source data and never exposing raw source samples. DVD encodes each source feature’s label information into its latent vicinity by fitting a Gaussian prior over its k-nearest neighbors and training the diffusion network to drift noisy samples back to label-consistent representations. During adaptation, we sample from each target feature’s latent vicinity, apply the frozen diffusion module to generate source-like cues, and use a simple InfoNCE loss to align the target encoder to these cues, explicitly transferring decision boundaries without source access. Across standard SFDA benchmarks, DVD outperforms state-of-the-art methods. We further show that the same latent diffusion module enhances the source classifier’s accuracy on in-domain data and boosts performance in supervised classification and domain generalization experiments. DVD thus reinterprets LDMs as practical, privacy-preserving bridges for explicit knowledge transfer, addressing a core challenge in source-free domain adaptation that prior methods have yet to solve.



Paperid:4172
Authors:Xinyu Yang, Yuwei An, Hongyi Liu, Tianqi Chen, Beidi Chen
Abstract:
Abstract:Autoregressive Large Language Models (AR-LLMs) frequently exhibit implicit parallelism in sequential generation. Inspired by this, we introduce **Multiverse**, a new generative model that enables natively parallel generation. Multiverse internalizes a MapReduce paradigm, generating automatically through three stages: (i) a *Map* stage for adaptive task decomposition, (ii) a *Process* stage for parallel subtask execution, and (iii) a *Reduce* stage for lossless result synthesis. Next, we build a real-world Multiverse reasoning model with co-design of data, algorithm, and system, enabling rapid and seamless transfer from frontier AR-LLMs. Starting from sequential reasoning chains, we create **Multiverse 1K** by converting them into structured training data using an automated LLM-assisted pipeline, avoiding costly human annotations. Algorithmically, we design **Multiverse Attention** to separate parallel reasoning steps while keeping compatibility with causal attention for efficient training. Systematically, we implement **Multiverse Engine** to enable parallel inference. It features a dedicated scheduler that dynamically switches between sequential and parallel generation, triggered directly by the model. After a 3-hour fine-tuning with 1K examples, *our Multiverse-32B stands as the only non-AR model achieving performance on par with leading AR-LLMs of the same scale*, evidenced by AIME24 & 25 scores of 52% and 43%, respectively. Moreover, our budget control experiments show that Multiverse-3B exhibits superior scaling, outperforming AR-LLMs by 1.87% on average using the same context length. Such parallel scaling further leads to practical efficiency gain, achieving up to 2$\times$ speedup across varying batch sizes.



Paperid:4173
Authors:Linjian Meng, Youzhi Zhang, Shangdong Yang, Tianyu Ding, Zhenxing Ge, Wenbin Li, Tianpei Yang, Bo An, Yang Gao
Abstract:
Abstract:To establish last-iterate convergence for Counterfactual Regret Minimization (CFR) algorithms in learning a Nash equilibrium (NE) of extensive-form games (EFGs), recent studies reformulate learning an NE of the original EFG as learning the NEs of a sequence of (perturbed) regularized EFGs. Hence, proving last-iterate convergence in solving the original EFG reduces to proving last-iterate convergence in solving (perturbed) regularized EFGs. However, these studies only establish last-iterate convergence for Online Mirror Descent (OMD)-based CFR algorithms instead of Regret Matching (RM)-based CFR algorithms in solving perturbed regularized EFGs, resulting in a poor empirical convergence rate, as RM-based CFR algorithms typically outperform OMD-based CFR algorithms. In addition, as solving multiple perturbed regularized EFGs is required, fine-tuning across multiple perturbed regularized EFGs is infeasible, making parameter-free algorithms highly desirable. This paper show that CFR$^+$, a classical parameter-free RM-based CFR algorithm, achieves last-iterate convergence in learning an NE of perturbed regularized EFGs. This is the first parameter-free last-iterate convergence for RM-based CFR algorithms in perturbed regularized EFGs. Leveraging CFR$^+$ to solve perturbed regularized EFGs, we get Reward Transformation CFR$^+$ (RTCFR$^+$). Importantly, we extend prior work on the parameter-free property of CFR$^+$, enhancing its stability, which is vital for the empirical convergence of RTCFR$^+$. Experiments show that RTCFR$^+$ exhibits a significantly faster empirical convergence rate than existing algorithms that achieve theoretical last-iterate convergence.



Paperid:4174
Authors:Fengmiao Bian, Jinyang ZHENG, Ziyun Liu, Jianzhou Luo, Jian-Feng CAI
Abstract:
Finding low-rank matrix weights is a key technique for addressing the high memory usage and computational demands of large models. Most existing algorithms rely on the factorization of the low-rank matrix weights, which is non-unique and redundant. Their convergence is slow especially when the target low-rank matrices are ill-conditioned, because the convergence rate depends on the condition number of the Jacobian operator for the factorization and the Hessian of the loss function with respect to the weight matrix. To address this challenge, we adopt the Riemannian gradient descent (RGD) algorithm on the Riemannian manifold of fixed-rank matrices to update the entire low-rank weight matrix. This algorithm completely avoids the factorization, thereby eliminating the negative impact of the Jacobian condition number. Furthermore, by leveraging the geometric structure of the Riemannian manifold and selecting an appropriate metric, it mitigates the negative impact of the Hessian condition number. Ultimately, this results in our two plug-and-play optimizers: RAdaGrad and RAdamW, which are RGD with metrics adapted from AdaGrad and AdamW and restricted to the manifold. Our algorithms can be seamlessly integrated with various deep neural network architectures without any modifications. We evaluate the effectiveness of our algorithms through fine-tuning experiments on large language models and diffusion models. Experimental results consistently demonstrate that our algorithms provide superior performance compared to state-of-the-art methods. Additionally, our algorithm is not only effective for fine-tuning large models but is also applicable to deep neural network (DNN) compression.



Paperid:4175
Authors:Gen Li, Changxiao Cai
Abstract:
Abstract:Diffusion models have emerged as a powerful paradigm for modern generative modeling, demonstrating strong potential for large language models (LLMs). Unlike conventional autoregressive (AR) models that generate tokens sequentially, diffusion models allow for parallel sampling of tokens, leading to faster sampling and eliminating the left-to-right generation constraints. Despite their empirical success, theoretical understandings of diffusion model approaches remain underdeveloped. In this work, we develop convergence guarantees for diffusion language models from an information-theoretic perspective. Our analysis demonstrates that the sampling error, measured by the Kullback-Leibler (KL) divergence, decays inversely with the number of iterations $T$ and scales linearly with the mutual information between tokens in the target text sequence. In particular, we establish matching upper and lower bounds, up to some constant factor, that shows the tightness of our convergence analysis. These results offer novel theoretical insights into the practical effectiveness of diffusion language models.



Paperid:4176
Authors:Chaochen Gao, Xing W, Zijia Lin, Debing Zhang, Songlin Hu
Abstract:
High-quality long-context instruction data is essential for aligning long-context large language models (LLMs). Despite the public release of models like Qwen and Llama, their long-context instruction data remains proprietary. Human annotation is costly and challenging, while template-based synthesis methods limit scale, diversity, and quality. We introduce LongMagpie, a self-synthesis framework that automatically generates large-scale long-context instruction data. Our key insight is that aligned long-context LLMs, when presented with a document followed by special tokens preceding a user turn, auto-regressively generate contextually relevant queries. By harvesting these document-query pairs and the model's responses, LongMagpie produces high-quality instructions without human effort. Experiments on HELMET, RULER, and Longbench v2 demonstrate that LongMagpie achieves leading performance on long-context tasks while maintaining competitive performance on short-context tasks, establishing it as a simple and effective approach for open, diverse, and scalable long-context instruction data synthesis.



Paperid:4177
Authors:Jing Wen, Alex Schwing, Shenlong Wang
Abstract:
We tackle the task of recovering an animatable 3D human avatar from a single or a sparse set of images. For this task, beyond a set of images, many prior state-of-the-art methods use accurate “ground-truth” camera poses and human poses as input to guide reconstruction at test-time. We show that pose‑dependent reconstruction degrades results significantly if pose estimates are noisy.To overcome this, we introduce NoPo-Avatar, which reconstructs avatars solely from images, without any pose input. By removing the dependence of test-time reconstruction on human poses, NoPo-Avatar is not affected by noisy human pose estimates, making it more widely applicable. Experiments on challenging THuman2.0, XHuman, and HuGe100K data show that NoPo-Avatar outperforms existing baselines in practical settings (without ground‑truth poses) and delivers comparable results in lab settings (with ground‑truth poses).



Paperid:4178
Authors:Xiaoqian Shen, Wenxuan Zhang, Jun Chen, Mohamed Elhoseiny
Abstract:
Abstract:Understanding and reasoning over long videos pose significant challenges for large video language models (LVLMs) due to the difficulty in processing intensive video tokens beyond context window and retaining long-term sequential information. Retrieval-Augmented Generation (RAG) has demonstrated effectiveness in processing long context for Large Language Models (LLMs); however, applying RAG to long video faces challenges such as disrupted temporal dependencies and inclusion of irrelevant information that can hinder accurate reasoning. To address these limitations, we propose Vgent, a novel \textbf{graph-based retrieval-reasoning-augmented generation framework} to enhance LVLMs for long video understanding. Our approach introduces two key innovations: (i) It represents videos by structured graphs with semantic relationships across video clips preserved to improve retrieval effectiveness. (ii) It introduces an intermediate reasoning step to mitigate the reasoning limitation of LVLMs, which leverages structured verification to reduce retrieval noise and facilitate the explicit aggregation of relevant information across clips, resulting in more accurate and context-aware responses.We comprehensively evaluate our framework with various open-source LVLMs on three long-video understanding benchmarks. Our approach yielded an overall performance improvement of $3.0\%\sim 5.4\%$ over base models on MLVU, and outperformed state-of-the-art video RAG methods by $8.6\%$. Our code will be made publicly available\footnote{Please refer to the \href{https://anonymous.4open.science/r/Vgent-83E7}{ anonymous GitHub link} for access to the code.}.



Paperid:4179
Authors:Shengju Yu, Pei Zhang, Siwei Wang, Suyuan Liu, Xinhang Wan, Zhibin Dong, Tiejun Li, Xinwang Liu
Abstract:
Although receiving notable improvements, current multi-view clustering (MVC) techniques generally rely on feature library mechanisms to propagate accumulated knowledge from historical views to newly-arrived data, which overlooks the information pertaining to basis embedding within each view. Moreover, the mapping paradigm inevitably alters the values of learned landmarks and built affinities due to the uninterruption nature, accordingly disarraying the hierarchical cluster structures. To mitigate these two issues, we in the paper provide a named BSTM algorithm. Concretely, we firstly synchronize with the distinct dimensions by introducing a group of specialized projectors, and then establish unified anchors for all views collected so far to capture intrinsic patterns. Afterwards, departing from per-view architectures, we devise a shared bipartite graph construction via indicators to quantify similarity, which not only avoids redundant data-recalculations but alleviates the representation distortion caused by fusion. Crucially, there two components are optimized within an integrated framework, and collectively facilitate knowledge transfer upon encountering incoming views. Subsequently, to flexibly do transformation on anchors and meanwhile maintain numerical consistency, we develop a bit-swapping scheme operating exclusively on 0 and 1. It harmonizes anchors on current view and that on previous views through one-hot encoded row and column attributes, and the graph structures are correspondingly reordered to reach a matched configuration. Furthermore, a computationally efficient four-step updating strategy with linear complexity is designed to minimize the associated loss. Extensive experiments organized on publicly-available benchmark datasets with varying missing percentages confirm the superior effectiveness of our BSTM.



Paperid:4180
Authors:Mircea Lică, Ojas Shirekar, Baptiste Colle, Chirag Raman
Abstract:
Abstract:Embodied agents powered by large language models (LLMs), such as Voyager, promise open-ended competence in worlds such as Minecraft. However, when powered by open-weight LLMs they still falter on elementary tasks after domain-specific fine-tuning. We propose MindForge, a generative-agent framework for cultural lifelong learning through explicit perspective taking. We introduce three key innovations: (1) a structured theory of mind representation linking percepts, beliefs, desires, and actions; (2) natural inter-agent communication; and (3) a multi-component memory system. In an instructive setting within Minecraft, MindForge agents powered by open-weight LLMs significantly outperform their Voyager counterparts in basic tasks yielding $3\times$ more tech-tree milestones and collecting $2.3\times$ more unique items than the Voyager baseline without GPT-4. Furthermore, In fully collaborative settings, two agents that each fail in isolation succeed together, showing performance scales with the number of communication rounds, echoing the Condorcet Jury Theorem. MindForge agents demonstrate sophisticated behaviors, including expert-novice knowledge transfer, collaborative problem solving, and adaptation to out-of-distribution tasks through accumulated cultural experiences.



Paperid:4181
Authors:Liangjian Wen, Qun Dai, Yong Dai, Jianzhuang Liu, Jiangtao Zheng, Dongkai Wang, Zhao Kang, Jun Wang, Zenglin Xu, Jiang Duan
Abstract:
In multimodal representation learning, synergistic interactions between modalities not only provide complementary information but also create unique outcomes through specific interaction patterns that no single modality could achieve alone. Existing methods may struggle to effectively capture the full spectrum of synergistic information, leading to suboptimal performance in tasks where such interactions are critical. This is particularly problematic because synergistic information constitutes the fundamental value proposition of multimodal representation. To address this challenge, we introduce InfMasking, a contrastive synergistic information extraction method designed to enhance synergistic information through an Infinite Masking strategy. InfMasking stochastically occludes most features from each modality during fusion, preserving only partial information to create representations with varied synergistic patterns. Unmasked fused representations are then aligned with masked ones through mutual information maximization to encode comprehensive synergistic information. This infinite masking strategy enables capturing richer interactions by exposing the model to diverse partial modality combinations during training. As computing mutual information estimates with infinite masking is computationally prohibitive, we derive an InfMasking loss to approximate this calculation. Through controlled experiments, we demonstrate that InfMasking effectively enhances synergistic information between modalities. In evaluations on large-scale real-world datasets, InfMasking achieves state-of-the-art performance across seven benchmarks.



Paperid:4182
Authors:Wentao Wang, Hang Ye, Fangzhou Hong, Xue Yang, Jianfu Zhang, Yizhou Wang, Ziwei Liu, Liang Pan
Abstract:
Given a single in-the-wild human photo, it remains a challenging task to reconstruct a high-fidelity 3D human model. Existing methods face difficulties including a) the varying body proportions captured by in-the-wild human images; b) diverse personal belongings within the shot; and c) ambiguities in human postures and inconsistency in human textures. In addition, the scarcity of high-quality human data intensifies the challenge. To address these problems, we propose a Generalizable image-to-3D huMAN reconstruction framework, dubbed GeneMAN, building upon a comprehensive multi-source collection of high-quality human data, including 3D scans, multi-view videos, single photos, and our generated synthetic human data. GeneMAN encompasses three key modules. 1) Without relying on parametric human models (e.g., SMPL), GeneMAN first trains a human-specific text-to-image diffusion model and a view-conditioned diffusion model, serving as GeneMAN 2D human prior and 3D human prior for reconstruction, respectively. 2) With the help of the pretrained human prior models, the Geometry Initialization-&-Sculpting pipeline is leveraged to recover high-quality 3D human geometry given a single image. 3) To achieve high-fidelity 3D human textures, GeneMAN employs the Multi-Space Texture Refinement pipeline, consecutively refining textures in the latent and the pixel spaces. Extensive experimental results demonstrate that GeneMAN could generate high-quality 3D human models from a single image input, outperforming prior state-of-the-art methods. Notably, GeneMAN could reveal much better generalizability in dealing with in-the-wild images, often yielding high-quality 3D human models in natural poses with common items, regardless of the body proportions in the input images.



Paperid:4183
Authors:Ruitao Wu, Yifan Zhao, Guangyao Chen, Jia Li
Abstract:
Few-Shot Class-Incremental Learning (FSCIL) challenges models to sequentially learn new classes from minimal examples without forgetting prior knowledge, a task complicated by the stability-plasticity dilemma and data scarcity. Current FSCIL methods often struggle with generalization due to their reliance on limited datasets. While diffusion models offer a path for data augmentation, their direct application can lead to semantic misalignment or ineffective guidance. This paper introduces Diffusion-Classifier Synergy (DCS), a novel framework that establishes a mutual boosting loop between diffusion model and FSCIL classifier. DCS utilizes a reward-aligned learning strategy, where a dynamic, multi-faceted reward function derived from the classifier's state directs the diffusion model. This reward system operates at two levels: the feature level ensures semantic coherence and diversity using prototype-anchored maximum mean discrepancy and dimension-wise variance matching, while the logits level promotes exploratory image generation and enhances inter-class discriminability through confidence recalibration and cross-session confusion-aware mechanisms. This co-evolutionary process, where generated images refine the classifier and an improved classifier state yields better reward signals, demonstrably achieves state-of-the-art performance on FSCIL benchmarks, significantly enhancing both knowledge retention and new class learning.



Paperid:4184
Authors:Wenyue Chen, Peng Li, Wangguandong Zheng, Chengfeng Zhao, Mengfei Li, Yaolong Zhu, Zhiyang Dou, Ronggang Wang, Yuan Liu
Abstract:
Photorealistic 3D full-body human reconstruction from a single image is a critical yet challenging task for applications in films and video games due to inherent ambiguities and severe self-occlusions. While recent approaches leverage SMPL estimation and SMPL-conditioned image diffusion models to hallucinate novel views, they suffer from inaccurate 3D priors estimated from SMPL meshes and have difficulty in handling difficult human poses and reconstructing fine details.In this paper, we propose SyncHuman, a novel framework that combines 2D multiview diffusion and 3D native diffusion for the first time, enabling high-quality clothed human mesh reconstruction from single-view images even under challenging human poses.Multiview diffusion excels at capturing fine 2D details but struggles with structural consistency, whereas 3D native diffusion generates coarse yet structurally consistent 3D shapes. By integrating the complementary strengths of these two approaches, we develop a more effective generation framework. Specifically, we first jointly fine-tune the multiview diffusion model and the 3D native diffusion model with proposed pixel-aligned 2D-3D synchronization attention to produce geometrically aligned 3D shapes and 2D multiview images. To further improve details, we introduce a feature injection mechanism that lifts fine details from 2D multiview images onto the aligned 3D shapes, enabling accurate and high-fidelity reconstruction.Extensive experiments demonstrate that SyncHuman achieves robust and photorealistic 3D human reconstruction, even for images with challenging poses. Our method outperforms baseline methods in geometric accuracy and visual fidelity, demonstrating a promising direction for future 3D generation models.



Paperid:4185
Authors:Neeraj Kumar, Chad Vanderbilt
Abstract:
Pathology foundation models (PFMs) have emerged as powerful tools for analyzing whole slide images (WSIs). However, adapting these pretrained PFMs for specific clinical tasks presents considerable challenges, primarily due to the availability of only weak (WSI-level) labels for gigapixel images, necessitating multiple instance learning (MIL) paradigm for effective WSI analysis. This paper proposes a novel approach for single-GPU \textbf{T}ask \textbf{A}daptation of \textbf{PFM}s (TAPFM) that uses vision transformer (\vit) attention for MIL aggregation while optimizing both for feature representations and attention weights. The proposed approach maintains separate computational graphs for MIL aggregator and the PFM to create stable training dynamics that align with downstream task objectives during end-to-end adaptation. Evaluated on mutation prediction tasks for bladder cancer and lung adenocarcinoma across institutional and TCGA cohorts, TAPFM consistently outperforms conventional approaches, with H-Optimus-0 (TAPFM) outperforming the benchmarks. TAPFM effectively handles multi-label classification of actionable mutations as well. Thus, TAPFM makes adaptation of powerful pre-trained PFMs practical on standard hardware for various clinical applications.



Paperid:4186
Authors:Yuda Song, Dhruv Rohatgi, Aarti Singh, J. Bagnell
Abstract:
Partial observability is a notorious challenge in reinforcement learning (RL), due to the need to learn complex, history-dependent policies. Recent empirical successes have usedprivileged expert distillation-- which leverages availability of latent state information during training (e.g., from a simulator) to learn and imitate the optimal latent, Markovian policy -- to disentangle the task of ''learning to see'' from ''learning to act''. While expert distillation is more computationally efficient than RL without latent state information, it also has well-documented failure modes. In this paper -- through a simple but instructive theoretical model called theperturbed Block MDP, and controlled experiments on challenging simulated locomotion tasks -- we investigate the algorithmic trade-off between privileged expert distillation and standard RL without privileged information. Our main findings are:(1)The trade-off empirically hinges on thestochasticityof the latent dynamics, as theoretically predicted by contrastingapproximate decodabilitywithbelief contractionin the perturbed Block MDP; and(2)The optimal latent policy is not always the best latent policy to distill. Our results suggest new guidelines for effectively exploiting privileged information, potentially advancing the efficiency of policy learning across many practical partially observable domains.



Paperid:4187
Authors:Wei Fu, Jiaxuan Gao, Xujie Shen, Chen Zhu, Zhiyu Mei, Chuyi He, Shusheng Xu, Guo Wei, Jun Mei, Jiashu Wang, Tongkai Yang, Binhang Yuan, YI WU
Abstract:
Reinforcement learning (RL) has become a key technique for fine-tuning large language reasoning models (LRMs), yet scaling RL training to support long sequences and massive models introduces significant system and algorithmic challenges. Existing frameworks tightly couple generation and training on the same hardware, limiting scalability and causing inefficiencies. In this paper, we present AReaL, a scalable asynchronous RL framework that decouples text generation and training across disjoint GPU sets. To address challenges unique to asynchronous pipelines—such as data staleness and rollout interruption—we propose a principled algorithm-system co-design. Our method introduces (1) Staleness Control to bound policy lag, (2) a Decoupled PPO Objective for stable learning under mild off-policy conditions, and (3) Interruptible Generation to reduce GPU idle time via chunked rollouts. Experiments on models up to 14B parameters across math and coding tasks demonstrate that AReaL outperforms or matches strong baselines in performance, while significantly improving training throughput and scaling efficiently across context lengths (up to 32K) and GPU clusters (up to 1024 GPUs). Our results establish a foundation for robust, high-throughput RL fine-tuning of next-generation LRMs.



Paperid:4188
Authors:Weizhi Fei, Xueyan Niu, XIE GUOQING, Yingqing Liu, Bo Bai, Wei Han
Abstract:
Although applications involving long-context inputs are crucial for the effective utilization of large language models (LLMs), they also result in increased computational costs and reduced performance. To address this challenge, we propose an efficient, training-free prompt compression method that retains key information within compressed prompts. We identify specific attention heads in transformer-based LLMs, which we designate as evaluator heads, that are capable of selecting tokens in long inputs that are most significant for inference. Building on this discovery, we develop EHPC, an Evaluator Head-based Prompt Compression method, which enables LLMs to rapidly "skim through'' input prompts by leveraging only the first few layers with evaluator heads during the pre-filling stage, subsequently passing only the important tokens to the model for inference. EHPC achieves state-of-the-art results across two mainstream benchmarks: prompt compression and long-context inference acceleration. Consequently, it effectively improves performance with the reduced costs associated with commercial API calls compared to prompt compressing methods. We further demonstrate that EHPC attains competitive results compared to key-value cache-based acceleration methods, thereby highlighting its potential to enhance the efficiency of LLMs for long-context tasks.



Paperid:4189
Authors:Jiale Ma, Wenzheng Pan, Yang Li, Junchi Yan
Abstract:
Combinatorial problems on graphs have attracted extensive efforts from the machine learning community over the past decade. Despite notable progress in this area under the umbrella of ML4CO, a comprehensive categorization, unified reproducibility, and transparent evaluation protocols are still lacking for the emerging and immense pool of neural CO solvers. In this paper, we establish a modular and streamlined framework benchmarking prevalent neural CO methods, dissecting their design choices via a tri-leveled "paradigm-model-learning'' taxonomy to better characterize different approaches. Further, we integrate their shared features and respective strengths to form 3 unified solvers representing global prediction (GP), local construction (LC), and adaptive expansion (AE) mannered neural solvers. We also collate a total of 34 datasets for 7 mainstream CO problems (including both edge-oriented tasks: TSP, ATSP, CVRP, as well as edge-oriented: MIS, MCut, MVC, MCut) across scales to facilitate more comparable results among literature. Extensive experiments upon our benchmark reveal a fair and exact performance exhibition indicative of the raw contribution of the learning components in each method, rethinking and insisting that pre- and post-inference heuristic tricks are not supposed to compensate for sub-par capability of the data-driven counterparts. Under this unified benchmark, an up-to-date replication of typical ML4CO methods is maintained, hoping to provide convenient reference and insightful guidelines for both engineering development and academic exploration of the ML4CO community in the future. Code is available at https://github.com/Thinklab-SJTU/ML4CO-Bench-101, and the dataset is at https://huggingface.co/datasets/ML4CO/ML4CO-Bench-101-SL.



Paperid:4190
Authors:Wenyuan Zhao, Adithya Balachandran, Chao Tian, Paul Pu Liang
Abstract:
The study of multimodality has garnered significant interest in fields where analyzing interactions among multiple information sources can enhance predictive modeling, data fusion, and interpretability. Partial information decomposition (PID) has emerged as a useful information-theoretic framework to quantify the degree to which individual modalities independently, redundantly, or synergistically convey information about a target variable. However, existing PID methods depend on optimizing over a joint distribution constrained by estimated pairwise probability distributions, which are costly and inaccurate for continuous and high-dimensional modalities. Our first key insight is that the problem can be solved efficiently when the pairwise distributions are multivariate Gaussians, and we refer to this problem as Gaussian PID (GPID). We propose a new gradient-based algorithm that substantially enhances computational efficiency for GPID based on an alternative formulation of the underlying optimization problem. To generalize the applicability to non-Gaussian data, we learn information-preserving encoders to transform random variables of arbitrary input distributions into pairwise Gaussian random variables. Along the way, we resolved an open problem regarding the optimality of joint Gaussian solutions for GPID. Empirical validation on diverse synthetic examples demonstrates that our proposed method provides more accurate and efficient PID estimates than existing baselines. We further evaluate on a series of large-scale multimodal benchmarks to show its utility in real-world applications of quantifying PID in multimodal datasets and selecting high-performing models.



Paperid:4191
Authors:Jayanta Mandi, Marianne Defresne, Senne Berden, Tias Guns
Abstract:
When some parameters of a constrained optimization problem (COP) are uncertain, this gives rise to a predict-then-optimize (PtO) problem, comprising two stages -- the prediction of the unknown parameters from contextual information and the subsequent optimization using those predicted parameters. Decision-focused learning (DFL) implements the first stage by training a machine learning (ML) model to optimize the quality of the decisions made using the predicted parameters. When parameters in the constraints of a COP are predicted, the predicted parameters can lead to infeasible solutions. Therefore, it is important to simultaneously manage both feasibility and decision quality. We develop a DFL framework for predicting constraint parameters in a generic COP. While prior works typically assume that the underlying optimization problem is a linear program (LP) or integer linear program (ILP), our approach makes no such assumption. We derive two novel loss functions based on maximum likelihood estimation (MLE): the first one penalizes infeasibility (by penalizing when the predicted parameters lead to infeasible solutions), and the second one penalizes suboptimal decisions (by penalizing when the true optimal solution is infeasible under the predicted parameters). We introduce a single tunable parameter to form a weighted average of the two losses, allowing decision-makers to balance suboptimality and feasibility. We experimentally demonstrate that adjusting this parameter provides a decision-maker the control over the trade-off between the two. Moreover, across several COP instances, we find that for a single value of the tunable parameter, our method matches the performance of the existing baselines on suboptimality and feasibility.



Paperid:4192
Authors:Wenxuan Bao, Ruxi Deng, Jingrui He
Abstract:
Pretrained vision-language models such as CLIP achieve strong zero-shot generalization but remain vulnerable to distribution shifts caused by input corruptions. In this work, we investigate how corruptions affect CLIP’s image embeddings and uncover a consistent phenomenon we term as embedding variance collapse, where both intra-class and inter-class variances shrink as corruption severity increases. We find that this collapse is closely tied to performance degradation, with inter-class variance strongly correlated with classification accuracy. To explain this phenomenon, we analyze how corruptions alter the structure of the embedding space. Our theoretical results suggest that the visual encoder tends to encode corruption-related signals, which dilute class-discriminative features and compress the representation geometry. We further show that maximizing inter-class variance, even when estimated from pseudo-labels, can provably enhance embedding quality. Based on this insight, we propose Mint, a simple test-time adaptation method that maximizes pseudo-label-based inter-class variance on the fly using cumulative prototypes and gradient estimates. Mint operates effectively with small batch sizes and consistently improves performance across multiple corruption benchmarks and CLIP architectures.



Paperid:4193
Authors:Jiaheng Dong, Hong Jia, Soumyajit Chatterjee, Abhirup Ghosh, James Bailey, Ting Dang
Abstract:
Abstract:Speech Foundation Models encounter significant performance degradation when deployed in real-world scenarios involving acoustic domain shifts, such as background noise and speaker accents. Test-time adaptation (TTA) has recently emerged as a viable strategy to address such domain shifts at inference time without requiring access to source data or labels. However, existing TTA approaches, particularly those relying on backpropagation, are memory-intensive, limiting their applicability in speech tasks and resource-constrained settings. Although backpropagation-free methods offer improved efficiency, existing ones exhibit poor accuracy. This is because they are predominantly developed for vision tasks, which fundamentally differ from speech task formulations, noise characteristics, and model architecture, posing unique transferability challenges.In this paper, we introduce E-BAT, first Efficient BAckpropagation-free TTA framework designed explicitly for speech foundation models. E-BAT achieves a balance between adaptation effectiveness and memory efficiency through three key components: (i) lightweight prompt adaptation for a forward-pass-based feature alignment, (ii) a multi-scale loss to capture both global (utterance-level) and local distribution shifts (token-level) and (iii) a test-time exponential moving average mechanism for stable adaptation across utterances. Experiments conducted on four noisy speech datasets spanning sixteen acoustic conditions demonstrate consistent improvements, with 4.1\%--13.5% accuracy gains over backpropogation-free baselines and 2.0$\times$–6.4$\times$ GPU memory savings compared to backpropogation-based methods. By enabling scalable and robust adaptation under acoustic variability, this work paves the way for developing more efficient adaptation approaches for practical speech processing systems in real-world environments. Our code will be publicly available upon acceptance.



Paperid:4194
Authors:Jiyang Xia, Fenghua Ling, Zhenhui Jessie Li, Junjie Yu, Hongliang Zhang, David Topping, LEI BAI, Zhonghua Zheng
Abstract:
Abstract:Urban warming differs markedly from regional background trends, highlighting the unique behavior of urban climates and the challenges they present. Accurately predicting local urban climate necessitates modeling the interactions between urban surfaces and atmospheric forcing. Although off-the-shelf machine learning (ML) algorithms offer considerable accuracy for climate prediction, they often function as black boxes, learning data mappings rather than capturing physical evolution. As a result, they struggle to capture key land-atmosphere interactions and may produce physically inconsistent predictions. To address these limitations, we propose UCformer, a novel multi-task, physics-guided Transformer architecture designed to emulate nonlinear urban climate processes. UCformer jointly estimates 2-m air temperature $\(T\)$, specific humidity $\(q\)$, and dew point temperature $\(t\)$ in urban areas, while embedding domain and physical priors into its learning structure. Experimental results demonstrate that incorporating domain and physical knowledge leads to significant improvements in emulation accuracy and generalizability under future urban climate scenarios. Further analysis reveals that learning shared correlations across cities enables the model to capture transferable urban surface–atmosphere interaction patterns, resulting in improved accuracy in urban climate emulation. Finally, UCformer shows strong potential to fit real-world data: when fine-tuned with limited observational data, it achieves competitive performance in estimating urban heat fluxes compared to a physics-based model.



Paperid:4195
Authors:Yinjie Min, Furong Xu, Xinyao Li, Changliang Zou, Yongdao Zhou
Abstract:
Active learning (AL) reduces annotation costs by selecting the most informative samples based on both model sensitivity and predictive uncertainty. While sensitivity can be measured through parameter gradients in an unsupervised manner, predictive uncertainty can hardly be estimated without true labels especially for regression tasks, reducing the informativeness of actively selected samples. This paper proposes the concept of \textit{auxiliary data} to aid the uncertainty estimation for regression tasks. With detailed theoretical analysis, we reveal that auxiliary data, despite potential distribution shifts, can provide a promising uncertainty surrogate when properly weighted. Such finding inspires our design of AGBAL, a novel AL framework that recalibrates auxiliary data losses through density ratio weighting to obtain reliable uncertainty estimates for sample selection. Extensive experiments show that AGBAL consistently outperforms existing approaches without auxiliary data across diverse synthetic and real-world datasets.



Paperid:4196
Authors:Guobin Shen, Dongcheng Zhao, Yiting Dong, Yang Li, Feifei Zhao, Yi Zeng
Abstract:
The evolution of the human brain has led to the development of complex synaptic plasticity, enabling dynamic adaptation to a constantly evolving world. This progress inspires our exploration into a new paradigm for Spiking Neural Networks (SNNs): a Plasticity-Driven Learning Framework (PDLF). This paradigm diverges from traditional neural network models that primarily focus on direct training of synaptic weights, leading to static connections that limit adaptability in dynamic environments. Instead, our approach delves into the heart of synaptic behavior, prioritizing the learning of plasticity rules themselves. This shift in focus from weight adjustment to mastering the intricacies of synaptic change offers a more flexible and dynamic pathway for neural networks to evolve and adapt. Our PDLF does not merely adapt existing concepts of functional and Presynaptic-Dependent Plasticity but redefines them, aligning closely with the dynamic and adaptive nature of biological learning. This reorientation enhances key cognitive abilities in artificial intelligence systems, such as working memory and multitasking capabilities, and demonstrates superior adaptability in complex, real-world scenarios. Moreover, our framework sheds light on the intricate relationships between various forms of plasticity and cognitive functions, thereby contributing to a deeper understanding of the brain's learning mechanisms. Integrating this groundbreaking plasticity-centric approach in SNNs marks a significant advancement in the fusion of neuroscience and artificial intelligence. It paves the way for developing AI systems that not only learn but also adapt in an ever-changing world, much like the human brain.



Paperid:4197
Authors:Xuehui Wang, Chongjie Si, Xue Yang, Yuzhi Zhao, Wenhai Wang, Xiaokang Yang, Wei Shen
Abstract:
Open-vocabulary semantic segmentation assigns every pixel a label drawn from an open-ended, text-defined space. Vision–language models such as CLIP excel at zero-shot recognition, yet their image-level pre-training hinders dense prediction. Current approaches either fine-tune CLIP—at high computational cost—or adopt training-free attention refinements that favor local smoothness while overlooking global semantics. In this paper, we present OPMapper, a lightweight, plug-and-play module that injects both local compactness and global connectivity into attention maps of CLIP. It combines Context-aware Attention Injection, which embeds spatial and semantic correlations, and Semantic Attention Alignment, which iteratively aligns the enriched weights with textual prompts. By jointly modeling token dependencies and leveraging textual guidance, OPMapper enhances visual understanding. OPMapper is highly flexible and can be seamlessly integrated into both training-based and training-free paradigms with minimal computational overhead. Extensive experiments demonstrate its effectiveness, yielding significant improvements across 8 open-vocabulary segmentation benchmarks.



Paperid:4198
Authors:Yunlong Lin, ZiXu Lin, Kunjie Lin, Jinbin Bai, Panwang Pan, Chenxin Li, Haoyu Chen, Zhongdao Wang, Xinghao Ding, Wenbo Li, Shuicheng Yan
Abstract:
Photo retouching has become integral to contemporary visual storytelling, enabling users to capture aesthetics and express creativity. While professional tools such as Adobe Lightroom offer powerful capabilities, they demand substantial expertise and manual effort. In contrast, existing AI-based solutions provide automation but often suffer from limited adjustability and poor generalization, failing to meet diverse and personalized editing needs. To bridge this gap, we introduce JarvisArt, a multi-modal large language model (MLLM)-driven agent that understands user intent, mimics the reasoning process of professional artists, and intelligently coordinates over 200 retouching tools within Lightroom. JarvisArt undergoes a two-stage training process: an initial Chain-of-Thought supervised fine-tuning to establish basic reasoning and tool-use skills, followed by Group Relative Policy Optimization for Retouching (GRPO-R) to further enhance its decision-making and tool proficiency. We also propose the Agent-to-Lightroom Protocol to facilitate seamless integration with Lightroom. To evaluate performance, we develop MMArt-Bench, a novel benchmark constructed from real-world user edits. JarvisArt demonstrates user-friendly interaction, superior generalization, and fine-grained control over both global and local adjustments, paving a new avenue for intelligent photo retouching. Notably, it outperforms GPT-4o with a 60\% improvement in average pixel-level metrics on MMArt-Bench for content fidelity, while maintaining comparable instruction-following capabilities.



Paperid:4199
Authors:Junsoo Oh, Jerry Song, Chulhee Yun
Abstract:
Weak-to-strong generalization refers to the phenomenon where a stronger model trained under supervision from a weaker one can outperform its teacher. While prior studies aim to explain this effect, most theoretical insights are limited to abstract frameworks or linear/random feature models. In this paper, we provide a formal analysis of weak-to-strong generalization from a linear CNN (weak) to a two-layer ReLU CNN (strong). We consider structured data composed of label-dependent signals of varying difficulty and label-independent noise, and analyze gradient descent dynamics when the strong model is trained on data labeled by the pretrained weak model. Our analysis identifies two regimes—data-scarce and data-abundant—based on the signal-to-noise characteristics of the dataset, and reveals distinct mechanisms of weak-to-strong generalization. In the data-scarce regime, generalization occurs via benign overfitting or fails via harmful overfitting, depending on the amount of data, and we characterize the transition boundary. In the data-abundant regime, generalization emerges in the early phase through label correction, but we observe that overtraining can subsequently degrade performance.



Paperid:4200
Authors:Mingfeng Fan, Jianan Zhou, Yifeng Zhang, Yaoxin Wu, Jinbiao Chen, Guillaume Sartoretti
Abstract:
Recent deep reinforcement learning methods have achieved remarkable success in solving multi-objective combinatorial optimization problems (MOCOPs) by decomposing them into multiple subproblems, each associated with a specific weight vector. However, these methods typically treat all subproblems equally and solve them using a single model, hindering the effective exploration of the solution space and thus leading to suboptimal performance. To overcome the limitation, we propose POCCO, a novel plug-and-play framework that enables adaptive selection of model structures for subproblems, which are subsequently optimized based on preference signals rather than explicit reward values. Specifically, we design a conditional computation block that routes subproblems to specialized neural architectures. Moreover, we propose a preference-driven optimization algorithm that learns pairwise preferences between winning and losing solutions. We evaluate the efficacy and versatility of POCCO by applying it to two state-of-the-art neural methods for MOCOPs. Experimental results across four classic MOCOP benchmarks demonstrate its significant superiority and strong generalization.



Paperid:4201
Authors:Jin Zhang, Ruiheng Zhang, Zhe Cao, Kaizheng Chen
Abstract:
Current camouflaged object detection methods predominantly follow discriminative segmentation paradigms and heavily rely on predefined categories present in the training data, limiting their generalization to unseen or emerging camouflage objects. This limitation is further compounded by the labor-intensive and time-consuming nature of collecting camouflage imagery. Although Large Vision-Language Models (LVLMs) show potential to improve such issues with their powerful generative capabilities, their understanding of camouflage scenes is still insufficient. To bridge this gap, we introduce MMCSBench, the first comprehensive multimodal benchmark designed to evaluate and advance LVLM capabilities in camouflage scenes. MMCSBench comprises 22,537 images and 76,843 corresponding image-text pairs across five fine-grained camouflage tasks. Additionally, we propose a new task, Camouflage Efficacy Assessment (CEA), aimed at quantitatively evaluating the camouflage effectiveness of objects in images and enabling automated collection of camouflage images from large-scale databases. Extensive experiments on 26 LVLMs reveal significant shortcomings in models' ability to perceive and interpret camouflage scenes. These findings highlight the fundamental differences between natural and camouflaged visual inputs, offering insights for future research in advancing LVLM capabilities within this challenging domain.



Paperid:4202
Authors:Andrew Lowy, Daogao Liu
Abstract:
Bilevel optimization, in which one optimization problem is nested inside another, underlies many machine learning applications with a hierarchical structure---such as meta-learning and hyperparameter optimization. Such applications often involve sensitive training data, raising pressing concerns about individual privacy. Motivated by this, we study differentially private bilevel optimization. We first focus on settings where the outer-level objective is convex, and provide novel upper and lower bounds on the excess risk for both pure and approximate differential privacy, covering both empirical and population-level loss. These bounds are nearly tight and essentially match the optimal rates for standard single-level differentially private ERM and stochastic convex optimization (SCO), up to additional terms that capture the intrinsic complexity of the nested bilevel structure. The bounds are achieved in polynomial time via efficient implementations of the exponential and regularized exponential mechanisms. A key technical contribution is a new method and analysis of log-concave sampling under inexact function evaluations, which may be of independent interest. In the non-convex setting, we develop novel algorithms with state-of-the-art rates for privately finding approximate stationary points. Notably, our bounds do not depend on the dimension of the inner problem.



Paperid:4203
Authors:Da Ma, Gonghu Shang, Zhi Chen, Libo Qin, Yijie LUO, Hongshen Xu, Lei Pan, Shuai Fan, Lu Chen, Kai Yu
Abstract:
Instruction tuning improves the ability of large language models (LLMs) to follow diverse human instructions, but achieving strong performance on specific target tasks remains challenging. A critical bottleneck is selecting the most relevant data to maximize task-specific performance. Existing data selection approaches include unstable influence-based methods and more stable distribution alignment methods, the latter of which critically rely on the underlying sample representation. In practice, most distribution alignment methods, from shallow features (e.g., BM25) to neural embeddings (e.g., BGE, LLM2Vec), may fail to capture how the model internally processes samples. To bridge this gap, we adopt a model-centric strategy in which each sample is represented by its neuronal activation pattern in the model, directly reflecting internal computation. However, directly using raw neuron activations leads to spurious similarity between unrelated samples due to neuron polysemanticity, where a single neuron may respond to multiple, unrelated concepts. To address this, we employ sparse autoencoders to disentangle polysemantic activations into sparse, monosemantic representations, and introduce a dedicated similarity metric for this space to better identify task-relevant data. Comprehensive experiments across multiple instruction datasets, models, tasks, and selection ratios show that our approach consistently outperforms existing data selection baselines in both stability and task-specific performance.



Paperid:4204
Authors:Christopher Chiu, Silviu Pitis, Mihaela van der Schaar
Abstract:
Abstract:Clinical reasoning in medicine is a hypothesis-driven process where physicians refine diagnoses from limited information through targeted history, physical examination, and diagnostic investigations. In contrast, current medical benchmarks for large language models (LLMs) primarily assess knowledge recall through single-turn questions, where complete clinical information is provided upfront. To address this gap, we introduce VivaBench, a multi-turn benchmark that evaluates sequential clinical reasoning in LLM agents. Our dataset consists of 1762 physician-curated clinical vignettes structured as interactive scenarios that simulate a $ \textit{viva voce}$ (oral) examination in medical training, requiring agents to actively probe for relevant findings, select appropriate investigations, and synthesize information across multiple steps to reach a diagnosis. While current LLMs demonstrate competence in diagnosing conditions from well-described clinical presentations, their performance degrades significantly when required to navigate iterative diagnostic reasoning under uncertainty in our evaluation. Our analysis identified several failure modes that mirror common cognitive errors in clinical practice, including: (1) fixation on initial hypotheses, (2) inappropriate investigation ordering, (3) premature diagnostic closure, and (4) failing to screen for critical conditions. These patterns reveal fundamental limitations in how current LLMs reason and make decisions under uncertainty. Through VivaBench, we provide a standardized benchmark for evaluating conversational medical AI systems for real-world clinical decision support. Beyond medical applications, we contribute to the larger corpus of research on agentic AI by demonstrating how sequential reasoning trajectories can diverge in complex decision-making environments.



Paperid:4205
Authors:Mónika Farsang, Radu Grosu
Abstract:
Abstract:We present LrcSSM, a $\textit{nonlinear}$ recurrent model that processes long sequences as fast as today’s linear state-space layers. By forcing the state-transition matrix to be diagonal and learned at every step, the full sequence can be solved in parallel with a single prefix-scan, giving $\mathcal{O}(TD)$ time and memory and only $\mathcal{O}(\log T)$ sequential depth, for input-sequence length $T$ and a state dimension $D$. Moreover, LrcSSM offers a formal gradient-stability guarantee that other input-varying systems such as Liquid-S4 and Mamba do not provide. Lastly, for network depth $L$, as the forward and backward passes cost $\Theta(T\,D\,L)$ FLOPs, with its low sequential depth and parameter count $\Theta(D\,L)$, the model follows the compute-optimal scaling law regime ($\beta \approx 0.42$) recently observed for Mamba, outperforming quadratic-attention Transformers at equal compute while avoiding the memory overhead of FFT-based long convolutions. We show that on a series of long-range forecasting tasks, LrcSSM outperforms LRU, S5 and Mamba.



Paperid:4206
Authors:Mingxuan Yan, Wang, Zechun Liu, Jiachen Li
Abstract:
To enable robots to achieve long-horizon tasks, recent hierarchical vision-language-action (VLAs) frameworks typically adopt vision-language model (VLM)-based planners to decompose complex manipulation tasks into simple sub-tasks that low-level visuomotor policies can easily handle. Typically, to finetune the VLM planner and let it learn to decompose the target task, a few human demonstrations are provided and will be segmented into sub-tasks by either human annotation or heuristic rules, which are less efficient, and the heuristic sub-tasks could largely deviate from the training data of visuomotor policy, which degrades the task performance. To address these issues, we propose a Retrieval-based Demonstration Decomposer (RDD) that automatically decomposes demonstrations into sub-tasks by aligning the visual features of decomposed sub-task intervals with the training data of low-level visuomotor policies to fully exploit its capability. Our method shows superior performance compared to the state-of-the-art sub-task decomposer on the RLBench benchmark and demonstrates robustness under various settings. Our code and demo videos are available in the supplementary materials.



Paperid:4207
Authors:Qiang Zhang, Fanrui Zhang, Jiawei Liu, Ming Hu, Junjun He, Zheng-Jun Zha
Abstract:
The dynamic nature of real-world information demands efficient knowledge editing in multimodal large language models (MLLMs) to ensure continuous knowledge updates. However, existing methods often struggle with precise matching in large-scale knowledge retrieval and lack multi-level guidance for coordinated editing, leading to less reliable outcomes. To tackle these challenges, we propose CARML, a novel retrieval-augmented editing framework that integrates conflict-aware dynamic retrieval with multi-level implicit and explicit guidance for reliable lifelong multimodal editing. Specifically, CARML introduces intra-modal uncertainty and inter-modal conflict quantification to dynamically integrate multi-channel retrieval results, so as to pinpoint the most relevant knowledge to the incoming edit samples. Afterwards, an edit scope classifier discerns whether the edit sample semantically aligns with the edit scope of the retrieved knowledge. If deemed in-scope, CARML refines the retrieved knowledge into information-rich continuous prompt prefixes, serving as the implicit knowledge guide. These prefixes not only include static knowledge prompt that capture key textual semantics but also incorporate token-level, context-aware dynamic prompt to explore fine-grained cross-modal associations between the edit sample and retrieved knowledge. To further enhance reliability, CARML incorporates a "hard correction" mechanism, leveraging explicit label knowledge to adjust the model’s output logits. Extensive experiments across multiple MLLMs and datasets indicate the superior performance of CARML in lifelong multimodal editing scenarios.



Paperid:4208
Authors:Xiaoling Zhou, Mingjie Zhang, Zhemg Lee, YUNCHENG HUA, chengli xing, Wei Ye, Flora Salim, Shikun Zhang
Abstract:
Large language models (LLMs) have achieved remarkable achievements across diverse applications; however, they remain plagued by spurious correlations and the generation of hallucinated content. Despite extensive efforts to enhance the resilience of LLMs, existing approaches either rely on indiscriminate fine-tuning of all parameters, resulting in parameter inefficiency and lack of specificity, or depend on post-processing techniques that offer limited adaptability and flexibility. This study introduces a novel Causality-driven Robust Optimization (CdRO) approach that selectively updates model components sensitive to causal reasoning, enhancing model causality while preserving valuable pretrained knowledge to mitigate overfitting. Our method begins by identifying the parameter components within LLMs that capture causal relationships, achieved through comparing the training dynamics of parameter matrices associated with the original samples, as well as augmented counterfactual and paraphrased variants. These comparisons are then fed into a lightweight logistic regression model, optimized in real time to dynamically identify and adapt the causal components within LLMs. The identified parameters are subsequently optimized using an enhanced policy optimization algorithm, where the reward function is designed to jointly promote both model generalization and robustness. Extensive experiments across various tasks using twelve different LLMs demonstrate the superior performance of our framework, underscoring its significant effectiveness in reducing the model’s dependence on spurious associations and mitigating hallucinations.



Paperid:4209
Authors:Minheng Ni, Zhengyuan Yang, Linjie Li, Chung-Ching Lin, Kevin Lin, Wangmeng Zuo, Lijuan Wang
Abstract:
Recent advances in large language models have significantly improved textual reasoning through the effective use of Chain-of-Thought (CoT) and reinforcement learning. However, extending these successes to vision-language tasks remains challenging due to inherent limitations in text-only CoT, such as visual hallucinations and insufficient multimodal integration. In this paper, we introduce Point-RFT, a multimodal reasoning framework explicitly designed to leverage visually grounded CoT reasoning for visual document understanding. Our approach consists of two stages: First, we conduct format finetuning using a curated dataset of 71K diverse visual reasoning problems, each annotated with detailed, step-by-step rationales explicitly grounded to corresponding visual elements. Second, we employ reinforcement finetuning targeting visual document understanding. On ChartQA, our approach improves accuracy from 70.88% (format-finetuned baseline) to 90.04%, surpassing the 83.92% accuracy achieved by reinforcement finetuning relying solely on text-based CoT. The result shows that our grounded CoT is more effective for multimodal reasoning compared with the text-only CoT. Moreover, Point-RFT exhibits superior generalization capability across several out-of-domain visual document reasoning benchmarks, including CharXiv, PlotQA, IconQA, TabMWP, etc., and highlights its potential in complex real-world scenarios.



Paperid:4210
Authors:Hainuo Wang, Qiming Hu, Xiaojie Guo
Abstract:
Restoring images degraded by adverse weather remains a significant challenge due to the highly non-uniform and spatially heterogeneous nature of weather-induced artifacts, \emph{e.g.}, fine-grained rain streaks versus widespread haze. Accurately estimating the underlying degradation can intuitively provide restoration models with more targeted and effective guidance, enabling adaptive processing strategies. To this end, we propose a Morton-Order Degradation Estimation Mechanism (MODEM) for adverse weather image restoration. Central to MODEM is the Morton-Order 2D-Selective-Scan Module (MOS2D), which integrates Morton-coded spatial ordering with selective state-space models to capture long-range dependencies while preserving local structural coherence. Complementing MOS2D, we introduce a Dual Degradation Estimation Module (DDEM) that disentangles and estimates both global and local degradation priors. These priors dynamically condition the MOS2D modules, facilitating adaptive and context-aware restoration. Extensive experiments and ablation studies demonstrate that MODEM achieves state-of-the-art results across multiple benchmarks and weather types, highlighting its effectiveness in modeling complex degradation dynamics. Our code will be released soon.



Paperid:4211
Authors:Jiaming Zhou, Ke Ye, Jiayi Liu, Teli Ma, Zifan Wang, Ronghe QIU, Kun-Yu Lin, Zhilin Zhao, Junwei Liang
Abstract:
Abstract:The generalization capabilities of vision-language-action (VLA) models to unseen tasks are crucial to achieving general-purpose robotic manipulation in open-world settings.However, the cross-task generalization capabilities of existing VLA models remain significantly underexplored.To address this gap, we introduce **AGNOSTOS**, a novel simulation benchmark designed to rigorously evaluate cross-task zero-shot generalization in manipulation. AGNOSTOS comprises 23 unseen manipulation tasks for test—distinct from common training task distributions—and incorporates two levels of generalization difficulty to assess robustness. Our systematic evaluation reveals that current VLA models, despite being trained on diverse datasets, struggle to generalize effectively to these unseen tasks. To overcome this limitation, we propose **Cross-Task In-Context Manipulation (X-ICM)**, a method that conditions large language models (LLMs) on in-context demonstrations from seen tasks to predict action sequences for unseen tasks.Additionally, we introduce a **dynamics-guided sample selection** strategy that identifies relevant demonstrations by capturing cross-task dynamics. On AGNOSTOS, X-ICM significantly improves cross-task zero-shot generalization performance over leading VLAs, achieving improvements of 6.0\% over $\pi_0$ and 7.9\% over VoxPoser.We believe AGNOSTOS and X-ICM will serve as valuable tools for advancing general-purpose robotic manipulation.



Paperid:4212
Authors:Zaiquan Yang, Yuhao LIU, Gerhard Hancke, Rynson Lau
Abstract:
Spatio-temporal video grounding (STVG) aims at localizing the spatio-temporal tube of a video, as specified by the input text query.In this paper, we utilize multimodal large language models (MLLMs) to explore a zero-shot solution in STVG.We reveal two key insights about MLLMs: (1) MLLMs tend to dynamically assign special tokens, referred to as \textit{grounding tokens}, for grounding the text query; and(2) MLLMs often suffer from suboptimal grounding due to the inability to fully integrate the cues in the text query (\textit{e.g.}, attributes, actions) for inference. Based on these insights, we propose a MLLM-based zero-shot framework for STVG, which includes novel decomposed spatio-temporal highlighting (DSTH) and temporal-augmented assembling (TAS) strategies to unleash the reasoning ability of MLLMs.The DSTH strategy first decouples the original query into attribute and action sub-queries for inquiring the existence of the target both spatially and temporally.It then uses a novel logit-guided re-attention (LRA) module to learn latent variables as spatial and temporal prompts, by regularizing token predictions for each sub-query.These prompts highlight attribute and action cues, respectively, directing the model's attention to reliable spatial and temporal related visual regions.In addition, as the spatial grounding by the attribute sub-query should be temporally consistent,we introduce the TAS strategy to assemble the predictions using the original video frames and the temporal-augmented frames as inputs to help improve temporal consistency.We evaluate our method on various MLLMs, and show that it outperforms SOTA methods on three common STVG benchmarks.



Paperid:4213
Authors:Xiong Zhang, Zhenli He, Changlong Fu, Cheng Xie
Abstract:
Generalist Graph Anomaly Detection (GGAD) extends traditional Graph Anomaly Detection (GAD) from one-for-one to one-for-all scenarios, posing significant challenges due to Feature Space Shift (FSS) and Graph Structure Shift (GSS). This paper first formalizes these challenges and proposes quantitative metrics to measure their severity. To tackle FSS, we develop an anomaly-driven graph invariant learning module that learns domain-invariant node representations. To address GSS, a novel structure-insensitive affinity learning module is introduced, capturing cross-domain structural correspondences via affinity-based features. Our unified framework, IA-GGAD, integrates these modules, enabling anomaly prediction on unseen graphs without target-domain retraining or fine-tuning. Extensive experiments on benchmark datasets from varied domains demonstrate IA-GGAD’s superior performance, significantly outperforming state-of-the-art methods (e.g., achieving up to +12.28\% AUROC over ARC on ACM). Ablation studies further confirm the effectiveness of each proposed module. The source code of IA-GGAD is available at \url{https://anonymous.4open.science/r/GGAD-8F0F/}.



Paperid:4214
Authors:Muhammad Adnan, Nithesh Kurella, Akhil Arunkumar, Prashant Nair
Abstract:
Diffusion Transformers (DiTs) achieve state-of-the-art results in text-to-image, text-to-video generation, and editing. However, their large model size and the quadratic cost of spatial-temporal attention over multiple denoising steps make video generation computationally expensive. Static caching mitigates this by reusing features across fixed steps but fails to adapt to generation dynamics, leading to suboptimal trade-offs between speed and quality.We propose Foresight, an adaptive layer-reuse technique that reduces computational redundancy across denoising steps while preserving baseline performance. Foresight dynamically identifies and reuses DiT block outputs for all layers across steps, adapting to generation parameters such as resolution and denoising schedules to optimize efficiency. Applied to OpenSora, Latte, and CogVideoX, Foresight achieves up to 1.63x end-to-end speedup, end-to-end speedup, while maintaining video quality.



Paperid:4215
Authors:Jan Sobotka, Luca Baroni, Ján Antolík
Abstract:
Decoding visual stimuli from neural population activity is crucial for understanding the brain and for applications in brain-machine interfaces. However, such biological data is often scarce, particularly in primates or humans, where high-throughput recording techniques, such as two-photon imaging, remain challenging or impossible to apply. This, in turn, poses a challenge for deep learning decoding techniques. To overcome this, we introduce MEIcoder, a biologically informed decoding method that leverages neuron-specific most exciting inputs (MEIs), a structural similarity index measure loss, and adversarial training. MEIcoder achieves state-of-the-art performance in reconstructing visual stimuli from single-cell activity in primary visual cortex (V1), especially excelling on small datasets with fewer recorded neurons. Using ablation studies, we demonstrate that MEIs are the main drivers of the performance, and in scaling experiments, we show that MEIcoder can reconstruct high-fidelity natural-looking images from as few as 1,000-2,500 neurons and less than 1,000 training data points. We also propose a unified benchmark with over 160,000 samples to foster future research. Our results demonstrate the feasibility of reliable decoding in early visual system and provide practical insights for neuroscience and neuroengineering applications.



Paperid:4216
Authors:Shiwei Li, Xiandi Luo, Haozhao Wang, Xing Tang, Ziqiang Cui, Dugang Liu, Yuhua Li, Xiuqiang He, Ruixuan Li
Abstract:
Abstract:Low-rank adaptation (LoRA) is a parameter-efficient fine-tuning (PEFT) method widely used in large language models (LLMs). LoRA essentially describes the projection of an input space into a low-dimensional output space, with the dimensionality determined by the LoRA rank.In standard LoRA, all input tokens share the same weights and undergo an identical input-output projection.This limits LoRA's ability to capture token-specific information due to the inherent semantic differences among tokens.To address this limitation, we propose **Token-wise Projected Low-Rank Adaptation (TopLoRA)**, which dynamically adjusts LoRA weights according to the input token, thereby learning token-wise input-output projections in an end-to-end manner.Formally, the weights of TopLoRA can be expressed as $B\Sigma_X A$, where $A$ and $B$ are low-rank matrices (as in standard LoRA), and $\Sigma_X$ is a diagonal matrix generated from each input token $X$.Notably, TopLoRA does not increase the rank of LoRA weights but achieves more granular adaptation by learning token-wise LoRA weights (i.e., token-wise input-output projections).Extensive experiments across multiple models and datasets demonstrate that TopLoRA consistently outperforms LoRA and its variants. The code is available at the anonymous repository https://anonymous.4open.science/r/TopLoRA.



Paperid:4217
Authors:Weifeng Lin, Xinyu Wei, Ruichuan An, Tianhe Ren, Tingwei Chen, Renrui Zhang, Ziyu Guo, Wentao Zhang, Lei Zhang, Hongsheng Li
Abstract:
Abstract:While powerful vision foundation models like Segment Anything Model 2 (SAM 2) excel at object segmentation in images and videos, how to achieve deep semantic understanding of these regions remains a critical challenge. To address this, we introduce Perceive Anything Model (PAM), an end-to-end vision-language model that efficiently integrates SAM 2 and Large Language Models (LLMs), enabling the simultaneous segmentation of objects while generating diverse semantic outputs for each region in both images and videos. Specifically, PAM introduces a Semantic Perceiver that acts as a crucial bridge. This component efficiently utilizes rich intermediate features from the SAM 2 backbone, thereby incorporating general vision, localization, and semantic priors into the visual tokens, which are subsequently fed into LLMs for understanding. To ensure PAM's robustness in understanding multi-dimensional semantic granularity, we develop a dedicated data augmentation and refinement pipeline, which yields 1.8M high-quality image data and 0.6M video data. Experimental results demonstrate that, even with a lightweight 1.5/3B LLM as the semantic decoder, PAM achieves strong performance across a diverse range of tasks, including category prediction, brief and detailed regional captioning, video captioning, and streaming region captioning. Furthermore, PAM exhibits significant inference efficiency, running 1.2$-$2.4$\times$ faster while consuming less GPU memory compared to prior approaches, marking a significant advancement for real-world applications.



Paperid:4218
Authors:Michael Plainer, Hao Wu, Leon Klein, Stephan Günnemann, Frank Noe
Abstract:
Diffusion models have recently gained significant attention due to their effectiveness in various scientific domains, including biochemistry. When trained on equilibrium molecular distributions, diffusion models provide both: a generative procedure to sample equilibrium conformations and associated forces derived from the model's scores. However, using the forces for coarse-grained molecular dynamics simulations uncovers inconsistencies in the samples generated via classical diffusion inference and simulation, despite both originating from the same model. Particularly at the small diffusion timesteps required for simulations, diffusion models fail to satisfy the Fokker-Planck equation, which governs how the score should evolve over time. We interpret this deviation as an indication of the observed inconsistencies and propose an energy-based diffusion model with a Fokker–Planck-derived regularization term enforcing consistency. We demonstrate the effectiveness of our approach on toy systems, alanine dipeptide, and introduce a state-of-the-art transferable Boltzmann emulator for dipeptides that supports simulation and demonstrates enhanced consistency and efficient sampling.



Paperid:4219
Authors:Zheng-An Chen, Tao Luo
Abstract:
Although transformer-based models have shown exceptional empirical performance, the fundamental principles governing their training dynamics are inadequately characterized beyond configuration-specific studies. Inspired by empirical evidence showing improved reasoning capabilities under small initialization scales in language models, we employ the gradient flow analytical framework established in \cite{zhou2022towards} to systematically investigate linearized Transformer training dynamics. Our theoretical analysis dissects the dynamics of attention modules into two distinct stages. In the first stage, asymmetric weight perturbations from random initialization sustain non-degenerate gradient dynamics in parameter matrices, facilitating systematic escape from small initialization regimes. Subsequently, these matrices undergo condensation, progressively aligning toward the target orientation. In the second stage, the previously static key-query matrices actively participate in training, driving the normalized matrices toward asymptotic rank collapse. This two-stage framework generalizes classical directional convergence results.



Paperid:4220
Authors:Keuntae Kim, Eunhye Jeong, Sehyeon Lee, Seohee Yoon, Yong Suk Choi
Abstract:
Recent research on the increasing of Large Language Model(LLM)'s reasoning ability has been very successful. LLMs trained with Reinforcement Learning (RL) for reasoning ability show strong performance in challenging tasks like math and coding, even at small model sizes. However, despite the remarkable increase in accuracy of task, the assessment of creativity of LLM's generations has been overlooked in the reasoning task, unlike in the writing task. The main reason why creativity assessment research has not been actively conducted in the reasoning task is that, firstly, it was difficult to define the ‘range of creativity’ and secondly, human evaluation was essential in the process of assessment creativity. To overcome these difficulties, we proposed CLAWS, a novel method that can classify mathematical solutions into ‘Typical, Creative, and Hallucinated’ solutions without human evaluation by using Attention weight by prompt section. CLAWS showed superior performance in five 7-8B math RL models (Deepseek, QWEN, Mathstral, Openmath, OREAL) than five existing white-box detection methods (Perplexity, Logit Entropy, Window Entropy, Hidden Score, Attention Score). We validated this on 4545 math problems from 181 math contests (A(J)HSME, AMC, AIME). The code will be released on github after publication.



Paperid:4221
Authors:Takeshi Koshizuka, Issei Sato
Abstract:
Physics-informed machine learning is gaining significant traction for enhancing statistical performance and sample efficiency through the integration of physical knowledge. However, current theoretical analyses often presume complete prior knowledge in non-hybrid settings, overlooking the crucial integration of observational data, and are frequently limited to linear systems, unlike the prevalent nonlinear nature of many real-world applications.To address these limitations, we introduce a discrete weak form that unifies collocation and variational methods, enabling the incorporation of incomplete and complex physical constraints in hybrid learning settings.Within this formulation, we establish that the generalization performance of physics-informed regression in such hybrid settings is governed by the dimension of the affine variety associated with the physical constraint, rather than by the number of parameters. This enables a unified analysis that is applicable to both linear and nonlinear equations. We also present a method to approximate this dimension and provide experimental validation of our theoretical findings.



Paperid:4222
Authors:Haoyang Liu, Jie Wang, Yuyang Cai, Xiongwei Han, Yufei Kuang, Jianye Hao
Abstract:
Optimization modeling is one of the most crucial but technical parts of operations research (OR). To automate the modeling process, existing works have leveraged large language models (LLMs), prompting them to break down tasks into steps for generating variables, constraints, and objectives. However, due to the highly complex mathematical structures inherent in OR problems, standard fixed-step decomposition often fails to achieve high performance. To address this challenge, we introduce OptiTree, a novel tree search approach designed to enhance modeling capabilities for complex problems through adaptive problem decomposition into simpler subproblems. Specifically, we develop a modeling tree that organizes a wide range of OR problems based on their hierarchical problem taxonomy and complexity, with each node representing a problem category and containing relevant high-level modeling thoughts. Given a problem to model, we recurrently search the tree to identify a series of simpler subproblems and synthesize the global modeling thoughts by adaptively integrating the hierarchical thoughts. Experiments show that OptiTree significantly improves the modeling accuracy compared to the state-of-the-art, achieving over 10% improvements on the challenging benchmarks.



Paperid:4223
Authors:Risheng Liu, Zhu Liu, Weihao Mao, Wei Yao, Jin Zhang
Abstract:
Adversarial learning is a widely used paradigm in machine learning, often formulated as a min-max optimization problem where the inner maximization imposes adversarial constraints to guide the outer learner toward more robust solutions. This framework underlies methods such as Sharpness-Aware Minimization (SAM) and Generative Adversarial Networks (GANs). However, traditional gradient-based approaches to such problems often face challenges in balancing accuracy and efficiency due to second-order complexities. In this paper, we propose a bilevel optimization framework that reformulates these adversarial learning problems by leveraging the tractability of the lower-level problem. The bilevel framework introduces no additional complexity and enables the use of advanced bilevel tools. We further develop a provably convergent single-loop stochastic algorithm that effectively balances learning accuracy and computational cost. Extensive experiments show that our method improves generation quality in terms of FID and JS scores for GANs, and consistently achieves higher accuracy for SAM under label noise and across various backbones, while promoting flatter loss landscapes. Overall, this work provides a practical and theoretically grounded framework for solving adversarial learning tasks through bilevel optimization.



Paperid:4224
Authors:Tong Yang, Yu Huang, Yingbin Liang, Yuejie Chi
Abstract:
Transformers have demonstrated remarkable capabilities in multi-step reasoning tasks. However, understandings of the underlying mechanisms by which they acquire these abilities through training remain limited, particularly from a theoretical standpoint. This work investigates how transformers learn to solve symbolic multi-step reasoning problems through chain-of-thought processes, focusing on path-finding in trees. We analyze two intertwined tasks: a backward reasoning task, where the model outputs a path from a goal node to the root, and a more complex forward reasoning task, where the model implements two-stage reasoning by first identifying the goal-to-root path and then reversing it to produce the root-to-goal path. Our theoretical analysis, grounded in the dynamics of gradient descent, shows that trained one-layer transformers can provably solve both tasks with generalization guarantees to unseen trees. In particular, our multi-phase training dynamics for forward reasoning elucidate how different attention heads learn to specialize and coordinate autonomously to solve the two subtasks in a single autoregressive path. These results provide a mechanistic explanation of how trained transformers can implement sequential algorithmic procedures. Moreover, they offer insights into the emergence of reasoning abilities, suggesting that when tasks are structured to take intermediate chain-of-thought steps, even shallow multi-head transformers can effectively solve problems that would otherwise require deeper architectures.



Paperid:4225
Authors:Yinsicheng Jiang, Yao Fu, Yeqi Huang, Ping Nie, Zhan Lu, Leyang Xue, Congjie He, Man-Kit Sit, Jilong Xue, Li Dong, Ziming Miao, DaYou Du, Tairan Xu, Kai Zou, Edoardo Maria Ponti, Luo Mai
Abstract:
The sparse Mixture-of-Experts (MoE) architecture is increasingly favored for scaling Large Language Models (LLMs) efficiently, but it depends on heterogeneous compute and memory resources. These factors jointly affect system Cost, Accuracy, and Performance (CAP), making trade-offs inevitable. Existing benchmarks often fail to capture these trade-offs accurately, complicating practical deployment decisions. To address this, we introduce MoE-CAP, a benchmark specifically designed for MoE systems. Our analysis reveals that achieving an optimal balance across CAP is difficult with current hardware; MoE systems typically optimize two of the three dimensions at the expense of the third—a dynamic we term the MoE-CAP trade-off. To visualize this, we propose the CAP Radar Diagram. We further introduce sparsity-aware performance metrics—Sparse Memory Bandwidth Utilization (S-MBU) and Sparse Model FLOPS Utilization (S-MFU)—to enable accurate performance benchmarking of MoE systems across diverse hardware platforms and deployment scenarios. This benchmark is available on Github: https://github.com/sparse-generative-ai/open-moe-llm-leaderboard.



Paperid:4226
Authors:Avisek Naug, Antonio Guillen-Perez, Vineet Kumar, Scott Greenwood, Wesley Brewer, Sahand Ghorbanpour, Ashwin Ramesh Babu, Vineet Gundecha, Ricardo Luna Gutierrez, Soumyendu Sarkar
Abstract:
Liquid cooling is critical for thermal management in high-density data centers with the rising AI workloads. However, machine learning-based controllers are essential to unlock greater energy efficiency and reliability, promoting sustainability. We present SustainLC, a Sustainable Liquid Cooling benchmark environment, for reinforcement learning (RL) control strategies in energy-efficient liquid cooling of high-performance computing (HPC) systems. Built on the baseline of a high-fidelity digital twin of Oak Ridge National Lab's Frontier Supercomputer cooling system, SustainLC provides detailed Modelica-based end-to-end models spanning site-level cooling towers to data center cabinets and server blade groups. RL agents optimize critical thermal controls like liquid supply temperature, flow rate, and granular valve actuation at the IT cabinet level, as well as cooling tower setpoints through a Gymnasium interface, with dynamic changes in workloads. This environment creates a multi-objective real-time optimization challenge balancing local thermal regulation and global energy efficiency, and also supports additional components like heat recovery. We benchmark centralized and decentralized multi-agent RL approaches, demonstrate policy distillation for explainable control, and present RL-LLM hybrid and other LLM approaches that explain control actions in natural language to foster user trust. SustainLC democratizes access to detailed, customizable liquid cooling models, enabling the ML community, operators, and vendors to develop sustainable data center liquid cooling control solutions.



Paperid:4227
Authors:Nandan Thakur, Jimmy Lin, Samuel Havens, Michael Carbin, Omar Khattab, Andrew Drozdov
Abstract:
We introduce FreshStack, a holistic framework for automatically building information retrieval (IR) evaluation benchmarks by incorporating challenging questions and answers.FreshStack conducts the following steps:(1) automatic corpus collection from code and technical documentation,(2) nugget generation from community-asked questions and answers, and(3) nugget-level support, retrieving documents using a fusion of retrieval techniques and hybrid architectures.We use FreshStack to build five datasets on fast-growing, recent, and niche topics to ensure the tasks are sufficiently challenging. On FreshStack, existing retrieval models, when applied out-of-the-box, significantly underperform oracle approaches on all five topics, denoting plenty of headroom to improve IR quality. In addition, we identify cases where rerankers do not improve first-stage retrieval accuracy (two out of five topics), and oracle context helps an LLM generator generate a high-quality RAG answer.We hope FreshStack will facilitate future work toward constructing realistic, scalable, and uncontaminated IR and RAG evaluation benchmarks.



Paperid:4228
Authors:Xuyang Zhong, Yixiao Huang, Chen Liu
Abstract:
Abstract:This work studies fast adversarial training against sparse adversarial perturbations bounded by $l_0$ norm. We first demonstrate the unique challenges of employing $1$-step attacks on $l_0$ bounded perturbations, especially catastrophic overfitting (CO) that cannnot be properly addressed by existing fast adversarial training method for other $l_p$ norms ($p \geq 1$). We highlight that CO in $l_0$ adversarial training arises from sub-optimal perturbation locations of $1$-step attack. Some strategies like multi-$\epsilon$ can mitigate this sub-optimality to some extent, they lead to unstable training in turn. Theoretical and numerical analyses also reveal that the loss landscape of $l_0$ adversarial training is more craggy than its $l_\infty$, $l_2$ and $l_1$ counterparts, which exaggerates CO. To address this issue, we adopt soft labels and the trade-off loss function to smooth the adversarial loss landscape. Extensive experiments demonstrate our method can overcome the challenge of CO, achieve state-of-the-art performance, and narrow the performance gap between $1$-step and multi-step adversarial training against sparse attacks.



Paperid:4229
Authors:Andrew Nam, Henry Conklin, Yukang Yang, Tom Griffiths, Jonathan D Cohen, Sarah-Jane Leslie
Abstract:
We present causal head gating (CHG), a scalable method for interpreting the functional roles of attention heads in transformer models. CHG learns soft gates over heads and assigns them a causal taxonomy—facilitating, interfering, or irrelevant—based on their impact on task performance. Unlike prior approaches in mechanistic interpretability which are hypothesis-driven and require prompt templates or target labels, CHG applies directly to any dataset using standard next-token prediction. We evaluate CHG across multiple large language models (LLMs) in the Llama 3 model family and diverse tasks, including syntax, commonsense, and mathematical reasoning, and show that CHG scores indeed yield causal—not merely correlational—insight, validated via ablation and causal mediation analyses. We also introduce contrastive CHG, a variant that isolates sub-circuits for specific task components. Our findings reveal that LLMs contain multiple sparse, sufficient sub-circuits, that individual head roles depend on interactions with others (low modularity), and that instruction following and in-context learning rely on separable mechanisms.



Paperid:4230
Authors:Nina Vesseron, Louis Béthune, Marco Cuturi
Abstract:
Abstract:The canonical approach in generative modeling is to split model fitting into two blocks: define first how to sample noise (e.g. Gaussian) and choose next what to do with it (e.g. using a single map or flows). We explore in this work an alternative route that ties sampling and mapping. We find inspiration in moment measures, a result that states that for any measure $\rho$, there exists a unique convex potential $u$ such that $\rho=\nabla u \sharp e^{-u}$. While this does seem to tie effectively sampling (from log-concave distribution $e^{-u}$) and action (pushing particles through $\nabla u$), we observe on simple examples (e.g., Gaussians or 1D distributions) that this choice is ill-suited for practical tasks. We study an alternative factorization, where $\rho$ is factorized as $\nabla w^*\sharp e^{-w}$, where $w^*$ is the convex conjugate of a convex potential $w$. We call this approach conjugate moment measures, and show far more intuitive results on these examples. Because $\nabla w^*$ is the Monge map between the log-concave distribution $e^{-w}$ and $\rho$, we rely on optimal transport solvers to propose an algorithm to recover $w$ from samples of $\rho$, and parameterize $w$ as an input-convex neural network. We also address the common sampling scenario in which the density of $\rho$ is known only up to a normalizing constant, and propose an algorithm to learn $w$ in this setting.



Paperid:4231
Authors:Wenpu Li, Bangyan Liao, Yi Zhou, QiXu, Pian Wan, Peidong Liu
Abstract:
The estimation of optical flow and 6-DoF ego-motion—two fundamental tasks in 3-D vision—has typically been addressed independently. For neuromorphic vision (e.g., event cameras), however, the lack of robust data association makes solving the two problems separately an ill-posed challenge, especially in the absence of supervision via ground truth.Existing works mitigate this ill-posedness by either enforcing the smoothness of the flow field via an explicit variational regularizer or leveraging explicit structure-and-motion priors in the parametrization to improve event alignment.The former notably introduces bias in results and computational overhead, while the latter—which parametrizes the optical flow in terms of the scene depth and the camera motion—often converges to suboptimal local minima.To address these issues, we propose an unsupervised pipeline that jointly optimizes egomotion and flow via implicit spatial-temporal and geometric regularization. First, by modeling camera's egomotion as a continuous spline and optical flow as an implicit neural representation, our method inherently embeds spatial-temporal coherence through inductive biases. Second, we incorporate structure-and-motion priors through differential geometric constraints, bypassing explicit depth estimation while maintaining rigorous geometric consistency.As a result, our framework (called \textbf{E-MoFlow}) unifies egomotion and optical flow estimation via implicit regularization under a fully unsupervised paradigm. Experiments demonstrate its versatility to general 6-DoF motion scenarios, achieving state-of-the-art performance among unsupervised methods and competitive even with supervised approaches. Code will be released upon acceptance.



Paperid:4232
Authors:Ximing Li, Yuanchao Dai, Bing Wang, Changchun Li, Jianfeng Qu, Renchu Guan
Abstract:
Abstract:Positive and Unlabeled (PU) learning is a special case of binary classification with weak supervision, where only positive labeled and unlabeled data are available. Previous studies suggest several specific risk estimators of PU learning such as non-negative PU (nnPU), which are unbiased and consistent with the expected risk of supervised binary classification. In nnPU, the negative-class empirical risk is estimated by positive labeled and unlabeled data with a non-negativity constraint. However, its negative-class empirical risk estimator approaches 0, so the negative class is over-played, resulting in imbalanced error rates between positive and negative classes. To solve this problem, we suppose that the expected risks of the positive-class and negative-class should be close. Accordingly, we constrain that the negative-class empirical risk estimator is lower bounded by the positive-class empirical risk, instead of 0; and also incorporate an explicit equality constraint between them. we suggest a risk estimator of PU learning that balances positive and negative classification error rates, named $\mathrm{D{\small C-PU} }$, and suggest an efficient training method for $\mathrm{D{\small C-PU} }$ based on the augmented Lagrange multiplier framework. We theoretically analyze the estimation error of $\mathrm{D{\small C-PU} }$ and empirically validate that $\mathrm{D{\small C-PU} }$ achieves higher accuracy and converges more stable than other risk estimators of PU learning. Additionally, $\mathrm{D{\small C-PU} }$ also performs competitive accuracy performance with practical PU learning methods.



Paperid:4233
Authors:Kushagra Gupta, Surya Murthy, Mustafa Karabag, Ufuk Topcu, David Fridovich-Keil
Abstract:
Abstract:Cooperative bargaining games are widely used to model resource allocation and conflict resolution. Traditional solutions assume the mediator can access agents’ utility function values and gradients. However, there is an increasing number of settings, such as human-AI interactions, where utility values may be inaccessible or incomparable due to unknown, nonaffine transformations. To model such settings, we consider that the mediator has access only to agents' $\textit{most preferred directions}-$normalized utility gradients in the decision space. To this end, we propose a cooperative bargaining algorithm where a mediator has access to only the direction oracle of each agent. We prove that unlike popular approaches such as the Nash and Kalai-Smorodinsky bargaining solutions, our approach is invariant to monotonic nonaffine transformations, and that under strong convexity and smoothness assumptions, this approach enjoys global asymptotic convergence to Pareto stationary solutions. Moreover, we show that the bargaining solutions found by our algorithm also satisfy the axioms of symmetry and (under slightly stronger conditions) independence of irrelevant alternatives, which are popular in the literature. Finally, we conduct experiments in two domains, multi-agent formation assignment and mediated stock portfolio allocation, which validate these theoretical results.



Paperid:4234
Authors:Jianyuan Guo, Peike Li, Trevor Cohn
Abstract:
Sign Language Translation (SLT) aims to map sign language videos to spoken language text. A common approach relies on gloss annotations as an intermediate representation, decomposing SLT into two sub-tasks: video-to-gloss recognition and gloss-to-text translation. While effective, this paradigm depends on expert-annotated gloss labels, which are costly and rarely available in existing datasets, limiting its scalability. To address this challenge, we propose a gloss-free pseudo gloss generation framework that eliminates the need for human-annotated glosses while preserving the structured intermediate representation.Specifically, we prompt a Large Language Model (LLM) with a few example text-gloss pairs using in-context learning to produce draft sign glosses from spoken language text. To enhance the correspondence between LLM-generated pseudo glosses and the sign sequences in video, we correct the ordering in the pseudo glosses for better alignment via a weakly supervised learning process.This reordering facilitates the incorporation of auxiliary alignment objectives, and allows for the use of efficient supervision via a Connectionist Temporal Classification (CTC) loss.We train our SLT model—consisting of a vision encoder and a translator—through a three-stage pipeline, which progressively narrows the modality gap between sign language and spoken language.Despite its simplicity, our approach outperforms previous state-of-the-art gloss-free frameworks on two SLT benchmarks and achieves competitive results compared to gloss-based methods.



Paperid:4235
Authors:Gilles Bareilles, Peva Blanchard, Julien Fageot, Lê-Nguyên Hoang
Abstract:
Abstract:If you tell a learning model that you prefer an alternative $a$ over another alternative $b$, then you probably expect the model to be *monotone*, that is, the valuation of $a$ increases, and that of $b$ decreases. Yet, perhaps surprisingly, many widely deployed comparison-based preference learning models, including large language models, fail to have this guarantee. Until now, the only comparison-based preference learning algorithms that were proved to be monotone are the Generalized Bradley-Terry models. Yet, these models are unable to generalize to uncompared data. In this paper, we advance the understanding of the set of models with generalization ability that are *monotone*. Namely, we propose a new class of Linear Generalized Bradley-Terry models with Diffusion Priors, and identify sufficient conditions on alternatives' embeddings that guarantee monotonicity. Our experiments show that this monotonicity is far from being a general guarantee, and that our new class of generalizing models improves accuracy, especially when the dataset is limited.



Paperid:4236
Authors:Yaoyao Xu, Di Wang, Zihan Zhou, Tianshu Yu, Mingchen Chen
Abstract:
Understanding the dynamic behavior of proteins is critical to elucidating their functional mechanisms, yet generating realistic, temporally coherent trajectories of protein ensembles remains a significant challenge. In this work, we introduce a novel hierarchical autoregressive framework for modeling protein dynamics that leverages the intrinsic multi-scale organization of molecular motions. Unlike existing methods that focus on generating static conformational ensembles or treat dynamic sampling as an independent process, our approach characterizes protein dynamics as a Markovian process. The framework employs a two-scale architecture: a low-resolution model captures slow, collective motions driving major conformational transitions, while a high-resolution model generates detailed local fluctuations conditioned on these large-scale movements. This hierarchical design ensures that the causal dependencies inherent in protein dynamics are preserved, enabling the generation of temporally coherent and physically realistic trajectories. By bridging high-level biophysical principles with state-of-the-art generative modeling, our approach provides an efficient framework for simulating protein dynamics that balances computational efficiency with physical accuracy.



Paperid:4237
Authors:Xu Wan, Chao Yang, Cheng Yang, Jie Song, Mingyang Sun
Abstract:
Safe Reinforcement Learning (RL) is crucial for achieving high performance while ensuring safety in real-world applications. However, the complex interplay of multiple uncertainty sources in real environments poses significant challenges for interpretable risk assessment and robust decision-making. To address these challenges, we propose Fuz-RL, a novel fuzzy-guided robust framework for safe RL. Specifically, our framework: (1) introduces an adaptive neuro-fuzzy network that constructs interpretable fuzzy measures for different uncertainty levels, and (2) develops a novel fuzzy Bellman operator for estimating robust value functions using Choquet integrals. Theoretically, we prove that solving the Fuz-RL problem (in Constrained Markov Decision Process (CMDP) form) is equivalent to solving distributionally robust safe RL problems (in robust CMDP form), effectively avoiding min-max optimization. Empirical analyses on safe-control-gym scenarios demonstrate that Fuz-RL effectively integrates with existing safe RL baselines in a model-free manner, significantly improving both safety and control performance under various types of uncertainties in observation, action, and dynamics.



Paperid:4238
Authors:Cheng Luo, Jianghui Wang, Bing Li, Siyang Song, Bernard Ghanem
Abstract:
In this paper, we introduce Online Multimodal Conversational Response Generation (OMCRG), a novel task that aims to generate synchronized verbal and non-verbal listener feedback in real-time, conditioned on the speaker's multimodal input. OMCRG reflects natural dyadic interactions and poses new challenges in achieving synchronization between the generated audio and facial responses of the listener.To address these challenges, we innovatively introduce text as an intermediate modality to bridge the audio and facial responses. We hence propose OmniResponse, a Multimodal Large Language Model (MLLM) that autoregressively generates high-quality multi-modal listener responses. OmniResponse leverages a pretrained LLM enhanced with two novel components: Chrono-Text, which temporally anchors generated text tokens, and TempoVoice, a controllable online TTS module that produces speech synchronized with facial reactions. To support further OMCRG research, we present ResponseNet, a new dataset comprising 696 high-quality dyadic interactions featuring synchronized split-screen videos, multichannel audio, transcripts, and facial behavior annotations. Comprehensive evaluations conducted on ResponseNet demonstrate that OmniResponse significantly outperforms baseline models in terms of semantic speech content, audio-visual synchronization, and generation quality.Our dataset, code, and models will be made publicly available.



Paperid:4239
Authors:Chen Chen, Pengsheng Guo, Liangchen Song, Jiasen Lu, Rui Qian, Tsu-Jui Fu, Xinze Wang, Wei Liu, Yinfei Yang, Alex Schwing
Abstract:
Conditional generative modeling aims to learn a conditional data distribution from samples containing data-condition pairs. For this, diffusion and flow-based methods have attained compelling results. These methods use a learned (flow) model to transport an initial standard Gaussian noise that ignores the condition to the conditional data distribution. The model is hence required to learn both mass transport \emph{and} conditional injection. To ease the demand on the model, we propose \emph{Condition-Aware Reparameterization} (CAR)--a lightweight, learned \emph{shift} that conditions the source, the target, or both distributions. By relocating these distributions, CAR shortens the probability path the model must learn, leading to faster training in practice. On low-dimensional synthetic data, we visualize and quantify the effects of CAR. On higher-dimensional natural image data (ImageNet-256), we show that adding CAR to SiT-XL/2 reduces FID from 2.07 to 1.68, while introducing less than (0.6\%) additional parameters.



Paperid:4240
Authors:Minseok Jeong, Giup Seo, Euiseok Hwang
Abstract:
Abstract:Recently, physics-informed neural networks (PINNs) have gained attention in the scientific community for their potential to solve partial differential equations (PDEs). However, they face challenges related to resource efficiency and slow convergence. Adaptive sampling methods, which prioritize collocation points with high residuals, improve both efficiency and accuracy. However, these methods often neglect points with medium or low residuals, which can affect stability as the complexity of the model increases. In this paper, we investigate this limitation and show that high residual-based approaches require stricter learning rate bounds to ensure stability. To address this, we propose a Langevin dynamics-based Adaptive Sampling (LAS) framework that is robust to various learning rates and model complexities. Our experiments demonstrate that the proposed method outperforms existing approaches in terms of relative $L^{2}$ error, and stability across a range of enviroments, including high-dimensional PDEs where Monte Carlo integration-based methods typically suffer from instability.



Paperid:4241
Authors:Xihang Qiu, Jiarong Cheng, Yuhao Fang, Wanpeng Zhang, Yao Lu, Ye Zhang, Chun Li
Abstract:
Multimodal Emotion Recognition in Conversations (MERC) enhances emotional understanding through the fusion of multimodal signals. However, unpredictable modality absence in real-world scenarios significantly degrades the performance of existing methods. Conventional missing-modality recovery approaches, which depend on training with complete multimodal data, often suffer from semantic distortion under extreme data distributions, such as fixed-modality absence. To address this, we propose the Federated Dialogue-guided and Semantic-Consistent Diffusion (FedDISC) framework, pioneering the integration of federated learning into missing-modality recovery. By federated aggregation of modality-specific diffusion models trained on clients and broadcasting them to clients missing corresponding modalities, FedDISC overcomes single-client reliance on modality completeness. Additionally, the DISC-Diffusion module ensures consistency in context, speaker identity, and semantics between recovered and available modalities, using a Dialogue Graph Network to capture conversational dependencies and a Semantic Conditioning Network to enforce semantic alignment. We further introduce a novel Alternating Frozen Aggregation strategy, which cyclically freezes recovery and classifier modules to facilitate collaborative optimization. Extensive experiments on the IEMOCAP, CMUMOSI, and CMUMOSEI datasets demonstrate that FedDISC achieves superior emotion classification performance across diverse missing modality patterns, outperforming existing approaches.



Paperid:4242
Authors:Amir Ahsaei, Rickard Ewetz, Hao Zheng
Abstract:
Abstract:Graph Neural Networks (GNNs) success leverages the homophily principle, where connected nodes share similar features and labels. However, this assumption breaks down in heterophilic graphs, where same-class nodes are often distributed across distant neighborhoods rather than immediate connections. Recent attempts expand the receptive field through multi-hop aggregation schemes that explicitly preserve intermediate representations from each hop distance. While effective at capturing heterophilic patterns, these methods require separate weight matrices per hop and feature concatenation, causing parameters and memory to scale linearly with hop count, leading to high computational complexity and GPU memory consumption. We propose Gated Multi-hop Message Passing (GAMMA), where nodes assess how relevant the aggregated information is from their k-hop neighbors. This assessment occurs through multiple refinement steps where the node compares each hop's embedding with its current representation, allowing it to focus on the most informative hops. During the forward pass, GAMMA finds the optimal mix of multi-hop information local to each node using a single feature vector without needing separate representations for each hop, thereby maintaining dimensionality comparable to single hop GNNs. In addition, we propose a weight sharing scheme that leverages a unified transformation for aggregated features from multiple hops so the global heterophilic patterns specific to each hop are learned during training. As such, GAMMA captures both global (per-hop) and local (per-node) heterophily patterns without high computation and memory overhead. Experiments show GAMMA matches or exceeds state-of-the-art heterophilic GNN accuracy, achieving up to $\approx20\times$ faster inference.



Paperid:4243
Authors:Huanlin Gao, Ping Chen, Fuyuan Shi, Chao Tan, Fang Zhao, Zhaoxiang Liu, Kai Wang, Shiguo Lian
Abstract:
We present LeMiCa, a training-free and efficient acceleration framework for diffusion-based video generation. While existing caching strategies primarily focus on reducing local heuristic errors, they often overlook the accumulation of global errors, leading to noticeable content degradation between accelerated and original videos. To address this issue, we formulate cache scheduling as a directed graph with error-weighted edges and introduce a Lexicographic Minimax Path Optimization strategy that explicitly bounds the worst-case path error. This approach substantially improves the consistency of global content and style across generated frames. Extensive experiments on multiple text-to-video benchmarks demonstrate that LeMiCa delivers dual improvements in both inference speed and generation quality. Notably, our method achieves a 2.9× speedup on the Latte model and reaches an LPIPS score of 0.05 on Open-Sora, outperforming prior caching techniques. Importantly, these gains come with minimal perceptual quality degradation, making LeMiCa a robust and generalizable paradigm for accelerating diffusion-based video generation. We believe this approach can serve as a strong foundation for future research on efficient and reliable video synthesis.



Paperid:4244
Authors:Qinwei Yang, Jingyi Li, Peng Wu
Abstract:
Randomized controlled trials (RCTs) often exhibit limited inferential efficiency in estimating treatment effects due to small sample sizes. In recent years, the combination of external controls has gained increasing attention as a means of improving the efficiency of RCTs. However, external controls are not always comparable to RCTs, and direct borrowing without careful evaluation can introduce substantial bias and reduce the efficiency of treatment effect estimation. In this paper, we propose a novel influence-based adaptive sample borrowing approach that effectively quantifies the "comparability'' of each sample in the external controls using influence function theory. Given a selected set of borrowed external controls, we further derive a semiparametric efficient estimator under an exchangeability assumption. Recognizing that the exchangeability assumption may not hold for all possible borrowing sets, we conduct a detailed analysis of the asymptotic bias and variance of the proposed estimator under violations of exchangeability. Building on this bias-variance trade-off, we further develop a data-driven approach to select the optimal subset of external controls for borrowing. Extensive simulations and real-world applications demonstrate that the proposed approach significantly enhances treatment effect estimation efficiency in RCTs, outperforming existing approaches.



Paperid:4245
Authors:Yisong Xiao, Aishan Liu, Siyuan Liang, Zonghao Ying, Xianglong Liu, Dacheng Tao
Abstract:
Large Language Models (LLMs) have demonstrated impressive performance across various tasks, yet they remain vulnerable to generating toxic content, necessitating detoxification strategies to ensure safe and responsible deployment. Test-time detoxification methods, which typically introduce static or dynamic interventions into LLM representations, offer a promising solution due to their flexibility and minimal invasiveness. However, current approaches often suffer from imprecise interventions, primarily due to their insufficient exploration of the transition space between toxic and non-toxic outputs. To address this challenge, we propose \textsc{A}utoregressive \textsc{R}eward \textsc{G}uided \textsc{R}epresentation \textsc{E}diting (ARGRE), a novel test-time detoxification framework that explicitly models toxicity transitions within the latent representation space, enabling stable and precise reward-guided editing. ARGRE identifies non-toxic semantic directions and interpolates between toxic and non-toxic representations to reveal fine-grained transition trajectories. These trajectories transform sparse toxicity annotations into dense training signals, enabling the construction of an autoregressive reward model that delivers stable and precise editing guidance. At inference, the reward model guides an adaptive two-step editing process to obtain detoxified representations: it first performs directional steering based on expected reward gaps to shift representations toward non-toxic regions, followed by lightweight gradient-based refinements. Extensive experiments across 8 widely used LLMs show that ARGRE significantly outperforms leading baselines in effectiveness (-62.21\% toxicity) and efficiency (-47.58\% inference time), while preserving the core capabilities of the original model with minimal degradation. Our code is available at the \href{https://anonymous.4open.science/r/ARGRE-6291}{anonymous website}.



Paperid:4246
Authors:Ruichuan An, Sihan Yang, Renrui Zhang, zijun shen, Ming Lu, Gaole Dai, Hao Liang, Ziyu Guo, Shilin Yan, Yulin Luo, Bocheng Zou, Chaoqun Yang, Wentao Zhang
Abstract:
Abstract:Personalized models have demonstrated remarkable success in understanding and generating concepts provided by users. However, existing methods use separate concept tokens for understanding and generation, treating these tasks in isolation. This may result in limitations for generating images with complex prompts. For example, given the concept $\langle bo\rangle$, generating "$\langle bo\rangle$ wearing its hat" without additional textual descriptions of its hat. We call this kind of generation personalized knowledge-driven generation. To address the limitation, we present UniCTokens, a novel framework that effectively integrates personalized information into a unified vision language model (VLM) for understanding and generation. UniCTokens trains a set of unified concept tokens to leverage complementary semantics, boosting two personalized tasks. Moreover, we propose a progressive training strategy with three stages: understanding warm-up, bootstrapping generation from understanding, and deepening understanding from generation to enhance mutual benefits between both tasks. To quantitatively evaluate the unified VLM personalization, we present UnifyBench, the first benchmark for assessing concept understanding, concept generation, and knowledge-driven generation. Experimental results on UnifyBench indicate that UniCTokens shows competitive performance compared to leading methods in concept understanding, concept generation, and achieving state-of-the-art results in personalized knowledge-driven generation. Our research demonstrates that enhanced understanding improves generation, and the generation process can yield valuable insights into understanding. Our code and dataset will be released.



Paperid:4247
Authors:Kangkang Deng, Jiachen Jin, Jiang Hu, Hongxia Wang
Abstract:
Abstract:We study the problem of minimizing the sum of a smooth function and a nonsmooth convex regularizer over a compact Riemannian submanifold embedded in Euclidean space. By introducing an auxiliary splitting variable, we propose an adaptive Riemannian alternating direction method of multipliers (ARADMM), which, for the first time, achieves convergence without requiring smoothing of the nonsmooth term. In contrast to conventional Riemannian ADMM methods that require exactly solving a nested subproblem at each iteration, our approach involves only one Riemannian gradient evaluation and one proximal update per iteration. Through careful and adaptive coordination of the stepsizes and penalty parameters, we establish an optimal iteration complexity of order $\mathcal{O}(\epsilon^{-3})$ for finding an $\epsilon$-approximate KKT point, matching the complexity of existing smoothing technique-based Riemannian ADMM methods. Extensive numerical experiments on sparse PCA and robust subspace recovery demonstrate that our ARADMM consistently outperforms state-of-the-art Riemannian ADMM variants in convergence speed and solution quality.



Paperid:4248
Authors:Yifan Li, Yuhang Chen, Anh Dao, Lichi Li, Zhongyi Cai, Zhen Tan, Tianlong Chen, Yu Kong
Abstract:
Existing Embodied Question Answering (EQA) benchmarks primarily focus on household environments, often overlooking safety-critical aspects and reasoning processes pertinent to industrial settings. This drawback limits the evaluation of agent readiness for real-world industrial applications. To bridge this, we introduce IndustryEQA, the first benchmark dedicated to evaluating embodied agent capabilities within safety-critical industrial warehouse scenarios. Built upon the NVIDIA Isaac Sim platform, IndustryEQA provides high-fidelity episodic memory videos featuring diverse industrial assets, dynamic human agents, and carefully designed hazardous situations inspired by real-world safety guidelines. The benchmark includes rich annotations covering six categories: equipment safety, human safety, object recognition, attribute recognition, temporal understanding, and spatial understanding. Besides, it also provides extra reasoning evaluation based on these categories. Specifically, it comprises 971 question-answer pairs generated from small warehouse scenarios and 373 pairs from large ones, incorporating scenarios with and without human. We further propose a comprehensive evaluation framework, including various baseline models, to assess their general perception and reasoning abilities in industrial environments. IndustryEQA aims to steer EQA research towards developing more robust, safety-aware, and practically applicable embodied agents for complex industrial environments.



Paperid:4249
Authors:Xiaowen Cai, Daizong Liu, Xiaoye Qu, Xiang Fang, Jianfeng Dong, Keke Tang, Pan Zhou, Lichao Sun, Wei Hu
Abstract:
Although Large Vision-Language Models (LVLMs) exhibit impressive multimodal capabilities, their vulnerability to adversarial examples has raised serious security concerns.Existing LVLM attackers simply optimize adversarial images that easily overfit a certain model/prompt, making them ineffective once they are transferred to attack a different model/prompt. Motivated by this research gap, this paper aims to develop a more powerful attack that is transferable to black-box LVLM models of different structures and task-aware prompts of different semantics. Specifically, we introduce a new perspective of information theory to investigate LVLMs' transferable characteristics by exploring the relative dependence between outputs of the LVLM model and input adversarial samples. Our empirical observations suggest that enlarging/decreasing the mutual information between outputs and the disentangled adversarial/benign patterns of input images helps to generate more agnostic perturbations for misleading LVLMs' perception with better transferability. In particular, we formulate the complicated calculation of information gain as an estimation problem and incorporate such informative constraints into the adversarial learning process.Extensive experiments on various LVLM models/prompts demonstrate our significant transfer-attack performance.



Paperid:4250
Authors:Yizhou Xu, Florent Krzakala, Lenka Zdeborová
Abstract:
The Restricted Boltzmann Machine (RBM) is one of the simplest generative neural networks capable of learning input distributions. Despite its simplicity, the analysis of its performance in learning from the training data is only well understood in cases that essentially reduce to singular value decomposition of the data. Here, we consider the limit of a large dimension of the input space and a constant number of hidden units. In this limit, we simplify the standard RBM training objective into a form that is equivalent to the multi-index model with non-separable regularization. This opens a path to analyze training of the RBM using methods that are established for multi-index models, such as Approximate Message Passing (AMP) and its state evolution, and the analysis of Gradient Descent (GD) via the dynamical mean-field theory. We then give rigorous asymptotics of the training dynamics of RBM on data generated by the spiked covariance model as a prototype of a structure suitable for unsupervised learning. We show in particular that RBM reaches the optimal computational weak recovery threshold, aligning with the BBP transition, in the spiked covariance model.



Paperid:4251
Authors:Weifei Jin, Yuxin Cao, Junjie Su, Minhui Xue, Jie Hao, Ke Xu, Jin Song Dong, Derui Wang
Abstract:
Recent advances in Audio-Language Models (ALMs) have significantly improved multimodal understanding capabilities. However, the introduction of the audio modality also brings new and unique vulnerability vectors. Previous studies have proposed jailbreak attacks that specifically target ALMs, revealing that defenses directly transferred from traditional audio adversarial attacks or text-based LLM jailbreaks are largely ineffective against these ALM-specific threats.To address this issue, we propose ALMGuard, the first defense framework tailored to ALMs. Based on the assumption that safety-aligned shortcuts naturally exist in ALMs, we design a method to identify universal Shortcut Activation Perturbations (SAPs) that serve as triggers that activate the safety shortcuts to safeguard ALMs at inference time. To better sift out effective triggers while preserving the model’s availability on benign tasks, we further propose Mel-Gradient Sparse Mask (M-GSM), which restricts perturbations to Mel-frequency bins that are sensitive to jailbreaks but insensitive to speech understanding. Both theoretical analyses and empirical results demonstrate the robustness of our method against both seen and unseen attacks. ALMGuard reduces the average success rate of the most advanced ALM-specific jailbreak attacks to 4.6% across four models, establishing it as the new state-of-the-art in the field. Furthermore, evaluations on benign benchmarks confirm that our method does not cause a significant degradation in model availability.



Paperid:4252
Authors:Felipe Maia Polo, Seamus Somerstep, Leshem Choshen, Yuekai Sun, Mikhail Yurochkin
Abstract:
Scaling laws for large language models (LLMs) predict model performance based on parameters like size and training data. However, differences in training configurations and data processing across model families lead to significant variations in benchmark performance, making it difficult for a single scaling law to generalize across all LLMs. On the other hand, training family-specific scaling laws requires training models of varying sizes for every family. In this work, we propose Skills Scaling Laws (SSLaws, pronounced as \texttt{Sloth}), a novel scaling law that leverages publicly available benchmark data and assumes LLM performance is driven by low-dimensional latent skills, such as reasoning and instruction following. These latent skills are influenced by computational resources like model size and training tokens, but with varying efficiencies across model families. \texttt{Sloth} exploits correlations across benchmarks to provide more accurate and interpretable predictions while alleviating the need to train multiple LLMs per family. We present both theoretical results on parameter identification and empirical evaluations on 12 prominent benchmarks, from Open LLM Leaderboard v1/v2, demonstrating that \texttt{Sloth} predicts LLM performance accurately and offers insights into scaling behaviors for complex downstream tasks, increased test-time compute, and compute-optimal scaling of skills.



Paperid:4253
Authors:Jinsu Kim, Yunhun Nam, Minseon Kim, Sangpil Kim, Jongheon Jeong
Abstract:
Recent advances in text-to-image models have increased the exposure of powerful image editing techniques as a tool, raising concerns about their potential for malicious use. An emerging line of research to address such threats focuses on implanting “protective” adversarial noise into images before their public release, so future attempts to edit them using text-to-image models can be impeded. However, subsequent works have shown that these adversarial noises are often easily “reversed,” e.g., with techniques as simple as JPEG compression, casting doubt on the practicality of the approach. In this paper, we argue that adversarial noise for image protection should not only be imperceptible, as has been a primary focus of prior work, but also irreversible, viz., it should be difficult to detect as noise provided that the original image is hidden. We propose a surprisingly simple method to enhance the robustness of image protection methods against noise reversal techniques. Specifically, it applies an adaptive per-region Gaussian blur on the noise to adjust the overall frequency spectrum. Through extensive experiments, we show that our method significantly improves the per-sample worst-case protection performance of existing methods against a wide range of reversal techniques, while also reducing quality degradation due to noise in terms of perceptual metrics.



Paperid:4254
Authors:Lutz Oettershagen, Othon Michail
Abstract:
Abstract:Balancing resource efficiency and fairness is critical in networked systems that support modern learning applications. We introduce the _Fair Minimum Labeling_ (FML) problem: the task of designing a minimum-cost temporal edge activation plan that ensures each group of nodes in a network has sufficient access to a designated target set, according to specified coverage requirements. FML captures key trade-offs in systems where edge activations incur resource costs and equitable access is essential, such as distributed data collection, update dissemination in edge-cloud systems, and fair service restoration in critical infrastructure. We show that FML is NP-hard and $\Omega(\log |V|)$-hard to approximate, and we present probabilistic approximation algorithms that match this bound, achieving the best possible guarantee for the activation cost. We demonstrate the practical utility of FML in a fair multi-source data aggregation task for training a shared model. Empirical results show that FML enforces group-level fairness with substantially lower activation cost than baseline heuristics, underscoring its potential for building resource-efficient, equitable temporal reachability in learning-integrated networks.



Paperid:4255
Authors:Yuxuan Yang, Dalin Zhang, Yuxuan Liang, Hua Lu, Gang Chen, Huan Li
Abstract:
Time Series Forecasting (TSF) is a crucial task in various domains, yet existing TSF models rely heavily on high-quality data and insufficiently exploit all available data. This paper explores a novel self-supervised approach to re-label time series datasets by inherently constructing candidate datasets. During the optimization of a simple reconstruction network, intermediates are used as pseudo labels in a self-supervised paradigm, improving generalization for any predictor. We introduce the Self-Correction with Adaptive Mask (SCAM), which discards overfitted components and selectively replaces them with pseudo labels generated from reconstructions. Additionally, we incorporate Spectral Norm Regularization (SNR) to further suppress overfitting from a loss landscape perspective. Our experiments on eleven real-world datasets demonstrate that SCAM consistently improves the performance of various backbone models. This work offers a new perspective on constructing datasets and enhancing the generalization of TSF models through self-supervised learning. The code is available at https://anonymous.4open.science/r/SCAM-BDD3.



Paperid:4256
Authors:Yifeng Peng, Xinyi Li, Yen-Chi Chen, Kaining Zhang, Zhiding Liang, Ying Wang, Yuxuan Du
Abstract:
Abstract:Variational quantum Eigensolver (VQE) is a leading candidate for harnessing quantum computers to advance quantum chemistry and materials simulations, yet its training efficiency deteriorates rapidly for large Hamiltonians. Two issues underlie this bottleneck: (i) the no-cloning theorem imposes a linear growth in circuit evaluations with the number of parameters per gradient step; and (ii) deeper circuits encounter barren plateaus (BPs), leading to exponentially increasing measurement overheads. To address these challenges, here we propose a deep learning framework, dubbed Titan, which identifies and freezes inactive parameters of a given ansätze at initialization for a specific class of Hamiltonians, reducing the optimization overhead without sacrificing accuracy. The motivation of Titan starts with our empirical findings that a subset of parameters consistently has negligible influence on training dynamics. Its design combines a theoretically grounded data construction strategy, ensuring each training example is informative and BP-resilient, with an adaptive neural architecture that generalizes across ansätze of varying sizes. Across benchmark transverse-field Ising models, Heisenberg models, and multiple molecule systems up to $30$ qubits, Titan achieves up to $3\times$ faster convergence and $40$–$60\%$ fewer circuit evaluations than state-of-the-art baselines, while matching or surpassing their estimation accuracy. By proactively trimming parameter space, Titan lowers hardware demands and offers a scalable path toward utilizing VQE to advance practical quantum chemistry and materials science.



Paperid:4257
Authors:Xinghan Li, Haodong Wen, Kaifeng Lyu
Abstract:
Abstract:Despite the popularity of Adam optimizer in practice, most theoretical analyses study SGD as a proxy and little is known about how the solutions found by Adam differ. In this paper, we show that Adam reduces a specific form of sharpness measure shaped by its adaptive updates, leading to qualitatively different solutions from SGD. When the training loss is small, Adam wanders around the manifold of minimizers and takes semi-gradients to minimize this sharpness measure in an adaptive manner, a behavior we rigorously characterize via a continuous-time approximation using stochastic differential equations. We further illustrate how this behavior differs from that of SGD in a well-studied setting: when training overparameterized models with label noise, SGD has been shown to minimize the trace of the Hessian matrix, $\text{tr}(\textbf{H})$, whereas we prove that Adam minimizes $\text{tr}(\text{diag}(\textbf{H})^{1/2})$ instead. In solving sparse linear regression with diagonal linear networks, Adam provably achieves better sparsity and generalization than SGD due to this difference. Finally, we note that our proof framework applies not only to Adam but also to a broad class of adaptive gradient methods, including but not limited to RMSProp, Adam-mini, and Adalayer. This provides a unified perspective for analyzing how adaptive optimizers reduce sharpness and may offer insights for future optimizer design.



Paperid:4258
Authors:Yangxuan Zhou, Sha Zhao, Jiquan Wang, Haiteng Jiang, Shijian Li, Tao Li, Gang Pan
Abstract:
Human brain achieves dynamic stability-plasticity balance through synaptic homeostasis, a self-regulatory mechanism that stabilizes critical memory traces while preserving optimal learning capacities. Inspired by this biological principle, we propose SPICED: a neuromorphic framework that integrates the synaptic homeostasis mechanism for unsupervised continual EEG decoding, particularly addressing practical scenarios where new individuals with inter-individual variability emerge continually. SPICED comprises a novel synaptic network that enables dynamic expansion during continual adaptation through three bio-inspired neural mechanisms: (1) critical memory reactivation, which mimics brain functional specificity, selectively activates task-relevant memories to facilitate adaptation; (2) synaptic consolidation, which strengthens these reactivated critical memory traces and enhances their replay prioritizations for further adaptations and (3) synaptic renormalization, which are periodically triggered to weaken global memory traces to preserve learning capacities. The interplay within synaptic homeostasis dynamically strengthens task-discriminative memory traces and weakens detrimental memories. By integrating these mechanisms with continual learning system, SPICED preferentially replays task-discriminative memory traces that exhibit strong associations with newly emerging individuals, thereby achieving robust adaptations. Meanwhile, SPICED effectively mitigates catastrophic forgetting by suppressing the replay prioritization of detrimental memories during long-term continual learning. Validated on three EEG datasets, SPICED show its effectiveness. More importantly, SPICED bridges biological neural mechanisms and artificial intelligence through synaptic homeostasis, providing insights into the broader applicability of bio-inspired principles.



Paperid:4259
Authors:Shengkun Tang, Cong Zeng, Yuanzhou Chen, Zhiqiang Shen, Wenchao Yu, Xujiang Zhao, Haifeng Chen, Wei Cheng, Zhiqiang Xu
Abstract:
The rapid advancement of large language models (LLMs) such as ChatGPT, DeepSeek, and Claude has significantly increased the presence of AI-generated text in digital communication. This trend has heightened the need for reliable detection methods to distinguish between human-authored and machine-generated content. Existing approaches both zero-shot methods and supervised classifiers largely conceptualize this task as a binary classification problem, often leading to poor generalization across domains and models. In this paper, we argue that such a binary formulation fundamentally mischaracterizes the detection task by assuming a coherent representation of human-written texts. In reality, human texts do not constitute a unified distribution, and their diversity cannot be effectively captured through limited sampling. This causes previous classifiers to memorize observed OOD characteristics rather than learn the essence of `non-ID' behavior, limiting generalization to unseen human-authored inputs. Based on this observation, we propose reframing the detection task as an out-of-distribution (OOD) detection problem, treating human-written texts as distributional outliers while machine-generated texts are in-distribution (ID) samples. To this end, we develop a detection framework using one-class learning method including DeepSVDD and HRN, and score-based learning techniques such as energy-based method, enabling robust and generalizable performance. Extensive experiments across multiple datasets validate the effectiveness of our OOD-based approach. Specifically, the OOD-based method achieves 98.3\% AUROC and AUPR with only 8.9\% FPR95 on DeepFake dataset. Moreover, we test our detection framework on multilingual, attacked, and unseen-model and -domain text settings, demonstrating the robustness and generalizability of our framework. Code will be released openly and also available in the supplementary materials.



Paperid:4260
Authors:Nicolas Zucchet, Francesco D'Angelo, Andrew Lampinen, Stephanie Chan
Abstract:
Emergence is a fascinating property of large language models and neural networks more broadly: as models scale and train for longer, they sometimes develop new abilities in sudden ways. Despite initial studies, we still lack a comprehensive understanding of how and when these abilities emerge. To address this gap, we study the emergence over training of sparse attention, a critical and frequently observed attention pattern in Transformers. By combining theoretical analysis of a toy model with empirical observations on small Transformers trained on a linear regression variant, we uncover the mechanics driving sparse attention emergence and reveal that emergence timing follows power laws based on task structure, architecture, and optimizer choice. We additionally find that repetition can greatly speed up emergence. Finally, we confirm these results on a well-studied in-context associative recall task. Our findings provide a simple, theoretically grounded framework for understanding how data distributions and model design influence the learning dynamics behind one form of emergence.



Paperid:4261
Authors:Gokul Rajaraman, Debasish Chatterjee
Abstract:
Abstract:The problem of optimally covering a given compact subset of $\mathbb{R}^N$ with a preassigned number $n$ of Euclidean metric balls has a long-standing history and it is well-recognized to be computationally hard. This article establishes a numerically viable algorithm for obtaining optimal covers of compact sets via two key contributions. The first is a foundational result establishing Lipschitz continuity of the marginal function of a certain parametric non-convex maximization problem in the optimal covering problem, and it provides the substrate for numerical gradient algorithms to be employed in this context. The second is an adaptation of a stochastically smoothed numerical gradient-based (zeroth-order) algorithm for a non-convex minimization problem, that, equipped with randomized restarts, spurs global convergence to an optimal cover. Several numerical experiments with complicated nonconvex compact sets demonstrate the excellent performance of our techniques.



Paperid:4262
Authors:Yiyuan Pan, Zhe Liu, Hesheng Wang
Abstract:
Unlocking autonomous exploration in complex Multi-Agent Reinforcement Learning (MARL) systems, especially under sparse rewards, hinges on endowing agents with effective motivation. While artificial curiosity offers a powerful self-supervised drive, its efficacy is undermined when agents cannot robustly discern meaningful novelty from environmental stochasticity. Existing methods overlook the crucial need for agents to dynamically calibrate their intrinsic curiosity based on a learned understanding of their peers, leading to suboptimal performance in truly decentralized and communication-less settings. Inspired by how human children rapidly master new social games by observing peers and adaptively calibrating their own exploratory behaviors, we propose a novel approach to enhance multi-agent exploration. We introduce CERMIC, a principled framework that empowers agents to robustly filter noisy surprise signals and guide exploration by dynamically calibrating their intrinsic curiosity with inferred multi-agent contextual awareness. Additionally, CERMIC generates theoretically-grounded intrinsic rewards, encouraging agents to explore state transitions with high information gain. Finally, we evaluate CERMIC on benchmark suites including VMAS, MeltingPot, and SMACv2. Empirical results demonstrate that exploration with CERMIC significantly outperforms SoTA algorithms in sparse-reward environments.



Paperid:4263
Authors:Fares Mehouachi, Saif Eddin Jabari
Abstract:
We introduce FlowMixer, a neural architecture that leverages constrained matrix operations to model structured spatiotemporal patterns. At its core, FlowMixer incorporates non-negative matrix mixing layers within a reversible mapping framework—applying transforms before mixing and their inverses afterward. This shape-preserving design enables a Kronecker-Koopman eigenmode framework that bridges statistical learning with dynamical systems theory, providing interpretable spatiotemporal patterns and facilitating direct algebraic manipulation of prediction horizons without retraining. Extensive experiments across diverse domains demonstrate FlowMixer's robust long-horizon forecasting capabilities while effectively modeling physical phenomena such as chaotic attractors and turbulent flows. These results suggest that architectural constraints can simultaneously enhance predictive performance and mathematical interpretability in neural forecasting systems.



Paperid:4264
Authors:Yunpeng Jiang, Jianshu Hu, Paul Weng, Yutong Ban
Abstract:
Symmetry is pervasive in robotics and has been widely exploited to improve sample efficiency in deep reinforcement learning (DRL). However, existing approaches primarily focus on spatial symmetries—such as reflection, rotation, and translation—while largely neglecting temporal symmetries. To address this gap, we explore time reversal symmetry, a form of temporal symmetry commonly found in robotics tasks such as door opening and closing. We propose Time Reversal symmetry enhanced Deep Reinforcement Learning (TR-DRL), a framework that combines trajectory reversal augmentation and time reversal guided reward shaping to efficiently solve temporally symmetric tasks. Our method generates reversed transitions from fully reversible transitions, identified by a proposed dynamics-consistent filter, to augment the training data. For partially reversible transitions, we apply reward shaping to guide learning, according to successful trajectories from the reversed task. Extensive experiments on the Robosuite and MetaWorld benchmarks demonstrate that TR-DRL is effective in both single-task and multi-task settings, achieving higher sample efficiency and stronger final performance compared to baseline methods.



Paperid:4265
Authors:Wang Yang, Zirui Liu, Hongye Jin, Qingyu Yin, Vipin Chaudhary, Xiaotian Han
Abstract:
Recent language models exhibit strong reasoning capabilities, yet the influence of long-context capacity on reasoning remains underexplored. In this work, we hypothesize that current limitations in reasoning stem, in part, from insufficient long-context capacity, motivated by empirical observations such as \textit{i)} higher context window length often leads to stronger reasoning performance, and \textit{ii)} failed reasoning cases resemble failed long-context cases. To test this hypothesis, we examine whether enhancing a model’s long-context ability before Supervised Fine-Tuning (SFT) leads to improved reasoning performance. Specifically, we compared models with identical architectures and fine-tuning data but varying levels of long-context capacity. Our results reveal a consistent trend: models with stronger long-context capacity achieve significantly higher accuracy on reasoning benchmarks after SFT. Notably, these gains persist even on tasks with short input lengths, indicating that long-context training offers generalizable benefits for reasoning performance. These findings suggest that long-context modeling is not just essential for processing lengthy inputs, but also serves as a critical foundation for reasoning. We advocate for treating long-context capacity as a first-class objective in the design of future language models. Our code is anonymously available at \url{https://anonymous.4open.science/r/LCTMerge-64B8/}.



Paperid:4266
Authors:Gyudong Kim, Hyukju Na, Jin Hyeon Kim, Hyunsung Jang, Jaemin Park, Jaegi Hwang, NAMKOO HA, Seungryong Kim, Young Geun Kim
Abstract:
As training billion-scale transformers becomes increasingly common, employing multiple distributed GPUs along with parallel training methods has become a standard practice. However, existing transformer designs suffer from significant communication overhead, especially in Tensor Parallelism (TP), where each block’s MHA–MLP connection requires an all-reduce communication. Through our investigation, we show that the MHA-MLP connections can be bypassed for efficiency, while the attention output of the first layer can serve as an alternative signal for the bypassed connection. Motivated by the observations, we propose FAL (First Attentions Last), an efficient transformer architecture that redirects the first MHA output to the MLP inputs of the following layers, eliminating the per-block MHA-MLP connections. This removes the all-reduce communication and enables parallel execution of MHA and MLP on a single GPU. We also introduce FAL+, which adds the normalized first attention output to the MHA outputs of the following layers to augment the MLP input for the model quality. Our evaluation shows that FAL reduces multi-GPU training time by up to 44%, improves single-GPU throughput by up to 1.18×, and achieves better perplexity compared to the baseline GPT. FAL+ achieves even lower perplexity without increasing the training time than the baseline.



Paperid:4267
Authors:Ruiyang Lin, Yongyi Guo, Kyra Gan
Abstract:
The weighted controlled direct effect (WCDE) generalizes the standard controlled direct effect (CDE) by averaging over the mediator distribution, providing a robust estimate when treatment effects vary across mediator levels. This makes the WCDE especially relevant in fairness analysis, where it isolates the direct effect of an exposure on an outcome, independent of mediating pathways. In this work, we first establish necessary and sufficient conditions for the unique identifiability of the WCDE, clarifying when it diverges from the CDE. Next, we derive the efficient influence function for the WCDE and consider the class of regular and asymptotically linear estimators. We characterize the optimal covariate adjustment set that minimizes asymptotic variance, demonstrating how mediator-confounder interactions introduce distinct requirements compared to average treatment effect estimation. Our results offer a principled framework for efficient estimation of direct effects in complex causal systems, with practical applications in fairness and mediation analysis.



Paperid:4268
Authors:Chunru Lin, Haotian Yuan, Yian Wang, Xiaowen Qiu, Tsun-Hsuan Johnson Wang, Minghao Guo, Bohan Wang, Yashraj Narang, Dieter Fox, Chuang Gan
Abstract:
Endowing robots with tool design abilities is critical for enabling them to solve complex manipulation tasks that would otherwise be intractable. While recent generative frameworks can automatically synthesize task settings—such as 3D scenes and reward functions—they have not yet addressed the challenge of tool-use scenarios. Simply retrieving human-designed tools might not be ideal since many tools (e.g., a rolling pin) are difficult for robotic manipulators to handle. Furthermore, existing tool design approaches either rely on predefined templates with limited parameter tuning or apply generic 3D generation methods that are not optimized for tool creation.To address these limitations, we proposeRobotSmith, an automated pipeline that leverages the implicit physical knowledge embedded in vision-language models (VLMs) alongside the more accurate physics provided by physics simulations to design and use tools for robotic manipulation. Our system (1) iteratively proposes tool designs using collaborative VLM agents, (2) generates low-level robot trajectories for tool use, and (3) jointly optimizes tool geometry and usage for task performance.We evaluate our approach across a wide range of manipulation tasks involving rigid, deformable, and fluid objects. Experiments show that our method consistently outperforms strong baselines in both task success rate and overall performance. Notably, our approach achieves a 50.0\% average success rate, significantly surpassing other baselines such as 3D generation (21.4\%) and tool retrieval (11.1\%). Finally, we deploy our system in real-world settings, demonstrating that the generated tools and their usage plans transfer effectively to physical execution, validating the practicality and generalization capabilities of our approach.



Paperid:4269
Authors:Ramchalam Kinattinkara Ramakrishnan, Zhaocong Yuan, Jay Zhuo, Chen Feng, Yicheng Lin, Chenzheng Su, Parker Zhang
Abstract:
Speculative decoding generally dictates having a small, efficient draft model that is either pretrained or distilled offline to a particular target model series, for instance, Llama or Qwen models. However, within online deployment settings, there are two major challenges: 1) usage of a target model that is incompatible with the draft model; 2) expectation of latency improvements over usage and time. In this work, we propose OmniDraft, a unified framework that enables a single draft model to operate with any target model and adapt dynamically to user data. We introduce an online n-gram cache with hybrid distillation fine-tuning to address the cross-vocabulary mismatch across draft and target models; and further improve decoding speed by leveraging adaptive drafting techniques. OmniDraft is particularly suitable for on-device LLM applications where model cost, efficiency and user customization are the major points of contention. This further highlights the need to tackle the above challenges and motivates the “one drafter for all” paradigm. We showcase the proficiency of the OmniDraft framework by performing online learning on math reasoning, coding and text generation tasks. Notably, OmniDraft enables a single Llama-68M model to pair with various target models including Vicuna-7B, Qwen2-7B and Llama3-8B models for speculative decoding; and additionally provides up to 1.5-2x speedup.



Paperid:4270
Authors:Huimin Han, Shaolin Ji
Abstract:
Abstract:Many \textit{Deep Reinforcement Learning} (DRL) algorithms are sensitive to time discretization, which reduces their performance in real-world scenarios. We propose Continuous Soft Actor-Critic, an off-policy actor-critic DRL algorithm in continuous time and space. It is robust to environment time discretization. We also extend the framework to multi-agent scenarios. This \textit{Multi-Agent Reinforcement Learning} (MARL) algorithm is suitable for both competitive and cooperative settings. Policy evaluation employs stochastic control theory, with loss functions derived from martingale orthogonality conditions. We establish scaling principles for hyperparameters of the algorithm as the environment time discretization $\delta t$ changes ($\delta t \rightarrow 0$). We provide theoretical proofs for the relevant theorems. To validate the algorithm's effectiveness, we conduct comparative experiments between the proposed algorithm and other mainstream methods across multiple tasks in \textit{Virtual Multi-Agent System} (VMAS). Experimental results demonstrate that the proposed algorithm achieves robust performance across various environments with different time discretization parameter settings, outperforming other methods.



Paperid:4271
Authors:Bryan Sangwoo Kim, Jeongsol Kim, Jong Chul Ye
Abstract:
Abstract:Modern single-image super-resolution (SISR) models deliver photo-realistic results at the scale factors on which they are trained, but collapse when asked to magnify far beyond that regime. We address this scalability bottleneck with Chain-of-Zoom (CoZ), a model-agnostic framework that factorizes SISR into an autoregressive chain of intermediate scale-states with multi-scale-aware prompts. CoZ repeatedly re-uses a backbone SR model, decomposing the conditional probability into tractable sub-problems to achieve extreme resolutions without additional training. Because visual cues diminish at high magnifications, we augment each zoom step with multi-scale-aware text prompts generated by a vision-language model (VLM). The prompt extractor itself is fine-tuned using Generalized Reward Policy Optimization (GRPO) with a critic VLM, aligning text guidance towards human preference. Experiments show that a standard $4\times$ diffusion SR model wrapped in CoZ attains beyond $256\times$ enlargement with high perceptual quality and fidelity.



Paperid:4272
Authors:Jiacheng Xie, Yang Yu, Ziyang Zhang, Shuai Zeng, Jiaxuan He, Ayush Vasireddy, Xiaoting tang, Congyu Guo, Lening Zhao, Congcong Jing, Guanghui An, Dong Xu
Abstract:
Traditional Chinese Medicine (TCM), as an effective alternative medicine, has been receiving increasing attention. In recent years, the rapid development of large language models (LLMs) tailored for TCM has underscored the need for an objective and comprehensive evaluation framework to assess their performance on real-world tasks. However, existing evaluation datasets are limited in scope and primarily text-based, lacking a unified and standardized multimodal question-answering (QA) benchmark. To address this issue, we introduce TCM-Ladder, the first multimodal QA dataset specifically designed for evaluating large TCM language models. The dataset spans multiple core disciplines of TCM, including fundamental theory, diagnostics, herbal formulas, internal medicine, surgery, pharmacognosy, and pediatrics. In addition to textual content, TCM-Ladder incorporates various modalities such as images and videos. The datasets were constructed using a combination of automated and manual filtering processes and comprise 52,000+ questions in total. These questions include single-choice, multiple-choice, fill-in-the-blank, diagnostic dialogue, and visual comprehension tasks. We trained a reasoning model on TCM-Ladder and conducted comparative experiments against 9 state-of-the-art general domain and 5 leading TCM-specific LLMs to evaluate their performance on the datasets. Moreover, we propose Ladder-Score, an evaluation method specifically designed for TCM question answering that effectively assesses answer quality regarding terminology usage and semantic expression. To our knowledge, this is the first work to evaluate mainstream general domain and TCM-specific LLMs on a unified multimodal benchmark. The datasets and leaderboard are publicly available at https://tcmladder.com or https://54.211.107.106 and will be continuously updated. The source code is available at https://github.com/orangeshushu/TCM-Ladder.



Paperid:4273
Authors:Wei Sun, Wen Yang, Pu Jian, Qianlong Du, Fuwei Cui, Shuo Ren, Jiajun Zhang
Abstract:
Recent advances have demonstrated that integrating reinforcement learning with rule-based rewards can significantly enhance the reasoning capabilities of large language models (LLMs), even without supervised fine-tuning (SFT). However, prevalent reinforcement learning algorithms such as GRPO and its variants like DAPO, suffer from a coarse granularity issue when computing the advantage. Specifically, they compute rollout-level advantages that assign identical values to every token within a sequence, failing to capture token-specific contributions. To address this limitation, we propose Key-token Advantage Estimation (KTAE)—a novel algorithm that estimates fine-grained, token-level advantages without introducing additional models. KTAE leverages the correctness of sampled rollouts and applies statistical analysis to quantify the importance of individual tokens within a sequence to the final outcome. This quantified token-level importance is then combined with the rollout-level advantage to obtain a more fine-grained token-level advantage estimation. Empirical results show that models trained with GRPO+KTAE and DAPO+KTAE outperform baseline methods across five mathematical reasoning benchmarks. Notably, they achieve higher accuracy with shorter responses and even surpass R1-Distill-Qwen-1.5B using the same base model.



Paperid:4274
Authors:Danni Yang, Zhikang Chen, Sen Cui, Mengyue Yang, Ding Li, Abudukelimu Wuerkaixi, Haoxuan Li, Jinke Ren, Mingming Gong
Abstract:
Federated continual learning (FCL) has garnered increasing attention for its ability to support distributed computation in environments with evolving data distributions. However, the emergence of new tasks introduce both temporal and cross-client shifts, making catastrophic forgetting a critical challenge. Most existing works aggregate knowledge from clients into a global model, which may not enhance client performance since irrelevant knowledge could introduce interference, especially in heterogeneous scenarios. Additionally, directly applying decentralized approaches to FCL suffer from ineffective group formation caused by task changes. To address these challenges, we propose a decentralized dynamic cooperation framework for FCL, where clients establish dynamic cooperative learning coalitions to balance the acquisition of new knowledge and the retention of prior learning, thereby obtaining personalized models. To maximize model performance, each client engages in selective cooperation, dynamically allying with others who offer meaningful performance gains. This results in non-overlapping, variable coalitions at each stage of the task. Moreover, we use coalitional affinity game to simulate coalition relationships between clients. By assessing both client gradient coherence and model similarity, we quantify the client benefits derived from cooperation. We also propose a merge-blocking algorithm and a dynamic cooperation evolution algorithm to achieve cooperative and dynamic equilibrium. Comprehensive experiments demonstrate the superiority of our method compared to various baselines. Code is available at: https://anonymous.4open.science/r/DCFCL-0372.



Paperid:4275
Authors:Vida Adeli, Ivan Klabučar, Javad Rajabi, Benjamin Filtjens, Soroush Mehraban, Diwei Wang, Trung Hieu Hoang, Minh Do, Hyewon Seo, Candice MULLER, Daniel Coelho, Claudia de Oliveira, Pieter Ginis, Moran Gilat, Alice Nieuwboer, Joke Spildooren, J. Mckay, Hyeokhyen Kwon, Gari Clifford, Christine Esper, Stewart Factor, Imari Genias, Amirhossein Dadashzadeh, Leia Shum, Alan Whone, Majid Mirmehdi, Andrea Iaboni, Babak Taati
Abstract:
Objective gait assessment in Parkinson’s Disease (PD) is limited by the absence of large, diverse, and clinically annotated motion datasets. We introduce Care-PD, the largest publicly available archive of 3D mesh gait data for PD, and the first multi-site collection spanning 9 cohorts from 8 clinical centers. All recordings (RGB video or motion capture) are converted into anonymized SMPL meshes via a harmonized preprocessing pipeline. Care-PD supports two key benchmarks: supervised clinical score prediction (estimating Unified Parkinson’s Disease Rating Scale, UPDRS, gait scores) and unsupervised motion pretext tasks (2D-to-3D keypoint lifting and full-body 3D reconstruction). Clinical prediction is evaluated under four generalization protocols: within-dataset, cross-dataset, leave-one-dataset-out, and multi-dataset in-domain adaptation.To assess clinical relevance, we compare state-of-the-art motion encoders with a traditional gait-feature baseline, finding that encoders consistently outperform handcrafted features. Pretraining on Care-PD reduces MPJPE (from 60.8mm to 7.5mm) and boosts PD severity macro-F1 by 17\%, underscoring the value of clinically curated, diverse training data. Care-PD and all benchmark code are released for non-commercial research (Code, Data).



Paperid:4276
Authors:Nadhir Hassen, Zhen Zhang, Johan Verjans
Abstract:
Probabilistic time series forecasting, essential in domains like healthcare and neuroscience, requires models capable of capturing uncertainty and intricate temporal dependencies. While deep learning has advanced forecasting, generating calibrated probability distributions over continuous future values remains challenging. We introduce Temporal Generative Flow Networks (Temporal GFNs), adapting Generative Flow Networks (GFNs) – a powerful framework for generating compositional objects – to this sequential prediction task. GFNs learn policies to construct objects (eg. forecast trajectories) step-by-step, sampling final objects proportionally to a reward signal. However, applying GFNs directly to continuous time series necessitates addressing their inherently discrete action spaces and ensuring differentiability. Our framework tackles this by representing time series segments as states and sequentially generating future values via quantized actions chosen by a forward policy. We introduce two key innovations: (1) An adaptive, curriculum-based quantization strategy that dynamically adjusts the number of discretization bins based on reward improvement and policy entropy, balancing precision and exploration throughout training. (2) A straight-through estimator mechanism enabling the forward policy to output both discrete (hard) samples for trajectory construction and continuous (soft) samples for stable gradient propagation. Training utilizes a trajectory balance loss objective, ensuring flow consistency, augmented by an entropy regularizer. We provide rigorous theoretical bounds on the quantization error's impact and the adaptive factor's range. We demonstrate how Temporal GFNs offer a principled way to leverage the structured generation capabilities of GFNs for probabilistic forecasting in continuous domains.



Paperid:4277
Authors:Haonan Wang, Jingyu Lu, Hongrui Li, Xiaomeng Li
Abstract:
Recent advances in neural decoding have enabled the reconstruction of visual experiences from brain activity, positioning fMRI-to-image reconstruction as a promising bridge between neuroscience and computer vision. However, current methods predominantly rely on subject-specific models or require subject-specific fine-tuning, limiting their scalability and real-world applicability. In this work, we introduce ZEBRA, the first zero-shot brain visual decoding framework that eliminates the need for subject-specific adaptation. ZEBRA is built on the key insight that fMRI representations can be decomposed into subject-related and semantic-related components. By leveraging adversarial training, our method explicitly disentangles these components to isolate subject-invariant, semantic-specific representations. This disentanglement allows ZEBRA to generalize to unseen subjects without any additional fMRI data or retraining. Extensive experiments show that ZEBRA significantly outperforms zero-shot baselines and achieves performance comparable to fully finetuned models on several metrics. Our work represents a scalable and practical step toward universal neural decoding. Code and weights will be publicly available upon acceptance.



Paperid:4278
Authors:François Rozet, Ruben Ohana, Michael McCabe, Gilles Louppe, Francois Lanusse, Shirley Ho
Abstract:
Abstract:The steep computational cost of diffusion models at inference hinders their use as fast physics emulators. In the context of image and video generation, this computational drawback has been addressed by generating in the latent space of an autoencoder instead of the pixel space. In this work, we investigate whether a similar strategy can be effectively applied to the emulation of dynamical systems and at what cost. We find that the accuracy of latent-space emulation is surprisingly robust to a wide range of compression rates (up to $1000\times$). We also show that diffusion-based emulators are consistently more accurate than non-generative counterparts and compensate for uncertainty in their predictions with greater diversity. Finally, we cover practical design choices, spanning from architectures to optimizers, that we found critical to train latent-space emulators.



Paperid:4279
Authors:Tianxing Wu, Shutong Zhu, Jingting Wang, Ning Xu, Guilin Qi, Haofen Wang
Abstract:
Uncertain knowledge graphs (UKGs) associate each triple with a confidence score to provide more precise knowledge representations. Recently, since real-world UKGs suffer from the incompleteness, uncertain knowledge graph (UKG) completion attracts more attention, aiming to complete missing triples and confidences. Current studies attempt to learn UKG embeddings to solve this problem, but they neglect the extremely imbalanced distributions of triple confidences. This causes that the learnt embeddings are insufficient to high-quality UKG completion. Thus, in this paper, to address the above issue, we propose a new semi-supervised Confidence Distribution Learning (ssCDL) method for UKG completion, where each triple confidence is transformed into a confidence distribution to introduce more supervision information of different confidences to reinforce the embedding learning process. ssCDL iteratively learns UKG embedding by relational learning on labeled data (i.e., existing triples with confidences) and unlabeled data with pseudo labels (i.e., unseen triples with the generated confidences), which are predicted by meta-learning to augment the training data and rebalance the distribution of triple confidences. Experiments on two UKG datasets demonstrate that ssCDL consistently outperforms the state-of-the-art baselines in different evaluation metrics.



Paperid:4280
Authors:Hanchen Su, Xuyuan Li, Yan Zhou, zhuoyi lu, Ziwei Chai, Haozheng Wang, Chen Zhang, YANG YANG
Abstract:
The increasing utilization of graph databases across various fields stems from their capacity to represent intricate interconnections. Nonetheless, exploiting the full capabilities of graph databases continues to be a significant hurdle, largely because of the inherent difficulty in translating natural language into Cypher. Recognizing the critical role of schema selection in database query generation and drawing inspiration from recent progress in reasoning-augmented approaches trained through reinforcement learning to enhance inference capabilities and generalization, we introduce Cypher-RI, a specialized framework for the Text-to-Cypher task. Distinct from conventional approaches, our methodology seamlessly integrates schema selection within the Cypher generation pipeline, conceptualizing it as a critical element in the reasoning process. The schema selection mechanism is guided by textual context, with its outcomes recursively shaping subsequent inference processes. Impressively, our 7B-parameter model, trained through this RL paradigm, demonstrates superior performance compared to baselines, exhibiting a 9.41\% accuracy improvement over GPT-4o on CypherBench. These results underscore the effectiveness of our proposed reinforcement learning framework, which integrates schema selection to enhance both the accuracy and reasoning capabilities in Text-to-Cypher tasks.



Paperid:4281
Authors:Zhengqing Wang, Yuefan Wu, Jiacheng Chen, Fuyang Zhang, Yasutaka Furukawa
Abstract:
This paper proposes a neural rendering approach that represents a scene as "compressed light-field tokens (CLiFTs)", retaining rich appearance and geometric information of a scene. CLiFT enables compute-efficient rendering by compressed tokens, while being capable of changing the number of tokens to represent a scene or render a novel view with one trained network. Concretely, given a set of images, multi-view encoder tokenizes the images with the camera poses. Latent-space K-means selects a reduced set of rays as cluster centroids using the tokens. The multi-view ``condenser'' compresses the information of all the tokens into the centroid tokens to construct CLiFTs. At test time, given a target view and a compute budget (i.e., the number of CLiFTs), the system collects the specified number of nearby tokens and synthesizes a novel view using a compute-adaptive renderer. trained to handle a variable number of tokens. Extensive experiments on RealEstate10K and DL3DV datasets quantitatively and qualitatively validate our approach, achieving significant data reduction with comparable rendering quality and the highest overall rendering score, while providing trade-offs of data size, rendering quality, and rendering speed.



Paperid:4282
Authors:Omer Moussa, Mariya Toneva
Abstract:
Pretrained language models are remarkably effective in aligning with human brain responses (e.g., fMRI) elicited by natural language stimuli, positioning them as promising model organisms for studying language processing in the brain. However, existing approaches for both estimating and improving this brain alignment are participant-dependent and highly affected by the amount of data available per participant, hindering both generalization to new participants and population-level analyses. In this work, we address these limitations by introducing a scalable, generalizable brain-tuning method, in which we fine-tune pretrained speech language models to jointly predict fMRI responses from multiple participants who listen to the same natural stories. We demonstrate that the resulting brain-tuned models exhibit strong individual brain alignment while generalizing across participants. Specifically, our method leads to 1) a 5-fold decrease in the amount of fMRI data needed to predict brain data from new participants and 2) up to a 50\% increase in the overall brain alignment. Furthermore, multi-brain-tuning additionally improves downstream performance on semantic tasks, suggesting that training using brain data from multiple participants leads to more generalizable semantic representations. Taken together, these findings demonstrate a bidirectional benefit between neuroscience and AI, helping bridge the gap between the two fields.



Paperid:4283
Authors:Susav Shrestha, Bradley Settlemyer, Nikoli Dryden, A. Reddy
Abstract:
Abstract:Accelerating large language model (LLM) inference is critical for real-world deployments requiring high throughput and low latency. Contextual sparsity, where each token dynamically activates only a small subset of the model parameters, shows promise but does not scale to large batch sizes due to union of active neurons quickly approaching dense computation. We introduce Polar Sparsity, highlighting a key shift in sparsity importance from MLP to Attention layers as we scale batch size and sequence length. While MLP layers become more compute-efficient under batching, their sparsity vanishes. In contrast, attention becomes increasingly more expensive at scale, while their head sparsity remains stable and batch-invariant. We develop hardware-efficient, sparsity-aware GPU kernels for selective MLP and Attention computations, delivering up to $2.2\times$ end-to-end speedups for models like OPT, LLaMA-2 \& 3, across various batch sizes and sequence lengths without compromising accuracy. To our knowledge, this is the first work to demonstrate that contextual sparsity can scale effectively to large batch sizes, delivering substantial inference acceleration with minimal changes, making Polar Sparsity practical for large-scale, high-throughput LLM deployment systems.



Paperid:4284
Authors:Patrick Seifner, Kostadin Cvejoski, David Berghaus, César Ali Ojeda Marin, Ramsés J. Sánchez
Abstract:
Stochastic differential equations (SDEs) describe dynamical systems where deterministic flows, governed by a drift function, are superimposed with random fluctuations, dictated by a diffusion function. The accurate estimation (or discovery) of these functions from data is a central problem in machine learning, with wide application across the natural and social sciences. Yet, current solutions either heavily rely on prior knowledge of the dynamics or involve intricate training procedures. We introduce FIM-SDE (Foundation Inference Model for SDEs), a pretrained recognition model capable of performing accuratein-context(or zero-shot) estimation of the drift and diffusion functions oflow-dimensionalSDEs from noisy time series data. Leveraging concepts from amortized inference and neural operators, we train FIM-SDE in a supervised fashion to map a large set of noisy, discretely observed SDE paths onto the space of drift and diffusion functions. We demonstrate thatone and the sameFIM-SDE model achieves robust in-context function estimation (i.e.without any parameter fine-tuning) across a wide range of synthetic and real-world processes --- from canonical SDE systems (e.g. double-well dynamics or weakly perturbed Lorentz attractors) to stock price recordings and oil-price and wind-speed fluctuations.Our pretrained model, repository and tutorials are available online (see link in .pdf file)



Paperid:4285
Authors:Robin Yadav, Qi Yan, Guy Wolf, Joey Bose, Renjie Liao
Abstract:
A fundamental challenge in organic chemistry is identifying and predicting the sequence of reactions that synthesize a desired target molecule. Due to the combinatorial nature of the chemical search space, single-step reactant prediction—i.e., single-step retrosynthesis—remains difficult, even for state-of-the-art template-free generative methods. These models often struggle to produce an accurate yet diverse set of feasible reactions in a chemically rational manner. In this paper, we propose RETRO SYNFLOW (RSF), a discrete flow-matching framework that formulates single-step retrosynthesis as a Markov bridge between a given product molecule and its corresponding reactants. Unlike prior approaches, RSF introduces a reaction center identification step to extract intermediate structures, or synthons, which serve as a more informative and structured source distribution for the discrete flow model. To further improve the diversity and chemical feasibility of generated samples, RSF incorporates Feynman-Kac (FK) steering with Sequential Monte Carlo (SMC) resampling at inference time. This approach leverages a learned forward-synthesis reward oracle to guide the generation process toward more promising reactant candidates. Empirically, RSF substantially outperforms the previous state-of-the-art methods in top-1 accuracy. In addition, FK-steering significantly improves round-trip accuracy, demonstrating stronger chemical validity and synthetic feasibility, all while maintaining competitive top-k performance. These results establish RSF as a new leading approach for single-step retrosynthesis prediction.



Paperid:4286
Authors:Fei Ye, Yulong Zhao, Qihe Liu, Junlin Chen, Adrian G. Bors, Jingling sun, Rongyao Hu, shijie zhou
Abstract:
Continual learning requires the model to continually capture novel information without forgetting prior knowledge. Nonetheless, existing studies predominantly address the catastrophic forgetting, often neglecting enhancements in model robustness. Consequently, these methodologies fall short in real-time applications, such as autonomous driving, where data samples frequently exhibit noise due to environmental and lighting variations, thereby impairing model efficacy and causing safety issues. In this paper, we address robustness in continual learning systems by introducing an innovative approach, the Dynamic Siamese Expansion Framework (DSEF) that employs a Siamese backbone architecture, comprising static and dynamic components, to facilitate the learning of both global and local representations over time. Specifically, the proposed framework dynamically generates a lightweight expert for each novel task, leveraging the Siamese backbone to enable rapid adaptation. A novel Robust Dynamic Representation Optimization (RDRO) approach is proposed to incrementally update the dynamic backbone by maintaining all previously acquired representations and prediction patterns of historical experts, thereby fostering new task learning without inducing detrimental knowledge transfer. Additionally, we propose a novel Robust Feature Fusion (RFF) approach to incrementally amalgamate robust representations from all historical experts into the expert construction process. A novel mutual information-based technique is employed to derive adaptive weights for feature fusion by assessing the knowledge relevance between historical experts and the new task, thus maximizing positive knowledge transfer effects. A comprehensive experimental evaluation, benchmarking our approach against established baselines, demonstrates that our method achieves state-of-the-art performance even under adversarial attacks.



Paperid:4287
Authors:Runze Yan, Xun Shen, Akifumi Wachi, Sebastien Gros, Anni Zhao, Xiao Hu
Abstract:
Abstract:When applying offline reinforcement learning (RL) in healthcare scenarios, the out-of-distribution (OOD) issues pose significant risks, as inappropriate generalization beyond clinical expertise can result in potentially harmful recommendations. While existing methods like conservative Q-learning (CQL) attempt to address the OOD issue, their effectiveness is limited by only constraining action selection by suppressing uncertain actions. This action-only regularization imitates clinician actions that prioritize short-term rewards, but it fails to regulate downstream state trajectories, thereby limiting the discovery of improved long-term treatment strategies. To safely improve policy beyond clinician recommendations while ensuring that state-action trajectories remain in-distribution, we propose Offline Guarded Safe Reinforcement Learning ($\mathsf{OGSRL}$), a theoretically grounded model-based offline RL framework. $\mathsf{OGSRL}$ introduces a novel dual constraint mechanism for improving policy with reliability and safety. First, the OOD guardian is established to specify clinically validated regions for safe policy exploration. By constraining optimization within these regions, it enables the reliable exploration of treatment strategies that outperform clinician behavior by leveraging the full patient state history, without drifting into unsupported state-action trajectories. Second, we introduce a safety cost constraint that encodes medical knowledge about physiological safety boundaries, providing domain-specific safeguards even in areas where training data might contain potentially unsafe interventions. Notably, we provide theoretical guarantees on safety and near-optimality: policies that satisfy these constraints remain in safe and reliable regions and achieve performance close to the best possible policy supported by the data. When evaluated on the MIMIC-III sepsis treatment dataset, $\mathsf{OGSRL}$ demonstrated significantly better OOD handling than baselines. $\mathsf{OGSRL}$ achieved a 78\% reduction in mortality estimates and a 51\% increase in reward compared to clinician decisions.



Paperid:4288
Authors:Ev Zisselman, Mirco Mutti, Shelly Francis-Meretzki, Elisei Shafer, Aviv Tamar
Abstract:
Abstract:Behavioral cloning is a simple yet effective technique for learning sequential decision-making from demonstrations. Recently, it has gained prominence as the core of foundation models for the physical world, where achieving generalization requires countless demonstrations of a multitude of tasks. Typically, a human expert with full information on the task demonstrates a (nearly) optimal behavior. In this paper, we propose to hide some of the task's information from the demonstrator. This ``blindfolded'' expert is compelled to employ non-trivial *exploration* to solve the task. We show that cloning the blindfolded expert generalizes better to unseen tasks than its fully-informed counterpart. We conduct experiments of real-world robot peg insertion tasks with (limited) human demonstrations, alongside a videogame from the Procgen benchmark. Additionally, we support our findings with theoretical analysis, which confirms that the generalization error scales with $\sqrt{I/m}$, where $I$ measures the amount of task information available to the demonstrator, and $m$ is the number of demonstrated tasks. Both theory and practice indicate that cloning blindfolded experts generalizes better with fewer demonstrated tasks. Project page with videos and code: [https://sites.google.com/view/blindfoldedexperts/home](https://sites.google.com/view/blindfoldedexperts/home)



Paperid:4289
Authors:Zhiyi Lyu, Jianguo Huang, Yanchen Deng, Steven Hoi, Bo An
Abstract:
Abstract:Large Language Models (LLMs) with inference-time scaling techniques show promise for code generation, yet face notable efficiency and scalability challenges. Construction-based tree-search methods suffer from rapid growth in tree size, high token consumption, and lack of anytime property. In contrast, improvement-based methods offer better performance but often struggle with uninformative reward signals and inefficient search strategies. In this work, we propose $\textbf{ReLoc}$, a unified local search framework which effectively performs step-by-step code revision. Specifically, ReLoc explores a series of local revisions through four key algorithmic components: initial code drafting, neighborhood code generation, candidate evaluation, and incumbent code updating, each of which can be instantiated with specific decision rules to realize different local search algorithms such as Hill Climbing (HC) or Genetic Algorithm (GA). Furthermore, we develop a specialized revision reward model that evaluates code quality based on revision distance to produce fine-grained preferences that guide the local search toward more promising candidates. Finally, our extensive experimental results demonstrate that our approach achieves superior performance across diverse code generation tasks, significantly outperforming both construction-based tree search as well as the state-of-the-art improvement-based code generation methods.



Paperid:4290
Authors:Valentin Schmutz, Ali Haydaroglu, Shuqi Wang, Yixiao Feng, Matteo Carandini, Kenneth D Harris
Abstract:
Computation in recurrent networks of neurons has been hypothesized to occur at the level of low-dimensional latent dynamics, both in artificial systems and in the brain. This hypothesis seems at odds with evidence from large-scale neuronal recordings in mice showing that neuronal population activity is high-dimensional. To demonstrate that low-dimensional latent dynamics and high-dimensional activity can be two sides of the same coin, we present an analytically solvable recurrent neural network (RNN) model whose dynamics can be exactly reduced to a low-dimensional dynamical system, but generates high-dimensional activity with a power-law covariance eigenspectrum. This raises the question: Do low-dimensional latents explain the high-dimensional activity observed in mouse visual cortex? Spectral theory tells us that the covariance eigenspectrum alone does not allow us to recover the dimensionality of the latents, which can be low or high, when neurons are nonlinear. To address this indeterminacy, we develop Neural Cross-Encoder (NCE), an interpretable, nonlinear latent variable modeling method for calcium recordings, and find that high-dimensional neuronal responses to drifting gratings and spontaneous activity in visual cortex can be reduced to low-dimensional latents, while the responses to natural images cannot. Together, our results show that taking into account neuronal nonlinearities suffices to reconcile apparently conflicting views on the dimensionality of neural activity in computational and systems neuroscience.



Paperid:4291
Authors:Kaleab Kinfu, Rene Vidal
Abstract:
Recent advances in multi-modal representation learning have led to unified embedding spaces that align modalities such as images, text, audio, and vision. However, human motion sequences, a modality that is fundamental to understanding dynamic human activities, remains largely unrepresented in these frameworks. Semantic understanding of actions requires multi-modal grounding: text conveys descriptive semantics, vision provides visual context, and audio provides environmental cues. To bridge this gap, we extend the LanguageBind embedding space to incorporate human motion. Specifically, we introduce Multi-Scale Temporal Motion Transformer (MuTMoT), a transformer-based encoder that maps motion sequences into semantically meaningful embeddings. Alignment is achieved via diverse cross-modal supervision: motion-text pairs from HumanML3D and KIT-ML, motion-video pairs rendered from AMASS, and motion-video-audio triplets from AIST++. Beyond alignment, we introduce REALM, a retrieval-augmented latent diffusion model for motion synthesis. Our approach achieves state-of-the-art or competitive performance across motion reconstruction, cross-modal retrieval, zero-shot action recognition, and text-to-motion generation benchmarks.



Paperid:4292
Authors:Zhenpeng Huang, Jiaqi Li, zihan jia, Xinhao Li, Desen Meng, Lingxue Song, Xi Chen, Liang Li, Limin Wang
Abstract:
We present LongVPO, a novel two‑stage Direct Preference Optimization framework that enables short‑context vision‑language models to robustly understand ultra‑long videos without any long‑video annotations. In Stage 1, we synthesize preference triples by anchoring questions to individual short clips, interleaving them with distractors, and applying visual‑similarity and question‑specificity filtering to mitigate positional bias and ensure unambiguous supervision. We also approximate the reference model’s scoring over long contexts by evaluating only the anchor clip, reducing computational overhead. In Stage 2, we employ a recursive captioning pipeline on genuine long videos to generate scene‑level metadata, then use a large language model to craft multi‑segment reasoning queries and dispreferred responses, allowing the target model to self‑generate preferred answers. With only 16K synthetic examples and no costly human labels, LongVPO outperforms state‑of‑the‑art open‑source models on multiple long‑video benchmarks, +5.4 on LVBench, +3.9 on LongVideoBench, +6.5 on MLVU, and +3.4 on Video-MME, while maintaining strong short‑video performance (+1.1 on MVBench), offering a scalable paradigm for efficient long‑form video understanding.



Paperid:4293
Authors:Han Xiao, Guozhi Wang, Yuxiang Chai, Zimu Lu, Weifeng Lin, Hao He, Lue Fan, Liuyang Bian, Rui Hu, Liang Liu, Shuai Ren, Yafei Wen, xiaoxin chen, Aojun Zhou, Hongsheng Li
Abstract:
In this paper, we introduce UI-Genie, a self-improving framework addressing two key challenges in GUI agents: verification of trajectory outcome is challenging and high-quality training data are not scalable. These challenges are addressed by a reward model and a self-improving pipeline, respectively. The reward model, UI-Genie-RM, features an image-text interleaved architecture that efficiently processes historical context and unifies action-level and task-level rewards. To support the training of UI-Genie-RM, we develop deliberately-designed data generation strategies including rule-based verification, controlled trajectory corruption, and hard negative mining. To address the second challenge, a self-improvement pipeline progressively expands solvable complex GUI tasks by enhancing both the agent and reward models through reward-guided exploration and outcome verification in dynamic environments. For training the model, we generate UI-Genie-RM-517k and UI-Genie-Agent-16k, establishing the first reward-specific dataset for GUI agents while demonstrating high-quality synthetic trajectory generation without manual annotation. Experimental results show that UI-Genie achieves state-of-the-art performance across multiple GUI agent benchmarks with three generations of data-model self-improvement. We will open-source our complete framework implementation to facilitate further research.



Paperid:4294
Authors:JIPENG LI, Yanning Shen
Abstract:
Explicit noise-level conditioning is widely regarded as essential for the effective operation of Graph Diffusion Models (GDMs). In this work, we challenge this assumption by investigating whether denoisers can implicitly infer noise levels directly from corrupted graph structures, potentially eliminating the need for explicit noise conditioning. To this end, we develop a theoretical framework centered on Bernoulli edge-flip corruptions and extend it to encompass more complex scenarios involving coupled structure-attribute noise. Extensive empirical evaluations on both synthetic and real-world graph datasets, using models such as GDSS and DiGress, provide strong support for our theoretical findings. Notably, unconditional GDMs achieve performance comparable or superior to their conditioned counterparts, while also offering reductions in parameters (4-6%) and computation time (8-10%). Our results suggest that the high-dimensional nature of graph data itself often encodes sufficient information for the denoising process, opening avenues for simpler, more efficient GDM architectures.



Paperid:4295
Authors:Yunbo Li, Jiaping Gui, Zhihang Deng, Fanchao Meng, Yue Wu
Abstract:
Federated learning (FL) enables collaborative model training across multiple parties without sharing raw data, with semi-asynchronous FL (SAFL) emerging as a balanced approach between synchronous and asynchronous FL. However, SAFL faces significant challenges in optimizing both gradient-based (e.g., FedSGD) and model-based (e.g., FedAvg) aggregation strategies, which exhibit distinct trade-offs in accuracy, convergence speed, and stability. While gradient aggregation achieves faster convergence and higher accuracy, it suffers from pronounced fluctuations, whereas model aggregation offers greater stability but slower convergence and suboptimal accuracy. This paper presents FedQS, the first framework to theoretically analyze and address these disparities in SAFL. FedQS introduces adivide-and-conquer strategyto handle client heterogeneity by classifying clients into four distinct types and adaptively optimizing their local training based on data distribution characteristics and available computational resources. Extensive experiments on computer vision, natural language processing, and real-world tasks demonstrate that FedQS achieves the highest accuracy, attains the lowest loss, and ranks among the fastest in convergence speed, outperforming state-of-the-art baselines. Our work bridges the gap between aggregation strategies in SAFL, offering a unified solution for stable, accurate, and efficient federated learning. The code and datasets are available at https://anonymous.4open.science/r/FedQS-EDD6.



Paperid:4296
Authors:Hongxu Chen, Ke Wei, Haishan Ye, Luo Luo
Abstract:
In this paper, we study the distributed convex-concave finite-sum minimax optimization over the network, and a decentralized variance-reduced optimistic gradient method with stochastic mini-batch sizes (DIVERSE) is proposed. For the strongly-convex-strongly-concave objective, it is shown that DIVERSE can achieve a linear convergence rate that depends on the global smoothness parameters, yielding sharper computation and communication complexity bounds than existing results. Furthermore, we also establish the lower complexity bounds, which show that our upper bounds are optimal up to a logarithmic factor in terms of the local incremental first-order oracle calls, the computation rounds, and the communication rounds. Numerical experiments demonstrate that our algorithm outperforms existing methods in practice.



Paperid:4297
Authors:Shaocong Ma, Heng Huang
Abstract:
Zeroth-order optimization (ZOO) is an important framework for stochastic optimization when gradients are unavailable or expensive to compute. A potential limitation of existing ZOO methods is the bias inherent in most gradient estimators unless the perturbation stepsize vanishes. In this paper, we overcome this biasedness issue by proposing a novel family ofunbiasedgradient estimators based solely on function evaluations. By reformulating directional derivatives as a telescoping series and sampling from carefully designed distributions, we construct estimators that eliminate bias while maintaining favorable variance. We analyze their theoretical properties, derive optimal scaling distributions and perturbation stepsizes of four specific constructions, and prove that SGD using the proposed estimators achieves optimal complexity for smooth non-convex objectives. Experiments on synthetic tasks and language model fine-tuning confirm the superior accuracy and convergence of our approach compared to standard methods.



Paperid:4298
Authors:Dilxat Muhtar, Enzhuo Zhang, Zhenshi Li, Feng Gu, Yanglangxing He, Pengfeng Xiao, Xueliang Zhang
Abstract:
Vision-Language Models (VLMs) have demonstrated great potential in interpreting remote sensing (RS) images through language-guided semantic. However, the effectiveness of these VLMs critically depends on high-quality image-text training data that captures rich semantic relationships between visual content and language descriptions. Unlike natural images, RS lacks large-scale interleaved image-text pairs from web data, making data collection challenging. While current approaches rely primarily on rule-based methods or flagship VLMs for data synthesis, a systematic framework for automated quality assessment of such synthetically generated RS vision-language data is notably absent. To fill this gap, we propose a novel score model trained on large-scale RS vision-language preference data for automated quality assessment. Our empirical results demonstrate that fine-tuning CLIP or advanced VLMs (e.g., Qwen2-VL) with the top 30% of data ranked by our score model achieves superior accuracy compared to both full-data fine-tuning and CLIP-score-based ranking approaches. Furthermore, we demonstrate applications of our scoring model for reinforcement learning (RL) training and best-of-N (BoN) test-time scaling, enabling significant improvements in VLM performance for RS tasks. Our code, model, and dataset will be made publicly available.



Paperid:4299
Authors:Shitong Xu, Yiyuan Yang, Niki Trigoni, Andrew Markham
Abstract:
Target speaker extraction focuses on isolating a specific speaker's voice from an audio mixture containing multiple speakers. To provide information about the target speaker's identity, prior works have utilized clean audio samples as conditioning inputs. However, such clean audio examples are not always readily available. For instance, obtaining a clean recording of a stranger's voice at a cocktail party without leaving the noisy environment is generally infeasible. Limited prior research has explored extracting the target speaker's characteristics from noisy enrollments, which may contain overlapping speech from interfering speakers. In this work, we explore a novel enrollment strategy that encodes target speaker information from the noisy enrollment by comparing segments where the target speaker is talking (Positive Enrollments) with segments where the target speaker is silent (Negative Enrollments). Experiments show the effectiveness of our model architecture, which achieves over 2.1 dB higher SI-SNRi compared to prior works in extracting the monaural speech from the mixture of two speakers. Additionally, the proposed two-stage training strategy accelerates convergence, reducing the number of optimization steps required to reach 3 dB SNR by 60\%. Overall, our method achieves state-of-the-art performance in the monaural target speaker extraction conditioned on noisy enrollments.



Paperid:4300
Authors:Agneet Chatterjee, Rahim Entezari, Maksym Zhuravinskyi, Maksim Lapin, Reshinth Adithyan, Amit Raj, Chitta Baral, 'YZ' Yezhou Yang, Varun Jampani
Abstract:
Recent advances in text-to-video (T2V) generation have enabled high-fidelity video synthesis from natural language prompts. However, existing models and benchmarks fail to capture the complexity and requirements of professional video generation. Towards that goal, we introduce Stable Cinemetrics (SCINE), a structured evaluation framework that formalizes filmmaking principles into four disentangled, hierarchical taxonomies: Setup, Event, Lighting, and Camera} Together, these taxonomies define 76 fine-grained control nodes grounded in industry practices. Using these taxonomies, we construct a benchmark of prompts aligned with professional use cases and develop an automated pipeline for prompt categorization and question generation, enabling independent evaluation of each control dimension. We conduct a large-scale human study spanning 10+ models and 20K videos, annotated by a pool of 80+ film professionals. Our analysis, both coarse and fine-grained reveal that even the strongest current models exhibit significant gaps, particularly in Events and Camera-related controls. To enable scalable evaluation, we train an automatic evaluator, a vision-language model aligned with expert annotations that outperforms existing zero-shot baselines. SCINE is the first approach to formalize professional video generation within the landscape of video generative models, introducing taxonomies centered around cinematic control and supporting them with structured evaluation pipelines and detailed analyses to guide future research.



Paperid:4301
Authors:Yang Peng, Kaicheng Jin, Liangyu Zhang, Zhihua Zhang
Abstract:
Abstract:In this paper, we study the finite-sample statistical rates of distributional temporal difference (TD) learning with linear function approximation.The aim of distributional TD learning is to estimate the return distribution of a discounted Markov decision process for a given policy $\pi$.Previous works on statistical analysis of distributional TD learning mainly focus on the tabular case.In contrast, we first consider the linear function approximation setting and derive sharp finite-sample rates.Our theoretical results demonstrate that the sample complexity of linear distributional TD learning matches that of classic linear TD learning.This implies that, with linear function approximation, learning the full distribution of the return from streaming data is no more difficult than learning its expectation (value function).To derive tight sample complexity bounds, we conduct a fine-grained analysis of the linear-categorical Bellman equation and employ the exponential stability arguments for products of random matrices.Our results provide new insights into the statistical efficiency of distributional reinforcement learning algorithms.



Paperid:4302
Authors:Jin Li, Zezhong Ding, Xike Xie
Abstract:
Abstract:Knowledge graphs (KGs) are vital for enabling knowledge reasoning across various domains. Recent KG reasoning methods that integrate both global and local information have achieved promising results. However, existing methods often suffer from score over-smoothing, which blurs the distinction between correct and incorrect answers and hinders reasoning effectiveness. To address this, we propose DuetGraph, a **coarse-to-fine** KG reasoning mechanism with **dual-pathway** global-local fusion. DuetGraph tackles over-smoothing by segregating—rather than stacking—the processing of local (via message passing) and global (via attention) information into two distinct pathways, preventing mutual interference and preserving representational discrimination. In addition, DuetGraph introduces a **coarse-to-fine** optimization, which partitions entities into high- and low-score subsets. This strategy narrows the candidate space and sharpens the score gap between the two subsets, which alleviates over-smoothing and enhances inference quality. Extensive experiments on various datasets demonstrate that DuetGraph achieves state-of-the-art (SOTA) performance, with up to an **8.7\%** improvement in reasoning quality and a **1.8$\times$** acceleration in training efficiency.



Paperid:4303
Authors:William Overman, Mohsen Bayati
Abstract:
Modern language‑model deployments must often balance competing objectives—for example, helpfulness versus harmlessness, cost versus accuracy, and reward versus safety. We introduce Conformal Arbitrage, a post‑hoc framework that learns a data‑driven threshold to mediate between a Primary model optimized for a primary objective and a more conservative Guardian—which could be another model or a human domain expert—aligned with a guardrail objective. The threshold is calibrated with conformal risk control, yielding finite‑sample, distribution‑free guarantees that the long‑run frequency of undesirable events (such as factual errors or safety violations) does not exceed a user‑specified quota. Because Conformal Arbitrage operates wholly at the API level—without requiring access to model logits or updating model weights—it complements weight‑based alignment techniques and integrates seamlessly with existing cost‑aware cascades. Empirically, Conformal Arbitrage traces an efficient frontier, allowing users to define an acceptable performance level for one objective while maximizing utility in another. We observe that our method outperforms (in terms of accuracy) cost-matched random routing between models. These properties make Conformal Arbitrage a practical, theoretically grounded tool for trustworthy and economical deployment of large language models across a broad range of potentially competing objectives.



Paperid:4304
Authors:Ashesh Ashesh, Florian Jug
Abstract:
Abstract:Fluorescence microscopy, while being a key driver for progress in the life sciences, is also subject to technical limitations. To overcome them, computational multiplexing techniques have recently been proposed, which allow multiple cellular structures to be captured in a single image and later be unmixed. Existing image decomposition methods are trained on a set of superimposed input images and the respective unmixed target images. It is critical to note that the relative strength (mixing ratio) of the superimposed images for a given input is a priori unknown. However, existing methods are trained on a fixed intensity ratio of superimposed inputs, making them not cognizant to the range of relative intensities that can occur in fluorescence microscopy. In this work, we propose a novel method called indiSplit that is cognizant of the severity of the above-mentioned mixing ratio. Our idea is based on InDI, a popular iterative method for image restoration, and an ideal starting point to embrace the unknown mixing ratio in any given input. We introduce $(i)$ a suitably trained regressor network that predicts the degradation level (mixing asymmetry) of a given input image and $(ii)$ a degradation-specific normalization module, enabling degradation-aware inference across all mixing ratios. We show that this method solves two relevant tasks in fluorescence microscopy, namely image splitting and bleedthrough removal and empirically demonstrate the applicability of indiSplit on $5$ public datasets.



Paperid:4305
Authors:Alexander Ratzan, Sidharth Goel, Junhao Wen, Christos Davatzikos, Erdem Varol
Abstract:
Spatial location and molecular interactions have long been linked to the connectivity patterns of neural circuits. Yet, at the macroscale of human brain networks, the interplay between spatial position, gene expression, and connectivity remains poorly understood. Recent efforts to map the human transcriptome and connectome have yielded spatially resolved brain atlases, however predicting connectivity from high-dimensional transcriptomic data while accounting for inherent spatial confounds presents a significant challenge. This paper introduces the Spatiomolecular Transformer (SMT), a new architecture designed to predict whole-brain functional connectivity solely from gene expression and regional spatial coordinates. SMT explicitly models biological context by tokenizing genes based on their transcription start site (TSS) order to capture local genomic organization, and incorporating regional 3D spatial location via a dedicated context [CLS] token within its multi-head self-attention mechanism. We rigorously benchmark SMT against established linear and non-linear baselines. Crucially, we employ custom-generated spatiomolecular null maps, which preserve both spatial and transcriptomic autocorrelation, to ensure that learned relationships are not mere artefacts of spatial proximity. Across multiple connectomics datasets (UKBiobank and MPI-LEMON) and parcellation resolutions, SMT demonstrates competitive predictive performance, particularly for long-range connections, and its predictions robustly exceed the expected contributions from spatial autocorrelation. Beyond prediction, SMT provides enhanced interpretability of gene-set contributions to functional brain architecture. Our findings suggest that the complex, multimodal patterns of functional brain networks have significant spatiomolecular foundations beyond chance. Code to reproduce our results can be found in the anonymous repository at [https://anonymous.4open.science/r/SMT-C56C/].



Paperid:4306
Authors:Nima Hosseini Dashtbayaz, Hesam Salehipour, Adrian Butscher, Nigel Morris
Abstract:
Abstract:Reduced-order modeling (ROM) of time-dependent and parameterized differential equations aims to accelerate the simulation of complex high-dimensional systems by learning a compact latent manifold representation that captures the characteristics of the solution fields and their time-dependent dynamics. Although high-fidelity numerical solvers generate the training datasets, they have thus far been excluded from the training process, causing the learned latent dynamics to drift away from the discretized governing physics. This mismatch often limits generalization and forecasting capabilities. In this work, we propose **Ph**ysics-**i**nformed **ROM** ($\Phi$-ROM) by incorporating differentiable PDE solvers into the training procedure. Specifically, the latent space dynamics and its dependence on PDE parameters are shaped directly by the governing physics encoded in the solver, ensuring a strong correspondence between the full and reduced systems. Our model outperforms state-of-the-art data-driven ROMs and other physics-informed strategies by accurately generalizing to new dynamics arising from unseen parameters, enabling long-term forecasting beyond the training horizon, maintaining continuity in both time and space, and reducing the data cost. Furthermore, $\Phi$-ROM learns to recover and forecast the solution fields even when trained or evaluated with sparse and irregular observations of the fields, providing a flexible framework for field reconstruction and data assimilation. We demonstrate the framework’s robustness across different PDE solvers and highlight its broad applicability by providing an open-source JAX implementation readily extensible to other PDE systems and differentiable solvers.



Paperid:4307
Authors:Tianyu Hua, Harper Hua, Violet Xiang, Benjamin Klieger, Sang Truong, Weixin Liang, Fan-Yun Sun, Nick Haber
Abstract:
Large language models (LLMs) have shown promise in transforming machine learning research, yet their capability to faithfully implement genuinely novel ideas from recent research papers—ideas unseen during pretraining—remains unclear. We introduce ResearchCodeBench, a benchmark that evaluates LLMs’ ability to translate cutting-edge ML contributions from top 2024-2025 research papers into executable code. We assessed 30+ proprietary and open-source LLMs, finding that even the best models correctly implement less than 40% of the code. We present empirical findings on performance comparison, contamination, and error patterns. By providing a rigorous evaluation platform, ResearchCodeBench enables continuous understanding and advancement of LLM-driven innovation in research code generation.



Paperid:4308
Authors:Jingwen Cheng, Ruikun Li, Huandong Wang, Yong Li
Abstract:
Predicting the behavior of complex systems is critical in many scientific and engineering domains, and hinges on the model’s ability to capture their underlying dynamics.Existing methods encode the intrinsic dynamics of high-dimensional observations through latent representations and predict autoregressively. However, these latent representations lose the inherent spatial structure of spatiotemporal dynamics, leading to the predictor's inability to effectively model spatial interactions and neglect emerging dynamics during long-term prediction.In this work, we propose SparseDiff, introducing a test-time adaptation strategy to dynamically update the encoding scheme to accommodate emergent spatiotemporal structures during the long-term evolution of the system. Specifically, we first design a codebook-based sparse encoder, which coarsens the continuous spatial domain into a sparse graph topology. Then, we employ a graph neural ordinary differential equation to model the dynamics and guide a diffusion decoder for reconstruction.SparseDiff autoregressively predicts the spatiotemporal evolution and adjust the sparse topological structure to adapt to emergent spatiotemporal patterns by adaptive re-encoding. Extensive evaluations on representative systems demonstrate that SparseDiff achieves an average prediction error reduction of 49.99\% compared to baselines, requiring only 1\% of the spatial resolution.



Paperid:4309
Authors:Zhengyu Hu, Jianxun Lian, Zheyuan Xiao, Seraphina Zhang, Tianfu Wang, Nicholas Jing Yuan, Xing Xie, Hui Xiong
Abstract:
Large language models (LLMs) have shown impressive capabilities across tasks such as mathematics, coding, and reasoning, yet their learning ability, which is crucial for adapting to dynamic environments and acquiring new knowledge, remains underexplored. In this work, we address this gap by introducing a framework inspired by cognitive psychology and education. Specifically, we decompose general learning ability into three distinct, complementary dimensions:Learning from Instructor(acquiring knowledge via explicit guidance),Learning from Concept(internalizing abstract structures and generalizing to new contexts), andLearning from Experience(adapting through accumulated exploration and feedback). We conduct a comprehensive empirical study across the three learning dimensions and identify several insightful findings, such as (i) interaction improves learning; (ii) conceptual understanding is scale-emergent and benefits larger models; and (iii) LLMs are effective few-shot learners but not many-shot learners. Based on our framework and empirical findings, we introduce a benchmark that provides a unified and realistic evaluation of LLMs' general learning abilities across three learning cognition dimensions. It enables diagnostic insights and supports evaluation and development of more adaptive and human-like models.



Paperid:4310
Authors:Hyungyung Lee, Geon Choi, Jung-Oh Lee, Hangyul Yoon, Hyuk Hong, Edward Choi
Abstract:
Recent progress in Large Vision-Language Models (LVLMs) has enabled promising applications in medical tasks, such as report generation and visual question answering. However, existing benchmarks focus mainly on the final diagnostic answer, offering limited insight into whether models engage in clinically meaningful reasoning. To address this, we present CheXStruct and CXReasonBench, a structured pipeline and benchmark built on the publicly available MIMIC-CXR-JPG dataset. CheXStruct automatically derives a sequence of intermediate reasoning steps directly from chest X-rays, such as segmenting anatomical regions, deriving anatomical landmarks and diagnostic measurements, computing diagnostic indices, and applying clinical thresholds. CXReasonBench leverages this pipeline to evaluate whether models can perform clinically valid reasoning steps and to what extent they can learn from structured guidance, enabling fine-grained and transparent assessment of diagnostic reasoning.The benchmark comprises 18,988 QA pairs across 12 diagnostic tasks and 1,200 cases, each paired with up to 4 visual inputs, and supports multi-path, multi-stage evaluation including visual grounding via anatomical region selection and diagnostic measurements.Even the strongest of 10 evaluated LVLMs struggle with structured reasoning and generalization, often failing to link abstract knowledge with anatomically grounded visual interpretation.



Paperid:4311
Authors:Seongsu Kim, Nayoung Kim, Dongwoo Kim, Sungsoo Ahn
Abstract:
Density functional theory (DFT) is a fundamental method for simulating quantum chemical properties, but it remains expensive due to the iterative self-consistent field (SCF) process required to solve the Kohn–Sham equations. Recently, deep learning methods are gaining attention as a way to bypass this step by directly predicting the Hamiltonian.However, they rely on deterministic regression and do not consider the highly structured nature of Hamiltonians.In this work, we propose QHFlow, a high-order equivariant flow matching framework that generates Hamiltonian matrices conditioned on molecular geometry. Flow matching models continuous-time trajectories between simple priors and complex targets, learning the structured distributions over Hamiltonians instead of direct regression. To further incorporate symmetry, we use a neural architecture that predicts SE(3)-equivariant vector fields, improving accuracy and generalization across diverse geometries. To further enhance physical fidelity, we additionally introduce a fine-tuning scheme to align predicted orbital energies with the target. QHFlow achieves state-of-the-art performance, reducing Hamiltonian error by 71% on MD17 and 53% on QH9. Moreover, we further show that QHFlow accelerates the DFT process without trading off the solution quality when initializing SCF iterations with the predicted Hamiltonian, significantly reducing the number of iterations and runtime.



Paperid:4312
Authors:Ben Schiffer, Mark Sellke
Abstract:
In the incentivized exploration model, a principal aims to explore and learn over time by interacting with a sequence of self-interested agents. It has been recently understood that the main challenge in designing incentive-compatible algorithms for this problem is to gather a moderate amount of initial data, after which one can obtain near-optimal regret via posterior sampling. With high-dimensional contexts, however, this \emph{initial exploration} phase requires exponential sample complexity in some cases, which prevents efficient learning unless initial data can be acquired exogenously. We show that these barriers to exploration disappear under mild geometric conditions on the set of available actions, in which case incentive-compatibility does not preclude regret-optimality. Namely, we consider the linear bandit model with actions in the Euclidean unit ball, and give an incentive-compatible exploration algorithm with sample complexity that scales polynomially with the dimension and other parameters.



Paperid:4313
Authors:Sohyun An, Ruochen Wang, Tianyi Zhou, Cho-Jui Hsieh
Abstract:
While recent success of large reasoning models (LRMs) significantly advanced LLMs' reasoning capability by optimizing the final answer accuracy using reinforcement learning, they may also drastically increase the output length due tooverthinking—characterized by unnecessarily complex reasoning paths that waste computation and potentially degrade the performance.We hypothesize that such inefficiencies stem from LRMs' limited capability to dynamically select the proper modular reasoning strategies, termedthinking patternsat the right position. To investigate this hypothesis, we propose a dynamic optimization framework that segments model-generated reasoning paths into distinct thinking patterns, systematically identifying and promoting beneficial patterns that improve the answer while removing detrimental ones. Empirical analysis confirms that our optimized thinking paths yield more concise yet sufficiently informative trajectories, enhancing reasoning efficiency by reducing attention FLOPs by up to 47% while maintaining accuracy for originally correct responses. Moreover, a non-trivial portion of originally incorrect responses are transformed into correct ones, achieving a 15.6% accuracy improvement with reduced length.Motivated by the improvement brought by the optimized thinking paths, we apply a preference optimization technique supported by a pairwise dataset contrasting suboptimal and optimal reasoning paths. Experimental evaluations across multiple mathematical reasoning benchmarks reveal that our method notably reduces computational overhead while simultaneously improving reasoning accuracy, achieving up to a 12% accuracy improvement and reducing token usage from approximately 5,000 to 3,000 tokens.



Paperid:4314
Authors:Yiting Wang, Wanghao Ye, Ping Guo, Yexiao He, Ziyao Wang, Bowei Tian, Shwai He, Guoheng Sun, Zheyu Shen, Sihan Chen, Ankur Srivastava, Qingfu Zhang, Gang Qu, Ang Li
Abstract:
Optimizing Register Transfer Level (RTL) code is crucial for improving the efficiency and performance of digital circuits in the early stages of synthesis. Manual rewriting, guided by synthesis feedback, can yield high-quality results but is time-consuming and error-prone. Most existing compiler-based approaches have difficulty handling complex design constraints. Large Language Model (LLM)-based methods have emerged as a promising alternative to address these challenges. However, LLM-based approaches often face difficulties in ensuring alignment between the generated code and the provided prompts. This paper introduces SymRTLO, a neuron-symbolic framework that integrates LLMs with symbolic reasoning for the efficient and effective optimization of RTL code. Our method incorporates a retrieval-augmented system of optimization rules and Abstract Syntax Tree (AST)-based templates, enabling LLM-based rewriting that maintains syntactic correctness while minimizing undesired circuit behaviors. A symbolic module is proposed for analyzing and optimizing finite state machine (FSM) logic, allowing fine-grained state merging and partial specification handling beyond the scope of pattern-based compilers. Furthermore, a fast verification pipeline, combining formal equivalence checks with test-driven validation, further reduces the complexity of verification. Experiments on the RTL-Rewriter benchmark with Synopsys Design Compiler and Yosys show that SymRTLO improves power, performance, and area (PPA) by up to 43.9%, 62.5%, and 51.1%, respectively, compared to the state-of-the-art methods. We will release the code as open source upon the paper's acceptance.



Paperid:4315
Authors:Herun Wan, Jiaying Wu, Minnan Luo, Zhi Zeng, Zhixiong Su
Abstract:
Misinformation detection models often rely on superficial cues (i.e., shortcuts) that correlate with misinformation in training data but fail to generalize to the diverse and evolving nature of real-world misinformation. This issue is exacerbated by large language models (LLMs), which can easily generate convincing misinformation through simple prompts. We introduce TruthOverTricks, a unified evaluation paradigm for measuring shortcut learning in misinformation detection. TruthOverTricks categorizes shortcut behaviors into intrinsic shortcut induction and extrinsic shortcut injection, and evaluates seven representative detectors across 14 popular benchmarks, along with two new factual misinformation datasets, NQ-Misinfo and Streaming-Misinfo. Empirical results reveal that existing detectors suffer severe performance degradation when exposed to both naturally occurring and adversarially crafted shortcuts. To address this, we propose SMF, an LLM-augmented data augmentation framework that mitigates shortcut reliance through paraphrasing, factual summarization, and sentiment normalization. SMF consistently enhances robustness across 16 benchmarks, encouraging models to rely on deeper semantic understanding rather than shortcut cues. To promote the development of misinformation detectors, we have published the resources publicly in the anonymous link.



Paperid:4316
Authors:Tianhao Li, Tingfa Xu, Ying Wang, Haolin Qin, Xu Lin, Jianan Li
Abstract:
Drone-based multi-object tracking is essential yet highly challenging due to small targets, severe occlusions, and cluttered backgrounds. Existing RGB-based multi-object tracking algorithms heavily depend on spatial appearance cues such as color and texture, which often degrade in aerial views, compromising tracking reliability. Multispectral imagery, capturing pixel-level spectral reflectance, provides crucial spectral cues that significantly enhance object discriminability under degraded spatial conditions. However, the lack of dedicated multispectral UAV datasets has hindered progress in this domain. To bridge this gap, we introduce \textbf{MMOT}, the first challenging benchmark for drone-based multispectral multi-object tracking dataset. It features three key characteristics: (i) \textbf{Large Scale} — 125 video sequences with over 488.8K annotations across eight object categories; (ii) \textbf{Comprehensive Challenges} — covering diverse real-world challenges such as extreme small targets, high-density scenarios, severe occlusions and complex platform motion; and (iii) \textbf{Precise Oriented Annotations} — enabling accurate localization and reduced object ambiguity under aerial perspectives. To better extract spectral features and leverage oriented annotations, we further present a multispectral and orientation-aware MOT scheme adapting existing MOT methods, featuring: (i) a lightweight Spectral 3D-Stem integrating spectral features while preserving compatibility with RGB pretraining; (ii) a orientation-aware Kalman filter for precise state estimation; and (iii) an end-to-end orientation-adaptive transformer architecture. Extensive experiments across representative trackers consistently show that multispectral input markedly improves tracking performance over RGB baselines, particularly for small and densely packed objects. We believe our work will benefit the community for advancing drone-based multispectral multi-object tracking research.



Paperid:4317
Authors:CHE WANG, Ziqi Zhang, Yinggui Wang, Tiantong Wang, Yurong Hao, Jianbo Gao, Tao Wei, Yang Cao, Zhong Chen, Wei Lim
Abstract:
On-device large models (LMs) reduce cloud dependency but expose proprietary model weights to the end-user, making them vulnerable to white-box model stealing (MS) attacks. A common defense is TEE-Shielded DNN Partition (TSDP), which places all trainable LoRA adapters (fine tuned on private data) inside a trusted execution environment (TEE). However, this design suffers from excessive host-to-TEE communication latency. We propose AegisGuard, a fine tuning and deployment framework that selectively shields the MS sensitive adapters while offloading the rest to the GPU, balancing security and efficiency. AegisGuard integrates two key components: i) RL-based Sensitivity Measurement (RSM), which injects Gaussian noise during training and applies a lightweight reinforcement learning to rank adapters based on their impact on model stealing; and (ii) Shielded-Adapter Compression (SAC), which structurally prunes the selected adapters to reduce both parameter size and intermediate feature maps, further lowering TEE computation and data transfer costs. Extensive experiments demonstrate that AegisGuard achieves black-box level MS resilience (surrogate accuracy around 39%, matching fully shielded baselines), while reducing end-to-end inference latency by 2–3× and cutting TEE memory usage by 4× compared to state-of-the-art TSDP methods. We make the code and artifacts publicly available (https://anonymous.4open.science/r/teeadapterprotection).



Paperid:4318
Authors:Xiaochuan Gong, Jie Hao, Mingrui Liu
Abstract:
Abstract:Hierarchical optimization refers to problems with interdependent decision variables and objectives, such as minimax and bilevel formulations. While various algorithms have been proposed, existing methods and analyses lack adaptivity in stochastic optimization settings: they cannot achieve optimal convergence rates across a wide spectrum of gradient noise levels without prior knowledge of the noise magnitude.In this paper, we propose novel adaptive algorithms for two important classes of stochastic hierarchical optimization problems: nonconvex-strongly-concave minimax optimization and nonconvex-strongly-convex bilevel optimization. Our algorithms achieve sharp convergence rates of $\widetilde{O}(1/\sqrt{T} + \sqrt{\bar{\sigma}}/T^{1/4})$ in $T$ iterations for the gradient norm, where $\bar{\sigma}$ is an upper bound on the stochastic gradient noise. Notably, these rates are obtained \emph{without requiring prior knowledge} of the noise level, thereby enabling automatic adaptivity in both low- and high-noise regimes. To our knowledge, this work provides the first adaptive and sharp convergence guarantees for stochastic hierarchical optimization. Our algorithm design and analysis rely on a novel synthesis of the momentum normalization technique for updating the upper-level variable with a carefully crafted adaptive stepsize scheme for both upper- and lower-level variables. Extensive experiments on synthetic and deep learning tasks demonstrate the effectiveness of our proposed algorithms



Paperid:4319
Authors:Philippe Wyder, Judah Goldfeder, Alexey Yermakov, Yue Zhao, Stefano Riva, Jan Williams, David Zoro, Amy Rude, Matteo Tomasetto, Joe Germany, Joseph Bakarji, Georg Maierhofer, Miles Cranmer, Nathan Kutz
Abstract:
Machine learning (ML) and artificial intelligence (AI) are transforming modeling and control in the physical, engineering, and biological sciences. However, rapid development has outpaced the creation of standardized, objective benchmarks—leading to weak baselines, reporting bias, and inconsistent evaluations across methods. This undermines reproducibility, misguides resource allocation, and obscures scientific progress. To address this, we propose a Common Task Framework (CTF) for scientific machine learning. The CTF features a curated set of datasets and task-specific metrics spanning forecasting, state reconstruction, generalization, and control under realistic constraints, including noise and limited data. Inspired by the success of CTFs in fields like natural language processing and computer vision, our framework provides a structured, rigorous foundation for head-to-head evaluation of diverse algorithms. As a first step, we benchmark methods on two canonical nonlinear systems: Kuramoto-Sivashinsky and Lorenz. These results illustrate the utility of the CTF in revealing method strengths, limitations, and suitability for specific classes of problems and diverse objectives. Our long-term vision is to replace ad hoc comparisons with standardized evaluations on hidden test sets that raise the bar for rigor and reproducibility in scientific ML.



Paperid:4320
Authors:Wenyuan Zhang, Jimin Tang, Weiqi Zhang, Yi Fang, Yu-Shen Liu, Zhizhong Han
Abstract:
Modeling reflections from 2D images is essential for photorealistic rendering and novel view synthesis. Recent approaches enhance Gaussian primitives with reflection-related material attributes to enable physically based rendering (PBR) with 3D Gaussian Splatting (3DGS). However, the material inference often lacks sufficient constraints, especially under limited environment modeling, resulting in illumination aliasing and reduced generalization. In this work, we revisit the problem from a multi-view perspective and show that multi-view consistent material inference with more physically-based environment modeling is key to learning accurate reflections with 3DGS. To this end, we enforce 3D Gaussians to produce multi-view consistent material maps during deferred shading. We also track photometric variations across views to identify highly reflective regions, which serve as strong priors for reflection strength terms. To handle indirect illumination caused by inter-object occlusions, we further introduce an environment modeling strategy through ray tracing with 3DGS, enabling photorealistic rendering of indirect radiance. Experiments on widely used benchmarks show that our method faithfully recovers both illumination and geometry, achieving state-of-the-art rendering quality in novel views synthesis.



Paperid:4321
Authors:Zeyuan Allen-Zhu
Abstract:
Abstract:Understanding architectural differences in language models is challenging, particularly at academic-scale pretraining (e.g., 1.3B params, 100B tokens), where results are often dominated by noise. We propose synthetic pretraining tasks to isolate and evaluate key model capabilities. Using this framework, we identify \emph{Canon layers}: lightweight components—named after the musical term—that enhance horizontal information flow across neighboring tokens. Canon layers compute weighted sum of nearby token representations and integrate seamlessly into Transformers, linear attention, state-space models, or any sequence architecture.We present 12 results, including how Canon enhances reasoning depth ($2\times$), reasoning breadth, knowledge manipulation, etc. Canon transforms weak architectures like NoPE to match RoPE and linear attention to rival state-space models (e.g., Mamba2), validated through synthetic tasks and real-world academic-scale pretraining. This synthetic framework isolates core capabilities often obscured at academic scales, offering an \emph{economical, principled path} to guide architecture design as pretraining pipelines improve, aiming to unlock deeper reasoning.



Paperid:4322
Authors:Jin Hu, Jiakai Wang, linna Jing, Haolin Li, Liu haodong, Haotong Qin, Aishan Liu, Ke Xu, Xianglong Liu
Abstract:
Abstract:Recently, semantically constrained adversarial examples (SemanticAE), which are directly generated from natural language instructions, have become a promising avenue for future research due to their flexible attacking forms, but have not been thoroughly explored yet.To generate SemanticAEs, current methods fall short of satisfactory attacking ability as the key underlying factors of semantic uncertainty in human instructions, such as $\textit{referring diversity}$, $\textit{descriptive incompleteness}$, and $\textit{boundary ambiguity}$, have not been fully investigated.To tackle the issues, this paper develops a multi-dimensional $\textbf{ins}$truction $\textbf{u}$ncertainty $\textbf{r}$eduction ($\textbf{InSUR}$) framework to generate more satisfactory SemanticAE, $\textit{i.e.}$, transferable, adaptive, and effective.Specifically, in the dimension of the sampling method, we propose the residual-driven attacking direction stabilization to alleviate the unstable adversarial optimization caused by the diversity of language references.By coarsely predicting the language-guided sampling process, the optimization process will be stabilized by the designed ResAdv-DDIM sampler, therefore releasing the transferable and robust adversarial capability of multi-step diffusion models.In task modeling, we propose the context-encoded attacking scenario constraint to supplement the missing knowledge from incomplete human instructions.Guidance masking and renderer integration are proposed to regulate the constraints of 2D/3D SemanticAE, activating stronger scenario-adapted attacks.Moreover, in the dimension of generator evaluation, we propose the semantic-abstracted attacking evaluation enhancement by clarifying the evaluation boundary based on the label taxonomy, facilitating the development of more effective SemanticAE generators.Extensive experiments demonstrate the superiority of the transfer attack performance of InSUR.Besides, it is worth highlighting that we realize the reference-free generation of semantically constrained 3D adversarial examples by utilizing language-guided 3D generation models for the first time.



Paperid:4323
Authors:Gongxu Luo, Haoyue Dai, Loka Li, Chengqian Gao, Boyang Sun, Kun Zhang
Abstract:
Gene regulatory network inference (GRNI) aims to discover how genes causally regulate each other from gene expression data. It is well-known that statistical dependencies in observed data do not necessarily imply causation, as spurious dependencies may arise fromlatent confounders, such as non-coding RNAs. Numerous GRNI methods have thus been proposed to address this confounding issue. However, dependencies may also result fromselection—only cells satisfying certain survival or inclusion criteria are observed—while these selection-induced spurious dependencies are frequently overlooked in gene expression data analyses. In this work, we show that such selection is ubiquitous and, when ignored or conflated with true regulations, can lead to flawed causal interpretation and misguided intervention recommendations. To address this challenge, a fundamental question arises: can we distinguish dependencies due to regulation, confounding, and crucially, selection? We show that gene perturbations offer a simple yet effective answer: selection-induced dependencies aresymmetricunder perturbation, while those from regulation or confounding are not. Building on this motivation, we propose GISL (Gene regulatory network Inference in the presence of Selection bias and Latent confounders), a principled algorithm that leverages perturbation data to uncover both true gene regulatory relations and non-regulatory mechanisms of selection and confounding up to the equivalence class. Experiments on synthetic and real-world gene expression data demonstrate the effectiveness of our method.



Paperid:4324
Authors:XianJun, Davin Choo, Yuqi Pan, Tonghan Wang, Milind Tambe, Alastair van Heerden, Cheryl Johnson
Abstract:
Abstract:We study a sequential decision-making problem on a $n$-node graph $\mathcal{G}$ where each node has an unknown label from a finite set $\mathbf{\Sigma}$, drawn from a joint distribution that is Markov with respect to $\mathcal{G}$. At each step, selecting a node reveals its label and yields a label-dependent reward. The goal is to adaptively choose nodes to maximize expected accumulated discounted rewards. We impose a frontier exploration constraint, where actions are limited to neighbors of previously selected nodes, reflecting practical constraints in settings such as contact tracing and robotic exploration. We design a Gittins index-based policy that applies to general graphs and is provably optimal when $\mathcal{G}$ is a forest. Our implementation runs in $\mathcal{O}(n^2 \cdot |\mathbf{\Sigma}|^2)$ time while using $\mathcal{O}(n \cdot |\mathbf{\Sigma}|^2)$ oracle calls to $\mathcal{P}$ and $\mathcal{O}(n^2 \cdot |\mathbf{\Sigma}|)$ space. Experiments on synthetic and real-world graphs show that our method consistently outperforms natural baselines, including in non-tree, budget-limited, and undiscounted settings. For example, in HIV testing simulations on real-world sexual interaction networks, our policy detects nearly all positive cases with only half the population tested, substantially outperforming other baselines.



Paperid:4325
Authors:WANTONG XIE, Yi-Xiang Hu, Jieyang Xu, Feng Wu, Xiangyang Li
Abstract:
Optimization plays a central role in operations research (OR) and numerous industrial applications, yet automating the end-to-end process of translating natural language descriptions into executable optimization programs remains a formidable challenge. While recent efforts have applied large language models (LLMs) to this task, existing approaches are hindered by high inference costs, limited robustness across domains, and weak verification mechanisms. In this work, we propose MURKA, a reinforcement learning and knowledge distillation-based framework that enhances LLM-driven optimization modeling via collaborative agent alignment. MURKA orchestrates three specialized agents---Extractor, Solver, and Checker---to achieve accurate problem understanding, robust formulation, and verifiable execution. The Extractor is trained using group relative policy optimization with a composite reward function that incorporates semantic correctness and execution fidelity. The Solver benefits from knowledge distillation from a powerful teacher model, yielding structurally valid and executable formulations in AMPL. The Checker iteratively verifies solution correctness via solver feedback. We validate MURKA's generalizability through extensive experiments across diverse OR benchmarks, demonstrating its robustness and scalability.Experimental results on eight diverse OR benchmarks, including NLP4LP, ComplexOR, and NL4Opt, demonstrate that MURKA, built on the LLaMa3-8B backbone, achieves a 5.9\% absolute improvement in solution accuracy and a 5.1\% increase in execution success rate compared to leading baselines. These results establish MURKA as an effective and scalable paradigm for LLM-driven optimization, with strong potential for deployment in real-world OR applications.



Paperid:4326
Authors:jusheng zhang, Yijia Fan, Wenjun Lin, Ruiqi Chen, Haoyi Jiang, Wenhao Chai, Jian Wang, Keze Wang
Abstract:
We proposeGAM-Agent, a game-theoretic multi-agent framework for enhancing vision-language reasoning. Unlike prior single-agent or monolithic models, GAM-Agent formulates the reasoning process as a non-zero-sum game between base agents—each specializing in visual perception subtasks—and a critical agent that verifies logic consistency and factual correctness. Agents communicate via structured claims, evidence, and uncertainty estimates. The framework introduces an uncertainty-aware controller to dynamically adjust agent collaboration, triggering multi-round debates when disagreement or ambiguity is detected. This process yields more robust and interpretable predictions. Experiments on four challenging benchmarks—MMMU, MMBench, MVBench, and V*Bench—demonstrate that GAM-Agent significantly improves performance across various VLM backbones. Notably, GAM-Agent boosts the accuracy of small-to-mid scale models (e.g., Qwen2.5-VL-7B, InternVL3-14B) by 5–6\%, and still enhances strong models like GPT-4o by up to 2–3\%. Our approach is modular, scalable, and generalizable, offering a path toward reliable and explainable multi-agent multimodal reasoning.



Paperid:4327
Authors:Haolin Yang, Hakaze Cho, Yiqiao Zhong, Naoya Inoue
Abstract:
The unusual properties of in-context learning (ICL) have prompted investigations into the internal mechanisms of large language models. Prior work typically focuses on either special attention heads or task vectors at specific layers, but lacks a unified framework linking these components to the evolution of hidden states across layers that ultimately produce the model’s output. In this paper, we propose such a framework for ICL in classification tasks by analyzing two geometric factors that govern performance: the separability and alignment of query hidden states. A fine-grained analysis of layer-wise dynamics reveals a striking two-stage mechanism—separability emerges in early layers, while alignment develops in later layers. Ablation studies further show that Previous Token Heads drive separability, while Induction Heads and task vectors enhance alignment. Our findings thus bridge the gap between attention heads and task vectors, offering a unified account of ICL’s underlying mechanisms.



Paperid:4328
Authors:Samuel Howard, Peter Potaptchik, George Deligiannidis
Abstract:
Recent advances in flow-based generative modelling have provided scalable methods for computing the Schrödinger Bridge (SB) between distributions, a dynamic form of entropy-regularised Optimal Transport (OT) for the quadratic cost. The successful Iterative Markovian Fitting (IMF) procedure solves the SB problem via sequential bridge-matching steps, presenting an elegant and practical approach with many favourable properties over the more traditional Iterative Proportional Fitting (IPF) procedure. Beyond the standard setting, optimal transport can be generalised to the multi-marginal case in which the objective is to minimise a cost defined over several marginal distributions. Of particular importance are costs defined over a tree structure, from which Wasserstein barycentres can be recovered as a special case. In this work, we extend the Iterative Markovian Fitting procedure to solve for the tree-structured SB problem. Our resulting algorithm inherits the many advantages of IMF over IPF approaches in the tree-based setting. In the specific case of Wasserstein barycentres, our approach can be viewed as extending fixed-point approaches for barycentre computation to the case of flow-based entropic OT solvers.



Paperid:4329
Authors:Florian A. Hölzl, Daniel Rueckert, Georgios Kaissis
Abstract:
Robust validation metrics remain essential in contemporary deep learning, not only to detect overfitting and poor generalization, but also to monitor training dynamics.In the supervised classification setting, we investigate whether interactions between training data and model weights can yield such a metric that both tracks generalization during training and attributes performance to individual training samples.We introduce Gradient-Weight Alignment (GWA), quantifying the coherence between per-sample gradients and model weights.We show that effective learning corresponds to coherent alignment, while misalignment indicates deteriorating generalization.GWA is efficiently computable during training and reflects both sample-specific contributions and dataset-wide learning dynamics.Extensive experiments show that GWA accurately predicts optimal early stopping, enables principled model comparisons, and identifies influential training samples, providing a validation-set-free approach for model analysis directly from the training data.



Paperid:4330
Authors:Anish Dhir, Cristiana Diaconu, Valentinian Lungu, James Requeima, Richard Turner, Mark van der Wilk
Abstract:
In scientific domains---from biology to the social sciences---many questions boil down to \textit{What effect will we observe if we intervene on a particular variable?} If the causal relationships (e.g.~a causal graph) are known, its possible to estimate the intervention distributions. In the absence of this domain knowledge, the causal structure must be discovered from the available observational data. However, observational data are often compatible with multiple causal graphs, making methods that commit to a single structure prone to overconfidence. A principled way to manage this structural uncertainty is via Bayesian inference, which averages over a posterior distribution on possible causal structures and functional mechanisms. Unfortunately, the number of causal structures grows super-exponentially with the number of nodes in the graph, making computations intractable. We propose to circumvent these challenges by using meta-learning to create an end-to-end model: the Model-Averaged Causal Estimation Transformer Neural Process (MACE-TNP). The model is trained to predict the Bayesian model-averaged interventional posterior distribution, and its end-to-end nature bypasses the need for expensive calculations. Empirically, we demonstrate that MACE-TNP outperforms strong Bayesian baselines. Our work established meta-learning as a flexible and scalable paradigm for approximating complex Bayesian causal inference, that can be scaled to increasingly challenging settings in the future.



Paperid:4331
Authors:Payal Mohapatra, Yueyuan Sui, Akash Pandey, Stephen Xia, Qi Zhu
Abstract:
From clinical healthcare to daily living, continuous sensor monitoring across multiple modalities has shown great promise for real-world intelligent decision-making but also faces various challenges. In this work, we argue for modeling such heterogeneous data sources under the multimodal paradigm and introduce a new framework, MAESTRO. We introduce MAESTRO, a novel framework that overcomes key limitations of existing multimodal learning approaches: (1) reliance on a single primary modality for alignment, (2) pairwise modeling of modalities, and (3) assumption of complete modality observations. These limitations hinder the applicability of these approaches in real-world multimodal time-series settings, where primary modality priors are often unclear, the number of modalities can be large (making pairwise modeling impractical), and sensor failures often result in arbitrary missing observations. At its core, MAESTRO facilitates dynamic intra- and cross-modal interactions based on task relevance, and leverages symbolic tokenization and adaptive attention budgeting to construct long multimodal sequences, which are processed via sparse cross-modal attention. The resulting cross-modal tokens are routed through a sparse Mixture-of-Experts (MoE) mechanism, enabling black-box specialization under varying modality combinations. We evaluate MAESTRO against 10 baselines on four diverse datasets spanning three applications, and observe average relative improvements of 4% and 8% over the best existing multimodal and multivariate approaches, respectively, under complete observations. Under partial observations—with up to 40% of missing modalities—MAESTRO achieves an average 9% improvement. Further analysis also demonstrates the robustness and efficiency of MAESTRO's sparse, modality-aware design for learning from dynamic time series.



Paperid:4332
Authors:Wei Huang, Jianshu Zhang, Leiyu Wang, Heyue Li, Luoyi Fan, Yichen Zhu, Nanyang Ye, Qinying Gu
Abstract:
Abstract:Offline reinforcement learning (offline RL) is increasingly approached as a sequence modeling task, with methods leveraging advanced architectures like Transformers to capture trajectory dependencies. Despite significant progress, the mechanisms underlying their effectiveness and limitations remain insufficiently understood. We conduct a thorough analysis on the representative Decision Transformer (DT) model using an entropy analysis and identify the inconsistencies in state-action-reward ($\langle s, a, R \rangle$) distributions causing attention ``dispersal". To address this, we propose a hierarchical framework that decomposes sequence modeling into intra-step relational modeling—handled by a Token Merger that fuses each $\langle s, a, R \rangle$ triplet—and inter-step modeling—handled by a Token Mixer across timesteps. We investigate several Token Merger designs and validate their effectiveness across various offline RL methods. Furthermore, our theoretical analysis and experimental results suggest that while Token Mixers are important, lightweight architecture can also achieve even better performance to more complex ones. We therefore propose a parameter-free Average Pooling Token Mixer, which, combined with a convolutional Token Merger, forms our final model, Decision HiFormer (DHi). DHi achieves a \textbf{73.6\%} improvement in inference speed and an \textbf{9.3\%} gain in policy performance on the D4RL benchmark compared to DT. DHi also generalizes well to real-world robotic manipulation tasks, offering both practical benefits and insights into sequence-based policy design for offline RL.



Paperid:4333
Authors:Sicheng Yang, Zhaohu Xing, Lei Zhu
Abstract:
Consistency learning with feature perturbation is a widely used strategy in semi-supervised medical image segmentation. However, many existing perturbation methods rely on dropout, and thus require a careful manual tuning of the dropout rate, which is a sensitive hyperparameter and often difficult to optimize and may lead to suboptimal regularization. To overcome this limitation, we propose VQ-seg, the first approach to employ vector quantization (VQ) to discretize the feature space and introduce a novel and controllable Quantized Perturbation Module (QPM) that replaces dropout. Our QPM perturbs discrete representations by shuffling the spatial locations of codebook indices, enabling effective and controllable regularization. To mitigate potential information loss caused by quantization, we design a dual-branch architecture where the post-quantization feature space is shared by both image reconstruction and segmentation tasks. Moreover, we introduce a Post-VQ Feature Adapter (PFA) to incorporate guidance from a foundation model (FM), supplementing the high-level semantic information lost during quantization. Furthermore, we collect a large-scale Lung Cancer (LC) dataset comprising 828 CT scans annotated for central-type lung carcinoma. Extensive experiments on the LC dataset and other public benchmarks demonstrate the effectiveness of our method, which outperforms state-of-the-art approaches. Code will be released.



Paperid:4334
Authors:Guanghao Li, Wenhao Jiang, Mingfeng Chen, Yan Li, Hao Yu, Shuting Dong, Tao Ren, Ming Tang, Chun Yuan
Abstract:
Chain-of-Thought (CoT) prompting improves the reasoning performance of large language models (LLMs) by encouraging step-by-step thinking. However, CoT-based methods depend on intermediate reasoning steps, which limits scalability and generalization. Recent work explores recursive reasoning, where LLMs reuse internal layers across iterations to refine latent representations without explicit CoT supervision. While promising, these approaches often require costly pretraining and lack a principled framework for how reasoning should evolve across iterations.We address this gap by introducingFlow Chain-of-Thought (Flow CoT), a reasoning paradigm that models recursive inference as a progressive trajectory of latent cognitive states. Flow CoT frames each iteration as a distinct cognitive stage—deepening reasoning across iterations without relying on manual supervision. To realize this, we proposeSCOUT(Stepwise Cognitive Optimization Using Teachers), a lightweight fine-tuning framework that enables Flow CoT-style reasoning without the need for pretraining. SCOUT uses progressive distillation to align each iteration with a teacher of appropriate capacity, and a cross-attention-based retrospective module that integrates outputs from previous iterations while preserving the model’s original computation flow.Experiments across eight reasoning benchmarks show that SCOUT consistently improves both accuracy and explanation quality, achieving up to 1.8\% gains under fine-tuning. Qualitative analyses further reveal that SCOUT enables progressively deeper reasoning across iterations—refining both belief formation and explanation granularity. These results not only validate the effectiveness of SCOUT, but also demonstrate the practical viability of Flow CoT as a scalable framework for enhancing reasoning in LLMs.



Paperid:4335
Authors:Enshen Zhou, Jingkun An, Cheng Chi, Yi Han, Shanyu Rong, Chi Zhang, Pengwei Wang, Zhongyuan Wang, Tiejun Huang, Lu Sheng, Shanghang Zhang
Abstract:
Spatial referring is a fundamental capability of embodied robots to interact with the 3D physical world. However, even with the powerful pretrained VLMs, recent approaches are still not qualified to accurately understand the complex 3D scenes and dynamically reason about the instruction-indicated locations for interaction. To this end, we propose RoboRefer, a 3D-aware vision language model (VLM) that can first achieve precise spatial understanding by integrating a disentangled but dedicated depth encoder via supervised fine-tuning (SFT). Moreover, RoboRefer advances generalized multi-step spatial reasoning via reinforcement fine-tuning (RFT), with metric-sensitive process reward functions tailored for spatial referring tasks. To support SFT and RFT training, we introduce RefSpatial, a large-scale dataset of 20M QA pairs (2x prior), covering 31 spatial relations (vs. 15 prior) and supporting complex reasoning processes (up to 5 steps). In addition, we introduce RefSpatial-Bench, a challenging benchmark filling the gap in evaluating spatial referring with multi-step reasoning. Experiments show that SFT-trained RoboRefer achieves state-of-the-art spatial understanding, with an average success rate of 89.6%. RFT-trained RoboRefer further outperforms all other baselines by a large margin, even surpassing Gemini-2.5-Pro by 12.4% in average accuracy on RefSpatial-Bench. Notably, RoboRefer can be integrated with various control policies to execute long-horizon, dynamic tasks across diverse robots (e,g., UR5, G1 humanoid) in cluttered real-world scenes.



Paperid:4336
Authors:Qixin Zhang, Yan Sun, Can Jin, Xikun Zhang, Yao Shu, Puning Zhao, Li Shen, Dacheng Tao
Abstract:
Abstract:In this paper, we present two effective policy learning algorithms for multi-agent online coordination(MA-OC) problem. The first one, **MA-SPL**, not only can achieve the optimal $(1-\frac{c}{e})$-approximation guarantee for the MA-OC problem with submodular objectives but also can handle the unexplored $\alpha$-weakly DR-submodular and $(\gamma,\beta)$-weakly submodular scenarios, where $c$ is the curvature of the investigated submodular functions, $\alpha$ denotes the diminishing-return(DR) ratio and the tuple$(\gamma,\beta)$ represents the submodularity ratios. Subsequently, in order to reduce the reliance on the unknown parameters $\alpha,\gamma,\beta$ inherent in the **MA-SPL** algorithm, we then introduce the second online algorithm named **MA-MPL**. This **MA-MPL** algorithm is entirely *parameter-free* and simultaneously can maintain the same approximation ratio as the first **MA-SPL** algorithm. The core of our **MA-SPL** and **MA-MPL** algorithms is a novel continuous-relaxation technique term as policy-based continuous extension. Compared with the well-established multi-linear extension, a notable advantage of this new policy-based continuous extension is its ability to provide a lossless rounding scheme for any set function, thereby enabling us to tackle the challenging weakly submodular objective functions. Finally, extensive simulations are conducted to demonstrate the effectiveness of our proposed algorithms.



Paperid:4337
Authors:Shohei Ohsawa
Abstract:
We introduce Support Vector Generation (SVG), a kernel-based framework that converts a frozen language model into an interpretable, training-free classifier for zero- and few-shot learning. SVG operates by combining Metropolis–Hastings sampling with support vector machine optimization in the reproducing kernel Hilbert space (RKHS) induced by the language model's embedding. Each classification decision is based on a weighted combination of at most 32 natural-language sentences, which serve as explicit support vectors and provide faithful rationales. Our theoretical analysis proves that SVG minimizes the empirical hinge loss over the span of the supports and admits a generalization bound independent of the language model size. Experiments on the GLUE benchmark show that SVG matches or surpasses prompting-based zero-shot baselines in accuracy across multiple tasks—without any fine-tuning or GPU acceleration. Notably, our CPU-only implementation completes training in under three minutes per task, and maintains competitive inference speed. These results suggest that SVG offers a viable path toward efficient, interpretable NLP systems under compute constraints.



Paperid:4338
Authors:Hongjin Qian, Zheng Liu, Chao Gao, Yankai Wang, Defu Lian, Zhicheng Dou
Abstract:
In real-world information-seeking scenarios, users have dynamic and diverse needs, requiring RAG systems to demonstrate adaptable resilience. To comprehensively evaluate the resilience of current RAG methods, we introduce HawkBench, a human-labeled, multi-domain benchmark designed to rigorously assess RAG performance across categorized task types. By stratifying tasks based on information-seeking behaviors, HawkBench provides a systematic evaluation of how well RAG systems adapt to diverse user needs.Unlike existing benchmarks, which focus primarily on specific task types (mostly factoid queries) and rely on varying knowledge bases, HawkBench offers: (1) systematic task stratification to cover a broad range of query types, including both factoid and rationale queries, (2) integration of multi-domain corpora across all task types to mitigate corpus bias, and (3) rigorous annotation for high-quality evaluation.HawkBench includes 1,600 high-quality test samples, evenly distributed across domains and task types. Using this benchmark, we evaluate representative RAG methods, analyzing their performance in terms of answer quality and response latency. Our findings highlight the need for dynamic task strategies that integrate decision-making, query interpretation, and global knowledge understanding to improve RAG generalizability. We believe HawkBench serves as a pivotal benchmark for advancing the resilience of RAG methods and their ability to achieve general-purpose information seeking.



Paperid:4339
Authors:Zuhair Hasan Shaik, Abdullah Mazhar, Aseem Srivastava, Md Shad Akhtar
Abstract:
Large Language Models have demonstrated impressive fluency across diverse tasks, yet their tendency to produce toxic content remains a critical challenge for AI safety and public trust. Existing toxicity mitigation approaches primarily manipulate individual neuron activations, but these methods suffer from instability, context dependence, and often compromise the model’s core language abilities. To address these shortcomings, we investigate three key questions: the stability of neuron-level toxicity indicators, the advantages of structural (layer-wise) representations, and the interpretability of mechanisms driving toxic generation. Through extensive experiments on Jigsaw and ToxiCN datasets, we show that aggregated layer-wise features provide more robust signals than single neurons. Moreover, we observe conceptual limitations in prior works that conflate toxicity detection experts and generation experts within neuron-based interventions. To mitigate this, we propose a novel principled intervention technique, EigenShift, based on eigen-decomposition of the language model’s final output layer. This method selectively targets generation-aligned components, enabling precise toxicity suppression without impairing linguistic competence. Our method requires no additional training or fine-tuning, incurs minimal computational cost, and is grounded in rigorous theoretical analysis.



Paperid:4340
Authors:Yuyan Chen, Nico Lang, B. Schmidt, Aditya Jain, Yves Basset, Sara Beery, Maxim Larrivee, David Rolnick
Abstract:
Global biodiversity is declining at an unprecedented rate, yet little information isknown about most species and how their populations are changing. Indeed, some90% Earth’s species are estimated to be completely unknown. Machine learning hasrecently emerged as a promising tool to facilitate long-term, large-scale biodiversitymonitoring, including algorithms for fine-grained classification of species fromimages. However, such algorithms typically are not designed to detect examplesfrom categories unseen during training – the problem of open-set recognition(OSR) – limiting their applicability for highly diverse, poorly studied taxa such asinsects. To address this gap, we introduce Open-Insect, a large-scale, fine-graineddataset to evaluate unknown species detection across different geographic regionswith varying difficulty. We benchmark 38 OSR algorithms across three categories:post-hoc, training-time regularization, and training with auxiliary data, finding thatsimple post-hoc approaches remain a strong baseline. We also demonstrate how toleverage auxiliary data to improve species discovery in regions with limited data.Our results provide timely insights to guide the development of computer visionmethods for biodiversity monitoring and species discovery.



Paperid:4341
Authors:Jinyu Cai, Yuan Xie, Glynnis Lim, Yifang Yin, Roger Zimmermann, See-Kiong Ng
Abstract:
Detecting anomalies in multivariate time-series data is an essential task across various domains, yet there are unresolved challenges such as (1) severe class imbalance between normal and anomalous data due to rare anomaly availability in the real world; (2) limited adaptability of the static graph-based methods to dynamically changing inter-variable correlations; and (3) neglect of subtle anomalies due to the overfitting of normal patterns in reconstruction-based methods. To tackle these issues, we propose Self-Perturbed Anomaly-Aware Graph Dynamics (SPAGD), a framework for time-series anomaly detection. SPAGD employs a self-perturbation module that generates self-perturbed time series from the reconstruction process of normal ones, which will provide auxiliary signals to alleviate class imbalance during training. Concurrently, an anomaly-aware graph construction module is proposed to dynamically adjust the graph structure by leveraging the reconstruction residuals of self-perturbed time series, thereby emphasizing the inter-variable disruptions induced by anomalous candidates. A unified spatio-temporal anomaly detection module then integrates both spatial and temporal convolutions to train a classifier that distinguishes the normal time series from the auxiliary self-perturbed samples. Extensive experiments across multiple benchmark datasets demonstrate the effectiveness of SPAGD compared to state-of-the-art baselines.



Paperid:4342
Authors:Kaixiang Yang, Xin Li, Yuxi Li, Qiang Li, Zhiwei Wang
Abstract:
Latent Diffusion-based Text-to-Image (T2I) is a free image editing tool that typically reverses an image into noise, reconstructs it using its original text prompt, and then generates an edited version under a new target prompt. To preserve unaltered image content, features from the reconstruction are directly injected to replace selected features in the generation.However, this direct replacement often leads to feature incompatibility, compromising editing fidelity and limiting creative flexibility, particularly for non-rigid edits (\emph{e.g.}, structural or pose changes).In this paper, we aim to address these limitations by proposing \textbf{FSI-Edit}, a novel framework using frequency- and stochasticity-based feature injection for flexible image editing.First, FSI-Edit enhances feature consistency by injecting \emph{high-frequency} components of reconstruction features into generation features, mitigating incompatibility while preserving the editing ability for major structures encoded in low-frequency information.Second, it introduces controlled \emph{noise} into the replaced reconstruction features, expanding the generative space to enable diverse non-rigid edits beyond the original image’s constraints.Experiments on non-rigid edits, \emph{e.g.}, addition, deletion, and pose manipulation, demonstrate that FSI-Edit outperforms existing baselines in target alignment, semantic fidelity and visual quality. Our work highlights the critical roles of frequency-aware design and stochasticity in overcoming rigidity in diffusion-based editing.



Paperid:4343
Authors:Xiang Liu, Zhenheng Tang, Peijie Dong, Zeyu Li, Liuyue, Bo Li, Xuming Hu, Xiaowen Chu
Abstract:
Large Language Models (LLMs) require significant GPU memory when processing long texts, with the key value (KV) cache consuming up to 70\% of total memory during inference. Although existing compression methods reduce memory by evaluating the importance of individual tokens, they overlook critical semantic relationships between tokens, resulting in fragmented context and degraded performance. We introduce ChunkKV, which fundamentally reimagines KV cache compression by treating semantic chunks - rather than isolated tokens - as basic compression units. This approach preserves complete linguistic structures and contextual integrity, ensuring that essential meaning is retained even under aggressive compression. Our innovation includes a novel layer-wise index reuse technique that exploits the higher cross-layer similarity of preserved indices in ChunkKV, reducing computational overhead and improving throughput by 26.5\%. Comprehensive evaluations on challenging benchmarks: LongBench, Needle-In-A-HayStack, GSM8K, and JailbreakV demonstrate that ChunkKV outperforms state-of-the-art methods by up to 8.7\% in precision while maintaining the same compression ratio. These results confirm that semantic-aware compression significantly enhances both efficiency and performance for long-context LLM inference, providing a simple yet effective solution to the memory bottleneck problem.



Paperid:4344
Authors:Gucongcong Fan, Chaoyue Niu, Chengfei Lyu, Fan Wu, Guihai Chen
Abstract:
Abstract:Mobile agents rely on Large Language Models (LLMs) to plan and execute tasks on smartphone user interfaces (UIs). While cloud-based LLMs achieve high task accuracy, they require uploading the full UI state at every step, exposing unnecessary and often irrelevant information. In contrast, local LLMs avoid UI uploads but suffer from limited capacity, resulting in lower task success rates. We propose $\textbf{CORE}$, a $\textbf{CO}$llaborative framework that combines the strengths of cloud and local LLMs to $\textbf{R}$educe UI $\textbf{E}$xposure, while maintaining task accuracy for mobile agents. CORE comprises three key components: (1) $\textbf{Layout-aware block partitioning}$, which groups semantically related UI elements based on the XML screen hierarchy; (2) $\textbf{Co-planning}$, where local and cloud LLMs collaboratively identify the current sub-task; and (3) $\textbf{Co-decision-making}$, where the local LLM ranks relevant UI blocks, and the cloud LLM selects specific UI elements within the top-ranked block. CORE further introduces a multi-round accumulation mechanism to mitigate local misjudgment or limited context. Experiments across diverse mobile apps and tasks show that CORE reduces UI exposure by up to 55.6\%, while maintaining task success within 5\% below cloud-only agents and significantly outperforming local-only baselines by up to 46.85\%.



Paperid:4345
Authors:Huadai Liu, Jialei Wang, Kaicheng Luo, Wen Wang, Qian Chen, Zhou Zhao, Wei Xue
Abstract:
While end-to-end video-to-audio generation has greatly improved, producing high-fidelity audio that authentically captures the nuances of visual content remains challenging. Like professionals in the creative industries, such generation requires sophisticated reasoning about items such as visual dynamics, acoustic environments, and temporal relationships. We present \textbf{ThinkSound}, a novel framework that leverages Chain-of-Thought (CoT) reasoning to enable stepwise, interactive audio generation and editing for videos. Our approach decomposes the process into three complementary stages: foundational foley generation that creates semantically coherent soundscapes, interactive object-centric refinement through precise user interactions, and targeted editing guided by natural language instructions. At each stage, a multimodal large language model generates contextually aligned CoT reasoning that guides a unified audio foundation model. Furthermore, we introduce \textbf{AudioCoT}, a comprehensive dataset with structured reasoning annotations that establishes connections between visual content, textual descriptions, and sound synthesis. Experiments demonstrate that ThinkSound achieves state-of-the-art performance in video-to-audio generation across both audio metrics and CoT metrics and excels in out-of-distribution Movie Gen Audio benchmark. The demo page is available at~\url{https://ThinkSound-Demo.github.io}



Paperid:4346
Authors:Jiang Lin, Xinyu Chen, Zhiqiu Zhang, Jizhi Zhang, Ye Wang, Qiang Tang, Qian Wang, Song Wu, Jian Yang, Zili Yi
Abstract:
Controlling the spatial and semantic structure of diffusion-generated images remains a challenge. Existing methods like ControlNet rely on handcrafted condition maps and retraining, limiting flexibility and generalization. Inversion-based approaches offer stronger alignment but incur high inference cost due to dual-path denoising.We present \textbf{FreeControl}, a training-free framework for semantic structural control in diffusion models. Unlike prior methods that extract attention across multiple timesteps, FreeControl performs \textit{one-step attention extraction} from a single, optimally chosen timestep and reuses it throughout denoising. This enables efficient structural guidance without inversion or retraining.To further improve quality and stability, we introduce \textit{Latent-Condition Decoupling (LCD)}: a principled separation of the timestep condition and the noised latent used in attention extraction. LCD provides finer control over attention quality and eliminates structural artifacts. FreeControl also supports compositional control via reference images assembled from multiple sources, enabling intuitive scene layout design and stronger prompt alignment. FreeControl introduces a new paradigm for test-time control—enabling structurally and semantically aligned, visually coherent generation directly from raw images, with the flexibility for intuitive compositional design and compatibility with modern diffusion models at ~5\% additional cost.



Paperid:4347
Authors:ZAIXI ZHANG, Ruofan Jin, Le Cong, Mengdi Wang
Abstract:
DNA language models have revolutionized our ability to design and manipulate DNA sequences—the fundamental language of life—with unprecedented precision, enabling transformative applications in therapeutics, synthetic biology, and gene editing. However, this capability also poses significant dual-use risks, including the potential creation of harmful biological agents. To address these biosecurity challenges, we introduce two innovative watermarking techniques: DNAMark and CentralMark. DNAMark employs synonymous codon substitutions to embed robust watermarks in DNA sequences while preserving the function of encoded proteins. CentralMark advances this by creating inheritable watermarks that transfer from DNA to translated proteins, leveraging protein embeddings to ensure detection across the central dogma. Both methods utilize state-of-the-art embeddings to generate watermark logits, enhancing resilience against natural mutations, synthesis errors, and adversarial attacks. Evaluated on a therapeutic DNA benchmark, DNAMark and CentralMark achieve F1 detection scores above 0.85 under diverse conditions, while maintaining over 60\% sequence similarity to ground truth and degeneracy scores below 15\%. A case study on a CRISPR-Cas9 system underscores CentralMark’s utility in real-world synthetic biology. This work establishes a vital framework for securing DNA language models, balancing innovation with accountability to mitigate biosecurity risks.



Paperid:4348
Authors:Jiaxi Cheng, Yuliang Xu, Shoupeng Wang, Tao Ma, Yuchen He, Jinghe Zhang, Sihang Cai, Jiawei Zhen, Jingyi Jia, Yao Wan, Yan Xia, Zhou Zhao
Abstract:
Industrial Anomaly Detection (IAD) is an indispensable quality control technology in modern production processes. Recently, on account of the outstanding visual comprehension and cross-domain knowledge transfer capabilities of multimodal large language models (MLLMs), existing studies have explored the application of MLLMs in the IAD domain and established some multimodal IAD datasets. However, although the latest datasets contain various fundamental IAD tasks, they formulate tasks in a general question-and-answer format lacking a rigorous reasoning process, and they are relatively limited in the diversity of scenarios, which restricts their reliability in practical applications. In this paper, we propose AnomalyCoT, a multimodal Chain-of-Thought (CoT) dataset for multi-scenario IAD tasks. It consists of 37,565 IAD samples with the CoT data and is defined by challenging composite IAD tasks. Meanwhile, the CoT data for each sample provides precise coordinates of anomaly regions, thereby improving visual comprehension of defects across different types. AnomalyCoT is constructed through a systematic pipeline and involves multiple manual operations. Based on AnomalyCoT, we conducted a comprehensive evaluation of various mainstream MLLMs and fine-tuned representative models in different ways. The final results show that Gemini-2.0-flash achieved the best performance in the direct evaluation with an accuracy rate of 59.6\%, while Llama 3.2-Vision achieves the best performance after LoRA fine-tuning with an accuracy rate of 94.0\%. Among all the fine-tuned models, the average accuracy improvement reaches 36.5\%, demonstrating the potential of integrating CoT datasets in future applications within the IAD field. The code and data are available at \url{https://github.com/Zhaolutuan/AnomalyCoT}.



Paperid:4349
Authors:Zherui Yang, Zhehao Li, Kangbo Lyu, Yixuan Li, Tao Du, Ligang Liu
Abstract:
Abstract:The conjugate gradient solver (CG) is a prevalent method for solving symmetric and positive definite linear systems $\mathbf{Ax} = \mathbf{b}$, where effective preconditioners are crucial for fast convergence. Traditional preconditioners rely on prescribed algorithms to offer rigorous theoretical guarantees, while limiting their ability to exploit optimization from data. Existing learning-based methods often utilize Graph Neural Networks (GNNs) to improve the performance and speed up the construction. However, their reliance on incomplete factorization leads to significant challenges: the associated triangular solve hinders GPU parallelization in practice, and introduces long-range dependencies which are difficult for GNNs to model. To address these issues, we propose a learning-based method to generate GPU-friendly preconditioners, particularly using GNNs to construct Sparse Approximate Inverse (SPAI) preconditioners, which avoids triangular solves and requires only two matrix-vector products at each CG step. The locality of matrix-vector product is compatible with the local propagation mechanism of GNNs. The flexibility of GNNs also allows our approach to be applied in a wide range of scenarios. Furthermore, we introduce a statistics-based scale-invariant loss function. Its design matches CG's property that the convergence rate depends on the condition number, rather than the absolute scale of $\mathbf{A}$, leading to improved performance of the learned preconditioner. Evaluations on three PDE-derived datasets and one synthetic dataset demonstrate that our method outperforms standard preconditioners (Diagonal, IC, and traditional SPAI) and previous learning-based preconditioners on GPUs. We reduce solution time on GPUs by 40%-53% (68%-113% faster), along with better condition numbers and superior generalization performance.



Paperid:4350
Authors:yuyang Hong, Jiaqi Gu, Yang Qi, Lubin Fan, Yue Wu, Ying Wang, Kun Ding, SHIMING XIANG, Jieping Ye
Abstract:
The task of Knowlegde-Based Visual Question Answering (KB-VQA) requires the model to understand visual features and retrieve external knowledge. Retrieval-Augmented Generation (RAG) have been employed to address this problem through knowledge base querying. However, existing work demonstrate two limitations: insufficient interactivity during knowledge retrieval and ineffective organization of retrieved information for Visual-Language Model (VLM). To address these challenges, we propose a three-stage visual language model with Process, Retrieve and Filter (VLM-PRF) framework. For interactive retrieval, VLM-PRF uses reinforcement learning (RL) to guide the model to strategically process information via tool-driven operations. For knowledge filtering, our method trains the VLM to transform the raw retrieved information into into task-specific knowledge. With a dual reward as supervisory signals, VLM-PRF successfully enable model to optimize retrieval strategies and answer generation capabilities simultaneously. Experiments on two datasets demonstrate the effectiveness of our framework.



Paperid:4351
Authors:Chenxu Niu, Jie Zhang, Yanbing Liu, Yunpeng Li, Jinta Weng, Yue Hu
Abstract:
Backdoor attacks pose a significant threat to large language models (LLMs) by embedding malicious triggers that manipulate model behavior. However, existing defenses primarily rely on prior knowledge of backdoor triggers or targets and offer only superficial mitigation strategies, thus struggling to fundamentally address the inherent reliance on unreliable features. To address these limitations, we propose a novel defense strategy, \textit{RepGuard}, that strengthens LLM resilience by adaptively separating abnormal features from useful semantic representations, rendering the defense agnostic to specific trigger patterns. Specifically, we first introduce a dual-perspective feature localization strategy that integrates local consistency and sample-wise deviation metrics to identify suspicious backdoor patterns. Based on this identification, an adaptive mask generation mechanism is applied to isolate backdoor-targeted shortcut features by decomposing hidden representations into independent spaces, while preserving task-relevant semantics. With a multi-objective optimization framework, our method can inherently mitigates backdoor attacks. Across \textit{Target Refusal} and \textit{Jailbreak} tasks under four types of attacks, RepGuard consistently reduced the attack success rate on poisoned data by nearly 80\% on average, while maintaining near-original task performance on clean data. Extensive experiments demonstrate that RepGuard provides a scalable and interpretable solution for safeguarding LLMs against sophisticated backdoor threats.



Paperid:4352
Authors:Junhao Xia, Haotian Zhu, Shuchao Pang, Zhigang Lu, Bing Li, Yongbin Zhou, Minhui Xue
Abstract:
Large Vision-Language Models (LVLMs) have demonstrated impressive capabilities in tasks requiring multimodal understanding. However, recent studies indicate that LVLMs are more vulnerable than LLMs to unsafe inputs and prone to generating harmful content. Existing defense strategies primarily include fine-tuning, input sanitization, and output intervention. Although these approaches provide a certain level of protection, they tend to be resource-intensive and struggle to effectively counter sophisticated attack techniques. In this work, we propose One-head Defense (Oh Defense), a novel yet simple approach utilizing LVLMs' internal safety capabilities. Through systematic analysis of attention mechanisms, we discover that LVLMs' safety capabilities are concentrated within specific attention heads that respond differently to safe or unsafe inputs. Further exploration reveals that a single critical attention head can effectively serve as a safety guard, providing a strong discriminative signal that amplifies the model's inherent safety capabilities. The Oh Defense requires no additional training or external modules, making it computationally efficient while effectively reactivating suppressed safety mechanisms. Extensive experiments across diverse LVLM architectures and unsafe datasets validate our approach. The Oh Defense achieves near-perfect defense success rates (>98%) for unsafe inputs while maintaining low false positive rates (<5%) for safe content.



Paperid:4353
Authors:Yilun Zhao, Kaiyan Zhang, Tiansheng Hu, Sihong Wu, Ronan Le Bras, Yixin Liu, Robert Tang, Joseph Chee Chang, Jesse Dodge, Jonathan Bragg, Chen Zhao, Hanna Hajishirzi, Doug Downey, Arman Cohan
Abstract:
We present SciArena, an open and collaborative platform for evaluating foundation models on scientific literature tasks. Unlike traditional benchmarks for scientific literature understanding and synthesis, SciArena engages the research community directly, following the Chatbot Arena evaluation approach of community voting on model comparisons, while tailoring the evaluation to open-ended scientific tasks that often require literature-grounded, long-form answers. By leveraging this collective intelligence, SciArena offers a more democratic assessment of model performance. The platform currently supports 15 open-source and proprietary foundation models and has collected over 8,000 votes from trusted researchers across diverse scientific domains over its first three months of operation. We analyze the collected data so far and confirm that the submitted questions are diverse, aligned with real-world literature needs, and that participating researchers demonstrate strong self-consistency and inter-annotator agreement in their pairwise evaluations. We discuss the results and insights based on the model ranking leaderboard. To further promote research in building model-based automated evaluation systems for literature tasks, we release SciArena-Eval, a meta-evaluation benchmark based on our collected preference data. The benchmark measures the accuracy of LLMs in judging answer quality by comparing their pairwise assessments with human votes. Our experiments highlight the benchmark’s challenges and emphasize the need for more reliable automated evaluation methods.



Paperid:4354
Authors:Shih-Hsin Wang, Yuhao Huang, Taos Transue, Justin Baker, Jonathan Forstater, Thomas Strohmer, Bao Wang
Abstract:
Abstract:Graph neural networks (GNNs) have emerged as powerful tools for learning protein structures by capturing spatial relationships at the residue level. However, existing GNN-based methods often face challenges in learning multiscale representations and modeling long-range dependencies efficiently. In this work, we propose an efficient multiscale graph-based learning framework tailored to proteins. Our proposed framework contains two crucial components: (1) It constructs a hierarchical graph representation comprising a collection of fine-grained subgraphs, each corresponding to a secondary structure motif (e.g., $\alpha$-helices, $\beta$-strands, loops), and a single coarse-grained graph that connects these motifs based on their spatial arrangement and relative orientation. (2) It employs two GNNs for feature learning: the first operates within individual secondary motifs to capture local interactions, and the second models higher-level structural relationships across motifs. Our modular framework allows a flexible choice of GNN in each stage. Theoretically, we show that our hierarchical framework preserves the desired maximal expressiveness, ensuring no loss of critical structural information. Empirically, we demonstrate that integrating baseline GNNs into our multiscale framework remarkably improves prediction accuracy and reduces computational cost across various benchmarks.



Paperid:4355
Authors:Kai Yan, Zhan Ling, Kang Liu, Yifan Yang, Ting-Han Fan, Lingfeng Shen, Zhengyin Du, Jiecao Chen
Abstract:
The ability to recognize patterns from examples and apply on new ones is a primal ability for general intelligence and widely studied by psychology and AI researchers. Many benchmarks have been proposed to measure such ability for Large Language Models (LLMs); however, they focus on few-shot (usually <10) setting and lack evaluation for aggregating many pieces of information from long contexts. On the other hand, the ever-growing context length of LLMs have brought forth the novel paradigm of many-shot In-Context Learning (ICL), which addresses new tasks with hundreds to thousands of examples without expensive and inefficient fine-tuning. However, many-shot evaluations often focus on classification, and popular long-context LLM tasks such as Needle-In-A-Haystack (NIAH) seldom require complicated intelligence for integrating many pieces of information. To fix the issues from both worlds, we propose MIR-Bench, the first many-shot in-context reasoning benchmark for pattern recognition that asks LLM to predict output via input-output examples from underlying functions with diverse data format. Based on MIR-Bench, we study many novel problems for many-shot in-context reasoning, and acquired many insightful findings including scaling effect, robustness, inductive vs. transductive reasoning, retrieval Augmented Generation (RAG), coding for inductive reasoning, cross-domain generalizability, etc.



Paperid:4356
Authors:Yichen Wang, Hangtao Zhang, Hewen Pan, Ziqi Zhou, Xianlong Wang, Peijin Guo, Lulu Xue, Shengshan Hu, Minghui Li, Leo Yu Zhang
Abstract:
Vision-Language Models (VLMs), with their strong reasoning and planning capabilities, are widely used in embodied decision-making (EDM) tasks such as autonomous driving and robotic manipulation. Recent research has increasingly explored adversarial attacks on VLMs to reveal their vulnerabilities. However, these attacks either rely on overly strong assumptions, requiring full knowledge of the victim VLM, which is impractical for attacking VLM-based EDM systems, or exhibit limited effectiveness. The latter stems from disrupting most semantic information in the image, which leads to a misalignment between the perception and the task context defined by system prompts. This inconsistency interrupts the VLM's reasoning process, resulting in invalid outputs that fail to affect interactions in the physical world. To this end, we propose a fine-grained adversarial attack framework, AdvEDM, which modifies the VLM's perception of only a few key objects while preserving the semantics of the remaining regions. This attack effectively reduces conflicts with the task context, making VLMs output valid but incorrect decisions and affecting the actions of entities, thus posing a more substantial safety threat in the physical world. We design two variants of based on this framework, AdvEDM-R and AdvEDM-A, which respectively remove the semantics of a specific object from the image and add the semantics of a new object into the image. The experimental results in both general scenarios and EDM tasks demonstrate fine-grained control and excellent attack performance.



Paperid:4357
Authors:Yantai Yang, Yuhao Wang, Zichen Wen, Luo Zhongwei, Chang Zou, Zhipeng Zhang, Chuan Wen, Linfeng Zhang
Abstract:
Abstract:Vision-Language-Action (VLA) models, particularly diffusion-based architectures, demonstrate transformative potential for embodied intelligence but are severely hampered by high computational and memory demands stemming from extensive inherent and inference-time redundancies. While existing acceleration efforts often target isolated inefficiencies, such piecemeal solutions typically fail to holistically address the varied computational and memory bottlenecks across the entire VLA pipeline, thereby limiting practical deployability. We introduce VLA-Pruner, a structured and training-free inference acceleration framework that systematically eliminates these barriers by cohesively exploiting multifaceted redundancies. VLA-Pruner synergistically integrates three targeted strategies: (1) pruning of functionally inconsequential layers from the language module, guided by an analysis of inter-layer redundancies; (2) optimizing the visual processing pathway through a task-aware strategy that selects a compact, diverse set of visual tokens, balancing task-criticality with informational coverage; and (3) alleviating temporal computational redundancy within the iterative diffusion-based action head by strategically caching and reusing key intermediate features.We apply our method to a standard VLA model CogACT, yielding a $1.93\times$ inference speedup and reduces FLOPs to $28.9\%$, with only a $0.6\%$ success rate drop in the SIMPLER benchmark. The code will be open-sourced and is available in the supplementary materials.



Paperid:4358
Authors:Min-Hsuan Yeh, Sharon Li
Abstract:
Human feedback plays a pivotal role in aligning large language models (LLMs) with human preferences. However, such feedback is often noisy or inconsistent, which can degrade the quality of reward models and hinder alignment. While various automated data cleaning methods have been proposed to mitigate this issue, a systematic evaluation of their effectiveness and generalizability remains lacking. To bridge this gap, we introduce the first comprehensive benchmark for evaluating 13 preference data cleaning methods in the context of LLM alignment. Our framework offers a standardized protocol to assess cleaning strategies in terms of alignment performance and generalizability across diverse datasets, model architectures, and optimization algorithms. By unifying disparate methods and rigorously comparing them, we uncover key factors that determine the success of data cleaning in alignment tasks. This benchmark lays the groundwork for principled and reproducible approaches to improving LLM alignment through better data quality—highlighting the crucial but underexplored role of data preprocessing in responsible AI development. We release modular implementations of all methods to catalyze further research.



Paperid:4359
Authors:Ahmed Masry, Juan Rodriguez, Tianyu Zhang, Suyuchen Wang, Chao Wang, Aarash Feizi, Akshay Kalkunte Suresh, Abhay Puri, Xiangru Jian, Pierre-André Noël, Sathwik Tejaswi Madhusudhan, Marco Pedersoli, Bang Liu, Nicolas Chapados, Yoshua Bengio, Enamul Hoque, Chris Pal, Issam Hadj Laradji, David Vazquez, Perouz Taslakian, Spandana Gella, Sai Rajeswar Mudumba
Abstract:
Aligning visual features with language embeddings is a key challenge in vision-language models (VLMs). The performance of such models hinges on having a good connector that maps visual features generated by a vision encoder to a shared embedding space with the LLM while preserving semantic similarity. Existing connectors, such as multilayer perceptrons (MLPs), often produce out-of-distribution or noisy inputs, leading to misalignment between the modalities. In this work, we propose a novel vision-text alignment method, AlignVLM, that maps visual features to a weighted average of LLM text embeddings. Our approach leverages the linguistic priors encoded by the LLM to ensure that visual features are mapped to regions of the space that the LLM can effectively interpret. AlignVLM is particularly effective for document understanding tasks, where scanned document images must be accurately mapped to their textual content. Our extensive experiments show that AlignVLM achieves state-of-the-art performance compared to prior alignment methods. We provide further analysis demonstrating improved vision-text feature alignment and robustness to noise.



Paperid:4360
Authors:Yiran Guo, Lijie Xu, Jie Liu, DAN YE, Shuang Qiu
Abstract:
Abstract:Enhancing the reasoning capabilities of large language models effectively using reinforcement learning (RL) remains a crucial challenge. Existing approaches primarily adopt two contrasting advantage estimation granularities: Token-level methods (e.g., PPO) aim to provide the fine-grained advantage signals but suffer from inaccurate estimation due to difficulties in training an accurate critic model. On the other extreme, trajectory-level methods (e.g., GRPO) solely rely on a coarse-grained advantage signal from the final reward, leading to imprecise credit assignment. To address these limitations, we propose Segment Policy Optimization (SPO), a novel RL framework that leverages segment-level advantage estimation at an intermediate granularity, achieving a better balance by offering more precise credit assignment than trajectory-level methods and requiring fewer estimation points than token-level methods, enabling accurate advantage estimation based on Monte Carlo (MC) without a critic model. SPO features three components with novel strategies: (1) flexible segment partition; (2) accurate segment advantage estimation; and (3) policy optimization using segment advantages, including a novel probability-mask strategy. We further instantiate SPO for two specific scenarios: (1) SPO-chain for short chain-of-thought (CoT), featuring novel probability-guided partition and chain-based advantage estimation, achieving $6$-$12$ percentage point improvements in accuracy over PPO and GRPO on GSM8K. (2) SPO-tree for long CoT, featuring novel tree-based advantage estimation, which significantly reduces the cost of MC estimation, achieving $7$-$11$ percentage point improvements over GRPO on MATH500 under 2K and 4K context evaluation.



Paperid:4361
Authors:Xingyu Ren, Pengwei Liu, Pengkai Wang, Guanyu Chen, Qinxin Wu, Dong Ni
Abstract:
Deep neural networks, particularly neural operators, provide an efficient alternative to costly simulations in surrogate modeling. However, their performance is often constrained by the need for large-scale labeled datasets, which are costly and challenging to acquire in many scientific domains.Semi-supervised learning reduces label reliance by leveraging unlabeled data yet remains vulnerable to noisy pseudo-labels that mislead training and undermine robustness.To address these challenges, we propose a novel framework, Uncertainty-Informed Meta Pseudo Labeling (UMPL).The core mechenism is to refine pseudo-label quality through uncertainty-informed feedback signals. Specifically, the teacher model generates pseudo labels via epistemic uncertainty, while the student model learns from these labels and provides feedback based on aleatoric uncertainty.This interplay forms a meta-learning loop where enhanced generalization and improved pseudo-label quality reinforce each other, enabling the student model to achieve more stable uncertainty estimation and leading to more robust training.Notably, This framework is model-agnostic and can be seamlessly integrated into various neural architectures, facilitating effective exploitation of unlabeled data to enhance generalization in distribution shifts and out-of-distribution scenarios.Extensive evaluations of four models across seven tasks covering steady state and transient prediction problems demonstrate that UMPL consistently outperforms the best existing semi-supervised regression methods. When using only 10% of the fully supervised training data, UMPL achieves a 14.18% improvement, highlighting its strong effectiveness under limited supervision. Our codes are available at https://anonymous.4open.science/r/UMPL-E25B.



Paperid:4362
Authors:Pengcheng Huang, Zhenghao Liu, Yukun Yan, Haiyan Zhao, Xiaoyuan Yi, Hao Chen, Zhiyuan Liu, Maosong Sun, Tong Xiao, Ge Yu, Chenyan Xiong
Abstract:
Large language models (LLMs) integrated with retrieval-augmented generation (RAG) have improved factuality by grounding outputs in external evidence. However, they remain susceptible to unfaithful generation, where outputs contradict retrieved context despite its relevance and accuracy. Existing approaches aiming to improve faithfulness primarily focus on enhancing the utilization of external context, but often overlook the persistent influence of internal parametric knowledge during generation. In this work, we investigate the internal mechanisms behind unfaithful generation and identify a subset of mid-to-deep feed-forward networks (FFNs) that are disproportionately activated in such cases. Building on this insight, we propose Parametric Knowledge Muting through FFN Suppression (ParamMute), a framework that improves contextual faithfulness by suppressing the activation of unfaithfulness-associated FFNs and calibrating the model toward retrieved knowledge. To evaluate our approach, we introduce CoFaithfulQA, a benchmark specifically designed to evaluate faithfulness in scenarios where internal knowledge conflicts with accurate external evidence. Experimental results show that ParamMute significantly enhances faithfulness across both CoFaithfulQA and the established ConFiQA benchmark, achieving substantial reductions in reliance on parametric memory. These findings underscore the importance of mitigating internal knowledge dominance and provide a new direction for improving LLM trustworthiness in RAG. All code will be released via GitHub.



Paperid:4363
Authors:Simone Ricci, Niccolò Biondi, Federico Pernici, Ioannis Patras, Alberto Del Bimbo
Abstract:
Abstract:Retrieval systems rely on representations learned by increasingly powerful models. However, due to the high training cost and inconsistencies in learned representations, there is significant interest in facilitating communication between representations and ensuring compatibility across independently trained neural networks.In the literature, two primary approaches are commonly used to adapt different learned representations: affine transformations, which adapt well to specific distributions but can significantly alter the original representation, and orthogonal transformations, which preserve the original structure with strict geometric constraints but limit adaptability. A key challenge is adapting the latent spaces of updated models to align with those of previous models on downstream distributions while preserving the newly learned representation spaces.In this paper, we impose a relaxed orthogonality constraint, namely $\lambda$-orthogonality regularization, while learning an affine transformation, to obtain distribution-specific adaptation while retaining the original learned representations.Extensive experiments across various architectures and datasets validate our approach, demonstrating that it preserves the model's zero-shot performance and ensures compatibility across model updates.



Paperid:4364
Authors:Zhen Yang, Ziwei Du, Minghan Zhang, Wei Du, Jie Chen, Fulan Qian, Shu Zhao
Abstract:
Table Question Answering (TableQA) combines natural language understanding and structured data reasoning, posing challenges in semantic interpretation and logical inference. Recent advances in Large Language Models (LLMs) have improved TableQA performance through Direct Prompting and Agent paradigms. However, these models often rely on spurious correlations, as they tend to overfit to token co-occurrence patterns in pretraining corpora, rather than perform genuine reasoning. To address this issue, we propose Causal Intervention TableQA (CIT), which is based on a structural causal graph and applies front-door adjustment to eliminate bias caused by token co-occurrence. CIT formalizes TableQA as a causal graph and identifies token co-occurrence patterns as confounders. By applying front-door adjustment, CIT guides question variant generation and reasoning to reduce confounding effects. Experiments on multiple benchmarks show that CIT achieves state-of-the-art performance, demonstrating its effectiveness in mitigating bias. Consistent gains across various LLMs further confirm its generalizability.



Paperid:4365
Authors:DONGZHI JIANG, Ziyu Guo, Renrui Zhang, ZHUOFAN ZONG, Hao Li, Le Zhuo, Shilin Yan, Pheng-Ann Heng, Hongsheng Li
Abstract:
Recent advancements in large language models have demonstrated how chain-of-thought (CoT) and reinforcement learning (RL) can improve performance. However, applying such reasoning strategies to the visual generation domain remains largely unexplored. In this paper, we presentT2I-R1, a novel reasoning-enhanced text-to-image generation model, powered by RL with a bi-level CoT reasoning process. Specifically, we identify two levels of CoT that can be utilized to enhance different stages of generation: (1) the semantic-level CoT for high-level planning of the prompt and (2) the token-level CoT for low-level pixel processing during patch-by-patch generation. To better coordinate these two levels of CoT, we introduceBiCoT-GRPOwith an ensemble of generation rewards, which seamlessly optimizes both generated CoTs within the same training step. By applying our reasoning strategies to the baseline model, Janus-Pro, we achieve superior performance with 13% improvement on T2I-CompBench and 19% improvement on the WISE benchmark, even surpassing the state-of-the-art model FLUX.1. All the training code is in the supplementary material and will be made public.



Paperid:4366
Authors:Sheng Wu, Lin Jin, Hui Feng, Bo Hu
Abstract:
Event cameras offer unique advantages in scenarios involving high speed, low light, and high dynamic range, yet their asynchronous and sparse nature poses significant challenges to efficient spatiotemporal representation learning. Specifically, despite notable progress in the field, effectively modeling the full spatiotemporal context, selectively attending to salient dynamic regions, and robustly adapting to the variable density and dynamic nature of event data remain key challenges. Motivated by these challenges, this paper proposes EventMG, a lightweight, efficient, multilevel Mamba-Graph architecture designed for learning high-quality spatiotemporal event representations. EventMG employs a multilevel approach, jointly modeling information at the micro (single event) and macro (event cluster) levels to comprehensively capture the multi-scale characteristics of event data. At the micro-level, it focuses on spatiotemporal details, employing State Space Model (SSM) based Mamba, to precisely capture long-range dependencies among numerous event nodes. Concurrently, at the macro-level, Component Graphs are introduced to efficiently encode the local semantics and global topology of dense event regions. Furthermore, to better accommodate the dynamic and sparse characteristics of data, we propose the Spatiotemporal-aware Event Scanning Technology (SEST), integrating the Attention-based Perturbation Network (APN) and Multidirectional Scanning Module (MSM), which substantially enhances the model's ability to perceive and focus on key spatiotemporal patterns. Experimental results demonstrate that EventMG achieves performance comparable or superior to state-of-the-art heavyweight models across multiple benchmarks, while maintaining an extremely low parameter count and linear complexity. This validates the effectiveness of the proposed architecture for efficient, multilevel spatiotemporal representation of event data. The code will be made publicly available upon acceptance.



Paperid:4367
Authors:Jing Wang, Weiting Peng, Jing Tang, Zeyu Gong, Xihua Wang, Bo Tao, Li cheng
Abstract:
Existing imitation learning methods decouple perception and action, which overlooks the causal reciprocity between sensory representations and action execution that humans naturally leverage for adaptive behaviors. To bridge this gap, we introduce Action-Guided Diffusion Policy (DP-AG), a unified representation learning that explicitly models a dynamic interplay between perception and action through probabilistic latent dynamics. DP-AG encodes latent observations into a Gaussian posterior via variational inference and evolves them using an action-guided SDE, where the Vector–Jacobian Product (VJP) of the diffusion policy's noise predictions serves as a structured stochastic force driving latent updates. To promote bidirectional learning between perception and action, we introduce a cycle-consistent contrastive loss that organizes the gradient flow of the noise predictor into a coherent perception–action loop, enforcing mutually consistent transitions in both latent updates and action refinements. Theoretically, we derive a variational lower bound for the action-guided SDE, and prove that the contrastive objective enhances continuity in both latent and action trajectories. Empirically, DP-AG significantly outperforms state-of-the-art methods across simulation benchmarks and real-world UR5 manipulation tasks. As a result, our DP-AG offers a promising step toward bridging biological adaptability and artificial policy learning.



Paperid:4368
Authors:Dian Meng, Zhiguang Cao, Jie Gao, Yaoxin Wu, Yaqing Hou
Abstract:
Neural solvers for the Vehicle Routing Problem (VRP) have typically relied on either node or edge inputs, limiting their flexibility and generalization in real-world scenarios. We propose UniteFormer, a unified neural solver that supports node-only, edge-only, and hybrid input types through a single model trained via joint edge-node modalities. UniteFormer introduces: (1) a mixed encoder that integrates graph convolutional networks and attention mechanisms to collaboratively process node and edge features, capturing cross-modal interactions between them; and (2) a parallel decoder enhanced with query mapping and a feed-forward layer for improved representation. The model is trained with REINFORCE by randomly sampling input types across batches. Experiments on TSP and CVRP demonstrate that UniteFormer achieves state-of-the-art performance and generalizes effectively to TSPLib and CVRPLib instances. These results underscore UniteFormer’s ability to handle diverse input modalities and its strong potential to improve performance across various VRP tasks.



Paperid:4369
Authors:Tingting Dan, Xinwei Huang, Won Hwa Kim, Guorong Wu
Abstract:
In graph neural networks (GNNs), feature representation learning evolves through iteratively updating node features and exchanging information based on graph topology. In this context, we conceptualize that the learning process in GNNs is a mean-field game (MFG), where each graph node is an agent, interacting with its topologically connected neighbors. However, current GNNs often employ the identical MFG strategy across different graph datasets, regardless of whether the graph exhibits homophilic or heterophilic characteristics. To address this challenge, we propose to formulate the learning mechanism into a variational framework of the MFG inverse problem, introducing an in-context selective message passing paradigm for each agent, which promotes the best overall outcome for the graph. Specifically, we seek for the application-adaptive transportation function (controlling information exchange throughout the graph) and reaction function (controlling feature representation learning on each agent),on the fly, which allows us to uncover the most suitable selective mechanism of message passing by solving an MFG variational problem through the lens of Hamiltonian flows. Taken together, our variational framework unifies existing GNN models into various mean-field games with distinct equilibrium states, each characterized by the learned in-context message passing operators. Furthermore, we present an agnostic end-to-end deep model, coinedGame-of-GNN, to jointly identify the message passing mechanism and fine-tune the GNN hyper-parameters on top of the elucidated message passing operators.Game-of-GNNhas achieved SOTA performance on diverse graph data, including popular benchmark datasets and human connectomes. More importantly, the mathematical insight of MFG framework provides a new window to understand the foundational principles of graph learning as an interactive dynamical system, which allows us to reshape the idea of designing next-generation GNN models.



Paperid:4370
Authors:Vivek Chavan, Yasmina Imgrund, Tung Dao, Sanwantri Bai, Bosong Wang, Ze Lu, Oliver Heimann, Jörg Krüger
Abstract:
We introduce the IndEgo dataset, a multimodal egocentric and exocentric video dataset addressing common industrial tasks, including assembly/disassembly, logistics and organisation, inspection and repair, woodworking, and others. The dataset contains 3,460 egocentric recordings (approximately 197 hours), along with 1,092 exocentric recordings (approximately 97 hours). A key focus of the dataset is collaborative work, where two workers work together on cognitively and physically intensive tasks. The egocentric recordings include rich multimodal data and added context via eye gaze, narration, sound, motion, and others. We provide detailed annotations (actions, summaries, mistake annotations, narrations), metadata, processed outputs (eye gaze, hand pose, semi-dense point cloud), and benchmarks on procedural and non-procedural task understanding, Mistake Detection, and reasoning-based Question Answering. Baseline evaluations for Mistake Detection, Question Answering and collaborative task understanding show that the dataset presents a challenge for the state-of-the-art multimodal models. Our dataset and code are available.



Paperid:4371
Authors:Haomiao Qiu, Miao Zhang, Ziyue Qiao, Liqiang Nie
Abstract:
Continual Learning (CL) aims to enable models to continuously acquire new knowledge from a sequence of tasks with avoiding the forgetting of learned information. However, existing CL methods only rely on the parameters of the most recent task for inference, which makes them susceptible to catastrophic forgetting. Inspired by the recent success of model merging techniques, we propose Perturb-and-Merge (P\&M), a novel continual learning framework that integrates model merging into the CL paradigm to mitigate forgetting. Specifically, after training on each task, P\&M constructs a new model by forming a convex combination of the previous model and the newly trained task-specific model. Through theoretical analysis, we minimize the total loss increase across all tasks and derive an analytical solution for the optimal merging coefficient. To further improve the performance of the merged model, we observe that the degradation introduced during merging can be alleviated by a regularization term composed of the task vector and the Hessian matrix of the loss function. Interestingly, we show that this term can be efficiently approximated using second-order symmetric finite differences, and a stochastic perturbation strategy along the task vector direction is accordingly devised which incurs no additional forward or backward passes while providing an effective approximation of the regularization term. Finally, we combine P\&M with LoRA, a parameter-efficient fine-tuning method, to reduce memory overhead. Our proposed approach achieves state-of-the-art performance on several continual learning benchmark datasets. The code is available at https://anonymous.4open.science/r/PaM-for-Continual-Learning-B63A.



Paperid:4372
Authors:Uliana Parkina, Maxim Rakhuba
Abstract:
Recent studies suggest that context-aware low-rank approximation is a useful tool for compression and fine-tuning of modern large-scale neural networks.In this type of approximation, a norm is weighted by a matrix of input activations, significantly improving metrics over the unweighted case. Nevertheless, existing methods for neural networks suffer from numerical instabilities due to their reliance on classical formulas involving explicit Gram matrix computation and their subsequent inversion. We demonstrate that this can degrade the approximation quality or cause numerically singular matrices.To address these limitations, we propose a novelinversion-free regularized frameworkthat is based entirely on stable decompositions and overcomes the numerical pitfalls of prior art. Our method can handle all possible challenging scenarios: (1) when calibration matrices exceed GPU memory capacity, (2) when input activation matrices are nearly singular, and even (3) when insufficient data prevents unique approximation.For the latter, we prove that our solution converges to a desired approximation and derive explicit error bounds.



Paperid:4373
Authors:Xin Ma, Yifan Wang, Siyu Yi, Wei Ju, Junyu Luo, Yusheng Zhao, Xiao Luo, Jiancheng Lv
Abstract:
Graph transfer learning, especially in unsupervised domain adaptation, focuses on transferring knowledge from a label-rich source domain graph to an unlabeled target domain graph. However, most existing methods assume a balanced label distribution in the source graph, often inconsistent with real-world scenarios where node labels are class-imbalanced. This presents significant challenges for graph transfer learning, as it is impacted by biased knowledge from the source graph and substantial domain distribution shifts. Given the effectiveness of class-specific prototypes in capturing domain semantics, in this paper, we propose a dual-branch prototype-enhanced contrastive framework for class-imbalanced graph domain adaptation. Specifically, we introduce a dual-branch graph encoder to capture both local and global information, generating class-specific prototypes from a distilled anchor set. Then, a prototype-enhance contrastive learning framework is introduced to address the aforementioned challenges. On the one hand, we encourage class alignment between the two branches based on constructed prototypes to mitigate the impact of class imbalance. On the other hand, we generate pseudo-labels for the target domain and align sample pairs across domains that share similar semantics to reduce domain discrepancies. Extensive experimental results show that our ImGDA outperforms the state-of-the-art methods across multiple datasets and settings.



Paperid:4374
Authors:Yang Qiu, Yixiong Zou, Jun Wang, Wei Liu, Xiangyu Fu, Ruixuan Li
Abstract:
Out-of-distribution generalization under distributional shifts remains a critical challenge for graph neural networks. Existing methods generally adopt the Invariant Risk Minimization (IRM) framework, requiring costly environment annotations or heuristically generated synthetic splits. To circumvent these limitations, in this work, we aim to develop an IRM-free method for capturing causal subgraphs. We first identify that causal subgraphs exhibit substantially smaller distributional variations than non-causal components across diverse environments, which we formalize as the Invariant Distribution Criterion and theoretically prove in this paper. Building on this criterion, we systematically uncover the quantitative relationship between distributional shift and representation norm for identifying the causal subgraph, and investigate its underlying mechanisms in depth. Finally, we propose an IRM-free method by introducing a norm-guided invariant distribution objective for causal subgraph discovery and prediction. Extensive experiments on two widely used benchmarks demonstrate that our method consistently outperforms state-of-the-art methods in graph generalization. Our codes will be released.



Paperid:4375
Authors:Mingyu Huang, Shasha Zhou, Ke Li
Abstract:
Machine learning models increasingly map biological sequence-fitness landscapes across various modalities to predict mutational effects. Effective evaluation of these models requires benchmarks derived from empirical data. However, current benchmarks often lack detailed topographical information regarding the underlying fitness landscape, which hampers interpretation and comparison of model performance. Here, we introduce GraphFLA, a Python framework that constructs and analyzes fitness landscapes from diverse modalities (DNA, RNA, protein, and beyond), accommodating datasets up to millions of mutants. GraphFLA calculates 20 biologically relevant features that characterize 4 fundamental aspects of landscape topography. By applying GraphFLA to over 5,300 landscapes from ProteinGym, RNAGym, and UniProbe, we demonstrate its utility in interpreting and comparing the performance of dozens of fitness prediction models, highlighting factors influencing model accuracy. Additionally, we release 155 combinatorially complete landscapes, encompassing over 2.2 million sequences across various modalities. All resources are available at https://github.com/COLA-Laboratory/GraphFLA.



Paperid:4376
Authors:Andrea Pugnana, Riccardo Massidda, Francesco Giannini, Pietro Barbiero, Mateo Espinosa Zarlenga, Roberto Pellungrini, Gabriele Dominici, Fosca Giannotti, Davide Bacciu
Abstract:
Concept Bottleneck Models (CBMs) are interpretable machine learning models that ground their predictions on human-understandable concepts, allowing for targeted interventions in their decision-making process. However, when intervened on, CBMs assume the availability of humans that can identify the need to intervene and always provide correct interventions. Both assumptions are unrealistic and impractical, considering labor costs and human error-proneness. In contrast, Learning to Defer (L2D) extends supervised learning by allowing machine learning models to identify cases where a human is more likely to be correct than the model, thus leading to deferring systems with improved performance. In this work, we gain inspiration from L2D and propose Deferring CBMs (DCBMs), a novel framework that allows CBMs to learn when an intervention is needed. To this end, we model DCBMs as a composition of deferring systems and derive a consistent L2D loss to train them. Moreover, by relying on a CBM architecture, DCBMs can explain why defer occurs on the final task. Our results show that DCBMs can achieve high predictive performance and interpretability by deferring only when strictly needed.



Paperid:4377
Authors:Zhenhao Zhang, Ye Shi, Lingxiao Yang, Suting Ni, Qi Ye, Jingya Wang
Abstract:
Understanding and synthesizing realistic 3D hand-object interactions (HOI) is critical for applications ranging from immersive AR/VR to dexterous robotics. Existing methods struggle with generalization, performing well on closed-set objects and predefined tasks but failing to handle unseen objects or open-vocabulary instructions. We introduce OpenHOI, the first framework for open-world HOI synthesis, capable of generating long-horizon manipulation sequences for novel objects guided by free-form language commands. Our approach integrates a 3D Multimodal Large Language Model (MLLM) fine-tuned for joint affordance grounding and semantic task decomposition, enabling precise localization of interaction regions (e.g., handles, buttons) and breakdown of complex instructions (e.g., “Find a water bottle and take a sip”) into executable sub-tasks. To synthesize physically plausible interactions, we propose an affordance-driven diffusion model paired with a training-free physics refinement stage that minimizes penetration and optimizes affordance alignment.Evaluations across diverse scenarios demonstrate OpenHOI’s superiority over state-of-the-art methods in generalizing to novel object categories, multi-stage tasks, and complex language instructions.



Paperid:4378
Authors:Lvmin Zhang, Shengqu Cai, Muyang Li, Gordon Wetzstein, Maneesh Agrawala
Abstract:
We present a neural network structure, FramePack, to train next-frame (or next-frame-section) prediction models for video generation. FramePack compresses input frame contexts with frame-wise importance so that more frames can be encoded within a fixed context length, with more important frames having longer contexts. The frame importance can be measured using time proximity, feature similarity, or hybrid metrics. The packing method allows for inference with thousands of frames and training with relatively large batch sizes. We also present drift prevention methods to address observation bias (error accumulation), including early-established endpoints, adjusted sampling orders, and discrete history representation. Ablation studies validate the effectiveness of the anti-drifting methods in both single-directional video streaming and bi-directional video generation. Finally, we show that existing video diffusion models can be finetuned with FramePack, and analyze the differences between different packing schedules.



Paperid:4379
Authors:Yuheng Tang, Hongwei Li, Kaijie Zhu, Michael Yang, Yangruibo Ding, Wenbo Guo
Abstract:
Motivated by the success of general‑purpose large language models (LLMs) in software patching, recent works started to train specialized patching models. Most works trained one model to handle the end‑to‑end patching pipeline (including issue localization, patch generation, and patch validation). However, it is hard for a small model to handle all tasks, as different sub-tasks have different workflows and require different expertise. As such, by using a 70 billion model, SOTA methods can only reach up to 41% resolved rate on SWE-bench-Verified. Motivated by the collaborative nature, we propose Co-PatcheR, the first collaborative patching system with small and specialized reasoning models for individual components. Our key technique novelties are the specific task designs and training recipes. First, we train a model for localization and patch generation. Our localization pinpoints the suspicious lines through a two-step procedure, and our generation combines patch generation and critique. We then propose a hybrid patch validation that includes two models for crafting issue-reproducing test cases with and without assertions and judging patch correctness, followed by a majority vote-based patch selection. Through extensive evaluation, we show that Co-PatcheR achieves 46% resolved rate on SWE-bench-Verified with only 3 x 14B models. This makes Co-PatcheR the best patcher with specialized models, requiring the least training resources and the smallest models. We conduct a comprehensive ablation study to validate our recipes, as well as our choice of training data number, model size, and testing-phase scaling strategy.



Paperid:4380
Authors:Weixin An, Yuanyuan Liu, Fanhua Shang, Han Yu, Junkang Liu, Hongying Liu
Abstract:
Abstract:Gradient clipping is increasingly important in centralized learning (CL) and federated learning (FL). Many works focus on its optimization properties under strong assumptions involving Gaussian noise and standard smoothness. However, practical machine learning tasks often only satisfy weaker conditions, such as heavy-tailed noise and $(L_0, L_1)$-smoothness. To bridge this gap, we propose a high-probability analysis for clipped Stochastic Gradient Descent (SGD) under these weaker assumptions. Our findings show a better convergence rate than existing ones can be achieved, and our high-probability analysis does not rely on the bounded gradient assumption. Moreover, we extend our analysis to FL, where a gap remains between expected and high-probability convergence, which the naive clipped SGD cannot bridge. Thus, we design a new \underline{Fed}erated \underline{C}lipped \underline{B}atched \underline{G}radient (FedCBG) algorithm, and prove the convergence and generalization bounds with high probability for the first time. Our analysis reveals the trade-offs between the optimization and generalization performance. Extensive experiments demonstrate that \methodname{} can generalize better to unseen client distributions than state-of-the-art baselines.



Paperid:4381
Authors:Tiezheng Zhang, Yitong Li, Yu-Cheng Chou, Jieneng Chen, Alan Yuille, Chen Wei, Junfei Xiao
Abstract:
Building state-of-the-art Vision-Language Models (VLMs) with strong captioning capabilities typically necessitates training on billions of high-quality image-text pairs, requiring millions of GPU hours. This paper introduces the Vision-Language-Vision(VLV)auto-encoder framework, which strategically leverages key pretrained components: a vision encoder, the decoder of a Text-to-Image (T2I) diffusion model, and subsequently, a Large Language Model (LLM). Specifically, we establish an information bottleneck by regularizing the language representation space, achieved through freezing the pretrained T2I diffusion decoder. Our VLV pipeline effectively distills knowledge from the text-conditioned diffusion model using continuous embeddings, demonstrating comprehensive semantic understanding via high-quality reconstructions. Furthermore, by fine-tuning a pretrained LLM to decode the intermediate language representations into detailed descriptions, we construct a state-of-the-art (SoTA) captioner comparable to leading models like GPT-4o and Gemini 2.0 Flash. Our method demonstrates exceptional cost-efficiency and significantly reduces data requirements; by primarily utilizing single-modal images for training and maximizing the utility of existing pretrained models (image encoder, T2I diffusion model, and LLM), it circumvents the need for massive paired image-text datasets, keeping the total training expenditure under $1,000 USD.



Paperid:4382
Authors:Fanhu Zeng, Haiyang Guo, Fei Zhu, Li Shen, Hao Tang
Abstract:
Fine-tuning pre-trained models with custom data leads to numerous expert models on specific tasks. Merging models into one universal model to empower multi-task ability refraining from data leakage has gained popularity. With the expansion in data and model size, parameter efficient tuning becomes the common practice for obtaining task-specific models efficiently. However, few methods are dedicated to efficient merging and existing methods designed for full fine-tuning merging fail under efficient merging. To address the issue, we analyze from low-rank decomposition and reveal that direction robustness during merging is crucial for merging efficient modules. We furthermore uncover that compensating for gap between stark singular values contributes to direction robustness. Therefore, we propose CoPA-Merging, a training-free parameter efficient merging method with complementary parameter adaptation to maintain directions. Specifically, we (1) prune parameters and scale coefficients from inter-parameter relation for singular values to maintain direction stability away from task interference and (2) perform cross-task normalization to enhance unseen task generalization. We establish a benchmark consisting of diverse multimodal tasks, on which we conduct experiments to certificate the outstanding performance and generalizability of our method. Additional study and extensive analyses further showcase the effectiveness.



Paperid:4383
Authors:Shenglong Zhou, Manjiang Yin, Leiyu Sun, Shicai Yang, Di Xie, Jiang Zhu
Abstract:
Test-time adaptation with pre-trained vision-language models shows impressive zero-shot classification abilities, and training-free methods further improve the performance without any optimization burden. However, existing training-free test-time adaptation methods typically rely on entropy criteria to select the visual features and update the visual caches, while ignoring the generalizable factors, such as shape-sensitive and style-insensitive factors. In this paper, we propose a novel shape and style guidance method (SSG) for training-free test-time adaptation in vision-language models, aiming to highlight the shape-sensitive (SHS) and style-insensitive (SIS) factors in addition to entropy criteria.Specifically, SSG perturbs the raw test image with shape and style corruption operations, and measures the prediction difference between the raw and corrupted one as perturbed prediction difference (PPD). Based on the PPD measurement, SSG reweights the high-confidence visual features and corresponding predictions, aiming to highlight the effect of SHS and STI factors during the test-time procedure. Furthermore, SSG takes both PPD and entropy into consideration to update the visual cache, aiming to maintain the stored sample with high entropy and generalizable factors. Extensive experimental results on out-of-distribution and cross-domain benchmark datasets demonstrate that our proposed SSG consistently outperforms previous state-of-the-art methods while also exhibiting promising computational efficiency.



Paperid:4384
Authors:Tianyu Fu, Yi Ge, Yichen You, Enshu Liu, Zhihang Yuan, Guohao Dai, Shengen Yan, Huazhong Yang, Yu Wang
Abstract:
Large Language Models (LLMs) achieve impressive reasoning capabilities at the cost of substantial inference overhead, posing substantial deployment challenges. Although distilled Small Language Models (SLMs) significantly enhance efficiency, their performance suffers as they fail to follow LLMs' reasoning paths. Luckily, we reveal that only a small fraction of tokens genuinely diverge reasoning paths between LLMs and SLMs. Most generated tokens are either identical or exhibit neutral differences, such as minor variations in abbreviations or expressions. Leveraging this insight, we introduceRoads to Rome (R2R), a neural token router that selectively utilizes LLMs only for these critical, path-divergent tokens, while leaving the majority of token generation to the SLM. We also develop an automatic data generation pipeline that identifies divergent tokens and generates token-level routing labels to train the lightweight router. We apply R2R to combine R1-1.5B and R1-32B models from the DeepSeek family, and evaluate on challenging math, coding, and QA benchmarks. With an average activated parameter size of 5.6B, R2R surpasses the average accuracy of R1-7B by 1.6×, outperforming even the R1-14B model. Compared to R1-32B, it delivers a 2.8× wall-clock speedup with comparable performance, advancing the Pareto frontier of test-time scaling efficiency.



Paperid:4385
Authors:Yifan Wei, Xiaoyan Yu, Tengfei Pan, Angsheng Li, Li Du
Abstract:
Large language models (LLMs) have achieved unprecedented performance by leveraging vast pretraining corpora, yet their performance remains suboptimal in knowledge-intensive domains such as medicine and scientific research, where high factual precision is required.While synthetic data provides a promising avenue for augmenting domain knowledge, existing methods frequently generate redundant samples that do not align with the model’s true knowledge gaps. To overcome this limitation, we propose a novel Structural Entropy-guided Knowledge Navigator (SENATOR) framework that addresses the intrinsic knowledge deficiencies of LLMs. Our approach employs the Structure Entropy (SE) metric to quantify uncertainty along knowledge graph paths and leverages Monte Carlo Tree Search (MCTS) to selectively explore regions where the model lacks domain-specific knowledge. Guided by these insights, the framework generates targeted synthetic data for supervised fine-tuning, enabling continuous self-improvement. Experimental results on LLaMA-3 and Qwen2 across multiple domain-specific benchmarks show that SENATOR effectively detects and repairs knowledge deficiencies, achieving notable performance improvements.



Paperid:4386
Authors:Mete Erdogan, Cengiz Pehlevan, Alper Erdogan
Abstract:
We introduceError Broadcast and Decorrelation(EBD), a novel learning framework for neural networks that addresses credit assignment by directly broadcasting output errors to individual layers, circumventing weight transport of backpropagation. EBD is rigorously grounded in the stochastic orthogonality property of Minimum Mean Square Error estimators. This fundamental principle states that the error of an optimal estimator is orthogonal to functions of the input. Guided by this insight, EBD defines layerwise loss functions that directly penalize correlations between layer activations and output errors, thereby establishing a principled foundation for error broadcasting. This theoretically sound mechanism naturally leads to the experimentally observed three-factor learning rule and integrates with biologically plausible frameworks to enhance performance and plausibility. Numerical experiments demonstrate EBD’s competitive or better performance against other error-broadcast methods on benchmark datasets. Our findings establish EBD as an efficient, biologically plausible, and principled alternative for neural network training.



Paperid:4387
Authors:Victor Li, Baiting Chen, Yuzhen Mao, Qi Lei, Zhun Deng
Abstract:
Calibrating blackbox machine learning models to achieve risk control is crucial to ensure reliable decision-making. A rich line of literature has been studying how to calibrate a model so that its predictions satisfy explicit finite-sample statistical guarantees under afixed,static, and unknown data-generating distribution. However, prediction-supported decisions may influence the outcome they aim to predict, a phenomenon namedperformativityof predictions, which is commonly seen in social science and economics. In this paper, we introducePerformative Risk Control, a framework to calibrate models to achieve risk control under performativity with provable theoretical guarantees. Specifically, we provide an iteratively refined calibration process, where we ensure the predictions are improved and risk-controlled throughout the process. We also study different types of risk measures and choices of tail bounds. Lastly, we demonstrate the effectiveness of our framework by numerical experiments on the task of predicting credit default risk. To the best of our knowledge, this work is the first one to study statistically rigorous risk control under performativity, which will serve as an important safeguard against a wide range of strategic manipulation in decision-making processes.



Paperid:4388
Authors:Dongchan Cho, Jiho Han, Keumyeong Kang, Minsang Kim, Honggyu Ryu, Namsoon Jung
Abstract:
Real-world multivariate time series anomalies are rare and often unlabeled. Additionally, prevailing methods rely on increasingly complex architectures tuned to benchmarks, detecting only fragments of anomalous segments and overstating performance. In this paper, we introduce OracleAD, a simple and interpretable unsupervised framework for multivariate time series anomaly detection. OracleAD encodes each variable’s past sequence into a single causal embedding to jointly predict the present time point and reconstruct the input window, effectively modeling temporal dynamics. These embeddings then undergo self-attention mechanism to project them into a shared latent space and capture spatial relationships. These relationships are not static, since they are modeled by a property that emerges from each variable's temporal dynamics. The projected embeddings are aligned to a Stable Latent Structure (SLS) representing normal-state relationships. Anomalies are identified using a dual scoring mechanism based on prediction error and deviation from the SLS, enabling fine-grained anomaly diagnosis at each time point and across individual variables. Since any noticeable SLS deviation originates from embeddings that violate the learned temporal causality of normal data, OracleAD directly pinpoints the root-cause variables at the embedding level. OracleAD achieves state-of-the-art results across multiple real-world datasets and evaluation protocols, while remaining interpretable through SLS.



Paperid:4389
Authors:Shubo Li, Yizhe Ding, Lingzhou Xue, Runze Li
Abstract:
Abstract:Optimal transport (OT) provides a powerful framework for comparing and transforming probability distributions, with wide applications in generative modeling, AI4Science and statistical inference. However, existing estimation theory typically requires stringent smoothness conditions on the underlying Brenier potentials and assumes bounded distribution supports, limiting practical applicability. In this paper, we introduce a unified theoretical framework for semi-dual OT map estimation that relaxes both of these restrictions. Building on sieved convex conjugate, our framework has two key contributions: (i) a new map stability bounds that holds without any second-order regularity assumptions on the true Brenier potentials, and (ii) an oracle inequality that cleanly decomposes the estimation error into statistical error, sieved bias, and approximation error. Specifically, our approximation error is measured in the $L^\infty$ norm rather than Sobolev norm in the existing results, aligning more naturally with classical approximation theory. Leveraging these tools, we provide statistical error of semi-dual estimators with mild and verifiable conditions on the true OT map. Moreover, we establish the first theoretical guarantee for deep neural network OT map estimator between general distributions, with Tanh network function class as an example.



Paperid:4390
Authors:Xiaojun Ning, Jing Wang, Zhiyang Feng, Tianzuo Xin, Shuo Zhang, Shaoqi Zhang, Zheng Lian, Yi Ding, Youfang Lin, Ziyu Jia
Abstract:
Affective brain-computer interfaces (aBCIs) play a crucial role in personalized human–computer interaction and neurofeedback modulation. To develop practical and effective aBCI paradigms and to investigate the spatial-temporal dynamics of brain activity under emotional inducement, portable electroencephalography (EEG) signals have been widely adopted. To further enhance spatial-temporal perception, functional near-infrared spectroscopy (fNIRS) has attracted increasing interest in the aBCI field and has been explored in combination with EEG. However, existing datasets typically provide only static fixation labels, overlooking the dynamic changes in subjects' emotions. Notably, some studies have attempted to collect continuously annotated emotional data, but they have recorded only peripheral physiological signals without directly observing brain activity, limiting insight into underlying neural states under different emotions. To address these challenges, we present REFED, the first EEG-fNIRS dataset with real-time dynamic emotional annotations. REFED simultaneously records brain signals from both EEG and fNIRS modalities while providing continuous, real-time annotations of valence and arousal. The results of the data analysis demonstrate the effectiveness of emotion inducement and the reliability of real-time annotation. This dataset offers the first possibility for studying the neural-vascular coupling mechanism under emotional evolution and for developing dynamic, robust affective BCIs.



Paperid:4391
Authors:Xinyu Wang, Jonas M. Kübler, Kailash Budhathoki, Yida Wang, Matthäus Kleindessner
Abstract:
When serving a single base LLM with several different LoRA adapters simultaneously, the adapters cannot simply be merged with the base model’s weights as the adapter swapping would create overhead and requests using different adapters could not be batched. Rather, the LoRA computations have to be separated from the base LLM computations, and in a multi-device setup the LoRA adapters can be sharded in a way that is well aligned with the base model’s tensor parallel execution, as proposed in S-LoRA. However, the S-LoRA sharding strategy encounters some communication overhead, which may be small in theory, but can be large in practice. In this paper, we propose to constrain certain LoRA factors to be block-diagonal, which allows for an alternative way of sharding LoRA adapters that does not require any additional communication for the LoRA computations. We demonstrate in extensive experiments that our block-diagonal LoRA approach is similarly parameter efficient as standard LoRA (i.e., for a similar number of parameters it achieves similar downstream performance) and that it leads to significant end-to-end speed-up over S-LoRA. For example, when serving on eight A100 GPUs, we observe up to 1.79x (1.23x) end-to-end speed-up with 0.87x (1.74x) number of adapter parameters for Llama-3.1-70B, and up to 1.63x (1.3x) end-to-end speed-up with 0.86x (1.73x) number of adapter parameters for Llama-3.1-8B.



Paperid:4392
Authors:Quanjiang Li, Tianxiang Xu, Tingjin Luo, Yan Zhong, Yang Li, Yiyun Zhou, Chenping Hou
Abstract:
Multi-view multi-label learning typically suffers from dual data incompleteness due to limitations in feature storage and annotation costs.The interplay of heterogeneous features, numerous labels, and missing information significantly degrades model performance. To tackle the complex yet highly practical challenges, we propose a Theory-Driven Label-Specific Representation (TDLSR) framework. Through constructing the view-specific sample topology and prototype association graph, we develop proximity-aware imputation via attention-based relational propagation, while deriving class representatives that capture the label correlation semantic. To obtain semantically distinct view representations, we introduce principles of information shift, interaction and orthogonality, which promotes task-relevant feature disentanglement and mitigates message distortion and redundancy. Building upon the mutual information optimization model, we formulate its theoretical upper bound into its objective. Moreover, label-semantic-guided feature learning is employed to identify the discriminative shared and specific representations and refine the label preference across views. Finally, extensive experiments on public datasets and real-world applications validate the effectiveness of TDLSR.



Paperid:4393
Authors:Ahmad-Reza Ehyaei, Ali Shirali, Samira Samadi
Abstract:
Counterfactual explanations provide individuals with cost-optimal recommendations to achieve their desired outcomes. However, when a significant number of individuals seek similar state modifications, this individual-centric approach can inadvertently create competition and introduce unforeseen costs. Additionally, disregarding the underlying data distribution may lead to recommendations that individuals perceive as unusual or impractical.To address these challenges, we propose a novel framework that extends standard counterfactual explanations by incorporating a population dynamics model. This framework penalizes deviations from equilibrium after individuals follow the recommendations, effectively mitigating externalities caused by correlated changes across the population. By balancing individual modification costs with their impact on others, our method ensures a more equitable and efficient outcome.We show how this approach reframes the counterfactual explanation problem from an individual-centric task to a collective optimization problem. Augmenting our theoretical insights, we design and implement scalable algorithms for computing collective counterfactuals, showcasing their effectiveness and advantages over existing recourse methods, particularly in aligning with collective objectives.



Paperid:4394
Authors:Haobo Li, Eunseo Jung, Zixin CHEN, Zhaowei Wang, Yueya WANG, Huamin Qu, Alexis Lau
Abstract:
Multimodal time series forecasting is foundational in various fields, such as utilizing satellite imagery and numerical data for predicting typhoons in climate science. However, existing multimodal approaches primarily focus on utilizing text data to help time series forecasting, leaving the visual data in existing time series datasets untouched. Furthermore, it is challenging for models to effectively capture the physical information embedded in visual data, such as satellite imagery's temporal and geospatial context, which extends beyond images themselves. To address this gap, we propose physics-informed positional encoding (PIPE), a lightweight method that embeds physical information into vision language models (VLMs). PIPE introduces two key innovations: (1) a physics-informed positional indexing scheme for mapping physics to positional IDs, and (2) a variant-frequency positional encoding mechanism for encoding frequency information of physical variables and sequential order of tokens within the embedding space. By preserving both the physical information and sequential order information, PIPE significantly improves multimodal alignment and forecasting accuracy. Through the experiments on the most representative and the largest open-sourced satellite image dataset, PIPE achieves state-of-the-art performance in both deep learning forecasting and climate domain methods, demonstrating superiority across benchmarks, including a 12% improvement in typhoon intensity forecasting over prior works.Our code is provided in the supplementary material.



Paperid:4395
Authors:Wenxiang Guo, Changhao Pan, Zhiyuan Zhu, Xintong Hu, Yu Zhang, Li Tang, Rui Yang, Han Wang, Zongbao Zhang, Yuhan Wang, Yixuan Chen, Hankun Xu, Ke Xu, PengFei Fan, ZheTao Chen, Yanhao Yu, Qiange Huang, Fei Wu, Zhou Zhao
Abstract:
Humans rely on multisensory integration to perceive spatial environments, where auditory cues enable sound source localization in three-dimensional space. Despite the critical role of spatial audio in immersive technologies such as VR/AR, most existing multimodal datasets provide only monaural audio, which limits the development of spatial audio generation and understanding. To address these challenges, we introduce MRSAudio, a large-scale multimodal spatial audio dataset designed to advance research in spatial audio understanding and generation. MRSAudio spans four distinct components: MRSLife, MRSSpeech, MRSMusic, and MRSSing, covering diverse real-world scenarios. The dataset includes synchronized binaural and ambisonic audio, exocentric and egocentric video, motion trajectories, and fine-grained annotations such as transcripts, phoneme boundaries, lyrics, scores, and prompts.To demonstrate the utility and versatility of MRSAudio, we establish five foundational tasks: audio spatialization, and spatial text to speech, spatial singing voice synthesis, spatial music generation and sound event localization and detection. Results show that MRSAudio enables high-quality spatial modeling and supports a broad range of spatial audio research.Demos and dataset access are available at https://mrsaudio.github.io.



Paperid:4396
Authors:Yu Li, Xingyu Qiu, Yuqian Fu, Jie Chen, Tianwen Qian, Xu Zheng, Danda Pani Paudel, Yanwei Fu, Xuanjing Huang, Luc V Gool, Yu-Gang Jiang
Abstract:
Cross-Domain Few-Shot Object Detection (CD-FSOD) aims to detect novel objects with only a handful of labeled samples from previously unseen domains. While data augmentation and generative methods have shown promise in few-shot learning, their effectiveness for CD-FSOD remains unclear due to the need for both visual realism and domain alignment. Existing strategies, such as copy-paste augmentation and text-to-image generation, often fail to preserve the correct object category or produce backgrounds coherent with the target domain, making them non-trivial to apply directly to CD-FSOD. To address these challenges, we propose Domain-RAG, a training-free, retrieval-guided compositional image generation framework tailored for CD-FSOD. Domain-RAG consists of three stages: domain-aware background retrieval, domain-guided background generation, and foreground-background composition. Specifically, the input image is first decomposed into foreground and background regions. We then retrieve semantically and stylistically similar images to guide a generative model in synthesizing a new background, conditioned on both the original and retrieved contexts. Finally, the preserved foreground is composed with the newly generated domain-aligned background to form the generated image. Without requiring any additional supervision or training, Domain-RAG produces high-quality, domain-consistent samples across diverse tasks, including CD-FSOD, remote sensing FSOD, and camouflaged FSOD. Extensive experiments show consistent improvements over strong baselines and establish new state-of-the-art results. Codes will be released upon acceptance.



Paperid:4397
Authors:Andy Zou, Maxwell Lin, Eliot Jones, Micha Nowak, Mateusz Dziemian, Nick Winter, Valent Nathanael, Ayla Croft, Xander Davies, Jai Patel, Robert Kirk, Yarin Gal, Dan Hendrycks, J. Zico Kolter, Matt Fredrikson
Abstract:
Abstract:AI agents are rapidly being deployed across diverse industries, but can they adhere to deployment policies under attacks? We organized a one-month red teaming challenge---the largest of its kind to date---involving expert red teamers attempting to elicit policy violations from AI agents powered by $22$ frontier LLMs. Our challenge collected $1.8$ million prompt injection attacks, resulting in over $60,000$ documented successful policy violations, revealing critical vulnerabilities. Utilizing this extensive data, we construct a challenging AI agent red teaming benchmark, currently achieving near $100\%$ attack success rates across all tested agents and associated policies. Our further analysis reveals high transferability and universality of successful attacks, underscoring the scale and criticality of existing AI agent vulnerabilities. We also observe minimal correlation between agent robustness and factors such as model capability, size, or inference compute budget, highlighting the necessity of substantial improvements in defense. We hope our benchmark and insights drive further research toward more secure and reliable AI agents.



Paperid:4398
Authors:Valentin Bieri, Marie-Julie Rakotosaona, Keisuke Tateno, Francis Engelmann, Leonidas Guibas
Abstract:
Current 3D layout estimation models are predominantly trained on synthetic datasets biased toward simplistic, single-floor scenes. This prevents them from generalizing to complex, multi-floor buildings, often forcing a per-floor processing approach that sacrifices global context. Few works have attempted to holistically address multi-floor layouts. In this work, we introduce HouseLayout3D, a real-world benchmark dataset, which highlights the limitations of existing research when handling expansive, architecturally complex spaces. Additionally, we propose MultiFloor3D, a baseline method leveraging recent advances in 3D reconstruction and 2D segmentation. Our approach significantly outperforms state-of-the-art methods on both our new and existing datasets. Remarkably, it does not require any layout-specific training.



Paperid:4399
Authors:Xiaonan Si, Meilin Zhu, Lijia Yu, Lijun Zhang, Shuaitong Liu, Xiaojun Jia, Simeng Qin, Xinfeng Li, Ranjie Duan, Yang Liu
Abstract:
Retrieval-augmented generation (RAG) systems enhance large language models (LLMs) with external knowledge but are vulnerable to corpus poisoning and contamination attacks, which can compromise output integrity. Existing defenses often apply aggressive filtering, leading to unnecessary loss of valuable information and reduced reliability in generation.To address this problem, we propose a two-stage semantic filtering and conflict-free framework for trustworthy RAG. In the first stage, we perform a joint filter with semantic and cluster-based filtering which is guided by the Entity-intent-relation extractor (EIRE). EIRE extracts entities, latent objectives, and entity relations from both the user query and filtered documents, scores their semantic relevance, and selectively adds valuable documents into the clean retrieval database. In the second stage, we proposed an EIRE-guided conflict-aware filtering module, which analyzes semantic consistency between the query, candidate answers, and retrieved knowledge before final answer generation, filtering out internal and external contradictions that could mislead the model.Through this two-stage process, SeCon-RAG effectively preserves useful knowledge while mitigating conflict contamination, achieving significant improvements in both generation robustness and output trustworthiness.Extensive experiments across various LLMs and datasets demonstrate that the proposed SeCon-RAG markedly outperforms state-of-the-art defense methods.



Paperid:4400
Authors:Qiujiang Jin, Aryan Mokhtari
Abstract:
In this paper, we establish global non-asymptotic convergence guarantees for the BFGS quasi-Newton method without requiring strong convexity or the Lipschitz continuity of the gradient or Hessian. Instead, we consider the setting where the objective function is strictly convex and strongly self-concordant. For an arbitrary initial point and any arbitrary positive-definite initial Hessian approximation, we prove global linear and superlinear convergence guarantees for BFGS when the step size is determined using a line search scheme satisfying the weak Wolfe conditions. Moreover, all our global guarantees are affine-invariant, with the convergence rates depending solely on the initial error and the strongly self-concordant constant. Our results extend the global non-asymptotic convergence theory of BFGS beyond traditional assumptions and, for the first time, establish affine-invariant convergence guarantees—aligning with the inherent affine invariance of the BFGS method.



Paperid:4401
Authors:Ruixiang ZHANG, Shuangfei Zhai, Jiatao Gu, Yizhe Zhang, Huangjie Zheng, Tianrong Chen, Miguel Angel Bautista, Joshua Susskind, Navdeep Jaitly
Abstract:
Autoregressive models have driven remarkable progress in language modeling. Their foundational reliance on discrete tokens, unidirectional context, and single-pass decoding, while central to their success, also inspires the exploration of a design space that could offer new axes of modeling flexibility. In this work, we explore an alternative paradigm, shifting language modeling from a discrete token space to a continuous latent space. We propose a novel framework that employs transformer-based autoregressive normalizing flows to model these continuous representations. This approach unlocks substantial flexibility, enabling the construction of models that can capture global bi-directional context through stacked, alternating-direction autoregressive transformations, support block-wise generation with flexible token patch sizes, and facilitate a hierarchical multi-pass generation process. We further propose new mixture-based coupling transformations designed to capture complex dependencies within the latent space shaped by discrete data, and demonstrate theoretical connections to conventional discrete autoregressive models. Extensive experiments on language modeling benchmarks demonstrate strong likelihood performance and highlight the flexible modeling capabilities inherent in our framework.



Paperid:4402
Authors:Beitao Chen, Xinyu Lyu, Shengming Yuan, Jingkuan Song, Hengtao Shen, Lianli Gao
Abstract:
By incorporating visual inputs, Multimodal Large Language Models (MLLMs) extend LLMs to support visual reasoning. However, this integration also introduces new vulnerabilities, making MLLMs susceptible to multimodal jailbreak attacks and hindering their safe deployment. Existing defense methods, including Image-to-Text Translation, Safe Prompting, and Multimodal Safety Tuning, attempt to address this by aligning multimodal inputs with LLMs’ built-in safeguards. Yet, they fall short in uncovering root causes of multimodal vulnerabilities, particularly how harmful multimodal tokens trigger jailbreak in MLLMs? Consequently, they remain vulnerable to text-driven multimodal attacks, often exhibiting overdefensive behaviors and imposing heavy training overhead. To bridge this gap, we present an comprehensive analysis of where, how and which harmful multimodal tokens bypass safeguards in MLLMs.Surprisingly, we find that less than 1% tokens in early-middle layers are responsible for inducing unsafe behaviors, highlighting the potential of precisely removing a small subset of harmful tokens, without requiring safety tuning, can still effectively improve safety against jailbreaks.Motivated by this, we propose Safe Prune-then-Restore (SafePTR), an training-free defense framework that selectively prunes harmful tokens at vulnerable layers while restoring benign features at subsequent layers. Without incurring additional computational overhead, SafePTR significantly enhances the safety of MLLMs while preserving efficiency. Extensive evaluations across three MLLMs and five benchmarks demonstrate SafePTR’s state-of-the-art performance in mitigating jailbreak risks without compromising utility.



Paperid:4403
Authors:Peng Xue, Wei Fang, Zhengyu Ma, Zihan Huang, Zhaokun Zhou, Yonghong Tian, Timothée Masquelier, Huihui Zhou
Abstract:
Spiking Neural Networks (SNNs) are distinguished from Artificial Neural Networks (ANNs) for their complex neuronal dynamics and sparse binary activations (spikes) inspired by the biological neural system. Traditional neuron models use iterative step-by-step dynamics, resulting in serial computation and slow training speed of SNNs. Recently, parallelizable spiking neuron models have been proposed to fully utilize the massive parallel computing ability of graphics processing units to accelerate the training of SNNs. However, existing parallelizable spiking neuron models involve dense floating operations and can only achieve high long-term dependencies learning ability with a large order at the cost of huge computational and memory costs. To solve the dilemma of performance and costs, we propose the mul-free channel-wise Parallel Spiking Neuron, which is hardware-friendly and suitable for SNNs’ resource-restricted application scenarios. The proposed neuron imports the channel-wise convolution to enhance the learning ability, induces the sawtooth dilations to reduce the neuron order, and employs the bit-shift operation to avoid multiplications. The algorithm for the design and implementation of acceleration methods is discussed extensively. Our methods are validated in neuromorphic Spiking Heidelberg Digits voices, sequential CIFAR images, and neuromorphic DVS-Lip vision datasets, achieving superior performance over SOTA spiking neurons. Training speed results demonstrate the effectiveness of our acceleration methods, providing a practical reference for future research.



Paperid:4404
Authors:Yuhui Li, Fangyun Wei, Chao Zhang, Hongyang Zhang
Abstract:
The sequential nature of modern LLMs makes them expensive and slow, and speculative sam- pling has proven to be an effective solution to this problem. Methods like EAGLE perform autoregression at the feature level, reusing top- layer features from the target model to achieve better results than vanilla speculative sampling. A growing trend in the LLM community is scaling up training data to improve model intelligence without increasing inference costs. However, we observe that scaling up data provides limited improvements for EAGLE. We identify that this limitation arises from EAGLE’s feature prediction constraints. In this paper, we introduce EAGLE-3, which abandons feature prediction in favor of direct token prediction and replaces reliance on top-layer features with multi-layer feature fusion via a technique named training-time test. These improvements significantly enhance performance and enable the draft model to fully benefit from scaling up training data. Our experiments include both chat models and reasoning models, evaluated on five tasks. The results show that EAGLE-3 achieves a speedup ratio up to 6.5x, with about 1.4x improvement over EAGLE-2. In the SGLang framework, EAGLE- 3 achieves a 1.38x throughput improvement at a batch size of 64.



Paperid:4405
Authors:Litao Guo, Xinli Xu, Luozhou Wang, Jiantao Lin, Jinsong Zhou, Zixin Zhang, Bolan Su, Yingcong Chen
Abstract:
With the rapid advancement of generative models, general-purpose generation has gained increasing attention as a promising approach to unify diverse tasks across modalities within a single system. Despite this progress, existing open-source frameworks often remain fragile and struggle to support complex real-world applications due to the lack of structured workflow planning and execution-level feedback. To address these limitations, we present ComfyMind, a collaborative AI system designed to enable robust and scalable general-purpose generation, built on the ComfyUI platform. ComfyMind introduces two core innovations: Semantic Workflow Interface (SWI) that abstracts low-level node graphs into callable functional modules described in natural language, enabling high-level composition and reducing structural errors; Search Tree Planning mechanism with localized feedback execution, which models generation as a hierarchical decision process and allows adaptive correction at each stage. Together, these components improve the stability and flexibility of complex generative workflows. We evaluate ComfyMind on three public benchmarks: ComfyBench, GenEval, and Reason-Edit, which span generation, editing, and reasoning tasks. Results show that ComfyMind consistently outperforms existing open-source baselines and achieves performance comparable to GPT-Image-1. ComfyMind paves a promising path for the development of open-source general-purpose generative AI systems.



Paperid:4406
Authors:Yu Chen, Rolandos Alexandros Potamias, Evangelos Ververas, Jifei Song, Jiankang Deng, Gim Hee Lee
Abstract:
Neural radiance fields (NeRF) and 3D Gaussian Splatting (3DGS) are popular techniques to reconstruct and render photorealistic images. However, the prerequisite of running Structure-from-Motion (SfM) to get camera poses limits their completeness. Although previous methods can reconstruct a few unposed images, they are not applicable when images are unordered or densely captured. In this work, we propose a method to train 3DGS from unposed images. Our method leverages a pre-trained 3D geometric foundation model as the neural scene representation. Since the accuracy of the predicted pointmaps does not suffice for accurate image registration and high-fidelity image rendering, we propose to mitigate the issue by initializing and fine-tuning the pre-trained model from a seed image. The images are then progressively registered and added to the training buffer, which is used to train the model further. We also propose to refine the camera poses and pointmaps by minimizing a point-to-camera ray consistency loss across multiple views. When evaluated on diverse challenging datasets, our method outperforms state-of-the-art pose-free NeRF/3DGS methods in terms of both camera pose accuracy and novel view synthesis, and even renders higher fidelity images than 3DGS trained with COLMAP poses.



Paperid:4407
Authors:Amir Hossein Rahmati, Sanket Jantre, Weifeng Zhang, Yucheng Wang, Byung-Jun Yoon, Nathan Urban, Xiaoning Qian
Abstract:
Low-Rank Adaptation (LoRA) offers a cost-effective solution for fine-tuning large language models (LLMs), but it often produces overconfident predictions in data-scarce few-shot settings. To address this issue, several classical statistical learning approaches have been repurposed for scalable uncertainty-aware LoRA fine-tuning. However, these approaches neglect how input characteristics affect the predictive uncertainty estimates. To address this limitation, we propose Contextual Low-Rank Adaptation (C-LoRA) as a novel uncertainty-aware and parameter efficient fine-tuning approach, by developing new lightweight LoRA modules contextualized to each input data sample to dynamically adapt uncertainty estimates. Incorporating data-driven contexts into the parameter posteriors, C-LoRA mitigates overfitting, achieves well-calibrated uncertainties, and yields robust predictions. Extensive experiments demonstrate that C-LoRA consistently outperforms the state-of-the-art uncertainty-aware LoRA methods in both uncertainty quantification and model generalization. Ablation studies further confirm the critical role of our contextual modules in capturing sample-specific uncertainties. C-LoRA sets a new standard for robust, uncertainty-aware LLM fine-tuning in few-shot regimes.



Paperid:4408
Authors:Giyeong Oh, Woohyun Cho, 시열 김, Suhwan Choi, Youngjae Yu
Abstract:
Residual connections are pivotal for deep neural networks, enabling greater depth by mitigating vanishing gradients. However, in standard residual updates, the module's output is directly added to the input stream. This can lead to updates that predominantly reinforce or modulate the existing stream direction, potentially underutilizing the module's capacity for learning entirely novel features.In this work, we introduceOrthogonal Residual Update: we decompose the module's output relative to the input stream and add only the component orthogonal to this stream.This design aims to guide modules to contribute primarily new representational directions, fostering richer feature learning and more efficient training.We demonstrate that our orthogonal update strategy improves generalization accuracy and training stability across diverse architectures (ResNetV2, Vision Transformers) and datasets (CIFARs, TinyImageNet, ImageNet-1k), achieving, for instance, a +4.3\%p top-1 accuracy gain for ViT-B on ImageNet-1k.



Paperid:4409
Authors:Jie Ma, NING QU, Zhitao Gao, Xing Rui, Jun Liu, Hongbin Pei, Jiang Xie, Lingyun Song, Pinghui Wang, Jing Tao, su zhou
Abstract:
Knowledge graph-based retrieval-augmented generation seeks to mitigate hallucinations in Large Language Models (LLMs) caused by insufficient or outdated knowledge. However, existing methods often fail to fully exploit the prior knowledge embedded in knowledge graphs (KGs), particularly their structural information and explicit or implicit constraints. The former can enhance the faithfulness of LLMs' reasoning, while the latter can improve the reliability of response generations. Motivated by these, we propose a trustworthy reasoning framework, termed Deliberation over Priors (\texttt{DP}), which sufficiently utilizes the priors contained in KGs. Specifically, \texttt{DP} adopts a progressive knowledge distillation strategy that integrates structural priors into LLMs through a combination of supervised fine-tuning and Kahneman-Tversky Optimization, thereby improving the faithfulness of relation path generation. Furthermore, our framework employs a reasoning-introspection strategy, which guides LLMs to perform refined reasoning verification based on extracted constraint priors, ensuring the reliability of response generation. Extensive experiments on three benchmark datasets demonstrate that \texttt{DP} achieves new state-of-the-art performance, especially a Hit@1 improvement of 13% on the ComplexWebQuestions dataset, and generates highly trustworthy responses. We also conduct various analyses to verify its flexibility and practicality. Code is available athttps://anonymous.4open.science/r/Deliberation-on-Priors.



Paperid:4410
Authors:Yu Gui, Cong Ma, Zongming Ma
Abstract:
Multi-modal contrastive learning as a self-supervised representation learning technique has achieved great success in foundation model training, such as CLIP~\citep{radford2021learning}. In this paper, we study the theoretical properties of the learned representations from multi-modal contrastive learning beyond linear representations and specific data distributions. Our analysis reveals that, enabled by temperature optimization, multi-modal contrastive learning not only maximizes mutual information between modalities but also adapts to intrinsic dimensions of data, which can be much lower than user-specified dimensions for representation vectors. Experiments on both synthetic and real-world datasets demonstrate the ability of contrastive learning to learn low-dimensional and informative representations, bridging theoretical insights and practical performance.



Paperid:4411
Authors:Silpa Vadakkeeveetil Sreelatha, Sauradip Nag, Muhammad Awais, Serge Belongie, Anjan Dutta
Abstract:
Abstract:The rapid advancement of diffusion models has enabled high-fidelity and semantically rich text-to-image generation; however, ensuring fairness and safety remains an open challenge. Existing methods typically improve fairness and safety at the expense of semantic fidelity and image quality. In this work, we propose RespoDiff, a novel framework for responsible text-to-image generation that incorporates a dual-module transformation on the intermediate bottleneck representations of diffusion models. Our approach introduces two distinct learnable modules: one focused on capturing and enforcing responsible concepts, such as fairness and safety, and the other dedicated to maintaining semantic alignment with neutral prompts. To facilitate the dual learning process, we introduce a novel score-matching objective that enables effective coordination between the modules. Our method outperforms state-of-the-art methods in responsible generation by ensuring semantic alignment while optimizing both objectives without compromising image fidelity. Our approach improves responsible and semantically coherent generation by $\textasciitilde20\%$ across diverse, unseen prompts. Moreover, it integrates seamlessly into large-scale models like SDXL, enhancing fairness and safety. Code will be released upon acceptance.



Paperid:4412
Authors:Chaoyang Li, Runze Ye, Jianyang Qin, Jinhao Cui, Lingzhi Wang, Ning Hu, Qing Liao
Abstract:
Current parameter-efficient fine-tuning (PEFT) methods have shown superior performance in continual learning. However, most existing PEFT-based methods focus on mitigating catastrophic forgetting by limiting modifications to the old task model caused by new tasks. This hinders backward knowledge transfer, as when new tasks have a strong positive correlation with old tasks, appropriately training on new tasks can transfer beneficial knowledge to old tasks. Critically, achieving backward knowledge transfer faces two fundamental challenges: (1) some parameters may be ineffective on task performance, which constrains the task solution space and model capacity; (2) since old task data are inaccessible, modeling task correlation via shared data is infeasible. To address these challenges, we propose CaLoRA, a novel \textbf{c}ausal-\textbf{a}ware \textbf{lo}w-\textbf{r}ank \textbf{a}daptation framework that is the first PEFT-based continual learning work with backward knowledge transfer. Specifically, we first propose \textbf{p}ar\textbf{a}meter-level \textbf{c}ounterfactual \textbf{a}ttribution (PaCA) that estimates the causal effect of LoRA parameters via counterfactual reasoning, identifying effective parameters from a causal view. Second, we propose \textbf{c}ross-t\textbf{a}sk \textbf{g}radient \textbf{a}daptation (CaGA) to quantify task correlation by gradient projection and evaluate task affinity based on gradient similarity. By incorporating causal effect, task correlation, and affinity, CaGA adaptively adjusts task gradients, facilitating backward knowledge transfer without relying on data replay. Extensive experiments across multiple benchmarks and continual learning settings show that CaLoRA outperforms state-of-the-art methods. In particular, CaLoRA better mitigates catastrophic forgetting by enabling positive backward knowledge transfer.



Paperid:4413
Authors:Alicia Zeng, Jack Gallant
Abstract:
Encoding models that use word embeddings or artificial neural network (ANN) representations to predict brain activity achieve high predictive performance but offer limited interpretability. A key challenge is that dense feature representations encode many distinct concepts along overlapping directions due to superposition, where multiple semantic features are compressed into correlated directions in the embedding space. This structure creates ambiguity in interpreting model weights, as different combinations of semantic directions can produce the same predicted brain response. To address this challenge, we introduce the Sparse Concept Encoding Model, which transforms dense embeddings into a higher-dimensional, sparse space composed of learned, interpretable concept atoms. Each stimulus is re-encoded as a sparse, non-negative combination of these atoms, enabling voxel responses to be modeled in an axis-aligned semantic basis where each dimension corresponds to a distinct concept. Applied to naturalistic fMRI data, the Sparse Concept Encoding Model matches the prediction accuracy of conventional dense models while offering substantially greater interpretability. It enables novel neuroscientific analyses that are infeasible with dense features, such as identifying overlapping cortical representations of time, space, and number. This work offers a more interpretable framework for linking language representations in ANNs to conceptual representations in the brain.



Paperid:4414
Authors:Jie Cheng, Ruixi Qiao, Lijun Li, Chao Guo, Junle Wang, Gang Xiong, Yisheng Lv, Fei-Yue Wang
Abstract:
Process reward model (PRM) has been proven effective in test-time scaling of LLM on challenging reasoning tasks. However, the reward hacking induced by PRM hinders its successful applications in reinforcement fine-tuning. We find the primary cause of reward hacking induced by PRM is that: the canonical summation-form credit assignment in reinforcement learning (RL), i.e. cumulative gamma-decayed future rewards, causes the LLM to hack steps with high rewards. Therefore, to unleashing the power of PRM in training-time, we propose PURE: Process sUpervised Reinforcement lEarning. The core of PURE is the min-form credit assignment that defines the value function as the minimum future rewards. This method unifies the optimization objective with respect to process rewards during test-time and training-time, and significantly alleviates reward hacking due to the limits on the range of values of value function and more rational assignment of advantages. Through extensively experiments on 3 base models, we achieve similar reasoning performance using PRM-based approach compared with verifiable reward-based approach if enabling min-form credit assignment. In contrast, the canonical sum-form credit assignment even collapses training at the beginning. Moreover, when we incorporate 1/10th verifiable rewards to auxiliary the PRM-based fine-tuning, it further alleviate reward hacking and results in the best fine-tuned model based on Qwen2.5-Math-7B with 82.5% accuracy on AMC23 and 53.3% average accuracy across 5 benchmarks. Furthermore, we summary the reward hacking cases we encountered during training and analysis the cause of training collapse.



Paperid:4415
Authors:Siyu Chen, Ting Han, Chengzheng Fu, Changshe Zhang, Chaolei Wang, Jinhe Su, Guorong Cai, Meiliu Wu
Abstract:
Open-Vocabulary semantic segmentation (OVSS) and domain generalization in semantic segmentation (DGSS) highlight a subtle complementarity that motivates Open-Vocabulary Domain-Generalized Semantic Segmentation (OV-DGSS). OV-DGSS aims to generate pixel-level masks for unseen categories while maintaining robustness across unseen domains, a critical capability for real-world scenarios such as autonomous driving in adverse conditions. We introduce Vireo, a novel single-stage framework for OV-DGSS that unifies the strengths of OVSS and DGSS for the first time. Vireo builds upon frozen Visual Foundation Models (VFMs) and incorporates scene geometry via Depth VFMs to extract domain-invariant structural features. To bridge the gap between visual and textual modalities under domain shift, we propose three key components: (1) GeoText Prompts, which align geometric features with language cues and progressively refine encoder representations; (2) Coarse Mask Prior Embedding (CMPE) for enhancing gradient flow for faster convergence and stronger textual influence; and (3) the Domain-Open-Vocabulary Vector Embedding Head (DOV-VEH), which fuses refined structural and semantic features for robust prediction. Comprehensive evaluation on these components demonstrates the effectiveness of our designs. Our proposed Vireo achieves the state-of-the-art performance and surpasses existing methods by a large margin in both domain generalization and open-vocabulary recognition, offering a unified and scalable solution for robust visual understanding in diverse and dynamic environments. Code is available at https://github.com/anonymouse-9c53tp182bvz/Vireo.



Paperid:4416
Authors:Hongyi Zhou, Weiran Liao, Xi Huang, Yucheng Tang, Fabian Otto, Xiaogang Jia, Xinkai Jiang, Simon Hilber, Ge Li, Qian Wang, Ömer Yağmurlu, Nils Blank, Moritz Reuss, Rudolf Lioutikov
Abstract:
We present the B-spline Encoded Action Sequence Tokenizer(BEAST), a novel action tokenizer that encodes action sequences into compact discrete or continuous tokens using B-splines. In contrast to existing action tokenizers based on vector quantization or byte pair encoding, BEAST requires no separate tokenizer training and consistently produces tokens of uniform length, enabling fast action sequence generation via parallel decoding. Leveraging our B-spline formulation, BEAST inherently ensures generating smooth trajectories without discontinuities between adjacent segments. We extensively evaluate BEAST by integrating it with three distinct model architectures: a Variational Autoencoder (VAE) with continuous tokens, a decoder-only Transformer with discrete tokens, and Florence-2, a pretrained Vision-Language Model with an encoder-decoder architecture, demonstrating BEAST's compatibility and scalability with large pretrained models. We evaluate BEAST across three established benchmarks consisting of 166 simulated tasks and on three distinct robot settings with a total of 8 real-world tasks. Experimental results demonstrate that BEAST (i) significantly reduces both training and inference computational costs, and (ii) consistently generates smooth, high-frequency control signals suitable for continuous control tasks while (iii) reliably achieves competitive task success rates compared to state-of-the-art methods.



Paperid:4417
Authors:Junhao Dong, Cong Zhang, Xinghua Qu, Piotr Koniusz, Zejun MA, Yew Soon Ong
Abstract:
A spectrum of well-established studies has demonstrated the superhuman capabilities of modern Vision-Language Models (VLMs) across a wide range of tasks, inducing a growing doubt about continuing to obtain reliable supervision from human annotators. Therefore, concerns are growing about stagnating model performance due to scarce high-quality labelled data. To address this challenge, ``superalignment'' motivates the study of weak-to-strong generalization, where the supervision from a weak model can elicit generalizable knowledge from a strong one. While effective in aligning capabilities, we reveal that the standard weak-to-strong generalization scheme typically fails to confer robustness, exposing strong VLMs to underlying security threats brought by adversarial examples. Through systematic analysis, we attribute this failure mode to whether adversaries are incorporated into the learning objective during alignment. To address this, we propose the first adversarially robust weak-to-strong generalization method to elicit the zero-shot robustness from large-scale models in an unsupervised scheme by mitigating the unreliable supervision from two perspectives: alignment re-weighting and reference refining. We provide theoretical analyses characterizing the condition under which robustness generalization is achievable. Extensive experiments across various vision-language benchmarks validate the effectiveness and generalizability of our method, demonstrating its plug-and-play application to large-scale VLMs.



Paperid:4418
Authors:Maya Varma, Jean-Benoit Delbrouck, Sophie Ostmeier, Akshay Chaudhari, Curtis Langlotz
Abstract:
Vision-language models (VLMs) have made great strides in addressing temporal understanding tasks, which involve characterizing visual changes across a sequence of images. However, recent works have suggested that when making predictions, VLMs may rely on static feature biases, such as background or object features, rather than dynamic visual changes. Static feature biases are a type of shortcut and can contribute to systematic prediction errors on downstream tasks; as a result, identifying and characterizing error-inducing static feature biases is critical prior to real-world model deployment. Existing approaches for identifying such systematic failure modes in trained models (i) are typically designed for non-temporal settings and (ii) are challenging to evaluate in temporal settings due to the lack of quantitative evaluation frameworks. In this work, we address these challenges by introducing TRoVe, an automated approach for discovering error-inducing static feature biases learned by temporal VLMs. Given a trained VLM and an annotated validation dataset associated with a downstream classification task, TRoVe extracts candidate static features from the dataset and scores each feature by (i) the effect of the feature on classification errors as well as (ii) the extent to which the VLM relies on the feature when making predictions. In order to quantitatively evaluate TRoVe, we introduce an evaluation framework consisting of 101 trained temporal VLMs paired with ground-truth annotations for learned static feature biases. We use this framework to demonstrate that TRoVe can accurately identify error-inducing static feature biases in VLMs, achieving a 28.6% improvement over the closest baseline. Finally, we apply TRoVe to 7 off-the-shelf VLMs and 2 temporal understanding tasks, surfacing previously-unknown static feature biases and demonstrating that knowledge of learned biases can aid in improving model performance at test time.



Paperid:4419
Authors:Zhida Zhao, Talas Fu, Yifan Wang, Lijun Wang, Huchuan Lu
Abstract:
Despite remarkable progress in driving world models, their potential for autonomous systems remains largely untapped: the world models are mostly learned for world simulation and decoupled from trajectory planning. While recent efforts aim to unify world modeling and planning in a single framework, the synergistic facilitation mechanism of world modeling for planning still requires further exploration. In this work, we introduce a new driving paradigm named Policy World Model (PWM), which not only integrates world modeling and trajectory planning within a unified architecture, but is also able to benefit planning using the learned world knowledge through the proposed action-free future state forecasting scheme. Through collaborative state-action prediction, PWM can mimic the human-like anticipatory perception, yielding more reliable planning performance. To facilitate the efficiency of video forecasting, we further introduce a new image tokenizer with context-guided compression and decoding alongside a dynamic focal loss. Despite utilizing only front camera input, our method matches or exceeds state-of-the-art approaches that rely on multi-view and multi-modal inputs. Code and model weights will be released.



Paperid:4420
Authors:Ziye Wang, Li Kang, Yiran Qin, Jiahua Ma, zhanglin peng, LEI BAI, Ruimao Zhang
Abstract:
Despite significant advances in robotic policy generation, effective coordination in embodied multi-agent systems remains a fundamental challenge—particularly in scenarios where agents must balance individual perspectives with global environmental awareness.Existing approaches often struggle to balance fine-grained local control with comprehensive scene understanding, resulting in limited scalability and compromised collaboration quality.In this paper, we present GauDP, a novel Gaussian-image synergistic representation that facilitates scalable, perception-aware imitation learning in multi-agent collaborative systems.Specifically, GauDP reconstructs a globally consistent 3D Gaussian field from local-view RGB images, allowing all agents to dynamically query task-relevant features from a shared scene representation. This design facilitates both fine-grained control and globally coherent behavior without requiring additional sensing modalities. We evaluate GauDP on the RoboFactory benchmark, which includes diverse multi-arm manipulation tasks. Our method achieves superior performance over existing image-based methods and approaches the effectiveness of point-cloud-driven methods, while maintaining strong scalability as the number of agents increases. Extensive ablations and visualizations further demonstrate the robustness and efficiency of our unified local-global perception framework for multi-agent embodied learning.



Paperid:4421
Authors:Pei Zhou, Wanting Yao, Qian Luo, Xunzhe Zhou, Yanchao Yang
Abstract:
Goal-conditioned policy learning for robotic manipulation presents significant challenges in maintaining performance across diverse objectives and environments. We introduceHyper-GoalNet, a framework that generates task-specific policy network parameters from goal specifications using hypernetworks. Unlike conventional methods that simply condition fixed networks on goal-state pairs, our approach separates goal interpretation from state processing -- the former determines network parameters while the latter applies these parameters to current observations. To enhance representation quality for effective policy generation, we implement two complementary constraints on the latent space: (1) a forward dynamics model that promotes state transition predictability, and (2) a distance-based constraint ensuring monotonic progression toward goal states. We evaluate our method on a comprehensive suite of manipulation tasks with varying environmental randomization. Results demonstrate significant performance improvements over state-of-the-art methods, particularly in high-variability conditions. Real-world robotic experiments further validate our method's robustness to sensor noise and physical uncertainties. Our code and trained models will be made publicly available.



Paperid:4422
Authors:Xiyuan Jin, Jing Wang, Ziwei Lin, QIANRU JIA, Yuqing Huang, Xiaojun Ning, Zhonghua Shi, Youfang Lin
Abstract:
Medical time-series analysis differ fundamentally from general ones by requiring specialized domain knowledge to interpret complex signals and clinical context. Large language models (LLMs) hold great promise for augmenting medical time-series analysis by complementing raw series with rich contextual knowledge drawn from biomedical literature and clinical guidelines. However, this potential can only be realized with prompts that are constructed precisely and meaningfully, guiding the LLM’s attention to the most relevant aspects of the data. Yet, determining what constitutes effective prompt content remains non-trivial—especially in medical settings where signal interpretation often hinges on subtle, expert-defined decision-making indicators. To address this challenge, we introduce InDiGO, a knowledge-aware evolutionary learning framework that systematically integrates clinical signals and decision-making indicators, continuously optimizing through iterative learning. Experiments across four real-world medical benchmarks show that InDiGO consistently outperforms existing state-of-the-art methods. The code is available at: https://anonymous.4open.science/r/InDiGO-5034.



Paperid:4423
Authors:Wenjun Ji, Yuxiang Fu, Luyang Ying, Deng-Ping Fan, Yuyi Wang, Ming-Ming Cheng, Ivor Tsang, Qing Guo
Abstract:
Cutting-edge works have demonstrated that text-to-image (T2I) diffusion models can generate adversarial patches that mislead state-of-the-art object detectors in the physical world, revealing detectors' vulnerabilities and risks. However, these methods neglect the adversarial patches' attack effectiveness when observed from different views in the physical world (\ie, angle robustness of the adversarial patches). In this paper, for the first time, we study the angle robustness of generated patches comprehensively, revealing the angle-robust issues of existing works and demonstrating that input texts affect the angle robustness of generated patches significantly. Motivated by the studies, we introduce Angle-Robust Concept Learning (AngleRoCL), a novel approach that learns a generalizable concept (\ie, specialized text embeddings in implementation) representing the capability of generating angle-robust patches. The learned concept can be incorporated into text prompts and guides T2I models to generate patches with their attack effectiveness inherently resistant to viewpoint variations. Through extensive simulation and physical-world experiments across multiple observation views, we demonstrate that AngleRoCL significantly enhances the angle robustness of generated patches compared to baseline methods. Our patches maintain high attack success rates even under challenging viewing conditions, with an average improvement of xxx in attack effectiveness across multiple angles. This research advances the understanding of physically angle-robust patches and provides insights into the relationship between textual concepts and physical properties in T2I-generated contents.



Paperid:4424
Authors:Mengjiao Ma, Qi Ma, Yue Li, Jiahuan Cheng, Runyi Yang, Bin Ren, Nikola Popovic, Mingqiang Wei, Nicu Sebe, Ender Konukoglu, Luc V Gool, Theo Gevers, Martin R. Oswald, Danda Pani Paudel
Abstract:
3D Gaussian Splatting (3DGS) serves as a highly performant and efficient encoding of scene geometry, appearance, and semantics. Moreover, grounding language in 3D scenes has proven to be an effective strategy for 3D scene understanding. Current Language Gaussian Splatting line of work fall into three main groups: (i) per-scene optimization-based, (ii) per-scene optimization-free, and (iii) generalizable approach. However, most of them are evaluated only on rendered 2D views of a handful of scenes and viewpoints close to the training views, limiting ability and insight into holistic 3D understanding. To address this gap, we propose the first large-scale benchmark that systematically assesses these three groups of methods directly in 3D space, evaluating on 1060 scenes across three indoor datasets and one outdoor dataset. Benchmark results demonstrate a clear advantage of the generalizable paradigm, particularly in relaxing the scene-specific limitation, enabling fast feed-forward inference on novel scenes, and achieving superior segmentation performance. We further introduce GaussianWorld-49K -- a carefully curated 3DGS dataset comprising of around 49K diverse indoor and outdoor scenes trained from multiple sources, with which we demonstrate generalizable approach could harness strong data priors. Our code, benchmark and datasets will be made public to accelerate research in generalizable 3DGS scene understanding.



Paperid:4425
Authors:Yanwei Ren, Haotian Zhang, Fuxiang Wu, Jiayan Qiu, Jiaxing Huang, Baosheng Yu, Liu Liu
Abstract:
Enhancing large language models by simply scaling up datasets has begun to yield diminishing returns, shifting the spotlight to data quality. Monte Carlo Tree Search (MCTS) has emerged as a powerful technique for generating high-quality chain-of-thought data, yet conventional approaches typically retain only the top-scoring trajectory from the search tree, discarding sibling nodes that often contain valuable partial insights, recurrent error patterns, and alternative reasoning strategies. This unconditional rejection of non-optimal reasoning branches may waste vast amounts of informative data in the whole search tree. We propose SIGMA (Sibling Guided Monte Carlo Augmentation), a novel framework that reintegrates these discarded sibling nodes to refine LLM reasoning. SIGMA forges semantic links among sibling nodes along each search path and applies a two-stage refinement: a critique model identifies overlooked strengths and weaknesses across the sibling set, and a revision model conducts text-based backpropagation to refine the top-scoring trajectory in light of this comparative feedback. By recovering and amplifying the underutilized but valuable signals from non-optimal reasoning branches, SIGMA substantially improves reasoning trajectories. On the challenging MATH benchmark, our SIGMA-tuned 7B model achieves 54.92\% accuracy using only 30K samples, outperforming state-of-the-art models trained on 590K samples. This result highlights that our sibling-guided optimization not only significantly reduces data usage but also significantly boosts LLM reasoning.



Paperid:4426
Authors:Borong Zhang, Yuhao Zhang, Jiaming Ji, Yingshan Lei, Juntao Dai, Yuanpei Chen, Yaodong Yang
Abstract:
Vision-language-action models (VLAs) show potential as generalist robot policies. However, these models pose extreme safety challenges during real-world deployment, including the risk of harm to the environment, the robot itself, and humans.How can safety constraints be explicitly integrated into VLAs?We address this by exploring an integrated safety approach (ISA), systematicallymodelingsafety requirements, then activelyelicitingdiverse unsafe behaviors, effectivelyconstrainingVLA policies via safe reinforcement learning, and rigorouslyassuringtheir safety through targeted evaluations. Leveraging the constrained Markov decision process (CMDP) paradigm, ISA optimizes VLAs from a min-max perspective against elicited safety risks. Thus, policies aligned through this comprehensive approach achieve the following key features: (I) effectivesafety-performance trade-offs, this exploration yields an 83.58\% safety improvement compared to the current state-of-the-art method, while also maintaining task performance (+3.85\%). (II) strongsafety assurance, with the ability to mitigate long-tail risks and handle extreme failure scenarios. (III) robustgeneralizationof learned safety behaviors to various out-of-distribution perturbations.



Paperid:4427
Authors:Idan Attias, Steve Hanneke, Arvind Ramaswami
Abstract:
Abstract:We study online and transductive online learning in settings where the learner can interact with the concept class only via Empirical Risk Minimization (ERM) or weak consistency oracles on arbitrary subsets of the instance domain. This contrasts with standard online models, where the learner has full knowledge of the concept class. The ERM oracle returns a hypothesis that minimizes the loss on a given subset, while the weak consistency oracle returns only a binary signal indicating whether the subset is realizable by a concept in the class. The learner's performance is measured by the number of mistakes and oracle calls.In the standard online setting with ERM access, we establish tight lower bounds in both the realizable and agnostic cases: $\Omega(2^{d_\mathrm{LD}})$ mistakes and $\Omega(\sqrt{T 2^{d_\mathrm{LD}}})$ regret, respectively, where $T$ is the number of timesteps and $d_\mathrm{LD}$ is the Littlestone dimension of the class. We further show how existing results for online learning with ERM access translate to the setting with a weak consistency oracle, at the cost of increasing the number of oracle calls by $O(T)$.We then consider the transductive online model, where the instance sequence is known in advance but labels are revealed sequentially. For general Littlestone classes, we show that the optimal mistake bound in the realizable case and in the agnostic case can be achieved using $O(T^{d_\mathrm{VC}+1})$ oracle calls, where $d_\mathrm{VC}$ is the VC dimension of the class. On the negative side, we show that a mistake bound of $\Omega(2^{d_\mathrm{LD}})$ is unavoidable for deterministic learners restricted to $O(T)$ ERM queries, and a mistake bound of $\Omega(T)$ is unavoidable for randomized learners restricted to $O(T)$ weak consistency queries.%Finally, for special families of concept classes, we demonstrate how to reduce the number of oracle calls using randomized algorithms while maintaining similar mistake bounds. In particular, for Thresholds on an unknown ordering, $O(\log T)$ ERM queries suffice, and for $k$-Intervals, $O(T^3 2^{2k})$ weak consistency queries suffice.



Paperid:4428
Authors:Sang Hoon Han, Seonho Lee, Hyeok Nam, Jae Hyeon Park, Min Hee Cha, Min Geol Kim, Hyunse Lee, Sangyeon Ahn, Moonju Chae, Sung In Cho
Abstract:
Skeleton-based hand gesture recognition plays a crucial role in enabling intuitive human–computer interaction. Traditional skeleton-based gesture recognition methods have primarily relied on hand-crafted features—such as distances between joint coordinates within the same frame or positional changes of joints across different frames—to alleviate issues arising from variations in sensor viewpoints or individual body proportions. However, hand-crafted features fail to capture the comprehensive information embedded in raw skeleton data required for effective hand gesture recognition, and they exhibit poor interpretability. Moreover, the normalization process in traditional methods, which depends on training data, introduces a domain gap between training and testing data, thereby posing challenges in applying the trained model to diverse test environments. To overcome these issues, our study proposes a novel approach for real-time gesture detection in skeleton-based hand gesture recognition. Specifically, we exclude traditional handcrafted features and introduce a new method called Skeleton Kinematics Extraction Through Coordinated grapH (SKETCH), which transforms four-dimensional (time, x, y, and z) skeleton raw data into a simple and intuitive visual representation. By applying our novel Dynamic Range Embedding (DRE), which encodes axis-wise relative movement range information from each x, y, and z coordinate, SKETCH produces a graph image that richly captures the raw data’s inherent information. The pre-trained vision model is then fine-tuned using the SKETCH to detect and classify hand gestures. The visual graph facilitates direct interpretation of the model’s focus within the raw input data. Moreover, the proposed method performs independent min-max normalization on a fixed-length segment of the incoming sequence in real time, mitigating performance degradation due to fluctuations in absolute coordinates caused by differences in sensor viewpoints and individual body proportions. Experimental results demonstrate that our approach outperforms existing methods on the SHREC’19 and SHREC’22 benchmarks while ensuring both robustness and interpretability.



Paperid:4429
Authors:Yongqiang Yao, Jingru Tan, Kaihuan Liang, Feizhao Zhang, Jiahao Hu, Yazhe Niu, Shuo Wu, Ruihao Gong, Dahua Lin, Ningyi Xu
Abstract:
Abstract:Training Long-Context Large Language Models (LLMs) is challenging, as hybrid training with long-context and short-context data often leads to workload imbalances. Existing works mainly use data packing to alleviate this issue, but fail to consider imbalanced attention computation and wasted communication overhead. This paper proposes Hierarchical Balance Packing (HBP), which designs a novel batch-construction method and training recipe to address those inefficiencies. In particular, the HBP constructs multi-level data packing groups, each optimized with a distinct packing length. It assigns training samples to their optimal groups and configures each group with the most effective settings, including sequential parallelism degree and gradient checkpointing configuration. To effectively utilize multi-level groups of data, we design a dynamic training pipeline specifically tailored to HBP, including curriculum learning, adaptive sequential parallelism, and stable loss. Our extensive experiments demonstrate that our method significantly reduces training time over multiple datasets and open-source models while maintaining strong performance. For the largest DeepSeek-V2 (236B) MOE model, our method speeds up the training by 2.4$\times$ with competitive performance.



Paperid:4430
Authors:Yun Xing, Yue Cao, Nhat Chung, Jie Zhang, Ivor Tsang, Ming-Ming Cheng, Yang Liu, Lei Ma, Qing Guo
Abstract:
Depth estimation is a critical task in autonomous driving and robotics, where inaccuracies (such as misidentifying nearby objects as distant) can lead to dangerous situations. Adversarial attacks against stereo depth estimation can help reveal vulnerabilities before deployment. Previous work has shown that repeating optimized textures within patches can effectively mislead stereo depth estimation in digital settings, i.e., when digitally inserted into images. However, our research reveals that these naively repeated texture structures perform poorly in physical world implementations, limiting their practical utility for testing depth estimation systems. In this work, for the first time, we discover that introducing regular intervals between repeated textures, creating a striped structure, significantly enhances physical-world effectiveness. Through extensive experimentation, we analyze how variations of this novel structure influence attack performance. Based on these insights, we develop a novel stereo depth attack that jointly optimizes both the striped structure and texture elements. We also discover that binary black and white textures demonstrate substantially higher effectiveness than colorful textures. Our generated adversarial patches can be inserted into any scenes and successfully attack state-of-the-art stereo depth estimation methods and even commercial RGB-D cameras (Intel RealSense) in real-world conditions, demonstrating their practical relevance for security assessment of depth estimation systems.



Paperid:4431
Authors:Marius Potfer, Vianney Perchet
Abstract:
Abstract:Repeated multi-unit auctions, where a seller allocates multiple identical items over many rounds, are common mechanisms in electricity markets and treasury auctions. We compare the two predominant formats: uniform-price and discriminatory auctions, focusing on the perspective of a single bidder learning to bid against stochastic adversaries. We characterize the learning difficulty in each format, showing that the regret scales similarly for both auction formats under both full-information and bandit feedback, as $\tilde{\Theta} ( \sqrt{T} )$ and $\tilde{\Theta} ( T^{2/3} )$, respectively. However, analysis beyond worst-case regret reveals structural differences: uniform-price auctions may admit faster learning rates, with regret scaling as $\tilde{\Theta} ( \sqrt{T} )$ in settings where discriminatory auctions remain at $\tilde{\Theta} ( T^{2/3} )$. Finally, we provide a specific analysis for auctions in which the other participants are symmetric and have unit-demand, and show that in these instances a similar regret rate separation appears.



Paperid:4432
Authors:Wenyi Yu, Siyin Wang, Xiaoyu Yang, Xianzhao Chen, Xiaohai Tian, Jun Zhang, Guangzhi Sun, Lu Lu, Yuxuan Wang, Chao Zhang
Abstract:
In order to enable fluid and natural human-machine speech interaction, existing full-duplex conversational systems often adopt modular architectures with auxiliary components such as voice activity detectors, interrupters, conversation state predictors, or multiple LLMs. These systems, however, suffer from error accumulation across modules and struggle with key challenges such as context-dependent barge-in and echo cancellation. Recent approaches, most notably Moshi, simplify the pipeline by injecting audio codecs into the token space of a single LLM. However, such methods still incur significant performance degradation when operating on the speech rather than text modality.In this paper, we introduce SALMONN-omni, the first single, standalone full-duplex speech LLM that operates without audio codecs in its token space. It features a novel dynamic thinking mechanism within the LLM backbone, enabling the model to learn when to transition between speaking and listening states. Experiments on widely used benchmarks for spoken question answering and open-domain dialogue show that SALMONN-omni achieves at least 30\% relative performance improvement over existing open-source full-duplex models and performs highly competitively to half-duplex and turn-based systems, despite using substantially less training data. Moreover, SALMONN-omni demonstrates strong performance in complex conversational scenarios, including turn-taking, backchanneling, echo cancellation and context-dependent barge-in, with further improvements achieved through reinforcement learning. Some demo conversations between user and SALMONN-omni are provided in the following repository https://anonymous.4open.science/r/SALMONN-omni-4BA8.



Paperid:4433
Authors:Yukun Jiang, Mingjie Li, Michael Backes, Yang Zhang
Abstract:
Abstract:Despite their superior performance on a wide range of domains, large language models (LLMs) remain vulnerable to misuse for generating harmful content, a risk that has been further amplified by various jailbreak attacks.Existing jailbreak attacks mainly follow sequential logic, where LLMs understand and answer each given task one by one.However, concurrency, a natural extension of the sequential scenario, has been largely overlooked.In this work, we first propose a word-level method to enable task concurrency in LLMs, where adjacent words encode divergent intents.Although LLMs maintain strong utility in answering concurrent tasks, which is demonstrated by our evaluations on mathematical and general question-answering benchmarks, we notably observe that combining a harmful task with a benign one significantly reduces the probability of it being filtered by the guardrail, showing the potential risks associated with concurrency in LLMs.Based on these findings, we introduce $\texttt{JAIL-CON}$, an iterative attack framework that $\underline{\text{JAIL}}$breaks LLMs via task $\underline{\text{CON}}$currency. Experiments on widely-used LLMs demonstrate the strong jailbreak capabilities of $\texttt{JAIL-CON}$ compared to existing attacks.Furthermore, when the guardrail is applied as a defense, compared to the sequential answers generated by previous attacks, the concurrent answers in our $\texttt{JAIL-CON}$ exhibit greater stealthiness and are less detectable by the guardrail, highlighting the unique feature of task concurrency in jailbreaking LLMs.



Paperid:4434
Authors:Shi Fu, Yingjie Wang, Yuzhu Chen, Li Shen, Dacheng Tao
Abstract:
Large generative models are increasingly trained on synthetic data from earlier generations, raising concerns aboutmodel collapse, a progressive performance decline consistently observed in empirical studies. However, theoretical understanding of recursive training dynamics and their failure modes remains limited. In this work, we theoretically show that recursive training inherently leads to exponential error growth unless mitigated by sufficient real data. Addressing the growing scarcity of real data, we introduce a self-verification mechanism enabling models to filter their outputs based on internal confidence scores without external validation. Through rigorous analysis, we derive finite-sample error bounds demonstrating that self-verification alone can prevent collapse, even in fully synthetic training regimes. Our theoretical framework extends to large language models (LLMs), characterizing the conditions under which recursive training can maintain stability without performance degradation.



Paperid:4435
Authors:Weixiang Yan, Haitian Liu, Tengxiao Wu, Qian Chen, Wen Wang, Haoyuan Chai, Jiayi Wang
Abstract:
Large language models (LLMs) have achieved significant performance progress in various natural language processing applications. However, LLMs still struggle to meet the strict requirements for accuracy and reliability in the medical field and face many challenges in clinical applications. Existing clinical diagnostic evaluation benchmarks for evaluating medical agents powered by LLMs have severe limitations. To address these limitations, we introduce ClinicalLab, a comprehensive clinical diagnosis agent alignment suite. ClinicalLab includes ClinicalBench, an end-to-end multi-departmental clinical diagnostic evaluation benchmark for evaluating medical agents and LLMs. ClinicalBench is based on real cases that cover 24 departments and 150 diseases. We ensure that ClinicalBench does not have data leakage. ClinicalLab also includes four novel metrics (ClinicalMetrics) for evaluating the effectiveness of LLMs in clinical diagnostic tasks. We evaluate 17 general and medical-domain LLMs and find that their performance varies significantly across different departments. Based on these findings, in ClinicalLab, we propose ClinicalAgent, an end-to-end clinical agent that aligns with real-world clinical diagnostic practices. We systematically investigate the performance and applicable scenarios of variants of ClinicalAgent on ClinicalBench. Our findings demonstrate the importance of aligning with modern medical practices in designing medical agents.



Paperid:4436
Authors:Yuanze Lin, Yi-Wen Chen, Yi-Hsuan Tsai, Ronald Clark, Ming-Hsuan Yang
Abstract:
Although diffusion-based models can generate high-quality and high-resolution video sequences from textual or image inputs, they lack explicit integration of geometric cues when controlling scene lighting and visual appearance across frames. To address this limitation, we propose IllumiCraft, an end-to-end diffusion framework accepting three complementary inputs: (1) high-dynamic-range (HDR) video maps for detailed lighting control; (2) synthetically relit frames with randomized illumination changes (optionally paired with a static background reference image) to provide appearance cues; and (3) 3D point tracks that capture precise 3D geometry information. By integrating the lighting, appearance, and geometry cues within a unified diffusion architecture, IllumiCraft generates temporally coherent videos aligned with user-defined prompts. It supports the background-conditioned and text-conditioned video relighting and provides better fidelity than existing controllable video generation methods.



Paperid:4437
Authors:florence regol, Joseph Cotnareanu, Theodore Glavas, Mark Coates
Abstract:
Abstract:Recent years have witnessed the emergence of a spectrum of foundation models, covering a broad range of capabilities and costs. Often, we effectively use foundation models as feature generators and train classifiers that use the outputs of these models to make decisions. In this paper, we consider an increasingly relevant setting where we have two classifier stages. The first stage has access to features $x$ and has the option to make a classification decision or defer, while incurring a cost, to a second classifier that has access to features $x$ and $z$. This is similar to the ``learning to defer'' setting, with the important difference that we train both classifiers jointly, and the second classifier has access to more information. The natural loss for this setting is an $\ell_{01c}$ loss, where a penalty is paidfor incorrect classification, as in $\ell_{01}$, but an additional penalty $c$ is paidfor consulting the second classifier. The $\ell_{01c}$ loss is unwieldyfor training. Our primary contribution in this paper is the derivationof a hinge-based surrogate loss $\ell^c_{hinge}$ that is much moreamenable to training but also satisfies the property that$\ell^c_{hinge}$-consistency implies $\ell_{01c}$-consistency.



Paperid:4438
Authors:Ruoyu Feng, Yunpeng Qi, Jinming Liu, Yixin Gao, Xin Li, Xin Jin, Zhibo Chen
Abstract:
Image compression methods are usually optimized isolatedly for human perception or machine analysis tasks. We reveal fundamental commonalities between these objectives: preserving accurate semantic information is paramount, as it directly dictates the integrity of critical information for intelligent tasks and aids human understanding. Concurrently, enhanced perceptual quality not only improves visual appeal but also, by ensuring realistic image distributions, benefits semantic feature extraction for machine tasks.Based on this insight, we propose Diff-ICMH, a generative image compression framework aiming for harmonizing machine and human vision in image compression. It ensures perceptual realism by leveraging generative priors and simultaneously guarantees semantic fidelity through the incorporation of Semantic Consistency loss (SC loss) during training.Additionally, we introduce the Tag Guidance Module (TGM) that leverages highly semantic image-level tags to stimulate the pre-trained diffusion model's generative capabilities, requiring minimal additional bit rates. Consequently, Diff-ICMH supports multiple intelligent tasks through a single codec and bitstream without any task-specific adaptation, while preserving high-quality visual experience for human perception. Extensive experimental results demonstrate Diff-ICMH's superiority and generalizability across diverse tasks, while maintaining visual appeal for human perception.



Paperid:4439
Authors:Weijie Zhou, Xuantang Xiong, Yi Peng, Manli Tao, Chaoyang Zhao, Honghui Dong, Ming Tang, Jinqiao Wang
Abstract:
Visual reasoning in multimodal large language models (MLLMs) has primarily been studied in static, fully observable settings, limiting their effectiveness in real-world environments where information is often incomplete due to occlusion or limited field of view. Humans, in contrast, actively explore and interact with their environment—moving, examining, and manipulating objects—to gather information through a closed-loop process integrating perception, reasoning, and action. Inspired by this human capability, we introduce the Active Visual Reasoning (AVR) task, extending visual reasoning to partially observable, interactive environments. AVR necessitates agents to: (1) actively acquire information via sequential physical actions, (2) integrate observations across multiple steps for coherent reasoning, and (3) dynamically adjust decisions based on evolving visual feedback. To rigorously evaluate AVR, we introduce CLEVR-AVR, a simulation benchmark featuring multi-round interactive environments designed to assess both reasoning correctness and information-gathering efficiency. We present AVR-152k, a large-scale dataset offers rich Chain-of-Thought (CoT) annotations detailing iterative reasoning for uncertainty identification, action-conditioned information gain prediction, and information-maximizing action selection, crucial for training agents in a higher-order Markov Decision Process. Building on this, we develop PhysVLM-AVR, an MLLM achieving state-of-the-art performance on CLEVR-AVR, embodied reasoning (OpenEQA, RoboVQA), and passive visual reasoning (GeoMath, Geometry30K). Our analysis also reveals that current embodied MLLMs, despite detecting information incompleteness, struggle to actively acquire and integrate new information through interaction, highlighting a fundamental gap in active reasoning capabilities.



Paperid:4440
Authors:Xiaohui Wang, Peng Ye, Chenyu Huang, Shenghe Zheng, Bo Zhang, LEI BAI, Wanli Ouyang, Tao Chen
Abstract:
Abstract:With the rise of the fine-tuned–pretrained paradigm, storing numerous fine-tuned models for multi-tasking creates significant storage overhead.Delta compression alleviates this by storing only the pretrained model and the highly compressed delta weights (the differences between fine-tuned and pretrained model weights). However, existing methods fail to maintain both high compression and performance, and often rely on data.To address these challenges, we propose UltraDelta, the first data-free delta compression pipeline that achieves both ultra-high compression and strong performance. UltraDelta is designed to minimize redundancy, maximize information, and stabilize performance across inter-layer, intra-layer, and global dimensions, using three key components:(1) Variance-Based Mixed Sparsity Allocation assigns sparsity based on variance, giving lower sparsity to high-variance layers to preserve inter-layer information.(2) Distribution-Aware Compression applies uniform quantization and then groups parameters by value, followed by group-wise pruning, to better preserve intra-layer distribution.(3) Trace-Norm-Guided Rescaling uses the trace norm of delta weights to estimate a global rescaling factor, improving model stability under higher compression.Extensive experiments across (a) large language models (fine-tuned on LLaMA-2 7B and 13B) with up to 133$\times$, (b) general NLP models (RoBERTa-base, T5-base) with up to 800$\times$,(c) vision models (ViT-B/32, ViT-L/14) with up to 400$\times$, and(d) multi-modal models (BEiT-3) with 40$\times$ compression ratio, demonstrate that UltraDelta consistently outperforms existing methods, especially under ultra-high compression.



Paperid:4441
Authors:Vishnu Sarukkai, Zhiqiang Xie, Kayvon Fatahalian
Abstract:
Improving Large Language Model (LLM) agents for sequential decision-making tasks typically requires extensive task-specific knowledge engineering—custom prompts, curated examples, and specialized observation/action spaces. We investigate a different approach where agents automatically improve by learning from their own successful experiences without human intervention. Our method constructs and refines a database of self-generated trajectories that serve as in-context examples for future tasks. Even naive accumulation of successful trajectories yields substantial performance gains across three diverse benchmarks: ALFWorld (73\% to 89\%), Wordcraft (55\% to 64\%), and InterCode-SQL (75\% to 79\%). These improvements exceed those achieved by upgrading from gpt-4o-mini to gpt-4o and match the performance of allowing multiple attempts per task. We further enhance this approach with two innovations: database-level curation using population-based training to propagate high-performing example collections, and exemplar-level curation that selectively retains trajectories based on their empirical utility as in-context examples. With these enhancements, our method achieves 93\% success on ALFWorld—surpassing approaches that use more powerful LLMs and hand-crafted components. Our trajectory bootstrapping technique demonstrates that agents can autonomously improve through experience, offering a scalable alternative to labor-intensive knowledge engineering.



Paperid:4442
Authors:Sreyan Ghosh, Arushi Goel, Jaehyeon Kim, Sonal Kumar, Zhifeng Kong, Sang-gil Lee, Chao-Han Yang, Ramani Duraiswami, Dinesh Manocha, Rafael Valle, Bryan Catanzaro
Abstract:
We present Audio Flamingo 3 (AF3), a fully open state-of-the-art (SOTA) large audio-language model that advances reasoning and understanding across speech, sound, and music. AF3 introduces: (i) AF-Whisper, a unified audio encoder trained using a novel strategy for joint representation learning across all 3 modalities of speech, sound, and music; (ii) flexible, on-demand thinking, allowing the model to deliberately think before answering; (iii) multi-turn, multi-audio chat; (iv) long audio understanding and reasoning (including speech) up to 10 minutes; and (v) voice-to-voice interaction. To enable these capabilities, we propose several large-scale training datasets curated using novel strategies, including AudioSkills-XL, LongAudio-XL, AF-Think, and AF-Chat, and train AF3 with a novel five-stage curriculum-based training strategy. AF3 achieves new SOTA results on over 20+ (long) audio understanding and reasoning benchmarks. We will open-source all our code, data, and checkpoints upon paper acceptance. Technical Appendix is in Supplementary Material. Demo: https://audioflamingo3.github.io.



Paperid:4443
Authors:Feichen Gan, Lu Youcun, Yingying Zhang, Yukun Liu
Abstract:
Reliable uncertainty quantification is crucial for reinforcement learning (RL) in high-stakes settings. We propose a unified conformal prediction framework for infinite-horizon policy evaluation that constructs distribution-free prediction intervals for returns in both on-policy and off-policy settings. Our method integrates distributional RL with conformal calibration, addressing challenges such as unobserved returns, temporal dependencies, and distributional shifts. We propose a modular pseudo-return construction based on truncated rollouts and a time-aware calibration strategy using experience replay and weighted subsampling. These innovations mitigate model bias and restore approximate exchangeability, enabling uncertainty quantification even under policy shifts. Our theoretical analysis provides coverage guarantees that account for model misspecification and importance weight estimation. Empirical results, including experiments in synthetic and benchmark environments like Mountain Car, show that our method significantly improves coverage and reliability over standard distributional RL baselines.



Paperid:4444
Authors:Hao Kang, Qingru Zhang, Han Cai, Weiyuan Xu, Tushar Krishna, Yilun Du, Tsachy Weissman
Abstract:
Large language models (LLMs) have shown remarkable performance across diverse reasoning and generation tasks, and are increasingly deployed as agents in dynamic environments such as code generation and recommendation systems. However, many real-world applications, such as high-frequency trading and real-time competitive gaming, require decisions under strict latency constraints, where faster responses directly translate into higher rewards. Despite the importance of this latency–quality trade-off, it remains underexplored in the context of LLM-based agents. In this work, we present the first systematic study of this trade-off in real-time decision-making tasks. To support our investigation, we introduce two new benchmarks: HFTBench, a high-frequency trading simulation, and StreetFighter, a competitive gaming platform. Our analysis reveals that optimal latency–quality balance varies by task, and that sacrificing quality for lower latency can significantly enhance downstream performance. To address this, we propose FPX, an adaptive framework that dynamically selects model size and quantization level based on real-time demands. Our method achieves the best performance on both benchmarks, improving win rate by up to 80% in Street Fighter and boosting daily yield by up to 26.52% in trading, underscoring the need for latency-aware evaluation and deployment strategies for LLM-based agents. These results demonstrate the critical importance of latency-aware evaluation and deployment strategies for real-world LLM-based agents.



Paperid:4445
Authors:Xuelin Shen, Quan Liu, Jiayin Xu, Wenhan Yang
Abstract:
Learned Image Compression (LIC) models face critical challenges in real-world scenarios due to various environmental degradations, such as fog and rain. Due to the distribution mismatch between degraded inputs and clean training data, well-trained LIC models suffer from reduced compression efficiency, while retraining dedicated models for diverse degradation types is costly and impractical. Our method addresses the above issue by leveraging prompt learning under the information bottleneck principle, enabling compact extraction of shared components between degraded and clean images for improved latent alignment and compression efficiency. In detail, we propose an Information Bottleneck-constrained Latent Representation Unifying (IB-LRU) scheme, in which a Probabilistic Prompt Generator (PPG) is deployed to simultaneously capture the distribution of different degradations. Such a design dynamically guides the latent-representation process at the encoder through a gated modulation process. Moreover, to promote the degradation distribution capture process, the probabilistic prompt learning is guided by the Information Bottleneck (IB) principle. That is,IB constrains the information encoded in the prompt to focus solely on degradation characteristics while avoiding the inclusion of redundant image contextual information. We apply our IB-LRU method to a variety of state-of-the-art LIC backbones, and extensive experiments under various degradation scenarios demonstrate the effectiveness of our design. Our code will be publicly available.



Paperid:4446
Authors:Xinwen Zhang, Hongchang Gao
Abstract:
Abstract:Multi-level compositional optimization is a fundamental framework in machine learning with broad applications. While recent advances have addressed the stochastic optimization of minimization problems, the multi-level compositional minimax setting introduces significant new challenges—the biased nature of stochastic gradients for both the primal and dual variables. In this work, we address this gap by proposing a stochastic multi-level compositional gradient descent ascent algorithm with variance reduction, incorporating smoothed technique under the nonconvex-PL condition. We establish a convergence rate to an $(\epsilon, \epsilon/\sqrt{\kappa})$-stationary point with improved dependence on the condition number at $O\left(\kappa^{3/2}\right)$. To obtain an $\epsilon$-stationary point, we further design a stage-wise algorithm under the two-sided PL condition that enables a translation between two stationarity measures. Extensive experiments validate the effectiveness of our approach.



Paperid:4447
Authors:Jiwon Song, Dongwon Jo, Yulhwa Kim, jae-joon kim
Abstract:
Abstract:Recent reasoning-focused language models achieve high accuracy by generating lengthy intermediate reasoning paths before producing final answers.While this approach is effective in solving problems that require logical thinking, long reasoning paths significantly increase memory usage and throughput of token generation, limiting the practical deployment of such models.We propose Reasoning Path Compression (RPC), a training-free method that accelerates inference by leveraging the semantic sparsity of reasoning paths.RPC periodically compresses the KV cache by retaining KV cache that receive high importance score, which are computed using a selector window composed of recently generated queries.Experiments show that RPC improves generation throughput of QwQ-32B by up to 1.60$\times$ compared to the inference with full KV cache, with an accuracy drop of 1.2\% on the AIME 2024 benchmark.Our findings demonstrate that semantic sparsity in reasoning traces can be effectively exploited for compression, offering a practical path toward efficient deployment of reasoning LLMs.



Paperid:4448
Authors:Qinglong Hu, Qingfu Zhang
Abstract:
Large language model-assisted Evolutionary Search (LES) has emerged as a promising approach for Automated Algorithm Discovery (AAD). While many evolutionary search strategies have been developed for classic optimization problems, LES operates in abstract language spaces, presenting unique challenges for applying these strategies effectively. To address this, we propose a general LES framework that incorporates feature-assisted niche construction within abstract search spaces, enabling the seamless integration of niche-based search strategies from evolutionary computation. Building on this framework, we introduce PartEvo, an LES method that combines niche collaborative search and advanced prompting strategies to improve algorithm discovery efficiency. Experiments on both synthetic and real-world optimization problems show that PartEvo outperforms human-designed baselines and surpasses prior LES methods, such as Eoh and Funsearch. In particular, on resource scheduling tasks, PartEvo generates meta-heuristics with low design costs, achieving up to 90.1\% performance improvement over widely-used baseline algorithms, highlighting its potential for real-world applications.



Paperid:4449
Authors:Yang Cai, Haipeng Luo, Chen-Yu Wei, Weiqiang Zheng
Abstract:
Abstract:The convergence of online learning algorithms in games under self-play is a fundamental question in game theory and machine learning. Among various notions of convergence, last-iterate convergence is particularly desirable, as it reflects the actual decisions made by the learners and captures the day-to-day behavior of the learning dynamics. While many algorithms are known to converge in the average-iterate, achieving last-iterate convergence typically requires considerably more effort in both the design and the analysis of the algorithm. Somewhat surprisingly, we show in this paper that for a large family of games, there exists a simple black-box reduction that transforms the average iterates of an uncoupled learning dynamics into the last iterates of a new uncoupled learning dynamics, thus also providing a reduction from last-iterate convergence to average-iterate convergence. Our reduction applies to games where each player’s utility is linear in both their own strategy and the joint strategy of all opponents. This family includes two-player bimatrix games and generalizations such as multi-player polymatrix games. By applying our reduction to the Optimistic Multiplicative Weights Update algorithm, we obtain new state-of-the-art last-iterate convergence rates for uncoupled learning dynamics in two-player zero-sum normal-form games: (1) an $O(\frac{\log d}{T})$ last-iterate convergence rate under gradient feedback, representing an exponential improvement in the dependence on the dimension $d$ (i.e., the maximum number of actions available to either player); and (2) an $\tilde{O}(d^{\frac{1}{5}}T^{-\frac{1}{5}})$ last-iterate convergence rate under bandit feedback, improving upon the previous best rates of $\tilde{O}(\sqrt{d}T^{-\frac{1}{8}})$ and $\tilde{O}(\sqrt{d}T^{-\frac{1}{6}})$.



Paperid:4450
Authors:Shohei Enomoto
Abstract:
Visual prompting (VP) has emerged as a promising parameter-efficient fine-tuning approach for adapting pre-trained vision models to downstream tasks without modifying model parameters. Despite offering advantages like negligible computational overhead and compatibility with black-box models, conventional VP methods typically achieve lower accuracy than other adaptation approaches. Our analysis reveals two critical limitations: the restricted expressivity of simple additive transformation and a tendency toward overfitting when the parameter count increases. To address these challenges, we propose ACAVP (Affine, Color, and Additive Visual Prompting), which enhances VP's expressive power by introducing complementary transformation operations: affine transformation for creating task-specific prompt regions while preserving original image information, and color transformation for emphasizing task-relevant visual features. Additionally, we identify that overfitting is a critical issue in VP training and introduce TrivialAugment as an effective data augmentation, which not only benefits our approach but also significantly improves existing VP methods, with performance gains of up to 12 percentage points on certain datasets. This demonstrates that appropriate data augmentation is universally beneficial for VP training. Extensive experiments across twelve diverse image classification datasets with two different model architectures demonstrate that ACAVP achieves state-of-the-art accuracy among VP methods, surpasses linear probing in average accuracy, and exhibits superior robustness to distribution shifts, all while maintaining minimal computational overhead during inference.



Paperid:4451
Authors:Vineeth Dorna, Anmol Mekala, Wenlong Zhao, Andrew McCallum, J. Zico Kolter, Zachary Lipton, Pratyush Maini
Abstract:
Robust unlearning is crucial for safely deploying large language models (LLMs) in environments where data privacy, model safety, and regulatory compliance must be ensured. Yet the task is inherently challenging, partly due to difficulties in reliably measuring whether unlearning has truly occurred. Moreover, fragmentation in current methodologies and inconsistent evaluation metrics hinder comparative analysis and reproducibility. To unify and accelerate research efforts, we introduce OpenUnlearning, a standardized and extensible framework designed explicitly for benchmarking both LLM unlearning methods and metrics. OpenUnlearning integrates 9 state-of-the-art unlearning algorithms and 16 diverse evaluations across 3 leading benchmarks (TOFU, MUSE, and WMDP) and also enables analyses of forgetting behaviors across 450+ publicly released checkpoints. Leveraging OpenUnlearning, we propose a novel meta-evaluation benchmark focused specifically on assessing the faithfulness and robustness of evaluation metrics themselves. We also benchmark diverse unlearning methods and provide a comparative analysis against an extensive evaluation suite. Overall, we establish a clear, community-driven pathway toward rigorous development in LLM unlearning research.



Paperid:4452
Authors:Yi Zhang, Yushen Long, Liping Huang, Yun Ni, Xiaohong Wang, Jun Liu
Abstract:
Online ride-hailing platforms aim to deliver efficient mobility-on-demand services, often facing challenges in balancing dynamic and spatially heterogeneous supply and demand. Existing methods typically fall into two categories: reinforcement learning (RL) approaches, which suffer from data inefficiency, oversimplified modeling of real-world dynamics, and difficulty enforcing operational constraints; or decomposed online optimization methods, which rely on manually designed high-level objectives that lack awareness of low-level routing dynamics. To address this issue, we propose a novel hybrid framework that integrates large language model (LLM) with mathematical optimization in a dynamic hierarchical system: (1) it is training-free, removing the need for large-scale interaction data as in RL, and (2) it leverages LLM to bridge cognitive limitations caused by problem decomposition by adaptively generating high-level objectives. Within this framework, LLM serves as a meta-optimizer, producing semantic heuristics that guide a low-level optimizer responsible for constraint enforcement and real-time decision execution. These heuristics are refined through a closed-loop evolutionary process, driven by harmony search, which iteratively adapts the LLM prompts based on feasibility and performance feedback from the optimization layer. Extensive experiments based on scenarios derived from both the New York and Chicago taxi datasets demonstrate the effectiveness of our approach, achieving an average improvement of 16% compared to state-of-the-art baselines.



Paperid:4453
Authors:Li Liang, Bo Miao, Xinyu Wang, NAVEED AKHTAR, Jordan Vice, Ajmal Mian
Abstract:
Outdoor 3D semantic scene generation produces realistic and semantically rich environments for applications such as urban simulation and autonomous driving. However, advances in this direction are constrained by the absence of publicly available, well-annotated datasets. We introduce SketchSem3D, the first large‑scale benchmark for generating 3D outdoor semantic scenes from abstract freehand sketches and pseudo‑labeled annotations of satellite images. SketchSem3D includes two subsets, Sketch-based SemanticKITTI and Sketch-based KITTI-360 (containing LiDAR voxels along with their corresponding sketches and annotated satellite images), to enable standardized, rigorous, and diverse evaluations. We also propose Cylinder Mamba Diffusion (CymbaDiff) that significantly enhances spatial coherence in outdoor 3D scene generation. CymbaDiff imposes structured spatial ordering, explicitly captures cylindrical continuity and vertical hierarchy, and preserves both physical neighborhood relationships and global context within the generated scenes. Extensive experiments on SketchSem3D demonstrate that CymbaDiff achieves superior semantic consistency, spatial realism, and cross-dataset generalization. An anonymous download link for SketchSem3D is available here. We will make the benchmark and code public.



Paperid:4454
Authors:Benjamin Herdeanu, Juan Nathaniel, Carla Roesch, Jatan Buch, Gregor Ramien, Johannes Haux, Pierre Gentine
Abstract:
Causal discovery for dynamical systems poses a major challenge in fields where active interventions are infeasible. Most methods used to investigate these systems and their associated benchmarks are tailored to deterministic, low-dimensional and weakly nonlinear time-series data. To address these limitations, we presentCausalDynamics, a large-scale benchmark and extensible data generation framework to advance the structural discovery of dynamical causal models. Our benchmark consists of true causal graphs derived from thousands of coupled ordinary and stochastic differential equations as well as two idealized climate models. We perform a comprehensive evaluation of state-of-the-art causal discovery algorithms for graph reconstruction on systems with noisy, confounded, and lagged dynamics.CausalDynamicsconsists of a plug-and-play, build-your-own coupling workflow that enables the construction of a hierarchy of physical systems. We anticipate that our framework will facilitate the development of robust causal discovery algorithms that are broadly applicable across domains while addressing their unique challenges. We provide a user-friendly implementation and documentation on https://kausable.github.io/CausalDynamics.



Paperid:4455
Authors:Anastasia Vepreva, Julia Razlivina, Mariia Eremeyeva, Nina Gubina, Anastasia Orlova, Aleksei Dmitrenko, Kapranova Xenia, Susan Jyakhwo, Nikita Vasilev, Arsen Sarkisyan, Ivan Chernyshov, Vladimir Vinogradov, Andrei Dmitrenko
Abstract:
Despite recent advances in machine learning, many scientific discoveries in chemistry still rely on manually curated datasets extracted from the scientific literature. Automation of information extraction in specialized chemistry domains has the potential to scale up machine learning applications and improve the quality of predictions, enabling data-driven scientific discoveries at a faster pace. In this paper, we present ChemX, a collection of 10 benchmarking datasets across several domains of chemistry providing a reliable basis for evaluating and fine-tuning automated information extraction methods. The datasets encompassing various properties of small molecules and nanomaterials have been manually extracted from peer-reviewed publications and systematically validated by domain experts through a cross-verification procedure allowing for identification and correction of errors at sources. In order to demonstrate the utility of the resulting datasets, we evaluate the extraction performance of the state-of-the-art large language models (LLMs). Moreover, we design our own agentic approach to take full control of the document preprocessing before LLM-based information extraction. Finally, we apply the recently emerged multi-agent systems specialized in chemistry to compare performance against the strong baselines. Our empirical results highlight persistent challenges in chemical information extraction, particularly in handling domain-specific terminology, complex tabular and schematic formats, and context-dependent ambiguities. We discuss the importance of expert data validation, the nuances of the evaluation pipeline, and the prospects of automated information extraction in chemistry. Finally, we provide open documentation including standardized schemas and provenance metadata, as well as the code and other materials to ensure reproducibility. ChemX is poised to advance automatic information extraction in chemistry by challenging the quality and generalization capabilities of existing methods, as well as providing insights into evaluation strategies.



Paperid:4456
Authors:Haotian Ling, Zequn Chen, Qiuying Chen, Donglin Di, Yongjia Ma, Hao Li, Chen Wei, Zhulin Tao, Xun Yang
Abstract:
Consistent pose‐driven character animation has achieved remarkable progress in single‐character scenarios. However, extending these advances to multi‐character settings is non‐trivial, especially when position swap is involved. Beyond mere scaling, the core challenge lies in enforcing correct Identity Correspondence (IC) between characters in reference and generated frames. To address this, we introduce EverybodyDance, which, to our knowledge, is the first systematic solution targeting IC correctness in multi-character animation. EverybodyDance is built around the Identity Matching Graph (IMG), which models characters in the generated and reference frames as two node sets in a weighted complete bipartite graph. Edge weights, computed via our proposed Mask–Query Attention (MQA) mechanism, quantify the affinity between each pair of characters. Our key insight is to formalize IC correctness as a graph structural metric and to optimize it during training. To enhance robust and accurate IC, we propose a series of targeted strategies, including identity-embedded guidance, a multi-scale matching strategy, and pre-classified sampling, which work synergistically. Finally, to evaluate IC performance, we curate the Identity Correspondence Evaluation benchmark, dedicated to multi‐character IC correctness. Extensive experiments demonstrate that EverybodyDance substantially outperforms state‐of‐the‐art baselines in both IC and visual fidelity.



Paperid:4457
Authors:Yilan Chen, Zhichao Wang, Wei Huang, Andi Han, Taiji Suzuki, Arya Mazumdar
Abstract:
Gradient-based optimization methods have shown remarkable empirical success, yet their theoretical generalization properties remain only partially understood. In this paper, we establish a generalization bound for gradient flow that aligns with the classical Rademacher complexity bounds for kernel methods—specifically those based on the RKHS norm and kernel trace—through a data-dependent kernel called the loss path kernel (LPK). Unlike static kernels such as NTK, the LPK captures the entire training trajectory, adapting to both data and optimization dynamics, leading to tighter and more informative generalization guarantees. Moreover, the bound highlights how the norm of the training loss gradients along the optimization trajectory influences the final generalization performance. The key technical ingredients in our proof combine stability analysis of gradient flow with uniform convergence via Rademacher complexity. Our bound recovers existing kernel regression bounds for overparameterized neural networks and shows the feature learning capability of neural networks compared to kernel methods. Numerical experiments on real-world datasets validate that our bounds correlate well with the true generalization gap.



Paperid:4458
Authors:Rui Zhao, Yuze Fan, Ziguo Chen, Fei Gao, Zhenhai Gao
Abstract:
End-to-end learning has emerged as a transformative paradigm in autonomous driving research. However, the inherently multimodal nature of driving behaviors and the generalization challenges in long-tail scenarios remain critical obstacles to robust deployment. We propose DiffE2E, a diffusion-based end-to-end autonomous driving framework. This framework first performs multi-scale alignment of multi-sensor perception features through a hierarchical bidirectional cross-attention mechanism. It then introduces a novel class of hybrid diffusion-supervision decoders based on the Transformer architecture, and adopts a collaborative training paradigm that seamlessly integrates the strengths of both diffusion and supervised policies. DiffE2E models structured latent spaces, where diffusion captures the distribution of future trajectories and supervision enhances controllability and robustness. A global condition integration module enables deep fusion of perception features with high-level targets, significantly improving the quality of trajectory generation. Subsequently, a cross-attention mechanism facilitates efficient interaction between integrated features and hybrid latent variables, promoting the joint optimization of diffusion and supervision objectives for structured output generation, ultimately leading to more robust control. Experiments demonstrate that DiffE2E achieves state-of-the-art performance in both CARLA closed-loop evaluations and NAVSIM benchmarks. The proposed integrated diffusion-supervision policy offers a generalizable paradigm for hybrid action representation, with strong potential for extension to broader domains including embodied intelligence.



Paperid:4459
Authors:Hao Tang, Chen-Wei Xie, Haiyang Wang, Xiaoyi Bao, Tingyu Weng, Pandeng Li, Yun Zheng, Liwei Wang
Abstract:
Generalist models have achieved remarkable success in both language and vision-language tasks, showcasing the potential of unified modeling. However, effectively integrating fine-grained perception tasks like detection and segmentation into these models remains a significant challenge. This is primarily because these tasks often rely heavily on task-specific designs and architectures that can complicate the modeling process. To address this challenge, we present UFO, a framework that unifies fine-grained visual perception tasks through an open-ended language interface. By transforming all perception targets into the language space, UFO unifies object-level detection, pixel-level segmentation, and image-level vision-language tasks into a single model. Additionally, we introduce a novel embedding retrieval approach that relies solely on the language interface to support segmentation tasks. Our framework bridges the gap between fine-grained perception and vision-language tasks, significantly simplifying architectural design and training strategies while achieving comparable or superior performance to methods with intricate task-specific designs. After multi-task training on five standard visual perception datasets, UFO outperforms the previous state-of-the-art generalist models by 12.3 mAP on COCO instance segmentation and 3.3 mIoU on ADE20K semantic segmentation. Furthermore, our method seamlessly integrates with existing MLLMs, effectively combining fine-grained perception capabilities with their advanced language abilities, thereby achieving superior performance on the challenging reasoning segmentation. Code and models will be publicly available.



Paperid:4460
Authors:Alan Amin, Nate Gruver, Andrew Wilson
Abstract:
Discrete diffusion models, like continuous diffusion models, generate high-quality samples by gradually undoing noise applied to datapoints with a Markov process. Gradual generation in theory comes with many conceptual benefits; for example, inductive biases can be incorporated into the noising Markov process. In practice, however, the consistently best performing discrete diffusion model is masking diffusion, which does not denoise gradually. Here we explain the superior performance of masking diffusion by noting that it makes use of a fundamental difference between continuous and discrete Markov processes: discrete Markov processes evolve by discontinuous jumps at a fixed rate and, unlike other discrete diffusion models, masking diffusion builds in the known distribution of jump times and only learns where to jump to. We show that we can similarly bake in the known distribution of jump times into any discrete diffusion model. The resulting models -- schedule-conditioned diffusion (SCUD) -- generalize classical discrete diffusion and masking diffusion. By applying SCUD to models with noising processes that incorporate inductive biases on images, text, and protein data, we build diffusion models that outperform masking.



Paperid:4461
Authors:Zhi Zhou, Tan Yuhao, Zenan Li, Yuan Yao, Lan-Zhe Guo, Yu-Feng Li, Xiaoxing Ma
Abstract:
Test-time sampling, which aims to sample multiple reasoning paths for a given input during inference, is a widely adopted technique to improve the reasoning performance of large language models (LLMs). However, despite its practical success, the theoretical foundations remain underexplored. In this paper, we provide the first theoretical framework for analyzing test-time sampling strategies, grounded in the perspective of confidence estimation. Based on the framework, we analyze two dominant paradigms: self-consistency and perplexity, and reveal key limitations: self-consistency suffers from high estimation error while perplexity exhibits substantial modeling error and possible degradation of the estimation error convergence. To address these limitations, we introduce RPC, a hybrid method that leverages our theoretical insights through two key components:Perplexity ConsistencyandReasoning Pruning.Perplexity Consistencycombines the strengths of self-consistency and perplexity, boosting the convergence rate of estimation error from linear to exponential while preserving model error.Reasoning Pruningprevents degradation by eliminating low-probability reasoning paths. Both theoretical analysis and empirical results across seven benchmark datasets demonstrate that RPC has a strong potential for reducing reasoning error. Notably, RPC achieves reasoning performance comparable to self-consistency while not only enhancing confidence reliability but also reducing sampling costs by 50%.



Paperid:4462
Authors:Longshen Ou, Jingwei Zhao, Ziyu Wang, Gus Xia, Qihao Liang, Torin Hopkins, Ye Wang
Abstract:
We present a unified framework for automatic multitrack music arrangement that enables a single pre-trained symbolic music model to handle diverse arrangement scenarios, including reinterpretation, simplification, and additive generation. At its core is a segment-level reconstruction objective operating on token-level disentangled content and style, allowing for flexible any-to-any instrumentation transformations at inference time. To support track-wise modeling, we introduce REMI-z, a structured tokenization scheme for multitrack symbolic music that enhances modeling efficiency and effectiveness for both arrangement tasks and unconditional generation. Our method outperforms task-specific state-of-the-art models on representative tasks in different arrangement scenarios---band arrangement, piano reduction, and drum arrangement, in both objective metrics and perceptual evaluations. Taken together, our framework demonstrates strong generality and suggests broader applicability in symbolic music-to-music transformation.



Paperid:4463
Authors:Xinpeng Lv, Haotian Wang, Yunxin Mao, KE LIANG, Haoxuan Li, Wanrong Huang, Long Lan, Haoang Chi, Huan Chen, Jinxuan Yang, Cyuanlong, Wenjing Yang
Abstract:
Strategic classification (SC) explores how individuals or entities modify their features strategically to achieve favorable classification outcomes. However, existing SC methods, which are largely based on linear models or shallow neural networks, face significant limitations in terms of scalability and capacity when applied to real-world datasets with significantly increasing scale, especially in financial services and the internet sector. In this paper, we investigate how to leverage large language models to design a more scalable and efficient SC framework, especially in the case of growing individuals engaged with decision-making processes. Specifically, we introduce GLIM, a gradient-free SC method grounded in in-context learning. During the feed-forward process of self-attention, GLIM implicitly simulates the typical bi-level optimization process of SC, including both the feature manipulation and decision rule optimization. Without fine-tuning the LLMs, our proposed GLIM enjoys the advantage of cost-effective adaptation in dynamic strategic environments. Theoretically, we prove GLIM can support pre-trained LLMs to adapt to a broad range of strategic manipulations. We validate our approach through experiments with a bunch of pre-trained LLMs on real-world and synthetic datasets in financial and internet domains, demonstrating that our GLIM exhibits both robustness and efficiency, and offering an effective solution for large-scale SC tasks.



Paperid:4464
Authors:Shibin Mei, Bingbing Ni
Abstract:
We present PhysDiff-VTON, a diffusion-based framework for image-based virtual try-on that systematically addresses the dual challenges of garment deformation modeling and high-frequency detail preservation. The core innovation lies in integrating physics-inspired mechanisms into the diffusion process: a pose-guided deformable warping module simulates fabric dynamics by predicting spatial offsets conditioned on human pose semantics, while wavelet-enhanced feature decomposition explicitly preserves texture fidelity through frequency-aware attention. Further enhancing generation quality, a novel sampling strategy optimizes the denoising trajectory via least action principles, enforcing temporal coherence, spatial smoothness, and multi-scale structural consistency. Comprehensive evaluations across multiple datasets demonstrate significant improvements in both geometric plausibility and perceptual quality compared to existing approaches. The framework establishes a new paradigm for synthesizing photorealistic try-on images that adhere to physical constraints while maintaining intricate garment details, advancing the practical applicability of diffusion models in fashion technology.



Paperid:4465
Authors:Ann Huang, Satpreet Harcharan Singh, Flavio Martinelli, Kanaka Rajan
Abstract:
Task-trained recurrent neural networks (RNNs) are widely used in neuroscience and machine learning to model dynamical computations. To gain mechanistic insight into how neural systems solve tasks, prior work often reverse-engineers individual trained networks. However, different RNNs trained on the same task and achieving similar performance can exhibit strikingly different internal solutions—a phenomenon known as solution degeneracy. Here, we develop a unified framework to systematically quantify and control solution degeneracy across three levels: behavior, neural dynamics, and weight space. We apply this framework to 3,400 RNNs trained on four neuroscience-relevant tasks—flip-flop memory, sine wave generation, delayed discrimination, and path integration—while systematically varying task complexity, learning regime, network size, and regularization. We find that increased task complexity and stronger feature learning reduce degeneracy in neural dynamics but increase it in weight space, with mixed effects on behavior. In contrast, larger networks and structural regularization reduce degeneracy at all three levels. These findings empirically validate the Contravariance Principle and provide practical guidance for researchers aiming to tailor RNN solutions—whether to uncover shared neural mechanisms or to model individual variability observed in biological systems. This work provides a principled framework for quantifying and controlling solution degeneracy in task-trained RNNs, offering new tools for building more interpretable and biologically grounded models of neural computation.



Paperid:4466
Authors:Yuanyi Wang, Zhaoyi Yan, Yiming Zhang, Qi Zhou, Yanggan Gu, Fei Wu, Hongxia Yang
Abstract:
Abstract:Recent advances in large language models (LLMs) have intensified efforts to fuse heterogeneous open-source models into a unified system that inherits their complementary strengths. Existing logit-based fusion methods maintain inference efficiency but treat vocabulary dimensions independently, overlooking semantic dependencies encoded by cross-dimension interactions. These dependencies reflect how token types interact under a model's internal reasoning and are essential for aligning models with diverse generation behaviors. To explicitly model these dependencies, we propose \textbf{InfiGFusion}, the first structure-aware fusion framework with a novel \textit{Graph-on-Logits Distillation} (GLD) loss. Specifically, we retain the top-$k$ logits per output and aggregate their outer products across sequence positions to form a global co-activation graph, where nodes represent vocabulary channels and edges quantify their joint activations. To ensure scalability and efficiency, we design a sorting-based closed-form approximation that reduces the original $O(n^4)$ cost of Gromov-Wasserstein distance to $O(n \log n)$, with provable approximation guarantees. Experiments across multiple fusion settings show that GLD consistently improves fusion quality and stability. InfiGFusion outperforms SOTA models and fusion baselines across 11 benchmarks spanning reasoning, coding, and mathematics. It shows particular strength in complex reasoning tasks, with +35.6 improvement on Multistep Arithmetic and +37.06 on Causal Judgement over SFT, demonstrating superior multi-step and relational inference.



Paperid:4467
Authors:Ryien Hosseini, Filippo Simini, Venkatram Vishwanath, Rebecca Willett, Henry Hoffmann
Abstract:
Graph Neural Networks learn on graph-structured data by iteratively aggregating local neighborhood information. While this local message passing paradigm imparts a powerful inductive bias and exploits graph sparsity, it also yields three key challenges: (i) oversquashing of long-range information, (ii) oversmoothing of node representations, and (iii) limited expressive power. In this work we inject randomized global embeddings of node features, called Sketched Random Features, into standard GNNs, enabling them to efficiently capture long-range dependencies. The embeddings are unique, distance-sensitive, and topology-agnostic—properties which we analytically and empirically show alleviate the aforementioned limitations when injected into GNNs. Experimental results on real-world graph learning tasks confirm that this strategy consistently improves performance metrics over baseline GNNs, offering both a standalone solution and a complementary enhancement to existing techniques such as graph positional encodings.



Paperid:4468
Authors:Yunke Ao, Masoud Moghani, Mayank Mittal, Manish Prajapat, Luohong Wu, Frederic Giraud, Fabio Carrillo, Andreas Krause, Philipp Fürnstahl
Abstract:
Ultrasound (US) is a widely used medical imaging modality due to its real-time capabilities, non-invasive nature, and cost-effectiveness. By reducing operator dependency and enhancing access to complex anatomical regions, robotic ultrasound can help improve workflow efficiency. Recent studies have demonstrated the potential of deep reinforcement learning (DRL) and imitation learning (IL) to enable more autonomous and intelligent robotic ultrasound navigation. However, the application of learning-based robotic ultrasound to computer-assisted surgical tasks, such as anatomy reconstruction and surgical guidance, remains largely unexplored. A key bottleneck for this is the lack of realistic and efficient simulation environments tailored to these tasks. In this work, we present SonoGym, a scalable simulation platform for robotic ultrasound, enabling parallel simulation across tens to hundreds of environments. Our framework supports realistic and real-time simulation of US data from CT-derived 3D models of the anatomy through both a physics-based and a Generative Adversarial Network (GAN) approach. Our framework enables the training of DRL and recent IL agents (vision transformers and diffusion policies) for relevant tasks in robotic orthopedic surgery by integrating common robotic platforms and orthopedic end effectors. We further incorporate submodular DRL---a recent method that handles history-dependent rewards---for anatomy reconstruction and safe reinforcement learning for surgery. Our results demonstrate successful policy learning across a range of scenarios, while also highlighting the limitations of current methods in clinically relevant environments. We believe our simulation can facilitate research in robot learning approaches for such challenging robotic surgery applications. Dataset, codes and videos are publicly available at https://sonogym.github.io/.



Paperid:4469
Authors:Xinghong Chen, Weilin Wu, Kunping Yang, Guannan Chen
Abstract:
Forecasting real-world time series requires modeling both short-term fluctuations and long-term evolutions, as these signals typically exhibit multiscale temporal structures. A core challenge lies in reconciling such dynamics: high-frequency seasonality demands local precision, while low-frequency trends require global robustness. However, most existing methods adopt a unified loss function across all temporal components, overlooking their structural differences. This misalignment often causes overfitting to seasonal noise or underfitting of long-term trends, leading to suboptimal forecasting performance. To address this issue, we propose a Physics-guided Multiscale Loss Framework (PMLF) that decomposes time series into seasonal and trend components and assigns component-specific objectives grounded in the distinct energy responses of oscillatory and drift dynamics. Specifically, we assign a quadratic loss to seasonal components, reflecting the quadratic potential energy profile of molecular vibration, while a logarithmic loss is used for trend components to capture the sublinear energy profile of molecular drift under sustained external forces. Furthermore, we introduce a softmax-based strategy that adaptively balances the unequal energetic responses of these two physical processes. Experiments on different public benchmarks show that PMLF improves the performance of diverse baselines, demonstrating the effectiveness of physics-guided loss design in modeling structural heterogeneity in time series forecasting.



Paperid:4470
Authors:Wen-Tse Chen, Jiayu Chen, Fahim Tajwar, Hao Zhu, Xintong Duan, Ruslan Salakhutdinov, Jeff Schneider
Abstract:
Learning from self-sampled data and sparse environmental feedback remains a fundamental challenge in training self-evolving agents. Temporal credit assignment mitigates this issue by transforming sparse feedback into dense supervision signals. However, previous approaches typically depend on domain-specific value functions for credit assignment, which suffer from poor sample efficiency and limited generalization. In this work, we propose to leverage pre-trained knowledge from large language models (LLMs) to transform sparse rewards into dense training signals (i.e., the advantage function) through retrospective in-context learning (RICL). We further propose an online learning framework, RICOL, which iteratively refines the policy based on the credit assignment results from RICL. We empirically demonstrate that RICL can accurately estimate the advantage function with limited samples and effectively identify critical states for temporal credit assignment. Extended evaluation on the BabyAI benchmark shows that RICOL significantly improves sample efficiency compared to traditional online RL algorithms while achieving performance comparable to imitation learning from expert demonstartions. Our findings highlight the potential of leveraging LLMs for temporal credit assignment, paving the way for more sample-efficient and generalizable RL paradigms.



Paperid:4471
Authors:Haodong Chen, Haojian Huang, Qifeng Chen, Harry Yang, Ser Nam Lim
Abstract:
Recent advancements in video generation have enabled the creation of high-quality, visually compelling videos. However, generating videos that adhere to the laws of physics remains a critical challenge for applications requiring realism and accuracy. In this work, we proposePhysHPO, a novel framework for Hierarchical Cross-Modal Direct Preference Optimization, to tackle this challenge by enabling fine-grained preference alignment for physically plausible video generation. PhysHPO optimizes video alignment across four hierarchical granularities: a)Instance Level, aligning the overall video content with the input prompt; b)State Level, ensuring temporal consistency using boundary frames as anchors; c)Motion Level, modeling motion trajectories for realistic dynamics; and d)Semantic Level, maintaining logical consistency between narrative and visuals. Recognizing that real-world videos are the best reflections of physical phenomena, we further introduce an automated data selection pipeline to efficiently identify and utilize"good data"from existing large-scale text-video datasets, thereby eliminating the need for costly and time-intensive dataset construction. Extensive experiments on both physics-focused and general capability benchmarks demonstrate that PhysHPO significantly improves physical plausibility and overall video generation quality of advanced models. To the best of our knowledge, this is the first work to explore fine-grained preference alignment and data selection for video generation, paving the way for more realistic and human-preferred video generation paradigms.



Paperid:4472
Authors:Zihuan Qiu, Yi Xu, Chiyuan He, Fanman Meng, Linfeng Xu, Qingbo Wu, Hongliang Li
Abstract:
Continual model merging integrates independently fine-tuned models sequentially without access to original training data, providing a scalable and efficient solution to continual learning. However, current methods still face critical challenges, notably parameter interference among tasks and limited adaptability to evolving test distributions. The former causes catastrophic forgetting of integrated tasks, while the latter hinders effective adaptation to new tasks. To address these, we propose MINGLE, a novel framework for test-time continual model merging, which leverages test-time adaptation using a small set of unlabeled test samples from the current task to dynamically guide the merging process.MINGLE employs a mixture-of-experts architecture composed of parameter-efficient, low-rank experts, enabling efficient adaptation and improving robustness to distribution shifts.To mitigate catastrophic forgetting, we propose Null-Space Constrained Gating, which restricts gating updates to subspaces orthogonal to prior task representations. This suppresses activations on old task inputs and preserves model behavior on past tasks.To further balance stability and adaptability, we design an Adaptive Relaxation Strategy, which dynamically adjusts the constraint strength based on interference signals captured during test-time adaptation.Extensive experiments on standard continual merging benchmarks demonstrate that MINGLE achieves robust generalization, reduces forgetting significantly, and consistently surpasses previous state-of-the-art methods by 7-9\% on average across diverse task orders.



Paperid:4473
Authors:Alexander Bodard, Panagiotis Patrinos
Abstract:
Abstract:We study generalized smoothness in nonconvex optimization, focusing on $(L_0, L_1)$-smoothness and anisotropic smoothness. The former was empirically derived from practical neural network training examples, while the latter arises naturally in the analysis of nonlinearly preconditioned gradient methods. We introduce a new sufficient condition that encompasses both notions, reveals their close connection, and holds in key applications such as phase retrieval and matrix factorization. Leveraging tools from dynamical systems theory, we then show that nonlinear preconditioning—including gradient clipping—preserves the saddle point avoidance property of classical gradient descent. Crucially, the assumptions required for this analysis are actually satisfied in these applications, unlike in classical results that rely on restrictive Lipschitz smoothness conditions. We further analyze a perturbed variant that efficiently attains second-order stationarity with only logarithmic dependence on dimension, matching similar guarantees of classical gradient methods.



Paperid:4474
Authors:Zeman Li, Yuan Deng, Peilin Zhong, Meisam Razaviyayn, Vahab Mirrokni
Abstract:
Abstract:Modern foundation models are trained on diverse datasets to enhance generalization across tasks and domains. A central challenge in this process is determining how to effectively mix and sample data from multiple sources. This naturally leads to a multi-task learning (MTL) perspective. While prior work in MTL has emphasized mitigating gradient conflicts, we observe that large-scale pretraining scenarios—such as multilingual or multi-domain training—often exhibit little to no gradient conflict. Motivated by this observation, we propose $\textbf{PiKE}$ ($\textbf{P}$ositive gradient $\textbf{i}$nteraction-based $\textbf{K}$-task weights $\textbf{E}$stimator), an adaptive data mixing algorithm that dynamically adjusts sampling weights during training. PiKE exploits non-conflicting gradient interactions to minimize a near-tight upper bound on the average loss decrease at each step, while incurring negligible computational overhead. We provide theoretical convergence guarantees and show that PiKE outperforms static and non-adaptive mixing baselines. Furthermore, we extend PiKE to promote balanced learning across tasks. Extensive experiments on large-scale language model pretraining confirm that PiKE achieves faster convergence and improved downstream performance compared to existing approaches.



Paperid:4475
Authors:Yi Zhang, Elynn Chen, Yujun Yan
Abstract:
Abstract:We study contextual dynamic pricing when a target market can leverage $K$ auxiliary markets—offline logs or concurrent streams—whose *mean utilities differ by a structured preference shift*. We propose *Cross-Market Transfer Dynamic Pricing (CM-TDP)*, the first algorithm that *provably* handles such model-shift transfer and delivers minimax-optimal regret for *both* linear and non-parametric utility models.For linear utilities of dimension $d$, where the *difference* between source- and target-task coefficients is $s_{0}$-sparse, CM-TDP attains regret $\tilde{\mathcal{O}}\bigl((dK^{-1}+s_{0})\log T\bigr)$. For nonlinear demand residing in a reproducing kernel Hilbert space with effective dimension $\alpha$, complexity $\beta$ and task-similarity parameter $H$, the regret becomes $\tilde{\mathcal{O}}\bigl(K^{-2\alpha\beta/(2\alpha\beta+1)}T^{1/(2\alpha\beta+1)} + H^{2/(2\alpha+1)}T^{1/(2\alpha+1)}\bigr)$, matching information-theoretic lower bounds up to logarithmic factors. The RKHS bound is the first of its kind for transfer pricing and is of independent interest.Extensive simulations show up to 38\% higher cumulative revenue and $6\times$ faster convergence relative to single-market pricing baselines. By bridging transfer learning, robust aggregation, and revenue optimization, CM-TDP moves toward pricing systems that *transfer faster, price smarter*.



Paperid:4476
Authors:Hanyang Wang, Juergen Branke, Matthias Poloczek
Abstract:
Many real-world black-box optimization problems have multiple conflicting objectives. Rather than attempting to approximate the entire set of Pareto-optimal solutions, interactive preference learning, i.e., optimization with a decision maker in the loop, allows to focus the search on the most relevant subset. However, few previous studies have exploited the fact that utility functions are usually monotonic. In this paper, we address the Bayesian Optimization with Preference Exploration (BOPE) problem and propose using a neural network ensemble as a utility surrogate model. This approach naturally integrates monotonicity and allows to learn the decision maker's preferences from pairwise comparisons. Our experiments demonstrate that the proposed method outperforms state-of-the-art approaches and exhibits robustness to noise in utility evaluations. An ablation study highlights the critical role of monotonicity in enhancing performance.



Paperid:4477
Authors:Qingyang Zhang, Haitao Wu, Changqing Zhang, Peilin Zhao, Yatao Bian
Abstract:
Existing methods to enhance the reasoning capability of large language models predominantly rely on supervised fine-tuning (SFT) followed by reinforcement learning (RL) on reasoning-specific data. These approaches critically depend on external supervisions--such as labeled reasoning traces, verified golden answers, or pre-trained reward models.In this work, we propose Entropy Minimized Policy Optimization (EMPO), which makes an early attempt at fully unsupervised LLM reasoning incentivization. By continuously minimizing the predictive entropy of LLMs on unlabeled questions in a latent semantic space, EMPO achieves competitive performance compared to supervised counterparts on both mathematical and free-form natural reasoning tasks. Specifically, without any supervised signals, EMPO boosts the accuracy of Qwen2.5-Math-7B Base from 30.7\% to 48.1\% on mathematical benchmarks and improves the accuracy of Qwen2.5-7B Base from 32.1\% to 50.1\% on MMLU-Pro. Primary experiments and analysis are also provided to interpret the effectiveness of EMPO. Code is available at anonymous github.



Paperid:4478
Authors:Nguyen Phuc, Ngoc-Hieu Nguyen, Duy M. H. Nguyen, Anji Liu, An Mai, Binh Nguyen, Daniel Sonntag, Khoa D Doan
Abstract:
Recently, Direct Alignment Algorithms (DAAs) such as Direct Preference Optimization (DPO) have emerged as alternatives to the standard Reinforcement Learning from Human Feedback (RLHF) for aligning large language models (LLMs) with human values. Surprisingly, while DAAs do not use a separate proxy reward model as in RLHF, their performance can still deteriorate over the course of training -- an over-optimization phenomenon found in RLHF where the learning policy exploits the overfitting to inaccuracies of the reward model to achieve high rewards.One attributed source of over-optimization in DAAs is the under-constrained nature of their offline optimization, which can gradually shift probability mass toward non-preferred responses not presented in the preference dataset. This paper proposes a novel importance-sampling approach to mitigate the distribution shift problem of offline DAAs.This approach, called (IS-DAAs), multiplies the DAA objective with an importance ratio that accounts for the reference policy distribution. IS-DAAs additionally avoid the high variance issue associated with importance sampling by clipping the importance ratio to a maximum value. Our extensive experiments demonstrate that IS-DAAs can effectively mitigate over-optimization, especially under low regularization strength, and achieve better performance than other methods designed to address this problem.



Paperid:4479
Authors:Thanh-Tung Le, Tuan Pham, Tung Nguyen, Deying Kong, Xiaohui Xie, Stephan Mandt
Abstract:
Novel view synthesis (NVS) seeks to render photorealistic, 3D‑consistent images of a scene from unseen camera poses given only a sparse set of posed views. Existing deterministic networks render observed regions quickly but blur unobserved areas, whereas stochastic diffusion‑based methods hallucinate plausible content yet incur heavy training‑ and inference‑time costs. In this paper, we propose a hybrid framework that unifies the strengths of both paradigms. A bidirectional transformer encodes multi‑view image tokens and Plücker‑ray embeddings, producing a shared latent representation. Two lightweight heads then act on this representation: (i) a feed‑forward regression head that renders pixels where geometry is well constrained, and (ii) a masked autoregressive diffusion head that completes occluded or unseen regions. The entire model is trained end‑to‑end with joint photometric and diffusion losses, without handcrafted 3D inductive biases, enabling scalability across diverse scenes. Experiments demonstrate that our method attains state‑of‑the‑art image quality while reducing rendering time by an order of magnitude compared with fully generative baselines.



Paperid:4480
Authors:Xiaoqi Zhao, Youwei Pang, Chenyang Yu, Lihe Zhang, Huchuan Lu, Shijian Lu, Georges Fakhri, Xiaofeng Liu
Abstract:
Multi-modal image segmentation faces real-world deployment challenges from incomplete/corrupted modalities degrading performance. While existing methods address training-inference modality gaps via specialized per-combination models, they introduce high deployment costs by requiring exhaustive model subsets and model-modality matching. In this work, we propose a unified modality-relax segmentation network (UniMRSeg) through hierarchical self-supervised compensation (HSSC). Our approach hierarchically bridges representation gaps between complete and incomplete modalities across input, feature and output levels. First, we adopt modality reconstruction with the hybrid shuffled-masking augmentation, encouraging the model to learn the intrinsic modality characteristics and generate meaningful representations for missing modalities through cross-modal fusion.Next, modality-invariant contrastive learning implicitly compensates the feature space distance among incomplete-complete modality pairs. Furthermore, the proposed lightweight reverse attention adapter explicitly compensates for the weak perceptual semantics in the frozen encoder. Last, UniMRSeg is fine-tuned under the hybrid consistency constraint to ensure stable prediction under all modality combinations without large performance fluctuations. Without bells and whistles, UniMRSeg significantly outperforms the state-of-the-art methods under diverse missing modality scenarios on MRI-based brain tumor segmentation, RGB-D semantic segmentation, RGB-D/T salient object segmentation. The code will be released to facilitate further research.



Paperid:4481
Authors:Mingze Dong, Leda Wang, Yuval Kluger
Abstract:
Abstract:Mask-based pretraining has become a cornerstone of modern large-scale models across natural language processing, computer vision, and recently biology. Despite its empirical success, its role and limits in learning data representations have been unclear. In this work, we show that the behavior of mask-based pretraining can be directly characterized by test risk in high-dimensional minimum-norm ("ridge-less") linear regression, without relying on further model specifications. Further analysis of linear models uncovers several novel aspects of mask-based pretraining. The theoretical framework and its implications have been validated across diverse neural architectures (including MLPs, CNNs, and Transformers) applied to both vision and language tasks. Guided by our theory, we propose an embarrassingly simple yet overlooked pretraining scheme named *Randomly Random Mask AutoEncoding* (**R$^2$MAE**), designed for learning universal representations capturing features of all scales from data. We implement **R$^2$MAE** in DNA sequence and single-cell self-supervised models, where it improves upon state-of-the-art models in terms of various performance metrics for MLP and transformer-based architectures, outperforming standard and more complicated mask pretraining schemes.



Paperid:4482
Authors:Shogo Iwazaki
Abstract:
Abstract:This paper addresses the Bayesian optimization problem (also referred to as the Bayesian setting of the Gaussian process bandit), where the learner seeks to minimize the regret under a function drawn from a known Gaussian process (GP). Under a Mat\'ern kernel with some extent of smoothness, we show that the Gaussian process upper confidence bound (GP-UCB) algorithm achieves $\tilde{O}(\sqrt{T})$ cumulative regret with high probability. Furthermore, our analysis yields $O(\sqrt{T \ln^4 T})$ regret under a squared exponential kernel. These results fill the gap between the existing regret upper bound of GP-UCB and the current best upper bound provided by Scarlett [2018]. The key idea in our proof is to capture the concentration behavior of the input sequence realized by GP-UCB, enabling us to handle GP's information gain in a refined manner.



Paperid:4483
Authors:Hanze Guo, Jing Yao, Xiao Zhou, Xiaoyuan Yi, Xing Xie
Abstract:
As large language models (LLMs) become increasingly integrated into applications serving users across diverse cultures, communities and demographics, it is critical to align LLMs with pluralistic human values beyond average norms (e.g. HHH). In psychological and social value theories such as Schwartz’s Value Theory, each value profile is represented by multi-dimensional values paired with priorities. However, existing methods encounter two challenges when aligning with such fine-grained value objectives: they often treat multiple values as independent and equally important, ignoring their interdependence and relative priorities (value complexity); they struggle to precisely control nuanced value priorities, especially those underrepresented ones (value steerability). To handle these challenges, we propose COUPLE, a COUnterfactual reasoning framework for PLuralistic valuE alignment. It introduces a Structural Causal Model (SCM) to feature the causal relationship between high-level value dimensions and behaviors. Moreover, it applies counterfactual reasoning to generate a new output aligned with any desired value objective. Benefitting from explicit causal modeling, COUPLE also provides better interpretability. We evaluate COUPLE on two datasets with different value systems and demonstrate that COUPLE advances other baselines across diverse types of value objectives. Our code is available at the anonymous repository: https://anonymous.4open.science/r/COUPLE-7D83.



Paperid:4484
Authors:Qingzhu Zhang, Jiani Zhong, Zongsheng Li, Xinke Shen, Quanying Liu
Abstract:
Task-specific pre-training is essential for tasks with limited and heterogeneous data. Existing task-general pre-training EEG models struggle with complex tasks like emotion recognition due to mismatches between task-specific features and broad pre-training approaches. This work aims to develop a task-specific multi-dataset joint pre-training framework for cross-dataset emotion recognition, tackling problems of large inter-dataset distribution shifts, inconsistent emotion category definitions, and substantial inter-subject variability. We introduce a Cross-Dataset covariance Alignment (CDA) loss to align second-order statistical properties across datasets, enabling robust generalization without the need for extensive labels or per-subject calibration. Additionally, we propose a hybrid encoder to capture the long-term dependency and complex dynamics of EEG, combining a Mamba-like linear attention channel encoder and a spatiotemporal dynamics model. Our method outperforms state-of-the-art large-scale EEG models by an average of 6.59% on F1 score in cross-dataset emotion recognition. Performance scales with the increase of datasets used in pre-training. Multi-dataset joint pre-training achieves a performance gain of 15.14% over single-dataset training. This work provides a scalable framework for task-specific pre-training and highlights its benefit in generalizable affective computing. Our code is available at https://anonymous.4open.science/status/CovarianceEEGEmotion-D1C1.



Paperid:4485
Authors:tao zhang, Cheng Da, Kun Ding, Huan Yang, kun jin, Yan Li, Tingting Gao, Di ZHANG, SHIMING XIANG, Chunhong Pan
Abstract:
Preference optimization for diffusion models aims to align them with human preferences for images. Previous methods typically use Vision-Language Models (VLMs) as pixel-level reward models to approximate human preferences. However, when used for step-level preference optimization, these models face challenges in handling noisy images of different timesteps and require complex transformations into pixel space. In this work, we show that pre-trained diffusion models are naturally suited for step-level reward modeling in the noisy latent space, as they are explicitly designed to process latent images at various noise levels. Accordingly, we propose theLatent Reward Model (LRM), which repurposes components of the diffusion model to predict preferences of latent images at arbitrary timesteps. Building on LRM, we introduceLatent Preference Optimization (LPO), a step-level preference optimization method conducted directly in the noisy latent space. Experimental results indicate that LPO significantly improves the model's alignment with general, aesthetic, and text-image alignment preferences, while achieving a 2.5-28x training speedup over existing preference optimization methods.



Paperid:4486
Authors:Menglin Yang, Ram Samarth B B, Aosong Feng, Bo Xiong, Jiahong Liu, Irwin King, Rex Ying
Abstract:
Large language models (LLMs) have demonstrated remarkable performance on various tasks. However, it remains an open question whether the default Euclidean space is the most suitable choice for embedding tokens in LLMs. In this study, we first investigate the non-Euclidean characteristics of LLMs. Our findings reveal that token frequency follows a power-law distribution, with high-frequency tokens clustering near the origin and low-frequency tokens positioned farther away. Additionally, token embeddings exhibit a high degree of hyperbolicity, indicating a latent tree-like structure in the embedding space. Building on the observation, we propose to efficiently fine-tune LLMs in hyperbolic space to better exploit the underlying complex structures. However, we found that this fine-tuning in hyperbolic space cannot be achieved with naive application of exponential and logarithmic maps when the embedding and weight matrices both reside in Euclidean space. To address this technical issue, we introduce a new method called hyperbolic low-rank efficient fine-tuning, HypLoRA, that performs low-rank adaptation directly on the hyperbolic manifold, avoiding the cancellation effect caused by the exponential and logarithmic maps, thus preserving the hyperbolic modeling capabilities. Extensive experiments on diverse benchmarks, encompassing both arithmetic and commonsense reasoning, demonstrate that HypLoRA significantly enhances LLM performance.



Paperid:4487
Authors:Lin Yuan, Xiaowan Li, Yan Zhang, Jiawei Zhang, Hongbo Li, Xinbo Gao
Abstract:
Advancements in image generation technologies have raised significant concerns about their potential misuse, such as producing misinformation and deepfakes. Therefore, there is an urgent need for effective methods to detect AI-generated images (AIGI). Despite progress in AIGI detection, achieving reliable performance across diverse generation models and scenes remains challenging due to the lack of source-invariant features and limited generalization capabilities in existing methods. In this work, we explore the potential of using image entropy as a cue for AIGI detection and propose Multi-granularity Local Entropy Patterns (MLEP), a set of entropy feature maps computed across shuffled small patches over multiple image scaled. MLEP comprehensively captures pixel relationships across dimensions and scales while significantly disrupting image semantics, reducing potential content bias. Leveraging MLEP, a robust CNN-based classifier for AIGI detection can be trained. Extensive experiments conducted in an open-world scenario, evaluating images synthesized by 32 distinct generative models, demonstrate significant improvements over state-of-the-art methods in both accuracy and generalization.



Paperid:4488
Authors:Xincheng Yao, Yan Luo, Zefeng Qian, Chongyang Zhang
Abstract:
The current mainstream and state-of-the-art anomaly detection (AD) methods are substantially established on pretrained feature networks yielded by ImageNet pretraining. However, regardless of supervised or self-supervised pretraining, the pretraining process on ImageNet does not match the goal of anomaly detection (i.e., pretraining in natural images doesn't aim to distinguish between normal and abnormal). Moreover, natural images and industrial image data in AD scenarios typically have the distribution shift. The two issues can cause ImageNet-pretrained features to be suboptimal for AD tasks. To further promote the development of the AD field, pretrained representations specially for AD tasks are eager and very valuable. To this end, we propose a novel AD pretraining framework specially designed for learning robust and discriminative pretrained representations for unsupervised anomaly detection. Specifically, closely surrounding the goal of anomaly detection (i.e., focus on discrepancies between normals and anomalies), we propose angle- and norm-oriented contrastive losses to maximize the angle size and norm difference between normal and abnormal features simultaneously. To avoid the distribution shift from natural images to AD images, our pretraining is performed on a recent large-scale AD dataset, RealIAD. To further alleviate the potential shift between pretraining data and downstream AD datasets, we learn the pretrained AD representations based on the class-generalizable representations, residual features. For evaluation, based on five embedding-based AD methods, we simply replace their original features with our pretrained representations. Extensive experiments on five AD datasets and four backbones consistently show the superiority of our pretrained features. Code and pretrained weights will be available online.



Paperid:4489
Authors:Ruilin Tong, Haodong Lu, Yuhang Liu, Dong Gong
Abstract:
Continual learning (CL) aims to adapt a model to a sequence of tasks while maintaining performance on previously seen ones. Despite their effectiveness in mitigating forgetting, data storage and replay are often infeasible due to privacy or security constraints, and are impractical or unavailable for arbitrary pre-trained models. Data-free CL aims to continually update models with new tasks without storing previous data. Model inversion enables data generation by extracting knowledge from a trained model, allowing replay of previous tasks without stored data. However, model inversion faces two key challenges. Inversely generating data, e.g., images, solely from highly-compressed class labels leads to feature drift between synthetic and real data, causing forgetting of real image knowledge after replay and degrading inversion quality in subsequent stages. And performing inversion is usually computationally expensive, as each iteration requires backpropagation through the full model and many steps are needed for convergence. These problems are more severe in pretrained foundation models such as Contrastive Language-Image Pre-training (CLIP) models. To improve model inversion efficiency, we propose Per-layer Model Inversion (PMI) approach inspired by the faster convergence of single-layer optimization. The inputs optimized from PMI provide strong initialization for full-model inversion, significantly reducing the number of iterations required for convergence. To address feature distribution shift, we model class-wise feature distribution using a Gaussian distribution and preserve distributional information with a contrastive model. Sampling features for inversion ensures alignment between synthetic and real feature distributions. Combining PMI and feature modeling, we demonstrate the feasibility of adapting models to new classes by generating data from pseudo image features mapped through semantic-aware feature projection. Our method shows strong effectiveness and compatibility across multiple CL settings.



Paperid:4490
Authors:Francesco Emanuele Stradi, Matteo Castiglioni, Alberto Marchesi, Nicola Gatti, Christian Kroer
Abstract:
Abstract:We study online decision making problems under resource constraints, where both reward and cost functions are drawn from distributions that may change adversarially over time. We focus on two canonical settings: $(i)$ online resource allocation where rewards and costs are observed before action selection, and $(ii)$ online learning with resource constraints where they are observed after action selection, under full feedback or bandit feedback. It is well known that achieving sublinear regret in these settings is impossible when the rewards and cost distributions may change arbitrarily over time. To address this challenge, we analyze a framework in which the learner is guided by a spending plan—a sequence prescribing expected resource usage across rounds. We design general (primal-)dual methods that achieve sublinear regret with respect to baselines that follow the spending plan. Crucially, the performance of our algorithms improves when the spending plan ensures a well-balanced distribution of the budget across rounds. We additionally provide a robust variant of our methods to handle worst-case scenarios where the spending plan is highly imbalanced. To conclude, we study the regret of our algorithms when competing against benchmarks that deviate from the prescribed spending plan.



Paperid:4491
Authors:Lishen Qu, Zhihao Liu, Jinshan Pan, Shihao Zhou, Jinglei Shi, Duosheng Chen, Jufeng Yang
Abstract:
Lens flare occurs when shooting towards strong light sources, significantly degrading the visual quality of images. Due to the difficulty in capturing flare-corrupted and flare-free image pairs in the real world, existing datasets are typically synthesized in 2D by overlaying artificial flare templates onto background images. However, the lack of flare diversity in templates and the neglect of physical principles in the synthesis process hinder models trained on these datasets from generalizing well to real-world scenarios. To address these challenges, we propose a new physics-informed method for flare data generation, which consists of three stages: parameterized template creation, the laws of illumination-aware 2D synthesis, and physical engine-based 3D rendering, which finally gives us a mixed flare dataset that incorporates both 2D and 3D perspectives, namely FlareX. This dataset offers 9,500 2D templates derived from 95 flare patterns and 3,000 flare image pairs rendered from 60 3D scenes. Furthermore, we design a masking approach to obtain real-world flare-free images from their corrupted counterparts to measure the performance of the model on real-world images. Extensive experiments demonstrate the effectiveness of our method and dataset. The code, dataset, and a 1-minute video demo are available in the supplementary materials.



Paperid:4492
Authors:Xiangkun Wang, Kejiang Chen, Lincong Li, Weiming Zhang, Nenghai Yu
Abstract:
Existing video steganography methods primarily embed secret information by modifying video content in the spatial or compressed domains. However, such methods are prone to distortion drift and are easily detected by steganalysis. Generative steganography, which avoids direct modification of the cover data, offers a promising alternative. Despite recent advances, most generative steganography studies focus on images and are difficult to extend to videos because of compression-induced distortions and the unique architecture of video generation models. To address these challenges, we propose LD-RoViS, a training-free and robust video steganography framework for the deterministic latent diffusion model. By modulating implicit conditional parameters during the diffusion process, LD-RoViS constructs a dedicated steganographic channel. Additionally, we introduce a novel multi-mask mechanism to mitigate errors caused by video compression and post-processing. The experimental results demonstrate that LD-RoViS can embed approximately 12,000 bits of data into a 5-second video with an extraction accuracy exceeding 99\%. Our implementation is available at https://anonymous.4open.science/r/LD-RoViS-7FB1.



Paperid:4493
Authors:zhicheng li, Shuoming Zhang, Jiacheng Zhao, Siqi Li, Xiyu Shi, Yangyu Zhang, Shuaijiang Li, Donglin Yu, Zheming Yang, YUAN WEN, Huimin Cui
Abstract:
Recent multimodal large language models (MLLMs) marry modality-specificvision or audio encoders with a shared text decoder. While the encoder is compute-intensive but memory-light, the decoder is the opposite, yet state-of-the-art servingstacks still time-multiplex these complementary kernels, idling SMs or HBM inturn. We introduce SpaceServe, a serving system that space-multiplexes MLLMs:it decouples all modality encoders from the decoder, and co-locates them on thesame GPU using fine-grained SM partitioning available in modern runtimes. Acost-model-guided Space-Inference Scheduler (SIS) dynamically assigns SM slices,while a Time-Windowed Shortest-Remaining-First (TWSRFT) policy batches en-coder requests to minimise completion latency and smooth decoder arrivals. Eval-uation shows that SpaceServe outperforms state-of-the-art by up to 29.71x onNvidia A100 GPUs



Paperid:4494
Authors:Zhihao Sun, Haoran Jiang, Haoran Chen, Yixin Cao, Xipeng Qiu, Zuxuan Wu, Yu-Gang Jiang
Abstract:
Multimodal large language models have unlocked new possibilities for various multimodal tasks. However, their potential in image manipulation detection remains unexplored. When directly applied to the IMD task, M-LLMs often produce reasoning texts that suffer from hallucinations and overthinking. To address this, in this work, we propose ForgerySleuth, which leverages M-LLMs to perform comprehensive clue fusion and generate segmentation outputs indicating specific regions that are tampered with. Moreover, we construct the ForgeryAnalysis dataset through a chain-of-clues process, which includes analysis and reasoning text to upgrade the image manipulation detection task. A data engine is also introduced to build a larger-scale dataset for the pre-training phase. Our extensive experiments demonstrate the effectiveness of ForgeryAnalysis and show that ForgerySleuth significantly outperforms existing methods in generalization, robustness, and explainability.



Paperid:4495
Authors:Linjian Meng, Youzhi Zhang, Zhenxing Ge, Tianyu Ding, Shangdong Yang, Zheng Xu, Wenbin Li, Yang Gao
Abstract:
Abstract:Regret Matching$^+$ (RM$^+$) variants are widely used to build superhuman Poker AIs, yet few studies investigate their last-iterate convergence in learning a Nash equilibrium (NE). Although their last-iterate convergence is established for games satisfying the Minty Variational Inequality (MVI), no studies have demonstrated that these algorithms achieve such convergence in the broader class of games satisfying the weak MVI. A key challenge in proving last-iterate convergence for RM$^+$ variants in games satisfying the weak MVI is that the proof of last-iterate convergence typically requires that the feedback of algorithms satisfies the weak MVI, which does not hold in RM$^+$ variants. To provide last-iterate convergence for RM$^+$ variants, we introduce a concise yet novel proof paradigm that involves: (i) transforming an RM$^+$ variant into an Online Mirror Descent (OMD) instance that updates within the original strategy space of the game to recover the weak MVI, and (ii) showing last-iterate convergence by proving the distance between accumulated regrets converges to zero via the recovered weak MVI of the feedback. Inspired by our proof paradigm, we propose Smooth Optimistic Gradient Based RM$^+$ (SOGRM$^+$) and show that it achieves last-iterate and finite-time best-iterate convergence in learning an NE of games satisfying the weak MVI, the weakest condition among all known RM$^+$ variants. Experiments show that SOGRM$^+$ significantly outperforms other algorithms.



Paperid:4496
Authors:Itzik Waizman, Yakov Gusakov, Itay Benou, Tammy Riklin Raviv
Abstract:
White matter tractography is an advanced neuroimaging technique that reconstructs the 3D white matter pathways of the brain from diffusion MRI data. It can be framed as a pathfinding problem aiming to infer neural fiber trajectories from noisy and ambiguous measurements, facing challenges such as crossing, merging, and fanning white-matter configurations.In this paper, we propose a novel tractography method that leverages Transformers to model the sequential nature of white matter streamlines, enabling the prediction of fiber directions by integrating both the trajectory context and current diffusion MRI measurements. To incorporate spatial information, we utilize CNNs that extract microstructural features from local neighborhoods around each voxel. By combining these complementary sources of information, our approach improves the precision and completeness of neural pathway mapping compared to traditional tractography models. We evaluate our method with the Tractometer toolkit, achieving competitive performance against state-of-the-art approaches, and present qualitative results on the TractoInferno dataset, demonstrating strong generalization to real-world data. The code attached to this submission will be made publicly available upon acceptance.



Paperid:4497
Authors:Yuchen Liang, Renxiang Huang, Lifeng LAI, Ness Shroff, Yingbin Liang
Abstract:
Abstract:Discrete state space diffusion models have shown significant advantages in applications involving discrete data, such as text and image generation. It has also been observed that their performance is highly sensitive to the choice of rate matrices, particularly between uniform and absorbing rate matrices. While empirical results suggest that absorbing rate matrices often yield better generation quality compared to uniform rate matrices, existing theoretical works have largely focused on the uniform rate matrices case. Notably, convergence guarantees and error analyses for absorbing diffusion models are still missing. In this work, we provide the first finite-time error bounds and convergence rate analysis for discrete diffusion models using absorbing rate matrices. We begin by deriving an upper bound on the KL divergence of the forward process, introducing a surrogate initialization distribution to address the challenge posed by the absorbing stationary distribution, which is a singleton and causes the KL divergence to be ill-defined. We then establish the first convergence guarantees for both the $\tau$-leaping and uniformization samplers under absorbing rate matrices, demonstrating improved rates over their counterparts using uniform rate matrices. Furthermore, under suitable assumptions, we provide convergence guarantees without early stopping. Our analysis introduces several new technical tools to address challenges unique to absorbing rate matrices. These include a Jensen-type argument for bounding forward process convergence, novel techniques for bounding absorbing score functions, and a non-divergent upper bound on the score near initialization that removes the need of early-stopping.



Paperid:4498
Authors:Wenbo Shang, Zihan Feng, Yang Yajun, Xin Huang
Abstract:
Abstract:Information diffusion prediction, which aims to forecast future infected users during the information spreading process on social platforms, is a challenging and critical task for public opinion analysis. With the development of social platforms, mass communication has become increasingly widespread. However, most existing methods based on GNNs and sequence models mainly focus on structural and temporal patterns in social networks, suffering from spurious diffusion connections and insufficient information for the diffusion analysis. We leverage strong reasoning capability of LLMs and develop a LL**M**-based causal framework for d**i**ffusion inf**l**uence **d**erivation (MILD). Comprehensively integrating four key factors of social diffusion, i.e., connections, active timelines, user profiles, and comments, MILD causally infers authentic diffusion links to construct a diffusion influence graph $G_I$. To validate the quality and reliability of our constructed graph $G_I$, we proposed a newly designed set of evaluation metrics w.r.t. diffusion prediction. We show MILD provides a reliable information diffusion structure that 12% absolutely better than the social network structure and achieves the state-of-the-art performance on diffusion prediction. MILD is expected to contribute to high-quality, more explainable, and more trustworthy public opinion analysis.



Paperid:4499
Authors:Zirui Wang, Wenjing Bian, Xinghui Li, Yifu Tao, Jianeng Wang, Maurice Fallon, Victor Prisacariu
Abstract:
Abstract:We introduce Oxford Day-and-Night, a large-scale, egocentric dataset for novel view synthesis (NVS) and visual relocalisation under challenging lighting conditions. Existing datasets often lack crucial combinations of features such as ground-truth 3D geometry, wide-ranging lighting variation, and full 6DoF motion. Oxford Day-and-Night addresses these gaps by leveraging Meta ARIA glasses to capture egocentric video and applying multi-session SLAM to estimate camera poses, reconstruct 3D point clouds, and align sequences captured under varying lighting conditions, including both day and night. The dataset spans over 30 km of recorded trajectories and covers an area of $40{,}000\mathrm{m}^2$, offering a rich foundation for egocentric 3D vision research. It supports two core benchmarks, NVS and relocalisation, providing a unique platform for evaluating models in realistic and diverse environments.



Paperid:4500
Authors:Qizhe Zhang, Mengzhen Liu, Lichen Li, Ming Lu, Yuan Zhang, Junwen Pan, Qi She, Shanghang Zhang
Abstract:
In multimodal large language models (MLLMs), the length of input visual tokens is often significantly greater than that of their textual counterparts, leading to a high inference cost. Many works aim to address this issue by removing redundant visual tokens. However, current approaches either rely on attention-based pruning, which retains numerous duplicate tokens, or use similarity-based pruning, overlooking the instruction relevance, consequently causing suboptimal performance. In this paper, we go beyond attention or similarity by proposing a novel visual token pruning method namedCDPruner, which maximizes the conditional diversity of retained tokens. We first define the conditional similarity between visual tokens conditioned on the instruction, and then reformulate the token pruning problem with determinantal point process (DPP) to maximize the conditional diversity of the selected subset. The proposed CDPruner is training-free and model-agnostic, allowing easy application to various MLLMs. Extensive experiments across diverse MLLMs show that CDPruner establishes new state-of-the-art on various vision-language benchmarks. By maximizing conditional diversity through DPP, the selected subset better represents the input images while closely adhering to user instructions, thereby preserving strong performance even with high reduction ratios. When applied to LLaVA, CDPruner reduces FLOPs by95\%and CUDA latency by78\%, while maintaining94\%of the original accuracy. Our code will be released.



Paperid:4501
Authors:Jun Jiang, Kejiang Chen, Zijin Yang, Weiming Zhang, Nenghai Yu
Abstract:
Abstract:Generative steganography has emerged as an active research area, yet its practical system is constrained by the inherent secret payload limitation caused by low entropy in generating stego texts. This payload limitation necessitates the use of lengthy stego texts or frequent transmissions, which increases the risk of suspicion by adversaries. Previous studies have mainly focused on payload enhancement through optimized entropy utilization while overlooking the crucial role of secret message processing. To address this gap, we propose StegoZip, a framework that leverages large language models to optimize secret message processing. StegoZip consists of two core components: semantic redundancy pruning and index-based compression coding. The former dynamically prunes the secret message to extract a low-semantic representation, whereas the latter further compresses it into compact binary codes. When integrated with state-of-the-art steganographic methods under lossless decoding, StegoZip achieves 2.5$\times$ the payload of the baselines while maintaining comparable processing time in practice. This enhanced payload significantly improves covertness by mitigating the risks associated with frequent transmissions while maintaining provable content security.



Paperid:4502
Authors:Yizhen Zhang, Yang Ding, Shuoshuo Zhang, Xinchen Zhang, Haoling Li, Zhong-Zhi Li, Peijie Wang, Jie Wu, Lei Ji, Yeyun Gong, yelong shen, Yujiu Yang
Abstract:
Inspired by the impressive reasoning capabilities demonstrated by reinforcement learning approaches like DeepSeek-R1, recent emerging research has begun exploring the use of reinforcement learning (RL) to enhance vision-language models (VLMs) for multimodal reasoning tasks. However, most existing multimodal reinforcement learning approaches remain limited to spatial reasoning within single-image contexts, yet still struggle to generalize to more complex and real-world scenarios involving multi-image positional reasoning, where understanding the relationships across images is crucial. To address this challenge, we propose a general reinforcement learning approach PeRL tailored for interleaved multimodal tasks, and a multi-stage strategy designed to enhance the exploration-exploitation trade-off, thereby improving learning efficiency and task performance. Specifically, we introduce permutation of image sequences to simulate varied positional relationships to explore more spatial and positional diversity. Furthermore, we design a rollout filtering mechanism for resampling to focus on trajectories that contribute most to learning optimal behaviors to exploit learned policies effectively. We evaluate our model on 5 widely-used multi-image benchmarks and 3 single-image benchmarks. Our experiments confirm that PeRL trained model consistently surpasses R1-related and interleaved VLM baselines by a large margin, achieving state-of-the-art performance on multi-image benchmarks, while preserving comparable performance on single-image tasks.



Paperid:4503
Authors:Dongwon Choi, Sunwoo Kim, Juyeon Kim, Kyungho Kim, Geon Lee, Shinhwan Kang, Myunghwan Kim, Kijung Shin
Abstract:
Relational databases (RDBs) are composed of interconnected tables, where relationships between them are defined through foreign keys. Recent research on applying machine learning to RDBs has explored graph-based representations of RDBs, where rows of tables are modeled as nodes, and foreign key relationships are modeled as edges.RDB-to-graph modeling helps capture cross-table dependencies, ultimately leading to enhanced performance across diverse tasks.However, there are numerous ways to model RDBs as graphs, and performance varies significantly depending on the chosen graph model. In our analysis, applying a common heuristic rule for graph modeling leads to up to a 10% drop in performance compared to the best-performing graph model, which remains non-trivial to identify.To foster research on intelligent RDB-to-graph modeling, we introduce RDB2G-Bench, the first benchmark framework for evaluating such methods.We construct extensive datasets covering 5 real-world RDBs and 12 predictive tasks, resulting in around 50k graph-performance pairs for efficient and reproducible evaluations. Thanks to our precomputed datasets, we were able to benchmark 9 automatic RDB-to-graph modeling methods on the 12 tasks over 600x faster than on-the-fly evaluation, which requires repeated model training.Our analysis of the datasets and benchmark results reveals key structural patterns affecting graph model effectiveness, along with practical implications for effective graph modeling.



Paperid:4504
Authors:Leying Zhang, Yao Qian, Xiaofei Wang, Manthan Thakker, Dongmei Wang, Jianwei Yu, Haibin Wu, Yuxuan Hu, Jinyu Li, Yanmin Qian, sheng zhao
Abstract:
Generating natural-sounding, multi-speaker dialogue is crucial for applications such as podcast creation, virtual agents, and multimedia content generation. However, existing systems struggle to maintain speaker consistency, model overlapping speech, and synthesize coherent conversations efficiently. In this paper, we introduce CodiFM, a fully non-autoregressive framework for zero-shot multi-talker dialogue generation. CodiFM directly predicts mel-spectrograms from multi-stream transcriptions using a flow-matching-based generative model, eliminating the reliance on intermediate token representations. To better capture realistic conversational dynamics, we propose transcription-level speaker disentanglement, sentence-level alignment, and prompt-level random masking strategies. Our approach achieves state-of-the-art performance, outperforming strong baselines like MoonCast and Sesame in speech quality, speaker consistency, and inference speed. Notably, CodiFM operates without requiring transcriptions for the prompt and supports controllable dialogue generation, including overlapping speech and precise timing control, demonstrating strong generalizability to real-world speech generation scenarios. Audio samples are available at the supplementary material.



Paperid:4505
Authors:Qingqing Yang, Hsin-Hung Li
Abstract:
Working memory (WM) supports the temporary retention of information but is limited in capacity and subject to noise. Flexibly allocating WM resource is therefore critical for adaptive behavior. While some studies show that humans can prioritize information based on attention, others suggest WM is insensitive to reward—raising open questions about whether humans can efficiently allocate WM resource based on utility or environmental statistics. To address this, we conducted behavioral experiments where participants learned stimulus–reward associations and performed a delayed-estimation WM task. WM precision was influenced by both environmental statistics and rewards. When rewards were uniform, memory was more precise for cardinal orientations (the oblique effect)—reflecting sensitivity to the naturalistic priors. When rewards were manipulated, high-reward information was maintained more stably, with effects increasing over time—indicating dynamic, value-guided resource allocation. To interpret these findings, we (1) evaluated an ideal observer model with efficient coding, showing that optimal WM allocation should reflect both stimulus probability and reward, and that these effects accumulate over delays. (2) We trained recurrent neural networks (RNNs) to maximize utility under a 2×2 design: prior (uniform vs. environmental) × reward policy (baseline vs. reward context). In the RNNs, memory is more stable for stimuli associated with higher probability or rewards, mirroring human behavior in leveraging both environmental statistics and rewards. Together, these results provide converging behavioral and computational evidence that WM resource allocation is shaped by both environmental statistics and rewards. This work offers insight into how intelligent systems can dynamically optimize memory under resource constraints.



Paperid:4506
Authors:Yidong Wu, Siyuan Chen, Binrui Wu, Fan Li, Jiechao Gao
Abstract:
In large-scale recommendation systems, multimodal content is increasingly introduced to enhance the generalization of ID features.The rise of Multimodal Large Language Models (MLLMs) enables the construction of unified user and item representations.However, the semantic distribution gap between MM and ID representations leads to \textit{convergence inconsistency} during joint training: the ID branch converges quickly, while the MM branch requires more epochs, thus limiting overall performance.To address this, we propose a two-stage framework including MM representation learning and joint training optimization.First, we fine-tune the MLLM to generate unified user and item representations, and introduce collaborative signals by post-aligning user ID representations to alleviate semantic differences.Then, we propose an Adaptive Gradient Masking (AGM) training strategy to dynamically regulate parameter updates between ID and MLLM branches.AGM estimates the contribution of each representation with mutual information, and applies non-uniform gradient masking at the sub-network level to balance optimization.We provide theoretical analysis of AGM's effectiveness and further introduce an unbiased variant, AGM*, to enhance training stability.Experiments on offline and online A/B tests validate the effectiveness of our approach in mitigating convergence inconsistency and improving performance.



Paperid:4507
Authors:HEE BIN YOO, Sungyoon Lee, Cheongjae Jang, Dong-Sig Han, Jaein Kim, Seunghyeon Lim, Byoung-Tak Zhang
Abstract:
Neural networks learn effective feature representations, enabling feature transfer to new tasks without additional training.While larger datasets are known to improve feature transfer, the theoretical conditions for their success remain unclear.This work investigates feature transfer in classifier-trained networks to identify the conditions that enable effective clustering in unseen classes.We first reveal higher similarity between training and unseen distributions to improve Cohesion and Separability.We then show enhanced feature expressiveness when inputs are similar to the training classes, while the features of irrelevant inputs are indistinguishable.We validate our analysis on synthetic and benchmark datasets, including CAR, CUB, SOP, ISC, and ImageNet.Our analysis highlights the importance of training class similarity to the input distribution for successful feature transfer.



Paperid:4508
Authors:Shi Qiu, Shaoyang Guo, Zhuo-Yang Song, Yunbo Sun, Zeyu Cai, Jiashen Wei, Tianyu Luo, Yixuan Yin, Zhang Haoxu, Yi Hu, Chenyang Wang, Chencheng Tang, Haoling Chang, Qi Liu, Ziheng Zhou, Tianyu Zhang, Jingtian Zhang, Zhangyi Liu, Minghao Li, Yuku Zhang, Boxuan Jing, Xianqi Yin, Yutong Ren, Zizhuo Fu, Jiaming Ji, Weike Wang, Xudong Tian, Anqi Lv, Laifu Man, Jianxiang Li, Feiyu Tao, Qihua Sun, Zhou Liang, Yushu Mu, Zhongxuan Li, Jing-Jun Zhang, Shutao Zhang, Xiaotian Li, Xingqi Xia, Jiawei Lin, Zheyu Shen, Jiahang Chen, Qiuhao Xiong, Binran Wang, Fengyuan Wang, Niziyang, Bohan Zhang, Fan Cui, shaochangkun, Qing-Hong Cao, Ming-xing Luo, Muhan Zhang, Hua Xing Zhu
Abstract:
Current benchmarks for evaluating the reasoning capabilities of Large Language Models (LLMs) face significant limitations: task oversimplification, data contamination, and flawed evaluation items. These deficiencies necessitate more rigorous assessment methods. To address these limitations, we introduce PHYBench, a benchmark of 500 original physics problems ranging from high school to Physics Olympiad difficulty. PHYBench addresses data contamination through original content and employs a systematic curation pipeline to eliminate flawed items. Evaluations show that PHYBench activates more tokens and provides stronger differentiation between reasoning models compared to other baselines like AIME 2024, OlympiadBench and GPQA. Even the best-performing model, Gemini 2.5 Pro, achieves only 36.9\% accuracy compared to human experts' 61.9\%. To further enhance evaluation precision, we introduce the Expression Edit Distance (EED) Score for mathematical expression assessment, which improves sample efficiency by 204\% over binary scoring. Moreover, PHYBench effectively elicits multi-step and multi-condition reasoning, providing a platform for examining models' reasoning robustness, preferences, and deficiencies. The benchmark results and dataset are publicly available at https://www.phybench.cn/.



Paperid:4509
Authors:Penghao Wang, Yiyang He, Xin Lv, Yukai Zhou, Lan Xu, Jingyi Yu, Jiayuan Gu
Abstract:
Understanding objects at the level of their constituent parts is fundamental to advancing computer vision, graphics, and robotics. While datasets like PartNet have driven progress in 3D part understanding, their reliance on untextured geometries and expert-dependent annotation limits scalability and usability. We introduce PartNeXt, a next-generation dataset addressing these gaps with over 23000 high-quality, textured 3D models annotated with fine-grained, hierarchical part labels across 50 categories. We benchmark PartNeXt on two tasks: (1) class-agnostic part segmentation, where state-of-the-art methods (e.g., PartField, SAMPart3D) struggle with fine-grained and leaf-level parts, and (2) 3D part-centric question answering, a new benchmark for 3D-LLMs that reveals significant gaps in open-vocabulary part grounding. Additionally, training Point-SAM on PartNeXt yields substantial gains over PartNet, underscoring the dataset’s superior quality and diversity. By combining scalable annotation, texture-aware labels, and multi-task evaluation, PartNeXt opens new avenues for research in structured 3D understanding.



Paperid:4510
Authors:Zihan Qiu, Zekun Wang, Bo Zheng, Zeyu Huang, Kaiyue Wen, Songlin Yang, Rui Men, Le Yu, Fei Huang, Suozhi Huang, Dayiheng Liu, Jingren Zhou, Junyang Lin
Abstract:
Gating mechanisms have been widely utilized, from early models like LSTMs and Highway Networks to recent state space models, linear attention, and also softmax attention.Yet, existing literature rarely examines the specific effects of gating.In this work, we conduct comprehensive experiments to systematically investigate gating-augmented softmax attention variants.Specifically, we perform a comprehensive comparison over 30 variants of 15B Mixture-of-Experts (MoE) models and 1.7B dense models trained on a 3.5 trillion token dataset.Our central finding is that a simple modification—applying a head-specific sigmoid gate after the Scaled Dot-Product Attention (SDPA)—consistently improves performance.This modification also enhances training stability, tolerates larger learning rates, and improves scaling properties.By comparing various gating positions and computational variants, we attribute this effectiveness to two key factors: (1) introducing non-linearity upon the low-rank mapping in the softmax attention, and (2) applying query-dependent sparse gating scores to modulate the SDPA output.Notably, we find this sparse gating mechanism mitigatesmassive activation,attention sinkand enhances long-context extrapolation performance. We will release related models to facilitate future research.



Paperid:4511
Authors:Xinyi Zhang, Bingyang Wei, Ruixuan Yu, Jian Sun
Abstract:
We propose a novel two-stage framework, Coarse-to-Fine Part Assembly (CFPA), for 3D shape assembly from basic parts. Effective part assembly demands precise local geometric reasoning for accurate component assembly, as well as global structural understanding to ensure semantic coherence and plausible configurations. CFPA addresses this challenge by integrating semantic abstraction and symmetry-aware reasoning into a unified pose prediction process. In the first stage, semantic super-parts are constructed via an optimal transport formulation to capture high-level object structure, which is then propagated to individual parts through a dual-range feature propagation mechanism. The second stage refines part poses via cross-stage feature interaction and instance-level geometric encoding, improving spatial precision and coherence. To enable diverse yet valid assemblies, we introduce a symmetry-aware loss that jointly models both self-symmetry and inter-part geometric similarity, allowing for diverse but structurally consistent assemblies. Extensive experiments on the PartNet benchmark demonstrate that CFPA achieves state-of-the-art performance in assembly accuracy, structural consistency, and diversity across multiple categories.



Paperid:4512
Authors:Jia-Hua Lee, Wei-Fang Sun, Bor-Jiun Lin, Chun-Yi Lee
Abstract:
World models represent a promising approach for training reinforcement learning agents with significantly improved sample efficiency. While most world model methods primarily rely on sequences of discrete latent variables to model environment dynamics, this compression often neglects critical visual details essential for reinforcement learning. Recent diffusion-based world models condition generation on a fixed context length of frames to predict the next observation, using separate recurrent neural networks to model rewards and termination signals. Although this architecture effectively enhances visual fidelity, the fixed context length approach inherently limits memory capacity.In this paper, we introduce EDELINE, a unified world model architecture that integrates state space models with diffusion models. Our approach outperforms existing baselines across visually challenging Atari 100k tasks, memory-demanding Crafter benchmark, and 3D first-person ViZDoom environments, demonstrating superior performance in all these diverse challenges.



Paperid:4513
Authors:Haeun Lee, Omin Kwon, Yeonhong Park, Jae Lee
Abstract:
Large Language Models (LLMs) are increasingly deployed in latency-critical, high-throughput services such as virtual assistants, code generation, and multi-modal applications. To satisfy stringent service-level objectives (SLOs), techniques such as continuous batching, paged attention, and hybrid scheduling of prefill and decode phases have been developed. While weight and activation quantization has been widely explored to accelerate inference, methods for adapting precision dynamically and efficiently at runtime remain largely underexplored. Recent hardware support for FP8 arithmetic—offering up to 2× the throughput of FP16—makes it an attractive option for interactive LLM serving. However, naively co-deploying separate FP8 and FP16 models increases storage overhead and may sacrifice accuracy or performance.We present Dual-FP LLM, a precision-adaptive serving technique that enables both FP8 and FP16 inference from a single 16-bit model representation. By decomposing each FP16 weight into two 8-bit components, our method enables efficient FP8 execution while preserving full FP16 accuracy—without incurring additional memory cost. We implement a custom CUTLASS-based GEMM kernel that reconstructs FP16 operands on-the-fly and integrated it into the vLLM serving framework. Dual-FP LLM delivers up to 1.55x throughput improvement in FP8 mode, with negligible accuracy degradation, and introduces only up to 4.51% average overhead for FP16 inference. This enables dynamic, SLO-aware precision selection and offers a flexible foundation for scalable LLM serving under bursty, heterogeneous workloads.



Paperid:4514
Authors:Can Rong, Xin Zhang, Yanxin Xi, HONGJIE SUI, Jingtao Ding, Yong Li
Abstract:
Commuting Origin-destination (OD) flows, capturing daily population mobility of citizens, are vital for sustainable development across cities around the world. However, it is challenging to obtain the data due to the high cost of travel surveys and privacy concerns. Surprisingly, we find that satellite imagery, publicly available across the globe, contains rich urban semantic signals to support high-quality OD flow generation, with over 98\% expressiveness of traditional multisource hard-to-collect urban sociodemographic, economics, land use, and point of interest data. This inspires us to design a novel data generator, GlODGen, which can generate OD flow data for any cities of interest around the world. Specifically, GlODGen first leverages Vision-Language Geo-Foundation Models to extract urban semantic signals related to human mobility from satellite imagery. These features are then combined with population data to form region-level representations, which are used to generate OD flows via graph diffusion models. Extensive experiments on 4 continents and 6 representative cities show that GlODGen has great generalizability across diverse urban environments on different continents and can generate OD flow data for global cities highly consistent with real-world mobility data. We implement GlODGen as an automated tool, seamlessly integrating data acquisition and curation, urban semantic feature extraction, and OD flow generation together. It has been released at https://github.com/tsinghua-fib-lab/generate-od-pubtools.



Paperid:4515
Authors:Shoutao Guo, Shaolei Zhang, Qingkai Fang, Zhengrui Ma, Min Zhang, Yang Feng
Abstract:
The rapid advancement of Large Language Models (LLMs) has spurred significant progress in Large Speech-Language Models (LSLMs), enhancing their capabilities in both speech understanding and generation. While existing LSLMs often concentrate on augmenting speech generation or tackling a diverse array of short-speech tasks, the efficient processing of long-form speech remains a critical yet underexplored challenge. This gap is primarily attributed to the scarcity of long-speech training datasets and the high computational costs associated with long sequences. To address these limitations, we introduce FastLongSpeech, a novel framework designed to extend LSLM capabilities for efficient long-speech processing without necessitating dedicated long-speech training data. FastLongSpeech incorporates an iterative fusion strategy that can compress excessively long-speech sequences into manageable lengths. To adapt LSLMs for long-speech inputs, it introduces a dynamic compression training approach, which exposes the model to short-speech sequences at varying compression ratios, thereby transferring the capabilities of LSLMs to long-speech tasks. To assess the long-speech capabilities of LSLMs, we develop a long-speech understanding benchmark called LongSpeech-Eval. Experiments show that our method exhibits strong performance in both long-speech and short-speech tasks, while greatly improving inference efficiency.



Paperid:4516
Authors:Yuwu Lu, Chunzhi Liu
Abstract:
Black-box Unsupervised Domain Adaptation (BUDA) aims to transfer source domain knowledge to an unlabeled target domain, without accessing the source data or trained source model. Recent diffusion models have significantly advanced the ability to generate images from texts. While they can produce realistic visuals across diverse prompts and demonstrate impressive compositional generalization, these diffusion-based domain adaptation methods focus solely on composition, overlooking their sensitivity to textual nuances. In this work, we propose a novel diffusion-based method, called Rectifying-reasoning Errors of Diffusion (RrED) for BUDA. RrED is a two-stage learning strategy under diffusion supervision to effectively enhance the target model via the decomposed text and visual encoders from the diffusion model. Specifically, RrED consists of two stages: Diffusion-Target model Rectification (DTR) and Self-rectifying Reasoning Model (SRM). In DTR, we decouple the image and text encoders within the diffusion model: the visual encoder integrates our proposed feature-sensitive module to generate inferentially-enhanced visuals, while the text encoder enables multi-modal joint fine-tuning. In SRM, we prioritize the BUDA task itself, leveraging the target model's differential reasoning capability to rectify errors during learning. Extensive experiments confirm that RrED significantly outperforms other methods on four benchmark datasets, demonstrating its effectiveness in enhancing reasoning and generalization abilities.



Paperid:4517
Authors:Jiawei Gu, Ziyue Qiao, Xinming Li, Zechao Li
Abstract:
Abstract:Global Covariance Pooling (GCP) has garnered increasing attention in visual recognition tasks, where second-order statistics frequently yield stronger representations than first-order approaches. However, two main streams of GCP---Newton--Schulz-based iSQRT-COV and exact or near-exact SVD methods---struggle at opposite ends of the training spectrum. While iSQRT-COV stabilizes early learning by avoiding large gradient explosions, it over-compresses significant eigenvalues in later stages, causing an \emph{over-flattening} phenomenon that stalls final accuracy. In contrast, SVD-based methods excel at preserving the high-eigenvalue structure essential for deep networks but suffer from sensitivity to small eigenvalue gaps early on. We propose \textbf{Halley-SVD}, a high-order iterative method that unites the smooth gradient advantages of iSQRT-COV with the late-stage fidelity of SVD. Grounded in Halley's iteration, our approach obviates explicit divisions by $(\lambda_i - \lambda_j)$ and forgoes threshold- or polynomial-based heuristics. As a result, it prevents both early gradient explosions and the excessive compression of large eigenvalues. Extensive experiments on CNNs and transformer architectures show that Halley-SVD consistently and robustly outperforms iSQRT-COV at large model scales and batch sizes, achieving higher overall accuracy without mid-training switches or custom truncations. This work provides a new solution to the long-standing dichotomy in GCP, illustrating how high-order methods can balance robustness and spectral precision to fully harness the representational power of modern deep networks.



Paperid:4518
Authors:Tianhong Zhou, xu yin, Yingtao Zhu, Chuxi Xiao, Haiyang Bian, Lei Wei, Xuegong Zhang
Abstract:
Vision–language models (VLMs) exhibit strong zero-shot generalization on natural images and show early promise in interpretable medical image analysis. However, existing benchmarks do not systematically evaluate whether these models truly reason like human clinicians or merely imitate superficial patterns. To address this gap, we propose DrVD-Bench, the first multimodal benchmark for clinical visual reasoning. DrVD-Bench consists of three modules: Visual Evidence Comprehension, Reasoning Trajectory Assessment, and Report Generation Evaluation, comprising a total of 7,789 image–question pairs. Our benchmark covers 20 task types, 17 diagnostic categories, and five imaging modalities—CT, MRI, ultrasound, radiography, and pathology. DrVD-Bench is explicitly structured to reflect the clinical reasoning workflow from modality recognition to lesion identification and diagnosis. We benchmark 19 VLMs, including general-purpose and medical-specific, open-source and proprietary models, and observe that performance drops sharply as reasoning complexity increases. While some models begin to exhibit traces of human-like reasoning, they often still rely on shortcut correlations rather than grounded visual understanding. DrVD-Bench offers a rigorous and structured evaluation framework to guide the development of clinically trustworthy VLMs.



Paperid:4519
Authors:Xingguang Wei, Haomin Wang, Shenglong Ye, Ruifeng Luo, Zhang, Lixin Gu, Jifeng Dai, Yu Qiao, Wenhai Wang, Hongjie Zhang
Abstract:
We study the task of panoptic symbol spotting, which involves identifying both individual instances of countable \textit{things} and the semantic regions of uncountable \textit{stuff} in computer-aided design (CAD) drawings composed of vector graphical primitives.Existing methods typically rely on image rasterization, graph construction, or point-based representation, but these approaches often suffer from high computational costs, limited generality, and loss of geometric structural information. In this paper, we propose \textit{VecFormer}, a novel method that addresses these challenges through \textit{line-based representation} of primitives. This design preserves the geometric continuity of the original primitive, enabling more accurate shape representation while maintaining a computation-friendly structure, making it well-suited for vector graphic understanding tasks. To further enhance prediction reliability, we introduce a \textit{Branch Fusion Refinement} module that effectively integrates instance and semantic predictions, resolving their inconsistencies for more coherent panoptic outputs. Extensive experiments demonstrate that our method establishes a new state-of-the-art, achieving 91.1 PQ, with Stuff-PQ improved by 9.6 and 21.2 points over the second-best results under settings with and without prior information, respectively—highlighting the strong potential of line-based representation as a foundation for vector graphic understanding.



Paperid:4520
Authors:Qianxun Xu, Zuchuan Li
Abstract:
The integration of big data, physical laws, and machine learning algorithms has shown potential to improve the estimation and understanding of complex real-world systems. However, integrating physical laws with uncertainties in machine learning algorithms remains understudied. In this work, we bridge this gap through developing the Partial-Physics Informed Diffusion Model (PPIDM), a novel framework that integrates known physical principles through a physics operator while reducing the impact of unknown dynamics by minimizing related discrepancies. We showcase PPIDM’s capabilities using surface chlorophyll concentration data, which is influenced by both physical and biological processes, while the latter is poorly constrained. Experimental results reveal that PPIDM achieves substantially improved prediction accuracy and stability, significantly outperforming baseline methods that either neglect physics entirely or misapply incomplete physical laws under an assumption of completeness.



Paperid:4521
Authors:Steffen Schotthöfer, Lexie Yang, Stefan Schnake
Abstract:
Deployment of neural networks on resource-constrained devices demands models that are both compact and robust to adversarial inputs. However, compression and adversarial robustness often conflict. In this work, we introduce a dynamical low-rank training scheme enhanced with a novel spectral regularizer that controls the condition number of the low-rank core in each layer. This approach mitigates the sensitivity of compressed models to adversarial perturbations without sacrificing clean accuracy. The method is model- and data-agnostic, computationally efficient, and supports rank adaptivity to automatically compress the network at hand. Extensive experiments across standard architectures, datasets, and adversarial attacks show the regularized networks can achieve over 94 compression while recovering or improving adversarial accuracy relative to uncompressed baselines.



Paperid:4522
Authors:J Rosser, Jakob Foerster
Abstract:
Scaffolding Large Language Models (LLMs) into multi-agent systems often improves performance on complex tasks, but the safety impact of such scaffolds has not been thoroughly explored. We introduce AgentBreeder, a framework for multi-objective self-improving evolutionary search over scaffolds. We evaluate discovered scaffolds on widely recognized reasoning, mathematics, and safety benchmarks and compare them with popular baselines. In 'blue' mode, we see a 79.4% average uplift in safety benchmark performance while maintaining or improving capability scores. In 'red' mode, we find adversarially weak scaffolds emerging concurrently with capability optimization. Our work demonstrates the risks of multi-agent scaffolding and provides a framework for mitigating them. Code is available at https://anonymous.4open.science/r/AgentBreeder-86AF .



Paperid:4523
Authors:Xudong Yan, Songhe Feng
Abstract:
Compositional Zero-Shot Learning (CZSL) aims to recognize novel attribute-object compositions based on the knowledge learned from seen ones. Existing methods suffer from performance degradation caused by the distribution shift of label space at test time, which stems from the inclusion of unseen compositions recombined from attributes and objects. To overcome the challenge, we propose a novel approach that accumulates comprehensive knowledge in both textual and visual modalities from unsupervised data to update multi-modal prototypes at test time. Building on this, we further design an adaptive update weight to control the degree of prototype adjustment, enabling the model to flexibly adapt to distribution shift during testing. Moreover, a dynamic priority queue is introduced that stores high-confidence images to acquire visual knowledge from historical images for inference. Considering the semantic consistency of multimodal knowledge, we align textual and visual prototypes by multimodal collaborative representation learning. Extensive experiments indicate that our approach achieves state-of-the-art performance on four challenging benchmark datasets under both closed-world and open-world settings. We will release the source code.



Paperid:4524
Authors:Yiming Gao, Zhen Wang, Jefferson Chen, Mark Antkowiak, Mengzhou Hu, JungHo Kong, Dexter Pratt, Jieyuan Liu, Enze Ma, Zhiting Hu, Eric Xing
Abstract:
We present scPilot, the first systematic framework to practice omics-native reasoning: a large language model (LLM) converses in natural language while directly inspecting single-cell RNA-seq data and on-demand bioinformatics tools. scPilot converts core analyses, i.e., cell-type annotation, developmental-trajectory reconstruction, and transcription-factor targeting, into step-by-step reasoning problems that the model must solve, justify, and, when needed, revise with new evidence. To measure progress, we release scBench, a suite of 9 expertly curated datasets and graders that faithfully evaluate the omics-native reasoning capability of scPilot w.r.t various LLMs. Experiments with o1 show that iterative omics-native reasoning lifts average accuracy by 11\% for annotation and Gemini 2.5 Pro cuts trajectory graph-edit distance by 30\% versus one-shot prompting, while revealing systematic failure modes in gene-regulatory prediction. Our results demonstrate that grounding LLMs in raw omics data yields transparent, auditable analyses and opens a path toward fully automated, interpretable, and scientifically robust single-cell analysis workflows.



Paperid:4525
Authors:Tobias Würth, Niklas Freymuth, Gerhard Neumann, Luise Kärger
Abstract:
Graph-based learned simulators have emerged as a promising approach for simulating physical systems on unstructured meshes, offering speed and generalization across diverse geometries. However, they often struggle with capturing global phenomena, such as bending or long-range correlations, and suffer from error accumulation over long rollouts due to their reliance local message passing and direct next-step prediction. We address these limitations by introducing the Rolling Diffusion-Batched Inference Network (ROBIN), a novel learned simulator that integrates two key innovations: (i) Rolling Diffusion, a parallelized inference scheme that amortizes the cost of diffusion-based refinement across physical time steps by overlapping denoising steps across a temporal window. (ii) A Hierarchical Graph Neural Network built on algebraic multigrid coarsening, enabling multiscale message passing across different mesh resolutions. This architecture, implemented via Algebraic-hierarchical Message Passing Networks, captures both fine-scale local dynamics and global structural effects critical for phenomena like beam bending or multi-body contact. We validate ROBIN on challenging 2D and 3D solid mechanics benchmarks involving geometric, material, and contact nonlinearities. ROBIN achieves state-of-the-art accuracy on all tasks, substantially outperforming existing next-step learned simulators while reducing inference time by up to an order of magnitude compared to standard diffusion simulators.



Paperid:4526
Authors:Yifan Yang, Zhen Zhang, Rupak Swaminathan, Jing Liu, Nathan Susanj, Zheng Zhang
Abstract:
Fine-tuning vision language models (VLMs) has achieved remarkable performance across various downstream tasks; yet, it requires access to model gradients through backpropagation (BP), making them unsuitable for memory-constrained, inference-only edge devices. To address this limitation, previous work has explored various BP-free fine-tuning methods. However, these approaches often rely on high-variance evolutionary strategies (ES) or zeroth-order (ZO) optimization, and often fail to achieve satisfactory performance. In this paper, we propose a hybrid Sharpness-aware Zeroth-order optimization (SharpZO) approach, specifically designed to enhance the performance of ZO VLM fine-tuning via a sharpness-aware warm-up training. SharpZO features a two-stage optimization process: a sharpness-aware ES stage that globally explores and smooths the loss landscape to construct a strong initialization, followed by a fine-grained local search via sparse ZO optimization. The entire optimization relies solely on forward passes. Detailed theoretical analysis and extensive experiments on CLIP models demonstrate that SharpZO significantly improves accuracy and convergence speed, achieving up to 7\% average gain over state-of-the-art forward-only methods.



Paperid:4527
Authors:Tianyi Tan, Yinan Zheng, Ruiming Liang, Zexu Wang, Kexin ZHENG, Jinliang Zheng, Jianxiong Li, Xianyuan Zhan, Jingjing Liu
Abstract:
Modeling interactive driving behaviors in complex scenarios remains a fundamental challenge for autonomous driving planning. Learning-based approaches attempt to address this challenge with advanced generative models, removing the dependency on over-engineered architectures for representation fusion. However, brute-force implementation by simply stacking transformer blocks lacks a dedicated mechanism for modeling interactive behaviors that is common in real driving scenarios. The scarcity of interactive driving data further exacerbates this problem, leaving conventional imitation learning methods ill-equipped to capture high-value interactive behaviors. We propose Flow Planner, which tackles these problems through coordinated innovations in data modeling, model architecture, and learning scheme. Specifically, we first introduce fine-grained trajectory tokenization, which decomposes the trajectory into overlapping segments to decrease the complexity of whole trajectory modeling. With a sophisticatedly designed architecture, we achieve efficient temporal and spatial fusion of planning and scene information, to better capture interactive behaviors. In addition, the framework incorporates flow matching with classifier-free guidance for multi-modal behavior generation, which dynamically reweights agent interactions during inference to maintain coherent response strategies, providing a critical boost for interactive scenario understanding. Experimental results on the large-scale nuPlan dataset demonstrate that Flow Planner achieves state-of-the-art performance among learning-based approaches while effectively modeling interactive behaviors in complex driving scenarios.



Paperid:4528
Authors:Hongyu Shen, Zhizhen Jane Zhao
Abstract:
Abstract:Recent advances in false discovery rate (FDR)-controlled methods have enhanced reliability by limiting false positives, making them particularly suitable for applications in complex scenarios. This paper identifies the limitation of a so-called "mirror statistics" that is introduced in a prominent FDR-controlled framework---data splitting. Particularly, the paper addresses a unit variance constraint on the mirror test statistics that potentially reduces the feature selection power. From another angle, we generalize the mirror statistics with the Gaussian mirror framework, considering the variance information in forming a new test statistics, and name it the "generalized Gaussian mirror" ($\text{G}^2\text{M}$) method. We demonstrate both theoretically and empirically that the proposed test statistics achieves higher power than those of Gaussian mirror and data splitting. Comparisons with other FDR-controlled frameworks across synthetic, semi-synthetic, and real datasets highlight the superior performance of the $\text{G}^2\text{M}$ method in achieving higher power while maintaining FDR control. These findings suggest the potential for the $\text{G}^2\text{M}$ method for practical applications in real-world problems.



Paperid:4529
Authors:Jisoo Kim, Sungmin Kang, Sunwoo Lee
Abstract:
Expensive communication cost is a common performance bottleneck in Federated Learning (FL), which makes it less appealing in real-world applications. Many communication-efficient FL methods focus on discarding a part of model updates mostly based on gradient magnitude. In this study, we find that recycling previous updates, rather than simply dropping them, more effectively reduces the communication cost while maintaining FL performance. We proposeFedLUAR, a Layer-wise Update Aggregation with Recycling scheme for communication-efficient FL. We first define a useful metric that quantifies the extent to which the aggregated gradients influences the model parameter values in each layer.FedLUARselects a few layers based on the metric and recycles their previous updates on the server side. Our extensive empirical study demonstrates that the update recycling scheme significantly reduces the communication cost while maintaining model accuracy. For example, our method achieves nearly the same AG News accuracy as FedAvg, while reducing the communication cost to just 17%.



Paperid:4530
Authors:Danny Driess, Jost Springenberg, Brian Ichter, LILI YU, Adrian Li-Bell, Karl Pertsch, Allen Ren, Homer Walke, Quan Vuong, Lucy Xiaoyang Shi, Sergey Levine
Abstract:
Vision-language-action (VLA) models provide a powerful approach to training control policies for physical systems, such as robots, by combining end-to-end learning with transfer of semantic knowledge from web-scale vision-language model (VLM) training. However, the constraints of real-time control are often at odds with the design of VLMs: the most powerful VLMs have tens or hundreds of billions of parameters, presenting an obstacle to real-time inference, and operate on discrete tokens rather than the continuous-valued outputs that are required for controlling robots. To address this challenge, recent VLA models have used specialized modules for efficient continuous control, such as action experts or continuous output heads, which typically require adding new untrained parameters to the pretrained VLM backbone. While these modules improve real-time and control capabilities, it remains an open question whether they preserve or degrade the semantic knowledge contained in the pretrained VLM, and what effect they have on the VLA training dynamics. In this paper, we study this question in the context of VLAs that include a continuous diffusion or flow matching action expert, showing that naively including such experts significantly harms both training speed and knowledge transfer. We provide an extensive analysis of various design choices, their impact on performance and knowledge transfer, and propose a technique for insulating the VLM backbone during VLA training that mitigates this issue. Videos are available at https://how-to-vla.github.io/ (anonymous website).



Paperid:4531
Authors:Abdelrahman Eldesokey, Aleksandar Cvejić, Bernard Ghanem, Peter Wonka
Abstract:
We propose a novel approach for disentangling visual and semantic features from the backbones of pre-trained diffusion models, enabling the detection of visually inconsistent regions in subject-driven image generation. While diffusion model backbones are known to encode semantically rich features, they should also contain visual features that capture appearance to support their image synthesis capabilities. However, isolating these visual features is non-trivial due to the absence of datasets with annotated visual correspondences.To address this, we design an automated dataset generation pipeline that produces image pairs with annotated semantic and visual correspondences based on existing subject-driven datasets. Using this dataset, we propose an architecture to disentangle semantic and visual features in a contrastive manner. We further propose a metric that leverages the disentangled features to quantify and localize inconsistencies in subject-driven generation.Experiments show that our approach significantly outperforms global feature-based metrics such as CLIP and DINO, as well as Vision-Language Models, in capturing visual inconsistencies. To the best of our knowledge, this is the first approach that enables both quantification and spatial localization of inconsistency in subject-driven image generation, offering a valuable tool for advancing the task.



Paperid:4532
Authors:Yuesen Liao, Zhiwei Li, Binrui Wu, Zihao Cheng, Su Zhao, Shuai Chen, Weizhong Zhang
Abstract:
Sparse and low-rank matrix composite approximation has emerged as a promising paradigm for compressing large language models (LLMs), offering a more flexible pruning structure than conventional methods based solely on sparse matrices. The significant variation in weight redundancy across layers, along with the differing rank and sparsity structures of weight matrices, makes identifying the globally optimal pruning structure extremely challenging. Existing methods often depend on uniform or manually designed heuristic rules to allocate weight sparsity across layers, subsequently compressing each matrix using matrix approximation techniques. Given the above theoretical difficulty in global compression of LLMs and the limited computational and data resources available compared to the training phase, we argue that a collaboration between learning and matrix approximation is essential for effective compression. In this paper, we propose a novel LLM compression framework based on generalized bilevel optimization that naturally formulates an effective collaborative mechanism. Specifically, the outer loop frames the weight allocation task as a probabilistic optimization problem, enabling the automatic learning of both layer-wise sparsities and matrix-wise retained ranks, while the inner loop solves the corresponding sparsity and rank-constrained model compression problem via matrix approximation. Our main technical contributions include two key innovations for efficiently solving this bilevel optimization problem. First, we introduce a truncated Gaussian prior-based probabilistic parameterization integrated with a policy gradient estimator, which avoids expensive backpropagation and stabilizes the optimization process. Second, we design an adapted QR-based matrix approximation algorithm that significantly accelerates inner loop computations. Extensive experiments on Phi-3 and the LLama-2/3 family demonstrate the effectiveness of our method. Notably, it maintains over 95\% zero-shot accuracy under 50\% sparsity and achieves up to 2× inference speedup.



Paperid:4533
Authors:Zhuonan Zheng, Yuanchen Bei, Zhiyao Zhou, Sheng Zhou, Yao Ma, Ming Gu, HONGJIA XU, Jiawei Chen, Jiajun Bu
Abstract:
Graph Neural Networks (GNNs) excel in graph mining tasks thanks to their message-passing mechanism, which aligns with the homophily assumption. However, connected nodes can also exhibit inconsistent behaviors, termed heterophilic patterns, sparking interest in heterophilic GNNs (HTGNNs). Although the message-passing mechanism seems unsuitable for heterophilic graphs owing to the propagation of dissimilar messages, it is still popular in HTGNNs and consistently achieves notable success. Some efforts have investigated such an interesting phenomenon, but are limited in the data perspective. The model-perspective understanding remains largely unexplored, which is conducive to guiding the designs of HTGNNs. To fill this gap, we build the connection between node discriminability and the compatibility matrix (CM). We reveal that the effectiveness of the message passing in HTGNNs may be credited to increasing the proposed Compatibility Matrix Discriminability (CMD). However, the issues of sparsity and noise pose great challenges to leveraging CM. Thus, we propose CMGNN, a novel approach to alleviate these issues while enhancing the CM and node embeddings explicitly. A thorough evaluation involving 13 datasets and comparison against 19 well-established baselines highlights the superiority of CMGNN.



Paperid:4534
Authors:Han Liu, Jiaqi Li, Zhi Xu, Xiaotong Zhang, Xiaoming Xu, Fenglong Ma, Yuanman Li, Hong Yu
Abstract:
Black-box adversarial attack on vision-language pre-training models is a practical and challenging task, as text and image perturbations need to be considered simultaneously, and only the predicted results are accessible. Research on this problem is in its infancy and only a handful of methods are available. Nevertheless, existing methods either rely on complex iterative cross-search strategy which inevitably consume numerous queries, or only consider to reduce the similarity of positive image-text pairs but ignore that of negative ones will also be implicitly diminished, thus inevitably affecting the attack performance. To alleviate the above issues, we propose a simple yet effective framework to generate high quality adversarial examples on vision-language pre-trained models, named HQA-VLAttack, which consists of text and image attack stages. For text perturbation generation, it leverages the counter-fitting word vector to generate the substitute word set, thus guaranteeing the semantic consistency between the substitute word and the original word. For image perturbation generation, it first initializes the image adversarial example via the layer-Importance guided strategy, and then utilizes contrastive learning to optimize the image adversarial perturbation, which ensures that the similarity of positive image-text pairs is decreased while that of negative image-text pairs is increased. In this way, the optimized adversarial images and texts are more likely to retrieve negative examples, thereby enhancing the attack success rate. Experimental results on three benchmark datasets demonstrate that HQA-VLAttack outperforms other strong baselines in attack success rate significantly.



Paperid:4535
Authors:Yuwei Cheng, Zifeng Zhao, Haifeng Xu
Abstract:
Abstract:Online advertising platforms use automated auctions to connect advertisers with potential customers, requiring effective bidding strategies to maximize profits. Accurate ad impact estimation requires considering three key factors: delayed and long-term effects, cumulative ad impacts such as reinforcement or fatigue, and customer heterogeneity. However, these effects are often not jointly addressed in previous studies. To capture these factors, we model ad bidding as a Contextual Markov Decision Process (CMDP) with delayed Poisson rewards. For efficient estimation, we propose a two-stage maximum likelihood estimator combined with data-splitting strategies, ensuring controlled estimation error based on the first-stage estimator's (in)accuracy. Building on this, we design a reinforcement learning algorithm to derive efficient personalized bidding strategies. This approach achieves a near-optimal regret bound of $\tilde{\mathcal{O}}(dH^2\sqrt{T})$, where $d$ is the contextual dimension, $H$ is the number of rounds, and $T$ is the number of customers.



Paperid:4536
Authors:Jitai Hao, Qiang Huang, Hao Liu, Xinyan Xiao, Zhaochun Ren, Jun Yu
Abstract:
Abstract:Training high-performing Small Language Models (SLMs) remains costly, even with knowledge distillation and pruning from larger teacher models. Existing work often faces three key challenges: (1) information loss from hard pruning, (2) inefficient alignment of representations, and (3) underutilization of informative activations, particularly from Feed-Forward Networks (FFNs). To address these challenges, we introduce \textbf{Low-Rank Clone (LRC)}, an efficient pre-training method that constructs SLMs aspiring to behavioral equivalence with strong teacher models. LRC trains a set of low-rank projection matrices that jointly enable soft pruning by compressing teacher weights, and activation clone by aligning student activations, including FFN signals, with those of the teacher. This unified design maximizes knowledge transfer while removing the need for explicit alignment modules. Extensive experiments with open-source teachers (e.g., Llama-3.2-3B-Instruct, Qwen2.5-3B/7B-Instruct) show that LRC matches or surpasses state-of-the-art models trained on trillions of tokens--while using only 20B tokens, achieving over \textbf{1,000$\times$} training efficiency. Our code is available at \url{https://anonymous.4open.science/r/LowRankClone}.



Paperid:4537
Authors:Frank Röder, Jan Benad, Manfred Eppe, Pradeep Banerjee
Abstract:
Real-world reinforcement learning demands adaptation to unseen environmental conditions without costly retraining. Contextual Markov Decision Processes model this challenge, but existing methods often require explicit context variables (e.g., friction, gravity), limiting their use when contexts are latent or hard to measure. We introduce Dynamics-Aligned Latent Imagination (DALI), a framework integrated within the Dreamer architecture that infers latent context representations from agent-environment interactions. By training a self-supervised encoder to predict forward dynamics, DALI generates actionable representations that condition the world model and policy, bridging perception and control. We theoretically prove that this encoder is essential for efficient context inference and robust generalization. DALI’s latent space enables counterfactual consistency: perturbing a gravity-encoding dimension alters imagined rollouts in physically plausible ways. On challenging cMDP benchmarks, DALI achieves significant gains over context-unaware baselines, often surpassing context-aware baselines in extrapolation tasks, enabling zero-shot generalization to unseen contextual variations.



Paperid:4538
Authors:Mingyuan Xia, Chunxu Zhang, Zijian Zhang, Hao Miao, Qidong Liu, Yuanshao Zhu, Bo Yang
Abstract:
Temporal non-stationarity, the phenomenon that time series distributions change over time, poses fundamental challenges to reliable time series forecasting. Intuitively, the complex time series can be decomposed into two factors, i.e., time-invariant and time-varying components, which indicate static and dynamic patterns, respectively. Nonetheless, existing methods often conflate the time-varying and time-invariant components, and jointly learn the combined long-term patterns and short-term fluctuations, leading to suboptimal performance facing distribution shifts. To address this issue, we initiatively propose a lightweight static-dynamic decomposition framework, TimeEmb, for time series forecasting. TimeEmb innovatively separates time series into two complementary components: (1) time-invariant component, captured by a novel global embedding module that learns persistent representations across time series, and (2) time-varying component, processed by an efficient frequency-domain filtering mechanism inspired by full-spectrum analysis in signal processing. Experiments on real-world datasets demonstrate that TimeEmb outperforms state-of-the-art baselines and requires fewer computational resources. We conduct comprehensive quantitative and qualitative analyses to verify the efficacy of static-dynamic disentanglement. This lightweight framework can also improve existing time-series forecasting methods with simple integration. To ease reproducibility, our code is available at https://anonymous.4open.science/r/TimeEmb-9CC3.



Paperid:4539
Authors:Jingyuan Zhang, Xin Wang, Lei Yu, Zhirong Huang, Li Yang, Fengjun Zhang
Abstract:
Transformers have been widely regarded as a promising direction for breaking through the performance bottlenecks of Graph Neural Networks (GNNs), primarily due to their global receptive fields. However, a recent empirical study suggests that tuned classical GNNs can match or even outperform state-of-the-art Graph Transformers (GTs) on standard node classification benchmarks. Motivated by this fact, we deconstruct several representative GTs to examine how global attention components influence node representations. We find that the global attention module does not provide significant performance gains and may even exacerbate test error oscillations. Consequently, we consider that the Transformer is barely able to learn connectivity patterns that meaningfully complement the original graph topology. Interestingly, we further observe that mitigating such oscillations enables the Transformer to improve generalization in GNNs. In a nutshell, we reinterpret the Transformer through the lens of graph spectrum and reformulate it as a global-aware graph filter with band-pass characteristics and linear complexity. This unique perspective introduces multi-channel filtering constraints that effectively suppress test error oscillations. Extensive experiments (17 homophilous, heterophilous graphs) provide comprehensive empirical evidence for our perspective. This work clarifies the role of Transformers in GNNs and suggests that advancing modern GNN research may still require a return to the graph itself.



Paperid:4540
Authors:Qian Tang, Yuwen Gu, Boxiang Wang
Abstract:
Binary classification with heavily imbalanced class distributions is a common and fundamental task, where standard machine learning methods often struggle to provide reliable predictive performance. Although numerous approaches have been proposed to address this issue, classification in low-sample-size and high-dimensional settings remains particularly challenging. The abundance of noisy features in high-dimensional data limits the effectiveness of classical methods due to overfitting, and the minority class is even difficult to detect because of its severe underrepresentation with low sample size. To address this challenge, we introduce Quantile-based Discriminant Analysis (QuanDA), which builds upon a novel connection with quantile regression and naturally accounts for class imbalance through the choice of appropriate quantile levels. We provide comprehensive theoretical analysis to validate QuanDA in ultra-high dimensional settings. Through extensive simulation studies and applications to high-dimensional benchmark datasets, we demonstrate that QuanDA consistently outperforms existing classification methods for imbalanced data, including cost-sensitive large-margin classifiers, random forests, and SMOTE.



Paperid:4541
Authors:Di Jin, Yuxiang Zhang, Bingdao Feng, Xiaobao Wang, Dongxiao He, Zhen Wang
Abstract:
Graph Neural Networks (GNNs) are vulnerable to backdoor attacks. Existing defenses primarily rely on detecting structural anomalies, distributional outliers, or perturbation-induced prediction instability, which struggle to handle the more subtle, feature-based attacks that do not introduce obvious topological changes. Our empirical analysis reveals that both structure-based and feature-based attacks not only cause early loss convergence of target nodes but also induce a class-coherent loss drift, where this early convergence gradually spreads to nearby clean nodes, leading to significant distribution overlap. To address this issue, we propose \emph{LoSplit}, a training-time defense framework that leverages this early-stage loss drift to accurately separate target nodes. Our method dynamically selects epochs with maximal loss divergence, clusters target nodes via Gaussian Mixture Models (GMM), and applies a Decoupling-Unlearning strategy to break the association between target nodes and malicious labels. Extensive experiments on multiple real-world datasets demonstrate the effectiveness of this approach, significantly reducing attack success rates while maintaining high clean accuracy across diverse backdoor attack strategies.



Paperid:4542
Authors:Fan Wang, Pengtao Shao, Yiming Zhang, Bo Yu, Shaoshan Liu, Ning Ding, Yang Cao, Yu Kang, Haifeng Wang
Abstract:
In-Context Reinforcement Learning (ICRL) enables agents to learn automatically and on-the-fly from their interactive experiences. However, a major challenge in scaling up ICRL is the lack of scalable task collections. To address this, we propose the procedurally generated tabular Markov Decision Processes, named AnyMDP. Through a carefully designed randomization process, AnyMDP is capable of generating high-quality tasks on a large scale while maintaining relatively low structural biases. To facilitate efficient meta-training at scale, we further introduce step-wise supervision and induce prior information in the ICRL framework. Our results demonstrate that, with a sufficiently large scale of AnyMDP tasks, the proposed model can generalize to tasks that were not considered in the training set. The scalable task set provided by AnyMDP also enables a more thorough empirical investigation of the relationship between data distribution and ICRL performance. We further show that the generalization of ICRL potentially comes at the cost of increased task diversity and longer adaptation periods. This finding carries critical implications for scaling robust ICRL capabilities, highlighting the necessity of diverse and extensive task design, and prioritizing asymptotic performance over few-shot adaptation.



Paperid:4543
Authors:Duong Nguyen, Nghia Hoang, Thanh Huynh, Quoc Viet Hung Nguyen, Phi Le Nguyen
Abstract:
Multimodal federated learning in real-world settings often encounters incomplete and heterogeneous data across clients, resulting in misaligned local feature representations that limit the effectiveness of direct model aggregation. Unlike prior work that assumes either differing modality sets without missing input features or a shared modality set with missing features across clients, we consider a more general and realistic setting where each client observes a different subset of modalities and may also have missing input features within each modality. To address the resulting misalignment in learned representations, we propose a new federated learning framework featuring locally adaptive representations based on learnable client-side embedding controls that encode each client’s data-missing patterns. These embeddings serve as reconfiguration signals that align the globally aggregated representation with each client's local context, enabling more effective use of shared information. Furthermore, the embedding controls can be algorithmically aggregated across clients with similar data-missing patterns to enhance the robustness of reconfiguration signals in adapting the global representation. Empirical results on multiple federated multimodal benchmarks with diverse data-missing patterns across clients demonstrate the efficacy of the proposed method, achieving up to 36.45\% performance improvement under severe data incompleteness. The method is also supported by theoretical analysis with an explicit performance bound that matches our empirical observations.



Paperid:4544
Authors:Thanh-Dat Truong, Huu-Thien Tran, Tran Son, Bhiksha Raj, Khoa Luu
Abstract:
Large multimodal models (LMMs) have gained impressive performance due to their outstanding capability in various understanding tasks. However, these models still suffer from some fundamental limitations related to robustness and generalization due to the alignment and correlation between visual and textual features. In this paper, we introduce a simple but efficient learning mechanism for improving the robust alignment between visual and textual modalities by solving shuffling problems. In particular, the proposed approach can improve reasoning capability, visual understanding, and cross-modality alignment by introducing two new tasks: reconstructing the image order and the text order into the LMM's pre-training and fine-tuning phases. In addition, we propose a new directed-token approach to capture visual and textual knowledge, enabling the capability to reconstruct the correct order of visual inputs. Then, we introduce a new Image-to-Response Guided loss to further improve the visual understanding of the LMM in its responses. The proposed approach consistently achieves state-of-the-art (SoTA) performance compared with prior LMMs on academic task-oriented and instruction-following LMM benchmarks.



Paperid:4545
Authors:Li Ju, Max Andersson, Stina Fredriksson, Edward Glöckner, Andreas Hellander, Ekta Vats, Prashant Singh
Abstract:
Vision-language models (VLMs) as foundation models have significantly enhanced performance across a wide range of visual and textual tasks, without requiring large-scale training from scratch for downstream tasks. However, these deterministic VLMs fail to capture the inherent ambiguity and uncertainty in natural language and visual data. Recent probabilistic post-hoc adaptation methods address this by mapping deterministic embeddings onto probability distributions; however, existing approaches do not account for the asymmetric uncertainty between modalities, and the constraint that meaningful deterministic embeddings reside on a unit hypersphere, potentially leading to suboptimal performance. In this paper, we address the asymmetric uncertainty structure inherent in textual and visual data, and propose AsymVLM to build probabilistic embeddings from pre-trained VLMs on the unit hypersphere, enabling uncertainty quantification. We validate the effectiveness of the probabilistic embeddings on established benchmarks, and present comprehensive ablation studies demonstrating the inherent nature of asymmetry in the uncertainty structure of textual and visual data.



Paperid:4546
Authors:Linfeng Tang, Yeda Wang, Zhanchuan Cai, Junjun Jiang, Jiayi Ma
Abstract:
Current image fusion methods struggle with real-world composite degradations and lack the flexibility to accommodate user-specific needs. To address this, we propose ControlFusion, a controllable fusion network guided by language-vision prompts that adaptively mitigates composite degradations. On the one hand, we construct a degraded imaging model based on physical mechanisms, such as the Retinex theory and atmospheric scattering principle, to simulate composite degradations and provide a data foundation for addressing realistic degradations. On the other hand, we devise a prompt-modulated restoration and fusion network that dynamically enhances features according to degradation prompts, enabling adaptability to varying degradation levels. To support user-specific preferences in visual quality, a text encoder is incorporated to embed user-defined degradation types and levels as degradation prompts. Moreover, a spatial-frequency collaborative visual adapter is designed to autonomously perceive degradations from source images, thereby reducing complete reliance on user instructions. Extensive experiments demonstrate that ControlFusion outperforms SOTA fusion methods in fusion quality and degradation handling, particularly under real-world and compound degradations.



Paperid:4547
Authors:Haotian Chi, Zeyu Feng, Yueming LYU, Chengqi Zheng, Linbo Luo, Yew Soon Ong, Ivor Tsang, Hechang Chen, Yi Chang, Haiyan Yin
Abstract:
Long-horizon planning in robotic manipulation tasks requires translating underspecified, symbolic goals into executable control programs satisfying spatial, temporal, and physical constraints. However, language model-based planners often struggle with long-horizon task decomposition, robust constraint satisfaction, and adaptive failure recovery. We introduce InstructFlow, a multi-agent framework that establishes a symbolic, feedback-driven flow of information for code generation in robotic manipulation tasks. InstructFlow employs a InstructFlow Planner to construct and traverse a hierarchical instruction graph that decomposes goals into semantically meaningful subtasks, while a Code Generator generates executable code snippets conditioned on this graph. Crucially, when execution failures occur, a Constraint Generator analyzes feedback and induces symbolic constraints, which are propagated back into the instruction graph to guide targeted code refinement without regenerating from scratch. This dynamic, graph-guided flow enables structured, interpretable, and failure-resilient planning, significantly improving task success rates and robustness across diverse manipulation benchmarks, especially in constraint-sensitive and long-horizon scenarios.



Paperid:4548
Authors:Dongyue Lu, Lingdong Kong, Gim Hee Lee, Camille Simon Chane, Wei Tsang Ooi
Abstract:
Event cameras offer unparalleled advantages for real-time perception in dynamic environments, thanks to the microsecond-level temporal resolution and asynchronous operation. Existing event detectors, however, are limited by fixed-frequency paradigms and fail to fully exploit the high-temporal resolution and adaptability of event data. To address these limitations, we propose FlexEvent, a novel framework that enables detection at varying frequencies. Our approach consists of two key components: FlexFuse, an adaptive event-frame fusion module that integrates high-frequency event data with rich semantic information from RGB frames, and FlexTune, a frequency-adaptive fine-tuning mechanism that generates frequency-adjusted labels to enhance model generalization across varying operational frequencies. This combination allows our method to detect objects with high accuracy in both fast-moving and static scenarios, while adapting to dynamic environments. Extensive experiments on large-scale event camera datasets demonstrate that our approach surpasses state-of-the-art methods, achieving significant improvements in both standard and high-frequency settings. Notably, our method maintains robust performance when scaling from 20 Hz to 90 Hz and delivers accurate detection up to 180 Hz, proving its effectiveness in extreme conditions. Our framework sets a new benchmark for event-based object detection and paves the way for more adaptable, real-time vision systems. Code will be publicly available.



Paperid:4549
Authors:Yifei Gao, Yong Chen, Chen Zhang
Abstract:
Functional data play a pivotal role across science and engineering, yet their infinite-dimensional nature makes representation learning challenging. Conventional statistical models depend on pre-chosen basis expansions or kernels, limiting the flexibility of data-driven discovery, while many deep-learning pipelines treat functions as fixed-grid vectors, ignoring inherent continuity. In this paper, we introduce Functional Attention with a Mixture-of-Experts (FAME), an end-to-end, fully data-driven framework for function-on-function regression. FAME forms continuous attention by coupling a bidirectional neural controlled differential equation with MoE-driven vector fields to capture intra-functional continuity, and further fuses change to inter-functional dependencies via multi-head cross attention. Extensive experiments on synthetic and real-world functional-regression benchmarks show that FAME achieves state-of-the-art accuracy, strong robustness to arbitrarily sampled discrete observations of functions.



Paperid:4550
Authors:Yuanhao Cai, HE Zhang, Xi Chen, Jinbo Xing, Yiwei Hu, Yuqian Zhou, Kai Zhang, Zhifei Zhang, Soo Ye Kim, Tianyu Wang, Yulun Zhang, Xiaokang Yang, Zhe Lin, Alan Yuille
Abstract:
Existing feedforward subject-driven video customization methods mainly study single-subject scenarios due to the difficulty of constructing multi-subject training data pairs. Another challenging problem that how to use the signals such as depth, mask, camera, and text prompts to control and edit the subject in the customized video is still less explored. In this paper, we first propose a data construction pipeline, VideoCus-Factory, to produce training data pairs for multi-subject customization from raw videos without labels and control signals such as depth-to-video and mask-to-video pairs. Based on our constructed data, we develop an Image-Video Transfer Mixed (IVTM) training with image editing data to enable instructive editing for the subject in the customized video. Then we propose a diffusion Transformer framework, OmniVCus, with two embedding mechanisms, Lottery Embedding (LE) and Temporally Aligned Embedding (TAE). LE enables inference with more subjects by using the training subjects to activate more frame embeddings. TAE encourages the generation process to extract guidance from temporally aligned control signals by assigning the same frame embeddings to the control and noise tokens. Experiments demonstrate that our method significantly surpasses state-of-the-art methods in both quantitative and qualitative evaluations.



Paperid:4551
Authors:Gerardo Duran-Martin, Leandro Sánchez-Betancourt, Alvaro Cartea, Kevin Murphy
Abstract:
We introduce scalable algorithms for online learning andgeneralized Bayesian inference of neural network parameters,designed for sequential decision making tasks.Our methods combine the strengths of frequentist and Bayesian filtering, which include fast low-rank updates via a block-diagonal approximation of the parameter error covariance, and a well-defined posterior predictive distribution that we use for decision making.More precisely, our main method updates a low-rank error covariance for the hidden layers parameters, and a full-rank errorcovariance for the final layer parameters.Although this characterizes an improper posterior,we show that the resulting posterior predictive distribution is well-defined.Our methods update all network parameters online, with no need for replay buffers or offline retraining.We show, empirically, that our methods achieve a competitive tradeoffbetween speed and accuracy on (non-stationary) contextual bandit problemsand Bayesian optimization problems.



Paperid:4552
Authors:Jiawei Zhou, Lei Chen
Abstract:
As retrieval-augmented generation (RAG) becomes widespread, the role of information retrieval shifts from retrieving information for user queries to retrieving contexts for LLMs, where relevance is difficult to define or annotate. In this paper, we empirically show that relevance learned in traditional IR settings often misaligns with the needs of RAG. To address this gap, we introduce OpenRAG, a RAG framework OPtimized ENd-to-end by tuning the retriever via online contrastive learning to capture complex relevance tailored for RAG. Extensive experiments across diverse tasks show that in RAG settings, OpenRAG improves performance by 5.2% over the original retriever and outperforms state-of-the-art retrievers by 4.9%. Our approach is both efficient and practical, requiring only 4 GPUs and completing training within a single day. Despite tuning only a 0.2B retriever, our approach consistently outperforms LLM-augmented retrieval on most tasks and even rivals the performance of fine-tuning an 8B LLM, making it a cost-effective alternative for improving RAG.



Paperid:4553
Authors:Fenil Doshi, Thomas Fel, Talia Konkle, George Alvarez
Abstract:
Humans are able to recognize objects based on both local texture cues and the configuration of object parts, yet contemporary vision models primarily harvest local texture cues, yielding brittle, non-compositional features. Work on shape-vs-texture bias has pitted shape and texture representations in opposition, measuring shape relative to texture, ignoring the possibility that models (and humans) can simultaneously rely on both types of cues, and obscuring the absolute quality of both types of representation. We therefore recast shape evaluation as a matter of absolute configural competence, operationalized by the Configural Shape Score (CSS), which (i) measures the ability to recognize both images in Object-Anagram pairs that preserve local texture while permuting global part arrangement to depict different object categories. Across 86 convolutional, transformer, and hybrid models, CSS (ii) uncovers a broad spectrum of configural sensitivity with fully self-supervised and language-aligned transformers -- exemplified by DINOv2, SigLIP2 and EVA-CLIP -- occupying the top end of the CSS spectrum. Mechanistic probes reveal that (iii) high-CSS networks depend on long-range interactions: radius-controlled attention masks abolish performance showing a distinctive U-shaped integration profile, and representational-similarity analyses expose a mid-depth transition from local to global coding. A BagNet control, whose receptive fields straddle patch seams, remains at chance (iv), ruling out any ``border-hacking'' strategies. Finally, (v) we show that configural shape score also predicts other shape-dependent evals (e.g., foreground bias, spectral and noise robustness). Overall, we propose that the path toward truly robust, generalizable, and human-like vision systems may not lie in forcing an artificial choice between shape and texture, but rather in architectural and learning frameworks that seamlessly integrate both local-texture and global configural shape.



Paperid:4554
Authors:Charvi Rastogi, Tian Huey Teh, Pushkar Mishra, Roma Patel, Ding Wang, Mark Díaz, Alicia Parrish, Aida Mostafazadeh Davani, Zoe Ashwood, Michela Paganini, Vinodkumar Prabhakaran, Verena Rieser, Lora Aroyo
Abstract:
Current text-to-image (T2I) models often fail to account for diverse human experiences, leading to misaligned systems. We advocate for pluralistic alignment, where an AI understands and is steerable towards diverse, and often conflicting, human values. Our work provides three core contributions to achieve this in T2I models. First, we introduce a novel dataset for Diverse Intersectional Visual Evaluation (DIVE) -- the first multimodal dataset for pluralistic alignment. It enables deep alignment to diverse safety perspectives through a large pool of demographically intersectional human raters who provided extensive feedback across 1000 prompts, with high replication, capturing nuanced safety perceptions. Second, we empirically confirm demographics as a crucial proxy for diverse viewpoints in this domain, revealing significant, context-dependent differences in harm perception that diverge from conventional evaluations. Finally, we discuss implications for building aligned T2I models, including efficient data collection strategies, LLM judgment capabilities, and model steerability towards diverse perspectives. This research offers foundational tools for more equitable and aligned T2I systems.Content Warning: The paper includes sensitive content that may be harmful.



Paperid:4555
Authors:Mingzhe Du, Anh Tuan Luu, Yue Liu, Yuhao QING, Dong HUANG, Xinyi He, Qian Liu, Zejun MA, See-Kiong Ng
Abstract:
Large Language Models (LLMs) generate functionally correct solutions but often fall short in code efficiency, a critical bottleneck for real-world deployment. In this paper, we introduce a novel test-time iterative optimization framework to address this, employing a closed-loop system where LLMs iteratively refine code based on empirical performance feedback from an execution sandbox. We explore three training strategies: Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Group Relative Policy Optimization~(GRPO). Experiments on our Venus dataset and the APPS benchmark show that SFT and DPO rapidly saturate in efficiency gains. In contrast, GRPO, using reinforcement learning (RL) with execution feedback, continuously optimizes code performance, significantly boosting both pass@1 (from 47% to 62%) and the likelihood of outperforming human submissions in efficiency (from 31% to 45%). Our work demonstrates effective test-time code efficiency improvement and critically reveals the power of RL in teaching LLMs to truly self-improve code efficiency.



Paperid:4556
Authors:Antonio Ribeiro, David Vävinggren, Dave Zachariah, Thomas Schön, Francis Bach
Abstract:
Adversarial training has emerged as a key technique to enhance model robustness against adversarial input perturbations. Many of the existing methods rely on computationally expensive min-max problems that limit their application in practice. We propose a novel formulation of adversarial training in reproducing kernel Hilbert spaces, shifting from input to feature-space perturbations. This reformulation enables a convex optimization problem that can be efficiently solved and provides a natural adaptive regularization to the noise level and properties of the underlying function. We establish theoretical conditions under which the relaxed formulation serves as a proper approximation of the original problem, prove convexity, and propose an efficient optimization algorithm based on iterative kernel ridge regression. We also provide generalization bounds that help to understand the properties of the method. Furthermore, we extend the formulation to multiple kernel learning. We evaluate our method on synthetic and real-world datasets.



Paperid:4557
Authors:Zhaorun Chen, Zichen Wen, Yichao Du, Yiyang Zhou, Chenhang Cui, Siwei Han, Jen Weng, Chaoqi Wang, Zhengwei Tong, Leria HUANG, Canyu Chen, Haoqin Tu, Qinghao Ye, Zhihong Zhu, Yuqing Zhang, Jiawei Zhou, Zhuokai Zhao, Rafael Rafailov, Chelsea Finn, Huaxiu Yao
Abstract:
While text-to-image models like GPT-4o-Image and FLUX are rapidly proliferating, they often encounter challenges such as hallucination, bias, and the production of unsafe, low-quality output. To effectively address these issues, it is crucial to align these models with desired behaviors based on feedback from a multimodal judge. Despite their significance, current multimodal judges frequently undergo inadequate evaluation of their capabilities and limitations, potentially leading to misalignment and unsafe fine-tuning outcomes. To address this issue, we introduce MJ-Bench, a novel benchmark which incorporates a comprehensive preference dataset to evaluate multimodal judges in providing feedback for image generation models across six key perspectives: alignment, safety, image quality, bias, composition, and visualization. Specifically, we evaluate a large variety of multimodal judges including smaller-sized CLIP-based scoring models, open-source VLMs, and close-source VLMs on each decomposed subcategory of our preference dataset. Experiments reveal that close-source VLMs generally provide better feedback, with GPT-4o outperforming other judges in average. Compared with open-source VLMs, smaller-sized scoring models can provide better feedback regarding text-image alignment and image quality, while VLMs provide more accurate feedback regarding safety and generation bias due to their stronger reasoning capabilities. Further studies in feedback scale reveal that VLM judges can generally provide more accurate and stable feedback in natural language than numerical scales. Notably, human evaluations on end-to-end and fine-tuned models using separate feedback from these multimodal judges provide similar conclusions, further confirming the effectiveness of MJ-Bench.



Paperid:4558
Authors:Huiyi Wang, Chun Kwang Tan, Balint Hodossy, Shirui Lyu, Pierre Schumacher, James Heald, Kai Biegun, Samo Hromadka, Maneesh Sahani, Gunwoo Park, Beomsoo Shin, JongHyeon Park, Seungbum Koo, Chenhui Zuo, Chengtian Ma, Yanan Sui, Nicklas Hansen, Stone Tao, Yuan Gao, Hao Su, Seungmoon Song, Letizia Gionfrida, Massimo Sartori, Guillaume Durandau, Vikash Kumar, Vittorio Caggiano
Abstract:
Recent advancements in bionic prosthetic technology offer transformative opportunities to restore mobility and functionality for individuals with missing limbs. Users of bionic limbs, or bionic humans, learn to seamlessly integrate prosthetic extensions into their motor repertoire, regaining critical motor abilities. The remarkable movement generalization and environmental adaptability demonstrated by these individuals highlight motor intelligence capabilities unmatched by current artificial intelligence systems. Addressing these limitations, MyoChallenge '24 at NeurIPS 2024 established a benchmark for human-robot coordination with an emphasis on joint control of both biological and mechanical limbs. The competition featured two distinct tracks: a manipulation task utilizing the myoMPL model, integrating a virtual biological arm and the Modular Prosthetic Limb (MPL) for a passover task; and a locomotion task using the novel myoOSL model, combining a bilateral virtual biological leg with a trans-femoral amputation and the Open Source Leg (OSL) to navigate varied terrains. Marking the third iteration of the MyoChallenge, the event attracted over 50 teams with more than 290 submissions all around the globe, with diverse participants ranging from independent researchers to high school students. The competition facilitated the development of several state-of-the-art control algorithms for bionic musculoskeletal systems, leveraging techniques such as imitation learning, muscle synergy, and model-based reinforcement learning that significantly surpassed our proposed baseline performance by a factor of 10. By providing the open-source simulation framework of MyoSuite, standardized tasks, and physiologically realistic models, MyoChallenge serves as a reproducible testbed and benchmark for bridging ML and biomechanics. The competition website is featured here: https://sites.google.com/view/myosuite/myochallenge/myochallenge-2024.



Paperid:4559
Authors:Miran Özdogan, Gilad Landau, Gereon Elvers, Dulhan Jayalath, Pratik Somaiya, Francesco Mantegna, Mark Woolrich, Oiwi Parker Jones
Abstract:
Abstract:LibriBrain represents the largest single-subject MEG dataset to date for speech decoding, with over 50 hours of recordings---5$\times$ larger than the next comparable dataset and 50$\times$ larger than most. This unprecedented `depth' of within-subject data enables exploration of neural representations at a scale previously unavailable with non-invasive methods. LibriBrain comprises high-quality MEG recordings together with detailed annotations from a single participant listening to naturalistic spoken English, covering nearly the full Sherlock Holmes canon. Designed to support advances in neural decoding, LibriBrain comes with a Python library for streamlined integration with deep learning frameworks, standard data splits for reproducibility, and baseline results for three foundational decoding tasks: speech detection, phoneme classification, and word classification. Baseline experiments demonstrate that increasing training data yields substantial improvements in decoding performance, highlighting the value of scaling up deep, within-subject datasets. By releasing this dataset, we aim to empower the research community to advance speech decoding methodologies and accelerate the development of safe, effective clinical brain-computer interfaces.



Paperid:4560
Authors:Jiaqi Wang, Zhengyu Ma, Xiongri Shen, Chenlin Zhou, Leilei Zhao, Han Zhang, Yi Zhong, Siqi Cai, Zhenxi Song, Zhiguo Zhang
Abstract:
Abstract:Auditory attention detection (AAD) aims to decode listeners' focus in complex auditory environments from electroencephalography (EEG) recordings, which is crucial for developing neuro-steered hearing devices. Despite recent advancements, EEG-based AAD remains hindered by the absence of synergistic frameworks that can fully leverage complementary EEG features under energy-efficiency constraints. We propose S$^2$M-Former, a novel ***s***piking ***s***ymmetric ***m***ixing framework to address this limitation through two key innovations: i) Presenting a spike-driven symmetric architecture composed of parallel spatial and frequency branches with mirrored modular design, leveraging biologically plausible token-channel mixers to enhance complementary learning across branches; ii) Introducing lightweight 1D token sequences to replace conventional 3D operations, reducing parameters by 14.7$\times$. The brain-inspired spiking architecture further reduces power consumption, achieving a 5.8$\times$ energy reduction compared to recent ANN methods, while also surpassing existing SNN baselines in terms of parameter efficiency and performance. Comprehensive experiments on three AAD benchmarks (KUL, DTU and AV-GC-AAD) across three settings (within-trial, cross-trial and cross-subject) demonstrate that S$^2$M-Former achieves comparable state-of-the-art (SOTA) decoding accuracy, making it a promising low-power, high-performance solution for AAD tasks. Code and models will be made publicly available.



Paperid:4561
Authors:Jingqiu Ding, Yiding Hua, Lucas Slot, David Steurer
Abstract:
Abstract:We investigate implications of the (extended) low-degree conjecture (recently formalized in [moitra et al2023]) in the context of the symmetric stochastic block model. Assuming the conjecture holds, we establish that no polynomial-time algorithm can weakly recover community labels below the Kesten-Stigum (KS) threshold. In particular, we rule out polynomial-time estimators that, with constant probability, achieve $n^{-0.49}$ correlation with the true communities. Whereas, above the KS threshold, polynomial-time algorithms are known to achieve constant correlation with the true communities with high probability [massoulie et al 2014,abbe et al 2015]. To our knowledge, we provide the first rigorous evidence for such sharp transition in recovery rate for polynomial-time algorithms at the KS threshold. Notably, under a stronger version of the low-degree conjecture, our lower bound remains valid even when the number of blocks diverges. Furthermore, our results provide evidence of a computational-to-statistical gap in learning the parameters of stochastic block models.In contrast, prior work either (i) rules out polynomial-time algorithms with $1 - o(1)$ success probability [Hopkins 18, bandeira et al 2021] under the low-degree conjecture, or (ii) degree-$\text{poly}(k)$ polynomials for learning the stochastic block model [Luo et al 2023].For this, we design a hypothesis test which succeeeds with constant probability under symmetric stochastic block model, and $1-o(1)$ probability under the distribution of \Erdos \Renyi random graphs.Our proof combines low-degree lower bounds from [Hopkins 18, bandeira et al 2021] with graph splitting and cross-validation techniques. In order to rule out general recovery algorithms, we employ the correlation preserving projection method developed in [Hopkins et al 17].



Paperid:4562
Authors:Barathi Subramanian, Rathinaraja Jeyaraj, Mitchell Peterson, Terry Guo, Nigam Shah, Curtis Langlotz, Andrew Ng, Jeanne Shen
Abstract:
Multi-class tissue-type classification of colorectal cancer (CRC) histopathologic images is a significant step in the development of downstream machine learning models for diagnosis and treatment planning. However, publicly available CRC datasets used to build tissue classifiers often suffer from insufficient morphologic diversity, class imbalance, and low-quality image tiles, limiting downstream model performance and generalizability. To address this research gap, we introduce STARC-9 (STAnford coloRectal Cancer), a large-scale dataset for multi-class tissue classification. STARC-9 comprises 630,000 histopathologic image tiles uniformly sampled across nine clinically relevant tissue classes (each represented by ~70,000 tiles), systematically extracted from hematoxylin & eosin-stained whole-slide images (WSI) from 200 CRC patients at the Stanford University School of Medicine. To construct STARC-9, we propose a novel framework, DeepCluster++, consisting of three primary steps to ensure diversity within each tissue class. First, an encoder from an auto-encoder trained specifically on histopathologic images is used to extract feature vectors from all tiles within a given input WSI. Next, K-means clustering groups morphologically similar tiles, followed by an equal-frequency binning method to sample diverse patterns within each tissue class. Finally, the selected tiles are verified by expert gastrointestinal pathologists to ensure classification accuracy. This semi-automated approach significantly reduces the manual effort required for dataset curation while producing high-quality training examples. To validate the utility of STARC-9, we benchmarked baseline convolutional neural networks, transformers, and pathology-specific foundation models on downstream multi-class CRC tissue classification and segmentation tasks when trained on STARC-9 versus publicly available datasets, demonstrating superior generalizability of models trained on STARC-9. Although we demonstrate the utility of DeepCluster++ on CRC as a pilot use-case, it is a flexible framework that can be used for constructing high-quality datasets from large WSI repositories across a wide range of cancer and non-cancer applications.



Paperid:4563
Authors:Ke Niu, Zhuofan Chen, Haiyang Yu, Yuwen Chen, Teng Fu, Mengyang Zhao, Bin Li, Xiangyang Xue
Abstract:
Computer-Aided Design (CAD) plays a pivotal role in industrial manufacturing. Orthographic projection reasoning underpins the entire CAD workflow, encompassing design, manufacturing, and simulation. However, prevailing deep‐learning approaches employ standard 3D reconstruction pipelines as an alternative, which often introduce imprecise dimensions and limit the parametric editability required for CAD workflows. Many researchers adopt the vision–language models (VLMs) method, particularly supervised fine-tuning (SFT), to tackle CAD-related challenges. SFT shows promise but often devolves into pattern memorization, yielding poor out‐of‐distribution performance on complex reasoning tasks. To address these gaps, we introduce CReFT-CAD, a two-stage fine-tuning paradigm that first employs a curriculum‐driven reinforcement learning stage with difficulty‐aware rewards to build reasoning ability steadily, and then applies supervised post-tuning to hone instruction following and semantic extraction. Complementing this, we release TriView2CAD, the first large-scale, open-source benchmark for orthographic projection reasoning, comprising 200,000 synthetic and 3,000 real-world orthographic projections with precise dimension annotations and six interoperable data modalities. We benchmark leading VLMs on orthographic projection reasoning and demonstrate that CReFT-CAD substantially improves reasoning accuracy and out-of-distribution generalizability in real-world scenarios, offering valuable insights for advancing CAD reasoning research.



Paperid:4564
Authors:Xi Zhang, Yanyi Li, Yisi Luo, Qi Xie, Deyu Meng
Abstract:
Online tensor decompositions are powerful and proven techniques that address the challenges in processing high-velocity streaming tensor data, such as traffic flow and weather system. The main aim of this work is to propose a novel online functional tensor decomposition (OFTD) framework, which represents a spatial-temporal continuous function using the CP tensor decomposition parameterized by coordinate-based implicit neural representations (INRs). The INRs allow for natural characterization of continually expanded streaming data by simply adding new coordinates into the network. Particularly, our method transforms the classical online tensor decomposition algorithm into a more dynamic continual learning paradigm of updating the INR weights to fit the new data without forgetting the previous tensor knowledge. To this end, we introduce a long-tail memory replay method that adapts to the local continuity property of INR. Extensive experiments for streaming tensor completion using traffic, weather, user-item, and video data verify the effectiveness of the OFTD approach for streaming data analysis. This endeavor serves as a pivotal inspiration for future research to connect classical online tensor tools with continual learning paradigms to better explore knowledge underlying streaming tensor data.



Paperid:4565
Authors:Xinyue Zhu, Binghao Huang, Yunzhu Li
Abstract:
Handheld grippers are increasingly used to collect human demonstrations due to their ease of deployment and versatility. However, most existing designs lack tactile sensing, despite the critical role of tactile feedback in precise manipulation. We present a portable, lightweight gripper with integrated tactile sensors that enables synchronized collection of visual and tactile data in diverse, real-world, and in-the-wild settings. Building on this hardware, we propose a cross-modal representation learning framework that integrates visual and tactile signals while preserving their distinct characteristics. The learned representations are interpretable and consistently emphasize contact regions during physical interactions. When used for downstream manipulation tasks, these representations enable more efficient and effective policy learning, supporting precise robotic manipulation based on multimodal feedback. We validate our approach on fine-grained tasks such as test tube insertion and pipette-based fluid transfer, demonstrating improved accuracy and robustness under external disturbances. Our project page is available at https://touchinthewild.github.io/ .



Paperid:4566
Authors:Daniel Cao, August Chen, Karthik Sridharan, Benjamin Tang
Abstract:
We study the optimization of non-convex functions that are not necessarily smooth (gradient and/or Hessian are Lipschitz) using first order methods. Smoothness is a restrictive assumption in machine learning in both theory and practice, motivating significant recent work on finding first order stationary points of functions satisfying generalizations of smoothness with first order methods. We develop a novel framework that lets us systematically study the convergence of a large class of first-order optimization algorithms (which we call decrease procedures) under generalizations of smoothness. We instantiate our framework to analyze the convergence of first order optimization algorithms to first andsecondorder stationary points under generalizations of smoothness. As a consequence, we establish the first convergence guarantees for first order methods to second order stationary points under generalizations of smoothness. We demonstrate that several canonical examples fall under our framework, and highlight practical implications.



Paperid:4567
Authors:Quentin Bertrand, Anne Gagneux, Mathurin Massias, Rémi Emonet
Abstract:
Modern deep generative models can now produce high-quality synthetic samples that are often indistinguishable from real training data. A growing body of research aims to understand why recent methods - such as diffusion and flow matching techniques - generalize so effectively. Among the proposed explanations are the inductive biases of deep learning architectures and the stochastic nature of the conditional flow matching loss.In this work, we rule out the latter -the noisy nature of the loss- as a primary contributor to generalization in flow matching. First, we empirically show that in high-dimensional settings, the stochastic and closed-form versions of the flow matching loss yield nearly equivalent losses. Then, using state-of-the-art flow matching models on standard image datasets, we demonstrate that both variants achieve comparable statistical performance, with the surprising observation that using the closed-form can even improve performance.



Paperid:4568
Authors:Kaiyuan Chen, Shuangyu Xie, Zehan Ma, Pannag R Sanketi, Ken Goldberg
Abstract:
Vision-Language Models (VLMs) acquire real-world knowledge and general reasoning ability through internet-scale image-text corpora. They have potential to augment robotic systems with scene understanding and task planning, and to assist visuomotor policies trained on robot trajectory data. We explore the reverse paradigm — using rich, real, multi-modal robot trajectory data to enhance and evaluate VLMs. In this paper, we present Robo2VLM, a Visual Question Answering (VQA) dataset generation framework for VLMs. Given a human tele-operated robot trajectory, Robo2VLM derives ground-truth from non-visual and non-descriptive sensory modalities, such as end-effector pose, gripper aperture, and force sensing. Based on these modalities, it segments the robot trajectory into a sequence of manipulation phases. At each phase, Robo2VLM uses scene and interaction understanding to identify 3D properties of the robot, task goal, and the target object. The properties are used to generate representative VQA queries – images with textural multiple-choice questions – based on spatial, goal-conditioned, and interaction reasoning question templates. We create Robo2VLM-1, a large-scale in-the-wild dataset with 684,710 questions covering 463 distinct scenes and 3,396 robotic manipulation tasks from 176k real robot trajectories. Results suggest that Robo2VLM-1 can benchmark and improve VLM capabilities in spatial and interaction reasoning.



Paperid:4569
Authors:Lu Zhang, Jiazuo Yu, Haomiao Xiong, Ping Hu, Yunzhi Zhuge, Huchuan Lu, You He
Abstract:
Multi-modal Large Language Models (MLLMs) have shown remarkable capabilities across a wide range of vision-language tasks. However, due to the restricted input resolutions, MLLMs face significant challenges in precisely understanding and localizing visual details in high-resolution images---particularly when dealing with extra-small objects embedded in cluttered contexts. To address this issue, we propose FineRS, a two-stage MLLM-based reinforcement learning framework for jointly reasoning and segmenting extremely small objects within high-resolution scenes. FineRS adopts a coarse-to-fine pipeline comprising Global Semantic Exploration (GSE) and Localized Perceptual Refinement (LPR). Specifically, GSE performs instruction-guided reasoning to generate a textural response and a coarse target region, while LPR refines this region to produce an accurate bounding box and segmentation mask. To couple the two stages, we introduce a locate-informed retrospective reward, where LPR's outputs are used to optimize GSE for more robust coarse region exploration. Additionally, we present FineRS-4k, a new dataset for evaluating MLLMs on attribute-level reasoning and pixel-level segmentation on subtle, small-scale targets in complex high-resolution scenes. Experimental results on FineRS-4k and public datasets demonstrate that our method consistently outperforms state-of-the-art MLLM-based approaches on both instruction-guided segmentation and visual reasoning tasks.



Paperid:4570
Authors:Dong Zhemeng, Junjun Jiang, Youyu Chen, Jiaxin Zhang, Kui Jiang, Xianming Liu
Abstract:
The essence of 3D Gaussian Splatting (3DGS) training is to smartly allocate Gaussian primitives, expressing complex regions with more primitives and vice versa. Prior researches typically mark out under-reconstructed regions in a rendering-loss-driven manner. However, such a loss-driven strategy is often dominated by low-frequency regions, which leads to insufficient modeling of high-frequency details in texture-rich regions. As a result, it yields a suboptimal spatial allocation of Gaussian primitives.This inspires us to excavate the loss-agnostic visual prior in training views to identify complex regions that need more primitives to model. Based on this insight, we propose Complexity-Density Consistent Gaussian Splatting (CDC-GS), which allocates primitives based on the consistency between visual complexity of training views and the density of primitives. Specifically, primitives involved in rendering high visual complexity areas are categorized as modeling high complexity regions, where we leverage the high frequency wavelet components of training views to measure the visual complexity. And the density of a primitive is computed with the inverse of geometric mean of its distance to the neighboring primitives. Guided by the positive correlation between primitive complexity and density, we determine primitives to be densified as well as pruned. Extensive experiments demonstrate that our CDC-GS surpasses the baseline methods in rendering quality by a large margin using the same amount of Gaussians. And we provide insightful analysis to reveal that our method serves perpendicularly to rendering loss in guiding Gaussian primitive allocation.



Paperid:4571
Authors:hongyong han, Wei Wang, Gaowei Zhang, Mingjie Li, Yi Wang
Abstract:
Coral reefs are vital yet vulnerable ecosystems that require continuous monitoring to support conservation. While coral reef images provide essential information in coral monitoring, interpreting such images remains challenging due to the need for domain expertise. Visual Question Answering (VQA), powered by Large Vision-Language Models (LVLMs), has great potential in user-friendly interaction with coral reef images. However, applying VQA to coral imagery demands a dedicated dataset that addresses two key challenges: domain-specific annotations and multidimensional questions. In this work, we introduce CoralVQA, the first large-scale VQA dataset for coral reef analysis. It contains 12,805 real-world coral images from 67 coral genera collected from 3 oceans, along with 277,653 question-answer pairs that comprehensively assess ecological and health-related conditions. To construct this dataset, we develop a semi-automatic data construction pipeline in collaboration with marine biologists to ensure both scalability and professional-grade data quality. CoralVQA presents novel challenges and provides a comprehensive benchmark for studying vision-language reasoning in the context of coral reef images. By evaluating several state-of-the-art LVLMs, we reveal key limitations and opportunities. These insights form a foundation for future LVLM development, with a particular emphasis on supporting coral conservation efforts.



Paperid:4572
Authors:Saeed Amizadeh, Sara Abdali, Yinheng Li, Kazuhito Koishida
Abstract:
Transformers and their attention mechanism have been revolutionary in the field of Machine Learning. While originally proposed for the language data, they quickly found their way to the image, video, graph, etc. data modalities with various signal geometries. Despite this versatility, generalizing the attention mechanism to scenarios where data is presented at different scales from potentially different modalities is not straightforward. The attempts to incorporate hierarchy and multi-modality within transformers are largely based on ad hoc heuristics, which are not seamlessly generalizable to similar problems with potentially different structures. To address this problem, in this paper, we take a fundamentally different approach: we first propose a mathematical construct to represent multi-modal, multi-scale data. We then mathematically derive the neural attention mechanics for the proposed construct from the first principle of entropy minimization. We show that the derived formulation is optimal in the sense of being the closest to the standard Softmax attention while incorporating the inductive biases originating from the hierarchical/geometric information of the problem. We further propose an efficient algorithm based on dynamic programming to compute our derived attention mechanism. By incorporating it within transformers, we show that the proposed hierarchical attention mechanism not only can be employed to train transformer models in hierarchical/multi-modal settings from scratch, but it can also be used to inject hierarchical information into classical, pre-trained transformer models post training, resulting in more efficient models in zero-shot manner.



Paperid:4573
Authors:Chengliang Liu, Que Yuanxi, Qihao Xu, Yabo Liu, Jie Wen, Jinghua Wang, Xiaoling Luo
Abstract:
Alzheimer's Disease (AD) poses a significant health threat to the aging population, underscoring the critical need for early diagnosis to delay disease progression and improve patient quality of life. Recent advances in heterogeneous multimodal artificial intelligence (AI) have facilitated comprehensive joint diagnosis, yet practical clinical scenarios frequently encounter incomplete modalities due to factors like high acquisition costs or radiation risks. Moreover, traditional convolution-based architecture face inherent limitations in capturing long-range dependencies and handling heterogeneous medical data efficiently. To address these challenges, in our proposed heterogeneous multimodal diagnostic framework (HAD), we develop a multi-view Hilbert curve-based Mamba block and a hierarchical spatial feature extraction module to simultaneously capture local spatial features and global dependencies, effectively alleviating spatial discontinuities introduced by voxel serialization. Furthermore, to balance semantic consistency and modal specificity, we build a unified mutual information learning objective in the heterogeneous multimodal embedding space, which maintains effective learning of modality-specific information to avoid modality collapse caused by model preference. Extensive experiments demonstrate that our HAD significantly outperforms state-of-the-art methods in various modality-missing scenarios, providing an efficient and reliable solution for early-stage AD diagnosis.



Paperid:4574
Authors:Zhihang Lin, Mingbao Lin, Yuan Xie, Rongrong Ji
Abstract:
Abstract:This paper introduces Completion Pruning Policy Optimization (CPPO) to accelerate the training of reasoning models based on Group Relative Policy Optimization (GRPO). GRPO, while effective, incurs high training costs due to the need to sample multiple completions for each question. Our experiment and theoretical analysis reveal that the number of completions impacts model accuracy yet increases training time multiplicatively, and not all completions contribute equally to policy training---their contribution depends on their relative advantage. To address these issues, we propose CPPO, which prunes completions with low absolute advantages, significantly reducing the number needed for gradient calculation and updates. Additionally, we introduce a dynamic completion allocation strategy to maximize GPU utilization by incorporating additional questions, further enhancing training efficiency. Experimental results demonstrate that CPPO achieves up to $8.32\times$ speedup on GSM8K and $3.51\times$ on Math while preserving or even enhancing the accuracy compared to the original GRPO. Code will be released.



Paperid:4575
Authors:Cuong Dao, The Hung Tran, Phi Le Nguyen, Truong Thao Nguyen, Nghia Hoang
Abstract:
This paper studies the black-box optimization task which aims to find the maxima of a black-box function using a static set of its observed input-output pairs. This is often achieved via learning and optimizing a surrogate function with that offline data. Alternatively, it can also be framed as an inverse modeling task that maps a desired performance to potential input candidates that achieve it. Both approaches are constrained by the limited amount of offline data. To mitigate this limitation, we introduce a new perspective that casts offline optimization as a distributional translation task. This is formulated as learning a probabilistic bridge transforming an implicit distribution of low-value inputs (i.e., offline data) into another distribution of high-value inputs (i.e., solution candidates). Such probabilistic bridge can be learned using low- and high-value inputs sampled from synthetic functions that resemble the target function. These synthetic functions are constructed as the mean posterior of multiple Gaussian processes fitted with different parameterizations on the offline data, alleviating the data bottleneck. The proposed approach is evaluated on an extensive benchmark comprising most recent methods, demonstrating significant improvement and establishing a new state-of-the-art performance.



Paperid:4576
Authors:Bowei He, Lihao Yin, Hui-Ling Zhen, Shuqi LIU, Han Wu, Xiaokun Zhang, Mingxuan Yuan, Chen Ma
Abstract:
Post-training compression has been a widely employed approach to scale down large language model (LLM) and facilitate efficient inference. In various proposed compression methods, including pruning and quantization, calibration data plays a vital role by informing the weight importance and activation dynamic ranges. However, how calibration data impacts the LLM capability after compression is less explored. Few of the existing works, though recognizing the significance of this study, only investigate the language modeling or commonsense reasoning performance degradation from limited angles, like the data sources or sample amounts. More systematic research is still needed to examine the impacts on different LLM capabilities in terms of compositional properties and domain correspondence of calibration data. In this work, we aim at bridging this gap and further analyze underlying influencing mechanisms from the activation pattern perspective. Especially, we explore the calibration data's impacts on high-level complex reasoning capabilities, like math problem solving and code generation. Delving into the underlying mechanism, we find that the representativeness and diversity in activation space more fundamentally determine the quality of calibration data. Finally, we propose a calibration data curation framework based on such observations and analysis, enhancing the performance of existing post-training compression methods on preserving critical LLM capabilities. Our code is provided inLink.



Paperid:4577
Authors:Zeqi Ye, Minshuo Chen
Abstract:
Imputation methods play a critical role in enhancing the quality of practical time-series data, which often suffer from pervasive missing values. Recently, diffusion-based generative imputation methods have demonstrated remarkable success compared to autoregressive and conventional statistical approaches. Despite their empirical success, the theoretical understanding of how well diffusion-based models capture complex spatial and temporal dependencies between the missing values and observed ones remains limited. Our work addresses this gap by investigating the statistical efficiency of conditional diffusion transformers for imputation and quantifying the uncertainty in missing values. Specifically, we derive statistical sample complexity bounds based on a novel approximation theory for conditional score functions using transformers, and, through this, construct tight confidence regions for missing values. Our findings also reveal that the efficiency and accuracy of imputation are significantly influenced by the missing patterns. Furthermore, we validate these theoretical insights through simulation and propose a mixed-masking training strategy to enhance the imputation performance.



Paperid:4578
Authors:Lijun Sheng, Jian Liang, Ran He, Zilei Wang, Tieniu Tan
Abstract:
Test-time adaptation (TTA) methods have gained significant attention for enhancing the performance of vision-language models (VLMs) such as CLIP during inference, without requiring additional labeled data. However, current TTA researches generally suffer from major limitations such as duplication of baseline results, limited evaluation metrics, inconsistent experimental settings, and insufficient analysis. These problems hinder fair comparisons between TTA methods and obscure their practical strengths and weaknesses. To address these challenges, we introduce TTA-VLM, a comprehensive benchmark for evaluating TTA methods on VLMs. Our benchmark implements 8 episodic TTA and 7 online TTA methods within a unified and reproducible framework, and evaluates them across 15 widely used datasets. Unlike prior studies focused solely on CLIP, we extend the evaluation to SigLIP—a model trained with a Sigmoid loss—and include training-time tuning methods such as CoOp, MaPLe, and TeCoA to assess generality. Beyond classification accuracy, TTA-VLM incorporates various evaluation metrics, including robustness, calibration, out-of-distribution detection, and stability, enabling a more holistic assessment of TTA methods. Through extensive experiments, we find that 1) existing TTA methods produce limited gains compared to the previous pioneering work; 2) current TTA methods exhibit poor collaboration with training-time fine-tuning methods; 3) accuracy gains frequently come at the cost of reduced model trustworthiness. We release TTA-VLM to provide fair comparison and comprehensive evaluation of TTA methods for VLMs, and we hope it encourages the community to develop more reliable and generalizable TTA strategies. The code is available in https://github.com/TomSheng21/tta-vlm.



Paperid:4579
Authors:Yinuo Ren, Haoxuan Chen, Yuchen Zhu, Wei Guo, Yongxin Chen, Grant Rotskoff, Molei Tao, Lexing Ying
Abstract:
Abstract:Discrete diffusion models have emerged as a powerful generative modeling framework for discrete data with successful applications spanning from text generation to image synthesis. However, their deployment faces challenges due to the high dimensionality of the state space, necessitating the development of efficient inference algorithms. Current inference approaches mainly fall into two categories: exact simulation and approximate methods such as $\tau$-leaping. While exact methods suffer from unpredictable inference time and redundant function evaluations, $\tau$-leaping is limited by its first-order accuracy. In this work, we advance the latter category by tailoring the first extension of high-order numerical inference schemes to discrete diffusion models, enabling larger step sizes while reducing error. We rigorously analyze the proposed schemes and establish the second-order accuracy of the $\theta$-trapezoidal method in KL divergence. Empirical evaluations on GPT-2 level text and ImageNet-level image generation tasks demonstrate that our method achieves superior sample quality compared to existing approaches under equivalent computational constraints.



Paperid:4580
Authors:Tianyi Yan, Wencheng Han, xia zhou, Xueyang Zhang, Kun Zhan, Cheng-Zhong Xu, Jianbing Shen
Abstract:
Synthetic data is crucial for advancing autonomous driving (AD) systems, yet current state-of-the-art video generation models, despite their visual realism, suffer from subtle geometric distortions that limit their utility for downstream perception tasks. We identify and quantify this critical issue, demonstrating a significant performance gap in 3D object detection when using synthetic versus real data.To address this, we introduce Reinforcement Learning with Geometric Feedback (RLGF), RLGF uniquely refines video diffusion models by incorporating rewards from specialized latent-space AD perception models. Its core components include an efficient Latent-Space Windowing Optimization technique for targeted feedback during diffusion, and a Hierarchical Geometric Reward (HGR) system providing multi-level rewards for point-line-plane alignment, and scene occupancy coherence. To quantify these distortions, we propose GeoScores. Applied to models like DiVE on nuScenes, RLGF substantially reduces geometric errors (e.g., VP error by 21\%, Depth error by 57\%) and dramatically improves 3D object detection mAP by 12.7\%, narrowing the gap to real-data performance. RLGF offers a plug-and-play solution for generating geometrically sound and reliable synthetic videos for AD development.



Paperid:4581
Authors:Lanyun Zhu, Deyi Ji, Tianrun Chen, Haiyang Wu, Shiqi Wang
Abstract:
The success of DeepSeek-R1 demonstrates the immense potential of using reinforcement learning (RL) to enhance LLMs' reasoning capabilities. This paper introduces Retrv-R1, the first R1-style MLLM specifically designed for multimodal universal retrieval, achieving higher performance by employing step-by-step reasoning to produce more accurate retrieval results. We find that directly applying the methods of DeepSeek-R1 to retrieval tasks is not feasible, mainly due to (1) the high computational cost caused by the large token consumption required for multiple candidates with reasoning processes, and (2) the instability and suboptimal results when directly applying RL to train for retrieval tasks. To address these issues, Retrv-R1 introduces an information compression module with a details inspection mechanism, which enhances computational efficiency by reducing the number of tokens while ensuring that critical information for challenging candidates is preserved. Additionally, a new training paradigm is proposed, including an activation stage using a retrieval-tailored synthetic CoT dataset for more effective optimization, followed by RL with a novel curriculum reward to improve both performance and efficiency. Incorporating these novel designs, Retrv-R1 achieves SOTA performance, high efficiency, and strong generalization ability, as demonstrated by extensive experiments across multiple benchmarks and tasks.



Paperid:4582
Authors:Long Li, Wanghu Chen, Wencheng Zhang, Shi Yuan, Hongle Guo
Abstract:
Unsupervised anomaly detection of multivariate time series remains challenging in complex nonstationary dynamics, due to the high false-positive rates and limited interpretability. We propose PhysDiff, combining physics-guided decomposition with diffusion-based reconstruction, to address these issues. The physics-guided signal decomposition is introduced to disentangle overlapping dynamics by isolating high frequency oscillations and low frequency trends, which can reduce interference and provide meaningful physical priors. The reconstruction through conditional diffusion modeling captures deviations from learned normal behavior, making anomalies more distinguishable. Notably, PhysDiff introduces an amplitude-sensitive permutation entropy criterion to adaptively determine the optimal decomposition depth, and automatically extract adaptive frequency components used as explicit physics-based constraints for the diffusion process. Furthermore, the proposed dual-path conditional diffusion network integrates the physics-guided decomposition and dynamically regulates denoising via a novel time frequency energy routing mechanism. By weighting reconstruction errors across frequency bands, our method improves anomaly localization and enhances interpretability. Extensive experiments on five benchmark datasets and two NeurIPS-TS scenarios demonstrate that PhysDiff outperforms 18 state-of-the-art baselines, with average F1-score improvements on both standard and challenging datasets. Experimental results validate the advantages of combining principled signal decomposition with diffusion-based reconstruction for robust, interpretable anomaly detection in complex dynamic systems.



Paperid:4583
Authors:Shengbo Gong, Juntong Ni, Noveen Sachdeva, Carl Yang, Wei Jin
Abstract:
Graph condensation (GC) is an emerging technique designed to learn a significantly smaller graph that retains the essential information of the original graph. This condensed graph has shown promise in accelerating graph neural networks while preserving performance comparable to those achieved with the original, larger graphs. Additionally, this technique facilitates downstream applications like neural architecture search and deepens our understanding of redundancies in large graphs. Despite the rapid development of GC methods, particularly for node classification, a unified evaluation framework is still lacking to systematically compare different GC methods or clarify key design choices for improving their effectiveness. To bridge these gaps, we introduce GC4NC, a comprehensive framework for evaluating diverse GC methods on node classification across multiple dimensions including performance, efficiency, privacy preservation, denoising ability, NAS effectiveness, and transferability. Our systematic evaluation offers novel insights into how condensed graphs behave and the critical design choices that drive their success. These findings pave the way for future advancements in GC methods, enhancing both performance and expanding their real-world applications. The code is available at https://github.com/Emory-Melody/GraphSlim/tree/main/benchmark.



Paperid:4584
Authors:Changlun Li, Yao SHI, Chen Wang, Qiqi Duan, Runke RUAN, Weijie Huang, Haonan Long, Lijun Huang, Yuyu Luo, Nan Tang
Abstract:
Large Language Models (LLMs) have demonstrated notable capabilities across financial tasks, including financial report summarization, earnings call transcript analysis, and asset classification. However, their real-world effectiveness in managing complex fund investment remains inadequately assessed. A fundamental limitation of existing benchmarks for evaluating LLM-driven trading strategies is their reliance on historical back-testing, inadvertently enabling LLMs to "time travel"—leveraging future information embedded in their training corpora, thus resulting in possible information leakage and overly optimistic performance estimates. To address this issue, we introduce DeepFund, a live fund benchmark tool designed to rigorously evaluate LLM in real-time market conditions. Utilizing a multi-agent architecture, DeepFund connects directly with real-time stock market data—specifically data published after each model’s pretraining cutoff—to ensure fair and leakage-free evaluations. Empirical tests on nine flagship LLMs from leading global institutions across multiple investment dimensions—including ticker-level analysis, investment decision-making, portfolio management, and risk control—reveal significant practical challenges. Notably, even cutting-edge models such as DeepSeek-V3 and Claude-3.7-Sonnet incur net trading losses within DeepFund real-time evaluation environment, underscoring the present limitations of LLMs for active fund management. Our code is available at https://github.com/HKUSTDial/DeepFund.



Paperid:4585
Authors:Kunyun Wang, Bohan Li, Kai Yu, Minyi Guo, Jieru Zhao
Abstract:
Abstract:Diffusion models have emerged as a powerful class of generative models across various modalities, including image, video, and audio synthesis. However, their deployment is often limited by significant inference latency, primarily due to the inherently sequential nature of the denoising process. While existing parallelization strategies attempt to accelerate inference by distributing computation across multiple devices, they typically incur high communication overhead, hindering deployment on commercial hardware. To address this challenge, we propose $\textbf{ParaStep}$, a novel parallelization method based on a reuse-then-predict mechanism that parallelizes diffusion inference by exploiting similarity between adjacent denoising steps. Unlike prior approaches that rely on layer-wise or stage-wise communication, ParaStep employs lightweight, step-wise communication, substantially reducing overhead. ParaStep achieves end-to-end speedups of up to $\textbf{3.88}$$\times$ on SVD, $\textbf{2.43}$$\times$ on CogVideoX-2b, and $\textbf{6.56}$$\times$ on AudioLDM2-large, while maintaining generation quality. These results highlight ParaStep as a scalable and communication-efficient solution for accelerating diffusion inference, particularly in bandwidth-constrained environments.



Paperid:4586
Authors:Hongjoon Ahn, Heewoong Choi, Jisu Han, Taesup Moon
Abstract:
Offline goal-conditioned reinforcement learning (GCRL) offers a practical learning paradigm where goal-reaching policies are trained from abundant unlabeled (reward-free) datasets without additional environment interaction. However, offline GCRL still struggles with long-horizon tasks, even with recent advances that employ hierarchical policy structures, such as HIQL. By identifying the root cause of this challenge, we observe the following insights: First, performance bottlenecks mainly stem from the high-level policy’s inability to generate appropriate subgoals. Second, when learning the high-level policy in the long-horizon regime, the sign of the advantage signal frequently becomes incorrect. Thus, we argue that improving the value function to produce a clear advantage signal for learning the high-level policy is essential. In this paper, we propose a simple yet effective solution:Option-aware Temporally Abstractedvalue learning, dubbedOTA, which incorporates temporal abstraction into the temporal-difference learning process. By modifying the value update to be option-aware, the proposed learning scheme contracts the effective horizon length, enabling better advantage estimates even in long-horizon regimes. We experimentally show that the high-level policy extracted using the OTA value function achieves strong performance on complex tasks from OGBench, a recently proposed offline GCRL benchmark, including maze navigation and visual robotic manipulation environments.



Paperid:4587
Authors:Rong Ye, Yongxin Zhang, yikai zhang, Haoyu Kuang, peng sun, zhongyu wei
Abstract:
Achieving Artificial General Intelligence (AGI) requires AI agents that can not only make strategic decisions but also engage in flexible and meaningful communication. Inspired by Wittgenstein's language game theory, we propose that language agents can learn through in-context interaction rather than traditional multi-stage frameworks that separate decision-making from language expression. Using Werewolf, a social deduction game that tests language understanding, strategic interaction, and adaptability, as a test bed, we develop the Multi-agent Kahneman-Tversky's Optimization (MaKTO). MaKTO engages diverse models in extensive gameplay to generate unpaired desirable and unacceptable responses, then employs KTO to refine the model's decision-making process. In 9-player Werewolf games, MaKTO achieves a 61% average win rate across various models, outperforming GPT-4o and two-stage RL agents by relative improvements of 23.0% and 10.9%, respectively. Notably, MaKTO also demonstrates human-like performance, winning 60% against expert players and showing only 48.9% detectability in Turing-style blind tests.



Paperid:4588
Authors:Chao Su, Likang Peng, Yuan Sun, Dezhong Peng, Xi Peng, Xu Wang
Abstract:
Cross-modal hashing aims to efficiently retrieve information across different modalities by mapping data into compact hash codes. However, most existing methods assume access to fully accurate supervision, which rarely holds in real-world scenarios. Instead, annotations are often redundant, i.e., each sample is associated with a set of candidate labels that includes both ground-truth and redundant noisy entries. Treating all annotated labels as equally valid introduces two critical issues: (1) the sparse presence of true labels within the label set is not explicitly addressed, leading to overfitting on redundant noisy annotations; (2) redundant noisy labels induce spurious similarities that distort semantic alignment across modalities and degrade the quality of the hash space. To address these challenges, we propose that effective cross-modal hashing requires explicitly identifying and leveraging the true label subset within the entire annotations. Based on this insight, we present Neighbor-aware Contrastive Disambiguation (NACD), a novel framework designed for robust learning under redundant supervision. NACD introduces two key components. The first, Neighbor-aware Confidence Reconstruction (NACR), refines label confidence by aggregating information from cross-modal neighbors to distinguish true labels from redundant noisy ones. The second, Class-aware Robust Contrastive Hashing (CRCH), constructs reliable positive and negative pairs based on label confidence scores, which significantly improves robustness against noisy supervision. Moreover, to effectively reduce the quantization error, we introduce an innovative quantization loss that enforces binary constraints on the learned hash representations. Extensive experiments on three large-scale multimodal benchmarks demonstrate that our method consistently achieves superior performance over state-of-the-art approaches, setting a new standard for cross-modal hashing with redundant annotations.



Paperid:4589
Authors:Dhruva Karkada, James Simon, Yasaman Bahri, Michael Deweese
Abstract:
Self-supervised word embedding algorithms provide a relevant and minimal setting for studying representation learning in natural language modeling. We examine the quartic approximation of word2vec and empirically show that it closely matches both the training dynamics of word2vec and its performance on downstream tasks.Our main contribution is to analytically solve for both the training dynamics and the final word embeddings in terms of only the corpus statistics and hyperparameters.The solutions reveal that these models learn orthogonal linear subspaces one at a time, each one incrementing the effective rank of the embeddings until model capacity is saturated.Training on Wikipedia, we find that each of the top linear subspaces represents an interpretable topic-level concept.Finally, we apply our theory to describe how linear representations of more abstract concepts emerge during training; these can be used to complete analogies via vector addition.



Paperid:4590
Authors:Imezadelajara, Cristian Rodriguez-Opazo, Damien Teney, Damith Ranasinghe, Ehsan Abbasnejad
Abstract:
Abstract:Out-of-distribution (OOD) detection is essential for reliably deploying machine learning models in the wild. Yet, most methods treat large pre-trained models as monolithic encoders and rely solely on their final-layer representations for detection. We challenge this wisdom. We reveal the intermediate layers of pre-trained models, shaped by residual connections that subtly transform input projections, can encode surprisingly rich and diverse signals for detecting distributional shifts. Importantly, to exploit latent representation diversity across layers, we introduce an entropy-based criterion to automatically identify layers offering the most complementary information in a training-free setting, without access to OOD data. We show that selectively incorporating these intermediate representations can increase the accuracy of OOD detection by up to $10\%$ in far-OOD and over $7\%$ in near-OOD benchmarks compared to state-of-the-art training-free methods across various model architectures and training objectives. Our findings reveal a new avenue for OOD detection research and uncover the impact of various training objectives and model architectures on confidence-based OOD detection methods.



Paperid:4591
Authors:Shizhe Diao, Yu Yang, Yonggan Fu, Xin Dong, Dan SU, Markus Kliegl, ZIJIA CHEN, Peter Belcak, Yoshi Suhara, Hongxu Yin, Mostofa Patwary, Yingyan (Celine) Lin, Jan Kautz, Pavlo Molchanov
Abstract:
Pre-training datasets are typically collected from web content and lack inherent domain divisions. For instance, widely used datasets like Common Crawl do not include explicit domain labels, while manually curating labeled datasets such as The Pile is labor-intensive. Consequently, identifying an optimal pre-training data mixture remains a challenging problem, despite its significant benefits for pre-training performance. To address these challenges, we propose CLustering-based Iterative Data Mixture Bootstrapping (CLIMB), an automated framework that discovers, evaluates, and refines data mixtures in a pre-training setting. Specifically, CLIMB embeds and clusters large-scale datasets in a semantic space and then iteratively searches for optimal mixtures using a smaller proxy model and apredictor. This strategy enables effective domain adaptation without relying solely on curated data. When continuously trained on 400B tokens with this mixture, our 1B model exceeds the state-of-the-art Llama-3.2-1B by 2.0%. Moreover, we observe that optimizing for a specific domain (e.g., Social Sciences) yields a 5% improvement over random sampling. Finally, we introduce ClimbLab, a filtered1.2-trillion-token corpus with 20 clusters as a research playground, and ClimbMix, a compact yet powerful 400-billion-token dataset designed for efficient pre-training that delivers superior performance under an equal token budget. We analyze the final data mixture, elucidating the characteristics of an optimal data mixture.



Paperid:4592
Authors:Haifeng Zhao, Haiyang Li, Lei-Lei Ma, Dengdi Sun
Abstract:
The scarcity of annotated medical imaging data has driven significant progress in semi-supervised learning to alleviate reliance on expensive expert labeling. While foundational vision models such as the Segment Anything Model (SAM) exhibit robust generalization in generic segmentation tasks, their direct application to medical images often results in suboptimal performance. To address this challenge, in this work, we propose a novel fully SAM-based semi-supervised medical image segmentation framework and develop the corresponding knowledge distillation-based learning strategy. Specifically, we first employ an efficient SAM variant as the backbone network of the semi‑supervised framework and update the default prompt embedding of SAM to unleash its full potential. Then, we utilize an original SAM, which is rich in prior knowledge, as the teacher to optimize our efficient student SAM backbone through hierarchical knowledge distillation and a dynamic loss weighting strategy. Extensive experiments on various medical datasets demonstrate that our method outperforms state-of-the-art semi-supervised segmentation approaches. Especially, our model requires less than 10% of the parameter size of the original SAM, enabling substantially lower deployment and storage overhead in real-world clinical settings.



Paperid:4593
Authors:Kaijun Zhou, Xingyu Yan, Xingda Wei, Xijun Li, Jinyu Gu
Abstract:
Abstract:Diffusion in transformer is central to advances in high-quality multimodal generation but suffer from high inference latency due to their iterative nature. Inspired by speculative decoding's success in accelerating large language models, we propose Speculative Diffusion with Speculative Verification (SDSV), a novel method to enhance the efficiency of diffusion models. Adapting speculative execution to diffusion processes presents unique challenges. First, the substantial computational cost of verifying numerous speculative steps for continuous, high-dimensional outputs makes traditional full verification prohibitively expensive. Second, determining the optimal number of speculative steps $K$ involves a trade-off between potential acceleration and verification success rates. To address these, SDSV introduces two key innovations: 1) A speculative verification technique, which leverages the observed temporal correlation between draft and target model outputs, efficiently validates $K$ speculative steps by only checking the alignment of the initial and final states, significantly reducing verification overhead. 2) A multi-stage speculative strategy that adjusts $K$ according to the denoising phase—employing smaller $K$ during volatile early stages and larger $K$ during more stable later stages to optimize the balance between speed and quality. We apply SDSV to state-of-the-art diffusion transformers, including Flux.1-dev for image generation and Wan2.1 for video generation. Extensive experiments demonstrate that SDSV achieves up to 1.77$\times$-3.01$\times$ speedup in model inference with a minimal 0.3\%-0.4\% drop in VBench score, showcasing its effectiveness in accelerating multimodal diffusion models without significant quality degradation. The code will be publicly available once the acceptance of the paper.



Paperid:4594
Authors:Marlon Tobaben, Hibiki Ito, Joonas Jälkö, Yuan He, Antti Honkela
Abstract:
Membership inference attacks (MIAs) are used to test practical privacy of machine learning models. MIAs complement formal guarantees from differential privacy (DP) under a more realistic adversary model. We analyse MIA vulnerability of fine-tuned neural networks both empirically and theoretically, the latter using a simplified model of fine-tuning. We show that the vulnerability of non-DP models when measured as the attacker advantage at fixed false positive rate reduces according to a simple power law as the number of examples per class increases, even for the most vulnerable points, but the dataset size needed for adequate protection of the most vulnerable points is very large.



Paperid:4595
Authors:James Roggeveen, Erik Wang, Will Flintoft, Peter Donets, Raglan Ward, Ahmed Roman, Anton Graf, Siddharth Dandavate, Ava Williamson, David Ettel, Felix Yeung, Kacper Migacz, Yijun Wang, Egemen Bostan, Duy Thuc Nguyen, Zhe He, Marc Descoteaux, Anne Mykland, Shida Liu, Jorge Garcia Ponce, Luke Zhu, Yuyang Chen, Ekaterina Ivshina, Miguel Fernandez, Minjae Kim, Kennan Gumbs, Matthew Tan, Russell Yang, Mai Hoang, David Brown, Isabella Silveira, Lavon Sykes, Arjun Nageswaran, William Fredenberg, Yiming Chen, Lucas Martin, Yixing Tang, Kelly Smith, Hongyu Liao, Logan Wilson, Alexander D. Cai, Lucy Nathwani, Nickholas Gutierrez, Andrea Elizabeth Biju, Michael Brenner
Abstract:
Abstract:Large language models (LLMs) have shown remarkable progress in mathematical problem-solving, but evaluation has largely focused on problems that have exact analytical solutions or involve formal proofs, often overlooking approximation-based problems ubiquitous in applied science and engineering. To fill this gap, we build on prior work and present $\textbf{HARDMath2}$, a dataset of 211 original problems covering the core topics in an introductory graduate applied math class, including boundary-layer analysis, WKB methods, asymptotic solutions of nonlinear partial differential equations, and the asymptotics of oscillatory integrals. This dataset was designed and verified by the students and instructors of a core graduate applied mathematics course at Harvard. We build the dataset through a novel collaborative environment that challenges students to write and refine difficult problems consistent with the class syllabus, peer-validate solutions, test different models, and automatically check LLM-generated solutions against their own answers and numerical ground truths. Evaluation results show that leading frontier models still struggle with many of the problems in the dataset, highlighting a gap in the mathematical reasoning skills of current LLMs. Importantly, students identified strategies to create increasingly difficult problems by interacting with the models and exploiting common failure modes. This back-and-forth with the models not only resulted in a richer and more challenging benchmark but also led to qualitative improvements in the students' understanding of the course material, which is increasingly important as we enter an age where state-of-the-art language models can solve many challenging problems across a wide domain of fields.



Paperid:4596
Authors:Jingmin Zhu, Anqi Zhu, Hossein Rahmani, Jun Liu, Mohammed Bennamoun, Qiuhong Ke
Abstract:
We introduce \textit{Skeleton-Cache}, the first training-free test-time adaptation framework for skeleton-based zero-shot action recognition (SZAR), aimed at improving model generalization to unseen actions during inference. Skeleton-Cache reformulates inference as a lightweight retrieval process over a non-parametric cache that stores structured skeleton representations, combining both global and fine-grained local descriptors. To guide the fusion of descriptor-wise predictions, we leverage the semantic reasoning capabilities of large language models (LLMs) to assign class-specific importance weights. By integrating these structured descriptors with LLM-guided semantic priors, Skeleton-Cache dynamically adapts to unseen actions without any additional training or access to training data. Extensive experiments on NTU RGB+D 60/120 and PKU-MMD II demonstrate that Skeleton-Cache consistently boosts the performance of various SZAR backbones under both zero-shot and generalized zero-shot settings. Code will be released.



Paperid:4597
Authors:Yang Xu, Washim Mondal, Vaneet Aggarwal
Abstract:
Abstract:We present the first finite-sample analysis of policy evaluation in robust average-reward Markov Decision Processes (MDPs). Prior work in this setting have established only asymptotic convergence guarantees, leaving open the question of sample complexity. In this work, we address this gap by showing that the robust Bellman operator is a contraction under a carefully constructed semi-norm, and developing a stochastic approximation framework with controlled bias. Our approach builds upon Multi-Level Monte Carlo (MLMC) techniques to estimate the robust Bellman operator efficiently. To overcome the infinite expected sample complexity inherent in standard MLMC, we introduce a truncation mechanism based on a geometric distribution, ensuring a finite expected sample complexity while maintaining a small bias that decays exponentially with the truncation level. Our method achieves the order-optimal sample complexity of $\tilde{\mathcal{O}}(\epsilon^{-2})$ for robust policy evaluation and robust average reward estimation, marking a significant advancement in robust reinforcement learning theory.



Paperid:4598
Authors:Nimrod Berman, Omkar Joglekar, Eitan Kosman, Dotan Di Castro, Omri Azencot
Abstract:
Recent advances in generative modeling have positioned diffusion models as state-of-the-art tools for sampling from complex data distributions. While these models have shown remarkable success across single-modality domains such as images and audio, extending their capabilities toModality Translation (MT)—translating information across different sensory modalities—remains an open challenge. Existing approaches often rely on restrictive assumptions, including shared dimensionality, Gaussian source priors, and modality-specific architectures, which limit their generality and theoretical grounding. In this work, we propose a general-purpose framework for modality translation based on a latent-variable extension of Denoising Diffusion Bridge Models. By operating in a shared latent space, our method learns bridge between arbitrary modalities without requiring aligned dimensions. We introduce a contrastive alignment loss to enforce semantic consistency between paired samples and design a domain-agnostic encoder-decoder architecture tailored for noise prediction in latent space. Additionally, we propose a predictive loss to guide training towards accurate cross-domain translation and explore several training strategies to improve stability. Our approach supports arbitrary modality pairs and demonstrates strong performance on diverse MT tasks, including multi-view to 3D shape generation, image super-resolution, and multi-view scene synthesis. Comprehensive experiments and ablations validate the effectiveness of our framework, establishing a new strong baseline in general modality translation.



Paperid:4599
Authors:Yehonathan Refael, Guy Smorodinsky, Tom Tirer, Ofir Lindenbaum
Abstract:
Low-rank gradient-based optimization methods have significantly improved memory efficiency during the training of large language models (LLMs), enabling operations within constrained hardware without sacrificing performance. However, these methods primarily emphasize memory savings, often overlooking potential acceleration in convergence due to their reliance on standard isotropic steepest descent techniques, which can perform suboptimally in the highly anisotropic landscapes typical of deep networks, particularly LLMs. In this paper, we propose SUMO (Subspace-Aware Moment-Orthogonalization), an optimizer that employs exact singular value decomposition (SVD) for moment orthogonalization within a dynamically adapted low-dimensional subspace, enabling norm-inducing steepest descent optimization steps. By explicitly aligning optimization steps with the spectral characteristics of the loss landscape, SUMO effectively mitigates approximation errors associated with commonly used methods like Newton-Schulz orthogonalization approximation. We theoretically establish an upper bound on these approximation errors, proving their dependence on the condition numbers of moments, conditions we analytically demonstrate are encountered during LLM training. Furthermore, we both theoretically and empirically illustrate that exact orthogonalization via SVD substantially improves convergence rates while reducing overall complexity. Empirical evaluations confirm that SUMO accelerates convergence, enhances stability, improves performance, and reduces memory requirements by up to 20\% compared to state-of-the-art methods.



Paperid:4600
Authors:Zonglin Yang, Wanhao Liu, Ben Gao, Yujie Liu, Wei Li, Tong Xie, Lidong Bing, Wanli Ouyang, Erik Cambria, Dongzhan Zhou
Abstract:
Large language models (LLMs) have shown promise in automating scientific hypothesis generation, yet existing approaches primarily yield coarse-grained hypotheses lacking critical methodological and experimental details. We introduce and formally define the novel task of fine-grained scientific hypothesis discovery, which entails generating detailed, experimentally actionable hypotheses from coarse initial research directions. We frame this as a combinatorial optimization problem and investigate the upper limits of LLMs’ capacity to solve it when maximally leveraged. Specifically, we explore three foundational questions: (1) how to best harness an LLM’s internal heuristics to formulate the fine-grained hypothesis it itself would judge as the most promising among all the possible hypotheses it might generate; (2) whether such LLM-judged better hypotheses exhibit stronger alignment with ground-truth hypotheses; and (3) whether guiding the hypothesis search with an ensemble of diverse LLMs of similar capacity yields better outcomes than using repeated instances of the strongest LLM among them. To address these questions, we propose a hierarchical search method that incrementally proposes and integrates details into the hypothesis, progressing from general concepts to specific experimental configurations. We show that this hierarchical process smooths the reward landscape and enables more effective optimization. Empirical evaluations on a new benchmark of expert-annotated fine-grained hypotheses from recent chemistry literature show that our method consistently outperforms strong baselines.



Paperid:4601
Authors:Ming Li, Jike Zhong, Shitian Zhao, Yuxiang Lai, Haoquan Zhang, Wang Bill Zhu, Kaipeng Zhang
Abstract:
This paper investigates the role of explicit thinking process in rule-based reinforcement fine-tuning (RFT) for multi-modal large language models (MLLMs). We first extend \textit{Thinking-RFT} to image classification task, using verifiable rewards for fine-tuning~(FT). Experiments show {Thinking-RFT} significantly outperforms supervised FT and yields a cross-dataset generalization effect. We then rethink and question whether explicit thinking in RFT is always necessary and beneficial. Challenging the convention that explicit thinking is crucial for the success of RFT, we introduce \textit{No-Thinking-RFT}, exploring RFT without thinking by introducing a simple equality accuracy reward. We evaluate No-Thinking-RFT on six diverse tasks across different model sizes and types. Experiment results reveal four key findings: \textbf{(1).} Visual perception tasks do not require thinking during RFT, as No-Thinking-RFT consistently outperforms or matches Thinking-RFT across model sizes and types. \textbf{(2).} Models with limited capabilities struggle to generate high-quality CoT for RFT, making Thinking-RFT less effective than No-Thinking-RFT. \textbf{(3).} There are inconsistencies between the answers in the thinking tags and answer tags for some responses of Thinking-RFT, which show lower average accuracy than the overall accuracy. \textbf{(4).} The performance gain of No-Thinking-RFT mainly stems from improved learning during no thinking FT and the avoidance of inference overthinking, as evidenced by the partial gains from appending empty thinking tags at inference time of Thinking-RFT. We hypothesize that explicit thinking before verifiable answers may hinder reward convergence and reduce performance in certain scenarios. To test this, we propose \textit{Think-After-Answer}, which places thinking after the answer to mitigate this effect for experimental verification. Lastly, we conduct a pilot study to explore whether MLLMs can learn when to think during RFT, introducing an \textit{Adaptive-Thinking} method. Experiments show that model converges to either thinking or not depending on model capability, achieving comparable or better performance than both Thinking and No-Thinking-RFT. Our findings suggest MLLMs can adaptively decide to think or not based on their capabilities and task complexity, offering insights into the thinking process in RFT.



Paperid:4602
Authors:Hanlei Zhang, zhuohang li, Yeshuang Zhu, Hua Xu, Peiwu Wang, Haige Zhu, Jie Zhou, Jinchao Zhang
Abstract:
Multimodal language analysis is a rapidly evolving field that leverages multiple modalities to enhance the understanding of high-level semantics underlying human conversational utterances. Despite its significance, little research has investigated the capability of multimodal large language models (MLLMs) to comprehend cognitive-level semantics. In this paper, we introduce MMLA, a comprehensive benchmark specifically designed to address this gap. MMLA comprises over 61K multimodal utterances drawn from both staged and real-world scenarios, covering six core dimensions of multimodal semantics: intent, emotion, dialogue act, sentiment, speaking style, and communication behavior. We evaluate eight mainstream branches of LLMs and MLLMs using three methods: zero-shot inference, supervised fine-tuning, and instruction tuning. Extensive experiments reveal that even fine-tuned models achieve only about 60~70% accuracy, underscoring the limitations of current MLLMs in understanding complex human language. We believe that MMLA will serve as a solid foundation for exploring the potential of large language models in multimodal language analysis and provide valuable resources to advance this field. The datasets and code are open-sourced at https://github.com/thuiar/MMLA.



Paperid:4603
Authors:Yuxuan Han, Jose Blanchet, Zhengyuan Zhou
Abstract:
Abstract:In this work, we consider the data-driven assortment optimization problem under the linear MNL choice model.We first establish a improved confidence region for the MLE of the $d$-dimensional linear MNL likelihood function that removing the explicit dependency on a problem-dependent parameter $\kappa^{-1}$ in previous result (Oh and Iyengar, 2021), which scales exponentially with the radius of the parameter set.Building on the confidence region result, we investigate the data-driven assortment optimization problem in both offline and online settings. In the offline setting, the previously best-known result scales as $\tilde{O}\left(\sqrt{\frac{d}{\kappa n_{S^\star}}}\right)$, where $n_{S^\star}$ the number of times that optimal assortment $S^\star$ is observed (Dong et al., 2023). We propose a new pessimistic-based algorithm that, under a burn-in condition, removes the dependency on $d,\kappa^{-1}$ in the leading order bound and works under a more relaxed coverage condition, without requiring the exact observation of $S^\star$. In the online setting, we propose the first algorithm to achieve $\tilde{O}(\sqrt{dT})$ regret without a multiplicative dependency on $\kappa^{-1}$. In both settings, our results nearly achieve the corresponding lower bound when reduced to the canonical $N$-item MNL problem, demonstrating their optimality.



Paperid:4604
Authors:Rishabh Agrawal, Kaushik Ranade, Aja Khanal, Kalyan Basu, Apurva Narayan
Abstract:
We present SpecMAS, a novel multi-agent system that autonomously constructs and formally verifies executable system models from natural language specifications. Given a Standard Operating Procedure (SOP) describing a target system, SpecMAS parses the specification, identifies relevant operational modes, variables, transitions, and properties, and generates a formal model in NuSMV code syntax, an industry-standard symbolic model checker. A dedicated reasoning agent extracts both explicit and implicit properties from the SOP, and verification is performed via temporal logic model checking. If any properties fail to verify, an autonomous debugging agent analyzes counterexamples and iteratively corrects the model until all properties are satisfied. This closed-loop system design guarantees provable correctness by construction and advances the state of the art in automated, interpretable, and deployable verification pipelines. We demonstrate the generality, correctness, and practical feasibility of SpecMAS across a set of representative case studies and propose a new benchmark dataset for the evaluation and comparison of model checking performance.



Paperid:4605
Authors:Frederik Kunstner, Francis Bach
Abstract:
Abstract:Recent works have highlighted the optimization difficulties encountered by gradient descent in training the first and last layer of transformer-based language models, which are overcome by optimizers such as Adam. The problem appears linked to the heavy-tailed distribution of words in text data, where the frequency of the $k$th most frequent word $\pi_k$ is proportional to $1/k$, following Zipf's law. To better understand the impact of the data distribution on training performance, we study a linear bigram model for next-token prediction when the tokens follow a power-law $\pi_k \propto 1/k^\alpha$ parameterized by the exponent $\alpha$. We derive optimization scaling laws for deterministic gradient descent and sign descent as a proxy for Adam as a function of the power $\alpha \geq 0$. This setting differs from existing theoretical investigations in scaling laws which assume that the eigenvalues of the data decay as a power with power $\alpha > 1$. This assumption effectively makes the problem "finite dimensional" as most of the loss comes from a few of the largest eigencomponents. In comparison, we show that the problem is more difficult when the data have heavier tails. The case $\alpha = 1$ as found in text is ``worst-case'' for gradient descent, in that the number of iterations required to reach a small relative error scales almost linearly with dimension. While the performance of sign descent also depends on the dimension, for Zipf-distributed data the number of iterations scales only with the square-root of the dimension, leading to a large improvement over gradient descent for large vocabularies.



Paperid:4606
Authors:Ruiping Liu, Junwei Zheng, Yufan Chen, Zirui Wang, Kunyu Peng, Kailun Yang, Jiaming Zhang, Marc Pollefeys, Rainer Stiefelhagen
Abstract:
Physical environments and circumstances are fundamentally dynamic, yet current 3D datasets and evaluation benchmarks tend to concentrate on either dynamic scenarios or dynamic situations in isolation, resulting in incomplete comprehension. To overcome these constraints, we introduce Situat3DChange, an extensive dataset supporting three situation-aware change understanding tasks following the perception-action model: 121K question-answer pairs, 36K change descriptions for perception tasks, and 17K rearrangement instructions for the action task. To construct this large-scale dataset, Situat3DChange leverages 11K human observations of environmental changes to establish shared mental models and shared situational awareness for human-AI collaboration. These observations, enriched with egocentric and allocentric perspectives as well as categorical and coordinate spatial relations, are integrated using an LLM to support understanding of situated changes. To address the challenge of comparing pairs of point clouds from the same scene with minor changes, we propose SCReasoner, an efficient 3D MLLM approach that enables effective point cloud comparison with minimal parameter overhead and no additional tokens required for the language decoder. Comprehensive evaluation on Situat3DChange tasks highlights both the progress and limitations of MLLMs in dynamic scene and situation understanding. Additional experiments on data scaling and cross-domain transfer demonstrate the task-agnostic effectiveness of using Situat3DChange as a training dataset for MLLMs. The established dataset and source code are publicly available at: https://github.com/RuipingL/Situat3DChange.



Paperid:4607
Authors:Feng Zhu, David Simchi-Levi
Abstract:
Thompson Sampling (TS) has emerged as a powerful algorithm for sequential decision-making, with strong empirical success and theoretical guarantees. However, it has been shown that its behavior under stringent safety and robustness criteria --- such as safety of cumulative regret distribution and robustness to model mis-specification --- can sometimes perform poorly. In this work, we try to address these aspects through the lens of adaptive variance inflation for Gaussian Thompson Sampling. Our one-line change introduces a time- and arm-dependent inflation factor into the sampling variance, and yields several compelling benefits. The resulting policy achieves provably worst-case optimal expected regret and worst-case optimal fast-decaying regret tail bounds, even in the presence of heavy-tailed (sub-exponential) noise or mis-specified environments. The policy is also robust to mis-specified noise variances. Beyond cumulative regret, we further demonstrate that our method ensures strong post-experiment guarantees: simple regret and estimation error per arm exhibit fast-decaying tail probabilities, contributing to more reliable and robust downstream decisions. Finally, we extend our policy to incorporate settings with unknown arm-specific variances and empirically validate the consistent performance of our approach across a range of environments.



Paperid:4608
Authors:Zichong Wang, Zhipeng Yin, Wenbin Zhang
Abstract:
Graph generation models play pivotal roles in many real-world applications, from data augmentation to privacy-preserving. Despite their deployment successes, existing approaches often exhibit fairness issues, limiting their adoption in high-risk decision-making applications. Most existing fair graph generation works are based on autoregressive models that suffer from ordering sensitivity, while primarily addressing structural bias and overlooking the critical issue of feature bias. To this end, we propose FairGEM, a novel one-shot graph generation framework designed to mitigate both graph structural bias and node feature bias simultaneously. Furthermore, our theoretical analysis establishes that FairGEM delivers substantially stronger fairness guarantees than existing models while preserving generation quality. Extensive experiments across multiple real-world datasets demonstrate that FairGEM achieves superior performance in both generation quality and fairness.



Paperid:4609
Authors:Tianxu Wang, Zhuofan Zhang, Ziyu Zhu, Yue Fan, Jing Xiong, Pengxiang Li, Xiaojian (Shawn) Ma, Qing Li
Abstract:
3D visual grounding has made notable progress in localizing objects within complex 3D scenes. However, grounding referring expressions beyond objects in 3D scenes remains unexplored. In this paper, we introduce Anywhere3D-Bench, a holistic 3D visual grounding benchmark consisting of 2,632 referring expression-3D bounding box pairs spanning four different grounding levels: human-activity areas, unoccupied space beyond objects, individual objects in the scene, and fine-grained object parts. We assess a range of state-of-the-art 3D visual grounding methods alongside large language models (LLMs) and multimodal LLMs (MLLMs) on Anywhere3D-Bench. Experimental results reveal that space-level and part-level visual grounding pose the greatest challenges: space-level tasks require a more comprehensive spatial reasoning ability, for example, modeling distances and spatial relations within 3D space, while part-level tasks demand fine-grained perception of object composition. Even the best performance model, OpenAI o4-mini, achieves only 22.94% accuracy on space-level tasks and 33.68% on part-level tasks, significantly lower than its performance on area-level and object-level tasks. These findings underscore a critical gap in current models’ capacity to understand and reason about 3D scene beyond object-level semantics.



Paperid:4610
Authors:Teng Hu, Zhentao Yu, Zhengguang Zhou, Jiangning Zhang, Yuan Zhou, Qinglin Lu, Ran Yi
Abstract:
Despite recent advances in video generation, existing models still lack fine-grained controllability, especially for multi-subject customization with consistent identity and interaction.In this paper, we propose PolyVivid, a multi-subject video customization framework that enables flexible and identity-consistent generation. To establish accurate correspondences between subject images and textual entities, we design a VLLM-based text-image fusion module that embeds visual identities into the textual space for precise grounding. To further enhance identity preservation and subject interaction, we propose a 3D-RoPE-based enhancement module that enables structured bidirectional fusion between text and image embeddings. Moreover, we develop an attention-inherited identity injection module to effectively inject fused identity features into the video generation process, mitigating identity drift. Finally, we construct an MLLM-based data pipeline that combines MLLM-based grounding, segmentation, and a clique-based subject consolidation strategy to produce high-quality multi-subject data, effectively enhancing subject distinction and reducing ambiguity in downstream video generation.Extensive experiments demonstrate that PolyVivid achieves superior performance in identity fidelity, video realism, and subject alignment, outperforming existing open-source and commercial baselines. More comprehensive video results and comparisons are shown on the project page in the supplementary material.



Paperid:4611
Authors:Junxi Xiao, Qinliang Su, Zexin Yuan
Abstract:
Abstract:Mean-field variational inference (VI), despite its scalability, is limited by the independence assumption, making it unsuitable for scenarios with correlated data instances. Existing structured VI methods either focus on correlations among latent dimensions which lack scalability for modeling instance-level correlations, or are restricted to simple first-order dependencies, limiting their expressiveness. In this paper, we propose a novel VI framework that explicitly models $k$-order instance-level correlations among latent variables. By expressing the global posterior through overlapping $k$-dimensional local marginals, our method enables efficient parameterized sampling via a sequential procedure. To ensure the validity of these marginals, we introduce a conditional correlation parameterization method that guarantees positive definiteness of their correlation matrices. We further extend our method with a tree-structured backbone to capture more flexible dependency patterns. Extensive experiments on time-series and graph-structured datasets demonstrate that modeling higher-order correlations leads to significantly improved posterior approximations and better performance across various downstream tasks.



Paperid:4612
Authors:Zongrui Zou, Chenglin Fan, Michael Dinitz, Jingcheng Liu, Jalaj Upadhyay
Abstract:
Abstract:We study the problem of approximating all-pair distances in a weighted undirected graph with differential privacy, introduced by Sealfon [Sea16]. Given a publicly known undirected graph, we treat the weights of edges as sensitive information, and two graphs are neighbors if their edge weights differ in one edge by at most one. We obtain efficient algorithms with significantly improved bounds on a broad class of graphs which we refer to as *recursively separable*. In particular, for any $n$-vertex $K_h$-minor-free graph, our algorithm achieve an additive error of $ \widetilde{O}(h(nW)^{1/3} ) $, where $ W $ represents the maximum edge weight; For grid graphs, the same algorithmic scheme achieve additive error of $ \widetilde{O}(n^{1/4}\sqrt{W}) $.Our approach can be seen as a generalization of the celebrated binary tree mechanism for range queries, as releasing range queries is equivalent to computing all-pair distances on a path graph. In essence, our approach is based on generalizing the binary tree mechanism to graphs that are *recursively separable*.



Paperid:4613
Authors:HAO SUN, Zhongyi Han, Hao Chen, Jindong Wang, Xin Gao, Yilong Yin
Abstract:
Foundation models pretrained on web-scale data drive contemporary transfer learning in vision, language, and multimodal tasks. Recent work shows that mild label noise in these corpora may lift in-distribution accuracy yet sharply reduce out-of-distribution generalization, an effect known as catastrophic inheritance. Medical data is especially sensitive because annotations are scarce, domain shifts are large, and pretraining sources are noisy.We present the first systematic analysis of catastrophic inheritance in medical models. Controlled label-corruption experiments expose a clear structural collapse: as noise rises, the skewness and kurtosis of feature and logit distributions decline, signaling a flattened representation space and diminished discriminative detail. These higher-order statistics form a compact, interpretable marker of degradation in fine-grained tasks such as histopathology.Guided by this finding, we introduce a fine-tuning objective that restores skewness and kurtosis through two scalar regularizers added to the task loss. The method leaves the backbone unchanged and incurs negligible overhead. Tests on PLIP models trained with Twitter pathology images, as well as other large-scale vision and language backbones, show consistent gains in robustness and cross-domain accuracy under varied noise levels.



Paperid:4614
Authors:Chunlong Xie, Kangjie Chen, Shangwei Guo, Shudong Zhang, Tianwei Zhang, Tao Xiang
Abstract:
Modern Text-to-Image (T2I) models deploy multi-layered defenses to block Not-Safe-For-Work (NSFW) content generation. These defenses typically include sequential layers such as prompt filters, concept erasers and image filters. While existing adversarial attacks have demonstrated vulnerabilities in isolated defense layers, they prove largely ineffective against multi-layered defenses deployed in real-world T2I systems. In this paper, we demonstrate that exploiting overlapping vulnerabilities across these distinct defense layers enables adversaries to systematically bypass the entire safeguard of T2I systems. We propose Transstratal Adversarial Attack (TAA, a novel black-box framework to compromise T2I models with multi-layered protection. It generates transstratal adversarial prompts to evade all defense layers simultaneously. This is accomplished through transstratal adversarial candidate generation using LLMs to fulfill implicit and subjective adversarial requirements against different defense layers, combined with adversarial genetic optimization for efficient black-box search to maximize the bypass rates and generated image harmfulness. Evaluated across 14 T2I models (e.g., Stable Diffusion, DALL·E, and Midjourney) and 17 safety modules, our attack achieves an average attack success rate of 85.6\%, surpassing state-of-the-art methods by 73.5\%. Our findings challenge the isolated design of safety mechanisms and establish the first benchmark for holistic robustness evaluation in multi-layered safeguarded T2I models. The code can be found in https://anonymous.4open.science/r/TAA-T2I-E572.



Paperid:4615
Authors:Xin Zou, Di Lu, Yizhou Wang, Yibo Yan, Yuanhuiyi Lyu, Xu Zheng, Linfeng Zhang, Xuming Hu
Abstract:
Despite their powerful capabilities, multimodal large language models (MLLMs) suffer from considerable computational overhead due to their reliance on massive visual tokens. Recent studies have explored token pruning to alleviate this problem, which typically uses text-vision cross-attention or [CLS] attention to assess and discard redundant visual tokens. In this work, we identify a critical limitation of such attention-first pruning approaches, i.e., they tend to preserve semantically similar tokens, resulting in pronounced performance drops under high pruning rates. To this end, we propose HoloV, a simple yet effective, plug-and-play visual token pruning framework for efficient inference. Distinct from previous attention-first schemes, HoloV rethinks token retention from a holistic perspective. By adaptively distributing the pruning budget across different spatial crops, HoloV ensures that the retained tokens capture the global visual context rather than isolated salient features. This strategy minimizes representational collapse and maintains task-relevant information even under aggressive pruning. Experimental results demonstrate that our HoloV achieves superior performance across various tasks, MLLM architectures, and pruning ratios compared to SOTA methods. For instance, LLaVA1.5 equipped with HoloV preserves 95.8% of the original performance after pruning 88.9% of visual tokens, achieving superior efficiency-accuracy trade-offs.



Paperid:4616
Authors:Jason Wu, Yuyang Yuan, Kang Yang, Lance Kaplan, Mani Srivastava
Abstract:
Multimodal deep learning systems are deployed in dynamic scenarios due to the robustness afforded by multiple sensing modalities. Nevertheless, they struggle with varying compute resource availability (due to multi-tenancy, device heterogeneity, etc.) and fluctuating quality of inputs (from sensor feed corruption, environmental noise, etc.). Statically provisioned multimodal systems cannot adapt when compute resources change over time, while existing dynamic networks struggle with strict compute budgets. Additionally, both systems often neglect the impact of variations in modality quality. Consequently, modalities suffering substantial corruption may needlessly consume resources better allocated towards other modalities. We propose ADMN, a layer-wise Adaptive Depth Multimodal Network capable of tackling both challenges - it adjusts the total number of active layers across all modalities to meet compute resource constraints, and continually reallocates layers across input modalities according to their modality quality. Our evaluations showcase ADMN can match the accuracy of state-of-the-art networks while reducing up to 75% of their floating-point operations.



Paperid:4617
Authors:Vaibhav Rathore, Divyam Gupta, Biplab Banerjee
Abstract:
Abstract:Generalized Category Discovery (GCD) aims to classify test-time samples into either seen categories—available during training—or novel ones, without relying on label supervision. Most existing GCD methods assume simultaneous access to labeled and unlabeled data during training and arising from the same domain, limiting applicability in open-world scenarios involving distribution shifts. Domain Generalization with GCD (DG-GCD) lifts this constraint by requiring models to generalize to unseen domains containing novel categories, without accessing target-domain data during training.The only prior DG-GCD method, DG$^2$CD-Net~\cite{dg2net}, relies on episodic training with multiple synthetic domains and task vector aggregation, incurring high computational cost and error accumulation. We propose \textsc{HiDISC}, a hyperbolic representation learning framework that achieves domain and category-level generalization without episodic simulation. To expose the model to minimal but diverse domain variations, we augment the source domain using GPT-guided diffusion, avoiding overfitting while maintaining efficiency.To structure the representation space, we introduce \emph{Tangent CutMix}, a curvature-aware interpolation that synthesizes pseudo-novel samples in tangent space, preserving manifold consistency. A unified loss—combining penalized Busemann alignment, hybrid hyperbolic contrastive regularization, and adaptive outlier repulsion—facilitates compact, semantically structured embeddings. A learnable curvature parameter further adapts the geometry to dataset complexity.\textsc{HiDISC} achieves state-of-the-art results on PACS~\cite{pacs}, Office-Home~\cite{officehome}, and DomainNet~\cite{domainnet}, consistently outperforming the existing Euclidean and hyperbolic (DG)-GCD baselines.



Paperid:4618
Authors:Runjian Chen, Hyoungseob Park, Bo Zhang, Wenqi Shao, Ping Luo, Alex Wong
Abstract:
Labeling LiDAR point clouds is notoriously time-and-energy-consuming, which spurs recent unsupervised 3D representation learning methods to alleviate the labeling burden in LiDAR perception via pretrained weights. Existing work focus on either masked auto encoding or contrastive learning on LiDAR point clouds, which neglects the temporal LiDAR sequence that naturally accounts for object motion (and their semantics). Instead, we propose TREND, short for Temporal REndering with Neural fielD, to learn 3D representation via forecasting the future observation in an unsupervised manner. TREND integrates forecasting for 3D pre-training through a Recurrent Embedding scheme to generate 3D embeddings across time and a Temporal LiDAR Neural Field specifically designed for LiDAR modality to represent the 3D scene, with which we compute the loss using differentiable rendering. We evaluate TREND on 3D object detection and LiDAR semantic segmentation tasks on popular datasets, including Once, Waymo, NuScenes, and SemanticKITTI. Experiment results show that TREND brings up to 400% more improvement as compared to previous SOTA unsupervised 3D pre-training methods and generally improve different downstream tasks across datasets, demonstrating the effectiveness of TREND. Codes and models will be released.



Paperid:4619
Authors:Wei Lin, Yiwei Zhou, Junkai Zhang, Rui Shao, Zhiyuan Zhao, Junyu Gao, Antoni Chan, Xuelong Li
Abstract:
Recent advancements in Vision-Language Models (VLMs) have enabled GUI agents to leverage visual features for interface understanding and operation in the digital world. However, limited research has addressed the interpretation and interaction with control panels in real-world settings. To bridge this gap, we propose the Panel Understanding and Operation (PUO) benchmark, comprising annotated panel images from appliances and associated vision-language instruction pairs. Experimental results on the benchmark demonstrate significant performance disparities between zero-shot and fine-tuned VLMs, revealing the lack of PUO-specific capabilities in existing language models. Furthermore, we introduce a Privacy-Preserving Framework (PPF) to address privacy concerns in cloud-based panel parsing and reasoning. PPF employs a dual-stage architecture, performing panel understanding on edge devices while delegating complex reasoning to cloud-based LLMs. Although this design introduces a performance trade-off due to edge model limitations, it eliminates the transmission of raw visual data, thereby mitigating privacy risks. Overall, this work provides foundational resources and methodologies for advancing interactive human-machine systems and robotic field in panel-centric applications.



Paperid:4620
Authors:Rushi Qiang, Yuchen Zhuang, Yinghao Li, Dingu Sagar V K, Rongzhi Zhang, ChangHao Li, Ian Wong, Sherry Yang, Percy Liang, Chao Zhang, Bo Dai
Abstract:
We introduce MLE-Dojo, a Gym-style framework for systematically reinforcement learning, evaluating, and improving autonomous large language model (LLM) agents in iterative machine learning engineering (MLE) workflows. Unlike existing benchmarks that primarily rely on static datasets or single-attempt evaluations, MLE-Dojo provides an interactive environment enabling agents to iteratively experiment, debug, and refine solutions through structured feedback loops. Built upon 200+ real-world Kaggle challenges, MLE-Dojo covers diverse, open-ended MLE tasks carefully curated to reflect realistic engineering scenarios such as data processing, architecture search, hyperparameter tuning, and code debugging. Its fully executable environment supports comprehensive agent training via both supervised fine-tuning and reinforcement learning, facilitating iterative experimentation, realistic data sampling, and real-time outcome verification. Extensive evaluations of eight frontier LLMs reveal that while current models achieve meaningful iterative improvements, they still exhibit significant limitations in autonomously generating long-horizon solutions and efficiently resolving complex errors. Furthermore, MLE-Dojo’s flexible and extensible architecture seamlessly integrates diverse data sources, tools, and evaluation protocols, uniquely enabling model-based agent tuning and promoting interoperability, scalability, and reproducibility. We open-source our framework and benchmarks to foster community-driven innovation towards next-generation MLE agents.



Paperid:4621
Authors:Chao Song, ZHIYUAN LIU, Han Huang, Liang Wang, Qiong Wang, Jian-Yu Shi, Hui Yu, Yihang Zhou, Yang Zhang
Abstract:
Designing enzyme backbones with substrate-specific functionality is a critical challenge in computational protein engineering. Current generative models excel in protein design but face limitations in binding data, substrate-specific control, and flexibility for de novo enzyme backbone generation. To address this, we introduceEnzyBind, a dataset with 11,100 experimentally validated enzyme-substrate pairs specifically curated from PDBbind. Building on this, we proposeEnzyControl, a method that enables functional and substrate-specific control in enzyme backbone generation. Our approach generates enzyme backbones conditioned on MSA-annotated catalytic sites and their corresponding substrates, which are automatically extracted from curated enzyme-substrate data. At the core of EnzyControl isEnzyAdapter, a lightweight, modular component integrated into a pretrained motif-scaffolding model, allowing it to become substrate-aware. A two-stage training paradigm further refines the model's ability to generate accurate and functional enzyme structures. Experiments show that our EnzyControl achieves the best performance across structural and functional metrics on EnzyBind and EnzyBench benchmarks, with particularly notable improvements of 25% in designability and 15% in catalytic efficiency compared to the baseline models.



Paperid:4622
Authors:Yinhan He, Wendy Zheng, Yaochen Zhu, Zaiyi Zheng, Lin Su, Sriram Vasudevan, Qi Guo, Liangjie Hong, Jundong Li
Abstract:
Chain-of-Thought (CoT) enhances the performance of Large Language Models (LLMs) on reasoning tasks by encouraging step-by-step solutions. However, the verbosity of CoT reasoning hinders its mass deployment in efficiency-critical applications. Recently, implicit CoT approaches have emerged, which encode reasoning steps within LLM's hidden embeddings (termed ``implicit reasoning'') rather than explicit tokens. This approach accelerates CoT reasoning by reducing the reasoning length and bypassing some LLM components. However, existing implicit CoT methods face two significant challenges: (1) they fail to preserve the semantic alignment between the implicit reasoning (when transformed to natural language) and the ground-truth reasoning, resulting in a significant CoT performance degradation, and (2) they focus on reducing the length of the implicit reasoning; however, they neglect the considerable time cost for an LLM to generate one individual implicit reasoning token. To tackle these challenges, we propose a novel semantically-aligned implicit CoT framework termedSemCoT. In particular, for the first challenge, we design a contrastively trained sentence transformer that evaluates semantic alignment between implicit and explicit reasoning, which is used to enforce semantic preservation during implicit reasoning optimization. To address the second challenge, we introduce an efficient implicit reasoning generator by finetuning a lightweight language model using knowledge distillation. This generator is guided by our sentence transformer to distill ground-truth reasoning into semantically aligned implicit reasoning, while also optimizing for accuracy. SemCoT is the first approach that enhances CoT efficiency by jointly optimizing token-level generation speed and preserving semantic alignment with ground-truth reasoning. Extensive experiments demonstrate the superior performance of SemCoT compared to state-of-the-art methods in both efficiency and effectiveness. Our code can be found at https://anonymous.4open.science/r/SemCoT.



Paperid:4623
Authors:Hancheng Min, Zhihui Zhu, Rene Vidal
Abstract:
Among many mysteries behind the success of deep networks lies the exceptional discriminative power of their learned representations as manifested by the intriguing Neural Collapse (NC) phenomenon, where simple feature structures emerge in the last layer of a trained neural network. Prior work on understanding NC theoretically has focused on analyzing the optimization landscape of matrix-factorization-like problems by considering the last-layer features as unconstrained free optimization variables and showing that their global minima exhibit NC. In this paper, we show that gradient flow on a two-layer ReLU network for classifying orthogonally separable data provably exhibits NC, thereby advancing prior results in two ways: First, we relax the assumption of unconstrained features, showing the effect of data structure and nonlinear activations on NC characterizations. Second, we reveal the role of the implicit bias of the training dynamics in facilitating the emergence of NC.



Paperid:4624
Authors:Arnav Kumar Jain, Vibhakar Mohta, Subin Kim, Atiksh Bhardwaj, Juntao Ren, Yunhai Feng, Sanjiban Choudhury, Gokul Swamy
Abstract:
Abstract:The fundamental limitation of the behavioral cloning (BC) approach to imitation learning is that it only teaches an agent what the expert did at states the expert visited. This means that when a BC agent makes a mistake which takes them out of the support of the demonstrations, they often don't know how to recover from it. In this sense, BC is akin to *giving the agent the fish* -- giving them dense supervision across a narrow set of states -- rather than teaching them *to fish*: to be able to reason independently about achieving the expert's outcome even when faced with unseen situations at test-time. In response, we explore *learning to search* (L2S) from expert demonstrations, i.e. learning the components required to, at test time, plan to match expert outcomes, even after making a mistake. These include *(1)* a world model and *(2)* a reward model. We carefully ablate the set of algorithmic and design decisions required to combine these and other components for stable and sample/interaction-efficient learning of recovery behavior without additional human corrections. Across a dozen visual manipulation tasks from three benchmarks, our approach $\texttt{SAILOR}$ consistently out-performs state-of-the-art Diffusion Policies trained via BC on the same data. Furthermore, scaling up the amount of demonstrations used for BC by 5-10$\times$ still leaves a performance gap. We find that $\texttt{SAILOR}$ can identify nuanced failures and is robust to reward hacking.



Paperid:4625
Authors:Ruicheng Ao, Jiashuo Jiang, David Simchi-Levi
Abstract:
Abstract:Dynamic pricing with resource constraints is a critical challenge in online learning, requiring a delicate balance between exploring unknown demand patterns and exploiting known information to maximize revenue. We propose three tailored algorithms to address this problem across varying levels of prior knowledge: (1) a Boundary Attracted Re-solve Method for the full information setting, achieving logarithmic regret without the restrictive non-degeneracy condition; (2) an online learning algorithm for the no information setting, delivering an optimal $O(\sqrt{T})$ regret; and (3) an estimate-then-select re-solve algorithm for the informed price setting, leveraging machine-learned prices with known error bounds to bridge the gap between full and no information scenarios. Moreover, through numerical experiments, we demonstrate the robustness and practical applicability of our approaches. This work advances dynamic pricing by offering scalable solutions that adapt to diverse informational contexts while relaxing classical assumptions.



Paperid:4626
Authors:Artur Back de Luca, George Giapitzakis, Kimon Fountoulakis
Abstract:
Neural networks are known for their ability to approximate smooth functions, yet they fail to generalize perfectly to unseen inputs when trained on discrete operations. Such operations lie at the heart of algorithmic tasks such as arithmetic, which is often used as a test bed for algorithmic execution in neural networks. In this work, we ask: can neural networks learn to execute binary-encoded algorithmic instructions exactly? We use the Neural Tangent Kernel (NTK) framework to study the training dynamics of two-layer fully connected networks in the infinite-width limit and show how a sufficiently large ensemble of such models can be trained to execute exactly, with high probability, four fundamental tasks: binary permutations, binary addition, binary multiplication, and Subtract and Branch if Negative (SBN) instructions. Since SBN is Turing-complete, our framework extends to computable functions. We show how this can be efficiently achieved using only logarithmically many training data. Our approach relies on two techniques: structuring the training data to isolate bit-level rules, and controlling correlations in the NTK regime to align model predictions with the target algorithmic executions.



Paperid:4627
Authors:Xiang Zhang, Jiaqi Wei, Zijie Qiu, Sheng Xu, Zhi Jin, ZhiQiang Gao, Nanqing Dong, Siqi Sun
Abstract:
Autoregressive (AT) models, common in sequence generation, are limited in biological tasks like \textit{de novo} peptide sequencing by their unidirectionality, failing to capture crucial global bidirectional dependencies. Non-Autoregressive Transformers (NATs) offer holistic, bidirectional processing but face challenges with generative coherence, scalability, and complex optimization.To transcend this, we propose a hybrid framework enhancing AT generation by dynamically integrating rich contextual information from non-autoregressive mechanisms. Our approach couples a shared spectrum encoder with two decoders: a non-autoregressive one learning latent bidirectional features, and an AT decoder synthesizing the sequence by leveraging these features. A novel cross-decoder attention module enables the AT decoder to iteratively query and integrate these bidirectional features, enriching its predictions. This synergy is cultivated via a tailored training strategy with importance annealing for balanced objectives and \cross-decoder gradient blocking for stable, focused learning.Evaluations on a demanding 9-species benchmark show our model substantially surpasses AT and NAT baselines. It uniquely harmonizes AT stability with NAT contextual awareness, delivering robust, superior performance on diverse proteomic data. This research advances de novo peptide sequencing and contributes a novel architectural paradigm for augmenting AT models with enhanced bidirectional understanding for complex sequence modeling.Our code is available for reproduction: https://anonymous.4open.science/r/CrossNovo-E263



Paperid:4628
Authors:Eray Erturk, Saba Hashemi, Maryam Shanechi
Abstract:
Local field potentials (LFPs) can be routinely recorded alongside spiking activity in intracortical neural experiments, measure a larger complementary spatiotemporal scale of brain activity for scientific inquiry, and can offer practical advantages over spikes, including greater long-term stability, robustness to electrode degradation, and lower power requirements. Despite these advantages, recent neural modeling frameworks have largely focused on spiking activity since LFP signals pose inherent modeling challenges due to their aggregate, population-level nature, often leading to lower predictive power for downstream task variables such as motor behavior. To address this challenge, we introduce a cross-modal knowledge distillation framework that transfers high-fidelity representational knowledge from pretrained multi-session spike transformer models to LFP transformer models. Specifically, we first train a teacher spike model across multiple recording sessions using a masked autoencoding objective with a session-specific neural tokenization strategy. We then align the latent representations of the student LFP model to those of the teacher spike model. Our results show that the distilled LFP models consistently outperform single- and multi-session LFP baselines in both fully unsupervised and supervised settings, and can generalize to other sessions without additional distillation while maintaining superior performance. These findings demonstrate that cross-modal knowledge distillation is a powerful and scalable approach for leveraging high-performing spike models to develop more accurate LFP models.



Paperid:4629
Authors:Xihang Yue, Yi Yang, Linchao Zhu
Abstract:
The limited availability of high-quality training data poses a major obstacle in data-driven PDE solving, where expensive data collection and resolution constraints severely impact the ability of neural operator networks to learn and generalize the underlying physical system. To address this challenge, we propose DeltaPhi, a novel learning framework that transforms the PDE solving task from learning direct input-output mappings to learning the residuals between similar physical states, a fundamentally different approach to neural operator learning. This reformulation provides implicit data augmentation by exploiting the inherent stability of physical systems where closer initial states lead to closer evolution trajectories. DeltaPhi is architecture-agnostic and can be seamlessly integrated with existing neural operators to enhance their performance. Extensive experiments demonstrate consistent and significant improvements across diverse physical systems including regular and irregular domains, different neural architectures, multiple training data amount, and cross-resolution scenarios, confirming its effectiveness as a general enhancement for neural operators in data-limited PDE solving.



Paperid:4630
Authors:Jens U. Brandt, Noah Pütz, Marcus Greiff, Thomas Lew, John Subosits, Marc Hilbert, Thomas Bartz-Beielstein
Abstract:
Machine learning models play a key role in safety-critical applications, such as autonomous vehicles and advanced driver assistance systems, where their robustness during inference is essential to ensure reliable operation. Sensor faults, however, can corrupt input signals, potentially leading to severe model failures that compromise reliability. In this context, pretraining emerges as a powerful approach for learning expressive representations applicable to various downstream tasks. Among existing techniques, masking represents a promising direction for learning representations that are robust to corrupted input data. In this work, we extend this concept by specifically targeting robustness to sensor outages during pretraining. We propose a self-supervised masking scheme that simulates common sensor failures and explicitly trains the model to recover the original signal. We demonstrate that the resulting representations significantly improve the robustness of predictions to seen and unseen sensor failures on a vehicle dynamics dataset, maintaining strong downstream performance under both nominal and various fault conditions. As a practical application, we deploy the method on a modified Lexus LC 500 and show that the pretrained model successfully operates as a substitute for a physical sensor in a closed-loop control system. In this autonomous racing application, we show that, under sensor failures at the wheelspeed sensors, a standard supervised baseline trained without sensor failures may cause the vehicle to leave the track, while our pretrained model enables reliable racing performance.



Paperid:4631
Authors:Felix Wagner, Pramit Saha, Harry Anthony, Alison Noble, Konstantinos Kamnitsas
Abstract:
Safe deployment of machine learning (ML) models in safety-critical domains such as medical imaging requires detecting inputs with characteristics not seen during training, known as out-of-distribution (OOD) detection, to prevent unreliable predictions. Effective OOD detection after deployment could benefit from access to the training data, enabling direct comparison between test samples and the training data distribution to identify differences. State-of-the-art OOD detection methods, however, either discard the training data after deployment or assume that test samples and training data are centrally stored together, an assumption that rarely holds in real-word settings. This is because shipping the training data with the deployed model is usually impossible due to the size of training databases, as well as proprietary or privacy constraints. We introduce the Isolation Network, an OOD detection framework that quantifies the difficulty of separating a target test sample from the training data by solving a binary classification task. We then propose Decentralized Isolation Networks (DIsoN), which enables the comparison of training and test data when data-sharing is impossible, by exchanging only model parameters between the remote computational nodes of training and deployment. We further extend DIsoN with class-conditioning, comparing a target sample solely with training data of its predicted class. We evaluate DIsoN on four medical imaging datasets (dermatology, chest X-ray, breast ultrasound, histopathology) across 12 OOD detection tasks. DIsoN performs favorably against existing methods while respecting data-privacy. This decentralized OOD detection framework opens the way for a new type of service that ML developers could provide along with their models: providing remote, secure utilization of their training data for OOD detection services. Code will be made available at:**



Paperid:4632
Authors:Weining Ren, Hongjun Wang, Xiao Tan, Kai Han
Abstract:
We present Fin3R, a simple, effective, and general fine-tuning method for feed-forward 3D reconstruction networks. This family of models regresses the pointmap of all input images under the reference frame coordinate within a single forward pass. However, current models struggle with fine geometry due to (i) the scarcity of high-fidelity depth-and-pose supervision and (ii) the inherent geometric misalignment from multi-view pointmap regression. Fin3R tackles both issues with an extra lightweight fine-tuning step. We freeze the decoder that handles view matching and fine-tune only the image encoder, the component dedicated to feature extraction. The encoder is enriched with fine geometry distilled from a strong monocular teacher on large unlabeled datasets using a custom, lightweight LoRA adapter. We test our method on three representative models—DUSt3R (two-view), CUT3R (sequential multi-view), and VGGT (parallel multi-view)—and fine-tuned models deliver sharper boundaries, recover complex structures, and achieve higher geometry accuracy in both single- and multi-view settings. At test time, Fin3R adds only the tiny LoRA weights, leaving memory and latency virtually unchanged.



Paperid:4633
Authors:Wanlong Liu, Junxiao Xu, Fei Yu, Yukang Lin, Ke Ji, Wenyu Chen, Lifeng Shang, Yasheng Wang, Yan Xu, Benyou Wang
Abstract:
Recent advancements in long chain-of-thought (CoT) reasoning models have improved performance on complex tasks, but they suffer from overthinking, which generates redundant reasoning steps, especially for simple problems. This paper revisits the reasoning patterns of long and short CoT models, observing that the short CoT patterns offer concise reasoning efficiently, while the long CoT patterns excel in challenging scenarios where the short CoT patterns struggles. To enable models to leverage both patterns, we propose Question-Free Fine-Tuning (QFFT), a distillation approach that removes the input question during training and learns exclusively from reasoning chains. This approach allows the model to adaptively employ both reasoning patterns: it prioritizes the short CoT patterns and activates the long CoT patterns only when necessary. Experiments on various mathematical datasets demonstrate that QFFT reduces average response length by more than 50\%, while achieving performance comparable to Supervised Fine-Tuning (SFT). Additionally, QFFT achieves superior generalization on out-of-domain tasks.



Paperid:4634
Authors:Arya Honarpisheh, Mustafa Bozdag, Octavia Camps, Mario Sznaier
Abstract:
State-space models (SSMs) have recently emerged as a compelling alternative to Transformers for sequence modeling tasks. This paper presents a theoretical generalization analysis of selective SSMs, the core architectural component behind the Mamba model. We derive a novel covering number-based generalization bound for selective SSMs, building upon recent theoretical advances in the analysis of Transformer models. Using this result, we analyze how the spectral abscissa of the continuous-time state matrix governs the model’s training dynamics and its ability to generalize across sequence lengths. We empirically validate our findings on a synthetic majority task and the IMDb sentiment classification benchmark, illustrating how our theoretical insights translate into practical model behavior.



Paperid:4635
Authors:Quang Anh Pham, Janaka Brahmanage, Tien Mai, Akshat Kumar
Abstract:
Abstract:Offline Learning from Observations (LfO) focuses on enabling agents to imitate expert behavior using datasets that contain only expert state trajectories and separate transition data with suboptimal actions. This setting is both practical and critical in real-world scenarios where direct environment interaction or access to expert action labels is costly, risky, or infeasible. Most existing LfO methods attempt to solve this problem through state or state-action occupancy matching. They typically rely on pretraining a discriminator to differentiate between expert and non-expert states, which could introduce errors and instability—especially when the discriminator is poorly trained. While recent discriminator-free methods have emerged, they generally require substantially more data, limiting their practicality in low-data regimes. In this paper, we propose IOSTOM ($\textit{Imitation from Observation via State Transition Occupancy Matching}$), a novel offline LfO algorithm designed to overcome these limitations. Our approach formulates a learning objective based on the joint state visitation distribution. A key distinction of IOSTOM is that it first excludes actions entirely from the training objective. Instead, we learn an $\textit{implicit policy}$ that models transition probabilities between states, resulting in a more compact and stable optimization problem. To recover the expert policy, we introduce an efficient action inference mechanism that $\textit{avoids training an inverse dynamics model}$. Extensive empirical evaluations across diverse offline LfO benchmarks show that IOSTOM substantially outperforms state-of-the-art methods, demonstrating both improved performance and data efficiency.



Paperid:4636
Authors:Hao Gao, Shaoyu Chen, Bo Jiang, Bencheng Liao, Yiang Shi, Xiaoyang Guo, Yuechuan Pu, haoran yin, Xiangyu Li, xinbang zhang, ying zhang, Wenyu Liu, Qian Zhang, Xinggang Wang
Abstract:
Existing end-to-end autonomous driving (AD) algorithms typically follow the Imitation Learning (IL) paradigm, which faces challenges such as causal confusion and an open-loop gap. In this work, we establish a 3DGS-based closed-loop Reinforcement Learning (RL) training paradigm. By leveraging 3DGS techniques, we construct a photorealistic digital replica of the real physical world, enabling the AD policy to extensively explore the state space and learn to handle out-of-distribution scenarios through large-scale trial and error. To enhance safety, we design specialized rewards to guide the policy in effectively responding to safety-critical events and understanding real-world causal relationships. To better align with human driving behavior, we incorporate IL into RL training as a regularization term. We introduce a closed-loop evaluation benchmark consisting of diverse, previously unseen 3DGS environments. Compared to IL-based methods, RAD achieves stronger performance in most closed-loop metrics, particularly exhibiting a 3x lower collision rate. Abundant closed-loop results are presented in the supplementary material. Code will be released to facilitate future research.



Paperid:4637
Authors:Zhenwen Liang, Linfeng Song, Yang Li, TAO YANG, Haitao Mi, Dong Yu
Abstract:
Automated Theorem Proving (ATP) in formal languages remains a formidable challenge in AI, demanding rigorous logical deduction and navigating vast search spaces. While large language models (LLMs) have shown promising performance, existing stepwise provers often suffer from biased search guidance, leading to inefficiencies and suboptimal proof strategies. This paper introduces the Multi-Perspective Search Prover (MPS-Prover), a novel stepwise ATP system designed to overcome these limitations. MPS-Prover incorporates two key innovations: a highly effective post-training data curation strategy that prunes approximately 40\% of redundant training data without sacrificing performance, and a multi-perspective tree search mechanism. This search integrates a learned critic model with strategically designed heuristic rules to diversify tactic selection, prevent getting trapped in unproductive states, and enhance search robustness. Extensive evaluations demonstrate that MPS-Prover achieves state-of-the-art performance on multiple challenging benchmarks, including miniF2F and ProofNet, outperforming prior 7B parameter models. Furthermore, our analyses reveal that MPS-Prover generates significantly shorter and more diverse proofs compared to existing stepwise and whole-proof methods, highlighting its efficiency and efficacy. Our work advances the capabilities of LLM-based formal reasoning and offers a robust framework and a comprehensive analysis for developing more powerful theorem provers.



Paperid:4638
Authors:YUNQING LIU, Nan Zhang, Fangjun Wang, Kengo Murata, Takuma Yamamoto, Osafumi Nakayama, Genta Suzuki, Zhiming Tan
Abstract:
Mamba, a lightweight sequence modeling framework offering near-liner complexity, presents a promising alternative to Transformers. In this work, we introduce MOGO (Mamba Only Glances Once), an end-to-end framework for efficient video action detection built entirely on the Mamba architecture. In MOGO, our newly designed Mamba-based decoder can even use just one Mamba layer to effectively perform action detection. It uses neither Transformer structures nor RCNN-like methods for proposal detection. Our framework introduces two key innovations. First, we propose a pure Mamba-based Encoder-Decoder architecture. The encoder processes cross-frame video information, while the decoder incorporates two novel Mamba-based structures that leverage Mamba’s intrinsic capabilities to detect actions. Theoretical analysis and ablation experiments confirm their synergy and the necessity of each structure. Second, we design a video token construction mechanism to improve the model's performance. The token importance block can ensure that the retained token information is highly relevant to the predicted targets. These two innovations make MOGO both efficient and accurate, as demonstrated on the JHMDB and UCF101-24 benchmark datasets. Compared to SOTA action detection methods, MOGO achieves superior performance in terms of GFLOPs, model parameters, and inference speed (latency) with comparable detection precision. Additionally, it requires significantly less GPU memory than some token reconstruction methods.



Paperid:4639
Authors:Yifan Pu, JIXUAN YING, Tianzhu Ye, Dongchen Han, Ziyi Wang, Qixiu Li, shao xinyu, Xiaochen Wang, Gao Huang, Xiu Li
Abstract:
Abstract:Vision Transformers (ViTs) have become a universal backbone for both image recognition and image generation. Yet their Multi–Head Self–Attention (MHSA) layer still performs a quadratic query–key interaction for \emph{every} token pair, spending the bulk of computation on visually weak or redundant correlations. We introduce \emph{Visual–Contrast Attention} (VCA), a drop-in replacement for MHSA that injects an explicit notion of discrimination while reducing the theoretical complexity from $\mathcal{O}(N^{2}C)$ to $\mathcal{O}(N n C)$ with $n\!\ll\!N$. VCA first distils each head’s dense query field into a handful of spatially pooled \emph{visual–contrast tokens}, then splits them into a learnable \emph{positive} and \emph{negative} stream whose differential interaction highlights what truly separates one region from another. The module adds fewer than $0.3$\,M parameters to a DeiT-Tiny backbone, requires no extra FLOPs, and is wholly architecture-agnostic. Empirically, VCA lifts DeiT-Tiny top-1 accuracy on ImageNet-1K from $72.2\%$ to \textbf{$75.5\%$} (+$3.3$) and improves three strong hierarchical ViTs by up to $2.9$\%, while in class-conditional ImageNet generation it lowers FID-50K by $2$ to $7.5$ points across both diffusion (DiT) and flow (SiT) models. Extensive ablations confirm that (i) spatial pooling supplies low-variance global cues, (ii) dual positional embeddings are indispensable for contrastive reasoning, and (iii) combining the two in both stages yields the strongest synergy. VCA therefore offers a simple path towards faster and sharper Vision Transformers.



Paperid:4640
Authors:Matthew Zurek, Guy Zamir, Yudong Chen
Abstract:
We study offline reinforcement learning in average-reward MDPs, which presents increased challenges from the perspectives of distribution shift and non-uniform coverage, and has been relatively underexamined from a theoretical perspective. While previous work obtains performance guarantees under single-policy data coverage assumptions, such guarantees utilize additional complexity measures which are uniform over all policies, such as the uniform mixing time. We develop sharp guarantees depending only on the target policy, specifically the bias span and a novel policy hitting radius, yielding the first fully single-policy sample complexity bound for average-reward offline RL. We are also the first to handle general weakly communicating MDPs, contrasting restrictive structural assumptions made in prior work. To achieve this, we introduce an algorithm based on pessimistic discounted value iteration enhanced by a novel quantile clipping technique, which enables the use of a sharper empirical-span-based penalty function. Our algorithm also does not require any prior parameter knowledge for its implementation. Remarkably, we show via hard examples that learning under our conditions requires coverage assumptions beyond the stationary distribution of the target policy, distinguishing single-policy complexity measures from previously examined cases. We also develop lower bounds nearly matching our main result.



Paperid:4641
Authors:蕾蕾 温, Liwei Zheng, Hongda Li, Lijun Sun, Zhihua Wei, Wen Shen
Abstract:
In recent years, large language models (LLMs) have made significant advancements in arithmetic reasoning. However, the internal mechanism of how LLMs solve arithmetic problems remains unclear.In this paper, we propose explaining arithmetic reasoning in LLMs using game-theoretic interactions.Specifically, we disentangle the output score of the LLM into numerous interactions between the input words.We quantify different types of interactions encoded by LLMs during forward propagation to explore the internal mechanism of LLMs for solving arithmetic problems.We find that (1) the internal mechanism of LLMs for solving simple one-operator arithmetic problems is their capability to encode operand-operator interactions and high-order interactions from input samples.Additionally, we find that LLMs with weak one-operator arithmetic capabilities focus more on background interactions.(2) The internal mechanism of LLMs for solving relatively complex two-operator arithmetic problems is their capability to encode operator interactions and operand interactions from input samples.(3) We explain the task-specific nature of the LoRA method from the perspective of interactions.



Paperid:4642
Authors:Zijian Dong, Li Ruilin, Joanna Chong, Niousha Dehestani, Yinghui Teng, Yi Lin, Zhizhou Li, Yichi Zhang, Yapei Xie, Leon Ooi, B.T. Yeo, Juan Helen Zhou
Abstract:
We presentBrain Harmony (BrainHarmonix), the first multimodal brain foundation model that unifies structural morphology and functional dynamics into compact 1D token representations. The model was pretrained on two of the largest neuroimaging datasets to date, encompassing 64,594 T1-weighted structural MRI 3D volumes (~14 million images) and 70,933 functional MRI (fMRI) time series. BrainHarmonix is grounded in two foundational neuroscience principles:structure complements function- structural and functional modalities offer distinct yet synergistic insights into brain organization;function follows structure- brain functional dynamics are shaped by cortical morphology. The modular pretraining process involves single-modality training with geometric pre-alignment followed by modality fusion through shared brain hub tokens. Notably, our dynamics encoder uniquely handles fMRI time series with heterogeneous repetition times (TRs), addressing a major limitation in existing models. BrainHarmonix is also the first to deeply compress high-dimensional neuroimaging signals into unified, continuous 1D tokens, forming a compact latent space of the human brain. BrainHarmonix achieves strong generalization across diverse downstream tasks, including neurodevelopmental and neurodegenerative disorder classification and cognition prediction - consistently outperforming previous approaches. Our models - pretrained on 8 H100 GPUs - will be made publicly available, aiming to catalyze a new era of AI-driven neuroscience powered by large-scale multimodal neuroimaging.



Paperid:4643
Authors:Jean-Baptiste Fermanian, Mohamed Hebiri, Joseph Salmon
Abstract:
Conformal prediction methods are statistical tools designed to quantify uncertainty and generate predictive sets with guaranteed coverage probabilities. This work introduces an innovative refinement to these methods, specifically tailored for scenarios where multiple predictive observations of a single instance are available. Our approach is particularly motivated by applications in citizen science, where multiple images of the same plant or animal are captured by individuals. Our method integrates the information from each observation into conformal prediction, enabling a reduction in the size of the predicted label set while preserving the required conditional coverage guarantee. The approach is based on the aggregation of conformal p-values computed from each view. By exploiting the exact distribution of these p-values, we refine standard aggregation strategies such as majority voting or template-based methods, and we provide a general framework for adapting our approach to other contexts. We evaluate our method on simulated and real data, with a particular focus on Pl@ntNet, a prominent citizen science platform that facilitates the collection and identification of plant species through user-submitted images.



Paperid:4644
Authors:Zidong Cao, Jinjing Zhu, Hao Ai, Lutao Jiang, Yuanhuiyi Lyu, Hui Xiong
Abstract:
360-degree monocular depth estimation plays a crucial role in scene understanding owing to its 180-degree by 360-degree field-of-view (FoV). To mitigate the distortions brought by equirectangular projection, existing methods typically divide 360-degree images into distortion-less perspective patches. However, since these patches are processed independently, depth inconsistencies are often introduced due to scale drift among patches. Recently, video depth estimation (VDE) models have leveraged temporal consistency for stable depth predictions across frames. Inspired by this, we propose to represent a 360-degree image as a sequence of perspective frames, mimicking the viewpoint adjustments users make when exploring a 360-degree scenario in virtual reality. Thus, the spatial consistency among perspective depth patches can be enhanced by exploiting the temporal consistency inherent in VDE models. To this end, we introduce a training-free pipeline for 360-degree monocular depth estimation, called ST²360D. Specifically, ST²360D transforms a 360-degree image into perspective video frames, predicts video depth maps using VDE models, and seamlessly merges these predictions into a complete 360-degree depth map. To generate sequenced perspective frames that align with VDE models, we propose two tailored strategies. First, a spherical-uniform sampling (SUS) strategy is proposed to facilitate uniform sampling of perspective views across the sphere, avoiding oversampling in polar regions typically with limited structural details. Second, a latitude-guided scanning (LGS) strategy is introduced to organize the frames into a coherent sequence, starting from the equator, prioritizing low-latitude slices, and progressively moving toward higher latitudes. Extensive experiments demonstrate that ST²360D achieves strong zero-shot capability on several datasets, supporting resolutions up to 4K.



Paperid:4645
Authors:Yang Chen, Menglin Zou, Jiaqi Zhang, Yitan Zhang, Junyi Yang, Gaël Gendron, Libo Zhang, Jiamou Liu, Michael Witbrock
Abstract:
Inverse Reinforcement Learning (IRL) aims to recover reward functions that explain expert behavior, offering interpretability and generalization beyond direct imitation. However, most IRL methods suffer from unstable reward learning due to adversarial reward-policy optimization. While recent non-adversarial methods offer improved empirical stability, a principled framework for stable reward learning remains elusive. This work surrounds a key insight: stability can be enforced by construction if each reward update provably reduces the divergence between the learned and expert policies. We formalize this idea through a general reward optimization objective and show that it admits a Majorization-Minimization (MM) procedure with guaranteed monotonic reduction in policy divergence. Building on this, we propose Proximal Inverse Reward Optimization (PIRO), a practical, non-adversarial IRL algorithm that approximates the theoretical guarantee. PIRO explicitly learns reward functions, scales to high-dimensional tasks, and has strong sample efficiency on expert data, while relying on on-policy rollouts for reward updates. Empirically, across OpenAI Gym tasks and a real-world animal behavior modeling task, PIRO substantially improves training stability while matching or exceeding the performance of state-of-the-art IRL methods. This paper demonstrates that reward-driven policy alignment can provide a basis for stable IRL.



Paperid:4646
Authors:Joshua Ashkinaze, Hua Shen, Saipranav Avula, Ceren Budak, Eric Gilbert
Abstract:
We introduce the Deep Value Benchmark (DVB), an evaluation framework that directly tests whether large language models (LLMs) learn fundamental human values or merely surface-level preferences. This distinction is critical for AI alignment: systems that capture deeper values are likely to generalize human intentions robustly, while those that only model statistical patterns in preference data may produce misaligned behavior. The DVB uses a novel experimental design with controlled confounding between deep values (e.g., moral principles) and shallow features (e.g., superficial attributes). In the training phase, we expose LLMs to human preference data with deliberately correlated deep and shallow features---for instance, where a user consistently prefers (formal language, non-maleficence) options over (informal language, justice) alternatives. The testing phase then breaks these correlations, presenting choices between (formal language, justice) and (informal language, non-maleficence) options. This design allows us to precisely measure a model's Deep Value Generalization Rate (DVGR)---the probability of generalizing based on the underlying value rather than the shallow feature. Overall, we find that across 9 different models the average DVGR is just 0.30 and all models generalize deep values less than chance. Bigger models have a (slightly) lower DVGR than smaller models. We are releasing our dataset (with three separate human validation experiments) so others can measure performance on future models and training paradigms. DVB is an interpretable measure of a core feature of alignment.



Paperid:4647
Authors:Chenchen Tan, Youyang Qu, Xinghao Li, Hui Zhang, Shujie Cui, Cunjian Chen, Longxiang Gao
Abstract:
The increase in computing power and the necessity of AI-assisted decision-making boost the growing application of large language models (LLMs). Along with this, the potential retention of sensitive data of LLMs has spurred increasing research into machine unlearning. However, existing unlearning approaches face a critical dilemma: Aggressive unlearning compromises model utility, while conservative strategies preserve utility but risk hallucinated responses. This significantly limits LLMs' reliability in knowledge-intensive applications. To address this, we introduce a novel Attention-Shifting (AS) framework for selective unlearning. AS is driven by two design objectives: (1) context-preserving suppression that attenuates attention to fact-bearing tokens without disrupting LLMs' linguistic structure; and (2) hallucination-resistant response shaping that discourages fabricated completions when queried about unlearning content. AS realizes these objectives through two attention-level interventions, which are importance-aware suppression applied to the unlearning set to reduce reliance on memorized knowledge and attention-guided retention enhancement that reinforces attention toward semantically essential tokens in the retained dataset to mitigate unintended degradation. These two components are jointly optimized via a dual-loss formulation. Experimental results show that AS improves performance preservation over the state-of-the-art unlearning methods, achieving up to 15\% higher accuracy on the ToFU benchmark and 10\% on the TDEC benchmark, while maintaining competitive hallucination-free unlearning effectiveness. Compared to existing methods, AS demonstrates a superior balance between unlearning effectiveness, generalization, and response reliability.



Paperid:4648
Authors:Qiao Li, Jie Li, Yukang Zhang, Lei Tan, Jing Chen, Jiayi Ji
Abstract:
Aerial-Ground person re-identification (AG-ReID) is an emerging yet challenging task that aims to match pedestrian images captured from drastically different viewpoints, typically from unmanned aerial vehicles (UAVs) and ground-based surveillance cameras. The task poses significant challenges due to extreme viewpoint discrepancies, occlusions, and domain gaps between aerial and ground imagery. While prior works have made progress by learning cross-view representations, they remain limited in handling severe pose variations and spatial misalignment. To address these issues, we propose a Geometric and Semantic Alignment Network (GSAlign) tailored for AG-ReID. GSAlign introduces two key components to jointly tackle geometric distortion and semantic misalignment in aerial-ground matching: a Learnable Thin Plate Spline (LTPS) Transformation Module and a Dynamic Alignment Module (DAM). The LTPS module adaptively warps pedestrian features based on a set of learned keypoints, effectively compensating for geometric variations caused by extreme viewpoint changes. In parallel, the DAM estimates visibility-aware representation masks that highlight visible body regions at the semantic level, thereby alleviating the negative impact of occlusions and partial observations in cross-view correspondence. Extensive experiments on the challenging CARGO benchmark demonstrate the effectiveness of GSAlign, achieving significant improvements of +18.8\% in mAP and +16.8\% in Rank-1 accuracy over previous state-of-the-art methods.



Paperid:4649
Authors:Yang Zhao, Kai Xiong, Xiao Ding, Li Du, Yangou Ouyang, Zhouhao Sun, Jiannan Guan, Wenbin Zhang, Bin Liu, Dong Hu, Bing Qin, Ting Liu
Abstract:
Abstract:A primary impediment to scaling reinforcement learning (RL) for large language model (LLM) training is the substantial computational cost, predominantly arising from the necessity of multi-sampling for policy optimization and evaluation. This underscores the critical yet challenging nature of efficient training data selection. Drawing inspiration from the Zone of Proximal Development (ZPD) theory, which posits that learners acquire knowledge more effectively from tasks of intermediate difficulty, we hypothesize that LLMs exhibit optimal learning from data they have not yet mastered but demonstrate the potential to comprehend. Conventional methodologies for assessing data difficulty or informativeness typically rely on computationally intensive multi-sampling or iterative procedures. To address this limitation, we introduce UFO-RL (**U**ncertainty-**F**ocused **O**ptimization for **R**einforcement **L**earning), a novel framework that employs a computationally efficient single-pass uncertainty estimation technique to identify informative training instances. This method, requiring only a single forward pass and obviating the need for iterative next-token computation, achieves a significant acceleration (up to 185$\times$) in data evaluation compared to multi-sampling approaches. UFO-RL leverages this efficient metric to select data within the model's estimated ZPD for training. Extensive experimentation across diverse LLMs and mathematical benchmarks demonstrates that training with a mere 10\% of the data, carefully selected by UFO-RL, yields performance comparable to or even surpassing that of full-data training. Furthermore, this targeted data selection results in up to a 16$\times$ reduction in overall training time, concurrently enhancing training stability and improving generalization capabilities. Thus, UFO-RL presents a practical and highly efficient strategy for scaling RL fine-tuning of LLMs by focusing learning efforts on the most informative and valuable data, thereby mitigating the computational bottlenecks associated with traditional RL training.



Paperid:4650
Authors:Zhongxing Xu, Chengzhi Liu, Qingyue Wei, Juncheng Wu, James Zou, Xin Wang, Yuyin Zhou, Sheng Liu
Abstract:
Test-time compute has empowered multimodal large language models to generate extended reasoning chains, yielding strong performance on tasks such as multimodal math reasoning. However, we observe that this improved reasoning ability often comes with increased hallucination: as generations become longer, models tend to drift away from image-grounded content and rely more on language priors. Attention analysis reveals that longer reasoning chains reduce focus on visual inputs, contributing to hallucination. To systematically study this phenomenon, we introduce RH-AUC, a metric that quantifies how a model's perception accuracy changes with reasoning length, enabling evaluation of whether the model preserves visual grounding while reasoning. We also release RH-Bench, a diagnostic benchmark covering diverse multimodal tasks, designed to jointly assess the balance of reasoning ability and hallucination. We find that (i) larger models generally exhibit a better balance between reasoning and perception; (ii) reasoning and perception balance depends more on the types and domains of the training data than its volume. Our findings highlight the need for evaluation frameworks that account for both reasoning quality and perceptual reliability.



Paperid:4651
Authors:Haixiang Lan, Luofeng Liao, Adam N. Elmachtoub, Christian Kroer, Henry Lam, Haofeng Zhang
Abstract:
Data-driven stochastic optimization is ubiquitous in machine learning and operational decision-making problems. Sample average approximation (SAA) and model-based approaches such as estimate-then-optimize (ETO) or integrated estimation-optimization (IEO) are all popular, with model-based approaches being able to circumvent some of the issues with SAA in complex context-dependent problems. Yet the relative performance of these methods is poorly understood, with most results confined to the dichotomous cases of the model-based approach being either well-specified or misspecified. We develop the first results that allow for a more granular analysis of the relative performance of these methods under a local misspecification setting, which models the scenario where the model-based approach is nearly well-specified. By leveraging tools from contiguity theory in statistics, we show that there is a bias-variance tradeoff between SAA, IEO, and ETO under local misspecification, and that the relative importance of the bias and the variance depends on the degree of local misspecification. Moreover, we derive explicit expressions for the decision bias, which allows us to characterize (un)impactful misspecification directions, and provide further geometric understanding of the variance.



Paperid:4652
Authors:Jiawei Zhang, Youmin Zhang, Fabio Tosi, Meiying Gu, Jiahe Li, Xiaohan Yu, Jin Zheng, Xiao Bai, Matteo Poggi
Abstract:
We present Eve3D, a novel framework for dense 3D reconstruction based on 3D Gaussian Splatting (3DGS). While most existing methods rely on imperfect priors derived from pre-trained vision models, Eve3D fully leverages these priors by jointly optimizing both them and the 3DGS backbone. This joint optimization creates a mutually reinforcing cycle: the priors enhance the quality of the 3DGS, which in turn refines the priors, further improving the reconstruction. Additionally, Eve3D introduces a novel optimization step based on bundle adjustment, overcoming the limitations of the highly local supervision typical in standard 3DGS pipelines.Eve3D achieves state-of-the-art results in surface reconstruction and novel view synthesis on the Tanks \& Temples, DTU, and Mip-NeRF360 datasets. while retaining fast convergence, highlighting an unprecedented trade-off between accuracy and speed.



Paperid:4653
Authors:Xirui Jin, Renbiao Jin, Boying Li, Danping Zou, Wenxian Yu
Abstract:
Three-dimensional Gaussian Splatting (3DGS) has recently emerged as an efficient representation for novel-view synthesis, achieving impressive visual quality. However, in scenes dominated by large and low-texture regions, common in indoor environments, the photometric loss used to optimize 3DGS yields ambiguous geometry and fails to recover high-fidelity 3D surfaces. To overcome this limitation, we introduce PlanarGS, a 3DGS-based framework tailored for indoor scene reconstruction. PlanarGS first employs a pretrained vision-language segmentation model to detect planar regions in images. It then refines these region proposals via cross-view fusion and integrates depth and surface-normal priors extracted from a pretrained 3D network. Finally, the 3D Gaussians are optimized with two additional terms: a planar-prior supervision term that enforces planar consistency, and a geometric-prior supervision term that steers the Gaussians toward the depth and normal cues. We have conducted extensive experiments on standard indoor benchmarks. The results show that PlanarGS reconstructs accurate and detailed 3D surfaces, consistently outperforming state-of-the-art methods by a large margin.



Paperid:4654
Authors:Kaixiang Chen, Pengfei Fang, hui xue
Abstract:
This paper firstly addresses the challenge of few-shot universal cross-domain retrieval (FS-UCDR), enabling machines trained with limited data to generalize to novel retrieval scenarios, with queries from entirely unknown domains and categories. To achieve this, we first formally define the FS-UCDR task and propose the Multi-Modal Interactive Agent Layer (MAIL), which enhances the cross-modal interaction in vision-language models (VLMs) by aligning the parameter updates of target layer pairs across modalities.Specifically, MAIL freezes the selected target layer pair and introduces a trainable agent layer pair to approximate localized parameter updates. A bridge function is then introduced to couple the agent layer pair, enabling gradient communication across modalities to facilitate update alignment. The proposed MAIL offers four key advantages: 1) its cross-modal interaction mechanism improves knowledge acquisition from limited data, making it highly effective in low-data scenarios; 2) during inference, MAIL integrates seamlessly into the VLM via reparameterization, preserving inference complexity; 3) extensive experiments validate the superiority of MAIL, which achieves substantial performance gains over data-efficient UCDR methods while requiring significantly fewer training samples; 4) beyond UCDR, MAIL also performs competitively on few-shot classification tasks, underscoring its strong generalization ability.



Paperid:4655
Authors:Tianhao Chen, Xin Xu, Zijing Liu, Pengxiang Li, Xinyuan Song, AJAY JAISWAL, Fan Zhang, Jishan Hu, Yang Wang, Hao CHEN, Shizhe Diao, Shiwei Liu, Yu Li, Lu Yin, Can Yang
Abstract:
Modern Large Language Models, such as the LLaMA, Qwen and DeepSeek series, predominantly adopt the Pre-LayerNorm (Pre-LN) transformer architecture. While Pre-LN is stable during pretraining and scales well to large model sizes, it can suffer from an exponential growth in activation variance across layers. A straightforward analysis of the layer-wise Jacobian, supported by empirical evidence, shows that this phenomenon causes deeper layers to contribute less to learning, thereby slowing the convergence speed and limiting model performance. To mitigate this issue, we propose Gradient-Preserving Activation Scaling (GPAS), a simple technique that can be used in combination with existing approaches. GPAS works by scaling down the intermediate activations while keeping their gradients unchanged. This leaves information in the activations intact, and avoids the gradient vanishing problem associated with gradient downscaling. Extensive experiments across various model sizes from 71M to 1B show that GPAS achieves consistent performance gain. Beyond enhancing Pre-LN transformers, GPAS also shows promise in improving alternative architectures such as Sandwich-LN and Post-LN, demonstrating its versatility and potential for improving training dynamics in a wide range of settings.



Paperid:4656
Authors:Lorenzo Magnino, Kai Shao, Zida Wu, Jiacheng Shen, Mathieu Lauriere
Abstract:
Mean field games (MFGs) have emerged as a powerful framework for modeling interactions in large-scale multi-agent systems. Despite recent advancements in reinforcement learning (RL) for MFGs, existing methods are typically limited to finite spaces or stationary models, hindering their applicability to real-world problems. This paper introduces a novel deep reinforcement learning (DRL) algorithm specifically designed for non-stationary continuous MFGs. The proposed approach builds upon a Fictitious Play (FP) methodology, leveraging DRL for best-response computation and supervised learning for average policy representation. Furthermore, it learns a representation of the time-dependent population distribution using a Conditional Normalizing Flow. To validate the effectiveness of our method, we evaluate it on three different examples of increasing complexity. By addressing critical limitations in scalability and density approximation, this work represents a significant advancement in applying DRL techniques to complex MFG problems, bringing the field closer to real-world multi-agent systems.



Paperid:4657
Authors:Jaihoon Kim, Taehoon Yoon, Jisung Hwang, Minhyuk Sung
Abstract:
We propose an inference-time scaling approach for pretrained flow models. Recently, inference-time scaling has gained significant attention in LLMs and diffusion models, improving sample quality or better aligning outputs with user preferences by leveraging additional computation. For diffusion models, particle sampling has allowed more efficient scaling due to the stochasticity at intermediate denoising steps. On the contrary, while flow models have gained popularity as an alternative to diffusion models--offering faster generation and high-quality outputs--efficient inference-time scaling methods used for diffusion models cannot be directly applied due to their deterministic generative process. To enable efficient inference-time scaling for flow models, we propose three key ideas: 1) SDE-based generation, enabling particle sampling in flow models, 2) Interpolant conversion, broadening the search space and enhancing sample diversity, and 3) Rollover Budget Forcing (RBF), an adaptive allocation of computational resources across timesteps to maximize budget utilization. Our experiments show that SDE-based generation and variance-preserving (VP) interpolant-based generation, improves the performance of particle sampling methods for inference-time scaling in flow models. Additionally, we demonstrate that RBF with VP-SDE achieves the best performance, outperforming all previous inference-time scaling approaches.



Paperid:4658
Authors:Nguyen Do, Bach Ngo, Youval Kashuv, Canh Pham, Hanghang Tong, My T. Thai
Abstract:
We study the Quality of Service Degradation (QoSD) problem, in which an adversary perturbs edge weights to degrade network performance. This setting arises in both network infrastructures and distributed ML systems, where communication quality, not just connectivity, determines functionality. While classical methods rely on combinatorial optimization, and recent ML approaches address only restricted linear variants with small-size networks, no prior model directly tackles the QoSD problem under nonlinear edge-weight functions. This work proposes Hephaestus, a self-reinforcing generative framework that synthesizes feasible solutions in latent space, to fill this gap. Our method includes three phases: (1) Forge: a Predictive Path-Stressing (PPS) algorithm that uses graph learning and approximation to produce feasible solutions with performance guarantee, (2) Morph: a new theoretically grounded training paradigm for Mixture of Conditional VAEs guided by an energy-based model to capture solution feature distributions, and (3) Refine: a reinforcement learning agent that explores this space to generate progressively near-optimal solutions using our designed differentiable reward function. Experiments on both synthetic and real-world networks show that our approach consistently outperforms classical and ML baselines, particularly in scenarios with nonlinear cost functions where traditional methods fail to generalize.



Paperid:4659
Authors:Hongyuan Dong, Dingkang Yang, LiangXiao, ChaoFeng, Ran Jiao
Abstract:
Learning rate is widely regarded as crucial for effective foundation model pretraining.Recent research explores and demonstrates the transferability of learning rate configurations across varying model and dataset sizes, etc. Nevertheless, these approaches are constrained to specific training scenarios and typically necessitate extensive hyperparameter tuning on proxy models.In this work, we proposeAdaLRS, a plug-in-and-play adaptive learning rate search algorithm that conducts online optimal learning rate search via optimizing loss descent velocities.We provide experiment results to show that the optimization of training loss and loss descent velocity in foundation model pretraining are both convex and share the same optimal learning rate. Relying solely on training loss dynamics, AdaLRS involves few extra computations to guide the search process, and its convergence is guaranteed via theoretical analysis. Experiments on both LLM and VLM pretraining show that AdaLRS adjusts suboptimal learning rates to the neighborhood of optimum with marked efficiency and effectiveness, with model performance improved accordingly. We also show the robust generalizability of AdaLRS across varying training scenarios, such as different model sizes, training paradigms, base learning rate scheduler choices, etc.



Paperid:4660
Authors:Leheng Cai, Qirui Hu, Juntao Sun, Shuyuan Wu
Abstract:
Abstract:In this paper, we develop a novel algorithm for constructing time-uniform, asymptotic confidence sequences for quantiles under local differential privacy (LDP). The procedure combines dynamically chained parallel stochastic gradient descent (P-SGD) with a randomized response mechanism, thereby guaranteeing privacy protection while simultaneously estimating the target quantile and its variance. A strong Gaussian approximation for the proposed estimator yields asymptotically anytime-valid confidence sequences whose widths obey the law of the iterated logarithm (LIL). Moreover, the method is fully online, offering high computational efficiency and requiring only $\mathcal{O}(\kappa)$ memory, where $\kappa$ denotes the number of chains and is much smaller than the sample size. Rigorous mathematical proofs and extensive numerical experiments demonstrate the theoretical soundness and practical effectiveness of the algorithm.



Paperid:4661
Authors:Quntian Fang, Zhen Huang, Zhiliang Tian, Minghao Hu, Dongsheng Li, Yiping Yao, Xinyue Fang, Menglong Lu, Guotong Geng
Abstract:
Large language models (LLMs) are now pivotal in real-world applications. Model editing has emerged as a promising paradigm for efficiently modifying LLMs without full retraining. However, current editing approaches face significant limitations due to parameter drift, which stems from inconsistencies between newly edited knowledge and the model's existing knowledge. In sequential editing scenarios, cumulative drifts progressively lead to model collapse characterized by general capability degradation and balance between acquiring new knowledge and catastrophic forgetting of existing knowledge. Drawing inspiration from the hippocampal trisynaptic circuit for continual memorizing and forgetting, we propose a Hippocampal-like Sequential Editing (HSE) framework that designs the unlearning of obsolete knowledge, domain-specific knowledge update separation and replay for edited knowledge. Specifically, the HSE framework designs three core mechanisms: (1) Machine unlearning selectively erases outdated knowledge to facilitate integration of new information, (2) Fisher Information Matrix-guided parameter updates prevents cross-domain knowledge interference, and (3) Parameter replay consolidates long-term editing memory through lightweight and global replay of editing data in a parametric form. Theoretical analysis demonstrates that HSE achieves smaller generalization error bounds, more stable convergence and higher computational efficiency. Experimental results validate its effective balance between acquiring new knowledge and mitigating catastrophic forgetting, maintaining or even slightly enhancing general capabilities. In practical applications, experiments confirm its effectiveness in multi-domain hallucination mitigation, healthcare knowledge injecting, and societal bias reduction.



Paperid:4662
Authors:Hyungki Im, Wyame Benslimane, Paul Grigas
Abstract:
We study an extension of contextual stochastic linear optimization (CSLO) that, in contrast to most of the existing literature, involves inequality constraints that depend on uncertain parameters predicted by a machine learning model. To handle the constraint uncertainty, we use contextual uncertainty sets constructed via methods like conformal prediction. Given a contextual uncertainty set method, we introduce the "Smart Predict-then-Optimize with Robust Constraints" (SPO-RC) loss, a feasibility-sensitive adaptation of the SPO loss that measures decision error of predicted objective parameters. We also introduce a convex surrogate, SPO-RC+, and prove Fisher consistency with SPO-RC. To enhance performance, we train on truncated datasets where true constraint parameters lie within the uncertainty sets, and we correct the induced sample selection bias using importance reweighting techniques. Through experiments on fractional knapsack and alloy production problem instances, we demonstrate that SPO-RC+ effectively handles uncertainty in constraints and that combining truncation with importance reweighting can further improve performance.



Paperid:4663
Authors:Guilin Zhu, Runmin Wang, Yuanjie Shao, Wei dong Yang, Nong Sang, Changxin Gao
Abstract:
Class incremental semantic segmentation (CISS) enables a model to continually segment new classes from non-stationary data while preserving previously learned knowledge. Recent top-performing approaches are prototype-based methods that assign a prototype to each learned class to reproduce previous knowledge. However, modeling each class distribution relying on only a single prototype, which remains fixed throughout the incremental process, presents two key limitations: (i) a single prototype is insufficient to accurately represent the complete class distribution when incoming data stream for a class is naturally multimodal; (ii) the features of old classes may exhibit anisotropy during the incremental process, preventing fixed prototypes from faithfully reproducing the matched distribution. To address the aforementioned limitations, we propose a Continual Gaussian Mixture Distribution (CoGaMiD) modeling method. Specifically, the means and covariance matrices of the Gaussian Mixture Models (GMMs) are estimated to model the complete feature distributions of learned classes. These GMMs are stored to generate pseudo-features that support the learning of novel classes in incremental steps. Moreover, we introduce a Dynamic Adjustment (DA) strategy that utilizes the features of previous classes within incoming data streams to update the stored GMMs. This adaptive update mitigates the mismatch between fixed GMMs and continually evolving distributions. Furthermore, a Gaussian-based Representation Constraint (GRC) loss is proposed to enhance the discriminability of new classes, avoiding confusion between new and old classes. Extensive experiments on Pascal VOC and ADE20K show that our method achieves superior performance compared to previous methods, especially in more challenging long-term incremental scenarios.



Paperid:4664
Authors:Kehan Long, Jorge Cortes, Nikolay Atanasov
Abstract:
We study the problem of certifying the stability of closed-loop systems under control policies derived from optimal control or reinforcement learning (RL). Classical Lyapunov methods require a strict step-wise decrease in the Lyapunov function but such a certificate is difficult to construct for a learned control policy. The value function associated with an RL policy is a natural Lyapunov function candidate but it is not clear how it should be modified. To gain intuition, we first study the linear quadratic regulator (LQR) problem and make two key observations. First, a Lyapunov function can be obtained from the value function of an LQR policy by augmenting it with a residual term related to the system dynamics and stage cost. Second, the classical Lyapunov decrease requirement can be relaxed to a generalized Lyapunov condition requiring only decrease on average over multiple time steps. Using this intuition, we consider the nonlinear setting and formulate an approach to learn generalized Lyapunov functions by augmenting RL value functions with neural network residual terms. Our approach successfully certifies the stability of RL policies trained on Gymnasium and DeepMind Control benchmarks. We also extend our method to jointly train neural controllers and stability certificates using a multi-step Lyapunov loss, resulting in larger certified inner approximations of the region of attraction compared to the classical Lyapunov approach. Overall, our formulation enables stability certification for a broad class of systems with learned policies by making certificates easier to construct, thereby bridging classical control theory and modern learning-based methods.



Paperid:4665
Authors:Wenjun Cui, Xuhao Li, Yidong Li, Yuxin Ma, Qiyu Kang, Xueyang Fu, Wee Peng Tay, Zheng-Jun Zha
Abstract:
The integration of differential equations with neural networks has created powerful tools for modeling complex dynamics effectively across diverse machine learning applications. While standard integer-order neural ordinary differential equations (ODEs) have shown considerable success, they are limited in their capacity to model systems with memory effects and historical dependencies. Fractional calculus offers a mathematical framework capable of addressing this limitation, yet most current fractional neural networks use static memory weightings that cannot adapt to input-specific contextual requirements.This paper proposes a generalized Kernel-ATtention Fractional neural ODE network (KatFDE), which combines the memory-retention capabilities of fractional calculus with contextual learnable attention mechanisms. Our approach replaces fixed kernel functions in fractional operators with neural attention kernels that adaptively weight historical states based on their contextual relevance to current predictions. This allows our framework to selectively emphasize important temporal dependencies while filtering less relevant historical information. Our theoretical analysis establishes solution boundedness, problem well-posedness, and numerical equation solver convergence properties of the proposed model.Furthermore, through extensive evaluation on graph learning problems and spatio-temporal traffic flow forecasting tasks, we demonstrate that our adaptive attention-based fractional framework outperforms both integer-order neural ODE models and existing fractional approaches. The results confirm that our framework provides superior modeling capacity for complex dynamics with varying temporal dependencies.



Paperid:4666
Authors:Yonghao Liu, Yajun Wang, Chunli Guo, Wei Pang, Ximing Li, Fausto Giunchiglia, Xiaoyue Feng, Renchu Guan
Abstract:
Graph few-shot learning has attracted increasing attention due to its ability to rapidly adapt models to new tasks with only limited labeled nodes. Despite the remarkable progress made by existing graph few-shot learning methods, several key limitations remain. First, most current approaches rely on predefined and unified graph filters (e.g., low-pass or high-pass filters) to globally enhance or suppress node frequency signals. Such fixed spectral operations fail to account for the heterogeneity of local topological structures inherent in real-world graphs. Moreover, these methods often assume that the support and query sets are drawn from the same distribution. However, under few-shot conditions, the limited labeled data in the support set may not sufficiently capture the complex distribution of the query set, leading to suboptimal generalization. To address these challenges, we propose GRACE, a novel Graph few-shot leaRning framework that integrates Adaptive spectrum experts with Cross-sEt distribution calibration techniques. Theoretically, the proposed approach enhances model generalization by adapting to both local structural variations and cross-set distribution calibration. Empirically, GRACE consistently outperforms state-of-the-art baselines across a wide range of experimental settings. Our anonymous code can be found here.



Paperid:4667
Authors:Hangwei Zhang, Chun Kang, Yan Wang, Difan Zou
Abstract:
Parameter-efficient fine-tuning (PEFT) powerful pre-trained models for complex downstream tasks has proven effective in vision and language processing, yet this paradigm remains unexplored in scientific machine learning, where the objective is to model complex physical systems. We conduct the first systematic study of PEFT for pre-trained Large Operator Models (LOMs) obtained by scaling variants of Fourier Neural Operator. We observe that the widely used Low-Rank Adaptation (LoRA) yields markedly poorer performance on LOMs than Adapter tuning. We further theoretically establish that stacked LoRA incurs a depth-amplified lower bound on approximation error within Fourier layers, whereas adapters retain universal approximation capacity and, by concentrating parameters on energy-dominant low-frequency modes, attain exponentially decaying error with bottleneck width in the Fourier domain. Motivated by the robust empirical gains of adapters and by our theoretical characterization of PDE solutions as spectrally sparse, we introduce Frequency-Adaptive Adapter (F-Adapter). F-Adapter allocates adapter capacity based on spectral complexity, assigning higher-dimension modules to low-frequency components and lower-dimension modules to high-frequency components. Our F-Adapters establish state-of-the-art results on multiple challenging 3D Navier–Stokes benchmarks, markedly enhancing both generalization and spectral fidelity over LoRA and other PEFT techniques commonly used in LLMs. To the best of our knowledge, this work is the first to explore PEFT for scientific machine-learning and establishes F-Adapter as an effective paradigm for this domain. The code will be made publicly available upon acceptance.



Paperid:4668
Authors:Etienne Boursier, Scott Pesme, Radu-Alexandru Dragomir
Abstract:
Abstract:We study the dynamics of gradient flow with small weight decay on general training losses $F: \mathbb{R}^d \to \mathbb{R}$. Under mild regularity assumptions and assuming convergence of the unregularised gradient flow, we show that the trajectory with weight decay $\lambda$ exhibits a two-phase behaviour as $\lambda \to 0$. During the initial fast phase, the trajectory follows the unregularised gradient flow and converges to a manifold of critical points of $F$. Then, at time of order $1/\lambda$, the trajectory enters a slow drift phase and follows a Riemannian gradient flow minimising the $\ell_2$-norm of the parameters. This purely optimisation-based phenomenon offers a natural explanation for the \textit{grokking} effect observed in deep learning, where the training loss rapidly reaches zero while the test loss plateaus for an extended period before suddenly improving. We argue that this generalisation jump can be attributed to the slow norm reduction induced by weight decay, as explained by our analysis. We validate this mechanism empirically on several synthetic regression tasks.



Paperid:4669
Authors:Bo Ling, Zhengyu Gan, Wanyuan Wang, Guanyu Gao, Weiwei Wu, Yan Lyu
Abstract:
Human-in-the-loop (HIL) imitation learning enables agents to learn complex behaviors safely through real-time human intervention. However, existing methods struggle to efficiently leverage agent-generated data due to dynamically evolving trajectory distributions and imperfections caused by human intervention delays, often failing to faithfully imitate the human expert policy. In this work, we propose Faithful Dynamic Imitation Learning (FaithDaIL) to address these challenges. We formulate HIL imitation learning as an online non-convex problem and employ dynamic regret minimization to adapt to the shifting data distribution and track high-quality policy trajectories. To ensure faithful imitation of the human expert despite training on mixed agent and human data, we introduce an unbiased imitation objective and achieve it by weighting the behavior distribution relative to the human expert's as a proxy reward.Extensive experiments on MetaDrive and CARLA driving benchmarks demonstrate that FaithDaIL achieves state-of-the-art performance in safety and task success with significantly reduced human intervention data compared to prior HIL baselines.



Paperid:4670
Authors:Yao Huang, Yitong Sun, Yichi Zhang, Ruochen Zhang, Yinpeng Dong, Xingxing Wei
Abstract:
Large Language Models (LLMs) have achieved remarkable proficiency across a wide range of tasks, yet with these advances come emergent behaviors, such as sophisticated deception, where models may subtly mislead or manipulate responses, posing significant risks to their trustworthiness. While LLM performance has been extensively studied, systematic evaluations of deceptive tendencies across diverse real-world contexts remain limited. To address this gap, we introduce a comprehensive evaluation framework spanning five critical domains: economy, healthcare, education, social interaction, and entertainment, selected for their profound societal relevance. Our benchmark encompasses 150 scenarios, supported by a dataset of 1.5k samples, ensuring a thorough and holistic analysis. Additionally, we investigate model behavior under varied feedback mechanisms: no feedback, positive reinforcement (reward), negative feedback (pressure), and multi-turn reinforced positive or negative feedback loops, to elucidate how these conditions influence deceptive tendencies. Through extensive experiments, we broadly evaluate a diverse set of models, including standard LLMs and reasoning LLMs, revealing critical vulnerabilities, such as amplified deceptive responses under specific feedback dynamics, highlighting the urgent need for advanced mitigation strategies.



Paperid:4671
Authors:Lyle Regenwetter, Yazan Abu Obaideh, Fabien Chiotti, Ioanna Lykourentzou, Faez Ahmed
Abstract:
Abstract:We introduce Bike-Bench, an engineering design benchmark for evaluating generative models on problems with multiple real-world objectives and constraints. As generative AI's reach continues to grow, evaluating its capability to understand physical laws, human guidelines, and hard constraints grows increasingly important. Engineering product design lies at the intersection of these difficult tasks, providing new challenges for AI capabilities. Bike-Bench evaluates AI models' capability to generate designs that not only resemble the dataset, but meet specific performance objectives and constraints. To do so, Bike-Bench quantifies a variety of human-centered and multiphysics performance characteristics, such as aerodynamics, ergonomics, structural mechanics, human-rated usability, and similarity to subjective text or image prompts. Supporting the benchmark are several datasets of simulation results, a dataset of 10K human-rated bicycle assessments, and a synthetically-generated dataset of 1.4M designs, each with a parametric, CAD/XML, SVG, and PNG representation. Bike-Bench is uniquely configured to evaluate tabular generative models, LLMs, design optimization, and hybrid algorithms side-by-side. Our experiments indicate that LLMs and tabular generative models fall short of optimization and optimization-augmented generative models in both validity and optimality scores, suggesting significant room for improvement. We hope Bike-Bench, a first-of-its-kind benchmark, will help catalyze progress in generative AI for constrained multi-objective engineering design problems.



Paperid:4672
Authors:Jingyao Li, Zhanshan Li, Shuai Lü
Abstract:
Unsupervised domain adaptive hashing has emerged as a promising approach for efficient and memory-friendly cross-domain retrieval. It leverages the model learned on labeled source domains to generate compact binary codes for unlabeled target domain samples, ensuring that semantically similar samples are mapped to nearby points in the Hamming space. Existing methods typically apply domain adaptation techniques to the feature space or the Hamming space, especially pseudo-labeling and feature alignment. However, the inherent noise of pseudo-labels and the insufficient exploration of complementary knowledge across spaces hinder the ability of the adapted model. To address these challenges, we propose a Vision-language model assisted Pseudo-labeling and Dual Space adaptation (VPDS) method. Motivated by the strong zero-shot generalization capabilities of pre-trained vision-language models (VLMs), VPDS leverages VLMs to calibrate pseudo-labels, thereby mitigating pseudo-label bias. Furthermore, to simultaneously utilize the semantic richness of high-dimensional feature space and preserve discriminative efficiency of low-dimensional Hamming space, we introduce a dual space adaptation approach that performs independent alignment within each space. Extensive experiments on three benchmark datasets demonstrate that VPDS consistently outperforms existing methods in both cross-domain and single-domain retrieval tasks, highlighting its effectiveness and superiority.



Paperid:4673
Authors:Guangting Yu, Shiwei Lan
Abstract:
Solving and learning partial differential equations (PDEs) lies at the core of physics-informed machine learning. Traditional numerical methods, such as finite difference and finite element approaches, are rooted in domain-specific techniques and often lack scalability. Recent advances have introduced neural networks and Gaussian processes (GPs) as flexible tools for automating PDE solving and incorporating physical knowledge into learning frameworks. While GPs offer tractable predictive distributions and a principled probabilistic foundation, they may be suboptimal in capturing complex behaviors such as sharp transitions or non-smooth dynamics. To address this limitation, we propose the use of the Q-exponential process (Q-EP), a recently developed generalization of GPs designed to better handle data with abrupt changes and to more accurately model derivative information. We advocate for Q-EP as a superior alternative to GPs in solving PDEs and associated inverse problems. Leveraging sparse variational inference, our method enables principled uncertainty quantification -- a capability not naturally afforded by neural network-based approaches. Through a series of experiments, including the Eikonal equation, Burgers’ equation, and an inverse Darcy flow problem, we demonstrate that the variational Q-EP method consistently yields more accurate solutions while providing meaningful uncertainty estimates.



Paperid:4674
Authors:Zimu Lu, Yunqiao Yang, Houxing Ren, Haotian Hou, Han Xiao, Ke Wang, Weikang Shi, Aojun Zhou, Mingjie Zhan, Hongsheng Li
Abstract:
LLM‑based agents have demonstrated great potential in generating and managing code within complex codebases. In this paper, we introduce WebGen-Bench, a novel benchmark designed to measure an LLM-based agent's ability to create multi-file website codebases from scratch. It contains diverse instructions for website generation, created through the combined efforts of human annotators and GPT-4o. These instructions span three major categories and thirteen minor categories, encompassing nearly all important types of web applications.To assess the quality of the generated websites, we generate test cases targeting each functionality described in the instructions. These test cases are then manually filtered, refined, and organized to ensure accuracy, resulting in a total of 647 test cases. Each test case specifies an operation to be performed on the website and the expected outcome of the operation.To automate testing and improve reproducibility, we employ a powerful web-navigation agent to execute test cases on the generated websites and determine whether the observed responses align with the expected results.We evaluate three high-performance code-agent frameworks—Bolt.diy, OpenHands, and Aider—using multiple proprietary and open-source LLMs as engines. The best-performing combination, Bolt.diy powered by DeepSeek-R1, achieves only 27.8\% accuracy on the test cases, highlighting the challenging nature of our benchmark.Additionally, we construct WebGen-Instruct, a training set consisting of 6,667 website-generation instructions. Training Qwen2.5-Coder-32B-Instruct on Bolt.diy trajectories generated from a subset of the training set achieves an accuracy of 38.2\%, surpassing the performance of the best proprietary model.We release our data-generation, training, and testing code, along with both the datasets and model weights at https://github.com/mnluzimu/WebGen-Bench.



Paperid:4675
Authors:Xueqi Ma, Jun Wang, Yanbei Jiang, Sarah Erfani, Tongliang Liu, James Bailey
Abstract:
Large language models (LLMs) have achieved state-of-the-art performance in a variety of tasks, but remain largely opaque in terms of their internal mechanisms. Understanding these mechanisms is crucial to improve their reasoning abilities. Drawing inspiration from the interplay between neural processes and human cognition, we propose a novel interpretability framework to systematically analyze the roles and behaviors of attention heads, which are key components of LLMs. We introduce CogQA, a dataset that decomposes complex questions into step-by-step subquestions with a chain-of-thought design, each associated with specific cognitive functions such as retrieval or logical reasoning. By applying a multi-label probing method, we identify the attention heads responsible for these functions. Our analysis across multiple LLM families reveals that attention heads exhibit functional specialization, characterized as cognitive heads. These cognitive heads exhibit several key properties: they are universally sparse, and vary in number and distribution across different cognitive functions, and they display interactive and hierarchical structures. We further show that cognitive heads play a vital role in reasoning tasks—removing them leads to performance degradation, while augmenting them enhances reasoning accuracy. These insights offer a deeper understanding of LLM reasoning and suggest important implications for model design, training and fine-tuning strategies.



Paperid:4676
Authors:Zhichao Deng, Zhikun Liu, Junxue Wang, Shengqian Chen, Xiang Wei, Qiang Yu
Abstract:
Spiking Neural Networks (SNNs) offer a biologically plausible and energy-efficient framework for temporal information processing. However, existing studies overlook a fundamental property widely observed in biological neurons—synaptic heterogeneity, which plays a crucial role in temporal processing and cognitive capabilities. To bridge this gap, we introduce HetSyn, a generalized framework that models synaptic heterogeneity with synapse-specific time constants. This design shifts temporal integration from the membrane potential to the synaptic current, enabling versatile timescale integration and allowing the model to capture diverse synaptic dynamics. We implement HetSyn as HetSynLIF, an extended form of the leaky integrate-and-fire (LIF) model equipped with synapse-specific decay dynamics. By adjusting the parameter configuration, HetSynLIF can be specialized into vanilla LIF neurons, neurons with threshold adaptation, and neuron-level heterogeneous models. We demonstrate that HetSynLIF not only improves the performance of SNNs across a variety of tasks—including pattern generation, delayed match-to-sample, speech recognition, and visual recognition—but also exhibits strong robustness to noise, enhanced working memory performance, efficiency under limited neuron resources, and generalization across timescales. In addition, analysis of the learned synaptic time constants reveals trends consistent with empirical observations in biological synapses. These findings underscore the significance of synaptic heterogeneity in enabling efficient neural computation, offering new insights into brain-inspired temporal modeling. Once accepted, the code will be released on a public GitHub repository.



Paperid:4677
Authors:Rui Li, Zixuan Hu, Wenxi Qu, Jinouwen Zhang, Zhenfei Yin, Sha Zhang, Xuantuo Huang, Hanqing Wang, Tai WANG, Jiangmiao Pang, Wanli Ouyang, LEI BAI, Wangmeng Zuo, LINGYU DUAN, Dongzhan Zhou, SHIXIANG TANG
Abstract:
Scientific embodied agents play a crucial role in modern laboratories by automating complex experimental workflows. Compared to typical household environments, laboratory settings impose significantly higher demands on perception of physical-chemical transformations and long-horizon planning, making them an ideal testbed for advancing embodied intelligence. However, its development has been long hampered by the lack of suitable simulator and benchmarks. In this paper, we address this gap by introducing LabUtopia, a comprehensive simulation and benchmarking suite designed to facilitate the development of generalizable, reasoning-capable embodied agents in laboratory settings. Specifically, it integrates i) LabSim, a high-fidelity simulator supporting multi-physics and chemically meaningful interactions; ii) LabScene, a scalable procedural generator for diverse scientific scenes; and iii) LabBench, a hierarchical benchmark spanning five levels of complexity from atomic actions to long-horizon mobile manipulation. LabUtopia supports 30 distinct tasks and includes more than 200 scene and instrument assets, enabling large-scale training and principled evaluation in high-complexity environments. We demonstrate that LabUtopia offers a powerful platform for advancing the integration of perception, planning, and control in scientific-purpose agents and provides a rigorous testbed for exploring the practical capabilities and generalization limits of embodied intelligence in future research. The benchmark and codes are available at https://sites.google.com/view/labutopia/ .



Paperid:4678
Authors:Tianshuo Zhang, Li Gao, Siran Peng, Xiangyu Zhu, Zhen Lei
Abstract:
The rise of realistic digital face generation/manipulation poses significant social risks. The primary challenge lies in the rapid and diverse evolution of generation techniques, which often outstrip the detection capabilities of existing models. To defend against the ever-evolving new types of forgery, we need to enable our model to quickly adapt to new domains with limited computation and data, while avoiding forgetting previously learned forgery types. In this work, we posit that genuine facial samples are abundant and relatively stable in acquisition methods, while forgery faces continuously evolve with the iteration of manipulation techniques. Given the practical infeasibility of exhaustively collecting all forgery variants, we frame face forgery detection as a continual learning problem and allow the model to scale in complexity as new forgery types emerge. Specifically, we employ a Developmental Mixture of Experts (MoE) architecture, utilizing LoRA models as the individual experts, allocating the experts into two groups: a Real-LoRA to refine the real face knowledge modeled by the backbone and Fake-LoRAs to capture incremental fake face information from different types for each sub-task. To prevent catastrophic forgetting, we ensure that the learning direction of Fake-LoRAs is orthogonal to the established subspace. Moreover, we integrate orthogonal gradients into the orthogonal loss of Fake-LoRAs to alleviate the interference of gradients on previously learned tasks during the early training phase. Experimental results under both the datasets and manipulation types incremental protocols demonstrate the effectiveness of our method.



Paperid:4679
Authors:Yunqi Gao, Bing Hu, Boloursaz Mashhadi, A-Long Jin, Yanfeng Zhang, Pei Xiao, Rahim Tafazolli, Merouane DEBBAH
Abstract:
The parameter size of modern large language models (LLMs) can be scaled up to the trillion-level via the sparsely-activated Mixture-of-Experts (MoE) technique to avoid excessive increase of the computational costs. To further improve training efficiency, pipelining computation and communication has become a promising solution for distributed MoE training. However, existing work primarily focuses on scheduling tasks within the MoE layer, such as expert computing and all-to-all (A2A) communication, while neglecting other key operations including multi-head attention (MHA) computing, gating, and all-reduce communication. In this paper, we propose FlowMoE, a scalable framework for scheduling multi-type task pipelines. First, FlowMoE constructs a unified pipeline to consistently scheduling MHA computing, gating, expert computing, and A2A communication. Second, FlowMoE introduces a tensor chunk-based priority scheduling mechanism to overlap the all-reduce communication with all computing tasks. We implement FlowMoE as an adaptive and generic framework atop PyTorch. Extensive experiments with 675 typical MoE layers and four real-world MoE models across two GPU clusters demonstrate that our proposed FlowMoE framework outperforms state-of-the-art MoE training frameworks, reducing training time by14%-57%, energy consumption by 10%-39%, and memory usage by 7%-32%. FlowMoE’s code is anonymously available at https://anonymous.4open.science/r/FlowMoE.



Paperid:4680
Authors:Shen Yuan, Yin Zheng, Taifeng Wang, BINBINLIU, Hongteng Xu
Abstract:
Adapting large-scale foundation models in multi-task scenarios often suffers from task conflict and oblivion.To mitigate such issues, we propose a novel ``model MoE-ization'' strategy that leads to a conflict- and oblivion-resistant multi-task adaptation method. Given a weight matrix of a pre-trained model, our method applies SVD to it and introduces a learnable router to adjust its singular values based on tasks and samples.Accordingly, the weight matrix becomes a Mixture of Orthogonal Rank-one Experts (MoORE), in which each expert corresponds to the outer product of a left singular vector and the corresponding right one.We can improve the model capacity by imposing a learnable orthogonal transform on the right singular vectors.Unlike low-rank adaptation (LoRA) and its MoE-driven variants, MoORE guarantees the experts' orthogonality and maintains the column space of the original weight matrix.These two properties eliminate the information redundancy of the experts and make the adapted model resistant to forgetting its original tasks, respectively. Experiments on various datasets demonstrate that MoORE outperforms existing multi-task adaptation methods consistently, showing its superiority in terms of efficiency and oblivion-resistance.



Paperid:4681
Authors:Edan Toledo, Karen Hambardzumyan, Martin Josifoski, RISHI HAZRA, Nicolas Baldwin, Alexis Audran-Reiss, Michael Kuchnik, Despoina Magka, Minqi Jiang, Alisia Lupidi, Andrei Lupu, Roberta Raileanu, Kelvin Niu, Tatiana Shavrina, Jean-Christophe Gagnon-Audet, Michael Shvartsman, Shagun Sodhani, Alexander Miller, Abhishek Charnalia, Derek Dunfield, Carole-Jean Wu, Pontus Lars Erik Saito Stenetorp, Nicola Cancedda, Jakob Foerster, Yoram Bachrach
Abstract:
AI research agents are demonstrating great potential to accelerate scientific progress by automating the design, implementation and training of machine learning models. We focus on methods for improving agents' performance on MLE-bench, a challenging benchmark where agents compete in Kaggle competitions to solve real-world machine learning problems. We formalize AI research agents as search policies that navigate a space of candidate solutions, iteratively modifying them using operators. By designing and systematically varying different operator sets and search policies (Greedy, MCTS, Evolutionary), we show that their interplay is critical for achieving high performance. Our best pairing of search strategy and operator set achieves a new state-of-the-art result on MLE-bench lite, increasing the success rate of achieving a Kaggle medal from 39.8 % to 47 %. Our investigation underscores the importance of jointly considering the search strategy, operator design, and evaluation methodology in advancing automated machine learning.



Paperid:4682
Authors:Xiangyu Guo, Zhanqian Wu, Kaixin Xiong, Ziyang Xu, Lijun Zhou, Gangwei Xu, Shaoqing Xu, Haiyang Sun, Bing Wang, Guang Chen, Hangjun Ye, Wenyu Liu, Xinggang Wang
Abstract:
We present Genesis, a unified framework for joint generation of multi-view driving videos and LiDAR sequences with spatio-temporal and cross-modal consistency. Genesis employs a two-stage architecture that integrates a DiT-based video diffusion model with 3D-VAE encoding, and a BEV-aware LiDAR generator with NeRF-based rendering and adaptive sampling. Both modalities are directly coupled through a shared latent space, enabling coherent evolution across visual and geometric domains. To guide the generation with structured semantics, we introduce DataCrafter, a captioning module built on vision-language models that provides scene-level and instance-level supervision. Extensive experiments on the nuScenes benchmark demonstrate that Genesis achieves state-of-the-art performance across video and LiDAR metrics (FVD 16.95, FID 4.24, Chamfer 0.611), and benefits downstream tasks including segmentation and 3D detection, validating the semantic fidelity and practical utility of the generated data.



Paperid:4683
Authors:Hong Wang, Jie Wang, Minghao Ma, Haoran Shao, Haoyang Liu
Abstract:
Matrix preconditioning is a critical technique to accelerate the solution of linear systems, where performance heavily depends on the selection of preconditioning parameters.Traditional parameter selection approaches often define fixed constants for specific scenarios.However, they rely on domain expertise and fail to consider the instance-wise features for individual problems, limiting their performance. In contrast, machine learning (ML) approaches, though promising, are hindered by high inference costs and limited interpretability. To combine the strengths of both approaches, we propose a symbolic discovery framework—namely,SymbolicMatrixPreconditioning (SymMaP)—to learn efficient symbolic expressions for preconditioning parameters. Specifically, we employ a neural network to search the high-dimensional discrete space for expressions that can accurately predict the optimal parameters. The learned expression allows for high inference efficiency and excellent interpretability (expressed in concise symbolic formulas), making it simple and reliable for deployment. Experimental results show that SymMaP consistently outperforms traditional strategies across various benchmarks.



Paperid:4684
Authors:Yipeng Li, Xinchen Lyu, Zhenyu Liu
Abstract:
We aim to provide a unified convergence analysis for permutation-based Stochastic Gradient Descent (SGD), where data examples are permuted before each epoch. By examining the relations among permutations, we categorize existing permutation-based SGD algorithms into three categories: Arbitrary Permutations, Independent Permutations (including Random Reshuffling and FlipFlop Rajput et al., 2022), Dependent Permutations (including GraBs Lu et al., 2022a; Cooper et al., 2023). Existing unified analyses failed to encompass the Dependent Permutations category due to the inter-epoch permutation dependency. In this work, we propose a generalized assumption that explicitly characterizes the dependence of permutations across epochs. Building upon this assumption, we develop a unified framework for permutation-based SGD with arbitrary permutations of examples, incorporating all the existing permutation-based SGD algorithms. Furthermore, we adapt our framework for Federated Learning (FL), developing a unified framework for regularized client participation FL with arbitrary permutations of clients.



Paperid:4685
Authors:Shinji Ito, Kevin Jamieson, Haipeng Luo, Arnab Maiti, Taira Tsuchiya
Abstract:
Abstract:We study online learning in finite-horizon episodic Markov decision processes (MDPs) under the challenging \textit{aggregate bandit feedback} model,where the learner observes only the cumulative loss incurred in each episode,rather than individual losses at each state-action pair.While prior work in this setting has focused exclusively on worst-case analysis,we initiate the study of \textit{best-of-both-worlds} (BOBW) algorithms that achieve low regret in both stochastic and adversarial environments.We propose the first BOBW algorithms for episodic tabular MDPs with aggregate bandit feedback.In the case of known transitions,our algorithms achieve $O(\log T)$ regret in stochastic settings and ${O}(\sqrt{T})$ regret in adversarial ones.Importantly, we also establish matching lower bounds, showing the optimality of our algorithms in this setting.We further extend our approach to unknown-transition settings by incorporating confidence-based techniques.Our results rely on a combination of FTRL over occupancy measures,self-bounding techniques,and new loss estimators inspired by recent advances in online shortest path problems.Along the way,we also provide the first individual-gap-dependent lower bounds and demonstrate near-optimal BOBW algorithms for shortest path problems with bandit feedback.



Paperid:4686
Authors:Mingjie Liu, Shizhe Diao, Ximing Lu, Jian Hu, Xin Dong, Yejin Choi, Jan Kautz, Yi Dong
Abstract:
Abstract:Recent advances in reasoning-centric language models have highlighted reinforcement learning (RL) as a promising method for aligning models with verifiable rewards. However, it remains contentious whether RL truly expands a model’s reasoning capabilities or merely amplifies high-reward outputs already latent in the base model’s distribution, and whether continually scaling up RL compute reliably leads to improved reasoning performance. In this work, we challenge prevailing assumptions by demonstrating that prolonged RL (ProRL) training can uncover novel reasoning strategies that are inaccessible to base models, even under extensive sampling. We introduce ProRL, a novel training methodology that incorporates KL divergence control, reference policy resetting, and a diverse suite of tasks. Our empirical analysis reveals that RL-trained models consistently outperform base models across a wide range of pass@$k$ evaluations, including scenarios where base models fail entirely regardless of the number of attempts. We further show that reasoning boundary improvements correlates strongly with task competence of base model and training duration, suggesting that RL can explore and populate new regions of solution space over time. These findings offer new insights into the conditions under which RL meaningfully expands reasoning boundaries in language models and establish a foundation for future work on long-horizon RL for reasoning. We will release model weights and data to support further research.



Paperid:4687
Authors:Darin Tsui, Aryan Musharaf, Yigit Efe Erginbas, Justin Kang, Amirali Aghazadeh
Abstract:
The growing adoption of machine learning models for biological sequences has intensified the need for interpretable predictions, with Shapley values emerging as a theoretically grounded standard for model explanation. While effective for local explanations of individual input sequences, scaling Shapley-based interpretability to extract global biological insights requires evaluating thousands of sequences—incurring exponential computational cost per query. We introduce SHAP zero, a novel algorithm that amortizes the cost of Shapley value computation across large-scale biological datasets. After a one-time model sketching step, SHAP zero enables near-zero marginal cost for future queries by uncovering an underexplored connection between Shapley values, high-order feature interactions, and the sparse Fourier transform of the model. Applied to models of guide RNA efficacy, DNA repair outcomes, and protein fitness, SHAP zero explains predictions orders of magnitude faster than existing methods, recovering rich combinatorial interactions previously inaccessible at scale. This work opens the door to principled, efficient, and scalable interpretability for black-box sequence models in biology.



Paperid:4688
Authors:jusheng zhang, Yijia Fan, Zimo Wen, Jian Wang, Keze Wang
Abstract:
Driven by the applications in autonomous driving, robotics, and augmented reality, 3D object annotation is a critical task compared to 2D annotation, such as spatial complexity, occlusion, and viewpoint inconsistency. The existing methods relying on single models often struggle with these issues. In this paper, we introduce Tri-MARF, a novel framework that integrates tri-modal inputs (i.e., 2D multi-view images, text descriptions, and 3D point clouds) with multi-agent collaboration to enhance the 3D annotation process. Our Tri-MARF consists of three specialized agents: a vision-language model agent that generates multi-view descriptions, an information aggregation agent that selects optimal descriptions, and a gating agent that aligns text descriptions with 3D geometries for more refined captioning. Extensive experiments on the Objaverse-LVIS, Objaverse-XL, and ABO datasets demonstrate the superiority of our Tri-MARF, which achieves a CLIPScore of 88.7 (compared to 78.6–82.4 for other SOTA methods), retrieval accuracy of 45.2/43.8 (ViLT R@5), and an impressive throughput of 12,000 objects per hour on a single NVIDIA A100 GPU.



Paperid:4689
Authors:Yifu Luo, Xinhao Hu, Keyu Fan, Haoyuan Sun, Zeyu Chen, Bo Xia, Tiantian Zhang, Yongzhe Chang, Xueqian Wang
Abstract:
Reinforcement learning (RL) has garnered increasing attention in text-to-image (T2I) generation. However, most existing RL approaches are tailored to either diffusion models or autoregressive models, overlooking an important alternative: mask generative models. In this work, we propose Mask-GRPO, the first method to incorporate Group Relative Policy Optimization (GRPO)-based online RL into this overlooked paradigm. Our core insight is to redefine the transition probability, which is different from current approaches, and formulate the unmasking process as a multi-step decision-making problem. To further enhance our method, we explore several useful strategies, including removing the Kullback–Leibler constraint, applying the reduction strategy, and filtering out low-quality samples. Using Mask-GRPO, we improve a base model, Show-o, with significant results, with a 38\% improvement on the GenEval benchmark and 10\% on MSCOCO-30K FID, outperforming existing state-of-the-art approaches.



Paperid:4690
Authors:Byunghyun Kim, Minyoung Bae, Jae-Gil Lee
Abstract:
Data synthesis has become increasingly crucial for long-tail instance segmentation tasks to mitigate class imbalance and high annotation costs. Previous methods have primarily prioritized the selection of data from a pre-generated image object pool, which frequently leads to the inefficient utilization of generated data. To address this inefficiency, we propose acollaborativeapproach that incorporates feedback from an instance segmentation model to guide the augmentation process. Specifically, the diffusion model uses feedback to generate objects that exhibit high uncertainty. The number and size of synthesized objects for each class are dynamically adjusted based on the model state to improve learning in underrepresented classes. This augmentation process is further strengthened by runningmultiple rounds, allowing feedback to be refined throughout training. In summary,multi-round collaborative augmentation (MRCA)enhances sample efficiency by providing optimal synthetic data at the right moment. Our framework requiresonly 6\%of the data generation needed by state-of-the-art methods while outperforming them.



Paperid:4691
Authors:Donghoon Ahn, Jiwon Kang, Sanghyun Lee, Minjae Kim, Wooseok Jang, Jaewon Min, Sangwu Lee, Sayak Paul, Seungryong Kim
Abstract:
Recent guidance methods steer diffusion sampling by perturbing the model to construct an implicit bad model and guide generation away from it. Among them, attention perturbation has demonstrated strong empirical performance, yet existing approaches remain fragmented, necessitating a unified framework. We propose a generalized attention perturbation guidance framework that subsumes prior methods such as PAG and SEG as special cases, and show that interpolation between the original attention distribution and perturbed attention distribution effectively mitigates the oversimplification issue of existing attention perturbation guidance. As previously observed, only a small subset of layers contributes meaningfully to image quality improvements, making careful layer selection critical, especially in large-scale architectures such as DiT, where exhaustive perturbation evaluation is computationally prohibitive. To better understand how attention perturbations affect generation, we decompose the perturbation units into individual attention heads. Surprisingly, our per-head analysis reveals that specific heads govern distinct visual factors such as structure, lighting, or texture. Building on this, we introduce a systematic framework for head selection that allows for user-defined objectives, such as enhancing aesthetic quality and reinforcing specific visual styles. We validate our approach on modern large-scale text-to-image diffusion and flow models, including the Stable Diffusion 3 family and FLUX.1. Our work provides a unified view of attention perturbation guidance, uncovers interpretable structure within attention layers, and enables practical design of effective perturbation strategies in diffusion and flow models.



Paperid:4692
Authors:Haochen Zhang, Zhong Zheng, Lingzhou Xue
Abstract:
Abstract:Motivated by real-world settings where data collection and policy deployment—whether for a single agent or across multiple agents—are costly, we study the problem of on-policy single-agent reinforcement learning (RL) and federated RL (FRL) with a focus on minimizing burn-in costs (the sample sizes needed to reach near-optimal regret) and policy switching or communication costs. In parallel finite-horizon episodic Markov Decision Processes (MDPs) with $S$ states and $A$ actions, existing methods either require superlinear burn-in costs in $S$ and $A$ or fail to achieve logarithmic switching or communication costs. We propose two novel model-free RL algorithms—Q-EarlySettled-LowCost and FedQ-EarlySettled-LowCost—that are the first in the literature to simultaneously achieve: (i) the best near-optimal regret among all known model-free RL or FRL algorithms, (ii) low burn-in cost that scales linearly with $S$ and $A$, and (iii) logarithmic policy switching cost for single-agent RL or communication cost for FRL. Additionally, we establish gap-dependent theoretical guarantees for both regret and switching/communication costs, improving or matching the best-known gap-dependent bounds.



Paperid:4693
Authors:Zihao Guo, Qingyun Sun, Ziwei Zhang, Haonan Yuan, HUIPING ZHUANG, Xingcheng Fu, Jianxin Li
Abstract:
Graph incremental learning (GIL), which continuously updates graph models by sequential knowledge acquisition, has garnered significant interest recently. However, existing GIL approaches focus on task-incremental and class-incremental scenarios within a single domain. Graph domain-incremental learning (Domain-IL), aiming at updating models across multiple graph domains, has become critical with the development of graph foundation models (GFMs), but remains unexplored in the literature. In this paper, we proposeGraphDomain-Incremental Learning viaKnowledge Disentanglement andPreservation (GraphKeeper), to address catastrophic forgetting in Domain-IL scenario from the perspectives of embedding shifts and decision boundary deviations. Specifically, to prevent embedding shifts and confusion across incremental graph domains, we first propose the domain-specific parameter-efficient fine-tuning together with intra- and inter-domain disentanglement objectives. Consequently, to maintain a stable decision boundary, we introduce deviation-free knowledge preservation to continuously fit incremental domains. Additionally, for graphs with unobservable domains, we perform domain-aware distribution discrimination to obtain precise embeddings. Extensive experiments demonstrate the proposed GraphKeeper achieves state-of-the-art results with 6.5%\~16.6% improvement over the runner-up with negligible forgetting. Moreover, we show GraphKeeper can be seamlessly integrated with various representative GFMs, highlighting its broad applicative potential.



Paperid:4694
Authors:Yanghao Zhang, Panagiotis Kouvaros, Alessio Lomuscio
Abstract:
Neural Network (NN) verification methods provide local robustnessguarantees for a NN in the dense perturbation space of an input point.A key challenge in this area lies in the scalability of the resultingproblem, leading to large models of interest in applications not beingaddressable by the methods.In this paper we introduce HOVer, a method for the verificationof NNs against geometric perturbations, that uniquely employs a Hilbertspace-filling construction to reduce multidimensional problems tosingle-dimensional ones. The underlying Hölder optimisation, which alsoiteratively refines the estimation for the Hölder constant for constructing the lower bound, theoretically it may converge to a localminimum, thereby resulting in arobustness result being incorrect. However, we show experimentally that thisrisk can be contained in practice by appropriately devised heuristicsin the global optimisation setup. Indeed, unlike recently reportederrors from theoretically sound implementations, we found no incorrectresult by running the technique on a large set of benchmarks fromSoundnessBench and VNN-COMP.To validate the scalability of the approach, we report on extensiveexperiments on large NNs ranging from Resnet34 to Resnet152 and ViTs.These demonstrate the SoA performance of the approach in evaluating withhigh reliability the local robustness of large NNs against geometric perturbations on theImageNet dataset.Beyond image tasks, we show that the method's scalabilityenables for the first time robustness assessments for large-scale 3D-NNs in video classification tasks against geometricperturbations for long-sequence input frames on Kinetics/UCF101 datasets.



Paperid:4695
Authors:Peter Richtarik, Simone Maria Giancola, Dymitr Lubczyk, Robin Yadav
Abstract:
We contribute to the growing body of knowledge on more powerful and adaptive stepsizes for convex optimization, empowered by local curvature information. We do not go the route of fully-fledged second-order methods, which require the expensive computation of the Hessian. Instead, our key observation is that, for some problems (e.g., when minimizing the sum of squares of absolutely convex functions), local curvature information is readily available, and can be used to obtain surprisingly powerful matrix-valued stepsizes, and meaningful theory. In particular, we develop three new methods — LCD1, LCD2, and LCD3 — where the abbreviation stands for local curvature descent. While LCD1 generalizes gradient descent with fixed stepsize, LCD2 generalizes gradient descent with Polyak stepsize. Our methods enhance these classical gradient descent baselines with local curvature information, and our theory recovers the known rates in the special case when no curvature information is used. Our last method, LCD3, is a variable-metric version of LCD2; this feature leads to a closed-form expression for the iterates. Our empirical results are encouraging and show that the local curvature descent improves upon gradient descent.



Paperid:4696
Authors:Harsh Parikh, Trang Nguyen, Elizabeth Stuart, Kara Rudolph, Caleb Miles
Abstract:
Data integration approaches are increasingly used to enhance the efficiency and generalizability of studies. However, a key limitation of these methods is the assumption that outcome measures are identical across datasets -- an assumption that often does not hold in practice. Consider the following opioid use disorder (OUD) studies: the XBOT trial and the POAT study, both evaluating the effect of medications for OUD on withdrawal symptom severity (not the primary outcome of either trial). While XBOT measures withdrawal severity using the subjective opiate withdrawal scale, POAT uses the clinical opiate withdrawal scale. We analyze this realistic yet challenging setting where outcome measures differ across studies and where neither study records both types of outcomes. Our paper studies whether and when integrating studies with disparate outcome measures leads to efficiency gains. We introduce three sets of assumptions -- with varying degrees of strength -- linking both outcome measures. Our theoretical and empirical results highlight a cautionary tale: integration can improve asymptotic efficiency only under the strongest assumption linking the outcomes. However, misspecification of this assumption leads to bias. In contrast, a milder assumption may yield finite-sample efficiency gains, yet these benefits diminish as sample size increases. We illustrate these trade-offs via a case study integrating the XBOT and POAT datasets to estimate the comparative effect of two medications for opioid use disorder on withdrawal symptoms. By systematically varying the assumptions linking the SOW and COW scales, we show potential efficiency gains and the risks of bias. Our findings emphasize the need for careful assumption selection when fusing datasets with differing outcome measures, offering guidance for researchers navigating this common challenge in modern data integration.



Paperid:4697
Authors:Zijie Lin, Yang Zhang, Xiaoyan Zhao, Fengbin ZHU, Fuli Feng, Tat-Seng Chua
Abstract:
Large Language Models (LLMs) have shown strong potential for recommendation by framing item prediction as a token-by-token language generation task. However, existing methods treat all item tokens equally, simply pursuing likelihood maximization during both optimization and decoding. This overlooks crucial token-level differences in decisiveness—many tokens contribute little to item discrimination yet can dominate optimization or decoding.To quantify token decisiveness, we propose a novel perspective that models item generation as a decision process, measuring token decisiveness by the Information Gain (IG) each token provides in reducing uncertainty about the generated item. Our empirical analysis reveals that most tokens have low IG but often correspond to high logits, disproportionately influencing training loss and decoding, which may impair model performance.Building on these insights, we introduce an Information Gain-based Decisiveness-aware Token handling (IGD) strategy that integrates token decisiveness into both tuning and decoding. Specifically, IGD downweights low-IG tokens during tuning and rebalances decoding to emphasize tokens with high IG. In this way, IGD moves beyond pure likelihood maximization, effectively prioritizing high-decisiveness tokens. Extensive experiments on four benchmark datasets with two LLM backbones demonstrate that IGD consistently improves recommendation accuracy, achieving significant gains on widely used ranking metrics compared to strong baselines. Our codes are available at \url{https://anonymous.4open.science/r/LLM4Rec-IGD-FFFE}.



Paperid:4698
Authors:Junyi Li, Hwee Tou Ng
Abstract:
Large language models (LLMs) have significantly advanced in reasoning tasks through reinforcement learning (RL) optimization, achieving impressive capabilities across various challenging benchmarks. However, our empirical analysis reveals a critical drawback: reasoning-oriented RL fine-tuning significantly increases the prevalence of hallucinations. We theoretically analyze the RL training dynamics, identifying high-variance gradient, entropy-induced randomness, and susceptibility to spurious local optima as key factors leading to hallucinations. To address this drawback, we propose Factuality-aware Step-wise Policy Optimization (FSPO), an innovative RL fine-tuning algorithm incorporating explicit factuality verification at each reasoning step. FSPO leverages automated verification against given evidence to dynamically adjust token-level advantage values, incentivizing factual correctness throughout the reasoning process. Experiments across mathematical reasoning and hallucination benchmarks using Qwen2.5 and Llama models demonstrate that FSPO effectively reduces hallucinations while enhancing reasoning accuracy, substantially improving both reliability and performance.



Paperid:4699
Authors:Shikuang Deng, Jiayuan Zhang, Yuhang Wu, Ting Chen, Shi Gu
Abstract:
Hebbian learning is a biological principle that intuitively describes how neurons adapt connections through repeated stimuli. However, when applied to machine learning, it suffers serious issues due to the unconstrained updates of the connection and lack of account for feedback mediation. Such shortcomings limit its effective scaling to complex network architectures and tasks. To this end, here we introduce the Structural Projection Hebbian Representation (SPHeRe), a novel unsupervised learning method that integrates orthogonality and structural information preservation through a local auxiliary nonlinear block. The loss for structural information preservation backward to the input through an auxiliary lightweight projection that conceptually serves as feedback mediation while the orthogonality constraints account for the boundness of updating magnititude. Extensive experimental results show that SPHeRe achieves SOTA performance among unsupervised synaptic plasticity approaches on standard image classification benchmarks, including CIFAR-10, CIFAR-100, and Tiny-ImageNet. Furthermore, the method exhibits strong effectiveness in continual learning and transfer learning scenarios, and image reconstruction tasks show the robustness and generalizability of the extracted features. This work demonstrates the competitiveness and potential of Hebbian unsupervised learning rules within modern deep learning frameworks, demonstrating the possibility of efficient and biologically plausible learning algorithms without the strong dependence on strict backpropagation.



Paperid:4700
Authors:Zhiqiang Zhong, Davide Mottin
Abstract:
Recent studies inMachine Learning(ML) for biological research focus on investigating molecular properties to accelerate drug discovery. However, limited labeled molecular data often hampers the performance of ML models. A common strategy to mitigate data scarcity is leveraging auxiliary learning tasks to provide additional supervision, but selecting effective auxiliary tasks requires substantial domain expertise and manual effort, and their inclusion does not always guarantee performance gains. To overcome these challenges, we introduceAutomatic Auxiliary Task Selection(AutAuT), a fully automated framework that seamlessly retrieves auxiliary tasks using large language models and adaptively integrates them through a novelgradient alignmentweighting mechanism. By automatically emphasizing auxiliary tasks aligned with the primary objective, AutAuT significantly enhances predictive accuracy while reducing negative impacts from irrelevant tasks. Extensive evaluations demonstrate that AutAuT outperforms10auxiliary task-based approaches and18advanced molecular property prediction models.



Paperid:4701
Authors:Jun Wang, Zhenglai Li, Chang Tang, Suyuan Liu, Hao Yu, Chuan Tang, Miaomiao Li, Xinwang Liu
Abstract:
Most existing multi-view clustering methods aim to generate a consensus partition across all views, based on the assumption that all views share the same sample arrangement. However, in real-world scenarios, the collected data across different views is often unsynchronized, making it difficult to ensure consistent sample correspondence between views. To address this issue, we propose a scalable sample-alignment-based multi-view clustering method, referred to as SSA-MVC. Specifically, we first employ a cluster-label matching (CLM) algorithm to select the view whose clustering labels best match those of the others as the benchmark view. Then, for each of the remaining views, we construct representations of non-aligned samples by computing their similarities with aligned samples. Based on these representations, we build a similarity graph between the non-aligned samples of each view and those in the benchmark view, which serves as the alignment criterion. This alignment criterion is then integrated into a late-fusion framework to enable clustering without requiring aligned samples. Notably, the learned sample alignment matrix can be used to enhance existing multi-view clustering methods in scenarios where sample correspondence is unavailable. The effectiveness of the proposed SSA-MVC algorithm is validated through extensive experiments conducted on eight real-world multi-view datasets.



Paperid:4702
Authors:Xiao Liang, Zhong-Zhi Li, Yeyun Gong, Yang Wang, Hengyuan Zhang, yelong shen, Ying Nian Wu, Weizhu Chen
Abstract:
Reinforcement Learning with Verifiable Rewards (RLVR) has proven effective for training large language models (LLMs) on complex reasoning tasks, such as mathematical problem solving. A prerequisite for the scalability of RLVR is a high-quality problem set with precise and verifiable answers. However, the scarcity of well-crafted human-labeled math problems and limited-verification answers in existing distillation-oriented synthetic datasets limit their effectiveness in RL. Additionally, most problem synthesis strategies indiscriminately expand the problem set without considering the model’s capabilities, leading to low efficiency in generating useful questions. To mitigate this issue, we introduce a Self-aware Weakness-driven problem Synthesis framework (SwS) that systematically identifies model deficiencies and leverages them for problem augmentation. Specifically, we define weaknesses as questions that the model consistently fails to learn through its iterative sampling during RL training. We then extract the core concepts from these failure cases and synthesize new problems to strengthen the model's weak areas in subsequent augmented training, enabling it to focus on and gradually overcome its weaknesses. Without relying on external knowledge distillation, our framework enables robust generalization by empowering the model to self-identify and address its weaknesses in RL, yielding average performance gains of 10% and 7.7% on 7B and 32B models across eight mainstream reasoning benchmarks. Our code and data are available at https://anonymous.4open.science/r/SwS-E6F5/



Paperid:4703
Authors:Navita Goyal, Hal Daumé III, Alexandre Drouin, Dhanya Sridhar
Abstract:
Language model activations entangle concepts that mediate their behavior, making it difficult to interpret these factors, which has implications for generalizability and robustness. We introduce an approach for disentangling these concepts without supervision. Existing methods for concept discovery often rely on external labels, contrastive prompts, or known causal structures, which limits their scalability and biases them toward predefined, easily annotatable features. In contrast, we propose a new unsupervised algorithm that identifies causal differentiating concepts—interpretable latent directions in LM activations that must be changed to elicit a different model behavior. These concepts are discovered using a constrained contrastive learning objective, guided by the insight that eliciting a target behavior requires only sparse changes to the underlying concepts. We formalize this notion and show that under a particular assumption about the sparsity of these causal differentiating concepts, our method learns disentangled representations that align with human-interpretable factors influencing LM decisions. We empirically show the ability of our method to recover ground-truth causal factors in synthetic and semi-synthetic settings. Additionally, we illustrate the utility of our method through a case study on refusal behavior in language models. Our approach offers a scalable and interpretable lens into the internal workings of LMs, providing a principled foundation for interpreting language model behavior.



Paperid:4704
Authors:Dongki Jung, Jaehoon Choi, Yonghan Lee, Dinesh Manocha
Abstract:
Abstract:The increasing use of 360$^\circ$ images across various domains has emphasized the need for robust depth estimation techniques tailored for omnidirectional images. However, obtaining large-scale labeled datasets for 360$^\circ$ depth estimation remains a significant challenge. In this paper, we propose RPG360, a training-free robust 360$^\circ$ monocular depth estimation method that leverages perspective foundation models and graph optimization. Our approach converts 360$^\circ$ images into six- face cubemap representations, where a perspective foundation model is employed to estimate depth and surface normals. To address depth scale inconsistencies across different faces of the cubemap, we introduce a novel depth scale alignment technique using graph-based optimization, which parameterizes the predicted depth and normal maps while incorporating an additional per-face scale parameter. This optimization ensures depth scale consistency across the six-face cubemap while preserving 3D structural integrity. Furthermore, as foundation models exhibit inherent robustness in zero-shot settings, our method achieves superior performance across diverse datasets, including Matterport3D, Stanford2D3D, and 360Loc. We also demonstrate the versatility of our depth estimation approach by validating its benefits in downstream tasks such as feature matching 3.2 ∼ 5.4% and Structure from Motion 0.2 ∼ 9.7% in AUC@5$^\circ$.



Paperid:4705
Authors:Atharva Gundawar, Som Sagar, Ransalu Senanayake
Abstract:
Vision-Language Models (VLMs) are increasingly pivotal for generalist robot ma-nipulation, enabling tasks such as physical reasoning, policy generation, and failuredetection. However, their proficiency in these high-level applications often assumesa deep understanding of low-level physical prerequisites, a capability that is largelyunverified. To perform actions reliably, robots must comprehend intrinsic objectproperties (e.g., material, weight), action affordances (e.g., graspable, stackable),and physical constraints (e.g., stability, reachability, or an object’s state like beingclosed). Despite their ubiquitous use in manipulation, we argue that off-the-shelfVLMs may lack this granular, physically-grounded understanding, as these specificprerequisites are often overlooked in their pre-training. Addressing this criticalgap, we introduce PAC Bench, a comprehensive benchmark designed to system-atically evaluate VLM comprehension of these core Properties, Affordances, andConstraints (PAC) from a task executability perspective. PAC Bench features adiverse dataset with over 30,000 annotations, comprising 673 real-world images(115 object classes, 15 property types, 1–3 affordances defined per class), 100real-world humanoid-view scenarios and 120 unique simulated constraint scenariosacross four tasks. Our evaluations reveal significant gaps in the ability of VLMsto grasp fundamental physical concepts, underscoring their current limitations forreliable robot manipulation and pointing to key areas that require targeted research.PAC Bench also serves as a standardized benchmark for rigorously evaluating VLMphysical reasoning and guiding the development of more robust and physicallygrounded models for robotic manipulation.



Paperid:4706
Authors:Andreas Auer, Patrick Podest, Daniel Klotz, Sebastian Böck, Günter Klambauer, Sepp Hochreiter
Abstract:
In-context learning, the ability of large language models to perform tasks using only examples provided in the prompt, has recently been adapted for time series forecasting. This paradigm enables zero-shot prediction, where past values serve as context for forecasting future values, making powerful forecasting tools accessible to non-experts and increasing the performance when training data are scarce.Most existing zero-shot forecasting approaches rely on transformer architectures, which, despite their success in language, often fall short of expectations in time series forecasting, where recurrent models like LSTMs frequently have the edge. Conversely, while LSTMs are well-suited for time series modeling due to their state-tracking capabilities, they lack strong in-context learning abilities.We introduceTiRexthat closes this gap by leveraging xLSTM, an enhanced LSTM with competitive in-context learning skills. Unlike transformers, state-space models, or parallelizable RNNs such as RWKV, TiRex retains state tracking, a critical property for long-horizon forecasting. To further facilitate its state-tracking ability, we propose a training-time masking strategy called CPM.TiRex sets a new state of the art in zero-shot time series forecasting on the Hugging Face benchmarksGiftEvalandChronos-ZS, outperforming significantly larger models includingTabPFN-TS(Prior Labs),Chronos Bolt(Amazon),TimesFM(Google), andMoirai(Salesforce) across both short- and long-term forecasts.



Paperid:4707
Authors:Hebaixu Wang, Jing Zhang, Haonan Guo, Di Wang, Jiayi Ma, Bo Du
Abstract:
Diffusion models have achieved remarkable progress in universal image restoration. However, existing methods perform naive inference in the reverse process, which leads to cumulative errors under limited sampling steps and large step intervals. Moreover, they struggle to balance the commonality of degradation representations with restoration quality, often depending on complex compensation mechanisms that enhance fidelity at the expense of efficiency. To address these challenges, we introduce DGSolver, a diffusion generalist solver with universal posterior sampling. We first derive the exact ordinary differential equations for generalist diffusion models to unify degradation representations and design tailored high-order solvers with a queue-based accelerated sampling strategy to improve both accuracy and efficiency. We then integrate universal posterior sampling to better approximate manifold-constrained gradients, yielding a more accurate noise estimation and correcting errors in inverse inference. Extensive experiments demonstrate that DGSolver outperforms state-of-the-art methods in restoration accuracy, stability, and scalability, both qualitatively and quantitatively.



Paperid:4708
Authors:Chris Kolb, Laetitia Frost, Bernd Bischl, David Rügamer
Abstract:
Abstract:Structured sparsity regularization offers a principled way to compact neural networks, but its non-differentiability breaks compatibility with conventional stochastic gradient descent and requires either specialized optimizers or additional post-hoc pruning without formal guarantees. In this work, we propose $D$-Gating, a fully differentiable structured overparameterization that splits each group of weights into a primary weight vector and multiple scalar gating factors. We prove that any local minimum under $D$-Gating is also a local minimum using non-smooth structured $L_{2,2/D}$ penalization, and further show that the $D$-Gating objective converges at least exponentially fast to the $L_{2,2/D}$–regularized loss in the gradient flow limit. Together, our results show that $D$-Gating is theoretically equivalent to solving the original structured sparsity problem, yet induces distinct learning dynamics that evolve from a non-sparse regime into sparse optimization. We validate our theory across vision, language, and tabular tasks, where $D$-Gating consistently delivers strong performance–sparsity tradeoffs and outperforms both direct optimization of structured penalties and conventional pruning baselines.



Paperid:4709
Authors:Yanjie Li, Wenxuan Zhang, Xinqi Lyu, Yihao LIU, Bin Xiao
Abstract:
Recently, text-to-image diffusion models have been widely used for style mimicry and personalized customization through methods such as DreamBooth and Textual Inversion. This has raised concerns about intellectual property protection and the generation of deceptive content.Recent studies, such as Glaze and Anti-DreamBooth, have proposed using adversarial noise to protect images from these attacks. However, recent purification-based methods, such as DiffPure and Noise Upscaling, have successfully attacked these latest defenses, showing the vulnerabilities of these methods. Moreover, present methods show limited transferability across models, making them less effective against unknown text-to-image models.To address these issues, we propose a novel anti-mimicry method, StyleGuard. We propose a novel style loss that optimizes the style-related features in the latent space to make it deviate from the original image, which improves model-agnostic transferability.Additionally, to enhance the perturbation's ability to bypass diffusion-based purification, we designed a novel upscale loss that involves ensemble purifiers and upscalers during training.Extensive experiments on the WikiArt and CelebA datasets demonstrate that StyleGuard outperforms existing methods in robustness against various transformations and purifications, effectively countering style mimicry in various models. Moreover, StyleGuard is effective on different style mimicry methods, including DreamBooth and Textual Inversion.



Paperid:4710
Authors:Yuanshao Zhu, James Yu, Xiangyu Zhao, Xun Zhou, Liang Han, Xuetao Wei, Yuxuan Liang
Abstract:
Building a universal trajectory foundation model is a promising solution to address the limitations of existing trajectory modeling approaches, such as task specificity, regional dependency, and data sensitivity.Despite its potential, data preparation, pre-training strategy development, and architectural design present significant challenges in constructing this model.Therefore, we introduceUniTraj, a Universal Trajectory foundation model that aims to address these limitations through three key innovations.First, we constructWorldTrace, an unprecedented dataset of 2.45 million trajectories with billions of GPS points spanning 70 countries, providing the diverse geographic coverage essential for region-independent modeling. Second, we develop novel pre-training strategies---Adaptive Trajectory Resampling and Self-supervised Trajectory Masking---that enable robust learning from heterogeneous trajectory data with varying sampling rates and quality. Finally, we tailor a flexible model architecture to accommodate a variety of trajectory tasks, effectively capturing complex movement patterns to support broad applicability.Extensive experiments across multiple tasks and real-world datasets demonstrate that UniTraj consistently outperforms existing methods, exhibiting superior scalability, adaptability, and generalization. These findings also position UniTraj as a versatile solution for a wide array of trajectory-related applications, with WorldTrace serving as an ideal yet non-exclusive training resource. The implementation codes and dataset samples are available in theSupplementary Material.



Paperid:4711
Authors:Yu Yang, Alan Liang, Jianbiao Mei, Yukai Ma, Yong Liu, Gim Hee Lee
Abstract:
Diffusion models are advancing autonomous driving by enabling realistic data synthesis, predictive end-to-end planning, and closed-loop simulation, with a primary focus on temporally consistent generation. However, the generation of large-scale 3D scenes that require spatial coherence remains underexplored. In this paper, we propose X-Scene, a novel framework for large-scale driving scene generation that achieves both geometric intricacy and appearance fidelity, while offering flexible controllability. Specifically, X-Scene supports multi-granular control, including low-level conditions such as user-provided or text-driven layout for detailed scene composition and high-level semantic guidance such as user-intent and LLM-enriched text prompts for efficient customization. To enhance geometrical and visual fidelity, we introduce a unified pipeline that sequentially generates 3D semantic occupancy and the corresponding multiview images, while ensuring alignment between modalities. Additionally, we extend the generated local region into a large-scale scene through consistency-aware scene outpainting, which extrapolates new occupancy and images conditioned on the previously generated area, enhancing spatial continuity and preserving visual coherence. The resulting scenes are lifted into high-quality 3DGS representations, supporting diverse applications such as scene exploration. Comprehensive experiments demonstrate that X-Scene significantly advances controllability and fidelity for large-scale driving scene generation, empowering data generation and simulation for autonomous driving.



Paperid:4712
Authors:Zhenbin Wang, Lei Zhang, Wei Huang, Zhao Zhang, Zizhou Wang
Abstract:
Abstract:Out-of-distribution (OOD) problems commonly occur when models process data with a distribution significantly deviates from the in-distribution (InD) training data. In this paper, we hypothesize that a $\textit{field}$ or $\textit{potential}$ more essential than features exists, and features are not the ultimate essence of the data but rather manifestations of them during training. we investigate OOD problems from the perspective of collective behavior dynamics. With this in mind, we first treat the output of the feature extractor as charged particles and investigate their collective behavior dynamics within a self-consistent electric field. Then, to characterize the relationship between OOD problems and dynamical equations, we introduce the $\textit{basin of attraction}$ and prove that its boundary can be represented as the zero level set of a differentiable function of the potential, $\textit{i.e.}$, the spatial integral of field. We further demonstrate that: $\textit{i)}$ InD and OOD inputs can be effectively separated based on whether they are steady state solutions for specific field conditions, enabling robust OOD detection and outperforming prior methods over three benchmarks. $\textit{ii)}$ the generalization capability correlates positively with the basin of attraction. By analyzing the dynamics of perturbations, we propose that the potential is well-characterized by a Fourier-domain form of the Poisson equation. Evaluated on six benchmark datasets, our method rivals the SoTA approaches for OOD generalization and can be seamlessly integrated with them to deliver additional gains.



Paperid:4713
Authors:Albert Matveev, Sanmitra Ghosh, Aamal Hussain, James-Michael Leahy, Michalis Michaelides
Abstract:
Operator learning is a powerful paradigm for solving partial differential equations, with Fourier Neural Operators serving as a widely adopted foundation. However, FNOs face significant scalability challenges due to overparameterization and offer no native uncertainty quantification -- a key requirement for reliable scientific and engineering applications. Instead, neural operators rely on post hoc UQ methods that ignore geometric inductive biases. In this work, we introduce DINOZAUR: a diffusion-based neural operator parametrization with uncertainty quantification.Inspired by the structure of the heat kernel, DINOZAUR replaces the dense tensor multiplier in FNOs with a dimensionality-independent diffusion multiplier that has a single learnable time parameter per channel, drastically reducing parameter count and memory footprint without compromising predictive performance. By defining priors over those time parameters, we cast DINOZAUR as a Bayesian neural operator to yield spatially correlated outputs and calibrated uncertainty estimates. Our method achieves competitive or superior performance across several PDE benchmarks while providing efficient uncertainty quantification.



Paperid:4714
Authors:Jie Zhang, Cezara Petrui, Kristina Nikolić, Florian Tramer
Abstract:
Existing benchmarks for evaluating mathematical reasoning in large language models (LLMs) rely primarily on competition problems, formal proofs, or artificially challenging questions---failing to capture the nature of mathematics encountered in actual research environments. We introduce \textsc{RealMath}, a novel benchmark derived directly from research papers and mathematical forums that assesses LLMs' abilities on authentic mathematical tasks. Our approach addresses three critical challenges: sourcing diverse research-level content, enabling reliable automated evaluation through verifiable statements, and designing a continually refreshable dataset to mitigate contamination risks. Experimental results across multiple LLMs reveal surprising capabilities in handling research mathematics compared to competition problems, suggesting current models may already serve as valuable assistants for working mathematicians despite limitations on highly challenging problems. The code and dataset for \textsc{RealMath} are publicly available.



Paperid:4715
Authors:Shilin Zhang, Zican Hu, Wenhao Wu, Xinyi Xie, Jianxiang Tang, Chunlin Chen, Daoyi Dong, Yu Cheng, Zhenhong Sun, Zhi Wang
Abstract:
Offline meta-RL usually tackles generalization by inferring task beliefs from high-quality samples or warmup explorations. The restricted form limits their generality and usability since these supervision signals are expensive and even infeasible to acquire in advance for unseen tasks. Learning directly from the raw text about decision tasks is a promising alternative to leverage a much broader source of supervision. In the paper, we proposeText-to-DecisionAgent (T2DA), a simple and scalable framework that supervises offline meta-RL with natural language. We first introduce a generalized world model to encode multi-task decision data into a dynamics-aware embedding space. Then, inspired by CLIP, we predict which textual description goes with which decision embedding, effectively bridging their semantic gap via contrastive language-decision pre-training and aligning the text embeddings to comprehend the environment dynamics. After training the text-conditioned generalist policy, the agent can directly realize zero-shot text-to-decision generation in response to language instructions. Comprehensive experiments on MuJoCo and Meta-World benchmarks show that T2DA facilitates high-capacity zero-shot generalization and outperforms various types of baselines.



Paperid:4716
Authors:Shenghai Yuan, Xianyi He, Yufan Deng, Yang Ye, Jinfa Huang, lin bin, Chongyang Ma, Jiebo Luo, Li Yuan
Abstract:
Subject-to-Video (S2V) generation aims to create videos that faithfully incorporate reference content, providing enhanced flexibility in the production of videos. To establish the infrastructure for S2V generation, we proposeOpenS2V-Nexus, consisting of (i)OpenS2V‑Eval, a fine‑grained benchmark, and (ii)OpenS2V‑5M, a million‑scale dataset.In contrast to existing S2V benchmarks inherited from VBench that focus on global and coarse-grained assessment of generated videos,OpenS2V-Evalfocuses on the model's ability to generate subject-consistent videos with natural subject appearance and identity fidelity. For these purposes,OpenS2V-Evalintroduces 180 prompts from seven major categories of S2V, which incorporate both real and synthetic test data. Furthermore, to accurately align human preferences with S2V benchmarks, we propose three automatic metrics, NexusScore, NaturalScore and GmeScore, to separately quantify subject consistency, naturalness, and text relevance in generated videos. Building on this, we conduct a comprehensive evaluation of 11 representative S2V models, highlighting their strengths and weaknesses across different content. Moreover, we create the first open-source large-scale S2V generation datasetOpenS2V-5M, which consists of five million high-quality 720P subject-text-video triplets. Specifically, we ensure subject‐information diversity in our dataset by (1) segmenting subjects and building pairing information via cross‐video associations and (2) prompting GPT-4o on raw frames to synthesize multi-view representations. ThroughOpenS2V-Nexus, we deliver a robust infrastructure to accelerate future S2V generation research.



Paperid:4717
Authors:Insu Lee, Wooje Park, Jaeyun Jang, Minyoung Noh, Kyuhong Shim, Byonghyo Shim
Abstract:
Large vision-language models (LVLMs) are increasingly deployed in interactive applications such as virtual and augmented reality, where first-person (egocentric) view captured by head-mounted cameras serves as key input.While this view offer fine-grained cues about user attention and hand–object interactions, their narrow field of view and lack of global context often lead to failures on spatially or contextually demanding queries.To address this, we introduce a framework that augments egocentric inputs with third-person (exocentric) views, providing complementary information such as global scene layout and object visibility to LVLMs.We present E3VQA, the first benchmark for multi-view question answering with 4K high-quality question–answer pairs grounded in synchronized ego–exo image pairs.Additionally, we propose M3CoT, a training-free prompting technique that constructs a unified scene representation by integrating scene graphs from three complementary perspectives.M3CoT enables LVLMs to reason more effectively across views, yielding consistent performance gains (4.84\% for GPT-4o and 5.94\% for Gemini 2.0 Flash) over a recent CoT baseline.Our extensive evaluation reveals key strengths and limitations of LVLMs in multi-view reasoning and highlights the value of leveraging both egocentric and exocentric inputs.



Paperid:4718
Authors:Jie Liu, Wenxuan Wang, Zizhan Ma, Guolin Huang, Yihang SU, Kao-Jung Chang, Wenting Chen, Haoliang Li, Linlin Shen, Michael R Lyu
Abstract:
Clinical decision making (CDM) is a complex, dynamic process crucial to healthcare delivery, yet it remains a significant challenge for artificial intelligence systems. While Large Language Model (LLM)-based agents have been tested on general medical knowledge using licensing exams and knowledge question-answering tasks, their performance in the CDM in real-world scenarios is limited due to the lack of comprehensive benchmark that mirror actual medical practice. To address this gap, we present MedChain, a dataset of 12,163 clinical cases that covers five key stages of clinical workflow. MedChain distinguishes itself from existing benchmarks with three key features of real-world clinical practice: personalization, interactivity, and sequentiality. Further, to tackle real-world CDM challenges, we also propose MedChain-Agent, an AI system that integrates a feedback mechanism and a MedCase-RAG module to learn from previous cases and adapt its responses. MedChain-Agent demonstrates remarkable adaptability in gathering information dynamically and handling sequential clinical tasks, significantly outperforming existing approaches. The relevant dataset and code will be released upon acceptance of this paper.



Paperid:4719
Authors:Qiyao Wei, Edward R Morrell, Lea Goetz, Mihaela van der Schaar
Abstract:
Evaluating the open-form textual responses generated by Large Language Models (LLMs) typically requires measuring the semantic similarity of the response to a (human generated) reference. However, there is evidence that current semantic similarity methods may capture syntactic or lexical forms over semantic content. While benchmarks exist for semantic equivalence, they often suffer from high generation costs due to reliance on subjective human judgment, limited availability for domain-specific applications, and unclear definitions of equivalence. This paper introduces a novel method for generating benchmarks to evaluate semantic similarity methods for LLM outputs, specifically addressing these limitations. Our approach leverages knowledge graphs (KGs) to generate pairs of natural-language statements that are semantically similar or dissimilar, with dissimilar pairs categorized into one of four sub-types. We generate benchmark datasets in four different domains (general knowledge, biomedicine, finance, biology), and conduct a comparative study of semantic similarity methods including traditional natural language processing scores and LLM-as-a-judge predictions. We observe that the sub-type of semantic variation, as well as the domain of the benchmark impact the performance of semantic similarity methods, with no method being consistently superior. Our results present important implications for the use of LLM-as-a-judge in detecting the semantic content of text.



Paperid:4720
Authors:Jun Zhao, Yongzhuo Yang, Xiang Hu, Jingqi Tong, Yi Lu, Wei Wu, Tao Gui, Qi Zhang, Xuanjing Huang
Abstract:
Retrieval-Augmented Generation (RAG) provides additional contextual knowledge to complement the parametric knowledge in Large Language Models (LLMs). These two knowledge interweave to enhance the accuracy and timeliness of LLM responses. However, the internal mechanisms by which LLMs utilize these knowledge remain unclear. We propose modeling the forward propagation of knowledge as an entity flow, employing this framework to trace LLMs' internal behaviors when processing mixed-source knowledge. Linear probing utilizes a trainable linear classifier to detect specific attributes in hidden layers. However, once trained, a probe cannot adapt to dynamically specified entities. To address this challenge, we construct an entity-aware probe, which introduces special tokens to mark probing targets and employs a small trainable rank-8 lora update to process these special markers. We first verify this approach through an attribution experiment, demonstrating that it can accurately detect information about ad-hoc entities from complex hidden states. Next, we trace entity flows across layers to understand how LLMs reconcile conflicting knowledge internally. Our probing results reveal that contextual and parametric knowledge are routed between tokens through distinct sets of attention heads, supporting attention competition only within knowledge types. While conflicting knowledge maintains a residual presence across layers, aligned knowledge from multiple sources gradually accumulates, with the magnitude of this accumulation directly determining its influence on final outputs.



Paperid:4721
Authors:Jiyao Zhang, Zhiyuan Ma, Tianhao Wu, Zeyuan Chen, Hao Dong
Abstract:
Dexterous grasping in cluttered environments presents substantial challenges due to the high degrees of freedom of dexterous hands, occlusion, and potential collisions arising from diverse object geometries and complex layouts. To address these challenges, we propose CADGrasp, a two-stage algorithm for general dexterous grasping using single-view point cloud inputs. In the first stage, we predict a scene-decoupled, contact- and collision-aware representation—sparse IBS—as the optimization target. Sparse IBS compactly encodes the geometric and contact relationships between the dexterous hand and the scene, enabling stable and collision-free dexterous grasp pose optimization. To enhance the prediction of this high-dimensional representation, we introduce an occupancy-diffusion model with voxel-level conditional guidance and force closure score filtering. In the second stage, we develop several energy functions and ranking strategies for optimization based on sparse IBS to generate high-quality dexterous grasp poses. Extensive experiments in both simulated and real-world settings validate the effectiveness of our approach, demonstrating its capability to mitigate collisions while maintaining a high grasp success rate across diverse objects and complex scenes.



Paperid:4722
Authors:Hengyu Liu, Chenxin Li, Zhengxin Li, Yipeng Wu, Wuyang Li, Zhiqin Yang, Zhenyuan Zhang, Yunlong Lin, Sirui Han, Brandon Feng
Abstract:
Vision-language models (VLMs) excel at descriptive tasks, but whether they truly understand scenes from visual observations remains uncertain. We introduce IR3D-Bench, a benchmark challenging VLMs to demonstrate understanding through active creation rather than passive recognition. Grounded in the analysis-by-synthesis paradigm, IR3D-Bench tasks Vision-Language Agents (VLAs) with actively using programming and rendering tools to recreate the underlying 3D structure of an input image, achieving agentic inverse rendering through tool use. This ''understanding-by-creating'' approach probes the tool-using generative capacity of VLAs, moving beyond the descriptive or conversational capacity measured by traditional scene understanding benchmarks. We provide a comprehensive suite of metrics to evaluate geometric accuracy, spatial relations, appearance attributes, and overall plausibility. Initial experiments on agentic inverse rendering powered by various state-of-the-art VLMs highlight current limitations, particularly in visual precision rather than basic tool usage. IR3D-Bench, including data and evaluation protocols, is released to facilitate systematic study and development of tool-using VLAs towards genuine scene understanding by creating.



Paperid:4723
Authors:Enneng Yang, Anke Tang, Li Shen, Guibing Guo, Xingwei Wang, Xiaochun Cao, Jie Zhang
Abstract:
Model merging integrates multiple expert models with diverse capabilities into a unified framework, facilitating collaborative learning. However, most existing methods assume simultaneous access to all models, which is often impractical in real-world scenarios where models are received sequentially. While some studies have investigated continual model merging (CMM)--which involves sequentially merging multiple models--the challenge of balancing prior knowledge (stability) and incorporating new tasks (plasticity) remains unresolved. This paper, for the first time, formally defines the stability and plasticity of CMM from the perspective of orthogonal projection. Subsequently, we analyze the relationships among the spaces spanned by task data, historical gradients, and accumulated gradients. Building on this, we propose a data-free \textbf{D}ual \textbf{O}rthogonal \textbf{P}rojection (DOP) method, which eliminates data dependence and mitigates interference between the merged model and models for old and new tasks by projecting their parameter differences onto their respective approximate data spaces. Finally, to solve potential conflicts between stability and plasticity, we reformulate DOP as a multi-objective optimization problem and employ a multi-gradient descent algorithm to obtain a Pareto-optimal solution. Extensive experiments across multiple architectures and task configurations validate that our approach significantly outperforms state-of-the-art CMM methods.



Paperid:4724
Authors:Guowei Wang, Fan Lyu, Changxing Ding
Abstract:
Existing test-time adaptation (TTA) methods primarily focus on scenarios involving domain shifts in a single modality. However, they often prove ineffective when multiple modalities simultaneously undergo domain shifts, as they struggle to identify and utilize reliable samples within testing batches amid severe prediction bias. To address this problem, we propose Partition-Then-Adapt (PTA), a novel approach combating prediction bias for TTA with multi-modal domain shifts. PTA comprises two key components: Partition and Debiased Reweighting (PDR) and multi-modal Attention-Guided Alignment (AGA). Specifically, PDR evaluates each sample’s predicted label frequency relative to the batch average, partitioning the batch into potential reliable and unreliable subsets. It then reweights each sample by jointly assessing its bias and confidence levels through a quantile-based approach. By applying weighted entropy loss, PTA simultaneously promotes learning from reliable subsets and discourages reliance on unreliable ones. Moreover, AGA regularizes PDR to focus on semantically meaningful multi-modal cues. Extensive experiments validate the effectiveness of PTA, surpassing state-of-the-art method by 6.1% on Kinetics50-C and 5.8% on VGGSound-C, respectively. Code of this paper will be released.



Paperid:4725
Authors:Daniel D. Richman, Jessica Karaguesian, Carl-Mikael Suomivuori, Ron Dror
Abstract:
The function of biomolecules, such as proteins, depends on their ability to interconvert between a wide range of structures or ``conformations.’’ Researchers have endeavored for decades to develop computational methods to predict the distribution of conformations, which is far harder to determine experimentally than a static folded structure. We present ConforMix, an inference-time algorithm that enhances sampling of conformational distributions using a combination of classifier guidance, filtering, and free energy estimation. Our approach upgrades diffusion models---whether trained for static structure prediction or conformational generation---to enable more efficient discovery of conformational variability without requiring prior knowledge of major degrees of freedom. ConforMix is orthogonal to improvements in model pretraining and would benefit even a hypothetical model that perfectly reproduced the Boltzmann distribution. Remarkably, when applied to a diffusion model trained for static structure prediction, ConforMix captures structural changes including domain motion, cryptic pocket flexibility, transporter cycling, and partial unfolding, while avoiding unphysical states. Case studies of large, biologically critical proteins, including GPCRs and transporters, demonstrate the scalability, accuracy, and utility of this method.



Paperid:4726
Authors:Janet Wang, Yunbei Zhang, Zhengming Ding, Jihun Hamm
Abstract:
Paucity of medical data severely limits the generalizability of diagnostic ML models, as the full spectrum of disease variability can not be represented by a small clinical dataset. To address this, diffusion models (DMs) have been considered as a promising avenue for synthetic image generation and augmentation. However, they frequently producemedically inaccurateimages, deteriorating the model performance. Expert domain knowledge is critical for synthesizing images that correctly encode clinical information, especially when data is scarce and quality outweighs quantity. Existing approaches for incorporating human feedback, such as reinforcement learning (RL) and Direct Preference Optimization (DPO), rely on robust reward functions or demand labor-intensive expert evaluations. Recent progress in Multimodal Large Language Models (MLLMs) reveals their strong visual reasoning capabilities, making them adept candidates as evaluators. In this work, we propose a novel framework, coined MAGIC (MedicallyAccurateGeneration ofImages through AI-ExpertCollaboration), that synthesizes clinically accurate skin disease images for data augmentation. Our method creatively translates expert-defined criteria into actionable feedback for image synthesis of DMs, significantly improving clinical accuracy while reducing the direct human workload. Experiments demonstrate that our method greatly improves the clinical quality of synthesized skin disease images, with outputs aligning with dermatologist assessments. Additionally, augmenting training data with these synthesized images improves diagnostic accuracy by +9.02% on a challenging 20-condition skin disease classification task, and by +13.89% in the few-shot setting.



Paperid:4727
Authors:Haitong Ma, Haoran Yu, Haobo Fu, Shuai Li
Abstract:
Reinforcement learning (RL) for collaborative agents capable of cooperating with humans to accomplish tasks has long been a central goal in the RL community. While prior approaches have made progress in adapting collaborative agents to diverse human partners, they often focus solely on optimizing task performance and overlook human preferences—despite the fact that such preferences often diverge from the reward-maximization objective of the environment.Addressing this discrepancy poses significant challenges: humans typically provide only a small amount of offline, preference-related feedback and are unable to engage in online interactions, resulting in a distributional mismatch between the agent’s online learning process and the offline human data. To tackle this, we formulate the problem as an online&offline reinforcement learning problem that jointly integrates online generalization and offline preference learning, entirely under an offline training regime.We propose a simple yet effective training framework built upon existing RL algorithms that alternates between offline preference learning and online generalization recovery, ensuring the stability and alignment of both learning objectives.We evaluate our approach on a benchmark built upon the Overcooked environment—a standard environment for human-agent collaboration—and demonstrate remarkable performance across diverse preference styles and cooperative scenarios.



Paperid:4728
Authors:Menghao Zhang, Pengfei Ren, Kangheng Lin, Qi Qi, Haifeng Sun, Huazheng Wang, Zirui Zhuang, Lei Zhang, Jianxin Liao, Jingyu Wang
Abstract:
Large Vision-Language Models (LVLMs) pretrained on large-scale multimodal data have shown promising capabilities in Video Anomaly Detection (VAD). However, their ability to reason about abnormal events based on scene semantics remains underexplored. In this paper, we investigate LVLMs’ behavior in VAD from a visual-textual co-occurrence perspective, focusing on whether their decisions are driven by statistical shortcuts between visual instances and textual phrases. By analyzing visual-textual co-occurrence in pretraining data and conducting experiments under different data settings, we reveal a hallucination phenomenon: LVLMs tend to rely on co-occurrence patterns between visual instances and textual phrases associated with either normality or abnormality, leading to incorrect predictions when these high-frequency objects appear in semantically mismatched contexts. To address this issue, we propose VAD-DPO, a direct preference optimization method supervised with counter-example pairs. By constructing visually similar but semantically contrasting video clips, VAD-DPO encourages the model to align its predictions with the semantics of scene rather than relying on co-occurrence patterns. Extensive experiments on six benchmark datasets demonstrate the effectiveness of VAD-DPO in enhancing both anomaly detection and reasoning performance, particularly in scene-dependent scenarios.



Paperid:4729
Authors:Yifeng Jiao, Yuchen Liu, Yu Zhang, Xin Guo, Yushuai Wu, Chen Jiang, Jiyang Li, Hongwei Zhang, LIMEI HAN, Xin Gao, Yuan Qi, Yuan Cheng
Abstract:
The advent of single-cell Assay for Transposase-Accessible Chromatin using sequencing (scATAC-seq) offers an innovative perspective for deciphering regulatory mechanisms by assembling a vast repository of single-cell chromatin accessibility data. While foundation models have achieved significant success in single-cell transcriptomics, there is currently no foundation model for scATAC-seq that supports zero-shot high-quality cell identification and comprehensive multi-omics analysis simultaneously. Key challenges lie in the high dimensionality and sparsity of scATAC-seq data, as well as the lack of a standardized schema for representing open chromatin regions (OCRs). Here, we present ChromFound, the pioneering foundation model tailored for scATAC-seq. ChromFound utilizes a hybrid architecture and genome-aware tokenization to effectively capture genome-wide long contexts and dynamic chromatin landscapes. Pretrained on a dataset of 1.97 million cells across 30 human tissue types and 6 disease categories, ChromFound showcases its versatility by excelling in 6 diverse tasks. Notably, it demonstrates robust zero-shot performance in generating universal cell representations and exhibits excellent transferability in cell-type annotation and cross-omics prediction. By accurately predicting gene-enhancer connections, ChromFound paves the way for interpreting the functions of various disease risk in the noncoding genome.



Paperid:4730
Authors:Tobias Lorenz, Marta Kwiatkowska, Mario Fritz
Abstract:
Data errors, corruptions, and poisoning attacks during training pose a major threat to the reliability of modern AI systems. While extensive effort has gone into empirical mitigations, the evolving nature of attacks and the complexity of data require a more principled, provable approach to robustly learn on such data - and to understand how perturbations influence the final model. Hence, we introduce MIBP-Cert, a novel certification method based on Mixed Integer Bilinear Programming (MIBP) that computes sound, deterministic bounds to provide provable robustness even under complex threat models. By computing the set of parameters reachable through perturbed or manipulated data, we can predict all possible outcomes and guarantee robustness. To make solving this optimization problem tractable, we propose a novel relaxation scheme that bounds each training step without sacrificing soundness. We demonstrate the applicability of our approach to continuous and discrete data, as well as different threat models - including complex ones that were previously out of reach.



Paperid:4731
Authors:Hao Chen, Jiaming Liu, Chenyang Gu, Zhuoyang Liu, Renrui Zhang, Xiaoqi Li, Xiao He, Yandong Guo, Chi-Wing Fu, Shanghang Zhang, Pheng-Ann Heng
Abstract:
Generalized policy and execution efficiency constitute the two critical challenges in robotic manipulation. While recent foundation policies benefit from the common-sense reasoning capabilities of internet-scale pretrained vision-language models (VLMs), they often suffer from low execution frequency. To mitigate this dilemma, dual-system approaches, inspired by Kahneman’s theory, have been proposed to leverage a VLM-based System 2 model handling high-level reasoning and a separate System 1 action model ensuring real-time control. However, existing designs maintain both systems as separate models, limiting System 1 from fully leveraging the rich pretrained knowledge from the VLM-based System 2. In this work, we propose Fast-in-Slow (FiS), a unified dual-system vision-language-action (VLA) model that embeds the System 1 execution module within the VLM-based System 2 by partially sharing parameters. This innovative paradigm not only enables high-frequency execution in System 1, but also facilitates coordination between the reasoning and execution components within a single foundation model of System 2. Given their fundamentally distinct roles within FiS-VLA, we design the two systems to incorporate heterogeneous modality inputs alongside asynchronous operating frequencies, enabling both fast and precise manipulation. To enable coordination between the two systems, a dual-aware co-training strategy is proposed that equips System 1 with action generation capabilities while preserving System 2’s contextual reasoning representation. For evaluation, FiS-VLA outperforms previous state-of-the-art methods by 8% in simulation and 11% in real-world tasks in terms of average success rate, while achieving a 21.9 Hz control frequency without action chunking mechanism.



Paperid:4732
Authors:Jinhua Yin, Peiru Yang, Chen Yang, Huili Wang, Zhiyang Hu, Shangguang Wang, Yongfeng Huang, Tao Qi
Abstract:
Large vision-language models (LVLMs) derive their capabilities from extensive training on vast corpora of visual and textual data. Empowered by large-scale parameters, these models often exhibit strong memorization of their training data, rendering them susceptible to membership inference attacks (MIAs).Existing MIA methods for LVLMs typically operate under white- or gray-box assumptions, by extracting likelihood-based features for the suspected data samples based on the target LVLMs. However, mainstream LVLMs generally only expose generated outputs while concealing internal computational features during inference, limiting the applicability of these methods.In this work, we propose the first black-box MIA framework for LVLMs, based on a prior knowledge-calibrated memory probing mechanism. The core idea is to assess the model memorization of the private semantic information embedded within the suspected image data, that is unlikely to be inferred from general world knowledge alone.We conduct extensive experiments across four LVLMs and three datasets.Empirical results demonstrate that our method effectively identifies training data of LVLMs in a purely black-box setting and even achieves performance comparable to gray-box and white-box methods.Further analysis reveals the robustness of our method against potential adversarial manipulations, and the effectiveness of the methodology designs.Our code and data are available at \url{https://anonymous.4open.science/r/KCMP-2D2C}



Paperid:4733
Authors:Jingxin Zhan, Yuchen Xin, Chenjie Sun, Zhihua Zhang
Abstract:
Abstract:We consider a common case of the combinatorial semi-bandit problem, the $m$-set semi-bandit, where the learner exactly selects $m$ arms from the total $d$ arms. In the adversarial setting, the best regret bound, known to be $\mathcal{O}(\sqrt{nmd})$ for time horizon $n$, is achieved by the well-known Follow-the-Regularized-Leader (FTRL) policy. However, this requires to explicitly compute the arm-selection probabilities via optimizing problems at each time step and sample according to them. This problem can be avoided by the Follow-the-Perturbed-Leader (FTPL) policy, which simply pulls the $m$ arms that rank among the $m$ smallest (estimated) loss with random perturbation. In this paper, we show that FTPL with a Fréchet perturbation also enjoys the near optimal regret bound $\mathcal{O}(\sqrt{nmd\log(d)})$ in the adversarial setting and approaches best-of-both-world regret bounds, i.e., achieves a logarithmic regret for the stochastic setting. Moreover, our lower bound shows that the logarithmic factor is unavoidable with our approach; any improvement would require a fundamentally different and more challenging method.



Paperid:4734
Authors:Yuzheng Hu, Fan Wu, Haotian Ye, David Forsyth, James Zou, Nan Jiang, Jiaqi Ma, Han Zhao
Abstract:
Online reinforcement learning (RL) excels in complex, safety-critical domains, yet it faces challenges such as sample inefficiency, training instability, and a lack of interpretability. Data attribution offers a principled way to trace model behavior back to individual training samples.However, in online RL, each training sample not only drives policy updates but also influences future data collection, violating the fixed dataset assumption in existing attribution methods. In this paper, we initiate the study of data attribution for online RL, focusing on the widely used Proximal Policy Optimization (PPO) algorithm. We start by establishing alocalattribution framework, interpreting model checkpoints with respect to the records in the recent training buffer. We design two target functions, capturing agent action and cumulative return respectively, and measure each record's contribution through gradient similarity between its training loss and these targets. We demonstrate the power of this framework through three concrete applications: diagnosis of learning, temporal analysis of behavior formation, and targeted intervention during training. Leveraging this framework, we further propose an algorithm, iterative influence-based filtering (IIF), for online RL training that iteratively performs experience filtering to refine policy updates. Across standard RL benchmarks (classic control, navigation, locomotion) to RLHF for large language models, IIF reduces sample complexity, speeds up training, and achieves higher returns. Overall, these results advance interpretability, efficiency, and effectiveness of online RL.



Paperid:4735
Authors:chaofan gan, Yuanpeng Tu, Xi Chen, Tieyuan Chen, Yuxi Li, Mehrtash Harandi, Weiyao Lin
Abstract:
Pre-trained stable diffusion models (SD) have shown great advances in visual correspondence. In this paper, we investigate the capabilities of Diffusion Transformers (DiTs) for accurate dense correspondence. Distinct from SD, DiTs exhibit a critical phenomenon in which very few feature activations exhibit significantly larger values than others, known as massive activations, leading to uninformative representations and significant performance degradation for DiTs.The massive activations consistently concentrate at very few fixed dimensions across all image patch tokens, holding little local information. We trace these dimension-concentrated massive activations and find that such concentration can be effectively localized by the zero-initialized Adaptive Layer Norm (AdaLN-zero).Building on these findings, we propose Diffusion Transformer Feature (DiTF), a training-free framework designed to extract semantic-discriminative features from DiTs. Specifically, DiTF employs AdaLN to adaptively localize and normalize massive activations with channel-wise modulation. In addition, we develop a channel discard strategy to further eliminate the negative impacts from massive activations. Experimental results demonstrate that our DiTF outperforms both DINO and SD-based models and establishes a new state-of-the-art performance for DiTs in different visual correspondence tasks (e.g., with +9.4% on Spair-71k and +4.4% on AP-10K-C.S.).



Paperid:4736
Authors:Shi Chen, Erik Sandström, Sandro Lombardi, Siyuan Li, Martin R. Oswald
Abstract:
Achieving truly practical dynamic 3D reconstruction requires online operation, global pose and map consistency, detailed appearance modeling, and the flexibility to handle both RGB and RGB-D inputs. However, existing SLAM methods typically merely remove the dynamic parts or require RGB-D input, while offline methods are not scalable to long video sequences, and current transformer-based feedforward methods lack global consistency and appearance details. To this end, we achieve online dynamic scene reconstruction by disentangling the static and dynamic parts within a SLAM system. The poses are tracked robustly with a novel motion masking strategy, and dynamic parts are reconstructed leveraging a progressive adaptation of a Motion Scaffolds graph. Our method yields novel view renderings competitive to offline methods and achieves on-par tracking with state-of-the-art dynamic SLAM methods.



Paperid:4737
Authors:Da Wang, Yi Ma, Ting Guo, Hongyao Tang, Wei Wei, Jiye Liang
Abstract:
Offline reinforcement learning (RL) aims to learn optimal policies from static datasets while enhancing generalization to out-of-distribution (OOD) data. To mitigate overfitting to suboptimal behaviors in offline datasets, existing methods often relax constraints on policy and data or extract informative patterns through data-driven techniques. However, there has been limited exploration into structurally guiding the optimization process toward flatter regions of the solution space that offer better generalization. Motivated by this observation, we present \textit{FANS}, a generalization-oriented structured network framework that promotes flatter and robust policy learning by guiding the optimization trajectory through modular architectural design. FANS comprises four key components: (1) Residual Blocks, which facilitate compact and expressive representations; (2) Gaussian Activation, which promotes smoother gradients; (3) Layer Normalization, which mitigates overfitting; and (4) Ensemble Modeling, which reduces estimation variance. By integrating FANS into a standard actor-critic framework, we highlight that this remarkably simple architecture achieves superior performance across various tasks compared to many existing advanced methods. Moreover, we validate the effectiveness of FANS in mitigating overestimation and promoting generalization, demonstrating the promising potential of architectural design in advancing offline RL.



Paperid:4738
Authors:Ziyi Zhang, Yorie Nakahira, Guannan Qu
Abstract:
Abstract:We study the problem of stabilizing an unknown partially observable linear time-invariant (LTI) system. For fully observable systems, leveraging an unstable/stable subspace decomposition approach, state-of-art sample complexity is independent from system dimension $n$ and only scales with respect to the dimension of the unstable subspace. However, it remains open whether such sample complexity can be achieved for partially observable systems because such systems do not admit a uniquely identifiable unstable subspace. In this paper, we propose LTS-P, a novel technique that leverages compressed singular value decomposition (SVD) on the ''lifted'' Hankel matrix to estimate the unstable subsystem up to an unknown transformation. Then, we design a stabilizing controller that integrates a robust stabilizing controller for the unstable mode and a small-gain-type assumption on the stable subspace. We show that LTS-P stabilizes unknown partially observable LTI systems with state-of-the-art sample complexity that is dimension-free and only scales with the number of unstable modes, which significantly reduces data requirements for high-dimensional systems with many stable modes.



Paperid:4739
Authors:Yunghee Lee, Byeonghyun Pak, Junwha Hong, Hoseong Kim
Abstract:
In this paper, we propose Tortoise and Hare Guidance (THG), a training-free strategy that accelerates diffusion sampling while maintaining high-fidelity generation. We demonstrate that the noise estimate and the additional guidance term exhibit markedly different sensitivity to numerical error by reformulating the classifier-free guidance (CFG) ODE as a multirate system of ODEs. Our error-bound analysis shows that the additional guidance branch is more robust to approximation, revealing substantial redundancy that conventional solvers fail to exploit. Building on this insight, THG significantly reduces the computation of the additional guidance: the noise estimate is integrated with the tortoise equation on the original, fine-grained timestep grid, while the additional guidance is integrated with the hare equation only on a coarse grid. We also introduce (i) an error-bound-aware timestep sampler that adaptively selects step sizes and (ii) a guidance-scale scheduler that stabilizes large extrapolation spans. THG reduces the number of function evaluations (NFE) by up to 30% with virtually no loss in generation fidelity (∆ImageReward ≤ 0.032) and outperforms state-of-the-art CFG-based training-free accelerators under identical computation budgets. Our findings highlight the potential of multirate formulations for diffusion solvers, paving the way for real-time high-quality image synthesis without any model retraining. The source code is available at https://github.com/Tortoise-and-Hare-Guidance/THG.



Paperid:4740
Authors:Sourav Ganguly, Arnob Ghosh, Kishan Panaganti, Adam Wierman
Abstract:
Abstract:Constrained decision-making is essential for designing safe policies in real-world control systems, yet simulated environments often fail to capture real-world adversities. We consider the problem of learning a policy that will maximize the cumulative reward while satisfying a constraint, even when there is a mismatch between the real model and an accessible simulator/nominal model. In particular, we consider the robust constrained Markov decision problem (RCMDP) where an agent needs to maximize the reward and satisfy the constraint against the worst possible stochastic model under the uncertainty set centered around an unknown nominal model. Primal-dual methods, effective for standard constrained MDP (CMDP), are not applicable here because of the lack of the strong duality property. Further, one cannot apply the standard robust value-iteration based approach on the composite value function, either, as the worst-case models may be different for the reward value function and the constraint value function. We propose a novel technique that effectively minimizes the constraint value function--to satisfy the constraints; on the other hand, when all the constraints are satisfied, it can simply maximize the robust reward value function. We prove that such an algorithm finds a policy with at most $\epsilon$ sub-optimality and a feasible policy after $O(\epsilon^{-2})$ iterations. In contrast to the state-of-the-art method, we do not need to employ a binary search, thus, we reduce the computation time by at least 4x for smaller values of the discount factor ($\gamma$) and by at least 6x for larger values of $\gamma$.



Paperid:4741
Authors:Erfan Baghaei Potraghloo, Armin Azizi, Souvik Kundu, Massoud Pedram
Abstract:
Large language models (LLMs), despite their impressive performance across a wide range of tasks, often struggle to balance two competing objectives in open-ended text generation: fostering diversity and creativity while preserving logical coherence. Existing truncated sampling techniques, including temperature scaling, top-p(nucleus) sampling, and min-psampling, aim to manage this trade-off. However, they exhibit limitations, particularly in the effective incorporation of the confidence of the model into the corresponding sampling strategy. For example, min-psampling relies on a single top token as a heuristic for confidence, eventually underutilizing the information of the probability distribution. Toward effective incorporation of the confidence of the model, in this paper, we presenttop-Hdecoding. We first establish the theoretical foundation of the interplay between creativity and coherence in truncated sampling by formulating anentropy-constrained minimum divergenceproblem. We then prove this minimization problem to be equivalent to anentropy-constrained mass maximization(ECMM) problem, which is NP-hard. Finally, we present top-H decoding, a computationally efficient greedy algorithm to solve the ECMM problem. Extensive empirical evaluations demonstrate that top-H outperforms the state-of-the-art (SoTA) alternative of min-psampling by up to25.63%on creative writing benchmarks, while maintaining robustness on question-answering datasets such as GPQA, GSM8K, and MT-Bench. Additionally, anLLM-as-judgeevaluation confirms that top-H indeed produces coherent outputs even at higher temperatures, where creativity is especially critical. In summary, top-H advances SoTA in open-ended text generation and can beeasily integratedinto creative writing applications. We will open-source the code soon.



Paperid:4742
Authors:Ziqi Wang, Jiashun Liu, Ling Pan
Abstract:
Traditional continuous deep reinforcement learning (RL) algorithms employ deterministic or unimodal Gaussian actors, which cannot express complex multimodal decision distributions. This limitation can hinder their performance in diversity-critical scenarios. There have been some attempts to design online multimodal RL algorithms based on diffusion or amortized actors. However, these actors are intractable, making existing methods struggle with balancing performance, decision diversity, and efficiency simultaneously. To overcome this challenge, we first reformulate existing intractable multimodal actors within a unified framework, and prove that they can be directly optimized by policy gradient via reparameterization. Then, we propose a distance-based diversity regularization that does not explicitly require decision probabilities. We identify two diversity-critical domains, namely multi-goal achieving and generative RL, to demonstrate the advantages of multimodal policies and our method, particularly in terms of few-shot robustness. In conventional MuJoCo benchmarks, our algorithm also shows competitive performance. Moreover, our experiments highlight that the amortized actor is a promising policy model class with strong multimodal expressivity and high performance.



Paperid:4743
Authors:Longtian Qiu, Shan Ning, Jiaxuan Sun, Xuming He
Abstract:
Reinforcement learning (RL) has shown promise in enhancing the general Chain-of-Thought (CoT) reasoning capabilities of multimodal large language models (MLLMs). However, when applied to improve general CoT reasoning, existing RL frameworks often struggle to generalize beyond the training distribution. To address this, we propose NoisyGRPO, a systematic multimodal RL framework that introduces controllable noise into visual inputs for enhanced exploration and explicitly models the advantage estimation process via a Bayesian framework.Specifically, NoisyGRPO improves RL training by: (1) \textbf{Noise-Injected Exploration Policy}: Perturbing visual inputs with Gaussian noise to encourage exploration across a wider range of visual scenarios; and (2) \textbf{Bayesian Advantage Estimation}: Formulating advantage estimation as a principled Bayesian inference problem, where the injected noise level serves as a prior and the observed trajectory reward as the likelihood. This Bayesian modeling fuses both sources of information to compute a robust posterior estimate of trajectory advantage, effectively guiding MLLMs to prefer visually grounded trajectories over noisy ones.Experiments on standard CoT quality, general capability, and hallucination benchmarks demonstrate that NoisyGRPO substantially improves generalization and robustness, especially in RL settings with small-scale MLLMs such as Qwen2.5-VL 3B.



Paperid:4744
Authors:Anas Barakat, Souradip Chakraborty, Peihong Yu, Pratap Tokekar, Amrit Singh Bedi
Abstract:
Reinforcement learning with general utilities (RLGU) offers a unifying framework to capture several problems beyond standard expected returns, including imitation learning, pure exploration, and safe RL. Despite recent fundamental advances in the theoretical analysis of policy gradient (PG) for standard RL and recent efforts in RLGU, the understanding of PG methods and their scope of application in RLGU still remain limited. In this work, we establish global optimality guarantees of PG methods for RLGU in which the objective is a general concave utility function of the state-action occupancy measure. In the tabular setting, we provide global optimality results using a new proof technique building on recent theoretical developments on the convergence of PG methods for standard RL using gradient domination. Our proof technique opens avenues for analyzing policy parameterizations beyond the direct policy parameterization for RLGU. In addition, we provide global optimality results for large state action space settings beyond prior work which has mostly focused on the tabular setting. In this large scale setting, we adapt PG methods by approximating occupancy measures within a function approximation class using maximum likelihood estimation. Our sample complexity only scales with the dimension of our function approximation class rather than the size of the state action space.



Paperid:4745
Authors:Yongmin Lee, Hye Won Chung
Abstract:
Multimodal dataset distillation aims to synthesize a small set of image-text pairs that enables efficient training of large-scale vision-language models. While dataset distillation has shown promise in unimodal tasks, extending it to multimodal contrastive learning presents key challenges: learning cross-modal alignment and managing the high computational cost of large encoders. Prior approaches address scalability by freezing the text encoder and update only the image encoder and text projection layer. However, we find this severely limits semantic alignment and becomes a bottleneck for performance scaling.We propose CovMatch, a scalable dataset distillation framework that aligns the cross-covariance of real and synthetic features while regularizing feature distributions within each modality. Unlike prior approaches, CovMatch enables joint optimization of both encoders, leading to stronger cross-modal alignment and improved performance. Evaluated on Flickr30K and COCO, CovMatch outperforms state-of-the-art multimodal distillation methods and achieves up to 6.8\% absolute gains in retrieval accuracy using only 500 synthetic pairs.



Paperid:4746
Authors:Jintao Zhang, Jia wei, Pengle Zhang, Xiaoming Xu, Haofeng Huang, Haoxu Wang, Kai Jiang, Jun Zhu, Jianfei Chen
Abstract:
Abstract:The efficiency of attention is important due to its quadratic time complexity. We enhance the efficiency of attention through two key contributions: First, we leverage the new $\texttt{FP4}$ Tensor Cores in Blackwell GPUs to accelerate attention computation. Our implementation achieves $\textbf{1038}$ $\texttt{TOPS}$ on $\texttt{RTX5090}$, which is a $\textbf{5}\times$ speedup over the fastest FlashAttention on $\texttt{RTX5090}$. Experiments show that our $\texttt{FP4}$ attention can accelerate inference of various models in a plug-and-play way. Second, we pioneer low-bit attention to training tasks. Existing low-bit attention works like FlashAttention3 and SageAttention focus only on inference. However, the efficiency of training large models is also important. To explore whether low-bit attention can be effectively applied to training tasks, we design an accurate and efficient $\texttt{8-bit}$ attention for both forward and backward propagation. Experiments indicate that $\texttt{8-bit}$ attention achieves lossless performance in fine-tuning tasks but exhibits slower convergence in pretraining tasks.



Paperid:4747
Authors:Shengming Yuan, Xinyu Lyu, Shuailong Wang, Jingkuan Song, Beitao Chen, Lianli Gao
Abstract:
Multimodal large language models (MLLMs) face an inherent trade-off between faithfulness and creativity, as different tasks require varying degrees of associative reasoning. However, existing methods lack the flexibility to modulate this reasoning strength, limiting MLLMs' adaptability across factual and creative scenarios. To bridge this gap, we propose equipping MLLMs with mechanisms that enable flexible control over associative reasoning. We begin by investigating the internal mechanisms underlying associative behavior in MLLMs and find that: (1) middle layers play a pivotal role in shaping model’s associative tendencies, (2) modifying representations in these layers effectively regulates associative reasoning strength, and (3) hallucinations can be exploited to derive steering vectors that guide this modulation. Building on these findings, we introduce Flexible Association Control (FlexAC), a lightweight and training-free framework for modulating associative behavior in MLLMs. FlexAC first induces hallucination-guided intermediate representations to encode associative directions. Then, it selects high-association instances to construct effective associative steering vectors, whose strengths are adaptively calibrated to balance creative guidance with output stability. Finally, recognizing the multi-dimensional nature of associative reasoning, FlexAC incorporates task-specific associative vectors derived from a forward pass on a few target-domain samples, enabling models to follow diverse associative directions and better adapt to creative tasks. Notably, our method achieves up to a 5.8× improvement in creativity on Creation-MMBench and a 29\% reduction in hallucination rate on CHAIR, surpassing existing baselines and demonstrating its effectiveness in enabling flexible control over associative reasoning in MLLMs.



Paperid:4748
Authors:Alireza Mousavi-Hosseini, Clayton Sanford, Denny Wu, Murat Erdogdu
Abstract:
Abstract:Theoretical efforts to prove advantages of Transformers in comparison with classical architectures such as feedforward and recurrent neural networks have mostly focused on representational power. In this work, we take an alternative perspective and prove that even with infinite compute, feedforward and recurrent networks may suffer from larger sample complexity compared to Transformers, as the latter can adapt to a form of dynamic sparsity. Specifically, we consider a sequence-to-sequence data generating model on sequences of length $N$, where the output at each position only depends on $q \ll N$ relevant tokens, and the positions of these tokens are described in the input prompt. We prove that a single-layer Transformer can learn this model if and only if its number of attention heads is at least $q$, in which case it achieves a sample complexity almost independent of $N$, while recurrent networks require $N^{\Omega(1)}$ samples on the same problem. If we simplify this model, recurrent networks may achieve a complexity almost independent of $N$, while feedforward networks still require $N$ samples. Our proposed sparse retrieval model illustrates a natural hierarchy in sample complexity across these architectures.



Paperid:4749
Authors:Sarah Cen, Salil Goyal, Zaynah Javed, Ananya Karthik, Percy Liang, Daniel Ho
Abstract:
Model audits play a critical role in holding AI actors accountable, and one branch of the law for which model audits are particularly salient is discrimination law. Several areas of discrimination law (including but extending beyond employment) implicate what is known as the "less discriminatory alternatives" (LDA) framework, in which a policy (i.e., model) is defensible if no LDA can be found with a reasonable amount of effort. Notably, the burden of proving that an LDA exists typically falls on the claimant (the party alleging discrimination) rather than the defendant. This creates a significant hurdle, as claimants would seemingly need to produce a less discriminatory model, a task requiring resources and expertise beyond most litigants. Moreover, model developers generally shield any information about the current model and training data as trade secrets. In this work, we propose a novel toolkit enabling claimants to establish the existence of LDAs even if they lack significant compute, training data, or information about the current model. Our method allows claimants to demonstrate that (obtainable) less discriminatory models exist without producing them, requiring substantially fewer resources and (potentially proprietary) information than model development. We do so by establishing a scaling law for the loss-fairness Pareto frontier (PF). As our main result, we provide a closed-form upper bound for the PF, where we focus on fairness as demographic parity as an illustrative example. Using this expression, the claimant can fit a curve to the PF in the "low resource-information regime," then extrapolate the PF that applies for the (large) model being contested. The expression thus serves as a scaling law for loss-fairness PFs.



Paperid:4750
Authors:Matthew Fahrbach, Srikumar Ramalingam, Morteza Zadimoghaddam, Sara Ahmadian, Gui Citovsky, Giulia DeSalvo
Abstract:
Abstract:This work studies a novel subset selection problem called *max-min diversification with monotone submodular utility* (MDMS), which has a wide range of applications in machine learning, e.g., data sampling and feature selection.Given a set of points in a metric space,the goal of MDMS is to maximize $f(S) = g(S) + \lambda \cdot \text{div}(S)$subject to a cardinality constraint $|S| \le k$,where$g(S)$ is a monotone submodular functionand$\text{div}(S) = \min_{u,v \in S : u \ne v} \text{dist}(u,v)$ is the *max-min diversity* objective.We propose the `GIST` algorithm, which gives a $\frac{1}{2}$-approximation guarantee for MDMSby approximating a series of maximum independent set problems with a bicriteria greedy algorithm.We also prove that it is NP-hard to approximate within a factor of $0.5584$.Finally, we show in our empirical study that `GIST` outperforms state-of-the-art benchmarksfor a single-shot data sampling task on ImageNet.



Paperid:4751
Authors:Chenao Li, Shuo Yan, Enyan Dai
Abstract:
Enzyme-catalyzed protein cleavage is essential for many biological functions. Accurate prediction of cleavage sites can facilitate various applications such as drug development, enzyme design, and a deeper understanding of biological mechanisms. However, most existing models are restricted to an individual enzyme, which neglects shared knowledge of enzymes and fails to generalize to novel enzymes. Thus, we introduce a unified protein cleavage site predictor named UniZyme, which can generalize across diverse enzymes. To enhance the enzyme encoding for the protein cleavage site prediction, UniZyme employs a novel biochemically-informed model architecture along with active-site knowledge of proteolytic enzymes. Extensive experiments demonstrate that UniZyme achieves high accuracy in predicting cleavage sites across a range of proteolytic enzymes, including unseen enzymes. The code is available in https://anonymous.4open.science/r/UniZyme-4A67



Paperid:4752
Authors:Patrick Kahardipraja, Reduan Achtibat, Thomas Wiegand, Wojciech Samek, Sebastian Lapuschkin
Abstract:
Large language models are able to exploit in-context learning to access external knowledge beyond their training data through retrieval-augmentation. While promising, its inner workings remain unclear. In this work, we shed light on the mechanism of in-context retrieval augmentation for question answering by viewing a prompt as a composition of informational components. We propose an attribution-based method to identify specialized attention heads, revealing in-context heads that comprehend instructions and retrieve relevant contextual information, and parametric heads that store entities’ relational knowledge. To better understand their roles, we extract function vectors and modify their attention weights to show how they can influence the answer generation process. Finally, we leverage the gained insights to trace the sources of knowledge used during inference, paving the way towards more safe and transparent language models.



Paperid:4753
Authors:Samet Demir, Zafer Dogan
Abstract:
Pretrained Transformers demonstrate remarkable in-context learning (ICL) capabilities, enabling them to adapt to new tasks from demonstrations without parameter updates. However, theoretical studies often rely on simplified architectures (e.g., omitting MLPs), data models (e.g., linear regression with isotropic inputs), and single-source training—limiting their relevance to realistic settings. In this work, we study ICL in pretrained Transformers with nonlinear MLP heads on nonlinear regression tasks drawn from multiple data sources with heterogeneous input, task, and noise distributions. We analyze a model where the MLP comprises two layers, with the first trained via a single gradient step and the second fully optimized. Under high-dimensional asymptotics, we prove that such models are equivalent in ICL error to structured polynomial predictors, leveraging results from Gaussian universality and polynomial approximation theory. This equivalence reveals that nonlinear MLPs meaningfully enhance ICL performance—particularly on nonlinear tasks—compared to linear baselines. It also enables a precise analysis of data mixing effects: we identify key properties of high-quality data sources (low noise, structured covariances) and show that feature learning emerges only when the task covariance exhibits sufficient structure. These results are validated empirically across various activation functions, model sizes, and data distributions. Overall, our work advances the theoretical foundations of ICL in Transformers and provides actionable insight into the role of architecture and data in ICL.



Paperid:4754
Authors:Dogyun Park, Taehoon Lee, Minseok Joo, Hyunwoo J. Kim
Abstract:
Abstract:Recently, Flow Matching models have pushed the boundaries of high-fidelity data generation across a wide range of domains. It typically employs a single large network to learn the entire generative trajectory from noise to data. Despite their effectiveness, this design struggles to capture distinct signal characteristics across timesteps simultaneously and incurs substantial inference costs due to the iterative evaluation of the entire model. To address these limitations, we propose Blockwise Flow Matching (BFM), a novel framework that partitions the generative trajectory into multiple temporal segments, each modeled by smaller but specialized velocity blocks. This blockwise design enables each block to specialize effectively in its designated interval, improving inference efficiency and sample quality. To further enhance generation fidelity, we introduce a Semantic Feature Guidance module that explicitly conditions velocity blocks on semantically rich features aligned with pretrained representations.Additionally, we propose a lightweight Feature Residual Approximation strategy that preserves semantic quality while significantly reducing inference cost.Extensive experiments on ImageNet 256$\times$256 demonstrate that BFM establishes a substantially improved Pareto frontier over existing Flow Matching methods, achieving 2.1$\times$ to 4.9$\times$ accelerations in inference complexity at comparable generation performance.



Paperid:4755
Authors:shengtian yang, Yue Feng, Yingshi Liu, Jingrou Zhang, Jie Qin
Abstract:
Video Anomaly Detection (VAD) aims to locate unusual activities or behaviors within videos. Recently, offline VAD has garnered substantial research attention, which has been invigorated by the progress in large language models (LLMs) and vision-language models (VLMs), offering the potential for a more nuanced understanding of anomalies. However, online VAD has seldomly received attention due to real-time constraints and computational intensity. In this paper, we introduce a novel \textbf{M}emory-based online scoring queue scheme for \textbf{T}raining-free VAD (MoniTor), to address the inherent complexities in online VAD. Specifically, MoniTor applies a streaming input to VLMs, leveraging the capabilities of pre-trained large-scale models. To capture temporal dependencies more effectively, we incorporate a novel prediction mechanism inspired by Long Short-Term Memory (LSTM) networks to ensure that the model can effectively model past states and leverage previous predictions to identify anomalous behaviors, thereby better understanding the current frame. Moreover, we design a scoring queue and an anomaly prior to dynamically store recent scores and cover all anomalies in the monitoring scenario, providing guidance for LLMs to distinguish between normal and abnormal behaviors over time.We evaluate MoniTor on two large datasets (i.e., UCF-Crime and XD-Violence) containing various surveillance and real-world scenarios. The results demonstrate that MoniTor outperforms state-of-the-art methods and is competitive with weakly supervised methods without training. Code will be available.



Paperid:4756
Authors:Myeongsoo Kim, Shweta Garg, Baishakhi Ray, Varun Kumar, Anoop Deoras
Abstract:
Programming assistants powered by large language models have transformed software development, yet most benchmarks focus narrowly on code generation tasks. Recent efforts like InfiBench and StackEval attempt to address this gap using Stack Overflow data but remain limited to single-turn interactions in isolated contexts, require significant manual curation, and fail to represent complete project environments. We introduce CodeAssistBench (CAB), the first benchmark framework for evaluating multi-turn programming assistance in realistic settings that address real-world questions about actual codebases. Unlike existing programming Q&A benchmarks, CAB automatically generates scalable datasets from question-related GitHub issues using configurable parameters (e.g., repository creation date, star count, programming languages), and includes automatic containerization of codebases for evaluation. It then evaluates models through simulated users in these containerized environments with full codebase access. Using this framework, we constructed a test set of 3,286 real-world programming questions across 231 repositories, spanning seven programming languages and diverse problem domains. Our evaluation of leading LLMs reveals a substantial capability gap: while models perform well on Stack Overflow questions with success rates of 70-83%, they resolve only up to 16.49% of CAB's recent issues. This discrepancy highlights the challenges of providing assistance in complex, project-specific contexts versus answering standalone questions. Our fully automated framework enables continuous benchmark expansion and is available at https://anonymous.4open.science/r/CAB-CBA3/.



Paperid:4757
Authors:Zekai Zhao, Qi Liu, Kun Zhou, Zihan Liu, Yifei Shao, Zhiting Hu, Biwei Huang
Abstract:
Despite the remarkable reasoning performance, eliciting the long chain-of-thought(CoT) ability in large language models(LLMs) typically requires costly reinforcement learning or supervised fine-tuning on high-quality distilled data. We investigate the internal mechanisms behind this capability and show that a small set of high-impact activations in the last few layers, greatly govern the long-form reasoning attributes, e.g. output length and self-reflection. Through simply amplifying these activations and adding ``wait'' tokens, the long CoT ability can be invoked without training, leading to significantly increased self-reflection rate and accuracy. In addition, we also find that the activation changes follow predictable trajectories, i.e. a sharp rise after special tokens and a subsequent exponential decay. Based on these insights, we introduce a general training-free activation control technique. It utilizes a few contrastive examples to identify the relevant activations, and then incorporates simple analytic functions to adjust their values at inference time to elicit long CoTs. Extensive experiments have verified the effectiveness of our methods in efficiently eliciting the long CoT ability of LLMs and improving the performance. Besides, we further propose a parameter-efficient fine-tuning method that trains only the last-layer activation amplification module and a few LoRA layers, outperforming LoRA on reasoning benchmarks with much fewer parameters. Our code and data will be fully public released.



Paperid:4758
Authors:Jingli Lin, Chenming Zhu, Runsen Xu, Xiaohan Mao, Xihui Liu, Tai WANG, Jiangmiao Pang
Abstract:
Recent advances in multimodal large language models (MLLMs) have shown remarkable capabilitiesin integrating vision and language for complex reasoning. While most existing benchmarks evaluate models under offline settings with a fixed set of pre-recorded inputs, we introduce OST-Bench, a benchmark designed to evaluate Online Spatio-Temporal understanding from the perspective of an agent actively exploring a scene. The “Online” aspect emphasizes the need to process and reason over incrementally acquired observations, while the “Spatio-Temporal” component requires integrating current visual inputs with historical memory to support dynamic spatial reasoning. OST-Bench better reflects the challenges of real-world embodied perception. Built on an efficient data collection pipeline, OST-Bench consists of 1.4k scenes and 10k question-answer pairs collected from ScanNet, Matterport3D, and ARKitScenes. We evaluate several leading MLLMs on OST-Bench and observe that they fall short on tasks requiring complex spatio-temporal reasoning. Under the online setting, their accuracy declines as the exploration horizon extends and the memory grows. Through further experimental analysis, we identify common error patterns across models and find that both complex clue-based spatial reasoning demands and long-term memory retrieval requirements significantly drop model performance along two separate axes, highlighting the core challenges that must be addressed to improve online embodied reasoning. To foster further research and development in the field, our codes, dataset, and benchmark are available at https://github.com/rbler1234/OST-Bench.



Paperid:4759
Authors:Chenyue Li, Wen Deng, Mengqian Lu, Binhang Yuan
Abstract:
The rapid advancements in large language models (LLMs), particularly in their reasoning capabilities, hold transformative potential for addressing complex challenges in atmospheric science. However, leveraging LLMs effectively in this domain requires a robust and comprehensive evaluation benchmark. Toward this end, we present AtmosSci-Bench, a novel benchmark designed to systematically assess LLM performance across five core categories of atmospheric science problems: hydrology, atmospheric dynamics, atmospheric physics, geophysics, and physical oceanography.AtmosSci-Bench features a dual-format design comprising both multiple-choice questions (MCQs) and open-ended questions (OEQs), enabling scalable automated evaluation alongside deeper analysis of conceptual understanding. We employ a template-based MCQ generation framework to create diverse, graduate-level problems with symbolic perturbation, while OEQs are used to probe open-ended reasoning.We conduct a comprehensive evaluation of representative LLMs, categorized into four groups: instruction-tuned models, advanced reasoning models, math-augmented models, and domain-specific climate models. Our analysis provides some interesting insights into the reasoning and problem-solving capabilities of LLMs in atmospheric science. We believe AtmosSci-Bench can serve as a critical step toward advancing LLM applications in climate service by offering a standard and rigorous evaluation framework.



Paperid:4760
Authors:Jinzhe Liu, Junshu Sun, Shufan Shen, Chenxue Yang, Shuhui Wang
Abstract:
Lifelong knowledge editing enables continuous, precise updates to outdated knowledge in large language models (LLMs) without computationally expensive full retraining. However, existing methods often accumulate errors throughout the editing process, causing a gradual decline in both editing accuracy and generalization. To tackle this problem, we propose Neuron-Specific Masked Knowledge Editing (NMKE), a novel fine-grained editing framework that combines neuron-level attribution with dynamic sparse masking. Leveraging neuron functional attribution, we identify two key types of knowledge neurons, with knowledge-general neurons activating consistently across prompts and knowledge-specific neurons activating to specific prompts.NMKE further introduces an entropy-guided dynamic sparse mask, locating relevant neurons to the target knowledge. This strategy enables precise neuron-level knowledge editing with fewer parameter modifications.Experimental results from thousands of sequential edits demonstrate that NKME outperforms existing methods in maintaining high editing success rates and preserving model general capabilities in lifelong editing.



Paperid:4761
Authors:Xiaochuan Gou, Ziyue Li, Tian Lan, Junpeng Lin, zhishuai Li, Bingyu Zhao, Chen Zhang, Di Wang, Xiangliang Zhang
Abstract:
Long-separated research has been conducted on two highly correlated tracks: traffic and incidents. Traffic track witnesses complicating deep learning models, e.g., to push the prediction a few percent more accurate, and the incident track only studies the incidents alone, e.g., to infer the incident risk. We, for the first time, spatiotemporally aligned the two tracks in a large-scale region (16,972 traffic nodes) from year 2022 to 2024: our TraffiDent dataset includes traffic, i.e., time-series indexes on traffic flow, lane occupancy, and average vehicle speed, and incident, whose records are spatiotemporally aligned with traffic data, with seven different incident classes. Additionally, each node includes detailed physical and policy-level meta-attributes of lanes. Previous datasets typically contain only traffic or incident data in isolation, limiting research to general forecasting tasks. TraffiDent integrates both, enabling detailed analysis of traffic-incident interactions and causal relationships. To demonstrate its broad applicability, we design: (1) post-incident traffic forecasting to quantify the impact of different incidents on traffic indexes; (2) incident classification using traffic indexes to determine the incidents types for precautions measures; (3) global causal analysis among the traffic indexes, meta-attributes, and incidents to give high-level guidance of the interrelations of various factors; (4) local causal analysis within road nodes to examine how different incidents affect the road segments' relations. The dataset is available at https://xaitraffic.github.io.



Paperid:4762
Authors:Thibaut Loiseau, Guillaume Bourmaud, Vincent Lepetit
Abstract:
Pre-training techniques have greatly advanced computer vision, with CroCo’s cross-view completion approach yielding impressive results in tasks like 3D reconstruction and pose regression. However, cross-view completion is ill-posed in non-covisible regions, limiting its effectiveness. We introduce Alligat0R, a novel pre-training approach that replaces cross-view learning with a covisibility segmentation task. Our method predicts whether each pixel in one image is covisible in the second image, occluded, or outside the field of view, making the pre-training effective in both covisible and non-covisible regions, and provides interpretable predictions. To support this, we present Cub3, a large-scale dataset with 5M image pairs and dense covisibility annotations derived from the nuScenes and ScanNet datasets. Cub3 includes diverse scenarios with varying degrees of overlap. The experiments show that our novel pre-training method Alligat0R significantly outperforms CroCo in relative pose regression. Alligat0R and Cub3 will be made publicly available.



Paperid:4763
Authors:Kevin Hayes, Micah Goldblum, Vikash Sehwag, Gowthami Somepalli, Ashwinee Panda, Tom Goldstein
Abstract:
Text-to-image (T2I) models are capable of generating visually impressive images, yet they often fail to accurately capture specific attributes in user prompts, such as the correct number of objects with the specified colors. The diversity of such errors underscores the need for a hierarchical evaluation framework that can compare prompt adherence abilities of different image generation models. Simultaneously, benchmarks of vision language models (VLMs) have not kept pace with the complexity of scenes that VLMs are used to annotate. In this work, we propose a structured methodology for jointly evaluating T2I models and VLMs by testing whether VLMs can identify 27 specific failure modes in the images generated by T2I models conditioned on challenging prompts. Our second contribution is a dataset of prompts and images generated by 5 T2I models (Flux, SD3-Medium, SD3-Large, SD3.5- Medium, SD3.5-Large) and the corresponding annotations from a VLM (Molmo) annotated by an LLM (Llama3) to test whether the VLM can correctly identify the failure mode in a generated image. By analyzing failure modes on a curated set of prompts, we reveal systematic errors in attribute fidelity and object representation. Our findings suggest that current metrics are insufficient to capture these nuanced errors, highlighting the importance of targeted benchmarks for advancing generative model reliability and interpretability.



Paperid:4764
Authors:Saurabh Kumar, Abhayjeet Singh, DEEKSHITHA G, Amartya veer, Jesuraj Bandekar, Savitha Murthy, Sumit Sharma, Sandhya Badiger, sathvik udupa, Amala Nagireddi, Srinivasa Raghavan K M, Rohan Saxena, Jai Nanavati, Raoul Nanavati, Janani Sridharan, Arjun Mehta, Ashish S, Sai Mora, Prashanthi Venkataramakrishnan, Gauri Date, Karthika P, Prasanta Ghosh
Abstract:
We introduce RESPIN-S1.0, the largest publicly available dialect-rich read speech corpus for Indian languages, comprising over 10,000 hours of validated audio spanning nine major languages -- Bengali, Bhojpuri, Chhattisgarhi, Hindi, Kannada, Magahi, Maithili, Marathi, and Telugu. Indian languages are characterized by high dialectal variation and are spoken by populations that are often digitally underserved. Existing speech corpora typically represent only standard dialects and lack domain relevance. RESPIN-S1.0 fills this critical gap by collecting speech across 38+ dialects and two high-impact domains: agriculture and finance. Text data was carefully composed by native dialect speakers and validated via a robust pipeline involving both automatic and manual checks. Over 200,000 utterances were recorded through a crowdsourced mobile application by native speakers and subsequently categorized into clean, semi-noisy, and noisy slabs based on transcription quality. The clean slab alone exceeds 10,000 hours. RESPIN also provides speaker metadata, phonetic lexicons, and dialect-aware train-dev-test splits to ensure reproducibility. To benchmark performance, we evaluate a range of ASR models -- TDNN-HMM, E-Branchformer, Whisper, IndicWav2Vec2, and SPRING SSL models -- and find that fine-tuning on RESPIN significantly improves recognition accuracy over existing pretrained models. A subset of RESPIN-S1.0 has already supported community efforts through challenges such as the SLT Code Hackathon 2022 and MADASR@ASRU 2023/2025, with over 1200 hours of data released publicly. This resource supports research in dialectal ASR, LID, DID, and speech-related areas, and sets a new standard for inclusive, dialect-rich corpora in multilingual, low-resource settings.



Paperid:4765
Authors:Yi Dong, Yuxi Wang, Xianhui Lin, Wenqi Ouyang, Zhiqi Shen, Peiran Ren, Ruoxi Fan, Rynson Lau
Abstract:
Color enhancement is a crucial yet challenging task in digital photography. It demands methods that are (i) expressive enough for fine-grained adjustments, (ii) adaptable to diverse inputs, and (iii) able to preserve texture. Existing approaches typically fall short in at least one of these aspects, yielding unsatisfactory results. We propose GenColor, a novel diffusion-based framework for sophisticated, texture-preserving color enhancement. GenColor reframes the task as conditional image generation. Leveraging ControlNet and a tailored training scheme, it learns advanced color transformations that adapt to diverse lighting and content. We train GenColor on ARTISAN, our newly collected large-scale dataset of 1.2M high-quality photographs specifically curated for enhancement tasks. To overcome texture preservation limitations inherent in diffusion models, we introduce a color-transfer network with a novel degradation scheme that simulates texture–color relationships. This network achieves pixel-perfect texture preservation while enabling fine-grained color matching with the diffusion-generated reference images. Extensive experiments show that GenColor produces visually compelling results comparable to those of expert colorists and surpasses state-of-the-art methods in both subjective and objective evaluations. We will release the code and dataset and kindly invite reviewers to view our supplementary video and website for a more intuitive demonstration.



Paperid:4766
Authors:Shlomi Hod, Lucas Rosenblatt, Julia Stoyanovich
Abstract:
Differentially private (DP) machine learning often relies on the availability of public data for tasks like privacy-utility trade-off estimation, hyperparameter tuning, and pretraining. While public data assumptions may be reasonable in text and image data, they are less likely to hold for tabular data due to tabular data heterogeneity across domains. We propose leveraging powerful priors to address this limitation; specifically, we synthesize realistic tabular data directly from schema-level specifications -- such as variable names, types, and permissible ranges -- without ever accessing sensitive records. To that end, this work introduces the notion of ``surrogate'' public data -- datasets generated independently of sensitive data, which consume no privacy loss budget and are constructed solely from publicly available schema or metadata. Surrogate public data are intended to encode plausible statistical assumptions (informed by publicly available information) into a dataset with many downstream uses in private mechanisms. We automate the process of generating surrogate public data with large language models (LLMs); in particular, we propose two methods: direct record generation as CSV files, and automated structural causal model (SCM) construction for sampling records. Through extensive experiments, we demonstrate that surrogate public tabular data can effectively replace traditional public data when pretraining differentially private tabular classifiers. To a lesser extent, surrogate public data are also useful for hyperparameter tuning of DP synthetic data generators, and for estimating the privacy-utility tradeoff.



Paperid:4767
Authors:Ehsan Latif, Zirak Khan, Xiaoming Zhai
Abstract:
Scientific sketches (e.g., models) offer a powerful lens into students' conceptual understanding, yet AI-powered automated assessment of such free-form, visually diverse artifacts remains a critical challenge. Existing solutions often treat sketch evaluation as either an image classification task or monolithic vision-language models, which lack interpretability, pedagogical alignment, and adaptability across cognitive levels. To address these limitations, we present SketchMind, a cognitively grounded, multi-agent framework for evaluating and improving student-drawn scientific sketches. SketchMind introduces Sketch Reasoning Graphs (SRGs), semantic graph representations that embed domain concepts and Bloom's taxonomy-based cognitive labels. The system comprises modular agents responsible for rubric parsing, sketch perception, cognitive alignment, and iterative feedback with sketch modification, enabling personalized and transparent evaluation. We evaluate SketchMind on a curated dataset of 3,575 student-generated sketches across six science assessment items with different highest order of Bloom's level that require students to draw models to explain phenomena. Compared to baseline GPT-4o performance without SRG (average accuracy: 55.6%), the model with SRG integration achieves 77.1% average accuracy (+21.4% average absolute gain). We also demonstrate that multi-agent orchestration with SRG enhances SketchMind performance, for example, GPT-4.1 gains an average 8.9% increase in sketch prediction accuracy, outperforming single-agent pipelines across all items. Human evaluators rated the feedback and co-created sketches generated by SketchMind with GPT-4.1, which achieved an average of 4.1 out of 5, significantly higher than those of baseline models (e.g., 2.3 for GPT-4o). Experts noted the system’s potential to meaningfully support conceptual growth through guided revision. Our code and (pending approval) dataset will be released to support reproducibility and future research in AI-driven education.



Paperid:4768
Authors:Huitong Yang, Zhuoxiao Chen, Fengyi Zhang, Zi Huang, Yadan Luo
Abstract:
Maintaining robust 3D perception under dynamic and unpredictable test-time conditions remains a critical challenge for autonomous driving systems. Existing test-time adaptation (TTA) methods often fail in high-variance tasks like 3D object detection due to unstable optimization and sharp minima. While recent model merging strategies based on linear mode connectivity (LMC) offer improved stability by interpolating between fine-tuned checkpoints, they are computationally expensive, requiring repeated checkpoint access and multiple forward passes. In this paper, we introduce CodeMerge, a lightweight and scalable model merging framework that bypasses these limitations by operating in a compact latent space. Instead of loading full models, CodeMerge represents each checkpoint with a low-dimensional fingerprint derived from the source model’s penultimate features and constructs a key-value codebook. We compute merging coefficients using regularized leverage scores on these fingerprints, enabling efficient model composition without compromising adaptation quality. Our method achieves strong performance across challenging benchmarks, improving end-to-end 3D detection 14.9\% NDS on nuScenes-C and LiDAR-based detection by over 7.6\% mAP on nuScenes-to-KITTI, while benefiting downstream tasks such as online mapping, motion prediction and planning even without training. Code and pretrained models are released in the supplementary material.



Paperid:4769
Authors:Xingjian Wu, Xiangfei Qiu, Hanyin Cheng, Zhengyu Li, Jilin Hu, Chenjuan Guo, Bin Yang
Abstract:
Time Series Forecasting has made significant progress with the help of Patching technique, which partitions time series into multiple patches to effectively retain contextual semantic information into a representation space beneficial for modeling long-term dependencies. However, conventional patching partitions a time series into adjacent patches, which causes a fixed representation space, thus resulting in insufficiently expressful representations. In this paper, we pioneer the exploration of constructing a selective representation space to flexibly include the most informative patches for forecasting. Specifically, we propose the Selective Representation Space (SRS) module, which utilizes the learnable Selective Patching and Dynamic Reassembly techniques to adaptively select and shuffle the patches from the contextual time series, aiming at fully exploiting the information of contextual time series to enhance the forecasting performance of patch-based models. To demonstrate the effectiveness of SRS module, we propose a simple yet effective SRSNet consisting of SRS and an MLP head, which achieves state-of-the-art performance on real-world datasets from multiple domains. Furthermore, as a novel plugin-and-play module, SRS can also enhance the performance of existing patch-based models.



Paperid:4770
Authors:MingYu Lu, Ethan Weinberger, Chanwoo Kim, Su-In Lee
Abstract:
High-content screening (HCS) assays based on high-throughput microscopy techniques such as Cell Painting have enabled the interrogation of cells' morphological responses to perturbations at an unprecedented scale. The collection of such data promises to facilitate a better understanding of the relationships between different perturbations and their effects on cellular state. Towards achieving this goal, recent advances in cross-modal contrastive learning could, in theory, be leveraged to learn a unified latent space that aligns perturbations with their corresponding morphological effects. However, the application of such methods to HCS data is not straightforward due to substantial differences in the semantics of Cell Painting images compared to natural images, and the difficulty of representing different classes of perturbations (e.g. small molecule vs CRISPR gene knockout) in a single latent space. In response to these challenges, here we introduce CellCLIP, a cross-modal contrastive learning framework for HCS data. CellCLIP leverages pre-trained image encoders coupled with a novel channel encoding scheme to better capture relationships between different microscopy channels in image embeddings, along with natural language encoders for representing perturbations. Our framework outperforms current open-source models, demonstrating the best performance in both cross-modal retrieval and biologically meaningful downstream tasks while also achieving significant reductions in computation time. Code for our reproducing our experiments is available at https://anonymous.4open.science/r/CellCLIP-4D1C.



Paperid:4771
Authors:Yue Huang, Zhengzhe Jiang, Xiaonan Luo, Kehan Guo, Haomin Zhuang, Yujun Zhou, Zhengqing Yuan, Xiaoqi Sun, Jules Schleinitz, Yanbo Wang, Shuhao Zhang, Mihir Surve, Nitesh Chawla, Olaf Wiest, Xiangliang Zhang
Abstract:
Empowering large language models (LLMs) with chemical intelligence remains a challenge due to the scarcity of high-quality, domain-specific instruction-response datasets and the misalignment of existing synthetic data generation pipelines with the inherently hierarchical and rule-governed structure of chemical information. To address this, we propose ChemOrch, a framework that synthesizes chemically grounded instruction–response pairs through a two-stage process: task-controlled instruction generation and tool-aware response construction. ChemOrch enables controllable diversity and levels of difficulty for the generated tasks and ensures response precision through tool planning \& distillation, and tool-based self-repair mechanisms. The effectiveness of ChemOrch is evaluated based on: 1) the \textbf{high quality} of generated instruction data, demonstrating superior diversity and strong alignment with chemical constraints; 2) the \textbf{dynamic generation of evaluation tasks} that more effectively reveal LLM weaknesses in chemistry; and 3) the significant \textbf{improvement of LLM chemistry capabilities} when the generated instruction data are used for fine-tuning. Our work thus represents a critical step toward scalable and verifiable chemical intelligence in LLMs. The code is available at \url{https://anonymous.4open.science/r/ChemOrch-854A}.



Paperid:4772
Authors:Ningyuan Huang, Richard Stiskalek, Jun-Young Lee, Adrian Bayer, Charles Margossian, Christian Kragh Jespersen, Lucia Perez, Lawrence Saul, Francisco Villaescusa
Abstract:
Cosmological simulations provide a wealth of data in the form of point clouds and directed trees. A crucial goal is to extract insights from this data that shed light on the nature and composition of the Universe. In this paper we introduce CosmoBench, a benchmark dataset curated from state-of-the-art cosmological simulations whose runs required more than 41 million core-hours and generated over two petabytes of data. CosmoBench is the largest dataset of its kind: it contains 34 thousands of point clouds from simulations of dark matter halos and galaxies at three different length scales, as well as 25 thousand directed trees that record the formation history of halos on two different time scales. The data in CosmoBench can be used for multiple tasks---to predict cosmological parameters from point clouds and merger trees, to predict the velocities of individual halos and galaxies from their collective positions, and to reconstruct merger trees on finer time scales from those on coarser time scales. We provide multiple baselines on these tasks, some based on established approaches from cosmological modeling and others rooted in machine learning. For the latter, we study different approaches---from simple linear models that are minimally constrained by symmetries to much larger and more computationally-demanding models in deep learning, such as graph neural networks. We find that least-squares fits with a handful of invariant features sometimes outperform deep architectures with many more parameters and far longer training times. Still there remains tremendous potential to improve these baselines by combining machine learning and cosmological modeling in a more principled way, one that fully exploits the structure in the data. CosmoBench sets the stage for bridging cosmology and geometric deep learning at scale. We invite the community to push the frontier of scientific discovery by engaging with this challenging, high-impact dataset. The data and code are available atthis URL.



Paperid:4773
Authors:Yang Xu, Swetha Ganesh, Washim Mondal, Qinbo Bai, Vaneet Aggarwal
Abstract:
Abstract:This paper investigates infinite-horizon average reward Constrained Markov Decision Processes (CMDPs) with general parametrization. We propose a Primal-Dual Natural Actor-Critic algorithm that adeptly manages constraints while ensuring a high convergence rate. In particular, our algorithm achieves global convergence and constraint violation rates of $\tilde{\mathcal{O}}(1/\sqrt{T})$ over a horizon of length $T$ when the mixing time, $\tau_{\mathrm{mix}}$, is known to the learner. In absence of knowledge of $\tau_{\mathrm{mix}}$, the achievable rates change to $\tilde{\mathcal{O}}(1/T^{0.5-\epsilon})$ provided that $T \geq \tilde{\mathcal{O}}\left(\tau_{\mathrm{mix}}^{2/\epsilon}\right)$. Our results match the theoretical lower bound for Markov Decision Processes and establish a new benchmark in the theoretical exploration of average reward CMDPs.



Paperid:4774
Authors:Longfei Li, Zhiwen Fan, Wenyan Cong, Xinhang Liu, Yuyang Yin, Matt Foutter, Panwang Pan, Chenyu You, Yue Wang, Zhangyang "Atlas" Wang, Yao Zhao, Marco Pavone, Yunchao Wei
Abstract:
The synthesis of realistic Martian landscape videos, essential for mission rehearsal and robotic simulation, presents unique challenges. These primarily stem from the scarcity of high-quality Martian data and the significant domain gap relative to terrestrial imagery.To address these challenges, we introduce a holistic solution comprising two main components: 1) a data curation framework, Multimodal Mars Synthesis (M3arsSynth), which processes stereo navigation images to render high-fidelity 3D video sequences. 2) a video-based Martian terrain generator (MarsGen), that utilizes multimodal conditioning data to accurately synthesize novel, 3D-consistent frames. Our data are sourced from NASA’s Planetary Data System (PDS), covering diverse Martian terrains and dates, enabling the production of physics-accurate 3D surface models at metric-scale resolution. During inference, MarsGen is conditioned on an initial image frame and can be guided by specified camera trajectories or textual prompts to generate new environments.Experimental results demonstrate that our solution surpasses video synthesis approaches trained on terrestrial data, achieving superior visual quality and 3D structural consistency.



Paperid:4775
Authors:shengchuan gao, Shuo Wang, Yabiao Wang, Ran Yi
Abstract:
This work tackles a challenging problem: stochastic human motion prediction (SHMP), which aims to forecast diverse and physically plausible future pose sequences based on a short history of observed motion. While autoregressive sequence models have excelled in related generation tasks, their reliance on vector‐quantized tokenization limits motion fidelity and training stability. To overcome these drawbacks, we introduce \textbf{WaveAR}, a novel AR based framework which is the first successful application of a continuous autoregressive generation paradigm to HMP to our best knowledge. WaveAR consists of two stages. In the first stage, a lightweight Spatio‐Temporal VAE (ST-VAE) compresses the raw 3D-joint sequence into a downsampled latent token stream, providing a compact yet expressive foundation. In the second stage, we apply masked autoregressive prediction directly in this continuous latent space, conditioning on both unmasked latents and multi‐scale spectral cues extracted via a 2D discrete wavelet transform. A fusion module consisting of alternating cross-attention and self-attention layers adaptively fuses temporal context with low- and high-frequency wavelet subbands, and a small MLP‐based diffusion head predicts per-token noise residuals under a denoising loss. By avoiding vector quantization and integrating localized frequency information, WaveAR preserves fine‐grained motion details while maintaining fast inference speed. Extensive experiments on standard benchmarks demonstrate that our approach delivers more accurate and computationally efficient predictions than prior state‐of-the-art methods.



Paperid:4776
Authors:Yunjia Qi, Hao Peng, Xiaozhi Wang, Amy Xin, Youfeng Liu, Bin Xu, Lei Hou, Juanzi Li
Abstract:
Abstract:Large Language Models (LLMs) have demonstrated advanced capabilities in real-world agentic applications. Growing research efforts aim to develop LLM-based agents to address practical demands, introducing a new challenge: agentic scenarios often involve lengthy instructions with complex constraints, such as extended system prompts and detailed tool specifications. While adherence to such instructions is crucial for agentic applications, whether LLMs can reliably follow them remains underexplored. In this paper, we introduce AgentIF, the first benchmark for systematically evaluating LLM instruction following ability in agentic scenarios. AgentIF features three key characteristics: (1) Realistic, constructed from $50$ real-world agentic applications. (2) Long, averaging $1,723$ words with a maximum of $15,630$ words. (3) Complex, averaging $11.9$ constraints per instruction, covering diverse constraint types, such as tool specifications and condition constraints.To construct AgentIF, we collect $707$ human-annotated instructions across $50$ agentic tasks from industrial application agents and open-source agentic systems. For each instruction, we annotate the associated constraints and corresponding evaluation metrics, including code-based evaluation, LLM-based evaluation, and hybrid code-LLM evaluation.% resulting in a final set of $707$ instructions.We use AgentIF to systematically evaluate existing advanced LLMs. We observe that current models generally perform poorly, especially in handling complex constraint structures and tool specifications. We further conduct error analysis and analytical experiments on instruction length and meta constraints, providing some findings about the failure modes of existing LLMs. We have released the code and data to facilitate future research.



Paperid:4777
Authors:Shuo Yang, Haocheng Xi, Yilong Zhao, Muyang Li, Jintao Zhang, Han Cai, Yujun Lin, Xiuyu Li, Chenfeng Xu, Kelly Peng, Jianfei Chen, Song Han, Kurt Keutzer, Ion Stoica
Abstract:
Abstract:Diffusion Transformers (DiTs) are essential for video generation but suffer from significant latency due to the quadratic complexity of attention. By computing only critical tokens, sparse attention reduces computational costs and offers a promising acceleration approach. However, we identify that existing methods fail to approach optimal generation quality under the same computation budget for two reasons: (1) Inaccurate critical token identification: current methods cluster tokens based on position rather than semantics, leading to imprecise aggregated representations. (2) Excessive computation waste: critical tokens are scattered among non-critical ones, leading to wasted computation on GPUs, which are optimized for processing contiguous tokens.In this paper, we propose SAPAttn, a training-free framework that maximizes identification accuracy and minimizes computation waste, achieving a Pareto frontier trade-off between generation quality and efficiency. The core of SAPAttn is semantic-aware permutation, which clusters and reorders tokens based on semantic similarity using k-means. This approach ensures both a precise cluster representation, improving identification accuracy, and a densified layout of critical tokens, enabling efficient computation without padding. Additionally, SAPAttn integrates Top-p dynamic budget control and customized kernel implementations, achieving up to $2.30\times$ and $1.89\times$ speedup while maintaining a PSNR of up to $30$ and $26$ on HunyuanVideo and Wan 2.1, respectively.



Paperid:4778
Authors:Yeongbins Seo, Dongha Lee, Jaehyung Kim, Jinyoung Yeo
Abstract:
Autoregressive (AR) language models generate text one token at a time, which limits their inference speed. Diffusion-based language models offer a promising alternative, as they can decode multiple tokens in parallel. However, we identify a key bottleneck in current diffusion LMs: the \textbf{long decoding-window problem}, where tokens generated far from the input context often become irrelevant or repetitive. Previous solutions like semi-autoregressive address this issue by splitting windows into blocks, but this sacrifices speed and bidirectionality, eliminating the main advantage of diffusion models. To overcome this, we propose Convolutional decoding (\textit{Conv}), a normalization-based method that narrows the decoding window without hard segmentation, preserving fluency and bidirectionality. Additionally, we introduce Rejecting Rule-based Fine-Tuning (R2FT), a post-hoc training scheme that better aligns tokens at positions far from context. Our methods achieve state-of-the-art results on open-ended generation benchmarks (e.g., AlpacaEval) among diffusion LM baselines, with significantly lower step size than previous works, demonstrating both speed and quality improvements. The code is available online (\url{https://anonymous.4open.science/r/conv-89D7}).



Paperid:4779
Authors:Yufeng Zou, Zijian Wang, Diego Klabjan, Han Liu
Abstract:
Times series foundation models (TSFMs) have emerged as a promising paradigm for time series analysis, showing remarkable generalization performance across different domains. While there has been research on hallucinations in foundation models, hallucinations in TSFMs are underexplored. In this paper, we formally define TSFM hallucinations in the zero-shot forecasting setting by examining whether a generated forecast exhibits different dynamics from those of the context. Our study reveals that TSFM hallucinations are mainly caused by the loss of context information in hidden states during forward propagation. As such, we propose methodologies to identify signal subspaces in TSFMs and magnify the signal subspace information through intervention. Extensive experiments demonstrate that our proposed intervention approach effectively mitigates hallucinations and improves forecast performance. Furthermore, the signal strength measure we compute from signal subspaces has strong predictive power of hallucinations and forecast performance of the model. Our work contributes to deeper understanding of TSFM trustworthiness that could foster future research in this direction.



Paperid:4780
Authors:Yue Jiang, Jichu Li, Yang Liu, Dingkang Yang, Feng Zhou, Quyu Kong
Abstract:
We introduce DanmakuTPPBench, a comprehensive benchmark designed to advance multi-modal Temporal Point Process (TPP) modeling in the era of Large Language Models (LLMs). While TPPs have been widely studied for modeling temporal event sequences, existing datasets are predominantly unimodal, hindering progress in models that require joint reasoning over temporal, textual, and visual information. To address this gap, DanmakuTPPBench comprises two complementary components:(1) DanmakuTPP-Events, a novel dataset derived from the Bilibili video platform, where user-generated bullet comments (Danmaku) naturally form multi-modal events annotated with precise timestamps, rich textual content, and corresponding video frames;(2) DanmakuTPP-QA, a challenging question-answering dataset constructed via a novel multi-agent pipeline powered by state-of-the-art LLMs and multi-modal LLMs (MLLMs), targeting complex temporal-textual-visual reasoning. We conduct extensive evaluations using both classical TPP models and recent MLLMs, revealing significant performance gaps and limitations in current methods’ ability to model multi-modal event dynamics. Our benchmark establishes strong baselines and calls for further integration of TPP modeling into the multi-modal language modeling landscape.



Paperid:4781
Authors:Yuxin Ma, Dmitriy Kunisky
Abstract:
Abstract:We introduce a new family of algorithms for detecting and estimating a rank-one signal from a noisy observation under prior information about that signal's direction, focusing on examples where the signal is known to have entries biased to be positive. Given a matrix observation $\mathbf{Y}$, our algorithms construct a *nonlinear Laplacian*, another matrix of the form $\mathbf{Y} + \mathrm{diag}(\sigma(\mathbf{Y 1})$ for a nonlinear $\sigma: \mathbb{R} \to \mathbb{R}$, and examine the top eigenvalue and eigenvector of this matrix. When $\mathbf{Y}$ is the (suitably normalized) adjacency matrix of a graph, our approach gives a class of algorithms that search for unusually dense subgraphs by computing a spectrum of the graph "deformed" by the degree profile $\mathbf{Y1}$. We study the performance of such algorithms compared to direct spectral algorithms (the case $\sigma = 0$) on models of sparse principal component analysis with biased signals, including the Gaussian planted submatrix problem. For such models, we rigorously characterize the critical threshold strength of rank-one signal, as a function of the nonlinearity $\sigma$, at which an outlier eigenvalue appears in the spectrum of a nonlinear Laplacian matrix. While identifying the $\sigma$ that minimizes this critical signal strength in closed form seems intractable, we explore three approaches to design $\sigma$ numerically: exhaustively searching over simple classes of $\sigma$, learning $\sigma$ from datasets of problem instances, and tuning $\sigma$ using black-box optimization of the critical signal strength. We find both theoretically and empirically that, if $\sigma$ is chosen appropriately, then nonlinear Laplacian spectral algorithms substantially outperform direct spectral algorithms, while avoiding the complexity of broader classes of algorithms like approximate message passing or general first order methods.



Paperid:4782
Authors:Ronghao Lin, Qiaolin He, Sijie Mai, Ying Zeng, Aolin Xiong, Li Huang, Yap-peng Tan, Haifeng Hu
Abstract:
Multimodal machine learning, mimicking the human brain’s ability to integrate various modalities has seen rapid growth. Most previous multimodal models are trained on perfectly paired multimodal input to reach optimal performance. In real‑world deployments, however, the presence of modality is highly variable and unpredictable, causing the pre-trained models in suffering significant performance drops and fail to remain robust with dynamic missing modalities circumstances. In this paper, we present a novel Cyclic INformative Learning framework (CyIN) to bridge the gap between complete and incomplete multimodal learning. Specifically, we firstly builds an informative latent space by adopting token- and label-level Information Bottleneck (IB) cyclically among various modalities. Capturing task-related features with variational approximation, the informative bottleneck latents are purified for more efficient cross-modal interaction and multimodal fusion. Moreover, to supplement the missing information caused by incomplete multimodal input, we propose cross-modal cyclic translation by reconstruct the missing modalities with the remained ones through forward and reverse propagation process. With the help of the extracted and reconstructed informative latents, CyIN succeeds in jointly optimizing complete and incomplete multimodal learning in one unified model. Extensive experiments on 4 datasets demonstrate the superior performance of our method in both complete and diverse incomplete scenarios.



Paperid:4783
Authors:Yibo Wang, Guangda Huzhang, Qingguo Chen, Zhao Xu, Weihua Luo, Kaifu Zhang, Lijun Zhang
Abstract:
Self-play fine-tuning has demonstrated promising abilities in adapting large language models (LLMs) to downstream tasks with limited real-world data. The basic principle is to iteratively refine the model with real samples and synthetic ones generated from itself. However, the existing methods primarily focus on the relative gaps between the rewards for two types of data, neglecting their absolute values. Through theoretical analysis, we identify that the gap-based methods suffer from unstable evolution, due to the potentially degenerated objectives. To address this limitation, we introduce a novel self-play fine-tuning method, namely \underline{S}elf-\underline{P}l\underline{A}y via Noise \underline{C}ontrastive \underline{E}stimation (SPACE), which leverages noise contrastive estimation to capture the real-world data distribution. Specifically, SPACE treats synthetic samples as auxiliary components, and discriminates them from the real ones in a binary classification manner. As a result, SPACE independently optimizes the absolute reward values for each type of data, ensuring a consistently meaningful objective and thereby avoiding the instability issue. Theoretically, we show that the optimal solution of the objective in SPACE aligns with the underlying distribution of real-world data, and SPACE guarantees a provably stable convergence to the optimal distribution. Empirically, we show that SPACE significantly improves the performance of LLMs over various tasks, and outperforms supervised fine-tuning that employs much more real-world samples. Compared to gap-based self-play fine-tuning methods, SPACE exhibits remarkable superiority and stable evolution.



Paperid:4784
Authors:Rafael Bischof, Michal Piovarci, Michael Kraus, Siddhartha Mishra, Bernd Bickel
Abstract:
Abstract:We present HyPINO, a multi-physics neural operator designed for zero-shot generalization across a broad class of parametric PDEs without requiring task-specific fine-tuning. Our approach combines a Swin Transformer-based hypernetwork with mixed supervision: (i) labeled data from analytical solutions generated via the Method of Manufactured Solutions (MMS), and (ii) unlabeled samples optimized using physics-informed objectives. The model maps PDE parametrizations to target Physics-Informed Neural Networks (PINNs) and can handle linear elliptic, hyperbolic, and parabolic equations in two dimensions with varying source terms, geometries, and mixed Dirichlet/Neumann boundary conditions, including interior boundaries. HyPINO achieves strong zero-shot accuracy on seven benchmark problems from PINN literature, outperforming U-Nets, Poseidon, and Physics-Informed Neural Operators (PINO). Further, we introduce an iterative refinement procedure that compares the physics of the generated PINN to the requested PDE and uses the discrepancy to generate a "delta" PINN. Summing their contributions and repeating this process forms an ensemble whose combined solution progressively reduces the error on six benchmarks and achieves over 100x gain in average $L_2$ loss in the best case, while retaining forward-only inference. Additionally, we evaluate the fine-tuning behavior of PINNs initialized by HyPINO and show that they converge faster and to lower final error than both randomly initialized and Reptile-meta-learned PINNs on five benchmarks, performing on par on the remaining two. Our results highlight the potential of this scalable approach as a foundation for extending neural operators toward solving increasingly complex, nonlinear, and high-dimensional PDE problems with significantly improved accuracy and reduced computational cost.



Paperid:4785
Authors:Tianxiong Zhong, Xingye Tian, Boyuan Jiang, Xuebo Wang, Xin Tao, Pengfei Wan, Zhiwei Zhang
Abstract:
Abstract:Modern video generation frameworks based on Latent Diffusion Models suffer from inefficiencies in tokenization due to the Frame-Proportional Information Assumption.Existing tokenizers provide fixed temporal compression rates, causing the computational cost of the diffusion model to scale linearly with the frame rate.The paper proposes the Duration-Proportional Information Assumption: the upper bound on the information capacity of a video is proportional to the duration rather than the number of frames.Based on this insight, the paper introduces VFRTok, a Transformer-based video tokenizer, that enables variable frame rate encoding and decoding through asymmetric frame rate training between the encoder and decoder.Furthermore, the paper proposes Partial Rotary Position Embeddings (RoPE) to decouple position and content modeling, which groups correlated patches into unified tokens.The Partial RoPE effectively improves content-awareness, enhancing the video generation capability.Benefiting from the compact and continuous spatio-temporal representation, VFRTok achieves competitive reconstruction quality and state-of-the-art generation fidelity while using only $1/8$ tokens compared to existing tokenizers.



Paperid:4786
Authors:HongXin Xu, Tianyu Guo, Xianwei Zhang
Abstract:
To accelerate large language model (LLM) inference, pipeline parallelism partitions model layers into sequential stages, each assigned to a different device for concurrent execution. However, this method often suffers from pipeline bubbles caused by imbalanced computation in the tail stage. While upstream stages focus solely on layer-forward operations, the final stage must also handle post-processing tasks like sampling, introducing significant latency. This uneven workload leads to pipeline misalignment, forcing upstream stages to idle and degrading overall performance. Existing frameworks typically distribute layers evenly across stages without accounting for computational load differences. To address this, we propose DynaPipe, a dynamic layer redistribution scheme that adaptively balances computation by predicting execution latency in real time. Moreover, we introduce an asynchronous key-value (KV) cache migration coordinator to enablenon-blocking layer redistribution during inference. Experiments on representative LLMs demonstrate that DynaPipe reduces average end-to-end request latency by 8% to 49% across diverse workloads, outperforming state-of-the-art pipeline parallelism systems.



Paperid:4787
Authors:Ziyang Wei, Jiaqi Li, Zhipeng Lou, Wei Biao Wu
Abstract:
We establish a comprehensive finite-sample and asymptotic theory for stochastic gradient descent (SGD) with constant learning rates. First, we propose a novel linear approximation technique to provide a quenched central limit theorem (CLT) for SGD iterates with refined tail properties, showing that regardless of the chosen initialization, the fluctuations of the algorithm around its target point converge to a multivariate normal distribution. Our conditions are substantially milder than those required in the classical CLTs for SGD, yet offering a stronger convergence result. Furthermore, we derive the first Berry-Esseen bound -- the Gaussian approximation error -- for the constant learning-rate SGD, which is sharp compared to the decaying learning-rate schemes in the literature. Beyond the moment convergence, we also provide the Nagaev-type inequality for the SGD tail probabilities by adopting the autoregressive approximation techniques, which entails non-asymptotic large-deviation guarantees. These results are verified via numerical simulations, paving the way for theoretically grounded uncertainty quantification, especially with non-asymptotic validity.



Paperid:4788
Authors:Peng Lai, Jianjie Zheng, Sijie Cheng, Yun Chen, Peng Li, Yang Liu, Guanhua Chen
Abstract:
The growing scale of evaluation tasks has led to the widespread adoption of automated evaluation using large language models, a paradigm known as "LLM-as-a-judge." However, improving its alignment with human preferences without complex prompts or fine-tuning remains challenging. In this work, motivated by preliminary findings that middle-to-upper layers encode semantically and task-relevant representations that are often more aligned with human judgments than the final layer, we propose LAGER, a lightweight and efficient framework for enhancing LLM-as-a-Judge alignment with human scoring, via internal representations. LAGER produces fine-grained judgment scores by aggregating cross-layer score-token logits and computing the expected score from a softmax-based distribution, withthe LLM backbone kept frozen. LAGER fully leverages the complementary information across different layers, overcoming the limitations of relying solely on the final layer. We evaluate our method on the standard alignment benchmarks Flask, HelpSteer, and BIGGen using Spearman correlation, and find that LAGER achieves improvements of up to 7.5% over the best baseline across these benchmarks. Without reasoning steps, LAGER matches or outperforms reasoning-based methods. Experiments on downstream applications, such as data selection and emotional understanding, further show the effectiveness of our method.



Paperid:4789
Authors:Kaixun Jiang, Zhaoyu Chen, HaiJing Guo, Jinglun Li, Jiyuan Fu, Pinxue Guo, Hao Tang, Bo Li, Wenqiang Zhang
Abstract:
Preference alignment in diffusion models has primarily focused on benign human preferences (e.g., aesthetic). In this paper, we propose a novel perspective: framing unrestricted adversarial example generation as a problem of aligning with adversary preferences. Unlike benign alignment, adversarial alignment involves two inherently conflicting preferences: visual consistency and attack effectiveness, which often lead to unstable optimization and reward hacking (e.g., reducing visual quality to improve attack success). To address this, we propose APA (Adversary Preferences Alignment), a two-stage framework that decouples conflicting preferences and optimizes each with differentiable rewards. In the first stage, APA fine-tunes LoRA to improve visual consistency using rule-based similarity reward. In the second stage, APA updates either the image latent or prompt embedding based on feedback from a substitute classifier, guided by trajectory-level and step-wise rewards. To enhance black-box transferability, we further incorporate a diffusion augmentation strategy. Experiments demonstrate that APA achieves significantly better attack transferability while maintaining high visual consistency, inspiring further research to approach adversarial attacks from an alignment perspective.



Paperid:4790
Authors:Tara Akhound-Sadegh, Jungyoon Lee, Joey Bose, Valentin De Bortoli, Arnaud Doucet, Michael Bronstein, Dominique Beaini, Siamak Ravanbakhsh, Kirill Neklyudov, Alexander Tong
Abstract:
Abstract:Sampling efficiently from a target unnormalized probability density remains a core challenge, with relevance across countless high-impact scientific applications. A promising approach towards this challenge is the design of amortized samplers that borrow key ideas, such as probability path design, from state-of-the-art generative diffusion models. However, all existing diffusion-based samplers remain unable to draw samples from distributions at the scale of even simple molecular systems. In this paper, we propose Progressive Inference-Time Annealing (PITA) a novel framework to learn diffusion-based samplers that combines two complementary interpolation techniques: I.) Annealing of the Boltzmann distribution and II.) Diffusion smoothing. PITA trains a sequence of diffusion models from high to low temperatures by sequentially training each model at progressively higher temperatures, leveraging engineered easy access to samples of the temperature-annealed target density. In the subsequent step, PITA enables simulating the trained diffusion model to *procure training samples at a lower temperature* for the next diffusion model through inference-time annealing using a novel Feynman-Kac PDE combined with Sequential Monte Carlo. Empirically, PITA enables, for the first time, equilibrium sampling of $N$-body particle systems, Alanine Dipeptide, and tripeptides in Cartesian coordinates with dramatically lower energy function evaluations.



Paperid:4791
Authors:Harris Nisar, Ozgur Kara, James Rehg
Abstract:
To better understand how the complex human attention system operates on images, numerous models have been developed for scanpath and saliency prediction. These models are typically trained on compressed representations of raw eye-tracking data, referred to as scanpaths, while the rich information contained in the raw trajectories is often discarded. Moreover, most existing approaches fail to capture the variability observed among human subjects viewing the same image. They generally predict a single scanpath of fixed, pre-defined length, which conflicts with the inherent diversity and stochastic nature of real-world visual attention. To address these challenges, we propose DiffEye, a diffusion-based training framework designed to model continuous and diverse eye movement trajectories during free viewing of natural images. Our method builds on a diffusion model conditioned on visual stimuli and introduces a novel component, namely Corresponding Positional Embedding (CPE), which aligns spatial gaze information with the patch-based semantic features of the visual input. By leveraging raw eye-tracking trajectories rather than relying on scanpaths, DiffEye captures the inherent variability in human gaze behavior and generates high-quality, realistic eye movement patterns, despite being trained on a comparatively small dataset. The generated trajectories can also be converted into scanpaths and saliency maps, resulting in outputs that more accurately reflect the distribution of human visual attention. DiffEye is the first method to tackle this task on natural images using a diffusion model while fully leveraging the richness of raw eye-tracking data. Our extensive evaluation shows that DiffEye not only achieves state-of-the-art performance in scanpath generation but also enables, for the first time, the generation of continuous eye movement trajectories.



Paperid:4792
Authors:Yihua Zhang, Changsheng Wang, Yiwei Chen, Chongyu Fan, Jinghan Jia, Sijia Liu
Abstract:
Abstract:Input saliency aims to quantify the influence of input tokens on the output of large language models (LLMs), which has been widely used for prompt engineering, model interpretability, and behavior attribution. Despite the proliferation of saliency techniques, the field lacks a standardized and rigorous evaluation protocol. In this work, we introduce a stress-testing framework inspired by the needle-in-a-haystack (NIAH) setting to systematically assess the reliability of seven popular input saliency methods. Our evaluation reveals a surprising and critical flaw: existing methods consistently assign non-trivial importance to irrelevant context, and this attribution error worsens as input length increases. To address this issue, we propose a novel saliency method based on Attention Bias Optimization (ours), which explicitly optimizes the attention bias associated with each input token to quantify its causal impact on target token generation. ABO robustly outperforms existing methods by $20\sim40\%$ in saliency accuracy across diverse NIAH tasks, maintains effectiveness up to 10K-token prompts, and enables practical applications including zero-shot detoxification, sentiment steering, and reasoning-error correction. Our findings highlight the limitations of prevalent attribution methods and establish ABO as a principled alternative for accurate token attribution.



Paperid:4793
Authors:Beibu Li, Qichao Shentu, Yang Shu, Hui Zhang, Ming Li, Ning Jin, Bin Yang, Chenjuan Guo
Abstract:
Time series anomaly detection plays a crucial role in a wide range of real-world applications. Given that time series data can exhibit different patterns at different sampling granularities, multi-scale modeling has proven beneficial for uncovering latent anomaly patterns that may not be apparent at a single scale. However, existing methods often model multi-scale information independently or rely on simple feature fusion strategies, neglecting the dynamic changes in cross-scale associations that occur during anomalies. Moreover, most approaches perform multi-scale modeling based on fixed sliding windows, which limits their ability to capture comprehensive contextual information. In this work, we propose CrossAD, a novel framework for time series Anomaly Detection that takes Cross-scale associations and Cross-window modeling into account. We propose a cross-scale reconstruction that reconstructs fine-grained series from coarser series, explicitly capturing cross-scale associations. Furthermore, we design a query library and incorporate global multi-scale context to overcome the limitations imposed by fixed window sizes. Extensive experiments conducted on seven real-world datasets using nine evaluation metrics validate the effectiveness of CrossAD, demonstrating state-of-the-art performance in anomaly detection.



Paperid:4794
Authors:Xinyu Luo, Kecheng Chen, Pao-Sheng Sun, Chris Xing TIAN, Arindam Basu, Haoliang Li
Abstract:
Spiking Neural Networks (SNNs), as a biologically plausible alternative to Artificial Neural Networks (ANNs), have demonstrated advantages in terms of energy efficiency, temporal processing, and biological plausibility. However, SNNs are highly sensitive to distribution shifts, which can significantly degrade their performance in real-world scenarios. Traditional test-time adaptation (TTA) methods designed for ANNs often fail to address the unique computational dynamics of SNNs, such as sparsity and temporal spiking behavior. To address these challenges, we propose SPike-Aware Consistency Enhancement (SPACE), the first source-free and single-instance TTA method specifically designed for SNNs. SPACE leverages the inherent spike dynamics of SNNs to maximize the consistency of spike-behavior-based local feature maps across augmented versions of a single test sample, enabling robust adaptation without requiring source data. We evaluate SPACE on multiple datasets. Furthermore, SPACE exhibits robust generalization across diverse network architectures, consistently enhancing the performance of SNNs on CNNs (such as VGG and ResNet), Transformer models, and ConvLSTM architectures. Experimental results show that SPACE outperforms state-of-the-art methods, highlighting its effectiveness and robustness in real-world settings.



Paperid:4795
Authors:Pragya Singh, Ankush Gupta, Somay Jalan, Mohan Kumar, Pushpendra Singh
Abstract:
Emotion recognition from physiological signals has substantial potential for applications in mental health and emotion-aware systems. However, the lack of standardized, large-scale evaluations across heterogeneous datasets limits progress and model generalization. We introduce FEEL (Framework for Emotion Evaluation), the first large-scale benchmarking study of emotion recognition using electrodermal activity (EDA) and photoplethysmography (PPG) signals across 19 publicly available datasets. We evaluate 16 architectures spanning traditional machine learning, deep learning, and self-supervised pretraining approaches, structured into four representative modeling paradigms. Our study includes both within-dataset and cross-dataset evaluations, analyzing generalization across variations in experimental settings, device types, and labeling strategies. Our results showed that fine-tuned contrastive signal-language pretraining (CLSP) models (73/114) achieve the highest F1 across arousal and valence classification tasks, while simpler models like Random Forests, LDA, and MLP remain competitive (36/114). Models leveraging handcrafted features (109/114) consistently outperform those trained on raw signal segments, underscoring the value of domain knowledge in low-resource, noisy settings. Further cross-dataset analyses reveal that models trained on real-life setting data generalize well to lab (F1 = 0.79) and constraint-based settings (F1 = 0.78). Similarly, models trained on expert-annotated data transfer effectively to stimulus-labeled (F1 = 0.72) and self-reported datasets (F1 = 0.76). Moreover, models trained on lab-based devices also showed high transferability to both custom wearable (F1 = 0.81) and Empatica E4 (F1 = 0.73), underscoring the influence of heterogeneity. Overall, FEEL provides a unified framework for benchmarking physiological emotion recognition, delivering insights to guide the development of robust, generalizable emotion-aware technologies. Code implementation is provided in this https://anonymous.4open.science/r/FFEL-AE78/README.md.



Paperid:4796
Authors:David Dang, Myoung-Gyun Suh, Maodong Gao, ByoungJun Park, Beyonce Hu, Yucheng Jin, Wilton Kort-Kamp, Ho Lee
Abstract:
Endoscopic imaging is indispensable for visualizing internal organs, yet conventional systems remain bulky and costly because they rely on large, multi-element optics, which limits their ability to access and image certain areas of the body. Achieving high-quality endomicroscopy with hundred micron-scale and inexpensive hardware remains a grand challenge. Optical fibers offer a sub-millimeter-scale imaging conduit that could meet this need, but existing fiber-based approaches typically require either raster scanning or multicore bundles, which limit resolution and speed of imaging. In this work, we overcome these limitations by combining dual-comb interferometry with optical ghost imaging and advanced algorithm. Optical frequency combs enable precise and parallel speckle illumination via wavelength-division multiplexing through a single-core fiber, while our dual-comb compressive ghost imaging approach enables snapshot detection of bucket-sum signals using a single-pixel detector, eliminating the need for both spatial and spectral scanning. To reconstruct images from these highly compressed measurements, we introduce Optical Ghost-GPT, a transformer-based image reconstruction model that enables fast, high-fidelity recovery at low sampling rates. Our dual-comb ghost imaging approach, combined with the novel algorithm, outperforms classical ghost imaging techniques in both speed and accuracy, enabling real-time, high-resolution endoscopic imaging with a significantly reduced device footprint. This advancement paves the way for non-invasive, high-resolution, low-cost endomicroscopy and other sensing applications constrained by hardware size and complexity.



Paperid:4797
Authors:Amin Heyrani Nobari, Lyle Regenwetter, Cyril Picard, Ligong Han, Faez Ahmed
Abstract:
Structural topology optimization (TO) is essential in engineering design, enabling efficient material distribution for optimized physical performance. As such, it serves as an exemplary representative challenge, encompassing stringent performance goals, hard constraints, and substantial nonlinearities typical of real-world engineering design and physics-based optimization problems. However, existing deep-learning solutions are confined to fixed low-resolution square grids, a few hand-coded boundary conditions and loads, and post-processing with gradient optimizers, which limits practical deployment. We present Optimize Any Topology (OAT), the first foundation model that predicts minimum-compliance layouts directly for arbitrary aspect ratios, resolutions, volume fractions, loads, and fixtures. OAT combines an autoencoder with an implicit neural-field decoder, all trained on a new open corpus of 2.2 million optimized structures spanning 2M unique boundary-condition configurations we introduce called the OpenTO dataset. On four public benchmarks and two challenging unseen tests, OAT lowers mean compliance up to 90% relative to the best prior models and delivers sub-1 s inference on a single GPU across resolutions from 64×64 to 256x256 and aspect ratios as high as 10:1. These results establish a way for fast, shape and resolution-free general topology optimization and provide a large-scale dataset to spur further research in physics-aware generative modeling.



Paperid:4798
Authors:Yuzhou Gu, Yanjun Han, Jian Qian
Abstract:
We study the evolution of information in interactive decision making through the lens of a stochastic multi-armed bandit problem. Focusing on a fundamental example where a unique optimal arm outperforms the rest by a fixed margin, we characterize the optimal success probability and mutual information over time. Our findings reveal distinct growth phases in mutual information---initially linear, transitioning to quadratic, and finally returning to linear---highlighting curious behavioral differences between interactive and non-interactive environments. In particular, we show that optimal success probability and mutual information can be decoupled, where achieving optimal learning does not necessarily require maximizing information gain. These findings shed new light on the intricate interplay between information and learning in interactive decision making.



Paperid:4799
Authors:Jamie Hayes, Adam Dziedzic, A. Feder Cooper, Christopher Choquette-Choo, Franziska Boenisch, Georgios Kaissis, Igor Shilov, I Shumailov, Katherine Lee, Matthew Jagielski, Matthieu Meeus, Meenatchi Sundaram Muthu Selva Annamalai, Niloofar Mireshghallah, Yves-Alexandre de Montjoye, Milad Nasr
Abstract:
State-of-the-art membership inference attacks (MIAs) typically require training many reference models, making it difficult to scale these attacks to large pre-trained language models (LLMs). As a result, prior research has either relied on weaker attacks that avoid training reference models (e.g., fine-tuning attacks), or on stronger attacks applied to small-scale models and datasets.However, weaker attacks have been shown to be brittle---achieving close-to-arbitrary success---and insights from strong attacks in simplified settings do not translate to today's LLMs. These challenges have prompted an important question: are the limitations observed in prior work due to attack design choices, or are MIAs fundamentally ineffective on LLMs? We address this question by scaling LiRA---one of the strongest MIAs--to GPT-2 architectures ranging from 10M to 1B parameters, training reference models on over 20B tokens from the C4 dataset. Our results advance the understanding of MIAs on LLMs in three key ways: (1) strong MIAs can succeed on pre-trained LLMs;(2) their effectiveness, however, remains limited (e.g., AUC < 0.7) in practical settings; and, (3) the relationship between MIA success and related privacy metrics is not as straightforward as prior work has suggested.



Paperid:4800
Authors:Hoang Pham, The Anh Ta, Tom Jacobs, Rebekka Burkholz, Long Tran-Thanh
Abstract:
Sparse neural networks promise efficiency, yet training them effectively remains a fundamental challenge. Despite advances in pruning methods that create sparse architectures, understanding why some sparse structures are better trainable than others with the same level of sparsity remains poorly understood. Aiming to develop a systematic approach to this fundamental problem, we propose a novel theoretical framework based on the theory of graph limits, particularly graphons, that characterizes sparse neural networks in the infinite-width regime. Our key insight is that connectivity patterns of sparse neural networks induced by pruning methods converge to specific graphons as networks' width tends to infinity, which encodes implicit structural biases of different pruning methods. We postulate theGraphon Limit Hypothesisand provide empirical evidence to support it. Leveraging this graphon representation, we derive aGraphon Neural Tangent Kernel (Graphon NTK)to study the training dynamics of sparse networks in the infinite width limit. Graphon NTK provides a general framework for the theoretical analysis of sparse networks. We empirically show that the spectral analysis of Graphon NTK correlates with observed training dynamics of sparse networks, explaining the varying convergence behaviours of different pruning methods. Our framework provides theoretical insights into the impact of connectivity patterns on the trainability of various sparse network architectures.



Paperid:4801
Authors:Tianxing Man, Yu Bai, Ganyu Wang, Jinjie Fang, Haoran Fang, Bin Gu, Yi Chang
Abstract:
Abstract:Vertical Federated Learning (VFL) enables participants to collaboratively train models on aligned samples while keeping their heterogeneous features private and distributed.Despite their utility, VFL models remain vulnerable to adversarial attacks during inference. Adversarial Training (AT), which generates adversarial examples at each training iteration, stands as the most effective defense for improving model robustness. However, applying AT in VFL settings (VFAL) faces significant computational efficiency challenges, as the distributed training framework necessitates iterative propagations across participants.To this end, we propose **_DecVFAL_** framework, which substantially accelerates **_VFAL_** training through a dual-level ***Dec***oupling mechanism applied during adversarial sample generation.Specifically, we first decouple the bottom modules of clients (directly responsible for adversarial updates) from the remaining networks, enabling efficient _lazy sequential propagations_ that reduce communication frequency through delayed gradients. We further introduce _decoupled parallel backpropagation_ to accelerate delayed gradient computation by eliminating idle waiting through parallel processing across modules.Additionally, we are the first to establish convergence analysis for VFAL, rigorously characterizing how our decoupling mechanism interacts with existing VFL dynamics, and prove that _DecVFAL_ achieves an $\mathcal{O}(1/\sqrt{K})$ convergence rate matching that of standard VFLs.Experimental results show that _DecVFAL_ ensures competitive robustness while significantly achieving about $3\sim10\times$ speed up.



Paperid:4802
Authors:Mengqi Guo, Bo Xu, Yanyan Li, Gim Hee Lee
Abstract:
Novel view synthesis from monocular videos of dynamic scenes with unknown camera poses remains a fundamental challenge in computer vision and graphics. While recent advances in 3D representations such as Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) have shown promising results for static scenes, they struggle with dynamic content and typically rely on pre-computed camera poses.We present 4D3R, a pose-free dynamic neural rendering framework that decouples static and dynamic components through a two-stage approach. Our method first leverages 3D foundational models for initial pose and geometry estimation, followed by motion-aware refinement. 4D3R introduces two key technical innovations: (1) a motion-aware bundle adjustment (MA-BA) module that combines transformer-based learned priors with SAM2 for robust dynamic object segmentation, enabling more accurate camera pose refinement; and (2) an efficient Motion-Aware Gaussian Splatting (MA-GS) representation that uses control points with a deformation field MLP and linear blend skinning to model dynamic motion, significantly reducing computational cost while maintaining high-quality reconstruction.Extensive experiments on real-world dynamic datasets demonstrate that our approach achieves up to 1.8dB PSNR improvement over state-of-the-art methods, particularly in challenging scenarios with large dynamic objects, while reducing computational requirements by 5× compared to previous dynamic scene representations.



Paperid:4803
Authors:Cheng Yan, Felix Mohr, Tom Viering
Abstract:
Sample-wise learning curves plot performance versus training set size. They are useful for studying scaling laws and speeding up hyperparameter tuning and model selection. Learning curves are often assumed to be well-behaved: monotone (i.e. improving with more data) and convex. By constructing the Learning Curves Database 1.1 (LCDB 1.1), a large-scale database with high-resolution learning curves, we show that learning curves are less often well-behaved than previously thought. Using statistically rigorous methods, we observe significant ill-behavior in approximately 14% of the learning curves, almost twice as much as in previous estimates. We also identify which learners are to blame and show that specific learners are more ill-behaved than others. Additionally, we demonstrate that different feature scalings rarely resolve ill-behavior. We evaluate the impact of ill-behavior on downstream tasks, such as learning curve fitting and model selection, and find it poses significant challenges, underscoring the relevance and potential of LCDB 1.1 as a challenging benchmark for future research.



Paperid:4804
Authors:Xiaomeng Yang, LEI LU, Qihui Fan, Changdi Yang, Juyi Lin, Yanzhi Wang, Xuan Zhang, Shangqian Gao
Abstract:
Abstract:Diffusion models have demonstrated exceptional capabilities in generating high-fidelity images. However, their iterative denoising process results in significant computational overhead during inference, limiting their practical deployment in resource-constrained environments. Existing acceleration methods often adopt uniform strategies that fail to capture the temporal variations during diffusion generation, while the commonly adopted sequential $\textit{pruning-then-fine-tuning strategy}$ suffers from sub-optimality due to the misalignment between pruning decisions made on pretrained weights and the model’s final parameters. To address these limitations, we introduce $\textbf{ALTER}$: $\textbf{A}$ll-in-One $\textbf{L}$ayer Pruning and $\textbf{T}$emporal $\textbf{E}$xpoert $\textbf{R}$outing, a unified framework that transforms diffusion models into a mixture of efficient temporal experts.ALTER achieves a single-stage optimization that unifies layer pruning, expert routing, and model fine-tuning by employing a trainable hypernetwork, which dynamically generates layer pruning decisions and manages timestep routing to specialized, pruned expert sub-networks throughout the ongoing fine-tuning of the UNet. This unified co-optimization strategy enables significant efficiency gains while preserving high generative quality. Specifically, ALTER achieves same-level visual fidelity to the original 50-step Stable Diffusion v2.1 model while utilizing only 25.9\% of its total MACs with just 20 inference steps and delivering a 3.64$\times$ speedup through 35\% sparsity.



Paperid:4805
Authors:Yifu Guo, Jiaye Lin, Huacan Wang, Yuzhen Han, Sen Hu, Ziyi Ni, Licheng Wang, Mingguang Chen
Abstract:
Large Language Model (LLM)-based agents have recently shown impressive capabilities in complex reasoning and tool use via multi-step interactions with their environments. While these agents have the potential to tackle complicated tasks, their problem-solving process—agents' interaction trajectory leading to task completion—remains underexploited. These trajectories contain rich feedback that can navigate agents toward the right directions for solving problems correctly. Although prevailing approaches, such as Monte Carlo Tree Search (MCTS), can effectively balance exploration and exploitation, they ignore the interdependence among various trajectories and lack the diversity of search spaces, which leads to redundant reasoning and suboptimal outcomes. To address these challenges, we propose SE-Agent, a \textbf{S}elf-\textbf{E}volution framework that enables \textbf{Agents} to optimize their reasoning processes iteratively. Our approach revisits and enhances former pilot trajectories through three key operations: revision, recombination, and refinement. This evolutionary mechanism enables two critical advantages: (1) it expands the search space beyond local optima by intelligently exploring diverse solution paths guided by previous trajectories, and (2) it leverages cross-trajectory inspiration to efficiently enhance performance while mitigating the impact of suboptimal reasoning paths. Through these mechanisms, SE-Agent achieves continuous self-evolution that incrementally improves reasoning quality. We evaluate SE-Agent on SWE-bench Verified to resolve real-world GitHub issues. Experimental results across five strong LLMs show that integrating SE-Agent delivers up to \textbf{55\%} relative improvement, achieving \textbf{state-of-the-art} performance among all open-source agents on SWE-bench Verified.



Paperid:4806
Authors:Ziqiao Wang, Wangbo Zhao, Yuhao Zhou, Zekai Li, Zhiyuan Liang, Mingjia Shi, Xuanlei Zhao, Pengfei Zhou, Kaipeng Zhang, Zhangyang "Atlas" Wang, Kai Wang, Yang You
Abstract:
Abstract:Diffusion Transformers (DiTs) deliver state-of-the-art image quality, yet their training remains notoriously slow. A recent remedy---representation alignment (REPA) that matches DiT hidden features to those of a non-generative teacher (e.g. DINO)---dramatically accelerates the early epochs but plateaus or even degrades performance later. We trace this failure to a capacity mismatch: once the generative student begins modelling the joint data distribution, the teacher's lower-dimensional embeddings and attention patterns become a straitjacket rather than a guide. We then introduce HASTE (Holistic Alignment with Stage-wise Termination for Efficient training), a two-phase schedule that keeps the help and drops the hindrance. Phase I applies a holistic alignment loss that simultaneously distills attention maps (relational priors) and feature projections (semantic anchors) from the teacher into mid-level layers of the DiT, yielding rapid convergence. Phase II then performs one-shot termination that deactivates the alignment loss, once a simple trigger such as a fixed iteration is hit, freeing the DiT to focus on denoising and exploit its generative capacity. HASTE speeds up training of diverse DiTs without architecture changes. On ImageNet $256{\times}256$, it reaches the vanilla SiT-XL/2 baseline FID in 50 epochs and matches REPA’s best FID in 500 epochs, amounting to a $\boldsymbol{28\times}$ reduction in optimization steps. HASTE also improves text-to-image DiTs on MS-COCO, demonstrating to be a simple yet principled recipe for efficient diffusion training across various tasks.



Paperid:4807
Authors:Yuante Li, Xu Yang, Xiao Yang, Xisen Wang, Weiqing Liu, Jiang Bian
Abstract:
Financial markets pose fundamental challenges for asset return prediction due to their high dimensionality, non-stationarity, and persistent volatility. Despite advances in large language models and multi-agent systems, current quantitative research pipelines suffer from limited automation, weak interpretability, and fragmented coordination across key components such as factor mining and model innovation. In this paper, we propose R&D-Agent for Quantitative Finance, in short RD-Agent(Q), the first data-centric multi-agent framework designed to automate the full-stack research and development of quantitative strategies via coordinated factor-model co-optimization. RD-Agent(Q) decomposes the quant process into two iterative stages: a Research stage that dynamically sets goal-aligned prompts, formulates hypotheses based on domain priors, and maps them to concrete tasks, and a Development stage that employs a code-generation agent, Co-STEER, to implement task-specific code, which is then executed in real-market backtests. The two stages are connected through a feedback stage that thoroughly evaluates experimental outcomes and informs subsequent iterations, with a multi-armed bandit scheduler for adaptive direction selection. Empirically, RD-Agent(Q) achieves up to 2× higher annualized returns than classical factor libraries using 70% fewer factors, and outperforms state-of-the-art deep time-series models on real markets. Its joint factor–model optimization delivers a strong balance between predictive accuracy and strategy robustness. Our code is available at: https://github.com/microsoft/RD-Agent.



Paperid:4808
Authors:Junxi Chen, Liang Li, Yunbin Tu, Li Su, Zhe Xue, Qingming Huang
Abstract:
Video Anomaly Detection (VAD) aims to identify abnormal frames from discrete events within video sequences. Existing VAD methods suffer from heavy annotation burdens in fully-supervised paradigm, omission of subtle anomalies in semi-supervised paradigm, and vulnerability to noise in weakly-supervised paradigm. To address these limitations, we propose a novel paradigm: Single-Frame supervised VAD (SF-VAD), which uses a single annotated abnormal frame per abnormal video. SF-VAD ensures annotation efficiency while offering precise anomaly reference, facilitating robust anomaly modeling, and enhancing the detection of subtle anomalies in complex visual contexts. To validate its effectiveness, we construct three SF-VAD benchmarks by manually re-annotating the ShanghaiTech, UCF-Crime, and XD-Violence datasets in a practical procedure. Further, we devise Frame-guided Progressive Learning (FPL), to generalize sparse frame supervision to event-level anomaly understanding. FPL first leverages evidential learning to estimate anomaly relevance guided by annotated frames. Then it extends anomaly supervision by mining discrete abnormal events based on anomaly relevance and feature similarity. Meanwhile, FPL decouples normal patterns by isolating distinct normal frames outside abnormal events, reducing false alarms. Extensive experiments show SF-VAD achieves state-of-the-art detection results while offering a favorable trade-off between performance and annotation cost



Paperid:4809
Authors:Haochen Zhang, Junze Yin, Guanchu Wang, Zirui Liu, Lin Yang, Tianyi Zhang, Anshumali Shrivastava, Vladimir Braverman
Abstract:
Low-rank optimization has emerged as a promising approach to enabling memory-efficient training of large language models (LLMs). Existing low-rank optimization methods typically project gradients onto a low-rank subspace, reducing the memory cost of storing optimizer states. A key challenge in these methods is selecting suitable subspaces to ensure an effective optimization trajectory. Most existing approaches select the dominant subspace to preserve gradient information, as this intuitively provides the best approximation. However, we find that in practice, the dominant subspace stops changing during pretraining, thereby constraining weight updates to similar subspaces. In this paper, we propose importance sampling for low-rank optimization in LLM pretraining with a provable convergence guarantee, which the dominant subspace approach does not have. Empirically, we demonstrate that our method significantly outperforms previous methods in LLM pretraining tasks.



Paperid:4810
Authors:Ryotaro Kawata, Yujin Song, Alberto Bietti, Naoki Nishikawa, Taiji Suzuki, Samuel Vaiter, Denny Wu
Abstract:
Transformers can implement both generalizable algorithms (e.g., induction heads) and simple positional shortcuts (e.g., memorizing fixed output positions). In this work, we study how the choice of pretraining data distribution steers a shallow transformer toward one behavior or the other. Focusing on a minimal trigger–output prediction task — copying the token immediately following a special trigger upon its second occurrence — we develop a rigorous analysis of gradient-based training of a single-layer transformer.In the infinite-sample regime, we prove a transition in the learned mechanism: if input sequences exhibit sufficient diversity, measured by a low “max–sum” ratio of trigger-to-trigger distances, the trained model implements an induction head that generalizes to unseen contexts; by contrast, when this ratio is large, the model resorts to positional memorization and fails to generalize out-of-distribution. Extending to finite samples, we quantify the number of pretraining sequences required to cross this transition and derive the optimal context-length distribution that minimizes sample complexity while ensuring induction-head learning. Finally, we validate our theoretical predictions with controlled synthetic experiments, demonstrating that broadening context distributions robustly induces induction heads and enables length extrapolation. Our results shed light on the algorithmic biases of pretrained transformers and offer practical guidelines for data-driven control of their learned behaviors.



Paperid:4811
Authors:Yuan Yao, Haitian Zheng, yongsheng yu, Yuqian Zhou, Zhe Lin, Jiebo Luo
Abstract:
Latent Diffusion Models (LDMs) have markedly advanced the quality of image inpainting and local editing. However, the inherent latent compression often introduces pixel-level inconsistencies, such as chromatic shifts, texture mismatches, and visible seams along editing boundaries. Existing remedies, including background-conditioned latent decoding and pixel-space harmonization, usually fail to fully eliminate these artifacts in practice and do not generalize well across different latent representations or tasks. We introduce PixPerfect, a pixel‐level refinement framework that delivers seamless, high-fidelity local edits across diverse LDM architectures and tasks. PixPerfect leverages (i) a differentiable discriminative pixel space that amplifies and suppresses subtle color and texture discrepancies, (ii) a comprehensive artifact simulation pipeline that exposes the refiner to realistic local editing artifacts during training, and (iii) a direct pixel-space refinement scheme that ensures broad applicability across diverse latent representations and tasks. Extensive experiments on inpainting, object removal, and insertion benchmarks demonstrate that PixPerfect substantially enhances perceptual fidelity and downstream editing performance, establishing a new standard for robust and high-fidelity localized image editing.



Paperid:4812
Authors:Long Ma, Zhiyuan Yan, Jin Xu, Yize Chen, Qinglang Guo, Zhen Bi, Yong Liao, Hui Lin
Abstract:
Detecting deepfakes has been an increasingly important topic, especially given the rapid development of AI generation techniques.In this paper, we ask: How can we build a universal detection framework that is effective for most facial deepfakes?One significant challenge is the wide variety of deepfake generators available, resulting in varying forgery artifacts (e.g., lighting inconsistency, color mismatch, etc).But should we ``teach" the detector to learn all these artifacts separately? It is impossible and impractical to elaborate on them all.So the core idea is to pinpoint the more common and general artifacts across different deepfakes.Accordingly, we categorize deepfake artifacts into two distinct yet complementary types: Face Inconsistency Artifacts (FIA) and Up-Sampling Artifacts (USA). FIA arise from the challenge of generating all intricate details, inevitably causing inconsistencies between the complex facial features and relatively uniform surrounding areas.USA, on the other hand, are the inevitable traces left by the generator's decoder during the up-sampling process.This categorization stems from the observation that all existing deepfakes typically exhibit one or both of these artifacts.To achieve this, we propose a new data-level pseudo-fake creation framework that constructs fake samples with only the FIA and USA, without introducing extra less-general artifacts.Specifically, we employ a super-resolution to simulate the USA, while utilise image-level self-blending on diverse facial regions to create the FIA.We surprisingly found that, with this intuitive design, a standard image classifier trained only with our pseudo-fake data can non-trivially generalize well to previously unseen deepfakes.



Paperid:4813
Authors:Lin Li, Chuhan ZHANG, Dong Zhang, Chong Sun, Chen Li, Long Chen
Abstract:
Open-vocabulary scene graph generation (OVSGG) extends traditional SGG by recognizing novel objects and relationships beyond predefined categories, leveraging the knowledge from pre-trained large-scale models. Existing OVSGG methods always adopt a two-stage pipeline: 1) Infusing knowledge into large-scale models via pre-training on large datasets; 2) Transferring knowledge from pre-trained models with fully annotated scene graphs during supervised fine-tuning. However, due to the lack of explicit interaction modeling, these methods struggle to distinguish between interacting and non-interacting instances of the same object category, which significantly exacerbates relation pair mismatches. To this end, in this paper, we propose an interACtion-Centric end-to-end OVSGG framework (ACC) in an interaction-driven paradigm to minimize these mismatches. For interaction-centric knowledge infusion, ACC employs a bidirectional interaction prompt for robust pseudo-supervision generation to enhance the model's interaction knowledge. For interaction-centric knowledge transfer, ACC first adopts interaction-guided query selection that prioritizes pairing interacting objects to reduce interference from non-interacting ones. Then, it integrates interaction-consistent knowledge distillation to bolster robustness by pushing relational foreground away from the background while retaining general knowledge. Extensive experimental results on three benchmarks show that ACC achieves state-of-the-art performance, demonstrating the potential of interaction-centric paradigms for real-world applications.



Paperid:4814
Authors:Yao Teng, Fu-Yun Wang, Xian Liu, Zhekai Chen, Han Shi, Yu Wang, Zhenguo Li, Weiyang Liu, Difan Zou, Xihui Liu
Abstract:
As a new paradigm of visual content generation, autoregressive text-to-image models suffer from slow inference due to their sequential token-by-token decoding process, often requiring thousands of model forward passes to generate a single image. To address this inefficiency, we propose Speculative Jacobi-Denoising Decoding (SJD2), a framework that incorporates the denoising process into Jacobi iterations to enable parallel token generation in autoregressive models. Our method introduces a next-clean-token prediction paradigm that enables the pre-trained autoregressive models to accept noise-perturbed token embeddings and predict the next clean tokens through low-cost fine-tuning. This denoising paradigm is beneficial to the stabilization of the Jacobi trajectories. During inference, our method initializes token sequences with Gaussian noise and performs iterative denoising in the embedding space. Concurrently with the denoising, we employ a probabilistic criterion to verify and accept multiple tokens in parallel. Experiments show that our method can accelerate generation by reducing model forward passes while maintaining the visual quality of generated images.



Paperid:4815
Authors:Yimu Zhang, Dongqi Han, Yansen Wang, Zhenning Lv, Yu Gu, Dongsheng Li
Abstract:
Spike sorting is an essential process in neural recording, which identifies and separates electrical signals from individual neurons recorded by electrodes in the brain, enabling researchers to study how specific neurons communicate and process information. Although there exist a number of spike sorting methods which have contributed to significant neuroscientific breakthroughs, many are heuristically designed, making it challenging to verify their correctness due to the difficulty of obtaining ground truth labels from real-world neural recordings. In this work, we explore a data-driven, deep learning-based approach. We begin by creating a large-scale dataset through electrophysiology simulations using biologically realistic computational models. We then present SimSort, a pretraining framework for spike sorting. Trained solely on simulated data, SimSort demonstrates zero-shot generalizability to real-world spike sorting tasks, yielding consistent improvements over existing methods across multiple benchmarks. These results highlight the potential of simulation-driven pretraining to enhance the robustness and scalability of spike sorting in experimental neuroscience.



Paperid:4816
Authors:Chaoya Jiang, Yongrui Heng, Wei Ye, Haiyang Xu, Ming Yan, Ji Zhang, Fei Huang, Shikun Zhang
Abstract:
Abstract:Recently, reasoning-based MLLMs have achieved a degree of success in generating long-form textual reasoning chains. However, they still struggle with complex tasks that necessitate dynamic and iterative focusing on and revisiting of visual regions to achieve precise grounding of textual reasoning in visual evidence. We introduce VLM-R³ (Visual Language Model with Region Recognition, Reasoning, and Refinement ), a framework that equips an MLLM with the ability to (i) decide when additional visual evidence is needed, (ii) determine where to ground within the image, and (iii) seamlessly weave the relevant sub-image content back into an interleaved chain-of-thought. The core of our method is \textbf{Region-Conditioned Reinforcement Policy Optimization (R-GRPO)}, a training paradigm that rewards the model for selecting informative regions, formulating appropriate transformations (e.g. crop, zoom), and integrating the resulting visual context into subsequent reasoning steps. To bootstrap this policy, we compile a modest but carefully curated Visuo-Lingual Interleaved Rationale (VLIR) corpus that provides step-level supervision on region selection and textual justification. Extensive experiments on MathVista, ScienceQA, and other benchmarks show that VLM-R$^3$ sets a new state of the art in zero-shot and few-shot settings, with the largest gains appearing on questions demanding subtle spatial reasoning or fine-grained visual cue extraction.



Paperid:4817
Authors:Qing Yu, Xiaobei Wang, Shuchang Liu, yandong.bai, Xiaoyu Yang, Xueliang Wang, Chang Meng, Shanshan Wu, HailanYang, Bin Wen, Huihui Xiao, Xiang Li, Fan Yang, Xiaoqiang Feng, Lantao Hu, Han Li, Kun Gai, Lixin Zou
Abstract:
Recommender systems filter contents/items valuable to users by inferring preferences from user features and historical behaviors. Mainstream approaches follow the learning-to-rank paradigm, which focus on discovering and modeling item topics (e.g.,categories), and capturing user preferences on these topics based on historical interactions.However, this paradigm often neglects the modeling of user characteristics and their social roles, which are logical confounders influencing the correlated interest and user preference transition. To bridge this gap, we introduce the user role identification task and the behavioral logic modeling task that aim to explicitly model user roles and learn the logical relations between item topics and user social roles. We show that it is possible to explicitly solve these tasks through an efficient integration framework of Large Language Model (LLM) and recommendation systems, for which we propose TagCF. On the one hand, TagCF exploits the (Multi-modal) LLM's world knowledge and logic inference ability to extract realistic tag-based virtual logic graphs that reveal dynamic and expressive knowledge of users, refining our understanding of user behaviors.On the other hand, TagCF presents empirically effective integration modules that take advantage of the extracted tag-logic information, augmenting the recommendation performance.We conduct both online experiments and offline experiments with industrial and public datasets as verification of TagCF's effectiveness, and we empirically show that the user role modeling strategy is potentially a better choice than the modeling of item topics. Additionally, we provide evidence that the extracted logic graphs are empirically a general and transferable knowledge that can benefit a wide range of recommendation tasks.



Paperid:4818
Authors:Sukanya Patra, Souhaib Ben Taieb
Abstract:
Unsupervised anomaly detection (AD) methods typically assume clean training data, yet real-world datasets often contain undetected or mislabeled anomalies, leading to significant performance degradation. Existing solutions require access to the training pipelines, data or prior knowledge of the proportions of anomalies in the data, limiting their real-world applicability. To address this challenge, we propose EPHAD, a simple yet effective inference-time adaptation framework that updates the outputs of AD models trained on contaminated datasets using evidence gathered at inference. Our approach formulates test-time adaptation as a Bayesian inference problem, integrating the prior knowledge captured by the AD model trained on contaminated datasets with auxiliary evidence derived from foundation models like CLIP, classical methods like the Latent Outlier Factor or domain-specific knowledge. We illustrate the intuition behind EPHAD using a synthetic toy example and validate its effectiveness through comprehensive experiments across eight image-based AD datasets, twenty-seven tabular datasets, and a real-world industrial dataset. Additionally, we conduct an ablation study to analyse hyperparameter influence and robustness to varying contamination levels, demonstrating the versatility and robustness of EPHAD across diverse AD models and evidence pairs. To ensure reproducibility, our code is publicly available https://anonymous.4open.science/r/EPAF-2025/.



Paperid:4819
Authors:Zeqin Yu, Haotao Xie, Jian Zhang, Jiangqun Ni, Wenkang Su, Jiwu Huang
Abstract:
Existing Text Image Forgery Localization (T-IFL) methods often suffer from poor generalization due to the limited scale of real-world datasets and the distribution gap caused by synthetic data that fails to capture real-world tampering complexity. To tackle this issue, we propose Fourier Series-based Tampering Synthesis (FSTS), a structured and interpretable framework for synthesizing tampered text images. TFSTS first collects 16,750 real-world tampering instances from five representative manipulation types, using a structured pipeline that records human-performed editing traces via multi-format logs (e.g., video, PSD, and editing logs). By analyzing these collected parameters and identify recurring behavioral patterns at both individual and population levels, we formulate a hierarchical modeling framework, where each individual’s tampering parameter is represented as a compact combination of basis operation–parameter configurations, and the population-level distribution is constructed by aggregating these behaviors. Since this formulation draws inspiration from Fourier series, it enable an interpretable approximation using basis functions and their learned weights. By sampling from this modeled distribution, FSTS synthesizes diverse and realistic training data that better reflect real-world forgery traces. Extensive experiments across four evaluation protocols demonstrate that models trained with FSTS data achieve significantly improved generalization on real-world datasets.



Paperid:4820
Authors:SangYong Lee, Sangjun Chung, Simon Woo
Abstract:
With increasing regulations on private data usage in AI systems, machine unlearning has emerged as a critical solution for selectively removing sensitive information from trained models while preserving their overall utility. While many existing unlearning methods rely on the \textit{retain data} to mitigate the performance decline caused by forgetting, such data may not always be available (\textit{retain-free}) in real-world scenarios. To address this challenge posed by retain-free unlearning, we introduce RUAGO utilizing adversarial soft labels to mitigate over-unlearning and a generative model pretrained on out-of-distribution (OOD) data to effectively distill the original model’s knowledge. We introduce a progressive sampling strategy to incrementally increase synthetic data complexity, coupled with an inversion-based alignment step that ensures the synthetic data closely matches the original training distribution. Our extensive experiments on multiple benchmark datasets and architectures demonstrate that our approach consistently outperforms existing retain-free methods and achieves comparable or superior performance relative to retain-based approaches, demonstrating its effectiveness and practicality in real-world, data-constrained environments.



Paperid:4821
Authors:Yue Feng, Jinwei Hu, Qijia Lu, Jiawei Niu, Li Tan, Shuo Yuan, Ziyi Yan, Yizhen Jia, Qingzhi He, Shiping Ge, Ethan Chen, Wentong Li, Limin Wang, Jie Qin
Abstract:
We propose the Multi-modal Untrimmed Video Retrieval task, along with a new benchmark (MUVR) to advance video retrieval for long-video platforms. MUVR aims to retrieve untrimmed videos containing relevant segments using multi-modal queries. It has the following features:1) Practical retrieval paradigm:MUVR supports video-centric multi-modal queries, expressing fine-grained retrieval needs through long text descriptions, video tag prompts, and mask prompts. It adopts a one-to-many retrieval paradigm and focuses on untrimmed videos, tailored for long-video platform applications.2) Multi-level visual correspondence:To cover common video categories (e.g., news, travel, dance) and precisely define retrieval matching criteria, we construct multi-level visual correspondence based on core video content (e.g., news events, travel locations, dance moves) which users are interested in and want to retrieve. It covers six levels: copy, event, scene, instance, action, and others.3) Comprehensive evaluation criteria:We develop 3 versions of MUVR (i.e., Base, Filter, QA). MUVR-Base/Filter evaluates retrieval models, while MUVR-QA assesses MLLMs in a question-answering format. We also propose a Reranking Score to evaluate the reranking ability of MLLMs. MUVR consists of 53K untrimmed videos from the video platform Bilibili, with 1,050 multi-modal queries and 84K matches. Extensive evaluations of 3 state-of-the-art video retrieval models, 6 image-based VLMs, and 10 MLLMs are conducted. MUVR reveals the limitations of retrieval methods in processing untrimmed videos and multi-modal queries, as well as MLLMs in multi-video understanding and reranking. Our code and benchmark will be open-sourced soon.



Paperid:4822
Authors:Zhilin Wang, Jiaqi Zeng, Olivier Delalleau, Hoo-Chang Shin, Felipe Soares, Alexander Bukharin, Ellie Evans, Yi Dong, Oleksii Kuchaiev
Abstract:
Preference datasets are essential for training general-domain, instruction-following language models with Reinforcement Learning from Human Feedback (RLHF). Each subsequent data release raises expectations for future data collection, meaning there is a constant need to advance the quality and diversity of openly available preference data. To address this need, we introduce HelpSteer3-Preference, a permissively licensed (CC-BY-4.0), high-quality, human-annotated preference dataset comprising of over 40,000 samples. These samples span diverse real-world applications of large language models (LLMs), including tasks relating to STEM, coding and multilingual scenarios. Using HelpSteer3-Preference, we train Reward Models (RMs) that achieve top performance on RM-Bench (82.4%) and JudgeBench (73.7%). This represents a substantial improvement (~10% absolute) over the previously best-reported results from existing RMs. We demonstrate HelpSteer3-Preference can also be applied to train Generative RMs and how policy models can be aligned with RLHF using our RMs.



Paperid:4823
Authors:Burouj Armgaan, Eshan Jain, Harsh Pandey, Mahesh Chandran, Sayan Ranu
Abstract:
Abstract:Graph Neural Networks (GNNs) are widely used for node classification, yet their opaque decision-making limits trust and adoption. While local explanations offer insights into individual predictions, global explanation methods—those that characterize an entire class—remain underdeveloped. Existing global explainers rely on motif discovery in small graphs, an approach that breaks down in large, real-world settings where subgraph repetition is rare, node attributes are high-dimensional, and predictions arise from complex structure-attribute interactions. We propose GnnXemplar, a novel global explainer inspired from Exemplar Theory from cognitive science. GnnXemplar identifies representative nodes in the GNN embedding space—exemplars—and explains predictions using natural language rules derived from their neighborhoods. Exemplar selection is framed as a coverage maximization problem over reverse $k$-nearest neighbors, for which we provide an efficient greedy approximation. To derive interpretable rules, we employ a self-refining prompt strategy using large language models (LLMs). Experiments across diverse benchmarks show that GnnXemplar significantly outperforms existing methods in fidelity, scalability, and human interpretability, as validated by a user study with 60 participants.



Paperid:4824
Authors:Lifan Zhao, Yanyan Shen, Zhaoyang Liu, xue wang, Jiaji Deng
Abstract:
Scaling laws motivate the development of Time Series Foundation Models (TSFMs) that pre-train vast parameters and achieve remarkable zero-shot forecasting performance. Surprisingly, even after fine-tuning, TSFMs cannot consistently outperform smaller, specialized models trained on full-shot downstream data. A key question is how to regularize the adaptation process of TSFMs for a target forecasting task. Through empirical studies on various TSFMs, the pre-trained models often exhibit inherent sparsity and redundancy in computation, suggesting that TSFMs have learned to activate task-relevant network substructures to accommodate diverse forecasting tasks. To preserve this valuable prior knowledge, we propose structured pruning method to regularize the subsequent fine-tuning process by focusing it on a more relevant and compact parameter space. Extensive experiments on seven TSFMs and six benchmarks demonstrate that fine-tuning a smaller, pruned TSFM significantly improves forecasting performance compared to fine-tuning original models. This ``prune-then-finetune'' paradigm often enables TSFMs to achieve state-of-the-art performance and surpass strong specialized baselines. We made our code available at https://anonymous.4open.science/r/Prune-then-Finetune/.



Paperid:4825
Authors:Manchao Bao, Shengjiang Fang, Tao Yue, Xuemei Hu
Abstract:
Long-distance depth imaging holds great promise for applications such as autonomous driving and robotics. Direct time-of-flight (dToF) imaging offers high-precision, long-distance depth sensing, yet demands ultra-short pulse light sources and high-resolution time-to-digital converters. In contrast, indirect time-of-flight (iToF) imaging often suffers from phase wrapping and low signal-to-noise ratio (SNR) as the sensing distance increases. In this paper, we introduce a novel ToF imaging paradigm, termed Burst-Encodable Time-of-Flight (BE-ToF), which facilitates high-fidelity, long-distance depth imaging. Specifically, the BE-ToF system emits light pulses in burst mode and estimates the phase delay of the reflected signal over the entire burst period, thereby effectively avoiding the phase wrapping inherent to conventional iToF systems. Moreover, to address the low SNR caused by light attenuation over increasing distances, we propose an end-to-end learnable framework that jointly optimizes the coding functions and the depth reconstruction network. A specialized double well function and first-order difference term are incorporated into the framework to ensure the hardware implementability of the coding functions. The proposed approach is rigorously validated through comprehensive simulations and real-world prototype experiments, demonstrating its effectiveness and practical applicability.



Paperid:4826
Authors:Miruna Oprescu, David Park, Xihaier Luo, Shinjae Yoo, Nathan Kallus
Abstract:
Estimating causal effects from spatiotemporal observational data is essential in public health, environmental science, and policy evaluation, where randomized experiments are often infeasible. However, existing approaches typically rely on strong structural assumptions or fail to account for key challenges such as interference, spatial confounding, temporal carryover, andtime-varying confounding— where covariates are influenced by past treatments and, in turn, affect future treatments and outcomes. In this work, we propose theGST-UNet(G-computationSpatio-TemporalUNet), an end-to-end neural framework that integrates a U-Net-based spatiotemporal embedding with regression-based iterative G-computation to estimate location-specific potential outcomes under complex intervention sequences. The GST-UNet explicitly adjusts for time-varying confounders and accommodates non-linear spatial and temporal dependencies, enabling valid causal inference in settings with limited data and complex feedback dynamics. We demonstrate its effectiveness in synthetic experiments and apply it to estimate the impact of wildfire smoke on respiratory hospitalizations during the 2018 California Camp Fire. Our results highlight GST-UNet as a ready-to-use framework for practitioners seeking reliable causal estimates from spatiotemporal data, with the potential to advance causal inference across a wide range of policy-driven and scientific applications.



Paperid:4827
Authors:Itamar Harel, Yonathan Wolanowsky, Gal Vardi, Nati Srebro, Daniel Soudry
Abstract:
Abstract:We analyze the generalization gap (gap between the training and test errors) when training a potentially over-parametrized model using a Markovian stochastic training algorithm, initialized from some distribution $\theta_0 \sim p_0$. We focus on Langevin dynamics with a positive temperature $\beta^{-1}$, i.e. gradient descent on a training loss $L$ with infinitesimal step size, perturbed with $\beta^{-1}$-variances Gaussian noise, and lightly regularized or bounded.There, we bound the generalization gap, *at any time during training*, by $\sqrt{(\beta\mathbb{E} L (\theta_0) + \log(1/\delta))/N}$ with probability $1-\delta$ over the dataset, where $N$ is the sample size, and $\mathbb{E} L(\theta_0)=O(1)$ with standard initialization scaling. In contrast to previous guarantees, we have no dependence on either training time or reliance on mixing, nor a dependence on dimensionality, gradient norms, or any other properties of the loss or model. This guarantee follows from a general analysis of any Markov process-based training that has a Gibbs-style stationary distribution. The proof is surprisingly simple, once we observe that the marginal distribution divergence from initialization remains bounded, as implied by a generalized second law of thermodynamics.



Paperid:4828
Authors:Lucas Piper, Arlindo L Oliveira, Tiago Marques
Abstract:
Convolutional neural networks (CNNs) trained on object recognition achieve high task performance but continue to exhibit vulnerability under a range of visual perturbations and out-of-domain images, when compared with biological vision. Prior work has demonstrated that coupling a standard CNN with a front-end block (VOneBlock) that mimics the primate primary visual cortex (V1) can improve overall model robustness. Expanding on this, we introduce Early Vision Networks (EVNets), a new class of hybrid CNNs that combine the VOneBlock with a novel SubcorticalBlock, whose architecture draws from computational models in neuroscience and is parameterized to maximize alignment with subcortical responses reported across multiple experimental studies. Without being optimized to do so, the assembly of the SubcorticalBlock with the VOneBlock improved V1 alignment across most standard V1 benchmarks, and better modeled extra-classical receptive field phenomena. In addition, EVNets exhibit stronger emergent shape bias and overperform the base CNN architecture by 8.5\% on an aggregate benchmark of robustness evaluations, including adversarial perturbations, common corruptions, and domain shifts. Finally, we show that EVNets can be further improved when paired with a state-of-the-art data augmentation technique, surpassing the performance of the isolated data augmentation approach by 7.3\% on our robustness benchmark. This result reveals complementary benefits between changes in architecture to better mimic biology and training-based machine learning approaches.



Paperid:4829
Authors:Junkun Chen, Aayush Bansal, Minh Vo, Yu-Xiong Wang
Abstract:
This paper proposes Virtual Fitting Room (VFR), a novel video generative model that produces arbitrarily long virtual try-on videos. Our VFR models long video generation tasks as an auto-regressive, segment-by-segment generation process, eliminating the need for resource-intensive generation and lengthy video data, while providing the flexibility to generate videos of arbitrary length. The key challenges of this task are twofold: ensuring local smoothness between adjacent segments and maintaining global temporal consistency across different segments. To address these challenges, we propose our VFR framework, which ensures smoothness through a prefix video condition and enforces consistency with the anchor video — a 360°-view video that comprehensively captures the human's whole-body appearance. Our VFR generates minute-scale virtual try-on videos with both local smoothness and global temporal consistency under various motions, making it a pioneering work in long virtual try-on video generation.



Paperid:4830
Authors:Yajuan Peng, Yi Zhao, Nguyen Cam-Tu, Zuchao Li, Xiaoliang Wang, Hai Zhao, Xiaoming Fu
Abstract:
KV cache eviction has emerged as an effective solution to alleviate resource constraints faced by LLMs in long-context scenarios. However, existing token-level eviction methods often overlook two critical aspects: (1) their irreversible eviction strategy fails to adapt to dynamic attention patterns during decoding (the saliency shift problem), and (2) they treat both marginally important tokens and truly unimportant tokens uniformly, despite the collective significance of marginal tokens to model performance (the marginal information over-compression problem). To address these issues, we design two compensation mechanisms based on the high similarity of attention matrices between LLMs with different scales. We propose SmallKV, a small model assisted compensation method for KV cache compression. SmallKV can maintain attention matching between different-scale LLMs to: 1) assist the larger model in perceiving globally important information of attention; and 2) use the smaller model’s attention scores to approximate those of marginal tokens in the larger model. Extensive experiments on benchmarks including GSM8K, BBH, MT-Bench, and LongBench demonstrate the effectiveness of SmallKV. Moreover, efficiency evaluations show that SmallKV achieves 1.75 - 2.56 times higher throughput than baseline methods, highlighting its potential for efficient and performant LLM inference in resource constrained environments.



Paperid:4831
Authors:Wentao Lu, Dong Nie, Pengcheng Xue, Zheng Cui, Piji Li, Daoqiang Zhang, Xuyun Wen
Abstract:
Decoding natural language text from non-invasive brain signals, such as functional magnetic resonance imaging (fMRI), remains a central challenge in brain-computer interface research. While recent advances in large language models (LLMs) have enabled open-vocabulary fMRI-to-text decoding, existing frameworks typically process the entire fMRI sequence in a single step, leading to performance degradation when handling long input sequences due to memory overload and semantic drift. To address this limitation, we propose a brain-inspired sequential fMRI-to-text decoding framework that mimics the human cognitive strategy of segmented and inductive language processing. Specifically, we divide long fMRI time series into consecutive segments aligned with optimal language comprehension length. Each segment is decoded incrementally, followed by a wrap-up mechanism that summarizes the semantic content and incorporates it as prior knowledge into subsequent decoding steps. This sequence-wise approach alleviates memory burden and ensures semantic continuity across segments. In addition, we introduce a text-guided masking strategy integrated with a masked autoencoder (MAE) framework for fMRI representation learning. This method leverages attention distributions over key semantic tokens to selectively mask the corresponding fMRI time points, and employs MAE to guide the model toward focusing on neural activity at semantically salient moments, thereby enhancing the capability of fMRI embeddings to represent textual information. Experimental results on the Narratives dataset demonstrate that our method significantly outperforms state-of-the-art approaches, with performance gains increasing as decoding length grows.



Paperid:4832
Authors:Qijia He, Minghan Wang, Xutong Liu, Zhiyong Wang, Fang Kong
Abstract:
The multi-armed bandit (MAB) is a fundamental online decision-making framework that has been extensively studied over the past two decades. To mitigate the high cost and slow convergence of purely online learning, modern MAB approaches have exploredhybridparadigms that leverage offline data to warm-start online learning. However, existing approaches face a significant limitation by assuming that the offline and online data are homogeneous—they share the same feedback structure and are drawn from the same underlying distribution. This assumption is often violated in practice, where offline data often originate from diverse sources and evolving environments, resulting in feedback heterogeneity and distributional shifts. In this work, we tackle the challenge of learning across this offline-online gap by developing a general hybrid bandit framework that incorporates heterogeneous offline data to improve online performance. We study two hybrid settings: (1) using reward-based offline data to accelerate online learning in preference-based bandits (i.e., dueling bandits), and (2) using preference-based offline data to improve online standard MAB algorithms. For both settings, we design novel algorithms and derive tight regret bounds that match or improve upon existing benchmarks despite heterogeneity. Empirical evaluations on both synthetic and real-world datasets show that our proposed methods significantly outperform baseline algorithms.



Paperid:4833
Authors:Sicheng Shen, Dongcheng Zhao, Linghao Feng, Zeyang Yue, Jindong Li, Tenglong Li, Guobin Shen, Yi Zeng
Abstract:
Spiking Transformers have recently emerged as promising architectures for combining the efficiency of spiking neural networks with the representational power of self-attention. However, the lack of standardized implementations, evaluation pipelines, and consistent design choices has hindered fair comparison and principled analysis. In this paper, we introduce \textbf{STEP}, a unified benchmark framework for Spiking Transformers that supports a wide range of tasks, including classification, segmentation, and detection across static, event-based, and sequential datasets. STEP provides modular support for diverse components such as spiking neurons, input encodings, surrogate gradients, and multiple backends (e.g., SpikingJelly, BrainCog). Using STEP, we reproduce and evaluate several representative models, and conduct systematic ablation studies on attention design, neuron types, encoding schemes, and temporal modeling capabilities. We also propose a unified analytical model for energy estimation, accounting for spike sparsity, bitwidth, and memory access, and show that quantized ANNs may offer comparable or better energy efficiency. Our results suggest that current Spiking Transformers rely heavily on convolutional frontends and lack strong temporal modeling, underscoring the need for spike-native architectural innovations. The full code is available at: https://github.com/Fancyssc/STEP.



Paperid:4834
Authors:Pritam Sarkar, Ali Etemad
Abstract:
Despite recent advances in Large Video Language Models (LVLMs), they still struggle with fine-grained temporal understanding, hallucinate, and often make simple mistakes on even simple video question-answering tasks, all of which pose significant challenges to their safe and reliable deployment in real-world applications. To address these limitations, we propose a self-alignment framework that enables LVLMs to learn from their own errors. Our proposed framework first obtains a training set of preferred and non-preferred response pairs, where non-preferred responses are generated by incorporating common error patterns that often occur due to inadequate spatio-temporal understanding, spurious correlations between co-occurring concepts, and over-reliance on linguistic cues while neglecting the vision modality, among others. To facilitate self-alignment of LVLMs with the constructed preferred and non-preferred response pairs, we introduce Refined Regularized Preference Optimization (RRPO), a novel preference optimization method that utilizes sub-sequence-level refined rewards and token-wise KL regularization to address the limitations of Direct Preference Optimization (DPO). We demonstrate that RRPO achieves more precise alignment and more stable training compared to DPO. Our experiments and analysis validate the effectiveness of our approach across diverse video tasks, including video hallucination, short- and long-video understanding, and fine-grained temporal reasoning.



Paperid:4835
Authors:Joy Hsu, Emily Jin, Jiajun Wu, Niloy Mitra
Abstract:
Real-world scenes, such as those in ScanNet, are difficult to capture, with highly limited data available. Generating realistic scenes with varied object poses remains an open and challenging task. In this work, we propose FactoredScenes, a framework that synthesizes realistic 3D scenes, by leveraging the underlying structure of rooms, while learning the variation of object poses from lived-in scenes. We propose a factored room representation that decomposes scenes into hierarchically organized concepts of programs and object poses. To encode structure, FactoredScenes learns a library of functions capturing reusable layout patterns from which scenes are drawn, then uses large language models to generate high-level programs, regularized by the learned library. To represent scene variations, FactoredScenes learns a program-conditioned model to hierarchically predict object poses, and retrieves and places 3D objects in a scene. We show that FactoredScenes generates realistic, real-world rooms that are difficult to distinguish from real ScanNet scenes.



Paperid:4836
Authors:Qichao Cao, Shangzhi Zeng, Jin Zhang
Abstract:
Bilevel optimization has garnered significant attention in the machine learning community recently, particularly regarding the development of efficient numerical methods. While substantial progress has been made in developing efficient algorithms for optimistic bilevel optimization, the study of methods for solving Pessimistic Bilevel Optimization (PBO) remains relatively less explored, especially the design of fully first-order, single-loop gradient-based algorithms. This paper aims to bridge this research gap. We first propose a novel smooth approximation to the PBO problem, using penalization and regularization techniques. Building upon this approximation, we then propose SiPBA (Single-loop Pessimistic Bilevel Algorithm), a new gradient-based method specifically designed for PBO which avoids second-order derivative information or inner-loop iterations for subproblem solving. We provide theoretical validation for the proposed smooth approximation scheme and establish theoretical convergence for the algorithm SiPBA. Numerical experiments on synthetic examples and practical applications demonstrate the effectiveness and efficiency of SiPBA.



Paperid:4837
Authors:Yuan Feng, Junlin Lv, Yukun Cao, Xike Xie, S. Kevin Zhou
Abstract:
Large Language Models have excelled in various domains but face efficiency challenges due to the growing Key-Value (KV) cache required for long-sequence inference. Recent efforts aim to reduce KV cache size by evicting vast non-critical cache elements during runtime while preserving generation quality. However, these methods typically allocate compression budgets uniformly across all attention heads, ignoring the unique attention patterns of each head. In this paper, we establish a theoretical loss upper bound between pre- and post-eviction attention output, explaining the optimization target of prior cache eviction methods, while guiding the optimization of adaptive budget allocation. Base on this, we propose {\it Ada-KV}, the first head-wise adaptive budget allocation strategy. It offers plug-and-play benefits, enabling seamless integration with prior cache eviction methods. Extensive evaluations on 13 datasets from Ruler and 16 datasets from LongBench, all conducted under both question-aware and question-agnostic scenarios, demonstrate substantial quality improvements over existing methods.



Paperid:4838
Authors:Tsai Hor Chan, Feng Wu, Yihang Chen, Guosheng Yin, Lequan Yu
Abstract:
Developing effective multimodal fusion approaches has become increasingly essential in many real-world scenarios, such as health care and finance. The key challenge is how to preserve the feature expressiveness in each modality while learning cross-modal interactions.Previous approaches primarily focus on the cross-modal alignment,while over-emphasis on the alignment of marginal distributions of modalities may impose excess regularization and obstruct meaningful representations within each modality.The Dirichlet process (DP) mixture model is a powerful Bayesian non-parametric method that can amplify the most prominent features by its richer-gets-richer property, which allocates increasing weights to them.Inspired by this unique characteristic of DP, we propose a new DP-driven multimodal learning framework that automatically achieves an optimal balance between prominent intra-modal representation learning and cross-modal alignment. Specifically, we assume that each modality follows a mixture of multivariate Gaussian distributions and further adopt DP to calculate the mixture weights for all the components. This paradigm allows DP to dynamically allocate the contributions of features and select the most prominent ones, leveraging its richer-gets-richer property, thus facilitating multimodal feature fusion.Extensive experiments on several multimodal datasets demonstrate the superior performance of our model over other competitors.Ablation analysis further validates the effectiveness of DP in aligning modality distributions and its robustness to changes in key hyperparameters.Code is anonymously available at https://anonymous.4open.science/r/DPMM-F15D



Paperid:4839
Authors:Simin Li, Zihao Mao, Hanxiao Li, Zonglei Jing, Zhuohang bian, Jun Guo, Li Wang, Zhuoran Han, Ruixiao Xu, Xin Yu, Chengdong Ma, Yuqing Ma, Bo An, Yaodong Yang, Weifeng Lv, Xianglong Liu
Abstract:
In cooperative Multi-Agent Reinforcement Learning (MARL), it is a common practice to tune hyperparameters in ideal simulated environments to maximize cooperative performance. However, policies tuned for cooperation often fail to maintain robustness and resilience under real-world uncertainties. Building trustworthy MARL systems requires a deep understanding of \emph{robustness}, which ensures stability under uncertainties, and \emph{resilience}, the ability to recover from disruptions—a concept extensively studied in control systems but largely overlooked in MARL. In this paper, we present a large-scale empirical study comprising over 82,620 experiments to evaluate cooperation, robustness, and resilience in MARL across 4 real-world environments, 13 uncertainty types, and 15 hyperparameters. Our key findings are: (1) Under mild uncertainty, optimizing cooperation improves robustness and resilience, but this link weakens as perturbations intensify. Robustness and resilience also varies by algorithm and uncertainty type. (2) Robustness and resilience do not generalize across uncertainty modalities or agent scopes: policies robust to action noise for all agents may fail under observation noise on a single agent. (3) Hyperparameter tuning is critical for trustworthy MARL: surprisingly, standard practices like parameter sharing, GAE, and PopArt can hurt robustness, while early stopping, high critic learning rates, and Leaky ReLU consistently help. By optimizing hyperparameters only, we observe substantial improvement in cooperation, robustness and resilience across all MARL backbones, with the phenomenon also generalizing to robust MARL methods across these backbones.



Paperid:4840
Authors:Songqi Zhou, Zeyuan Liu, Benben Jiang
Abstract:
Ensuring fairness in machine learning models is a critical challenge. Existing debiasing methods often compromise performance, rely on static correction strategies, and struggle with data sparsity, particularly within minority groups. Furthermore, their utilization of sensitive attributes is often suboptimal, either depending excessively on complete attribute labeling or disregarding these attributes entirely. To overcome these limitations, we propose FairNet, a novel framework for dynamic, instance-level fairness correction. FairNet integrates a bias detector with conditional low-rank adaptation (LoRA), which enables selective activation of the fairness correction mechanism exclusively for instances identified as biased, and thereby preserve performance on unbiased instances. A key contribution is a new contrastive loss function for training the LoRA module, specifically designed to minimize intra-class representation disparities across different sensitive groups and effectively address underfitting in minority groups. The FairNet framework can flexibly handle scenarios with complete, partial, or entirely absent sensitive attribute labels. Theoretical analysis confirms that, under moderate TPR/FPR for the bias detector, FairNet can enhance the performance of the worst group without diminishing overall model performance, and potentially yield slight performance improvements. Comprehensive empirical evaluations across diverse vision and language benchmarks validate the effectiveness of FairNet.



Paperid:4841
Authors:Yu Qian, Wilson Geisler, Xue-Xin Wei
Abstract:
Previous studies have compared the brain and deep neural networks trained on image classification. Intriguingly, while some suggest that their representations are highly similar, others argued the opposite. Here, we propose a new approach to characterize the similarity of the decision strategies of two observers (models or brains) using decision variable correlation (DVC). DVC quantifies the correlation between decoded decisions on individual samples in a classification task and thus can capture task-relevant information rather than general representational alignment. We evaluate this method using monkey V4/IT recordings and models trained on image classification tasks.We find that model–model similarity is comparable to monkey-monkey similarity, whereas model–monkey similarity is consistently lower and, surprisingly, decreases with increasing ImageNet-1k performance. While adversarial training enhances robustness, it does not improve model–monkey similarity in task-relevant dimensions; however, it markedly increases model–model similarity. Similarly, pre-training on larger datasets does not improve model–monkey similarity. These results suggest a fundamental divergence between the task-relevant representations in monkey V4/IT and those learned by models trained on image classification tasks.



Paperid:4842
Authors:Xiaotang Wang, Xuanwei Lin, Yun Zhu, Hao Li, Yongqi Zhang
Abstract:
Abstract:Crossmodal matching in single-cell omics is essential for explaining biological regulatory mechanisms and enhancing downstream analyses. However, current single-cell crossmodal models often suffer from three limitations: sparse modality signals, underutilization of biological attributes, and insufficient modeling of regulatory interactions. In this work, we present a novel framework which reformulates crossmodal matching as a graph classification task on Attributed Bipartite Graphs (ABGs). It models single-cell ATAC-RNA data as an ABG, where each expressed ATAC and RNA is treated as a distinct node with unique IDs and biological features. To model crossmodal interaction patterns on the constructed ABG, we propose $\text{Bi}^2\text{Former}$, a biologically-driven bipartite graph transformer that learns interpretable attention over ATAC–RNA pairs. This design enables the model to effectively learn and explain biological regulatory relationships between ATAC and RNA modalities.Extensive experiments demonstrate that $\text{Bi}^2\text{Former}$ achieves state-of-the-art performance in crossmodal matching across diverse datasets, generalizes well to unseen cell types, and reveals biologically meaningful regulatory patterns. This work pioneers an ABG-based approach for single-cell crossmodal matching, offering a powerful framework for uncovering regulatory interactions at the single-cell omics. Our code is available at: https://anonymous.4open.science/r/Bi2Former.



Paperid:4843
Authors:Noémie Bergues, Arthur Carré, Paul Join-Lambert, Brice Hoffmann, Arnaud Blondel, Hamza Tajmouati
Abstract:
Predicting the 3D conformation of small molecules within protein binding sites is a key challenge in drug design. When a crystallized reference ligand (template) is available, it provides geometric priors that can guide 3D pose prediction. We present a two-stage method for ligand conformation generation guided by such templates. In the first stage, we introduce a molecular alignment approach based on flow-matching to generate 3D coordinates for the ligand, using the template structure as a reference. In the second stage, a differentiable pose optimization procedure refines this conformation based on shape and pharmacophore similarities, internal energy, and, optionally, the protein binding pocket. We evaluate our approach on a new benchmark of ligand pairs co-crystallized with the same target and show that it outperforms standard docking tools and open-access alignment methods, especially in cases involving low similarity to the template or high ligand flexibility.



Paperid:4844
Authors:Jhanvi Garg, Krishnakumar Balasubramanian, Quan Zhou
Abstract:
Abstract:Simulated tempering is a widely used strategy for sampling from multimodal distributions. In this paper, we consider simulated tempering combined with an arbitrary local Markov chain Monte Carlo sampler and present a new decomposition theorem that provides a lower bound on the restricted spectral gap of the algorithm for sampling from mixture distributions. By working with the restricted spectral gap, the applicability of our results is extended to broader settings such as when the usual spectral gap is difficult to bound or becomes degenerate. We demonstrate the application of our theoretical results by analyzing simulated tempering combined with random walk Metropolis--Hastings for sampling from mixtures of Gaussian distributions. We show that in fixed-dimensional settings, the algorithm's complexity scales polynomially with the separation between modes and logarithmically with $1/\varepsilon$, where $\varepsilon$ is the target accuracy in total variation distance.



Paperid:4845
Authors:Woojin Kim, Jaeyoung Do
Abstract:
Classifier guidance is a widely adopted technique in diffusion language models, used to steer generation toward desired attributes. However, such guidance often introduces instability during the generation process, where token-level updates fluctuate across timesteps. We identify and formally characterize this phenomenon as update-forgetting. This instability disrupts the refinement process by overwriting semantic edits, ultimately degrading fluency and coherence, which is particularly problematic in tasks like controllable text generation. To address this, we propose TTA-Diffusion, a novel inference-time approach that dynamically allocates timesteps per token based on refinement needs. Unlike conventional diffusion models that apply uniform updates, TTA-Diffusion employs structured timestep allocation, preserving stable tokens while allowing uncertain tokens to undergo progressive adjustment. Experimental results across diverse tasks demonstrate that TTA-Diffusion significantly outperforms both diffusion-based and auto-regressive baselines in fluency and control accuracy while improving computational efficiency by reducing the number of required timesteps. On the sentiment control task, TTA-Diffusion achieves over 20\% higher accuracy and nearly half the perplexity of prior diffusion models, using less than one-fifth the denoising steps. This work highlights the importance of mitigating fluctuations in token updates and promoting a balanced refinement process, thereby enhancing stability and controllability in controllable language modeling.



Paperid:4846
Authors:Zheyu Zhang, Ziqi Pang, Shixing Chen, Xiang Hao, Vimal Bhat, Yu-Xiong Wang
Abstract:
Long video understanding is inherently challenging for vision-language models (VLMs) because of the extensive number of frames. With each video frame typically expanding into tens or hundreds of tokens, the limited context length of large language models (LLMs) forces the VLMs to perceive the frames sparsely and lose temporal information. To address this, we explore extreme video token compression towardsone token per frameat the final LLM layer. Our key insight is that heuristic-based compression, widely adopted by previous methods, is prone to information loss, and this necessitates supervising LLM layers intolearnableandprogressivemodules fortoken-level compression(LP-Comp). Such compression enables our VLM to digest 2x-4x more frames with improved performance. To further increase the token efficiency, we investigate \emph{frame-level compression}, which selects the frames most relevant to the queries via the internal attention scores of the LLM layers, namedquestion-conditioned compression(QC-Comp). As a notable distinction from previous studies, we mitigate the position bias of LLM attention in long contexts,i.e., the over-concentration on the beginning and end of a sequence, by splitting long videos into short segments and employing local attention. Collectively, our combinedtoken-levelandframe-levelleads to an extreme compression model for long video understanding, namedXComp, achieving a significantly larger compression ratio and enabling denser frame sampling. Our XComp is finetuned from VideoChat-Flash with a data-efficientsupervised compression tuningstage that only requires 2.5\% of the supervised fine-tuning data, yet boosts the accuracy from 42.9\% to 46.2\% on LVBench and enhances multiple other long video benchmarks.



Paperid:4847
Authors:Xinbiao Wang, Yuxuan Du, Zihan Lou, Yang Qian, Kaining Zhang, Yong Luo, Bo Du, Dacheng Tao
Abstract:
Abstract:Quantum many-body problems are central to various scientific disciplines, yet their ground-state properties are intrinsically challenging to estimate. Recent advances in deep learning (DL) offer potential solutions in this field, complementing prior purely classical and quantum approaches. However, existing DL-based models typically assume access to a large-scale and noiseless labeled dataset collected by infinite sampling. This idealization raises fundamental concerns about their practical utility, especially given the limited availability of quantum hardware in the near term. To unleash the power of these DL-based models, we propose AiDE-Q (\underline{a}utomat\underline{i}c \underline{d}ata \underline{e}ngine for \underline{q}uantum property estimation), an effective framework that addresses this challenge by iteratively generating high-quality synthetic labeled datasets. Specifically, AiDE-Q utilizes a confidence-check method to assess the quality of synthetic labels and continuously improves the employed DL models with the identified high-quality synthetic dataset. To verify the effectiveness of AiDE-Q, we conduct extensive numerical simulations on a diverse set of quantum many-body and molecular systems, with up to 50 qubits. The results show that AiDE-Q enhances prediction performance for various reference learning models, with improvements of up to $14.2\\%$. Moreover, we exhibit that a basic supervised learning model integrated with AiDE-Q outperforms advanced reference models, highlighting the importance of a synthetic dataset. Our work paves the way for more efficient and practical applications of DL for quantum property estimation.



Paperid:4848
Authors:Meilong Xu, Xiaoling Hu, Shahira Abousamra, Chen Li, Chao Chen
Abstract:
In semi-supervised segmentation, capturing meaningful semantic structures from unlabeled data is essential. This is particularly challenging in histopathology image analysis where objects are densely distributed. To address this issue, we propose a semi-supervised segmentation framework designed to robustly identify and preserve relevant topological features. Our method leverages multiple perturbed predictions obtained through stochastic dropouts and temporal training snapshots, enforcing topological consistency across these varied outputs. This consistency mechanism helps distinguish biologically meaningful structures from transient and noisy artifacts. A key challenge in this process is to accurately match the corresponding topological features across the predictions in the absence of ground truth. To overcome this, we introduce a novel matching strategy that integrates spatial overlap with global structural alignment, minimizing discrepancies among predictions. Extensive experiments demonstrate that our approach effectively reduces topological errors, resulting in more robust and accurate segmentations essential for reliable downstream analysis.



Paperid:4849
Authors:Zheng CHEN, Yushi Feng, Jisheng Dang, Changyang He, Yue Deng, Hongxi Pu, Haoxuan Li, Bo Li
Abstract:
Large Language Models (LLMs) have attained human-level fluency in text generation, which complicates the distinguishing between human-written and LLM generated texts. This increases the risk of misuse and highlights the need for reliable detectors. Yet, existing detectors exhibit poor robustness on out-of-distribution (OOD) data and attacked data, which is critical for real-world scenarios. Also, they struggle to provide interpretable evidence to support their decisions, thus undermining reliability. In light of these challenges, we propose IPAD (Inverse Prompt for AI Detection), a novel framework consisting of a Prompt Inverter that identifies predicted prompts that could have generated the input text, and two Distinguishers that examine the probability that the input texts align with the predicted prompts. Empirical evaluations demonstrate that IPAD outperforms the strongest baselines by 9.05% (Average Recall) on in-distribution data, 12.93% (AUROC) on out-of-distribution (OOD) data, and 5.48% (AUROC) on attacked data. IPAD also performs robust on structured datasets. Furthermore, an interpretability assessment is conducted to illustrate that IPAD enhances the AI detection trustworthiness by allowing users to directly examine the decision-making evidence, which provides interpretable support for its state-of-the-art detection results.



Paperid:4850
Authors:Haoyi Song, Ruihan Ji, Naichen Shi, Fan Lai, Raed AL Kontar
Abstract:
Large language models (LLMs) have transformed natural language processing, but their reliable deployment requires effective uncertainty quantification (UQ). Existing UQ methods are often heuristic and lack a fully probabilistic foundation. This paper begins by providing a theoretical justification for the role of perturbations in UQ for LLMs. We then introduce a dual random walk perspective, modeling input–output pairs as two Markov chains with transition probabilities defined by semantic similarity. Building on this, we propose a fully probabilistic framework based on an inverse model, which quantifies uncertainty by evaluating the diversity of the input space conditioned on a given output through systematic perturbations. Within this framework, we define a new uncertainty measure, Inv-Entropy. A key strength of our framework is its flexibility: it supports various definitions of uncertainty measures, embeddings, perturbation strategies, and similarity metrics. We also propose GAAP, a perturbation algorithm based on genetic algorithms, which enhances the diversity of sampled inputs. In addition, we introduce a new evaluation metric, Temperature Sensitivity of Uncertainty (TSU), which directly assesses uncertainty without relying on correctness as a proxy. Extensive experiments demonstrate that Inv-Entropy outperforms existing semantic UQ methods.



Paperid:4851
Authors:Ruifeng Luo, Zhengjie Liu, Tianxiao Cheng, Jie Wang, Tongjie Wang, Fei Cheng, Fu Chai, Yanpeng Li, Xingguang Wei, Haomin Wang, Shenglong Ye, Wenhai Wang, Zhang, Yu Qiao, Hongjie Zhang, Xianzhong Zhao
Abstract:
Recognizing symbols in architectural CAD drawings is critical for various advanced engineering applications. In this paper, we propose a novel CAD data annotation engine that leverages intrinsic attributes from systematically archived CAD drawings to automatically generate high-quality annotations, thus significantly reducing manual labeling efforts. Utilizing this engine, we construct ArchCAD-400K, a large-scale CAD dataset consisting of 413,062 chunks from 5538 highly standardized drawings, making it over 26 times larger than the largest existing CAD dataset. ArchCAD-400K boasts an extended drawing diversity and broader categories, offering line-grained annotations. Furthermore, we present a new baseline model for panoptic symbol spotting, termed Dual-Pathway Symbol Spotter (DPSS). It incorporates an adaptive fusion module to enhance primitive features with complementary image features, achieving state-of-the-art performance and enhanced robustness. Extensive experiments validate the effectiveness of DPSS, demonstrating the value of ArchCAD-400K and its potential to drive innovation in architectural design and construction.



Paperid:4852
Authors:Touqeer Ahmad, Mohammadreza Mousavi Kalan, François Portier, Gilles Stupfler
Abstract:
Abstract:Oversampling synthetic minority examples using $\textsc{Smote}$ and its variants is a leading strategy for addressing imbalanced classification problems. Despite the success of this approach in practice, its theoretical foundations remain underexplored. We develop a theoretical framework to analyze the behavior of $\textsc{Smote}$ and related methods when classifiers are trained on synthetic data. First, we establish an exponential inequality that characterizes the gap between the empirical risk computed on synthetic samples and the true population risk on the minority class. Second, we show that a kernel-based classification rule trained on synthetic data can achieve the minimax rate of convergence. This leads to practical guidelines for better parameter tuning of both $\textsc{Smote}$ and the downstream learning algorithm. Numerical experiments are provided to illustrate and support the theoretical findings.



Paperid:4853
Authors:Kijung Jeon, Michael Muehlebach, Molei Tao
Abstract:
Abstract:Sampling from constrained statistical distributions is a fundamental task in various fields including Bayesian statistics, computational chemistry, and statistical physics.This article considers the cases where the constrained distribution is described by an unconstrained density, as well as additional equality and/or inequality constraints, which often make the constraint set nonconvex. Existing methods for nonconvex constraint set $\Sigma \subset \mathbb{R}^d$ defined by equality or inequality constraints commonly rely on costly projection steps. Moreover, they cannot handle equality and inequality constraints simultaneously as each method only specialized in one case. In addition, rigorous and quantitative convergence guarantee is often lacking.In this paper, we introduce Overdamped Langevin with LAnding (OLLA), a new framework that can design overdamped Langevin dynamics accommodating both equality and inequality constraints. The proposed dynamics also deterministically corrects trajectories along the normal direction of the constraint surface, thus obviating the need for explicit projections. We show that, under suitable regularity conditions on the target density and $\Sigma$, OLLA converges exponentially fast in $W_2$ distance to the constrained target density $\rho_\Sigma(x) \propto \exp(-f(x))d\sigma_\Sigma$.Lastly, through experiments, we demonstrate the efficiency of OLLA compared to projection-based constrained Langevin algorithms and their slack variable variants, highlighting its favorable computational cost and reasonable empirical mixing.



Paperid:4854
Authors:Bing He, Yunuo Chen, Guo Lu, Qi Wang, Qunshan Gu, Rong Xie, Li Song, Wenjun Zhang
Abstract:
Dynamic scene reconstruction poses a persistent challenge in 3D vision. Deformable 3D Gaussian Splatting has emerged as an effective method for this task, offering real-time rendering and high visual fidelity.This approach decomposes a dynamic scene into a static representation in a canonical space and time-varying scene motion.Scene motion is defined as the collective movement of all Gaussian points, and for compactness, existing approaches commonly adopt implicit neural fields or sparse control points. However, these methods predominantly rely on gradient-based optimization for all motion information. Due to the high degree of freedom, they struggle to converge on real-world datasets exhibiting complex motion.To preserve the compactness of motion representation and address convergence challenges, this paper proposes heterogeneous 3D control points, termed \textbf{H3D control points}, whose attributes are obtained using a hybrid strategy combining optical flow back-projection and gradient-based methods. This design decouples directly observable motion components from those that are geometrically occluded.Specifically, components of 3D motion that project onto the image plane are directly acquired via optical flow back projection, while unobservable portions are refined through gradient-based optimization.Experiments on the Neu3DV and CMU-Panoptic datasets demonstrate that our method achieves superior performance over state-of-the-art 4D Gaussian splatting techniques. Remarkably, our method converges within just 100 iterations and achieves a per-frame processing speed of 2 seconds on a single NVIDIA RTX 4070 GPU.



Paperid:4855
Authors:Aladin Djuhera, Swanand Kadhe, Syed Zawad, Farhan Ahmed, Heiko Ludwig, Holger Boche
Abstract:
Recent work on large language models (LLMs) has increasingly focused on post-training and alignment with datasets curated to enhance instruction following, world knowledge, and specialized skills. However, post-training datasets used by both open- and closed-source LLMs remain inaccessible to the public, with limited information about their construction process. This lack of transparency has motivated the development of open-source post-training corpora. While training on these datasets can yield performance comparable to that of leading models, their large size and complexity make systematic comparisons challenging, particularly under common compute and memory constraints. As a result, it is difficult to attribute performance gains to individual samples, task categories, or curation strategies when assessing data quality. In this work, we conduct a comprehensive analysis of two widely used post-training datasets, Tulu-3-SFT-Mix and SmolTalk, to reveal their structural and qualitative similarities and differences. Using the MagPie framework, we annotate every sample with detailed quality metrics, including turn structure (single-turn vs. multi-turn), task category, input quality, and response quality. Through controlled experiments, we investigate three key questions: (1) How many high-quality examples are needed to achieve stable performance gains? (2) What is the marginal utility of multi-turn dialogues compared to single-turn prompts? (3) How do different data mixture ratios influence model outcomes? Our findings offer actionable insights for designing effective post-training datasets that improve model performance within practical resource limits. To support future research, we publicly release our annotated versions of both datasets.



Paperid:4856
Authors:Boheng Li, Renjie Gu, Junjie Wang, Leyi Qi, Yiming Li, Run Wang, Zhan Qin, Tianwei Zhang
Abstract:
Text-to-image (T2I) diffusion models have achieved impressive image generation quality and are increasingly fine-tuned for personalized applications. However, these models often inherit unsafe behaviors from toxic pretraining data, raising growing safety concerns. While recent safety-driven unlearning methods have made promising progress in suppressing model toxicity, they are identified to be fragile to downstream fine-tuning, where we reveal that state-of-the-art methods largely fail to retain their effectiveness even when fine-tuned on entirely benign datasets. To mitigate this problem, in this paper, we propose ResAlign, a safety-driven unlearning framework with enhanced resilience against downstream fine-tuning. By modeling downstream fine-tuning as an implicit optimization problem with a Moreau Envelope-based reformulation, ResAlign enables efficient gradient estimation to minimize the recovery of harmful behaviors. Additionally, a meta-learning strategy is proposed to simulate a diverse distribution of fine-tuning scenarios to improve generalization. Extensive experiments across a wide range of datasets, fine-tuning methods, and configurations demonstrate that ResAlign consistently outperforms prior unlearning approaches in retaining safety after downstream fine-tuning while preserving benign generation capability well.



Paperid:4857
Authors:Thomas Decker, Florian Buettner, Volker Tresp
Abstract:
Perturbation-based explanations are widely utilized to enhance the transparency of machine-learning models in practice. However, their reliability is often compromised by the unknown model behavior under the specific perturbations used. This paper investigates the relationship between uncertainty calibration - the alignment of model confidence with actual accuracy - and perturbation-based explanations. We show that models systematically produce unreliable probability estimates when subjected to explainability-specific perturbations and theoretically prove that this directly undermines global and local explanation quality. To address this, we introduce ReCalX, a novel approach to recalibrate models for improved explanations while preserving their original predictions. Empirical evaluations across diverse models and datasets demonstrate that ReCalX consistently reduces perturbation-specific miscalibration most effectively while enhancing explanation robustness and the identification of globally important input features.



Paperid:4858
Authors:Alexander Pluska, Sagar Malhotra
Abstract:
Local convergence has emerged as a fundamental tool for analyzing sparse random graph models. We introduce a new notion of local convergence,color convergence, based on the Weisfeiler–Leman algorithm. Color convergence fully characterizes the class of random graphs that are well-behaved in the limit for message-passing graph neural networks. Building on this, we propose theRefined Configuration Model(RCM), a random graph model that generalizes the configuration model. The RCM is universal with respect to local convergence among locally tree-like random graph models, including Erdős–Rényi, stochastic block and configuration models. Finally, this framework enables a complete characterization of the random trees that arise as local limits of such graphs.



Paperid:4859
Authors:Kyeongwon Lee, Lizhen Lin, Jaewoo Park, Seonghyun Jeong
Abstract:
This work establishes that sparse Bayesian neural networks achieve optimal posterior contraction rates over anisotropic Besov spaces and their hierarchical compositions. These structures reflect the intrinsic dimensionality of the underlying function, thereby mitigating the curse of dimensionality. Our analysis shows that Bayesian neural networks equipped with either sparse or continuous shrinkage priors attain the optimal rates which are dependent on the intrinsic dimension of the true structures. Moreover, we show that these priors enable rate adaptation, allowing the posterior to contract at the optimal rate even when the smoothness level of the true function is unknown. The proposed framework accommodates a broad class of functions, including additive and multiplicative Besov functions as special cases. These results advance the theoretical foundations of Bayesian neural networks and provide rigorous justification for their practical effectiveness in high-dimensional, structured estimation problems.



Paperid:4860
Authors:Akshara Prabhakar, Zuxin Liu, Ming Zhu, Jianguo Zhang, Tulika Manoj Awalgaonkar, Shiyu Wang, Zhiwei Liu, Haolin Chen, Thai Hoang, Juan Carlos Niebles, Shelby Heinecke, Weiran Yao, Huan Wang, Silvio Savarese, Caiming Xiong
Abstract:
Abstract:Training effective AI agents for multi-turn interactions requires high-quality data that captures realistic human-agent dynamics, yet such data is scarce and expensive to collect manually. We introduce APIGen-MT, a two-phase framework that generates verifiable and diverse multi-turn agent data. In the first phase, our agentic pipeline produces detailed task blueprints with ground-truth actions, leveraging a committee of LLM reviewers and iterative feedback loops. These blueprints are then transformed into complete interaction trajectories through simulated human-agent interplay. We train a family of models---the xLAM-2-fc-r series with sizes ranging from 1B to 70B parameters. Our models outperform frontier models such as GPT-4o and Claude 3.5 on $\tau$-bench and BFCL benchmarks, with the smaller models surpassing their larger counterparts, particularly in multi-turn settings, while maintaining superior consistency across multiple trials. Comprehensive experiments demonstrate that our verified blueprint-to-details approach yields high-quality training data, enabling the development of more reliable, efficient, and capable agents. We open-source both the synthetic data collected and the trained xLAM-2-fc-r models to advance research in AI agents.Dataset: https://huggingface.co/datasets/Salesforce/APIGen-MT-5k & Models: https://huggingface.co/collections/Salesforce/xlam-2-67ef5be12949d8dcdae354c4



Paperid:4861
Authors:Zijie Xu, Tong Bu, Zecheng Hao, Jianhao Ding, Zhaofei Yu
Abstract:
Abstract:Spiking Neural Networks (SNNs) offer low-latency and energy-efficient decision making through neuromorphic hardware, making them compelling for Reinforcement Learning (RL) in resource-constrained edge devices. Recent studies in this field directly replace Artificial Neural Networks (ANNs) by SNNs in existing RL frameworks, overlooking whether the RL algorithm is suitable for SNNs. However, most RL algorithms in continuous control are designed tailored to ANNs—including the target network soft updates mechanism—which conflict with the discrete, non-differentiable dynamics of SNN spikes. We identify that this mismatch destabilizes SNN training in continuous control tasks. To bridge this gap between discrete SNN and continuous control, we propose a novel proxy target framework. The continuous and differentiable dynamics of the proxy target enable smooth updates, bypassing the incompatibility of SNN spikes, stabilizing the RL algorithms. Since the proxy network operates only during training, the SNN retains its energy efficiency during deployment without inference overhead. Extensive experiments on continuous control benchmarks demonstrate that compared to vanilla SNNs, the proxy target framework enables SNNs to achieve up to $32\%$ higher performance across different spiking neurons. Notably, we are the first to surpass ANN performance in continuous control with simple Leaky-Integrate-and-Fire (LIF) neurons. This work motivates a new class of SNN-friendly RL algorithms tailored to SNN's characteristics, paving the way for neuromorphic agents that combine high performance with low power consumption.



Paperid:4862
Authors:Jiashun Liu, Zihao Wu, Johan Obando Ceron, Pablo Samuel Castro, Aaron Courville, Ling Pan
Abstract:
Abstract:Deep reinforcement learning (RL) agents frequently suffer from neuronal activity loss, which impairs their ability to adapt to new data and learn continually. A common method to quantify and address this issue is the $\tau$-dormant neuron ratio, which uses activation statistics to measure the expressive ability of neurons. While effective for simple MLP-based agents, this approach loses statistical power in more complex architectures. To address this, we argue that in advanced RL agents, maintaining a neuron's **learning capacity**, its ability to adapt via gradient updates, is more critical than preserving its expressive ability. Based on this insight, we shift the statistical objective from activations to gradients, and introduce **GraMa** (**Gra**dient **Ma**gnitude Neural Activity Metric), a lightweight, architecture-agnostic metric for quantifying neuron-level learning capacity. We show that **GraMa** effectively reveals persistent neuron inactivity across diverse architectures, including residual networks, diffusion models, and agents with varied activation functions. Moreover, **re**setting neurons guided by **GraMa** (**ReGraMa**) consistently improves learning performance across multiple deep RL algorithms and benchmarks, such as MuJoCo and the DeepMind Control Suite. **We make our code available.**



Paperid:4863
Authors:Eya Cherif, Arthur Ouaknine, Luke Brown, Phuong Dao, Kyle Kovach, Bing Lu, Daniel Mederer, Hannes Feilhauer, Teja Kattenborn, David Rolnick
Abstract:
Plant traits such as leaf carbon content and leaf mass are essential variables in the study of biodiversity and climate change. However, conventional field sampling cannot feasibly cover trait variation at ecologically meaningful spatial scales. Machine learning represents a transformative solution for plant trait prediction across ecosystems, leveraging hyperspectral data from remote sensing.Nevertheless, trait prediction from hyperspectral data is challenged by label scarcity and substantial domain shifts (e.g. across sensors, ecological distributions), requiring robust cross-domain methods.Here, we present GreenHyperSpectra, a pretraining dataset encompassing real-world cross-sensor and cross-ecosystem samples designed to benchmark trait prediction with semi- and self-supervised methods. We adopt an evaluation framework encompassing in-distribution and out-of-distribution scenarios. We successfully leverage GreenHyperSpectra to pretrain label-efficient multi-output regression models that outperform the state-of-the-art supervised baseline. Our empirical analyses demonstrate substantial improvements in learning spectral representations for trait prediction, establishing a comprehensive methodological framework to catalyze research at the intersection of representation learning and plant functional traits assessment.



Paperid:4864
Authors:Vansh Sharma, Harish Ganesh, Maryam Akram, Wanjiao Liu, Venkat Raman
Abstract:
This study presents a new high-fidelity multi-modal dataset containing 16000+ geometric variants of automotive hoods useful for machine learning (ML) applications such as engineering component design and process optimization, and multiphysics system surrogates. The dataset is centered on a practical multiphysics problem—hood deformation from fluid entrapment and inertial loading during rotary‑dip painting. Each hood is numerically modeled with a coupled Large-Eddy Simulation (LES)-Finite Element Analysis (FEA), using 1.2M cells in total to ensure spatial and temporal accuracy. The dataset provides time-resolved physical fields, along with STL meshes and structured natural language prompts for text-to-geometry synthesis. Existing datasets are either confined to 2D cases, exhibit limited geometric variations, or lack the multi‑modal annotations and data structures—shortcomings we address with AutoHood3D. We validate our numerical methodology, establish quantitative baselines across five neural architectures, and demonstrate systematic surrogate errors in displacement and force predictions. These findings motivate the design of novel approaches and multiphysics loss functions that enforce fluid–solid coupling during model training. By providing fully reproducible workflows, AutoHood3D enables physics‑aware ML development, accelerates generative‑design iteration, and facilitates the creation of new FSI benchmarks.



Paperid:4865
Authors:Xianhan Tan, Binli Luo, Yu Qi, Yueming Wang
Abstract:
Brain-computer interfaces have shown great potential in motor and speech rehabilitation, but still suffer from low performance stability across days, mostly due to the instabilities in neural signals. These instabilities, partially caused by neuron deaths and electrode shifts, leading to channel-level variabilities among different recording days. Previous studies mostly focused on aligning multi-day neural signals of onto a low-dimensional latent manifold to reduce the variabilities, while faced with difficulties when neural signals exhibit significant drift. Here, we propose to learn a channel-level invariant neural representation to address the variabilities in channels across days. It contains a channel-rearrangement module to learn stable representations against electrode shifts, and a channel reconstruction module to handle the missing neurons. The proposed method achieved the state-of-the-art performance with cross-day decoding tasks over two months, on multiple benchmark BCI datasets. The proposed approach showed good generalization ability that can be incorporated to different neural networks.



Paperid:4866
Authors:Sifan Wang, Ananyae bhartari, Bowen Li, Paris Perdikaris
Abstract:
Physics-informed neural networks (PINNs) have shown significant promise in computational science and engineering, yet they often face optimization challenges and limited accuracy. In this work, we identify directional gradient conflicts during PINN training as a critical bottleneck. We introduce a novel gradient alignment score to systematically diagnose this issue through both theoretical analysis and empirical experiments.Building on these insights, we show that (quasi) second-order optimization methods inherently mitigate gradient conflicts, thereby consistently outperforming the widely used Adam optimizer. Among them, we highlight the effectiveness of SOAP \cite{vyas2024soap} by establishing its connection to Newton’s method.Empirically, SOAP achieves state-of-the-art results on 10 challenging PDE benchmarks, including the first successful application of PINNs to turbulent flows at Reynolds numbers up to 10,000. It yields 2–10x accuracy improvements over existing methods while maintaining computational scalability, advancing the frontier of neural PDE solvers for real-world, multi-scale physical systems.



Paperid:4867
Authors:Anindya Sarkar, Binglin Ji, Yevgeniy Vorobeychik
Abstract:
In many scientific and engineering fields, where acquiring high-quality data is expensive—such as medical imaging, environmental monitoring, and remote sensing—strategic sampling of unobserved regions based on prior observations is crucial for maximizing discovery rates within a constrained budget. The rise of powerful generative models, such as diffusion models, has enabled active target discovery in partially observable environments by leveraging learned priors—probabilistic representations that capture underlying structure from data. With guidance from sequentially gathered task-specific observations, these models can progressively refine exploration and efficiently direct queries toward promising regions. However, in domains where learning a strong prior is infeasible due to extremely limited data or high sampling cost (such as rare species discovery, diagnostics for emerging diseases, etc.), these methods struggle to generalize. To overcome this limitation, we propose a novel approach that enables effective active target discovery even in settings with uninformative priors, ensuring robust exploration and adaptability in complex real-world scenarios. Our framework is theoretically principled and draws inspiration from neuroscience to guide its design. Unlike black-box policies, our approach is inherently interpretable, providing clear insights into decision-making. Furthermore, it guarantees a strong, monotonic improvement in prior estimates with each new observation, leading to increasingly accurate sampling and reinforcing both reliability and adaptability in dynamic settings. Through comprehensive experiments and ablation studies across various domains, including species distribution modeling and remote sensing, we demonstrate that our method substantially outperforms baseline approaches.



Paperid:4868
Authors:Vasiliki Tassopoulou, Charis Stamouli, Haochang Shou, George J. Pappas, Christos Davatzikos
Abstract:
We introduce a novel conformal prediction framework for constructing valid prediction bands around biomarker trajectories observed at subject-specific, randomly-timed follow-up visits. Existing conformal methods typically assume fixed time grids, limiting their applicability in longitudinal clinical studies. Our approach addresses this limitation by defining a time-adaptive nonconformity score that normalizes prediction errors using model-derived uncertainty estimates, enabling valid inference at arbitrary time points. We evaluate our method on two well-established brain biomarkers—hippocampal and ventricular volume—using a range of standard and state-of-the-art predictors. Across models, our conformalized predictors consistently achieve nominal coverage with tighter prediction intervals compared to baseline uncertainty estimates. To further accommodate population heterogeneity, we develop group-conditional conformal bands with formal coverage guarantees across clinically relevant and high-risk subgroups. Finally, we demonstrate the clinical utility of our approach in identifying subjects at risk of progression to Alzheimer’s disease. We introduce an uncertainty-aware progression metric based on the lower conformal bound and show that it enables the identification of 17.5\% more high-risk subjects compared to standard slope-based methods, highlighting the value of uncertainty calibration in real-world clinical decision making.



Paperid:4869
Authors:Jaehun Jung, Seungju Han, Ximing Lu, Skyler Hallinan, David Acuna, Shrimai Prabhumoye, Mostofa Patwary, Mohammad Shoeybi, Bryan Catanzaro, Yejin Choi
Abstract:
Abstract:Data diversity is crucial for training a strong language model. Yet metrics of diversity often diverge from this goal, measuring variations in heuristic features—like n-grams or embeddings—that are detached from how the model actually performs on a target task. This motivates us to ask: *Can we redefine data diversity—beyond measuring variations in heuristic features—in a way that better predicts model generalization?* Through large-scale empirical analyses spanning over 300 training runs, carefully controlled for data scale and quality, we show that data diversity can be a strong predictor of generalization in LLM reasoning—as measured by average model performance on unseen out-of-distribution benchmarks. We introduce **G-Vendi**, a metric that quantifies diversity via the entropy of model-induced loss gradients. G-Vendi scales to million-sample datasets and yet consistently outperforms heuristic alternatives, achieving strong correlation ($\text{Spearman's } \rho \approx 0.9$) with out-of-distribution (OOD) performance across both natural language inference (NLI) and math reasoning tasks. Building on this insight, we present **Prismatic Synthesis**, a framework for generating diverse synthetic data by targeting underrepresented regions in gradient space. Experimental results show that Prismatic Synthesis consistently improves model performance as we scale synthetic data—not just on in-distribution test but across unseen, out-of-distribution benchmarks—significantly outperforming state-of-the-art models in both domains. For example, PrismMath-7B, our model distilled from a 32B LLM without human verification, outperforms R1-Distill-Qwen-7B—trained on proprietary data generated by 671B R1—on 6 out of 7 challenging math benchmarks.



Paperid:4870
Authors:Jihoon Kwon, Kyle Min, Jy-yong Sohn
Abstract:
Despite recent advances, vision-language models trained with standard contrastive objectives still struggle with compositional reasoning -- the ability to understand structured relationships between visual and linguistic elements.This shortcoming is largely due to the tendency of the text encoder to focus on individual words rather than their relations, a limitation reinforced by contrastive training that primarily aligns words with visual objects.In this paper, we introduce REconstruction and Alignment of text Descriptions (READ), a fine-tuning method designed to enhance compositional reasoning by adding two auxiliary objectives to the contrastive learning: (1) a token-level reconstruction objective, where a frozen pre-trained decoder reconstructs paraphrased captions based on the embedding of the original caption; and (2) a sentence-level alignment objective, which explicitly aligns paraphrased sentences in the embedding space.We show that READ-CLIP, a model derived by applying the READ method to the pre-trained CLIP model, achieves the state-of-the-art performance across five major compositional reasoning benchmarks, outperforming the strongest conventional fine-tuning baseline by up to 4.1\%.Furthermore, applying READ to existing CLIP variants (including NegCLIP and FSC-CLIP) also improves performance on these benchmarks.Quantitative and qualitative analyses reveal that our proposed objectives -- reconstruction and alignment -- offer complementary benefits: the former encourages the encoder to capture relationships between words within a caption, while the latter ensures consistent representations for paraphrases expressed with different wording.



Paperid:4871
Authors:Jan Hagnberger, Daniel Musekamp, Mathias Niepert
Abstract:
Solving time-dependent Partial Differential Equations (PDEs) using a densely discretized spatial domain is a fundamental problem in various scientific and engineering disciplines, including modeling climate phenomena and fluid dynamics. However, performing these computations directly in the physical space often incurs significant computational costs. To address this issue, several neural surrogate models have been developed that operate in a compressed latent space to solve the PDE. While these approaches reduce computational complexity, they often use Transformer-based attention mechanisms to handle irregularly sampled domains, resulting in increased memory consumption. In contrast, convolutional neural networks allow memory-efficient encoding and decoding but are limited to regular discretizations. Motivated by these considerations, we propose CALM-PDE, a model class that efficiently solves arbitrarily discretized PDEs in a compressed latent space. We introduce a novel continuous convolution-based encoder-decoder architecture that uses an epsilon-neighborhood-constrained kernel and learns to apply the convolution operator to adaptive and optimized query points. We demonstrate the effectiveness of CALM-PDE on a diverse set of PDEs with both regularly and irregularly sampled spatial domains. CALM-PDE is competitive with or outperforms existing baseline methods while offering significant improvements in memory and inference time efficiency compared to Transformer-based methods.



Paperid:4872
Authors:Harsh Poonia, Felix Divo, Kristian Kersting, Devendra Singh Dhami
Abstract:
Causality in time series can be difficult to determine, especially in the presence of non-linear dependencies. The concept of Granger causality helps analyze potential relationships between variables, thereby offering a method to determine whether one time series can predict—Granger cause—future values of another. Although successful, Granger causal methods still struggle with capturing long-range relations between variables. To this end, we leverage the recently successful Extended Long Short-Term Memory (xLSTM) architecture and propose Granger causal xLSTMs (GC-xLSTM). It first enforces sparsity between the time series components by using a novel dynamic loss penalty on the initial projection. Specifically, we adaptively improve the model and identify sparsity candidates. Our joint optimization procedure then ensures that the Granger causal relations are recovered robustly. Our experimental evaluation on six diverse datasets demonstrates the overall efficacy of our proposed GC-xLSTM model.



Paperid:4873
Authors:Danfeng Li, Hui Zhang, Sheng Wang, Jiacheng Li, Zuxuan Wu
Abstract:
Despite recent advances in diffusion models, top-tier text-to-image (T2I) models still struggle to achieve precise spatial layout control,i.e.accurately generating entities with specified attributes and locations. Segmentation mask-to-image (S2I) generation has emerged as a promising solution by incorporating pixel-level spatial guidance and regional text prompts. However, existing S2I methods fail to simultaneously ensure semantic consistency and shape consistency.To address these challenges, we propose Seg2Any, a novel S2I framework built upon advanced multimodal diffusion transformers (e.g.FLUX). First, to achieve both semantic and shape consistency, we decouple segmentation mask conditions into regional semantic and high-frequency shape components. The regional semantic condition is introduced by a Semantic Alignment Attention Mask, ensuring that generated entities adhere to their assigned text prompts. The high-frequency shape condition, representing entity boundaries, is encoded as an Entity Contour Map and then introduced as an additional modality via multi-modal attention to guide image spatial structure. Second, to prevent attribute leakage across entities in multi-entity scenarios, we introduce an Attribute Isolation Attention Mask mechanism, which constrains each entity’s image tokens to attend exclusively to themselves during image self-attention.To support open-set S2I generation, we construct SACap-1M, a large-scale dataset containing 1 million images with 5.9 million segmented entities and detailed regional captions, along with a SACap-Eval benchmark for comprehensive S2I evaluation.Extensive experiments demonstrate that Seg2Any achieves state-of-the-art performance on both open-set and closed-set S2I benchmarks, particularly in fine-grained spatial and attribute control of entities.



Paperid:4874
Authors:Fan Li, Xuan Wang, Xuanbin Wang, Zhaoxiang Zhang, Yuelei Xu
Abstract:
Enhancing the cross-domain generalization of 3D semantic segmentation is a pivotal task in computer vision that has recently gained increasing attention. Most existing methods, whether using consistency regularization or cross-modal feature fusion, focus solely on individual objects while overlooking implicit semantic dependencies among them, resulting in the loss of useful semantic information. Inspired by the diffusion model's ability to flexibly compose diverse objects into high-quality images across varying domains, we seek to harness its capacity for capturing underlying contextual distributions and spatial arrangements among objects to address the challenging task of cross-domain 3D semantic segmentation. In this paper, we propose a novel cross-modal learning framework based on diffusion models to enhance the generalization of 3D semantic segmentation, named XDiff3D. XDiff3D comprises three key ingredients: (1) constructing object agent queries from diffusion features to aggregate instance semantic information; (2) decoupling fine-grained local details from object agent queries to prevent interference with 3D semantic representation; (3) leveraging object agent queries as an interface to enhance the modeling of object semantic dependencies in 3D representations. Extensive experiments validate the effectiveness of our method, achieving state-of-the-art performance across multiple benchmarks in different task settings. We will make the code publicly available.



Paperid:4875
Authors:Yunseon Choi, Minchan Jeong, Soobin Um, Kee-Eung Kim
Abstract:
Human experts employ diverse strategies to complete a task, producing to multi-modal demonstration data. Although traditional Adversarial Imitation Learning (AIL) methods have achieved notable success, they often collapse theses multi-modal behaviors into a single strategy, failing to replicate expert behaviors. To overcome this limitation, we propose DPAIL, an adversarial IL framework that leverages diffusion models as a policy class to enhance expressiveness. Building on the Adversarial Soft Advantage Fitting (ASAF) framework, which removes the need for policy optimization steps, DPAIL trains a diffusion policy using a binary cross-entropy objective to distinguish expert trajectories from generated ones. To enable optimization of the diffusion policy, we introduce a novel, tractable lower bound on the policy's likelihood. Through comprehensive quantitative and qualitative evaluations against various baselines, we demonstrate that our method not only captures diverse behaviors but also remains robust as the number of behavior modes increases.



Paperid:4876
Authors:Li Zhang, Zhongxuan Han, Chaochao Chen, XiaoHua Feng, Jiaming Zhang, Yuyuan Li
Abstract:
With emerging application of Federated Learning (FL) in decision-making scenarios, it is imperative to regulate model fairness to prevent disparities across sensitive groups (e.g., female, male).Current research predominantly focuses on two concepts of group fairness within FL:Global Fairness(overall model disparity across all clients) andLocal Fairness(the disparity within each client).However, the non-decomposable, non-differentiable nature of fairness criteria pose two fundamental, unresolved challenges for fair FL: (i)Harmonizing global and local fairness in multi-class classification; (ii)Enabling a controllable, optimal accuracy-fairness trade-off.To tackle the aforementioned challenges, we propose a novel controllable federated group-fairness calibration framework, named FedFACT.FedFACT identifies the Bayes-optimal classifiers under both global and local fairness constraints in multi-class case, yielding models with minimal performance decline while guaranteeing fairness.To effectively realize an adjustable, optimal accuracy-fairness balance, we derive specific characterizations of the Bayes-optimal fair classifiers for reformulating fair FL as personalized cost-sensitive learning problem for in-processing, and bi-level optimization for post-processing.Theoretically, we provide convergence and generalization guarantees for FedFACT to approach the near-optimal accuracy under given fairness levels.Extensive experiments on multiple datasets across various data heterogeneity demonstrate that FedFACT consistently outperforms baselines in balancing accuracy and global-local fairness.



Paperid:4877
Authors:Konstantinos Dafnis, Dimitris Metaxas
Abstract:
Vision–language models (VLMs) excel at zero-shot inference but often degrade under test-time domain shifts. For this reason, episodic test-time adaptation strategies have recently emerged as powerful techniques for adapting VLMs to a single unlabeled image. However, existing adaptation strategies, such as test-time prompt tuning, typically require backpropagating through large encoder weights or altering core model components. In this work, we introduce \textbf{S}pectrum-Aware \textbf{T}est-Time \textbf{S}teering (\textbf{STS}), a \textit{lightweight adaptation framework} that extracts a spectral subspace from the textual embeddings to define principal semantic directions, and learns to steer latent representations in a spectrum-aware manner by adapting a small number of per-sample shift parameters to minimize entropy across augmented views. STS operates entirely at inference in the latent space, without backpropagation through or modification of the frozen encoders. Building on standard evaluation protocols, our comprehensive experiments demonstrate that STS largely surpasses or compares favorably against state-of-the-art test-time adaptation methods, while introducing only a handful of additional parameters and achieving inference speeds up to 8× faster with a 12× smaller memory footprint than conventional test-time prompt tuning.



Paperid:4878
Authors:Dahao Xu, Jiahua Rao, Mingming Zhu, Jixian Zhang, Wei Lu, Shuangjia Zheng, Yuedong Yang
Abstract:
Abstract:Predicting changes in binding free energy ($\Delta\Delta G$) is essential for understanding protein-protein interactions, which are critical in drug design and protein engineering. However, existing methods often rely on pre-trained knowledge and heuristic features, limiting their ability to accurately model complex mutation effects, particularly higher-order and many-body interactions.To address these challenges, we propose H3-DDG, a Hypergraph-driven Hierarchical network to capture Higher-order many-body interactions across multiple scales. By introducing a hierarchical communication mechanism, H3-DDG effectively models both local and global mutational effects.Experimental results demonstrate state-of-the-art performance on multiple benchmarks. On the SKEMPI v2 dataset, H3-DDG achieves a Pearson correlation of 0.75, improving multi-point mutations prediction by 12.10%. On the challenging BindingGYM dataset, it outperforms Prompt-DDG and BA-DDG by 62.61% and 34.26%, respectively.Ablation and efficiency analyses demonstrate its robustness and scalability, while a case study on SARS-CoV-2 antibodies highlights its practical value in improving binding affinity for therapeutic design.



Paperid:4879
Authors:Ilias Diakonikolas, Chao Gao, Daniel Kane, John Lafferty, Ankit Pensia
Abstract:
Abstract:We study the task of noiseless linear regression under Gaussian covariates in the presence of additive oblivious contamination. Specifically, we are given i.i.d.\ samples from a distribution $(x, y)$ on $\mathbb R^d \times \mathbb R$ with $x \sim \mathcal N(0,I_d)$ and $y = x^\top \beta + z$, where $z$ is drawn from an unknown distribution that is independent of $x$. Moreover, $z$ satisfies $\mathbb P[z = 0] = \alpha>0$.The goal is to accurately recover the regressor $\beta$ to small $\ell_2$-error. Ignoring computational considerations, this problem is known to be solvable using $O(d/\alpha)$ samples. On the other hand, the best known polynomial-time algorithms require $\Omega(d/\alpha^2)$ samples. Here we provide formal evidence that the quadratic dependence in $1/\alpha$ is inherent for efficient algorithms. Specifically, we show that any efficient Statistical Query algorithmfor this task requires sample complexity at least $\tilde{\Omega}(d^{1/2}/\alpha^2)$.



Paperid:4880
Authors:Navid Hashemi, Samuel Sasaki, Ipek Oguz, Meiyi Ma, Taylor Johnson
Abstract:
Semantic segmentation neural networks (SSNs) are increasingly essential in high-stakes fields such as medical imaging, autonomous driving, and environmental monitoring, where robustness to input uncertainties and adversarial examples is crucial for ensuring safety and reliability. However, traditional probabilistic verification methods struggle to scale effectively with the size and depth of modern SSNs, especially when dealing with their high-dimensional, structured inputs/outputs. As the output dimension increases, these methods tend to become overly conservative, leading to pointless and overly restrictive safety guarantees. In this work, we propose a probabilistic, data-driven verification algorithm that is architecture-agnostic and scalable to handle the high-dimensional outputs of SSNs without introducing conservative and loose guarantees. We leverage efficient sampling-based reachability analysis to explore the space of possible outputs while maintaining computational feasibility. Our methodology is based on Conformal Inference (CI), which is known for its high data efficiency. However, CI tends to be overly conservative in high-dimensional spaces. To address this, in this paper, we introduce effective techniques that mitigate these sources of conservatism, enabling us to provide less conservative yet provable guarantees for SSNs. We validate our approach on deep segmentation models applied to benchmark datasets such as CamVid and Lung-Segmentation, showing that it can offer reliable safety guarantees while lowering the conservatism inherent in traditional methods.



Paperid:4881
Authors:Duo Wang, Yuan Zuo, Guangyue Lu, Junjie Wu
Abstract:
Generalizing to unseen graph tasks without task-specific supervision is challenging: conventional graph neural networks are typically tied to a fixed label space, while large language models (LLMs) struggle to capture graph structure. We introduce UniGTE, an instruction-tuned encoder–decoder framework that unifies structural and semantic reasoning. The encoder augments a pretrained autoregressive LLM with learnable alignment tokens and a structure-aware graph–text attention mechanism, enabling it to attend jointly to a tokenized graph and a natural-language task prompt while remaining permutation-invariant to node order. This yields compact, task-aware graph representations. Conditioned solely on these representations, a frozen LLM decoder predicts and reconstructs: it outputs the task answer and simultaneously paraphrases the input graph in natural language. The reconstruction objective regularizes the encoder to preserve structural cues. UniGTE is instruction-tuned on five datasets spanning node-, edge-, and graph-level tasks across diverse domains, yet requires no fine-tuning at inference. It achieves new state-of-the-art zero-shot results on node classification, link prediction, graph classification and graph regression under cross-task and cross-domain settings, demonstrating that tight integration of graph structure with LLM semantics enables robust, transferable graph reasoning.



Paperid:4882
Authors:Bryan Wong, Jong kim, Huazhu Fu, Mun Yi
Abstract:
Vision-language models (VLMs) have recently been integrated into multiple instance learning (MIL) frameworks to address the challenge of few-shot, weakly supervised classification of whole slide images (WSIs). A key trend involves leveraging multi-scale information to better represent hierarchical tissue structures. However, existing methods often face two key limitations: (1) insufficient modeling of interactions within the same modalities across scales (e.g., 5x and 20x) and (2) inadequate alignment between visual and textual modalities on the same scale. To address these gaps, we propose HiVE-MIL, a hierarchical vision-language framework that constructs a unified graph consisting of (1) parent–child links between coarse (5x) and fine (20x) visual/textual nodes to capture hierarchical relationships, and (2) heterogeneous intra-scale edges linking visual and textual nodes on the same scale. To further enhance semantic consistency, HiVE-MIL incorporates a two-stage, text-guided dynamic filtering mechanism that removes weakly correlated patch–text pairs, and introduces a hierarchical contrastive loss to align textual semantics across scales. Extensive experiments on TCGA breast, lung, and kidney cancer datasets demonstrate that HiVE-MIL consistently outperforms both traditional MIL and recent VLM-based MIL approaches, achieving gains of up to 4.1% in macro F1 under 16-shot settings. Our results demonstrate the value of jointly modeling hierarchical structure and multimodal alignment for efficient and scalable learning from limited pathology data. The code is available at https://anonymous.4open.science/r/HiVE-MIL



Paperid:4883
Authors:Dongzhe Zheng, Wenjie Mei
Abstract:
Stochastic optimal control methods often struggle in complex non-convex landscapes, frequently becoming trapped in local optima due to their inability to learn from historical trajectory data. This paper introduces Memory-Augmented Potential Field Theory, a unified mathematical framework that integrates historical experience into stochastic optimal control. Our approach dynamically constructs memory-based potential fields that identify and encode key topological features of the state space, enabling controllers to automatically learn from past experiences and adapt their optimization strategy. We provide a theoretical analysis showing that memory-augmented potential fields possess non-convex escape properties, asymptotic convergence characteristics, and computational efficiency. We implement this theoretical framework in a Memory-Augmented Model Predictive Path Integral (MPPI) controller that demonstrates significantly improved performance in challenging non-convex environments. The framework represents a generalizable approach to experience-based learning within control systems (especially robotic dynamics), enhancing their ability to navigate complex state spaces without requiring specialized domain knowledge or extensive offline training.



Paperid:4884
Authors:Pengrui Quan, Brian Wang, Kang Yang, Liying Han, Mani Srivastava
Abstract:
Spatiotemporal reasoning plays a key role in Cyber-Physical Systems (CPS). Despite advances in Large Language Models (LLMs) and Large Reasoning Models (LRMs), their capacity to reason about complex spatiotemporal signals remains underexplored. This paper proposes a hierarchical SpatioTemporal reAsoning benchmaRK, STARK, to systematically evaluate LLMs across three levels of reasoning complexity: state estimation (e.g., predicting field variables, localizing and tracking events in space and time), spatiotemporal reasoning over states (e.g., inferring spatial-temporal relationships), and world-knowledge-aware reasoning that integrates contextual and domain knowledge (e.g., intent prediction, landmark-aware navigation). We curate 26 distinct spatiotemporal tasks with diverse sensor modalities, comprising 14,552 challenges where models answer directly or by Python Code Interpreter. Evaluating 3 LRMs and 8 LLMs, we find LLMs achieve limited success in tasks requiring geometric reasoning (e.g., multilateration or triangulation), particularly as complexity increases. Surprisingly, LRMs show robust performance across tasks with various levels of difficulty, often competing or surpassing traditional first-principle-based methods. Our results show that in reasoning tasks requiring world knowledge, the performance gap between LLMs and LRMs narrows, with some LLMs even surpassing LRMs. However, the LRM o3 model continues to achieve leading performance across all evaluated tasks, a result attributed primarily to the larger size of the reasoning models. STARK motivates future innovations in model architectures and reasoning paradigms for intelligent CPS by providing a structured framework to identify limitations in the spatiotemporal reasoning of LLMs and LRMs.



Paperid:4885
Authors:Xiaohan Qin, Xiaoxing Wang, Ning Liao, Junchi Yan
Abstract:
Multi-Task Learning (MTL) enables a single model to learn multiple tasks simultaneously, leveraging knowledge transfer among tasks for enhanced generalization, and has been widely applied across various domains. However, task imbalance remains a major challenge in MTL. Although balancing the convergence speeds of different tasks is an effective approach to address this issue, it is highly challenging to accurately characterize the training dynamics and convergence speeds of multiple tasks within the complex MTL system. To this end, we attempt to analyze the training dynamics in MTL by leveraging Neural Tangent Kernel (NTK) theory and propose a new MTL method, NTKMTL. Specifically, we introduce an extended NTK matrix for MTL and adopt spectral analysis to balance the convergence speeds of multiple tasks, thereby mitigating task imbalance.Based on the approximation via shared representation, we further propose NTKMTL-SR, achieving training efficiency while maintaining competitive performance. Extensive experiments demonstrate that our methods achieve state-of-the-art performance across a wide range of benchmarks, including both multi-task supervised learning and multi-task reinforcement learning.



Paperid:4886
Authors:Fuyuan CAO, Shichang Qiao, Kui Yu, Jiye Liang
Abstract:
Many confounding-resistant domain generalization methods for image classification have been developed based on causal interventions. However, their reliance on strong assumptions limits their effectiveness in handling unobserved confounders. Although recent work introduces instrumental variables (IVs) to overcome this limitation, the reliance on manually predefined instruments, particularly in the context of visual data, may result in severe bias or invalidity when IV conditions are violated. To address these issues, we propose a novel approach to automatically learning Visual Instrumental Variables for confounding-resistant Domain Generalization (VIV-DG). We observe that certain non-causal visual attributes in image data naturally satisfy the basic conditions required for valid IVs. Motivated by this insight, we propose the visual instrumental variable, a novel concept that extends classical IV theory to the visual domain. Furthermore, we develop an automatic visual instrumental variable learner that enforces IV conditions on learned representations, enabling the automatic learning of valid visual instrumental variables from image data. Ultimately, VIV-DG inherits the strengths of classical IVs to mitigate unobserved confounding and avoids the significant bias caused by violations of IV conditions in predefined IVs. Extensive experiments on multiple benchmarks verify that VIV-DG achieves superior generalization ability.



Paperid:4887
Authors:Yihao LIU, Xinqi Lyu, Dong Wang, Yanjie Li, Bin Xiao
Abstract:
Large vision-language models (VLLMs) have driven significant progress in multi-modal systems, enabling a wide range of applications across domains such as healthcare, education, and content generation. Despite the success, the large-scale datasets used to train these models often contain sensitive or personally identifiable information, raising serious privacy concerns. To audit and better understand such risks, membership inference attacks (MIAs) have become a key tool. However, existing MIAs against VLLMs predominantly assume access to full-model logits, which are typically unavailable in many practical deployments. To facilitate MIAs in a more realistic and restrictive setting, we propose a novel framework: label-only membership inference attacks (LOMIA) targeting pre-trained VLLMs where only the model’s top-1 prediction is available. Within this framework, we propose three effective attack methods, all of which exploit the intuition that training samples are more likely to be memorized by the VLLMs, resulting in outputs that exhibit higher semantic alignment and lower perplexity. Our experiments show that our framework surpasses existing label-only attack adaptations for different VLLMs and competes with state-of-the-art logits-based attacks across all metrics on three widely used open-source VLLMs and GPT-4o.



Paperid:4888
Authors:Yuxuan Luo, Yuhui Yuan, Junwen Chen, Haonan Cai, Ziyi Yue, Yuwei Yang, Fatima Zohra Daha, Ji Li, Zhouhui Lian
Abstract:
Abstract:In this paper, we introduce knowledge image generation as a new task, alongside the Massive Multi-Discipline Multi-Tier Knowledge-Image Generation Benchmark (MMMG) to probe the reasoning capability of image generation models.Knowledge images have been central to human civilization and to the mechanisms of human learning—a fact underscored by dual-coding theory and the picture-superiority effect.Generating such images is challenging, demanding multimodal reasoning that fuses world knowledge with pixel-level grounding into clear explanatory visuals.To enable comprehensive evaluation, MMMG offers $4,456$ expert-validated (knowledge) image-prompt pairs spanning $10$ disciplines, $6$ educational levels, and diverse knowledge formats such as charts, diagrams, and mind maps. To eliminate confounding complexity during evaluation, we adopt a unified Knowledge Graph (KG) representation. Each KG explicitly delineates a target image’s core entities and their dependencies.We further introduce MMMG-Score to evaluate generated knowledge images. This metric combines factual fidelity, measured by graph-edit distance between KGs, with visual clarity assessment.Comprehensive evaluations of $18$ state-of-the-art text-to-image generation models expose serious reasoning deficits—low entity fidelity, weak relations, and clutter—with GPT-4o achieving an MMMG-Score of only $46.66$, underscoring the benchmark’s difficulty.To spur further progress, we release FLUX-Reason (MMMG-Score of $30.52$), an effective and open baseline that combines a reasoning LLM with diffusion models and is trained on $16,000$ curated knowledge image–prompt pairs.



Paperid:4889
Authors:Liang Ma, Jiajun Wen, Min Lin, Rongtao Xu, Xiwen Liang, Bingqian Lin, Jun Ma, Yongxin Wang, Ziming Wei, haokun lin, Mingfei Han, Meng Cao, Bokui Chen, Ivan Laptev, Xiaodan Liang
Abstract:
While vision-language models (VLMs) have demonstrated promising capabilities in reasoning and planning for embodied agents, their ability to comprehend physical phenomena, particularly within structured 3D environments, remains severely limited. To close this gap, we introduce PhyBlock, a progressive benchmark designed to assess VLMs on physical understanding and planning through robotic 3D block assembly tasks. PhyBlock integrates a novel four-level cognitive hierarchy assembly task alongside targeted Visual Question Answering (VQA) samples, collectively aimed at evaluating progressive spatial reasoning and fundamental physical comprehension, including object properties, spatial relationships, and holistic scene understanding. PhyBlock includes 2600 block tasks (400 assembly tasks, 2200 VQA tasks) and evaluates models across three key dimensions: partial completion, failure diagnosis, and planning robustness. We benchmark 21 state-of-the-art VLMs, highlighting their strengths and limitations in physically grounded, multi-step planning. Our empirical findings indicate that the performance of VLMs exhibits pronounced limitations in high-level planning and reasoning capabilities, leading to a notable decline in performance for the growing complexity of the tasks.Error analysis reveals persistent difficulties in spatial orientation and dependency reasoning. Surprisingly, chain-of-thought prompting offers minimal improvements, suggesting spatial tasks heavily rely on intuitive model comprehension. We position PhyBlock as a unified testbed to advance embodied reasoning, bridging vision-language understanding and real-world physical problem-solving.



Paperid:4890
Authors:Sungwoo Lee, Jungmin Lee, Sohee Kim, Hyebhin Yoon, Shinwon Park, Junhyeok Park, Jaehyuk Bae, Seok-Jun Hong, Choong-Wan Woo
Abstract:
Reinforcement learning (RL) agents have demonstrated strong performance in structured environments, yet they continue to struggle in real-world settings where goals are ambiguous, conditions change dynamically, and external supervision is limited. These challenges stem not primarily from the algorithmic limitations but from the characteristics of conventional training environments, which are usually static, task-specific, and externally defined. In contrast, biological agents develop autonomy and adaptivity by interacting with complex, dynamic environments, where most behaviors are ultimately driven by internal physiological needs. Inspired by these biological constraints, we introduce EVAAA (Essential Variables in Autonomous and Adaptive Agents), a 3D virtual environment for training and evaluating egocentric RL agents endowed with internal physiological state variables. In EVAAA, agents must maintain essential variables (EVs)—e.g., satiation, hydration, body temperature, and tissue integrity (the level of damage)—within viable bounds by interacting with environments that increase in difficulty at each stage. The reward system is derived from internal state dynamics, enabling agents to generate goals autonomously without manually engineered, task-specific reward functions. Built on Unity ML-Agents, EVAAA supports multimodal sensory inputs, including vision, olfaction, thermoception, collision, as well as egocentric embodiment. It features naturalistic survival environments for curricular training and a suite of unseen experimental testbeds, allowing for the evaluation of autonomous and adaptive behaviors that emerge from the interplay between internal state dynamics and environmental constraints. By integrating physiological regulation, embodiment, continual learning, and generalization, EVAAA offers a biologically inspired benchmark for studying autonomy, adaptivity, and internally driven control in RL agents. Our code is publicly available at https://anonymous.4open.science/r/evaaa-2486



Paperid:4891
Authors:Shayan Shekarforoush, David Lindell, Marcus Brubaker, David Fleet
Abstract:
Cryo-EM is a transformational paradigm in molecular biology where computational methods are used to infer 3D molecular structure at atomic resolution from extremely noisy 2D electron microscope images. At the forefront of research is how to model the structure when the imaged particles exhibit non-rigid conformational flexibility and compositional variation where parts are sometimes missing. We introduce a novel 3D reconstruction framework with a hierarchical Gaussian mixture model, inspired in part by Gaussian Splatting for 4D scene reconstruction. In particular, the structure of the model is grounded in an initial process that infers a part-based segmentation of the particle, providing essential inductive bias in order to handle both conformational and compositional variability. The framework, called CryoSPIRE, is shown to reveal biologically meaningful structures on complex experimental datasets, and establishes a new state-of-the-art on CryoBench, a benchmark for cryo-EM heterogeneity methods.



Paperid:4892
Authors:Qirui Mi, Mengyue Yang, Xiangning Yu, Zhiyu Zhao, Cheng Deng, Bo An, Haifeng Zhang, Xu Chen, Jun Wang
Abstract:
Simulating collective decision-making involves more than aggregating individual behaviors; it emerges from dynamic interactions among individuals. While large language models (LLMs) offer strong potential for social simulation, achieving quantitative alignment with real-world data remains a key challenge. To bridge this gap, we propose the \textbf{M}ean-\textbf{F}ield \textbf{LLM} (\textbf{MF-LLM}) framework, the first to incorporate mean field theory into LLM-based social simulation. MF-LLM models bidirectional interactions between individuals and the population through an iterative process, generating population signals to guide individual decisions, which in turn update the signals. This interplay produces coherent trajectories of collective behavior.To improve alignment with real-world data, we introduce \textbf{IB-Tune}, a novel fine-tuning method inspired by the \textbf{I}nformation \textbf{B}ottleneck principle, which retains population signals most predictive of future actions while filtering redundant history. Evaluated on a real-world social dataset, MF-LLM reduces KL divergence to human population distributions by \textbf{47\%} compared to non-mean-field baselines, enabling accurate trend forecasting and effective intervention planning. Generalizing across 7 domains and 4 LLM backbones, MF-LLM provides a scalable, high-fidelity foundation for social simulation.



Paperid:4893
Authors:Thummaluru Siddartha Reddy, Vempalli Naga Sai Saketh, Mahesh Chandran
Abstract:
Abstract:Graph Neural Networks \texttt{(GNNs)} excel at jointly modeling node features and topology, yet their \emph{black-box} nature limits their adoption in real-world applications where interpretability is desired. Inspired by the success of interpretable Neural Additive Model \texttt{(NAM)} for tabular data, Graph Neural Additive Network \texttt{(GNAN)} extends the additive modeling approach to graph data to overcome limitations of GNNs. While being interpretable, \texttt{GNAN} representation learning overlooks the importance of local aggregation and more importantly suffers from parameter complexity. To mitigate the above challenges, we introduce Graph Neural Additive Model with Random Fourier Features (\texttt{G-NAMRFF}), a lightweight, self‐interpretable graph additive architecture. \texttt{G-NAMRFF} represents each node embedding as the sum of feature‐wise contributions where each nonlinear mapping is modeled as a \emph{Gaussian process} \texttt{(GP)} with a graph- and feature-aware kernel. Specifically, we construct a kernel using Radial Basis Function (\texttt{RBF}) with graph structure induced by Laplacian and learnable Finite Impulse Response (\texttt{FIR}) filter. We approximate the kernel using Random Fourier Features \texttt{RFFs} which transforms the \texttt{GP} prior to a linear Bayesian formulation, which are subsequently learnt using a single layer neural network with size equal to number of \texttt{RFF} features. \texttt{G-NAMRFF} is light weight with $168\times$ fewer parameters compared to \texttt{GNAN}. Despite its compact size, \texttt{G-NAMRFF} matches or outperforms state-of-the-art \texttt{GNNs} and \texttt{GNAN} on node and graph classification tasks, delivering real-time interpretability without sacrificing accuracy.



Paperid:4894
Authors:Jiangjie Chen, Qianyu He, Siyu Yuan, Chen, Zhicheng Cai, Weinan Dai, Hongli Yu, Jiaze Chen, Xuefeng Li, Qiying Yu, Hao Zhou, Mingxuan Wang
Abstract:
Large Language Models (LLMs), such as OpenAI's o1 and DeepSeek's R1, excel at advanced reasoning tasks like math and coding via Reinforcement Learning with Verifiable Rewards (RLVR), but still struggle with puzzles solvable by humans without domain knowledge.We introduce Enigmata, the first comprehensive suite tailored for improving LLMs with puzzle reasoning skills. It includes 36 tasks across 7 categories, each with: 1) a generator that produces unlimited examples with controllable difficulty, and 2) a rule-based verifier for automatic evaluation.This generator-verifier design supports scalable, multi-task RL training, fine-grained analysis, and seamless RLVR integration.We further propose Enigmata-Eval, a rigorous benchmark, and develop optimized multi-task RLVR strategies.Our trained model, Qwen2.5-32B-Enigmata, consistently surpasses o3-mini-high and o1 on the puzzle reasoning benchmarks like Enigmata-Eval, ARC-AGI, and ARC-AGI 2.It also generalizes well to out-of-domain puzzle benchmarks and mathematical reasoning, with little multitasking trade-off.This work offers a unified, controllable framework for advancing logical reasoning in LLMs. Resources will be released upon publication.



Paperid:4895
Authors:Xiaojian Ding, Lin Zhao, Xian Li, Xiaoying Zhu
Abstract:
To address the challenges of insufficient semantic alignment across heterogeneous views, rigid fusion strategies, and the decoupling of missing view imputation from representation learning in incomplete multi-view clustering, this paper proposes HSACC, a novel framework based on Hierarchical Semantic Alignment and Cooperative Completion.HSACC introduces a dual-level semantic space design. In the low-level semantic space, consistency alignment is achieved by maximizing mutual information across views. In the high-level semantic space, adaptive view weights are dynamically assigned based on the distribution affinity between individual views and an initial fused representation. These weights are then used to perform weighted fusion, generating a unified global representation.Additionally, HSACC implicitly recovers missing views by projecting aligned latent representations into high-dimensional semantic spaces and jointly optimizes reconstruction and clustering to enable cooperative learning.Experimental results demonstrate that HSACC significantly outperforms state-of-the-art methods on five benchmark datasets. Ablation studies validate the effectiveness of the hierarchical alignment and dynamic weighting mechanisms, while parameter analysis confirms the model's robustness to hyperparameter variations.



Paperid:4896
Authors:Zongqian Li, Yixuan Su, Nigel Collier
Abstract:
Parameter-efficient fine-tuning (PEFT) methods have shown promise in adapting large language models, yet existing approaches exhibit counter-intuitive phenomena: integrating either matrix decomposition or mixture-of-experts (MoE) individually decreases performance across tasks, though decomposition improves results on specific domains despite reducing parameters, while MoE increases parameter count without corresponding decrease in training efficiency. Motivated by these observations and the modular nature of PT, we propose PT-MoE, a novel framework that integrates matrix decomposition with MoE routing for efficient PT. Evaluation results across 17 datasets demonstrate that PT-MoE achieves state-of-the-art performance in both question answering (QA) and mathematical problem solving tasks, improving F1 score by 1.49 points over PT and 2.13 points over LoRA in QA tasks, while improving mathematical accuracy by 10.75 points over PT and 0.44 points over LoRA, all while using 25% fewer parameters than LoRA. Our analysis reveals that while PT methods generally excel in QA tasks and LoRA-based methods in math datasets, the integration of matrix decomposition and MoE in PT-MoE yields complementary benefits: decomposition enables efficient parameter sharing across experts while MoE provides dynamic adaptation, collectively enabling PT-MoE to demonstrate cross-task consistency and generalization abilities. These findings, along with ablation studies on routing mechanisms and architectural components, provide insights for future PEFT methods.



Paperid:4897
Authors:Anirudh Bindiganavale Harish, Peikun Guo, Diya Gupta, Bhargav Ghanekar, Akilesh Rajavenkatanarayan, MANOJ SHARMA, Maureen August, Akane Sano, Ashok Veeraraghavan
Abstract:
Remote physiological sensing is an evolving area of research. As systems approach clinical precision, there is increasing focus on complex applications such as cognitive state estimation. Hence, there is a need for large datasets that facilitate research into complex downstream tasks such as remote cognitive load estimation. A first of its kind, our paper introduces an open-source multimodal multi-vital sign dataset consisting of concurrent recordings from RGB, NIR (near-infrared), thermal, and RF (radio-frequency) sensors alongside contact-based physiological signals, such as pulse oximeter and chest bands, providing a benchmark for cognitive state assessment. By adopting a multimodal approach to remote health sensing, our dataset and its associated hardware system excel at modeling the complexities of cognitive load. Here, cognitive load is defined as the mental effort exerted during tasks such as reading, memorizing, and solving math problems. By using the NASA-TLX survey, we set personalized thresholds for defining high/low cognitive levels, enabling a more reliable benchmark. Our benchmarking scheme bridges the gap between existing remote sensing strategies and cognitive load estimation techniques by using vital signs (such as PPG and respiratory waveforms) and physiological signals (blink waveforms) as an intermediary. Through this paper, we focus on replacing the need for intrusive contact-based physiological measurements with more user-friendly remote sensors.



Paperid:4898
Authors:Nate Gillman, Charles Herrmann, Michael Freeman, Daksh Aggarwal, Evan Luo, Deqing Sun, Chen Sun
Abstract:
Recent advances in video generation models have sparked interest in world models capable of simulating realistic environments.While navigation has been well-explored, physically meaningful interactions that mimic real-world forces remain largely understudied. In this work, we investigate using physical forces as a control signal for video generation and propose force prompts which enable users to interact with images through both localized point forces, such as poking a plant, and global wind force fields, such as wind blowing on fabric. We demonstrate that these force prompts can enable videos to respond realistically to physical control signals by leveraging the physical prior in the original pretrained model, without using any 3D asset or physics simulator at inference. The primary challenge of force prompting is the difficulty in obtaining high quality paired force-video training data, both in the real world due to the difficulty of obtaining force signals, and in synthetic data due to limitations in the visual quality and domain diversity of physics simulators. Our key finding is that video generation models cangeneralizeremarkably well when adapted to follow physical force conditioning from videos synthesized by Blender, even with limited demonstrations of few objects (e.g., flying flags, rolling balls, etc.). Our method can generate videos which simulate forces across diverse geometries, settings, and materials. We also try to understand the source of this generalization and perform ablations on the training data that reveal two key elements: visual diversity and the use of specific text keywords during training. Our approach is trained on only around 15k training examples for a single day on four A100 GPUs, and outperforms existing methods on force adherence and physics realism, bringing world models closer to real-world physics interactions. All datasets, code, and model weights will be open-sourced. Video examples can be found at https://sites.google.com/view/force-prompting-neurips2025



Paperid:4899
Authors:Kefei Zhu, Fengshuo Bai, YuanHao Xiang, Yishuai Cai, Xinglin Chen, Ruochong Li, Xingtao Wang, Yaodong Yang, Hao Dong, Xiaopeng Fan, Yuanpei Chen
Abstract:
Dexterous manipulation is critical to advancing robot capabilities in real-world applications, yet diverse and high-quality datasets remain scarce. Existing data collection methods either rely on human teleoperation or require significant human engineering, or are merely limited to grasping, restricting their scalability and generalization. In this paper, we introduce DexFlyWheel, a scalable data generation framework that employs a self-improving cycle to iteratively expand data diversity. Starting from efficient seed demonstrations warmup, our framework expands data diversity via multiple iterations in the self-improving cycle. Each iteration follows a closed-loop pipeline that combines imitation learning, reinforcement learning, rollout trajectory collection, and data augmentation. At each iteration, we first use imitation learning to extract behavioral priors from demonstrations and employ reinforcement learning to enhance generalization. Based on our policy, we rollout trajectories in simulation and then augment these across different environments and objects positions. As iterations progress, our framework generates more diverse data, including various objects, environments, and object positions. Experimental results show that policies trained on our dataset achieve an average success rate of 81.9\% on the challenge test sets, with a real-world transfer success rate of 78.3\% on dual-arm lift tasks. Videos can be found on our project website https://DexFlyWheel.github.io.



Paperid:4900
Authors:Susmit Agrawal, Krishn Vishwas Kher, Saksham Mittal, Swarnim Maheshwari, Vineeth N Balasubramanian
Abstract:
Abstract:Organisms constantly pivot between tasks such as evading predators, foraging, traversing rugged terrain, and socializing, often within milliseconds. Remarkably, they preserve knowledge of once-learned environments sans catastrophic forgetting, a phenomenon neuroscientists hypothesize, is due to a singular neural circuitry dynamically overlayed by neuromodulatory agents such as dopamine and acetylcholine.In parallel, deep learning research addresses analogous challenges via domain generalization ($\textbf{DG}$) and continual learning ($\textbf{CL}$), yet these methods remain siloed, despite the brain’s ability to perform them seamlessly. In particular, prior work has not explored architectures involving associative memories ($\textbf{AM}$s), which are an integral part of biological systems, to jointly address these tasks. We propose Memory-Integrated Reconfigurable Adapters ($\textbf{MIRA}$), a unified framework that integrates Hopfield-style associative memory modules atop a shared backbone. These memory modules store adapter-weight updates as values and retrieve them via learned keys. Associative memory keys are learned post-hoc to index and retrieve an affine combination of stored adapter updates for any given task or domain on a per-sample basis. By varying only the task-specific objectives, we demonstrate that $\textbf{MIRA}$ seamlessly accommodates domain shifts and sequential task exposures under one roof. Empirical evaluations on standard benchmarks confirm that our $\textbf{AM}$-augmented architecture significantly enhances adaptability and retention: in $\textbf{DG}$, $\textbf{MIRA}$ achieves SoTA out-of-distribution accuracy, and in incremental learning settings, it outperforms architectures explicitly designed to handle catastrophic forgetting using generic $\textbf{CL}$ algorithms. Extensive ablation studies validate the necessity of both associative memory storage and post-hoc key learning for robust interpolated retrieval of adapters. By unifying adapter-based modulation with biologically inspired associative memory, $\textbf{MIRA}$ delivers rapid task switching and enduring knowledge retention in a single extensible architecture, charting a path toward more versatile and memory-augmented AI systems.



Paperid:4901
Authors:DEOKJAE LEE, Hyun Oh Song
Abstract:
We study weight-only post-training quantization (PTQ), which quantizes the weights of a large language model (LLM) without retraining, using little or no calibration data. Weight-only PTQ is crucial for reducing the memory footprint and latency of LLM inference, especially in memory-bound, small-batch inference scenarios, such as personalized inference on edge devices. Despite its importance, irregular weight distributions with heavy-tailed outliers in LLMs complicate quantization, recently motivating rotation-based methods that transform weights into near-Gaussian distributions, which are more regular with fewer outliers, thereby reducing quantization error. In this work, we first derive the information-theoretically optimal bit allocation for Gaussianized weights under given bit budgets, revealing that fine-grained fractional-bit quantizers approaching the Gaussian distortion-rate bound are essential to achieve near-optimal quantization performance. To bridge this theoretical insight and practical implementation, we introduce Q-Palette, a versatile collection of fractional-bit quantizers that range from trellis-coded quantizers offering near-optimal distortion to simpler vector and scalar quantizers optimized for faster inference, all efficiently implemented with optimized CUDA kernels across various bitwidths. Furthermore, leveraging Q-Palette as a foundational component, we propose a novel mixed-scheme quantization framework, jointly optimizing quantizer choices and layer fusion decisions given resource constraints.



Paperid:4902
Authors:Ji Won Park, Kyunghyun Cho
Abstract:
Accurately estimatingsemanticaleatoric and epistemic uncertainties in large language models (LLMs) is particularly challenging in free-form question answering (QA), where obtaining stable estimates often requires many expensive generations. We introduce adiversity-steered samplerthat discourages semantically redundant outputs during decoding, covers both autoregressive and masked diffusion paradigms, and yields substantial sample-efficiency gains. The key idea is to inject a continuous semantic-similarity penalty into the model’s proposal distribution using a natural language inference (NLI) model lightly finetuned on partial prefixes or intermediate diffusion states. We debias downstream uncertainty estimates with importance reweighting and shrink their variance with control variates. Across four QA benchmarks, our method matches or surpasses baselines while covering more semantic clusters with the same number of samples. Being modular and requiring no gradient access to the base LLM, the framework promises to serve as a drop-in enhancement for uncertainty estimation in risk-sensitive model deployments.



Paperid:4903
Authors:Shuhong Zheng, Ashkan Mirzaei, Igor Gilitschenski
Abstract:
Current 3D/4D generation methods are usually optimized for photorealism, efficiency, and aesthetics. However, they often fail to preserve the semantic identity of the subject across different viewpoints. Adapting generation methods with one or few images of a specific subject (also known as Personalization or Subject-driven generation) allows generating visual content that align with the identity of the subject. However, personalized 3D/4D generation is still largely underexplored. In this work, we introduce TIRE (Track, Inpaint, REsplat), a novel method for subject-driven 3D/4D generation. It takes an initial 3D asset produced by an existing 3D generative model as input and uses video tracking to identify the regions that need to be modified. Then, we adopt a subject-driven 2D inpainting model for progressively infilling the identified regions. Finally, we resplat the modified 2D multi-view observations back to 3D while still maintaining consistency. Extensive experiments demonstrate that our approach significantly improves identity preservation in 3D/4D generation compared to state-of-the-art methods.



Paperid:4904
Authors:Yao Luan, Ni Mu, Yiqin Yang, Bo Xu, Qing-Shan Jia
Abstract:
Preference-based reinforcement learning (PbRL) bypasses complex reward engineering by learning rewards directly from human preferences, enabling better alignment with human intentions. However, its effectiveness in multi-stage tasks, where agents sequentially perform sub-tasks (e.g., navigation, grasping), is limited bystage misalignment: Comparing segments from mismatched stages, such as movement versus manipulation, results in uninformative feedback, thus hindering policy learning. In this paper, we validate the stage misalignment issue through theoretical analysis and empirical experiments. To address this issue, we proposeSTage-AlIgnedReward learning (STAIR), which first learns a stage approximation based on temporal distance, then prioritizes comparisons within the same stage. Temporal distance is learned via contrastive learning, which groups temporally close states into coherent stages, without predefined task knowledge, and adapts dynamically to policy changes. Extensive experiments demonstrate STAIR's superiority in multi-stage tasks and competitive performance in single-stage tasks. Furthermore, human studies show that stages approximated by STAIR are consistent with human cognition, confirming its effectiveness in mitigating stage misalignment.



Paperid:4905
Authors:Xiwen Wei, Mustafa Munir, Radu Marculescu
Abstract:
Unified Multimodal Generative Models (UMGMs) unify visual understanding and image generation within a single autoregressive framework. However, their ability to continually learn new tasks is severely hindered by catastrophic forgetting, both within a modality (intra-modal) and across modalities (inter-modal). While intra-modal forgetting has been studied in prior continual learning (CL) work, inter-modal forgetting remains largely unexplored. In this paper, we identify and empirically validate this phenomenon in UMGMs and provide a theoretical explanation rooted in gradient conflict between modalities. To address both intra- and inter-modal forgetting, we propose Modality-Decoupled Experts (MoDE), a lightweight and scalable architecture that isolates modality-specific updates to mitigate the gradient conflict and leverages knowledge distillation to prevent catastrophic forgetting and preserve pre-trained capabilities. Unlike previous CL methods that remain modality-coupled and suffer from modality gradient conflict, MoDE explicitly decouples modalities to prevent interference. Experiments across diverse benchmarks demonstrate that MoDE significantly mitigates both inter- and intra-modal forgetting, outperforming prior CL baselines in unified multimodal generation settings.



Paperid:4906
Authors:Qi Liu, Xinhao Zheng, Renqiu Xia, Qinxiang Cao, Junchi Yan
Abstract:
Abstract:Deductive formal problem-solving (D-FPS) enables process-verified, human-aligned problem-solving by implementing deductive solving processes within formal theorem proving (FTP) environments. However, current methods fail to address the misalignment between informal and formal reasoning granularity, and suffer from inefficiency due to backtracking and error propagation. Moreover, the extreme scarcity of formal problem-solution pairs further hinders progress.For the first gap, we propose **HAR** (_**H**ierarchical **A**utoregressive Formal **R**easoner_), a novel reasoning pipeline. HAR decouples informal-aligned drafting and detailed proving, and formulates solution construction as autoregressive generation with per-step feedback. Second, we propose **CoPA** (_**C**hain-**o**f-**P**roxy-**A**utoformalization_), a data generation pipeline that cascades statement autoformalization, proof drafting, and proof search as a proxy autoformalization path.Experiments demonstrate significant improvements: HAR achieves superior performance on FormalMath500 ($15.50\\% \mapsto 43.39\\%$) and MiniF2F-Solving ($21.87\\% \mapsto 55.68\\%$) with lower budget. CoPA shows consistent scalability through expert iteration. Explorations reveal promising directions in formal solution pruning and informal dataset denoising.



Paperid:4907
Authors:Dongnan Gui, Xun Guo, Wengang Zhou, Yan Lu
Abstract:
Generating an interactive visual world from a single image is both challenging and practically valuable, as single-view inputs are easy to acquire and align well with prompt-driven applications such as gaming and virtual reality. This paper introduces a novel unified framework, Image as a World (IaaW), which synthesizes high-quality 360-degree videos from a single image that are both controllable and temporally continuable. Our framework consists of three stages: world initialization, which jointly synthesizes spatially complete and temporally dynamic scenes from a single view; world exploration, which supports user-specified viewpoint rotation; and world continuation, which extends the generated scene forward in time with temporal consistency. To support this pipeline, we design a visual world model based on generative diffusion models modulated with spherical 3D positional encoding and multi-view composition to represent geometry and view semantics. Additionally, a vision-language model (IaaW-VLM) is fine-tuned to produce both global and view-specific prompts, improving semantic alignment and controllability. Extensive experiments demonstrate that our method produces panoramic videos with superior visual quality, minimal distortion and seamless continuation in both qualitative and quantitative evaluations. To the best of our knowledge, this is the first work to generate a controllable, consistent, and temporally expandable 360-degree world from a single image.



Paperid:4908
Authors:Yuanchao Dai, Zhengzhang Hou, Changchun Li, Yuanbo Xu, En Wang, Ximing Li
Abstract:
Positive-Unlabeled (PU) learning refers to a specific weakly-supervised learning paradigm that induces a binary classifier with a few positive labeled instances and massive unlabeled instances. To handle this task, the community has proposed dozens of PU learning methods with various techniques, demonstrating strong potential. In this paper, we conduct a comprehensive study to investigate the basic characteristics of current PU learning methods. We organize them into two fundamental families of PU learning, includingdisambiguation-free empirical risks, which approximate the expected risk of supervised learning, andpseudo-labeling methods, which estimate pseudo-labels for unlabeled instances. First, we make an empirical analysis on disambiguation-free empirical risks such as uPU, nnPU, and DistPU, and suggest a novel risk-consistent set-aware empirical risk from the perspective of aggregate supervision. Second, we make an empirical analysis of pseudo-labeling methods to evaluate the potential of pseudo-label estimation techniques and widely applied generic tricks in PU learning. Finally, based on those empirical findings, we propose a general framework of PU learning by integrating the set-aware empirical risk with pseudo-labeling. Compared with existing PU learning methods, the proposed framework can be a practical benchmark in PU learning.



Paperid:4909
Authors:Yongqi Ding, Lin Zuo, Mengmeng Jing, Kunshan Yang, Pei He, Tonglan Xie
Abstract:
Brain-inspired spiking neural networks (SNNs) promise to be a low-power alternative to computationally intensive artificial neural networks (ANNs), although performance gaps persist. Recent studies have improved the performance of SNNs through knowledge distillation, but rely on large teacher models or introduce additional training overhead. In this paper, we show that SNNs can be naturally deconstructed into multiple submodels for efficient self-distillation. We treat each timestep instance of the SNN as a submodel and evaluate its output confidence, thus efficiently identifying the strong and the weak. Based on this strong and weak relationship, we propose two efficient self-distillation schemes: (1) Strong2Weak: During training, the stronger "teacher" guides the weaker "student", effectively improving overall performance. (2) Weak2Strong: The weak serve as the "teacher", distilling the strong in reverse with underlying dark knowledge, again yielding significant performance gains. For both distillation schemes, we offer flexible implementations such as ensemble, simultaneous, and cascade distillation. Experiments show that our method effectively improves the discriminability and overall performance of the SNN, while its adversarial robustness is also enhanced, benefiting from the stability brought by self-distillation. This ingeniously exploits the temporal properties of SNNs and provides insight into how to efficiently train high-performance SNNs.



Paperid:4910
Authors:Mohammadsaleh Refahi, Mahdi Abavisani, Bahrad Sokhansanj, James R Brown, Gail Rosen
Abstract:
Abstract:Transformers have revolutionized nucleotide sequence analysis, yet capturing long‑range dependencies remains challenging. Recent studies show that autoregressive transformers often exhibit Markovian behavior by relying on fixed-length context windows for next-token prediction. However, standard self-attention mechanisms are computationally inefficient for long sequences due to their quadratic complexity and do not explicitly enforce global transition consistency.We introduce CARMANIA (Context-Aware Regularization with Markovian Integration for Attention-Based Nucleotide Analysis), a self-supervised pretraining framework that augments next-token (NT) prediction with a transition-matrix (TM) loss. The TM loss aligns predicted token transitions with empirically derived n-gram statistics from each input sequence, encouraging the model to capture higher-order dependencies beyond local context. This integration enables CARMANIA to learn organism-specific sequence structures that reflect both evolutionary constraints and functional organization.We evaluate CARMANIA across diverse genomic tasks, including regulatory element prediction, functional gene classification, taxonomic inference, antimicrobial resistance detection, and biosynthetic gene cluster classification. CARMANIA outperforms the previous best long-context model by at least 7\%, matches state-of-the-art on shorter sequences (exceeding prior results on 20/40 tasks while running $\sim$2.5$\times$ faster), and shows particularly strong improvements on enhancer and housekeeping gene classification tasks—including up to a 34\% absolute gain in Matthews correlation coefficient (MCC) for enhancer prediction. The TM loss boosts accuracy in 33 of 40 tasks, especially where local motifs or regulatory patterns drive prediction. This enables more effective modeling of sequence-dependent biological features while maintaining robustness across non-coding and low-signal regions.



Paperid:4911
Authors:Yupu Zhang, Zelin Xu, Tingsong Xiao, Gustavo Seabra, Yanjun Li, Chenglong Li, Zhe Jiang
Abstract:
Abstract:Predicting the binding affinity of protein-ligand complexes plays a vital role in drug discovery. Unfortunately, progress has been hindered by the lack of large-scale and high-quality binding affinity labels. The widely used PDBbind dataset has fewer than 20K labeled complexes. Self-supervised learning, especially graph contrastive learning (GCL), provides a unique opportunity to break the barrier by pre-training graph neural network models based on vast unlabeled complexes and fine-tuning the models on much fewer labeled complexes. However, the problem faces unique challenges, including a lack of a comprehensive unlabeled dataset with well-defined positive/negative complex pairs and the need to design GCL algorithms that incorporate the unique characteristics of such data. To fill the gap, we propose DecoyDB, a large-scale, structure-aware dataset specifically designed for self-supervised GCL on protein–ligand complexes. DecoyDB consists of high-resolution ground truth complexes ($\leq 2.5\text{Å}$) and diverse decoy structures with computationally generated binding poses that range from realistic to suboptimal (negative pairs). Each decoy is annotated with a Root Mean Squared Deviation (RMSD) from the native pose. We further design a customized GCL framework to pre-train graph neural networks based on DecoyDB and fine-tune the models with labels from PDBbind. Extensive experiments confirm that models pre-trained with DecoyDB achieve superior accuracy, label efficiency, and generalizability.



Paperid:4912
Authors:Jiawei Zhang, Ziyuan Liu, Leon Yan, Gen Li, Yuantao Gu
Abstract:
Diffusion-based inverse algorithms have shown remarkable performance across various inverse problems, yet their reliance on numerous denoising steps incurs high computational costs. While recent developments of fast diffusion ODE solvers offer effective acceleration for diffusion sampling without observations, their application in inverse problems remains limited due to the heterogeneous formulations of inverse algorithms and their prevalent use of approximations and heuristics, which often introduce significant errors that undermine the reliability of analytical solvers. In this work, we begin with an analysis of ODE solvers for inverse problems that reveals a linear combination structure of approximations for the inverse trajectory. Building on this insight, we propose a canonical form that unifies a broad class of diffusion-based inverse algorithms and facilitates the design of more generalizable solvers. Inspired by the linear subspace search strategy, we propose Learnable Linear Extrapolation (LLE), a lightweight approach that universally enhances the performance of any diffusion-based inverse algorithm conforming to our canonical form. LLE optimizes the combination coefficients to refine current predictions using previous estimates, alleviating the sensitivity of analytical solvers for inverse algorithms. Extensive experiments demonstrate consistent improvements of the proposed LLE method across multiple algorithms and tasks, indicating its potential for more efficient solutions and boosted performance of diffusion-based inverse algorithms with limited steps.



Paperid:4913
Authors:Tony Bonnaire, Raphaël Urfin, Giulio Biroli, Marc Mezard
Abstract:
Abstract:Diffusion models have achieved remarkable success across a wide range of generative tasks. A key challenge is understanding the mechanisms that prevent their memorization of training data and allow generalization. In this work, we investigate the role of the training dynamics in the transition from generalization to memorization. Through extensive experiments and theoretical analysis, we identify two distinct timescales: an early time $\tau_\mathrm{gen}$ at which models begin to generate high-quality samples, and a later time $\tau_\mathrm{mem}$ beyond which memorization emerges. Crucially, we find that $\tau_\mathrm{mem}$ increases linearly with the training set size $n$, while $\tau_\mathrm{gen}$ remains constant. This creates a growing window of training times with $n$ where models generalize effectively, despite showing strong memorization if training continues beyond it. It is only when $n$ becomes larger than a model-dependent threshold that overfitting disappears at infinite training times.These findings reveal a form of implicit dynamical regularization in the training dynamics, which allow to avoid memorization even in highly overparameterized settings. Our results are supported by numerical experiments with standard U-Net architectures on realistic and synthetic datasets, and by a theoretical analysis using a tractable random features model studied in the high-dimensional limit.



Paperid:4914
Authors:Erel Naor, Ofir Lindenbaum
Abstract:
Deep neural networks often under-perform on tabular data due to their sensitivity to irrelevant features and a spectral bias toward smooth, low-frequency functions. These limitations hinder their ability to capture the sharp, high-frequency signals that often define tabular structure, especially under limited labeled samples. While self-supervised learning (SSL) offers promise in such settings, it remains challenging in tabular domains due to the lack of effective data augmentations. We propose a hybrid autoencoder that combines a neural encoder with an oblivious soft decision tree (OSDT) encoder, each guided by its own stochastic gating network that performs sample-specific feature selection. Together, these structurally different encoders and model-specific gating networks implement model-based augmentation, producing complementary input views tailored to each architecture. The two encoders, trained with a shared decoder and cross-reconstruction loss, learn distinct yet aligned representations that reflect their respective inductive biases. During training, the OSDT encoder (robust to noise and effective at modeling localized, high-frequency structure) guides the neural encoder toward representations more aligned with tabular data. At inference, only the neural encoder is used, preserving flexibility and SSL compatibility. Spectral analysis highlights the distinct inductive biases of each encoder. Our method achieves consistent gains in low-label classification across diverse tabular datasets, outperforming deep and tree-based supervised baselines.



Paperid:4915
Authors:Cécile Rousseau, Tobia Boschi, Giandomenico Cornacchia, Dhaval Salwala, Alessandra Pascale, Juan Moreno
Abstract:
SDForger is a flexible and efficient framework for generating high-quality multivariate time series using LLMs. Leveraging a compact data representation, SDForger provides synthetic time series generation from a few samples and low-computation fine-tuning of any autoregressive LLM. Specifically, the framework transforms univariate and multivariate signals into tabular embeddings, which are then encoded into text and used to fine-tune the LLM. At inference, new textual embeddings are sampled and decoded into synthetic time series that retain the original data's statistical properties and temporal dynamics. Across a diverse range of datasets, SDForger outperforms existing generative models in many scenarios, both in similarity-based evaluations and downstream forecasting tasks. By enabling textual conditioning in the generation process, SDForger paves the way for multimodal modeling and the streamlined integration of time series with textual information.



Paperid:4916
Authors:Xi ruida
Abstract:
Multi-modality object Re-IDentification (ReID) targets to retrieve special objects by integrating complementary information from diverse visual sources.However, existing models that are trained on modality-complete datasets typically exhibit significantly degraded discriminationduring inference with modality-incomplete inputs.This disparity highlights the necessity of developing a robust multi-modality ReID model that remains effective in real-world applications. For that, this paper delivers a flexible framework tailored for more realistic multi-modality retrieval scenario, dubbed as Miss-ReID, which is the first work to friendly support both the modality-missing training and inference conditions. The core of Miss-ReID lies in compensating for missing visual cues via vision-text knowledge transfer driven by Vision-Language foundation Models (VLMs), effectively mitigating performance degradation. In brief, we capture diverse visual features from accessible modalities first, and then build memory banks to store heterogeneous prototypes for each identity, preserving multi-modality characteristics. Afterwards, we employ structure-aware query interactions to dynamically distill modality-invariant object structures from existing localized visual patches, which are further reversed into pseudo-word tokens that encapsulate the identity-relevant structural semantics.In tandem, the inverted tokens, integrated with learnable modality prompts, are embedded into crafted textual template to form the personalized linguistic descriptions tailored for diverse modalities.Ultimately, harnessing VLMs' inherent vision-text alignment capability, the resulting textual features effectively function as compensatory semantic representations for missing visual modalities, after being optimized with some memory-based alignment constraints.Extensive experiments demonstrate our model's efficacy and superiority over state-of-the-art methods in various modality-missing scenarios, and our endeavors further propel multi-modality ReID into real-world applications.



Paperid:4917
Authors:Siwei Wen, junyan ye, Peilin Feng, Hengrui Kang, Zichen Wen, Yize Chen, Jiang Wu, wenjun wu, Conghui He, Weijia Li
Abstract:
With the rapid advancement of Artificial Intelligence Generated Content (AIGC) technologies, synthetic images have become increasingly prevalent in everyday life, posing new challenges for authenticity assessment and detection. Despite the effectiveness of existing methods in evaluating image authenticity and locating forgeries, these approaches often lack human interpretability and do not fully address the growing complexity of synthetic data. To tackle these challenges, we introduce FakeVLM, a specialized large multimodal model designed for both general synthetic image and DeepFake detection tasks. FakeVLM not only excels in distinguishing real from fake images but also provides clear, natural language explanations for image artifacts, enhancing interpretability. Additionally, we present FakeClue, a comprehensive dataset containing over 100,000 images across seven categories, annotated with fine-grained artifact clues in natural language. FakeVLM demonstrates performance comparable to expert models while eliminating the need for additional classifiers, making it a robust solution for synthetic data detection. Extensive evaluations across multiple datasets confirm the superiority of FakeVLM in both authenticity classification and artifact explanation tasks, setting a new benchmark for synthetic image detection. The code and model weights are available at https://anonymous.4open.science/r/nips_2025-2B62.



Paperid:4918
Authors:Santiago Cadena, Andrea Merlo, Emanuel Laude, Alexander Bauer, Atul Agrawal, Maria Pascu, Marija Savtchouk, Lukas Bonauer, Enrico Guiraud, Stuart Hudson, Markus Kaiser
Abstract:
Stellarators are magnetic confinement devices that are being pursued to deliver steady-state carbon-free fusion energy. Their design involves a high-dimensional, constrained optimization problem that requires expensive physics simulations and significant domain expertise. Recent advances in plasma physics and open- source tools have facilitated the process of stellarator optimization. However, wide-spread community progress is currently bottlenecked by the lack of standardized optimization problems with strong baselines and datasets that enable data-driven approaches, particularly for quasi-isodynamic (QI) stellarator configurations, seen as a promising path to commercial fusion due to their intrinsic avoidance of current-driven disruptions. Here, we release an open dataset of diverse QI-like stellarator plasma boundary shapes, paired with their ideal magnetohydrodynamic (MHD) equilibria and performance metrics. We generated this dataset by sampling diverse QI fields and optimizing stellarator plasma boundaries to target key properties. We propose three optimization benchmarks of varying complexity involving a single-objective constrained geometric problem; a “simple-to-build” single-objective QI stellarator; and a multi-objective QI stellarator that is also MHD stable. For every benchmark, we provide reference code, evaluation scripts, and strong baselines based on classical optimization methods. By openly releasing the dataset (https://huggingface.co/datasets/proxima-fusion/ constellaration) together with benchmark problems and baselines (https://github.com/proximafusion/constellaration), we aim to lower the barrier for optimization and machine learning researchers to contribute to stellarator design, and to accelerate cross-disciplinary progress toward bringing fusion energy to the grid.



Paperid:4919
Authors:Dipendra Misra, Aldo Pacchiano, Ta-Chung Chi, Ge Gao
Abstract:
We study how to fine-tune LLMs using user-edit deployment data consisting of a set of context, an agent's response, and user edits. This deployment data is naturally generated by users in applications such as LLMs-based writing assistants and coding agents. Thenaturalorigin of user edits makes it a desired source for adapting and personalizing of LLMs. In this setup, there emerges a unification of various feedback types namely preferences, supervised labels, and cost that are typically studied separately in the literature. In this paper, we initiate the theoretical investigation of learning from user edits. We first derive bounds for learning algorithms that learn from each of these feedback types. We prove that these algorithms have different trade-offs depending upon the user, data distribution, and model class. We then propose a simple ensembling procedure to jointly learn from these feedback types. On two domains from Gao et al. 2024, we show our ensembling procedure outperforms these methods that learn from individual feedback. Further, we show that our proposed procedure can robustly adapt to different user-edit distributions at test time.



Paperid:4920
Authors:Ziquan Wei, Tingting Dan, Tianlong Chen, Guorong Wu
Abstract:
Given the large scale of public functional Magnetic Resonance Imaging (fMRI), e.g., UK Biobank (UKB) and Human Connectome Projects (HCP), brain foundation models are emerging. Although the amount of samples under rich environmental variables is unprecedented, existing brain foundation models learn from fMRI derived from a narrow range of cognitive states stimulated by similar environments, causing the limited robustness demonstrated in various applications and datasets acquired with different pipelines and limited sample size. By capitalizing on the variety of cognitive status as subjects performing explicit tasks, we present the mixture of brain experts, namely BrainMoE, pre-training on tasking fMRI with rich behavioral tasks in addition to resting fMRI for a robust brain foundation model. Brain experts are designed to produce embeddings for different behavioral tasks related to cognition. Afterward, these cognition embeddings are mixed by a cognition adapter via cross-attention so that BrainMoE can handle orthogonal embeddings and be robust on those boutique downstream datasets. We have pre-trained two existing self-regressive architectures and one new supervised architecture as brain experts on 68,251 fMRI scans among UKB and HCP, containing 12 different cognitive states. Then, BrainMoE is evaluated on a variety of applications, including sex prediction, human behavior recognition, and disease early diagnosis of Autism, Parkinson's disease, Alzheimer's disease, and Schizophrenia, where promising results in seven datasets from three different pipelines indicate great potential to facilitate current neuroimaging applications in clinical routines.



Paperid:4921
Authors:Zhichao Sun, Huazhang Hu, Yidong Ma, Gang Liu, Yibo Chen, Xu Tang, Yao Hu, Yongchao Xu
Abstract:
With the exponential growth of data, traditional object detection methods are increasingly struggling to handle vast vocabulary object detection tasks effectively. We analyze two key limitations of classification-based detectors: positive gradient dilution, where rare positive categories receive insufficient learning signals, and hard negative gradient dilution, where discriminative gradients are overwhelmed by numerous easy negatives. To address these challenges, we propose CQ-DINO, a category query-based object detection framework that reformulates classification as a contrastive task between object queries and learnable category queries. Our method introduces image-guided query selection, which reduces the negative space by adaptively retrieving top-K relevant categories per image via cross-attention, thereby rebalancing gradient distributions and facilitating implicit hard example mining.Furthermore, CQ-DINO flexibly integrates explicit hierarchical category relationships in structured datasets (e.g., V3Det) or learns implicit category correlations via self-attention in generic datasets (e.g., COCO). Experiments demonstrate that CQ-DINO achieves superior performance on the challenging V3Det benchmark (surpassing previous methods by 2.1% AP) while maintaining competitiveness in COCO. Our work provides a scalable solution for real-world detection systems requiring wide category coverage. The code is available in the supplemental material.



Paperid:4922
Authors:Wenqi Liu, Xuemeng Song, Jiaxi Li, Yinwei Wei, Na Zheng, Jianhua Yin, Liqiang Nie
Abstract:
Direct Preference Optimization (DPO) has emerged as an effective approach for mitigating hallucination in Multimodal Large Language Models (MLLMs). Although existing methods have achieved significant progress by utilizing vision-oriented contrastive objectives for enhancing MLLMs' attention to visual inputs and hence reducing hallucination, they suffer from non-rigorous optimization objective function and indirect preference supervision. To address these limitations, we propose a Symmetric Multimodal Preference Optimization (SymMPO), which conducts symmetric preference learning with direct preference supervision (i.e., response pairs) for visual understanding enhancement, while maintaining rigorous theoretical alignment with standard DPO. In addition to conventional ordinal preference learning, SymMPO introduces a preference margin consistency loss to quantitatively regulate the preference gap between symmetric preference pairs. Comprehensive evaluation across five benchmarks demonstrate SymMPO's superior performance, validating its effectiveness in hallucination mitigation of MLLMs.



Paperid:4923
Authors:Ishika Agarwal, Dilek Tur
Abstract:
Influence functions provide crucial insights into model training, but existing methods suffer from large computational costs and limited generalization. Particularly, recent works have proposed various metrics and algorithms to calculate the influence of data using language models, which do not scale well with large models and datasets. This is because of the expensive forward and backward passes required for computation, substantial memory requirements to store large models, and poor generalization of influence estimates to new data. In this paper, we explore the use of small neural networks -- which we refer to as the InfluenceNetwork -- to estimate influence values, achieving up to 99% cost reduction. Our evaluation demonstrates that influence values can be estimated with models just 0.0007% the size of full language models (we average across 1.5B-22B versions). We apply our algorithm of estimating influence values (called NN-CIFT: Neural Networks for effiCient Instruction Fine-Tuning) to the downstream task of subset selection for general instruction fine-tuning. In our study, we include four state-of-the-art influence functions and show no compromise in performance, despite large speedups, between NN-CIFT and the original influence functions. We provide an in-depth hyperparameter analyses of NN-CIFT. The code for our method can be found here: https://anonymous.4open.science/r/NN-CIFT-B065/.



Paperid:4924
Authors:Jianle Sun, Chaoqi Liang, Ran Wei, Peng Zheng, LEI BAI, Wanli Ouyang, Hongliang Yan, Peng Ye
Abstract:
Abstract:Advances in single-cell sequencing have enabled high-resolution profiling of diverse molecular modalities, while integrating unpaired multi-omics single-cell data remains challenging. Existing approaches either rely on pair information or prior correspondences, or require computing a global pairwise coupling matrix, limiting their scalability and flexibility. In this paper, we introduce a scalable and flexible generative framework called single-cell Multi-omics Regularized Disentangled Representations (scMRDR) for unpaired multi-omics integration. Specifically, we disentangle each cell’s latent representations into modality-shared and modality-specific components using a well-designed $\beta$-VAE architecture, which are augmented with isometric regularization to preserve intra-omics biological heterogeneity, adversarial objective to encourage cross-modal alignment, and masked reconstruction loss strategy to address the issue of missing features across modalities. Our method achieves excellent performance on benchmark datasets in terms of batch correction, modality alignment, and biological signal preservation. Crucially, it scales effectively to atlas-level datasets and supports integration of more than two omics, offering a powerful and flexible solution for large-scale multi-omics data integration and downstream biological discovery.



Paperid:4925
Authors:Dingqiang Ye, Chao Fan, Zhanbo Huang, Chengwen Luo, Jianqiang Li, Shiqi Yu, Xiaoming Liu
Abstract:
Large vision models (LVM) based gait recognition has achieved impressive performance.However, existing LVM-based approaches may overemphasize gait priors while neglecting the intrinsic value of LVM itself, particularly the rich, distinct representations across its multi-layers. To adequately unlock LVM's potential, this work investigates the impact of layer-wise representations on downstream recognition tasks.Our analysis reveals that LVM's intermediate layers offer complementary properties across tasks, integrating them yields an impressive improvement even without rich well-designed gait priors.Building on this insight, we propose a simple and universal baseline for LVM-based gait recognition, termed BiggerGait.Comprehensive evaluations on CCPG, CAISA-B*, SUSTech1K, and CCGR_MINI validate the superiority of BiggerGait across both within- and cross-domain tasks, establishing it as a simple yet practical baseline for gait representation learning.All the models and code will be publicly available.



Paperid:4926
Authors:Bowen Chen, Brynn zhao, Haomiao Sun, Li Chen, Xu Wang, Daniel Du, Xinglong Wu
Abstract:
Achieving fine-grained control over subject identity and semantic attributes (pose, style, lighting) in text-to-image generation, particularly for multiple subjects, often undermines the editability and coherence of Diffusion Transformers (DiTs). Many approaches introduce artifacts or suffer from attribute entanglement. To overcome these challenges, we propose a novel multi-subject controlled generation model XVerse. By transforming reference images into offsets for token-specific text-stream modulation, XVerse allows for precise and independent control for specific subject without disrupting image latents or features. Consequently, XVerse offers high-fidelity, editable multi-subject image synthesis with robust control over individual subject characteristics and semantic attributes. This advancement significantly improves personalized and complex scene generation capabilities.



Paperid:4927
Authors:Hengyuan Cao, Yutong Feng, Biao Gong, Yijing Tian, Yunhong Lu, Chuang Liu, Bin Wang
Abstract:
Video generative models can be regarded as world simulators due to their ability to capture dynamic, continuous changes inherent in real-world environments. These models integrate high-dimensional information across visual, temporal, spatial, and causal dimensions, enabling predictions of subjects in various status. A natural and valuable research direction is to explore whether a fully trained video generative model in high-dimensional space can effectively support lower-dimensional tasks such as controllable image generation. In this work, we propose a paradigm for video-to-image knowledge compression and task adaptation, termed \textit{Dimension-Reduction Attack} (\texttt{DRA-Ctrl}), which utilizes the strengths of video models, including long-range context modeling and flatten full-attention, to perform various generation tasks. Specially, to address the challenging gap between continuous video frames and discrete image generation, we introduce a mixup-based transition strategy that ensures smooth adaptation. Moreover, we redesign the attention structure with a tailored masking mechanism to better align text prompts with image-level control. Experiments across diverse image generation tasks, such as subject-driven and spatially conditioned generation, show that repurposed video models outperform those trained directly on images. These results highlight the untapped potential of large-scale video generators for broader visual applications. \texttt{DRA-Ctrl} provides new insights into reusing resource-intensive video models and lays foundation for future unified generative models across visual modalities.



Paperid:4928
Authors:Yarden Bakish, Itamar Zimerman, Hila Chefer, Lior Wolf
Abstract:
The development of effective explainability tools for Transformers is a crucial pursuit in deep learning research. One of the most promising approaches in this domain is Layer-wise Relevance Propagation (LRP), which propagates relevance scores backward through the network to the input space by redistributing activation values based on predefined rules. However, existing LRP-based methods for Transformer explainability entirely overlook a critical component of the Transformer architecture: its positional encoding (PE), resulting in violations of conservation, and the loss of an important and unique type of relevance, which is also associated with structural and positional features. To address this limitation, we reformulate the input space for Transformer explainability as a set of position-token pairs, rather than relying solely on the standard vocabulary space. This allows us to propose specialized theoretically-grounded LRP rules designed to propagate attributions across various positional encoding methods, including Rotary, Learned, and Absolute PE. Extensive experiments with both fine-tuned classifiers and zero-shot foundation models, such as LLaMA 3, demonstrate that our method significantly outperforms the SoTA in both vision and NLP explainability tasks. Our code is provided as a supplement.



Paperid:4929
Authors:Sara Cammarota, Matteo Ferrante, Nicola Toschi
Abstract:
Abstract:Decoding of visual stimuli from noninvasive neuroimaging techniques such as functional magnetic resonance has advanced rapidly in the last years; yet, most high-performing brain decoding models rely on complicated, non intelligible latent spaces. In this study we present an interpretable brain decoding framework that inserts a semantic bottleneck into BrainDiffuser, a well established, simple and linear decoding pipeline. We firstly produce a $214-\text{dimensional}$ binary interpretable space $\mathcal{L}$ for images, in which each dimension answers to a specific question about the image (e.g., "Is there a person?", "Is it outdoors?"). A first ridge regression maps voxel activity to this semantic space. Because this mapping is linear, its weight matrix can be visualized as maps of voxel importance for each dimension of $\mathcal{L}$, revealing which cortical regions influence mostly each semantic dimension. A second regression then transforms these concept vectors into CLIP embeddings required to produce the final decoded image, conditioning the BrainDiffuser model. We found that voxel-wise weight maps for individual questions are highly consistent with canonical category-selective regions in the visual cortex (face, bodies, places, words), simultaneously revealing that activation distributions, not merely location, bear semantic meaning in the brain. Visual brain decoding performances are only slightly lower compared to the original BrainDiffuser metrics (e.g., the CLIP similarity is decreased by $\leq 4$% for the four subjects), yet offering substantial gains in interpretability and neuroscientific insights. These results show that our interpretable brain decoding pipeline enables voxel-level analysis of semantic representations in the human brain without sacrificing decoding accuracy.



Paperid:4930
Authors:Hanyang Li, Yuheng Jia, Hui LIU, Junhui Hou
Abstract:
Abstract:Recent deep clustering models have produced impressive clustering performance. However, a common issue with existing methods is the disparity between global and local feature structures. While local structures typically show strong consistency and compactness within class samples, global features often present intertwined boundaries and poorly separated clusters. Motivated by this observation, we propose **DCBoost**, a parameter-free plug-in designed to enhance the global feature structures of current deep clustering models. By harnessing reliable local structural cues, our method aims to elevate clustering performance effectively. Specifically, we first identify high-confidence samples through adaptive $k$-nearest neighbors-based consistency filtering, aiming to select a sufficient number of samples with high label reliability to serve as trustworthy anchors for self-supervision. Subsequently, these samples are utilized to compute a discriminative loss, which promotes both intra-class compactness and inter-class separability, to guide network optimization. Extensive experiments across various benchmark datasets showcase that our DCBoost significantly improves the clustering performance of diverse existing deep clustering models. Notably, our method improves the performance of current state-of-the-art baselines (e.g., ProPos) by more than 3\% and amplifies the silhouette coefficient by over $7\times$. **The source code is included in the Supplementary Material.**



Paperid:4931
Authors:Sunay Joshi, Shayan Kiyani, George J. Pappas, Edgar Dobriban, Hamed Hassani
Abstract:
We consider the problem of conformal prediction under covariate shift. Given labeled data from a source domain and unlabeled data from a covariate shifted target domain, we seek to construct prediction sets with valid marginal coverage in the target domain. Most existing methods require estimating the unknown likelihood ratio function, which can be prohibitive for high-dimensional data such as images. To address this challenge, we introduce the likelihood ratio regularized quantile regression (LR-QR) algorithm, which combines the pinball loss with a novel choice of regularization in order to construct a threshold function without directly estimating the unknown likelihood ratio. We show that the LR-QR method has coverage at the desired level in the target domain, up to a small error term that we can control. Our proofs draw on a novel analysis of coverage via stability bounds from learning theory. Our experiments demonstrate that the LR-QR algorithm outperforms existing methods on high-dimensional prediction tasks, including a regression task for the Communities and Crime dataset, an image classification task from the WILDS repository, and an LLM question-answering task on the MMLU benchmark.



Paperid:4932
Authors:Hoang Son Nguyen, Xiao Fu
Abstract:
Latent component identification from unknownnonlinearmixtures is a foundational challenge in machine learning, with applications in tasks such as disentangled representation learning and causal inference. Prior work innonlinear independent component analysis(nICA) has shown that auxiliary signals---such as weak supervision---can supportidentifiabilityof conditionally independent latent components. More recent approaches explore structural assumptions, like sparsity in the Jacobian of the mixing function, to relax such requirements. In this work, we introduceDiverse Influence Component Analysis(DICA), a framework that exploits the convex geometry of the mixing function’s Jacobian. We propose aJacobian Volume Maximization(J-VolMax) criterion, which enables latent component identification by encouraging diversity in their influence on the observed variables. Under suitable conditions, this approach achieves identifiability without relying on auxiliary information, latent component independence, or Jacobian sparsity assumptions. These results extend the scope of identifiability analysis and offer a complementary perspective to existing methods.



Paperid:4933
Authors:Run Luo, Renke Shan, Longze Chen, Ziqiang Liu, Lu Wang, Min Yang, Xiaobo Xia
Abstract:
Large vision-language models (LVLMs) have emerged as foundational tools for real-world AI applications. Despite their remarkable capabilities, current LVLMs process entire images at the token level, leading to significant inefficiencies compared to human cognition, which selectively focuses on high-level vision concepts. This token-level redundancy becomes increasingly problematic for high-resolution images and long video sequences, resulting in large computational costs and limited scalability in practical applications. To address this limitation, we introduce the concept of a vision concept model, a novel paradigm that enables LVLMs to dynamically extract the most relevant vision concepts from complex inputs, based on task-specific instructions. To optimize this vision concept modeling process, we propose VCM, a self-supervised framework that leverages vision-language correlations across diverse instances. VCM is designed to learn meaningful vision concepts without the need for expensive concept-level annotations. At its core, it employs a forward-backward optimization algorithm that supports LVLMs to adjust concept granularity and spatial alignment dynamically. Experiments demonstrate that VCM remarkably reduces computational costs (e.g., achieving up to 85\% fewer FLOPs for LLaVA-1.5-7B), while maintaining strong performance across a series of vision-language tasks. The source code will be publicly available.



Paperid:4934
Authors:Yuyang Wanyan, Xi Zhang, Haiyang Xu, Haowei Liu, Junyang Wang, Jiabo Ye, Yutong Kou, Ming Yan, Fei Huang, Xiaoshan Yang, Weiming Dong, Changsheng Xu
Abstract:
In recent years, Multimodal Large Language Models (MLLMs) have been extensively utilized for multimodal reasoning tasks, including Graphical User Interface (GUI) automation. Unlike general offline multimodal tasks, GUI automation is executed in online interactive environments, necessitating step-by-step decision-making based on real-time status of the environment. This task has a lower tolerance for decision-making errors at each step, as any mistakes may cumulatively disrupt the process and potentially lead to irreversible outcomes like deletions or payments. To address these issues, we introduce a pre-operative critic mechanism that provides effective feedback prior to the actual execution, by reasoning about the potential outcome and correctness of actions. Specifically, we propose a Suggestion-aware Gradient Relative Policy Optimization (S-GRPO) strategy to construct our pre-operative critic model GUI-Critic-R1, incorporating a novel suggestion reward to enhance the reliability of the model's feedback. Furthermore, we develop a reasoning-bootstrapping based data collection pipeline to create a GUI-Critic-Train and a GUI-Critic-Test, filling existing gaps in GUI critic data. Static experiments on the GUI-Critic-Test across both mobile and web domains reveal that our GUI-Critic-R1 offers significant advantages in critic accuracy compared to current MLLMs. Dynamic evaluation on GUI automation benchmark further highlights the effectiveness and superiority of our model, as evidenced by improved success rates and operational efficiency.



Paperid:4935
Authors:Samuel Singh, Shirley Coyle, Mimi Zhang
Abstract:
We introduce FAEclust, a novel functional autoencoder framework for cluster analysis of multi-dimensional functional data, data that are random realizations of vector-valued random functions. Our framework features a functional encoder that captures complex nonlinear interdependencies among component functions and a universal approximator decoder capable of accurately reconstructing both Euclidean and manifold-valued functional data. Stability and robustness are enhanced through innovative regularization strategies applied to functional weights and biases. Additionally, we incorporate a clustering loss into the network's training objective, promoting the learning of latent representations that are conducive to effective clustering. A key innovation is our shape-informed clustering objective, ensuring that the clustering results are resistant to phase variations in the functions. We establish the universal approximation property of our non-linear decoder and validate the effectiveness of our model through extensive experiments.



Paperid:4936
Authors:Simone Azeglio, Olivier Marre, Peter Neri, Ulisse Ferrari
Abstract:
We propose a novel enhancement to Convolutional Neural Networks (CNNs) by incorporating learnable higher-order convolutions inspired by nonlinear biological visual processing. Our model extends the classical convolution operator using a Volterra-like expansion to capture multiplicative interactions observed in biological vision. Through extensive evaluation on standard benchmarks and synthetic datasets, we demonstrate that our architecture consistently outperforms traditional CNN baselines, achieving optimal performance with 3rd/4th order expansions. Systematic perturbation analysis and Representational Similarity Analysis reveal that different orders of convolution process distinct aspects of visual information, aligning with the statistical properties of natural images. This biologically-inspired approach offers both improved performance and deeper insights into visual information processing.



Paperid:4937
Authors:Bowen Song, Zecheng Zhang, Zhaoxu Luo, Jason Hu, Wei Yuan, Jing Jia, Zhengxu Tang, Guanyang Wang, Liyue Shen
Abstract:
Diffusion models have emerged as powerful tools for generative tasks, producing high-quality outputs across diverse domains. However, how the generated data responds to the initial noise perturbation in diffusion models remains under-explored, which hinders understanding the controllability of the sampling process. In this work, we first observe an interesting phenomenon: the relationship between the change of generation outputs and the scale of initial noise perturbation is highly linear through the diffusion ODE sampling. Then we provide both theoretical and empirical study to justify this linearity property of this input-output (noise-generation data) relationship. Inspired by these new insights, we propose a novel Controllable and Constrained Sampling method (CCS) together with a new controller algorithm for diffusion models to sample with desired statistical properties while preserving good sample quality. We perform extensive experiments to compare our proposed sampling approach with other methods on both sampling controllability and sampled data quality. Results show that our CCS method achieves more precisely controlled sampling while maintaining superior sample quality and diversity.



Paperid:4938
Authors:Antonios Valkanas, Soumyasundar Pal, Pavel Rumiantsev, Yingxue Zhang, Mark Coates
Abstract:
Large language models (LLMs) have achieved impressive results on complex reasoning tasks, but their high inference cost remains a major barrier to real-world deployment. A promising solution is to use cascaded inference, where small, cheap models handle easy queries, and only the hardest examples are escalated to more powerful models. However, existing cascade methods typically rely on supervised training with labeled data, offer no theoretical generalization guarantees, and provide limited control over test-time computational cost. We introduceC3PO(Cost Controlled Cascaded Prediction Optimization), a self-supervised framework for optimizing LLM cascades under probabilistic cost constraints. By focusing on minimizing regret with respect to the most powerful model, C3PO avoids the need for labeled data by constructing a cascade using only unlabeled model outputs. It leverages conformal prediction to bound the probability that inference cost exceeds a user-specified budget. We provide theoretical guarantees on both cost control and generalization error, and show that our optimization procedure is effective even with small calibration sets. Empirically, C3PO achievesstate-of-the-artperformance across a diverse set of reasoning benchmarks—including GSM8K, MATH500, and BigBench-Hard—outperforming strong LLM cascading baselines in both accuracy and cost-efficiency. Our results demonstrate that principled, label-free cascade optimization can enable scalable and reliable LLM deployment.



Paperid:4939
Authors:Alistair Benford, Per Kristian Lehre
Abstract:
Most methods for finding a Nash equilibrium rely on procedures that operate over the entire action space, making them infeasible for settings with too many actions to be searched exhaustively. Randomised search heuristics such as coevolutionary algorithms offer benefits in such settings, however they lack many of the theoretical guarantees established for exhaustive methods such as zero-regret learning. We address this by developing a method for proving necessary and sufficient conditions for a coevolutionary algorithm to be stable, in the sense that it reliably retains a Nash equilibrium following discovery. As the method provides bounds that are adapted to both application and algorithm instance, it can be used as a practical tool for parameter configuration. We additionally show how bounds on regret may be deduced from our results and undertake corresponding empirical analysis.



Paperid:4940
Authors:Zhuang Qi, Yu Pan, Lei Meng, Sijin Zhou, Han Yu, Xiaoxiao Li, Xiangxu Meng
Abstract:
Federated Prompt Learning (FPL) enables communication-efficient adaptation by tuning lightweight prompts on top of frozen pre-trained models. Existing FPL methods typically rely on global information, which is only available after the second training round, to facilitate collaboration among client models. Therefore, they are inherently dependent on multi-round communication to fully exhibit their strengths. Moreover, existing one-shot federated learning methods typically focus on fitting seen tasks, but lack cross-task generalization. To bridge this gap, we propose the global prompt refinement with non-interfering attention masking (GPR-NIAM) method for one-shot FPL. The core idea is to design a masking mechanism that restricts excessive interaction between the original text embeddings and the learnable prompt embeddings. GPR-NIAM achieves this through the collaboration of two key modules. Firstly, the attention isolation module suppresses attention from the learnable prompt tokens to the original text tokens, and reweights the reverse attention which preserves generalization across tasks. Secondly, the cross-silo collaborative refinement module integrates decentralized visual knowledge into a unified base and calibrates the global prompt through multi-source cross-modal knowledge alignment, further mitigating the inconsistency caused by data heterogeneity. Extensive experiments conducted on ten benchmark datasets under two tasks show that GPR-NIAM outperforms eight state-of-the-art methods in both class-level and domain-level generalization.



Paperid:4941
Authors:Tianhe Wu, Jian Zou, Jie Liang, Lei Zhang, Kede Ma
Abstract:
DeepSeek-R1 has demonstrated remarkable effectiveness in incentivizing reasoning and generalization capabilities of large language models (LLMs) through reinforcement learning. Nevertheless, the potential of reasoning-induced computational modeling has not been thoroughly explored in the context of image quality assessment (IQA), a task critically dependent on visual reasoning. In this paper, we introduce VisualQuality-R1, a reasoning-induced no-reference IQA (NR-IQA) model, and we train it with reinforcement learning to rank, a learning algorithm tailored to the intrinsically relative nature of visual quality. Specifically, for a pair of images, we employ group relative policy optimization to generate multiple quality scores for each image. These estimates are then used to compute comparativeprobabilities of one image having higher quality than the other under the Thurstone model. Rewards for each quality estimate are defined using continuous fidelity measures rather than discretized binary labels. Extensive experiments show that the proposed VisualQuality-R1 consistently outperforms discriminative deep learning-based NR-IQA models as well as a recent reasoning-induced quality regression method. Moreover, VisualQuality-R1 is capable of generating contextually rich, human-aligned quality descriptions, and supportsmulti-dataset training without requiring perceptual scale realignment. These features make VisualQuality-R1 especially well-suited for reliably measuring progress in a wide range of image processing tasks like super-resolution and image generation. Code will be made publicly available.



Paperid:4942
Authors:Zhoujun Cheng, Shibo Hao, Tianyang Liu, Fan Zhou, Feng Yao, Yuexin Bian, Yutao Xie, Nilabjo Dey, Yonghao Zhuang, Yuheng Zha, Yi Gu, Kun Zhou, Yuqi Wang, Yuan Li, Richard Fan, Jianshu She, Chengqian Gao, Abulhair Saparov, Taylor W. Killian, Haonan Li, Mikhail Yurochkin, Eric Xing, Zhengzhong Liu, Zhiting Hu
Abstract:
Reinforcement learning (RL) has shown promise in enhancing large language model (LLM) reasoning, yet progress towards broader capabilities is limited by the availability of high-quality, multi-domain datasets. This work introduces \ours, a 92K RL-for-reasoning dataset designed to address this gap, covering six reasoning domains: Math, Code, Science, Logic, Simulation, and Tabular, each with corresponding verifiers. We build \ours via a careful data-curation pipeline, including sourcing, deduplication, reward design, and domain-specific and difficulty-based filtering, to facilitate the systematic investigation of cross-domain RL generalization. Our study using \ours suggests the efficacy of a simple mixed-domain RL training approach and reveals several key aspects affecting cross-domain transferability. We further train two models {\ours}-7B and {\ours}-32B purely with RL on our curated data and observe largely improved performance over leading open RL reasoning model baselines, with gains of 7.3\% and 7.8\% respectively on an extensive 17-task, six-domain evaluation suite. We are releasing our dataset, code, and evaluation suite to the community, aiming to support further research and development of more general RL-enhanced reasoning models.



Paperid:4943
Authors:Zinan Zheng, Yang Liu, Jia Li
Abstract:
Graph neural networks have shown promising results in weather forecasting, which are critical for human activity such as agriculture planning and extreme weather preparation. However, most studies focus on finite and local areas for training, overlooking the influence of broader areas and limiting their ability to generalize effectively. Thus, in this work, we study global weather forecasting that is irregularly distributed and dynamically varying in practice, requiring the model to generalize to unobserved locations.To address such challenges, we propose a general \textbf{M}esh \textbf{I}nterpolation \textbf{G}raph \textbf{N}etwork (MIGN) that models the irregular weather station forecasting, consisting of two key designs: (1) learning spatially irregular data with regular mesh interpolation network to align the data; (2) leveraging parametric spherical harmonics location embedding to further enhance spatial generalization ability. Extensive experiments on up-to-date observation dataset show that MIGN significantly outperforms existing data-driven models. Besides, we show that MIGN has spatial generalization ability, and is capable of generalizing to previous unseen stations.



Paperid:4944
Authors:Arnav Mehra, Alexandros Psomas
Abstract:
Abstract:We consider the classic cake-cutting problem of producing fair allocations for $n$ agents, in the Robertson–Webb query model. In this model, it is known that: (i) proportional allocations can be computed using $O(n \log n)$ queries, and this is optimal for deterministic protocols; (ii) envy-free allocations (a subset of proportional allocations) can be computed using $O\left( n^{n^{n^{n^{n^{n}}}}} \right)$ queries, and the best known lower bound is $\Omega(n^2)$; (iii) perfect allocations (a subset of envy-free allocations) cannot be computed using a bounded (in $n$) number of queries.In this work, we introduce two hierarchies of new fairness notions: Complement Harmonically Bounded (CHB) and Complement Linearly Bounded (CLB). An allocation is CHB-$k$, if for any subset of agents $S$ of size at most $k$, and every agent $i \in S$, the value of $i$ for the union of all pieces allocated to agents outside of $S$ is at most $\frac{n-|S|}{n-|S|+1}$; for CLB-$k$ allocations, the upper bound becomes $\frac{n-|S|}{n}$. Intuitively, these notions of fairness ask that, for every agent $i$, the collective value that a group of agents has (from the perspective of agent $i$) is limited. Our hierarchies bridge the gap between proportionality, envy-freeness, and super envy-freeness.CHB-$k$ and CLB-$k$ coincide with proportionality for $k=1$. For all $k \leq n$, CHB-$k$ allocations are a superset of envy-free allocations (i.e., easier to find). On the other hand, for $k \in [2, \lceil n/2 \rceil - 1]$, CLB-$k$ allocations are incomparable to envy-free allocations. For $k \geq \lceil n/2 \rceil$, CLB-$k$ allocations are a subset of envy-free allocations (i.e., harder to find), while CLB-$n$ coincides with super envy-freeness: the value of each agent for her piece is at least $1/n$, and her value for the piece allocated to any other agent is at most $1/n$. We prove that CHB-$n$ allocations can be computed using $O(n^4)$ queries in the Robertson–Webb model. On the flip side, finding CHB-$2$ (and therefore all CHB-$k$ for $k \geq 2$) allocations requires $\Omega(n^2)$ queries, while CLB-$2$ (and therefore all CLB-$k$ for $k \geq 2$) allocations cannot be computed using a bounded (in $n$) number of queries. Our results reveal that envy-free allocations occupy a curious middle ground, between a computationally impossible notion of fairness, CLB-$\lceil n/2 \rceil$, and a computationally ``easy'' notion, CHB-$n$.



Paperid:4945
Authors:Jian Hu, Zixu Cheng, Shaogang Gong, Isabel Guan, Jianye Hao, Jun Wang, Kun Shao
Abstract:
Video Temporal Grounding (TG) aims to temporally locate video segments matching a natural language description (a query) in a long video. While Vision-Language Models (VLMs) are effective at holistic semantic matching, they often struggle with fine-grained temporallocalisation. Recently, Group Relative Policy Optimisation (GRPO) reformulates the inference process as a reinforcement learning task, enabling fine-grained grounding and achieving strong in-domain performance. However, GRPO relies on labelled data, making it unsuitable in unlabelled domains. Moreover, because videos are large and expensive to store and process, performing full-scale adaptation introduces prohibitive latency and computational overhead, making it impractical for real-time deployment. To overcome both problems, we introduce a Data-Efficient Unlabelled Cross-domain Temporal Grounding method, from which a model is first trained on a labelled source domain, then adapted to a target domain using only a small number of {\em unlabelled videos from the target domain}. This approach eliminates the need for target annotation and keeps both computational and storage overhead low enough to run in real time. Specifically, we introduce \textbf{U}ncertainty-quantified \textbf{R}ollout \textbf{P}olicy \textbf{A}daptation (\textbf{URPA}) for cross-domain knowledge transfer in learning video temporal grounding without target labels. URPA generates multiple candidate predictions using GRPO rollouts, averages them to form a pseudo label, and estimates confidence from the variance across these rollouts. This confidence then weights the training rewards, guiding the model to focus on reliable supervision. Experiments on three datasets across six cross-domain settings show that URPA generalises well using only a few unlabelled target videos. Codes are given in supplemental materials.



Paperid:4946
Authors:Wensong Bai, Chufan Chen, Yichao Fu, Qihang Xu, zhang chao, Hui Qian
Abstract:
Recent advancements in offline reinforcement learning (RL) have underscored the capabilities of conditional sequence modeling (CSM), a paradigm that models the action distribution conditioned on both historical trajectory and target return associated with each state. However, due to imbalanced return distribution caused by suboptimal datasets, CSM is grappling with a serious distributional shift problem when conditioning on high returns. While recent approaches attempt to empirically tackle this challenge through return rebalancing techniques such as weighted sampling and value-regularized supervision, the relationship between return rebalancing and the performance of CSM methods is not well understood. In this paper, we reveal that both expert-level and full-spectrum return-coverage critically influence the performance and sample efficiency of CSM policy. Building on this finding, we devise a simple yet effective return-coverage rebalancing mechanism that can be seamlessly integrated into common CSM frameworks, including the most widely used one, Decision Transformer (DT). The resulting DT algorithm, referred to as Rebalancing Value-regularized Decision Transformer (RVDT), integrates both implicit and explicit return-coverage rebalancing mechanisms, and achieves state-of-the-art performance in the D4RL experiments.



Paperid:4947
Authors:Jin Wang, Yao Lai, Aoxue Li, Shifeng Zhang, Jiacheng Sun, Ning Kang, Chengyue Wu, Zhenguo Li, Ping Luo
Abstract:
The rapid progress of large language models (LLMs) has catalyzed the emergence of multimodal large language models (MLLMs) that unify visual understanding and image generation within a single framework. However, most existing MLLMs rely on autoregressive (AR) architectures, which impose inherent limitations on future development, such as the raster-scan order in image generation and restricted reasoning abilities in causal context modeling. In this work, we challenge the dominance of AR-based approaches by introducing FUDOKI, a unified multimodal model purely based on discrete flow matching, as an alternative to conventional AR paradigms. By leveraging metric-induced probability paths with kinetic optimal velocities, our framework goes beyond the previous masking-based corruption process, enabling iterative refinement with self-correction capability and richer bidirectional context integration during generation. To mitigate the high cost of training from scratch, we initialize FUDOKI from pre-trained AR-based MLLMs and adaptively transition to the discrete flow matching paradigm. Experimental results show that FUDOKI achieves performance comparable to state-of-the-art AR-based MLLMs across both visual understanding and image generation tasks, highlighting its potential as a foundation for next-generation unified multimodal models. Furthermore, we show that applying test-time scaling techniques to FUDOKI yields significant performance gains, further underscoring its promise for future enhancement through reinforcement learning.



Paperid:4948
Authors:Binh Ho, Long Nguyen-Chi, TrungTin Nguyen, Van Hoang, Binh Nguyen, Chris Drovandi
Abstract:
Model-based clustering integrated with variable selection is a powerful tool for uncovering latent structures within complex data. However, its effectiveness is often hindered by challenges such as identifying relevant variables that define heterogeneous subgroups and handling data that are missing not at random, a prevalent issue in fields like transcriptomics. While several notable methods have been proposed to address these problems, they typically tackle each issue in isolation, thereby limiting their flexibility and adaptability. This paper introduces a unified framework designed to address these challenges simultaneously. Our approach incorporates a data-driven penalty matrix into penalized clustering to enable more flexible variable selection, along with a mechanism that explicitly models the relationship between missingness and latent class membership. We demonstrate that, under certain regularity conditions, the proposed framework achieves both asymptotic consistency and selection consistency, even in the presence of missing data. This unified strategy significantly enhances the capability and efficiency of model-based clustering, advancing methodologies for identifying informative variables that define homogeneous subgroups in the presence of complex missing data patterns. The performance of the framework, including its computational efficiency, is evaluated through simulations and demonstrated using both synthetic and real-world transcriptomic datasets.



Paperid:4949
Authors:Zhongao Zhou, Bin Yang, Wenke Huang, Jun Chen, Mang Ye
Abstract:
Abstract:In object re-identification (ReID) task , both cross-modal and multi-modal retrieval methods have achieved notable progress. However, existing approaches are designed for specific modality and category (person or vehicle) retrieval task, lacking generalizability to others. Acquiring multiple task-specific models would result in wasteful allocation of both training and deployment resources. To address the practical requirements for unified retrieval, we introduce Multi-Modal and Multi-Task object ReID ($\rm {M^3T}$-ReID). The $\rm {M^3T}$-ReID task aims to utilize a unified model to simultaneously achieve retrieval tasks across different modalities and different categories. Specifically,to tackle the challenges of modality distibution divergence and category semantics discrepancy posed in $\rm {M^3T}$-ReID, we design a novel Unbiased Prototype Consistency Learning (UPCL) framework, which consists of two main modules: Unbiased Prototypes-guided Modality Enhancement (UPME) and Cluster Prototypes Consistent Regularization (CPCR).UPME leverages modality-unbiased prototypes to simultaneously enhance cross-modal shared features and multi-modal fused features. Additionally, CPCR regulates discriminative semantics learning with category-consistent information through prototypes clustering.Under the collaborative operation of these two modules, our model can simultaneously learn robust cross-modal shared feature and multi-modal fused feature spaces, while also exhibiting strong category-discriminative capabilities. Extensive experiments on multi-modal datasets RGBNT201 and RGBNT100 demonstrates our UPCL framework showcasing exceptional performance for $\rm {M^3T}$-ReID.



Paperid:4950
Authors:Yuyao Zhang, Jinghao Li, Yu-Wing Tai
Abstract:
Text-to-image (T2I) generation has made remarkable progress, yet existing systems still lack intuitive control over spatial composition, object consistency, and multi-step editing. We presentLayerCraft, a modular framework that uses large language models (LLMs) as autonomous agents to orchestrate structured, layered image generation and editing. LayerCraft supports two key capabilities: (1)structured generationfrom simple prompts via chain-of-thought (CoT) reasoning, enabling it to decompose scenes, reason about object placement, and guide composition in a controllable, interpretable manner; and (2)layered object integration, allowing users to insert and customize objects---such as characters or props---across diverse images or scenes while preserving identity, context, and style. The system comprises a coordinator agent, theChainArchitectfor CoT-driven layout planning, and theObject Integration Network (OIN)for seamless image editing using off-the-shelf T2I models without retraining. Through applications like batch collage editing and narrative scene generation, LayerCraft empowers non-experts to iteratively design, customize, and refine visual content with minimal manual effort. Code will be released upon acceptance.



Paperid:4951
Authors:Xin Zhao, Xiaojun Chen, Bingshan Liu, Haoyu Gao, Zhendong Zhao, Yilong Chen
Abstract:
Large language models (LLMs) with Mixture-of-Experts (MoE) architectures achieve impressive performance and efficiency by dynamically routing inputs to specialized subnetworks, known as experts. However, this sparse routing mechanism inherently exhibits task preferences due to expert specialization, introducing a new and underexplored vulnerability to backdoor attacks. In this work, we investigate the feasibility and effectiveness of injecting backdoors into MoE-based LLMs by exploiting their inherent expert routing preferences.We thus propose \textbf{BadSwitch}, a novel backdoor framework that integrates task-coupled dynamic trigger optimization with a sensitivity-guided Top-S expert tracing mechanism. Our approach jointly optimizes trigger embeddings during pretraining while identifying S most sensitive experts, subsequently constraining the Top-K gating mechanism to these targeted experts. Unlike traditional backdoor attacks that rely on superficial data poisoning or model editing, BadSwitch primarily embeds malicious triggers into expert routing paths with strong task affinity, enabling precise and stealthy model manipulation. Through comprehensive evaluations across three prominent MoE architectures (Switch Transformer, QwenMoE, and DeepSeekMoE), we demonstrate that BadSwitch can efficiently hijack pre-trained models with up to 100\% success rate (ASR) while maintaining the highest clean accuracy (ACC) among all baselines.Furthermore, BadSwitch exhibits strong resilience against both text-level and model-level defense mechanisms, achieving 94.07\% ASR and 87.18\% ACC on the AGNews dataset. Our analysis of expert activation patterns reveals fundamental insights into MoE vulnerabilities. We anticipate this work will expose security risks in MoE systems and contribute to advancing AI safety.



Paperid:4952
Authors:wang lin, Wentao Hu, Liyu Jia, Kaihang Pan, Majun Zhang, Zhou Zhao, Fei Wu, Jingyuan Chen, Hanwang Zhang
Abstract:
With the continuous development of large language models and reasoning chain technologies, the potential of deep reasoning based on reinforcement learning has shown remarkable promise in multi-task scenarios. However, existing unified models have yet to achieve end-to-end integration in image generation and understanding tasks, limiting the model’s self-reflection ability and the realization of cross-modal reasoning chains. To address this, we propose Vinic, a novel framework designed to enable interleaved image generation and understanding through deep reasoning capabilities. We leverage a small amount of multimodal chain-of-thought (MCoT) data for cold-start and employ reinforcement learning to guide the integration of image generation and understanding tasks. Additionally, we introduce a momentum-based reward function, which dynamically adjusts the reward distribution by considering historical improvements, ensuring the stability of the model across multiple generations. Experimental results demonstrate that integrating MCoT can achieve a +22% improvement over the base model on Geneval, effectively enhancing both image generation quality and instruction alignment capabilities.



Paperid:4953
Authors:Evan Antoniuk, Shehtab Zaman, Tal Ben-Nun, Peggy Li, James Diffenderfer, Busra Sahin, Obadiah Smolenski, Everett Grethel, Tim Hsu, Anna Hiszpanski, Kenneth Chiu, Bhavya Kailkhura, Brian Van Essen
Abstract:
Abstract:Data-driven molecular discovery leverages artificial intelligence/machine learning (AI/ML) and generative modeling to filter and design novel molecules. Discovering novel molecules requires accurate out-of-distribution (OOD) predictions, but ML models struggle to generalize OOD. Currently, no systematic benchmarks exist for molecular OOD prediction tasks. We present BOOM, $\textbf{b}$enchmarks for $\textbf{o}$ut-$\textbf{o}f$-$\textbf{d}$istribution $\textbf{m}$olecular property predictions: a chemically-informed benchmark for OOD performance on common molecular property prediction tasks. We evaluate over 140 model-task combinations to benchmark deep learning models on OOD performance. Overall, we find that no existing model achieves strong generalization across all tasks: even the top-performing model exhibited an average OOD error 3$\times$ higher than in-distribution. Current chemical foundation models do not show strong OOD extrapolation, while models with high inductive bias can perform well on OOD tasks with simple, specific properties. We perform extensive ablation experiments, highlighting how data generation, pre-training, hyperparameter optimization, model architecture, and molecular representation impact OOD performance. Developing models with strong OOD generalization is a new frontier challenge in chemical ML. This open-source benchmark is available at https://github.com/FLASK-LLNL/BOOM



Paperid:4954
Authors:Mengbo Wang, Shourya Verma, Aditya Malusare, Luopin Wang, Yiyang Lu, Mario Sola, Vaneet Aggarwal, Ananth Grama, Nadia Lanman
Abstract:
Spatial transcriptomics technologies provide a great opportunity for biological discovery by aligning transcriptomics and histopathological morphology. We propose a novel method, "GeneFlow", the first end-to-end framework to tackle the inverse problem of translating single- and multi-cell gene expressions into paired cellular/tissue images. By combining an attention-based RNA encoder with a conditional UNet guided by rectified flow, we generate high-resolution images with various staining (e.g. H\&E, DAPI). Rectified flow with high-order ODE solvers creates a continuous, bijective mapping between expression and image manifolds, addressing the one-to-many relationship inherent in this problem. Our method provides the generation of realistic cellular morphology features and inter-cellular level spatially resolved interactions under genetic or chemical perturbations, assisting disease diagnosis with minimal invasion by revealing dysregulated patterns. Our rectified flow method outperformed diffusion baselines in all experiments.



Paperid:4955
Authors:Gaku Morio, Harri Rowlands, Dominik Stammbach, Christopher D Manning, Peter Henderson
Abstract:
Companies spend large amounts of money on public relations campaigns to project a positive brand image.However, sometimes there is a mismatch between what they say and what they do. Oil \& gas companies, for example, are accused of ``greenwashing'' with imagery of climate-friendly initiatives.Understanding the framing, and changes in framing, at scale can help better understand the goals and nature of public relation campaigns.To address this, we introduce a benchmark dataset of expert-annotated video ads obtained from Facebook and YouTube.The dataset provides annotations for 13 framing types for more than 50 companies or advocacy groups across 21 countries.Our dataset is especially designed for the evaluation of vision-language models (VLMs), distinguishing it from past text-only framing datasets.Baseline experiments show some promising results, while leaving room for improvement for future work: GPT-4.1 can detect environmental messages with 79\% F1 score, while our best model only achieves 46\% F1 score on identifying framing around green innovation.We also identify challenges that VLMs must address, such as implicit framing, handling videos of various lengths, or implicit cultural backgrounds.Our dataset contributes to research in multimodal analysis of strategic communication in the energy sector.



Paperid:4956
Authors:Vipul Sharma, Wesley Suttle, S Sivaranjani
Abstract:
We develop policy gradient algorithms with global optimality and convergence guarantees for reinforcement learning (RL) with proportional-integral-derivative (PID) parameterized control policies. RL enables learning control policies through direct interaction with a system, without explicit model knowledge that is typically assumed in classical control. The PID policy architecture offers built-in structural advantages, such as superior tracking performance, elimination of steady-state errors, and robustness to model error that have made it a widely adopted paradigm in practice. Despite these advantages, the PID parameterization has received limited attention in the RL literature, and PID control designs continue to rely on heuristic tuning rules without theoretical guarantees. We address this gap by rigorously integrating PID control with RL, offering theoretical guarantees while maintaining the practical advantages that have made PID control ubiquitous in practice. Specifically, we first formulate PID control design as an optimization problem with a control policy that is parameterized by proportional, integral, and derivative components. We derive exact expressions for policy gradients in these parameters, and leverage them to develop both model-based and model-free policy gradient algorithms for PID policies. We then establish gradient dominance properties of the PID policy optimization problem, and provide theoretical guarantees on convergence and global optimality in this setting. Finally, we benchmark the performance of our algorithms on the controlgym suite of environments.



Paperid:4957
Authors:Haizhong Zheng, Yang Zhou, Brian Bartoldson, Bhavya Kailkhura, Fan Lai, Jiawei Zhao, Beidi Chen
Abstract:
Reinforcement learning, such as PPO and GRPO, has powered recent breakthroughs in LLM reasoning. Scaling rollout to sample more prompts enables models to selectively use higher-quality data for training, which can stabilize RL training and improve model performance, but at the cost of significant computational overhead. In this paper, we first show that a substantial portion of this overhead can be avoided by skipping uninformative prompts before rollout. Our analysis of reward dynamics reveals a strong temporal consistency in prompt value: prompts that are uninformative in one epoch of training are likely to remain uninformative in near future epochs. Based on these insights, we propose GRESO (GRPO with Efficient Selective Rollout), an online, lightweight pre-rollout filtering algorithm that predicts and skips uninformative prompts using reward training dynamics. By evaluating GRESO on a broad range of math reasoning benchmarks and models, like Qwen2.5-Math-1.5B, DeepSeek-R1-Distill-Qwen-1.5B, and Qwen2.5-Math-7B, we show that GRESO achieves up to 2.4x wall-clock time speedup in rollout and up to 2.0x speedup in total training time without accuracy degradation.



Paperid:4958
Authors:Kaiyang Guo, Yinchuan Li, Zhitang Chen
Abstract:
Direct alignment methods typically optimize large language models (LLMs) by contrasting the likelihoods of preferred versus dispreferred responses. While effective in steering LLMs to match relative preference, these methods are frequently noted for decreasing the absolute likelihoods of example responses. As a result, aligned models tend to generate outputs that deviate from the expected patterns, exhibiting reward-hacking effect even without a reward model. This undesired consequence exposes a fundamental limitation in contrastive alignment, which we characterize as likelihood underdetermination. In this work, we revisit direct preference optimization (DPO)---the seminal direct alignment method---and demonstrate that its loss theoretically admits a decomposed reformulation. The reformulated loss not only broadens applicability to a wider range of feedback types, but also provides novel insights into the underlying cause of likelihood underdetermination. Specifically, the standard DPO implementation implicitly oversimplifies a regularizer in the reformulated loss, and reinstating its complete version effectively resolves the underdetermination issue. Leveraging these findings, we introduce PRoximalized PReference Optimization (PRO), which unifies alignment across diverse feedback types and addresses likelihood underdetermination through an efficient approximation of the complete regularizer. Comprehensive experiments show the superiority of PRO over existing methods in scenarios involving pairwise, binary and scalar feedback.



Paperid:4959
Authors:Die Chen, Zhiwen Li, Cen Chen, Yuexiang Xie, Xiaodan Li, Jinyan Ye, Yingda Chen, Yaliang Li
Abstract:
Text-to-image diffusion models have gained widespread application across various domains, demonstrating remarkable creative potential. However, the strong generalization capabilities of diffusion models can inadvertently lead to the generation of not-safe-for-work (NSFW) content, posing significant risks to their safe deployment. While several concept erasure methods have been proposed to mitigate the issue associated with NSFW content, a comprehensive evaluation of their effectiveness across various scenarios remains absent. To bridge this gap, we introduce a full-pipeline toolkit specifically designed for concept erasure and conduct the first systematic study of NSFW concept erasure methods. By examining the interplay between the underlying mechanisms and empirical observations, we provide in-depth insights and practical guidance for the effective application of concept erasure methods in various real-world scenarios, with the aim of advancing the understanding of content safety in diffusion models and establishing a solid foundation for future research and development in this critical area.We publicly release our code at https://anonymous.4open.science/r/ErasureBenchmark-7BBB to provide an open platform for further exploration and research.



Paperid:4960
Authors:Frank Shih, Zhenghao Jiang, Faming Liang
Abstract:
Uncertainty quantification (UQ) in scientific machine learning is increasingly critical as neural networks are widely adopted to tackle complex problems across diverse scientific disciplines. For physics-informed neural networks (PINNs), a prominent model in scientific machine learning, uncertainty is typically quantified using Bayesian or dropout methods. However, both approaches suffer from a fundamental limitation: the prior distribution or dropout rate required to construct honest confidence sets cannot be determined without additional information.In this paper, we propose a novel method within the framework of extended fiducial inference (EFI) to provide rigorous uncertainty quantification for PINNs. The proposed method leverages a narrow-neck hyper-network to learn the parameters of the PINN and quantify their uncertainty based on imputed random errors in the observations. This approach overcomes the limitations of Bayesian and dropout methods, enabling the construction of honest confidence sets based solely on observed data.This advancement represents a significant breakthrough for PINNs, greatly enhancing their reliability, interpretability, and applicability to real-world scientific and engineering challenges. Moreover, it establishes a new theoretical framework for EFI, extending its application to large-scale models, eliminating the need for sparse hyper-networks, and significantly improving the automaticity and robustness of statistical inference.



Paperid:4961
Authors:Guoliang He, Youhe Jiang, Wencong Xiao, Jiang Kaihua, Shuguang Wang, Jun Wang, Du Zixian, Zhuo Jiang, Xinlei Zhang, Binhang Yuan, Eiko Yoneki
Abstract:
Abstract:The scaling law for large language models (LLMs) depicts that the path towards machine intelligence necessitates training at large scale. Thus, companies continuously build large-scale GPU clusters, and launch training jobs that span over thousands of computing nodes. However, LLM pre-training presents unique challenges due to its complex communication patterns, where GPUs exchange data in sparse yet high-volume bursts within specific groups. Inefficient resource scheduling exacerbates bandwidth contention, leading to suboptimal training performance. This paper presents Arnold, a scheduling system summarizing our experience to effectively align LLM communication patterns to data center topology at scale. In-depth characteristic study is performed to identify the impact of physical network topology to LLM pre-training jobs. Based on the insights, we develop a scheduling algorithm to effectively align communication patterns to physical network topology in data centers. Through simulation experiments, we show the effectiveness of our algorithm in reducing the maximum spread of communication groups by up to $1.67$x. In production training, our scheduling system improves the end-to-end performance by $10.6\%$ when training with more than $9600$ Hopper GPUs, a significant improvement for our training pipeline.



Paperid:4962
Authors:Yingli Shen, Wen Lai, Shuo Wang, Xueren Zhang, Kangyang Luo, Alexander Fraser, Maosong Sun
Abstract:
The rapid development of multilingual large language models (LLMs) highlights the need for high-quality, diverse, and clean multilingual datasets. In this paper, we introduce DCAD-2000 (Data Cleaning as Anomaly Detection), a large-scale multilingual corpus built using newly extracted Common Crawl data and existing multilingual datasets. DCAD-2000 includes over 2,282 languages, 46.72TB of data, and 8.63 billion documents, spanning 155 high- and medium-resource languages and 159 writing scripts. To overcome the limitations of current data cleaning methods, which rely on manual heuristic thresholds, we propose reframing data cleaning as an anomaly detection task. This dynamic filtering approach significantly enhances data quality by identifying and removing noisy or anomalous content. By fine-tuning LLMs on DCAD-2000, we demonstrate the effectiveness of the data quality improvements, the robustness of the data cleaning pipeline, and enhanced performance for low-resource languages across multiple multilingual benchmarks.



Paperid:4963
Authors:Sam Bright-Thonney, Christina Reissel, Gaia Grosso, Nathaniel Woodward, Katya Govorkova, Andrzej Novak, Sangeon Park, Eric Moreno, Philip Harris
Abstract:
Novelty detection in large scientific datasets faces two key challenges: the intrinsic noise and high dimensionality of experimental data, and the necessity of making statistically robust statements about any observed outliers. While there is a wealth of literature on anomaly detection via dimensionality reduction, most methods do not produce outputs compatible with claims of scientific discovery. In this work we directly address these challenges, presenting the first step towards a unified pipeline for novelty detection adapted for the rigorous demands of scientific discovery. We introduce AutoSciDACT (Automated Scientific Discovery with Anomalous Contrastive Testing), a general purpose two-stage pipeline for detecting novelty in scientific data. The initial pre-training stage involves constructing a low-dimensional embedding of a dataset via contrastive learning, leveraging the abundance of high-quality simulated data in many scientific domains alongside specialist scientific expertise that can guide principled data augmentation strategies. Next, the discovery stage takes a sample of well-understood in-distribution points (i.e.\ a null hypothesis or background-only set) and performs a powerful machine learning-driven two-sample test against another dataset with unknown composition (i.e.\ observed data) using the New Physics Learning Machine (NPLM) framework. We perform experiments across a range of astronomical, physical, biological, image, and synthetic datasets, demonstrating extraordinary sensitivity to small fractions of anomalous data across all domains.



Paperid:4964
Authors:RUI CAO, Zifeng Ding, Zhijiang Guo, Michael Schlichtkrull, Andreas Vlachos
Abstract:
Abstract:Textual claims are often accompanied by images to enhance their credibility and spread on social media, but this also raises concerns about the spread of misinformation.Existing datasets for automated verification of image-text claims remain limited, as they often consist of synthetic claims and lack evidence annotations to capture the reasoning behind the verdict.In this work, we introduce AVerImaTeC, a dataset consisting of 1,297 real-world image-text claims. Each claim is annotated with question-answer (QA) pairs containing evidence from the web, reflecting a decomposed reasoning regarding the verdict.We mitigate common challenges in fact-checking datasets such as contextual dependence, temporal leakage, and evidence insufficiency, via claim normalization, temporally constrained evidence annotation, and a two-stage sufficiency check. We assess the consistency of the annotation in AVerImaTeC via inter-annotator studies, achieving a $\kappa=0.742$ on verdicts and $74.7\%$ consistency on QA pairs. We also propose a novel evaluation method for evidence retrieval and conduct extensive experiments to establish baselines for verifying image-text claims using open-web evidence.



Paperid:4965
Authors:Bo Du, Xuekang Zhu, Xiaochen Ma, Chenfan Qu, Kaiwen Feng, Zhe Yang, Chi-Man Pun, Jian liu, Ji-Zhe Zhou
Abstract:
The field of Fake Image Detection and Localization (FIDL) is highly fragmented, encompassing four domains: deepfake detection (Deepfake), image manipulation detection and localization (IMDL), artificial intelligence-generated image detection (AIGC), and document image manipulation localization (Doc). Although individual benchmarks exist in some domains, a unified benchmark for all domains in FIDL remains blank. The absence of a unified benchmark results in significant domain silos, where each domain independently constructs its datasets, models, and evaluation protocols without interoperability, preventing cross-domain comparisons and hindering the development of the entire FIDL field. To close the domain silo barrier, we propose ForensicHub, the first unified benchmark \& codebase for all-domain fake image detection and localization. Considering drastic variations on dataset, model, and evaluation configurations across all domains, as well as the scarcity of open-sourced baseline models and the lack of individual benchmarks in some domains, ForensicHub: i) proposes a modular and configuration-driven architecture that decomposes forensic pipelines into interchangeable components across datasets, transforms, models, and evaluators, allowing flexible composition across all domains; ii) fully implements 10 baseline models (3 of which are reproduced from scratch), 6 backbones, 2 new benchmarks for AIGC and Doc, and integrates 2 existing benchmarks of DeepfakeBench and IMDLBenCo through an adapter-based design; iii) establishes an image forensic fusion protocol evaluation mechanism that supports unified training and testing of diverse forensic models across tasks; iv) conducts indepth analysis based on the ForensicHub, offering 8 key actionable insights into FIDL model architecture, dataset characteristics, and evaluation standards. Specifically, ForensicHub includes 4 forensic tasks, 23 datasets, 42 baseline models, 6 backbones, 11 GPU-accelerated pixel- and image-level evaluation metrics, and realizes 16 kinds of cross-domain evaluations. ForensicHub represents a significant leap forward in breaking the domain silos in the FIDL field and inspiring future breakthroughs. Code is available at: https://github.com/scu-zjz/ForensicHub.



Paperid:4966
Authors:Jiayi Tian, Haiduo Huang, Tian Xia, Wenzhe zhao, Pengju Ren
Abstract:
We address the challenge of point cloud registration using color information, where traditional methods relying solely on geometric features often struggle in low-overlap and incomplete scenarios. To overcome these limitations, we propose GeGS-PCR, a novel two-stage method that combines geometric, color, and Gaussian information for robust registration. Our approach incorporates a dedicated color encoder that enhances color features by extracting multi-level geometric and color data from the original point cloud. We introduce the Geometric-3DGS module, which encodes the local neighborhood information of colored superpoints to ensure a globally invariant geometric-color context. Leveraging LORA optimization, we maintain high performance while preserving the expressiveness of 3DGS. Additionally, fast differentiable rendering is utilized to refine the registration process, leading to improved convergence. To further enhance performance, we propose a joint photometric loss that exploits both geometric and color features. This enables strong performance in challenging conditions with extremely low point cloud overlap. We validate our method by colorizing the Kitti dataset as ColorKitti and testing on both Color3DMatch and Color3DLoMatch datasets. Our method achieves state-of-the-art performance with Registration Recall at 99.9%, Relative Rotation Error as low as 0.013, and Relative Translation Error as low as 0.024, improving precision by at least a factor of 2.



Paperid:4967
Authors:Herbert Woisetschläger, Ryan Zhang, Shiqiang Wang, Hans Arno Jacobsen
Abstract:
Abstract:Open-weight LLM zoos provide access to numerous high-quality models, but selecting the appropriate model for specific tasks remains challenging and requires technical expertise. Most users simply want factually correct, safe, and satisfying responses without concerning themselves with model technicalities, while inference service providers prioritize minimizing operating costs. These competing interests are typically mediated through service level agreements (SLAs) that guarantee minimum service quality. We introduce MESS+, a stochastic optimization algorithm for cost-optimal LLM request routing while providing rigorous SLA compliance guarantees. MESS+ learns request satisfaction probabilities of LLMs in real-time as users interact with the system, based on which model selection decisions are made by solving a per-request optimization problem. Our algorithm includes a novel combination of virtual queues and request satisfaction prediction, along with a theoretical analysis of cost optimality and constraint satisfaction. Across a wide range of state-of-the-art LLM benchmarks, MESS+ achieves an average of $2\times$ cost savings compared to existing LLM routing techniques.



Paperid:4968
Authors:Shaocong Ma, Heng Huang
Abstract:
In financial applications, reinforcement learning (RL) agents are commonly trained on historical data, where their actions do not influence prices. However, during deployment, these agents trade in live markets where their own transactions can shift asset prices, a phenomenon known as market impact. This mismatch between training and deployment environments can significantly degrade performance. Traditional robust RL approaches address this model misspecification by optimizing the worst-case performance over a set of uncertainties, but typically rely on symmetric structures that fail to capture the directional nature of market impact. To address this issue, we develop a novel class of elliptic uncertainty sets. We establish both implicit and explicit closed-form solutions for the worst-case uncertainty under these sets, enabling efficient and tractable robust policy evaluation. Experiments on single-asset and multi-asset trading tasks demonstrate that our method achieves superior Sharpe ratio and remains robust under increasing trade volumes, offering a more faithful and scalable approach to RL in financial markets.



Paperid:4969
Authors:Jasmine Bayrooti, Sattar Vakili, Amanda Prorok, Carl Henrik Ek
Abstract:
Abstract:Thompson sampling (TS) is a powerful and widely used strategy for sequential decision-making, with applications ranging from Bayesian optimization to reinforcement learning (RL). Despite its success, the theoretical foundations of TS remain limited, particularly in settings with complex temporal structure such as RL. We address this gap by establishing no-regret guarantees for TS using models with Gaussian marginal distributions. Specifically, we consider TS in episodic RL with joint Gaussian process (GP) priors over rewards and transitions. We prove a regret bound of $\mathcal{\tilde{O}}(\sqrt{KH\Gamma(KH)})$ over $K$ episodes of horizon $H$, where $\Gamma(\cdot)$ captures the complexity of the GP model. Our analysis addresses several challenges, including the non-Gaussian nature of value functions and the recursive structure of Bellman updates, and extends classical tools such as the elliptical potential lemma to multi-output settings. This work advances the understanding of TS in RL and highlights how structural assumptions and model uncertainty shape its performance in finite-horizon Markov Decision Processes.



Paperid:4970
Authors:Hai Yan, Haijian Ma, Xiaowen Cai, Daizong Liu, Zenghui Yuan, Xiaoye Qu, Jianfeng Dong, Runwei Guan, Xiang Fang, Hongyang He, Yulai Xie, Pan Zhou
Abstract:
Although Multimodal Large Language Models (MLLMs) have demonstrated remarkable achievements in recent years, they remain vulnerable to adversarial examples that result in harmful responses. Existing attacks typically focus on optimizing adversarial perturbations for a certain multimodal image-prompt pair or fixed training dataset, which often leads to overfitting. Consequently, these perturbations fail to remain malicious once transferred to attack unseen image-prompt pairs, suffering from significant resource costs to cover the diverse multimodal inputs in complicated real-world scenarios. To alleviate this issue, this paper proposes a novel adversarial attack on MLLMs based on distribution approximation theory, which models the potential image-prompt input distribution and adds the same distribution-fitting adversarial perturbation on multimodal input pairs to achieve effective cross-image/prompt transfer attacks. Specifically, we exploit the Laplace approximation to model the Gaussian distribution of the image and prompt inputs for the MLLM, deriving an estimate of the mean and covariance parameters. By sampling from this approximated distribution with Monte Carlo mechanism, we efficiently optimize and fit a single input‑agnostic perturbation over diverse image‑prompt pairs, yielding strong universality and transferability. Extensive experiments are conducted to verify the strong adversarial capabilities of our proposed attack against prevalent MLLMs spanning a spectrum of images/prompts.



Paperid:4971
Authors:Yi Xiao, Hang Fan, Kun Chen, Ye Cao, Ben Fei, Wei Xue, LEI BAI
Abstract:
Accurate estimation of background error (i.e., forecast error) distribution is critical for effective data assimilation (DA) in numerical weather prediction (NWP). In state-of-the-art operational DA systems, it is common to account for the temporal evolution of background errors by employing hybrid methods, which blend a static climatological covariance with a flow-dependent ensemble-derived component. While effective to some extent, these methods typically assume Gaussian-distributed errors and rely heavily on hand-crafted covariance structures and domain expertise, limiting their ability to capture the complex, non-Gaussian nature of atmospheric dynamics. In this work, we propose LoRA-EnVar, a novel hybrid ensemble variational DA algorithm that integrates low-rank adaptation (LoRA) into a deep generative modeling framework. We first learn a climatological background error distribution using a variational autoencoder (VAE) trained on historical data. To incorporate flow-dependent uncertainty, we introduce LoRA modules that efficiently adapt the learned distribution in response to flow-dependent ensemble perturbations. Our approach supports online finetuning, enabling dynamic updates of the background error distribution without catastrophic forgetting. We validate LoRA-EnVar in high-resolution assimilation settings using the FengWu forecast model and simulated observations from ERA5 reanalysis. Experimental results show that LoRA-EnVar significantly improves assimilation accuracy over models assuming static background error distribution and achieves comparable or better performance than full finetuning while reducing the number of trainable parameters by three orders of magnitude. This demonstrates the potential of parameter-efficient adaptation for scalable, non-Gaussian DA in operational meteorology.



Paperid:4972
Authors:侯 云浩, Bochao Zou, Min Zhang, 燃 陈, Shangdong Yang, Yanmei Zhang, Junbao Zhuo, Siheng Chen, Jiansheng Chen, Huimin Ma
Abstract:
By sharing information across multiple agents, collaborative perception helps autonomous vehicles mitigate occlusions and improve overall perception accuracy. While most previous work focus on vehicle-to-vehicle and vehicle-to-infrastructure collaboration, with limited attention to aerial perspectives provided by UAVs, which uniquely offer dynamic, top-down views to alleviate occlusions and monitor large-scale interactive environments. A major reason for this is the lack of high-quality datasets for aerial-ground collaborative scenarios. To bridge this gap, we present AGC-Drive, the first large-scale real-world dataset for Aerial-Ground Cooperative 3D perception. The data collection platform consists of two vehicles, each equipped with five cameras and one LiDAR sensor, and one UAV carrying a forward-facing camera and a LiDAR sensor, enabling comprehensive multi-view and multi-agent perception. Consisting of approximately 120k LiDAR frames and 440k images, the dataset covers 14 diverse real-world driving scenarios, including urban roundabouts, highway tunnels, and on/off ramps. Notably, 19.5% of the data comprises dynamic interaction events, including vehicle cut-ins, cut-outs, and frequent lane changes. AGC-Drive contains 400 scenes, each with approximately 100 frames and fully annotated 3D bounding boxes covering 13 object categories. We provide benchmarks for two 3D perception tasks: vehicle-to-vehicle collaborative perception and vehicle-to-UAV collaborative perception. Additionally, we release an open-source toolkit, including spatiotemporal alignment verification tools, multi-agent visualization systems, and collaborative annotation utilities. The dataset and code are available at https://github.com/PercepX/AGC-Drive.



Paperid:4973
Authors:Chengxiu HUA, Jiawen Gu, Yushun Tang
Abstract:
Reinforcement learning (RL) has achieved significant success across a wide range of domains, however, most existing methods are formulated in discrete time. In this work, we introduce a novel RL method for continuous-time control, where state-action dynamics are governed by stochastic differential equations. Departing from traditional value function-based approaches, our key contribution is the characterization of continuous-time Q-functions via a martingale condition and the linking of diffusion policy scores to the action gradient of a learned continuous Q-function by the dynamic programming principle. This insight motivates Continuous Q-Score Matching (CQSM), a score-based policy improvement algorithm. Notably, our method addresses a long-standing challenge in continuous-time RL: preserving the action-evaluation capability of Q-functions without relying on time discretization. We further provide theoretical closed-form solutions for linear-quadratic (LQ) control problems within our framework. Numerical results in simulated environments demonstrate the effectiveness of our proposed CQSM compared to policy gradient and little q-learning in continuous-time settings.



Paperid:4974
Authors:Qianqian Qiao, DanDan Zheng, Yihang Bo, Bao Peng, Heng Huang, Longteng Jiang, HuayeWang, Jingdong Chen, Jun Zhou, Xin Jin
Abstract:
Video aesthetic assessment, a vital area in multimedia computing, integrates computer vision with human cognition. Its progress is limited by the lack of standardized datasets and robust models, as the temporal dynamics of video and multimodal fusion challenges hinder direct application of image-based methods. This study introduces VADB, the largest video aesthetic database with 10,490 diverse videos annotated by 37 professionals across multiple aesthetic dimensions, including overall and attribute-specific aesthetic scores, rich language comments and objective tags. We propose VADB-Net, a dual-modal pre-training framework with a two-stage training strategy, which outperforms existing video quality assessment models in scoring tasks and supports downstream video aesthetic assessment tasks. The dataset and source code are available at https://github.com/BestiVictory/VADB.



Paperid:4975
Authors:Ibuki Maeda, Yao, Atsushi Nitanda
Abstract:
Abstract:The Schrödinger bridge problem seeks the optimal stochastic process that connects two given probability distributions with minimal energy modification. While the Sinkhorn algorithm is widely used to solve the static optimal transport problem, a recent work (Pooladian and Niles-Weed, 2024) proposed the *Sinkhorn bridge*, which estimates Schrödinger bridges by plugging optimal transport into the time-dependent drifts of SDEs, with statistical guarantees in the one-sample estimation setting where the true source distribution is fully accessible. In this work, to further justify this method, we study the statistical performance of intermediate Sinkhorn iterations in the two-sample estimation setting, where only finite samples from both source and target distributions are available. Specifically, we establish a statistical bound on the squared total variation error of Sinkhorn bridge iterations: $\mathcal{O}(1/m+1/n + r^{2k})~(r \in (0,1))$, where $m$ and $n$ are the sample sizes from the source and target distributions, respectively, and $k$ is the number of Sinkhorn iterations. This result provides a theoretical guarantee for the finite-sample performance of the Schrödinger bridge estimator and offers practical guidance for selecting sample sizes and the number of Sinkhorn iterations. Notably, our theoretical results apply to several representative methods such as [SF]$^2$M, DSBM-IMF, BM2, and lightSB(-M) under specific settings, through the previously unnoticed connection between these estimators.



Paperid:4976
Authors:Zhizhuo Yin, Yuk Hang Tsui, Pan Hui
Abstract:
Generating full-body human gestures encompassing face, body, hands, and global movements from audio is crucial yet challenging for virtual avatar creation. Existing systems tokenize gestures frame-wise, predicting tokens of each frame from the input audio. However, expressive human gestures consist of varied patterns with different frame lengths, and different body parts exhibit motion patterns of varying durations. Existing systems fail to capture motion patterns across body parts and temporal scales due to the fixed frame-count setting of their gesture tokens. Inspired by the success of the feature pyramid technique in the multi-scale visual information extraction, we propose a novel framework named PyraMotion and an adaptive multi-scale feature capturing model called Attentive Pyramidal VQ-VAE (APVQ-VAE). Objective and subjective experiments demonstrate that the PyraMotion outperforms state-of-the-art methods in terms of generating natural and expressive full-body human gestures. Extensive ablation experiments highlight that the self-adaptiveness integration through attention maps contributes to performance.



Paperid:4977
Authors:Yanru Wu, Jianning Wang, Xiangyu Chen, Aurora, Yang Tan, Hanbing Liu, Yang Li
Abstract:
Continual learning (CL) has been a critical topic in contemporary deep neural network applications, where higher levels of both forward and backward transfer are desirable for an effective CL performance. Existing CL strategies primarily focus on task models — either by regularizing model updates or by separating task-specific and shared components — while often overlooking the potential of leveraging inter-task relationships to enhance transfer. To address this gap, we propose a transferability-aware task embedding, termed H-embedding, and construct a hypernet framework under its guidance to learn task-conditioned model weights for CL tasks. Specifically, H-embedding is derived from an information theoretic measure of transferability and is designed to be online and easy to compute. Our method is also characterized by notable practicality, requiring only the storage of a low-dimensional task embedding per task and supporting efficient end-to-end training. Extensive evaluations on benchmarks including CIFAR-100, ImageNet-R, and DomainNet show that our framework performs prominently compared to various baseline and SOTA approaches, demonstrating strong potential in capturing and utilizing intrinsic task relationships. Our code is publicly available at \url{https://anonymous.4open.science/r/H-embeddingguidedhypernet/}.



Paperid:4978
Authors:Moru Liu, Hao Dong, Jessica Kelly, Olga Fink, Mario Trapp
Abstract:
Abstract:Out-of-distribution (OOD) detection and segmentation are crucial for deploying machine learning models in safety-critical applications such as autonomous driving and robot-assisted surgery. While prior research has primarily focused on unimodal image data, real-world applications are inherently multimodal, requiring the integration of multiple modalities for improved OOD detection. A key challenge is the lack of supervision signals from unknown data, leading to overconfident predictions on OOD samples. To address this challenge, we propose Feature Mixing, an extremely simple and fast method for synthesizing multimodal outliers with theoretical support, which can be further optimized to help the model better distinguish between in-distribution (ID) and OOD data. Feature Mixing is modality-agnostic and applicable to various modality combinations. Additionally, we introduce CARLA-OOD, a new multimodal dataset for OOD segmentation, featuring synthetic OOD objects across diverse scenes and weather conditions. Extensive experiments on SemanticKITTI, nuScenes, CARLA-OOD datasets, and the MultiOOD benchmark demonstrate that Feature Mixing achieves state-of-the-art performance with a $10 \times$ to $370 \times$ speedup. Our source code and dataset will be publicly available.



Paperid:4979
Authors:Xue zhucun, Jiangning Zhang, Xie Xurong, Yuxuan Cai, Yong Liu, Xiangtai Li, Dacheng Tao
Abstract:
Multimodal Large Language Models (MLLMs) have demonstrated excellent performance in video understanding but suffer from degraded effectiveness when processing long videos due to fixed-length contexts and weaknesses in modeling long-term dependencies. Retrieval-Augmented Generation (RAG) technology can mitigate these limitations through dynamic knowledge expansion, but existing RAG schemes for video understanding employ fixed retrieval paradigms that use uniform structures regardless of input query difficulty. This introduces redundant computational overhead and latency (\textit{e.g.}, complex graph traversal operations) for simple queries (\textit{e.g.}, frame-level object recognition) while potentially causing critical information loss due to insufficient retrieval granularity for multi-hop reasoning. Such single-step retrieval mechanisms severely constrain the model's balance between resource efficiency and cognitive depth. To address this, we first propose a novel AdaVideoRAG framework for long-video understanding, which uses a lightweight intent classifier to dynamically and adaptively allocate appropriate retrieval schemes—ranging from the simplest to the most sophisticated—for different video understanding tasks based on query complexity. We introduce an Omni-Knowledge Indexing module to extract valuable information from multi-modal signals for context modeling and build corresponding databases, \textit{i.e.}, a text base from clip captions, ASR, and OCR; a visual base; and a graph for deep semantic understanding. This enables hierarchical knowledge access, integration, and generation from naive retrieval to graph retrieval, achieving an optimal balance between resource consumption and video understanding capabilities. Finally, we construct the HiVU benchmark for deep understanding evaluation. Extensive experiments show that our framework enhances the overall efficiency and accuracy of Video-QA for long videos and can be seamlessly integrated with existing MLLMs via lightweight API calls, establishing a new paradigm for adaptive retrieval augmentation in video analysis. Codes will be open-sourced soon.



Paperid:4980
Authors:Hongjie Chen, Jingqiu Ding, Yiding Hua, Stefan Tiegel
Abstract:
Abstract:We study the problem of robustly estimating the edge density of Erdos Renyi random graphs $\mathbb{G}(n, d^\circ/n)$ when an adversary can arbitrarily add or remove edges incident to an $\eta$-fraction of the nodes.We develop the first polynomial-time algorithm for this problem that estimates $d^\circ$ up to an additive error $O\left({[\sqrt{\log(n) / n} + \eta\sqrt{\log(1/\eta)} ] \cdot \sqrt{d^\circ} + \eta \log(1/\eta)}\right)$.Our error guarantee matches information-theoretic lower bounds up to factors of $\log(1/\eta)$.Moreover, our estimator works for all $d^\circ \geq \Omega(1)$ and achieves optimal breakdown point $\eta = 1/2$.Previous algorithms [Acharya et al 2022, Chen et al 2024], including inefficient ones, incur significantly suboptimal errors.Furthermore, even admitting suboptimal error guarantees, only inefficient algorithms achieve optimal breakdown point.Our algorithm is based on the sum-of-squares (SoS) hierarchy.A key ingredient is to construct constant-degree SoS certificates for concentration of the number of edges incident to small sets in $\mathbb{G}(n, d^\circ/n)$.Crucially, we show that these certificates also exist in the sparse regime, when $d^\circ = o(\log n)$, a regime in which the performance of previous algorithms was significantly suboptimal.



Paperid:4981
Authors:Qitao Tan, Jun Liu, Zheng Zhan, Caiwen Ding, Yanzhi Wang, Xiaolong Ma, Jaewoo Lee, Jin Lu, Geng Yuan
Abstract:
Large language models (LLMs) excel across various tasks, but standard first-order (FO) fine-tuning demands considerable memory, significantly limiting real-world deployment. Recently, zeroth-order (ZO) optimization stood out as a promising memory-efficient training paradigm, avoiding backward passes and relying solely on forward passes for gradient estimation, making it attractive for resource-constrained scenarios. However, ZO method lags far behind FO method in both convergence speed and accuracy. To bridge the gap, we introduce a novel layer-wise divergence analysis that uncovers the distinct update pattern of FO and ZO optimization. Aiming to resemble the learning capacity of FO method from the findings, we propose \textbf{Di}vergence-driven \textbf{Z}eroth-\textbf{O}rder (\textbf{DiZO}) optimization. DiZO conducts divergence-driven layer adaptation by incorporating projections to ZO updates, generating diverse-magnitude updates precisely scaled to layer-wise individual optimization needs. Our results demonstrate that DiZO significantly reduces the needed iterations for convergence without sacrificing throughput, cutting training GPU hours by up to 48\% on various datasets. Moreover, DiZO consistently outperforms the representative ZO baselines in fine-tuning RoBERTa-large, OPT-series, and Llama-series on downstream tasks and, in some cases, even surpasses memory-intensive FO fine-tuning. Our code is released at https://anonymous.4open.science/r/DiZO-E86D.



Paperid:4982
Authors:Julian Kranz, Davide Gallon, Steffen Dereich, Arnulf Jentzen
Abstract:
Abstract:We study gradient flows for loss landscapes of fully connected feed forward neural networks with commonly used continuously differentiable activation functions such as the logistic, hyperbolic tangent, softplus or GELU function. We prove that the gradient flow either converges to a critical point or diverges to infinity while the loss converges to an asymptotic critical value. Moreover, we prove the existence of a threshold $\varepsilon>0$ such that the loss value of any gradient flow initialized at most $\varepsilon$ above the optimal level converges to it. For polynomial target functions and sufficiently big architecture and data set, we prove that the optimal loss value is zero and can only be realized asymptotically. From this setting, we deduce our main result that any gradient flow with sufficiently good initialization diverges to infinity. Our proof heavily relies on the geometry of o-minimal structures. We confirm these theoretical findings with numerical experiments and extend our investigation to real-world scenarios, where we observe an analogous behavior.



Paperid:4983
Authors:Ashley Kurian, Aydin Aysu
Abstract:
Neural networks are valuable intellectual property due to the significant computational cost, expert labor, and proprietary data involved in their development. Consequently, protecting their parameters is critical not only for maintaining a competitive advantage but also for enhancing the model's security and privacy. Prior works have demonstrated the growing capability of cryptanalytic attacks to scale to deeper models. In this paper, we present the first defense mechanism against cryptanalytic parameter extraction attacks. Our key insight is to eliminate the neuron uniqueness necessary for these attacks to succeed. We achieve this by a novel, extraction-aware training method. Specifically, we augment the standard loss function with an additional regularization term that minimizes the distance between neuron weights within a layer. Therefore, the proposed defense has zero area-delay overhead during inference. We evaluate the effectiveness of our approach in mitigating extraction attacks while analyzing the model accuracy across different architectures and datasets. When re-trained with the same model architecture, the results show that our defense incurs a marginal accuracy change of less than 1\% with the modified loss function. Moreover, we present a theoretical framework to quantify the success probability of the attack. When tested comprehensively with prior attack settings, our defense demonstrated empirical success for sustained periods of extraction, whereas unprotected networks are extracted between 14 minutes to 4 hours.



Paperid:4984
Authors:Haolong Yan, Yeqing Shen, Xin Huang, Jia Wang, Kaijun Tan, Zhixuan Liang, Hongxin Li, Zheng Ge, Osamu Yoshie, Si Li, Xiangyu Zhang, Daxin Jiang
Abstract:
With the rapid development of Large Vision Language Models, the focus of Graphical User Interface (GUI) agent tasks shifts from single-screen tasks to complex screen navigation challenges. However, real-world GUI environments, such as PC software and mobile Apps, are often complex and proprietary, making it difficult to obtain the comprehensive environment information needed for agent training and evaluation. This limitation hinders systematic investigation and benchmarking of agent navigation capabilities. To address this limitation, we introduce GUI Exploration Lab, a simulation environment engine for GUI agent navigation research that enables flexible definition and composition of screens, icons, and navigation graphs, while providing full access to environment information for comprehensive agent training and evaluation. Through extensive experiments, we find that supervised fine-tuning enables effective memorization of fundamental knowledge, serving as a crucial foundation for subsequent training. Building on this, single-turn reinforcement learning further enhances generalization to unseen scenarios. Finally, multi-turn reinforcement learning encourages the development of exploration strategies through interactive trial and error, leading to further improvements in screen navigation performance. These findings demonstrate the advantages of reinforcement learning approaches in GUI navigation and offer practical guidance for building more capable and generalizable GUI agents.



Paperid:4985
Authors:John Vastola, Samuel J Gershman, Kanaka Rajan
Abstract:
Learning rules—prescriptions for updating model parameters to improve performance—are typically assumed rather than derived. Why do some learning rules work better than others, and under what assumptions can a given rule be considered optimal? We propose a theoretical framework that casts learning rules as policies for navigating (partially observable) loss landscapes, and identifies optimal rules as solutions to an associated optimal control problem. A range of well-known rules emerge naturally within this framework under different assumptions: gradient descent from short-horizon optimization, momentum from longer-horizon planning, natural gradients from accounting for parameter space geometry, non-gradient rules from partial controllability, and adaptive optimizers like Adam from online Bayesian inference of loss landscape shape. We further show that continual learning strategies like weight resetting can be understood as optimal responses to task uncertainty. By unifying these phenomena under a single objective, our framework clarifies the computational structure of learning and offers a principled foundation for designing adaptive algorithms.



Paperid:4986
Authors:Haonan Wang, Hanyu Zhou, Haoyue Liu, Luxin Yan
Abstract:
Optical flow estimation has achieved promising results in conventional scenes but faces challenges in high-speed and low-light scenes, which suffer from motion blur and insufficient illumination. These conditions lead to weakened texture and amplified noise and deteriorate the appearance saturation and boundary completeness of frame cameras, which are necessary for motion feature matching. In degraded scenes, the frame camera provides dense appearance saturation but sparse boundary completeness due to its long imaging time and low dynamic range. In contrast, the event camera offers sparse appearance saturation, while its short imaging time and high dynamic range gives rise to dense boundary completeness. Traditionally, existing methods utilize feature fusion or domain adaptation to introduce event to improve boundary completeness. However, the appearance features are still deteriorated, which severely affects the mostly adopted discriminative models that learn the mapping from visual features to motion fields and generative models that generate motion fields based on given visual features. So we introduce diffusion models that learn the mapping from noising flow to clear flow, which is not affected by the deteriorated visual features. Therefore, we propose a novel optical flow estimation framework Diff-ABFlow based on diffusion models with frame-event appearance-boundary fusion. Inspired by the appearance-boundary complementarity of frame and event, we propose an Attention-Guided Appearance-Boundary Fusion module to fuse frame and event. Based on diffusion models, we propose a Multi-Condition Iterative Denoising Decoder. Our proposed method can effectively utilize the respective advantages of frame and event, and shows great robustness to degraded input. In addition, we propose a dual-modal optical flow dataset for generalization experiments. Extensive experiments have verified the superiority of our proposed method. We will provide the code once the paper accepted.



Paperid:4987
Authors:Zebin Yang, Sunjian Zheng, Tong Xie, Tianshi Xu, Bo Yu, Fan Wang, Jie Tang, Shaoshan Liu, Meng Li
Abstract:
Abstract:Object-goal navigation (ObjNav) tasks an agent with navigating to the location of a specific object in an unseen environment. Embodied agents equipped with large language models (LLMs) and online constructed navigation maps can perform ObjNav in a zero-shot manner. However, existing agents heavily rely on giant LLMs on the cloud, e.g., GPT-4, while directly switching to small LLMs, e.g., LLaMA3.2-11b, suffer from significant success rate drops due to limited model capacity for understanding complex navigation maps, which prevents deploying ObjNav on local devices.At the same time, the long prompt introduced by the navigation map description will cause high planning latency on local devices.In this paper, we propose EfficientNav to enable on-device efficient LLM-based zero-shot ObjNav. To help the smaller LLMs better understand the environment, we propose semantics-aware memory retrieval to prune redundant information in navigation maps.To reduce planning latency, we propose discrete memory caching and attention-based memory clustering to efficiently save and re-use the KV cache.Extensive experimental results demonstrate that EfficientNavachieves 11.1\% improvement in success rate on HM3D benchmark over GPT-4-based baselines, and demonstrates 6.7$\times$ real-time latency reduction and 4.7$\times$ end-to-end latency reduction over GPT-4 planner. Our code is available on Anonymous Github.



Paperid:4988
Authors:Felix Chalumeau, Daniel Rajaonarivonivelomanantsoa, Ruan John de Kock, Juan Formanek, Sasha Abramowitz, Omayma Mahjoub, Wiem Khlifi, Simon Du Toit, Louay Nessir, Refiloe Shabe, Arnol Fokam, Siddarth Singh, Ulrich Armel Mbou Sob, Arnu Pretorius
Abstract:
Reinforcement learning (RL) systems have countless applications, from energy-grid management to protein design. However, such real-world scenarios are often extremely difficult, combinatorial in nature, and require complex coordination between multiple agents. This level of complexity can cause even state-of-the-art RL systems, trained until convergence, to hit a performance ceiling which they are unable to break out of with zero-shot inference. Meanwhile, many digital or simulation-based applications allow for an inference phase that utilises a specific time and compute budget to explore multiple attempts before outputting a final solution. In this work, we show that such an inference phase employed at execution time, and the choice of a corresponding inference strategy, are key to breaking the performance ceiling observed in complex multi-agent RL problems. Our main result is striking: we can obtain up to a 126% and, on average, a 45% improvement over the previous state-of-the-art across 17 tasks, using only a couple seconds of extra wall-clock time during execution. We also demonstrate promising compute scaling properties, supported by over 60k experiments, making it the largest study on inference strategies for complex RL to date. We make all of our experimental data and code available.



Paperid:4989
Authors:Shuwen Wei, Samuel Remedios, Blake Dewey, Zhangxing Bian, Shimeng Wang, Junyu Chen, Bruno Jedynak, shiv saidha, Peter Calabresi, Aaron Carass, Jerry L Prince
Abstract:
Medical image harmonization aims to reduce the differences in appearance caused by scanner hardware variations to allow for consistent and reliable comparisons across devices.Harmonization based on paired images from different devices has limited applicability in real-world clinical settings.On the other hand, unpaired harmonization typically does not guarantee anatomy consistency, which is problematic as anatomical information preservation is paramount.The Schrödinger bridge framework has achieved state-of-the-art style transfer performance with natural images by matching distributions of unpaired images, but this approach can also introduce anatomy changes when applied to medical images.We show that such changes occur because the Schrödinger bridge uses the square of the Euclidean distance between images as the transport cost in an entropy-regularized optimal transport problem.Such a transport cost is not appropriate for measuring anatomical distances, as medical images with the same anatomy need not have a small Euclidean distance between them.In this paper, we propose a latent metric Schrödinger bridge (LMSB) framework to improve the anatomical consistency for the harmonization of medical images.We develop an invertible network that maps medical images into a latent Euclidean metric space where the distances among images with the same anatomy are minimized using the pullback latent metric. Within this latent space, we train a Schrödinger bridge to match distributions. We show that the proposed LMSB is superior to the direct application of a Schrödinger bridge to harmonize optical coherence tomography (OCT) images.



Paperid:4990
Authors:Sameera Ramasinghe, Thalaiyasingam Ajanthan, Gil Avraham, Yan Zuo, Alexander Long
Abstract:
Scaling models has led to significant advancements in deep learning, but training these models in decentralized settings remains challenging due to communication bottlenecks. While existing compression techniques are effective in data-parallel, they do not extend to model parallelism. Unlike data-parallel training, where weight gradients are exchanged, model-parallel requires compressing activations and activation gradients as they propagate through layers, accumulating compression errors. We propose a novel compression algorithm that compresses both forward and backward passes, enabling up to 99% compression with no convergence degradation with negligible memory/compute overhead. By leveraging a recursive structure in transformer networks, we predefine a low-dimensional subspace to confine the activations and gradients, allowing full reconstruction in subsequent layers. Our method achieves up to 100x improvement in communication efficiency and enables training billion-parameter-scale models over low-end GPUs connected via consumer-grade internet speeds as low as 80Mbps, matching the convergence of centralized datacenter systems with 100Gbps connections with model parallel.



Paperid:4991
Authors:Emanuele Luconi, Dario Liscai, Carlo Baldassi, Alessandro Marin Vargas, Alessandro Sanzeni
Abstract:
Invariant object recognition-the ability to identify objects despite changes in appearance-is a hallmark of visual processing in the brain, yet its understanding remains a central challenge in systems neuroscience. Artificial neural networks trained to predict neural responses to visual stimuli (“digital twins”) could provide a powerful framework for studying such complex computations in silico. However, while current models accurately capture single-neuron responses within individual visual areas, their ability to reproduce how populations of neurons represent object identity, and how these representations transform across the cortical hierarchy, remains largely unexplored. Here we examine key functional signatures observed experimentally and find that current models account for hierarchical changes in basic single-neuron properties, such as receptive field size, but fail to capture more complex population-level phenomena, particularly invariant object representations. To address this gap, we introduce a biologically inspired hierarchical readout scheme that mirrors cortical anatomy, modeling each visual area as a projection from a distinct depth within a shared core network. This approach significantly improves the prediction of population-level representational transformations, outperforming standard models that use only the final layer, as well as alternatives with modified architecture, regularization, and loss function. Our results suggest that incorporating anatomical information provides a strong inductive bias in digital twin models, enabling them to better capture general principles of brain function.



Paperid:4992
Authors:Zhuo Chen, YIZHEN ZHENG, Huan Yee Koh, Hongxin Xiang, Linjiang Chen, Wenjie Du, Yang Wang
Abstract:
Molecular Relational Learning (MRL) aims to understand interactions between molecular pairs, playing a critical role in advancing biochemical research. With the recent development of large language models (LLMs), a growing number of studies have explored the integration of MRL with LLMs and achieved promising results. However, the increasing availability of diverse LLMs and molecular structure encoders has significantly expanded the model space, presenting major challenges for benchmarking. Currently, there is no LLM framework that supports both flexible molecular input formats and dynamic architectural switching. To address these challenges, reduce redundant coding, and ensure fair model comparison, we propose ModuLM, a framework designed to support flexible LLM-based model construction and diverse molecular representations. ModuLM provides a rich suite of modular components, including 8 types of 2D molecular graph encoders, 11 types of 3D molecular conformation encoders, 7 types of interaction layers, and 7 mainstream LLM backbones. Owing to its highly flexible model assembly mechanism, ModuLM enables the dynamic construction of over 50,000 distinct model configurations. In addition, we provide comprehensive benchmark results to demonstrate the effectiveness of ModuLM in supporting LLM-based MRL tasks.



Paperid:4993
Authors:Ziyuan Luo, Yangyi Zhao, Ka Chun Cheung, Simon See, Renjie Wan
Abstract:
The widespread adoption of Retrieval-Augmented Image Generation (RAIG) has raised significant concerns about the unauthorized use of private image datasets. While these systems have shown remarkable capabilities in enhancing generation quality through reference images, protecting visual datasets from unauthorized use in such systems remains a challenging problem. Traditional digital watermarking approaches face limitations in RAIG systems, as the complex feature extraction and recombination processes fail to preserve watermark signals during generation. To address these challenges, we propose ImageSentinel, a novel framework for protecting visual datasets in RAIG. Our framework synthesizes sentinel images that maintain visual consistency with the original dataset. These sentinels enable protection verification through randomly generated character sequences that serve as retrieval keys. To ensure seamless integration, we leverage vision-language models to generate the sentinel images. Experimental results demonstrate that ImageSentinel effectively detects unauthorized dataset usage while preserving generation quality for authorized applications.



Paperid:4994
Authors:Wenda Li, Huijie Zhang, Qing Qu
Abstract:
The widespread use of AI-generated content from diffusion models has raised significant concerns regarding misinformation and copyright infringement. Watermarking is a crucial technique for identifying these AI-generated images and preventing their misuse. In this paper, we introduceShallow Diffuse, a new watermarking technique that embeds robust and invisible watermarks into diffusion model outputs. Unlike existing approaches that integrate watermarking throughout the entire diffusion sampling process,Shallow Diffusedecouples these steps by leveraging the presence of a low-dimensional subspace in the image generation process. This method ensures that a substantial portion of the watermark lies in the null space of this subspace, effectively separating it from the image generation process. Our theoretical and empirical analyses show that this decoupling strategy greatly enhances the consistency of data generation and the detectability of the watermark. Extensive experiments further validate thatShallow Diffuseoutperforms existing watermarking methods in terms of consistency.



Paperid:4995
Authors:Andreas Floros, Seyed-Mohsen Moosavi-Dezfooli, Pier Luigi Dragotti
Abstract:
Diffusion models have transformed image synthesis through iterative denoising processes that define trajectories from noise to coherent data. While their generative capabilities are widely celebrated, a critical challenge remains unaddressed: ensuring responsible use by verifying whether an image originates from a model's training set, its novel generations, or external sources. We propose a framework that analyzes diffusion trajectories to trace data provenance. Unlike prior methods, we demonstrate that temporal dynamics across the entire trajectory encode discriminative signals for robust classification. This challenges the long-standing "Goldilocks zone" conjecture, which posits that membership inference is effective only within narrow denoising stages. More fundamentally, we expose critical flaws in current membership inference practices, showing how existing methods fail under distribution shifts or when model-generated data is present. For model attribution, we present the first approach applicable to diffusion that avoids foundation models and their potential data leakage. Ultimately, we unify membership inference and model attribution into a single, cohesive framework tailored to modern generative systems, making our assumptions explicit and establishing principled benchmarking standards. Our work prioritizes transparency and accountability in an era of increasingly opaque AI. Code and data given in the Supplementary Material.



Paperid:4996
Authors:Gefei Tan, Ali Shahin Shamsabadi, Ellen Kolesnikova, Hamed Haddadi, Xiao Wang
Abstract:
Abstract:Diffusion models are increasingly deployed in real-world text-to-image services. These models, however, encode implicit assumptions about the world based on web-scraped image-caption pairs used during training. Over time, such assumptions may become outdated, incorrect, or socially biased--leading to failures where the generated images misalign with users' expectations or evolving societal norms. Identifying and fixing such failures is challenging and, thus, a valuable asset for service providers, as failures often emerge post-deployment and demand specialized expertise and resources to resolve them. In this work, we introduce $\textit{SURE}$, the first end‑to‑end framework that $\textbf{S}$ec$\textbf{U}$rely $\textbf{RE}$pairs failures flagged by users of diffusion-based services. $\textit{SURE}$ enables the service provider to securely collaborate with an external third-party specialized in model repairing (i.e., Model Repair Institute) without compromising the confidentiality of user feedback, the service provider’s proprietary model, or the Model Repair Institute’s proprietary repairing knowledge. To achieve the best possible efficiency, we propose a co-design of a model editing algorithm with a customized two-party cryptographic protocol.Our experiments show that $\textit{SURE}$ is highly practical: $\textit{SURE}$ securely and effectively repairs all 32 layers of Stable Diffusion v1.4 in under 17 seconds (four orders of magnitude more efficient than a general baseline). Our results demonstrate that practical, secure model repair is attainable for large-scale, modern diffusion services.



Paperid:4997
Authors:Dahoon Lee, Chenglin Fan, Euiwoong Lee
Abstract:
Abstract:Correlation Clustering (CC) is a foundational problem in unsupervised learning that models binary similarity relations using labeled graphs. While classical CC has been well studied, many real-world applications involve more nuanced relationships—either multi-class categorical interactions or varying confidence levels in edge labels. To address these, two natural generalizations have been proposed: Chromatic Correlation Clustering (CCC), which assigns semantic colors to edge labels, and pseudometric-weighted CC, which allows edge weights satisfying the triangle inequality. In this paper, we develop improved approximation algorithms for both settings. Our approach leverages LP-based pivoting techniques combined with problem-specific rounding functions. For the pseudometric-weighted correlation clustering problem, we present a tight $\frac{10}{3}$-approximation algorithm, matching the best possible bound achievable within the framework of standard LP relaxation combined with specialized rounding. For the Chromatic Correlation Clustering (CCC) problem, we improve the approximation ratio from the previous best of $2.5$ to $2.15$, and we establish a lower bound of $2.11$ within the same analytical framework, highlighting the near-optimality of our result.



Paperid:4998
Authors:Tingjia Shen, Hao Wang, Chuhan Wu, Jin Yao Chin, Wei Guo, Yong Liu, Huifeng Guo, Defu Lian, Ruiming Tang, Enhong Chen
Abstract:
With the growing size of data and models in Sequential Recommendation (SR), the time required for debugging has become increasingly prohibitive, underscoring the urgent need for effective guidance in parameter configuration. The Scaling Law (SL) offers analogous guidance in the language domain, having achieved significant success by predicting model loss when scaling model size. However, the existing guidance from SL for SR remains qualitative, which is because quantitative analysis of SL on SR encounters challenges with quality measurement on redundant sequence along with loss-performance discrepancy. In response, we introduce the \textbf{Performance Law} for SR models, which predicts model performance across various settings, intending to provide a quantitative framework for guiding the parameter optimization of future models. Initially, Performance Law utilizes Real Entropy to measure data quality, aiming to remove the low-quality influence of low-entropy redundant sequences. Subsequently, Performance Law investigates a fitting decay term, which facilitated the prediction of the major loss-performance discrepancy phenomena of overfitting, ultimately achieving quantitative performance prediction. Extensive experiment on various datasets demonstrates the effectiveness of Performance Law by displaying exceptional quantitative prediction ability against the original and modified qualitative SL. Additional application experiments on optimal parameter prediction and model expansion potential prediction also demonstrated the broad applicability of the Performance Law. Our anonymized code is available at https://anonymous.4open.science/r/Performance-Law-DDE3/



Paperid:4999
Authors:Xue-Feng Zhu, Tianyang Xu, Yifan Pan, Jinjie Gu, Xi Li, Jiwen Lu, Xiaojun Wu, Josef Kittler
Abstract:
Existing multi-modal object tracking approaches primarily focus on dual-modal paradigms, such as RGB-Depth or RGB-Thermal, yet remain challenged in complex scenarios due to limited input modalities. To address this gap, this work introduces a novel multi-modal tracking task that leverages three complementary modalities, including visible RGB, Depth (D), and Thermal Infrared (TIR), aiming to enhance robustness in complex scenarios. To support this task, we construct a new multi-modal tracking dataset, coined RGBDT500, which consists of 500 videos with synchronised frames across the three modalities. Each frame provides spatially aligned RGB, depth, and thermal infrared images with precise object bounding box annotations. Furthermore, we propose a novel multi-modal tracker, dubbed RDTTrack. RDTTrack integrates tri-modal information for robust tracking by leveraging a pretrained RGB-only tracking model and prompt learning techniques. In specific, RDTTrack fuses thermal infrared and depth modalities under a proposed orthogonal projection constraint, then integrates them with RGB signals as prompts for the pre-trained foundation tracking model, effectively harmonising tri-modal complementary cues. The experimental results demonstrate the effectiveness and advantages of the proposed method, showing significant improvements over existing dual-modal approaches in terms of tracking accuracy and robustness in complex scenarios.



Paperid:5000
Authors:Yang Liu, Ming Ma, Xiaomin Yu, Pengxiang Ding, Han Zhao, Mingyang Sun, Siteng Huang, Donglin Wang
Abstract:
Despite impressive advancements in Visual-Language Models (VLMs) for multi-modal tasks, their reliance on RGB inputs limits precise spatial understanding. Existing methods for integrating spatial cues, such as point clouds or depth, either require specialized sensors or fail to effectively exploit depth information for higher-order reasoning. To this end, we propose a novel Spatial Sense and Reasoning method, dubbed SSR, a novel framework that transforms raw depth data into structured, interpretable textual rationales. These textual rationales serve as meaningful intermediate representations to significantly enhance spatial reasoning capabilities. Additionally, we leverage knowledge distillation to compress the generated rationales into compact latent embeddings, which facilitate resource-efficient and plug-and-play integration into existing VLMs without retraining. To enable comprehensive evaluation, we introduce a new dataset named SSR-CoT, a million-scale visual-language reasoning dataset enriched with intermediate spatial reasoning annotations, and present SSRBench, a comprehensive multi-task benchmark. Extensive experiments on multiple benchmarks demonstrate SSR substantially improves depth utilization and enhances spatial reasoning, thereby advancing VLMs toward more human-like multi-modal understanding.



Paperid:5001
Authors:Han Wang, Chao Ning
Abstract:
Conformal Prediction (CP) is a powerful statistical machine learning tool to construct uncertainty sets with coverage guarantees, which has fueled its extensive adoption in generating prediction regions for decision-making tasks, e.g., Trajectory Optimization (TO) in uncertain environments. However, existing methods predominantly employ a sequential scheme, where decisions rely unidirectionally on the prediction regions, and consequently the information from decision-making fails to be fed back to instruct CP. In this paper, we propose a novel Feedback-Based CP (Fb-CP) framework for shrinking-horizon TO with a joint risk constraint over the entire mission time. Specifically, a CP-based posterior risk calculation method is developed by fully leveraging the realized trajectories to adjust the posterior allowable risk, which is then allocated to future times to update prediction regions. In this way, the information in the realized trajectories is continuously fed back to the CP, enabling attractive feedback-based adjustments of the prediction regions and a provable online improvement in trajectory performance. Furthermore, we theoretically prove that such adjustments consistently maintain the coverage guarantees of the prediction regions, thereby ensuring provable safety. Additionally, we develop a decision-focused iterative risk allocation algorithm with theoretical convergence analysis for allocating the posterior allowable risk which closely aligns with Fb-CP. Furthermore, we extend the proposed method to handle distribution shift. The effectiveness and superiority of the proposed method are demonstrated through benchmark experiments.



Paperid:5002
Authors:Jiaqi Huang, Zunnan Xu, Jun Zhou, Ting Liu, Yicheng Xiao, Mingwen Ou, Bowen Ji, Xiu Li, Kehong Yuan
Abstract:
Leveraging multimodal large models for image segmentation has become a prominent research direction. However, existing approaches typically rely heavily on manually annotated datasets that include explicit reasoning processes, which are costly and time-consuming to produce. Recent advances suggest that reinforcement learning (RL) can endow large models with reasoning capabilities without requiring such reasoning-annotated data.In this paper, we propose SAM-R1, a novel framework that enables multimodal large models to perform fine-grained reasoning in image understanding tasks. Our approach is the first to incorporate fine-grained segmentation settings during the training of multimodal reasoning models. By integrating task-specific, fine-grained rewards with a tailored optimization objective, we further enhance the model's reasoning and segmentation alignment. We also leverage the Segment Anything Model (SAM) as a strong and flexible reward provider to guide the learning process.With only 3k training samples, SAM-R1 achieves strong performance across multiple benchmarks, demonstrating the effectiveness of reinforcement learning in equipping multimodal models with segmentation-oriented reasoning capabilities.



Paperid:5003
Authors:Xiang Hu, Jiaqi Leng, Jun Zhao, Kewei Tu, Wei Wu
Abstract:
Abstract:A key advantage of Recurrent Neural Networks (RNNs) over Transformers is their linear computational and space complexity enables faster training and inference for long sequences. However, RNNs are fundamentally unable to randomly access historical context, and simply integrating attention mechanisms may undermine their efficiency advantages.To overcome this limitation, we propose \textbf{H}ierarchical \textbf{S}parse \textbf{A}ttention (HSA), a novel attention mechanism that enhances RNNs with long-range random access flexibility while preserving their merits in efficiency and length generalization. HSA divides inputs into chunks, selecting the top-$k$ chunks and hierarchically aggregates information.The core innovation lies in learning token-to-chunk relevance based on fine-grained token-level information inside each chunk. This approach enhances the precision of chunk selection across both in-domain and out-of-domain context lengths.To make HSA efficient, we further introduce a hardware-aligned kernel design.By combining HSA with Mamba, we introduce RAMba, which achieves perfect accuracy in passkey retrieval across 64 million contexts despite pre-training on only 4K-length contexts, and significant improvements on various downstream tasks, with nearly constant memory footprint. These results show RAMba's huge potential in long-context modeling.



Paperid:5004
Authors:Wei Wu, Yuxing Lu, Zhengrui Guo, Chi Zhang, Can Liao, Yifan Bu, Fangxu Zhou, Jinzhuo Wang
Abstract:
In computational neuroscience, models representing single-neuron in-vivo activity have become essential for understanding the functional identities of individual neurons. These models, such as implicit representation methods based on Transformer architectures, contrastive learning frameworks, and variational autoencoders, aim to capture the invariant and intrinsic computational features of single neurons. The learned single-neuron computational role representations should remain invariant across changing environment and are affected by their molecular expression and location. Thus, the representations allow for in vivo prediction of the molecular cell types and anatomical locations of single neurons, facilitating advanced closed-loop experimental designs. However, current models face the problem of limited generalizability. This is due to batch effects caused by differences in experimental design, animal subjects, and recording platforms. These confounding factors often lead to overfitting, reducing the robustness and practical utility of the models across various experimental scenarios. Previous studies have not rigorously evaluated how well the models generalize to new animals or stimulus conditions, creating a significant gap in the field. To solve this issue, we present a comprehensive experimental protocol that explicitly evaluates model performance on unseen animals and stimulus types. Additionally, we propose a model-agnostic adversarial training strategy. In this strategy, a discriminator network is used to eliminate batch-related information from the learned representations. The adversarial framework forces the representation model to focus on the intrinsic properties of neurons, thereby enhancing generalizability. Our approach is compatible with all major single-neuron representation models and significantly improves model robustness. This work emphasizes the importance of generalization in single-neuron representation models and offers an effective solution, paving the way for the practical application of computational models in vivo. It also shows potential for building unified atlases based on single-neuron in vivo activity.



Paperid:5005
Authors:Jingyang Yi, Jiazheng Wang, Sida Li
Abstract:
Abstract:Recent models such as OpenAI o1 and DeepSeek-R1 have demonstrated strong performance on reasoning-intensive tasks by generating extended Chain-of-Thought (CoT) traces. While longer reasoning helps with thorough exploration of solution paths for complex problems, it also often leads to inefficient and redundant outputs—a phenomenon commonly described as $\textit{overthinking}$. In this paper, we propose $\texttt{ShorterBetter}$, a simple yet effective reinforcement learning method that enables reasoning models to learn their own optimal CoT lengths without manual supervision. We define the $\textit{Sample Optimal Length}$ (SOL) as the length of the shortest correct response among multiple generations, which serves as a dynamic reward signal to guide the model toward efficient reasoning. Applied to DeepSeek-Distill-Qwen-1.5B/7B as base models, $\texttt{ShorterBetter}$ achieves 50\%-80\% reduction in output lengths in both in-domain and out-of-domain reasoning tasks while maintaining accuracy. Our reasoning trace analysis shows that $\texttt{ShorterBetter}$ refines the structure of the reasoning traces by reducing unnecessary repetition, excessive self-verification, and over-exploration of alternatives.



Paperid:5006
Authors:Nianzu Yang, Jian Ma, Songlin Jiang, Huaijin Wu, Shuangjia Zheng, Wengong Jin, Junchi Yan
Abstract:
Diffusion models hold great potential for accelerating antibody design, but their performance is so far limited by the number of antibody-antigen complexes used for model training. Meanwhile, AlphaFold3-like protein folding models, pre-trained on a large corpus of crystal structures, have acquired a broad understanding of biomolecular interaction. Based on this insight, we develop a new antigen-conditioned antibody design model by adapting the diffusion module of AlphaFold3-like models for sequence-structure co-diffusion. Specifically, we extend their structure diffusion module with a sequence diffusion head and fine-tune the entire protein folding model for antibody sequence-structure co-design. Our benchmark results show that sequence-structure co-diffusion models not only surpass state-of-the-art antibody design methods in performance but also maintain structure prediction accuracy comparable to the original folding model. Notably, in the antibody co-design task, our method achieves a CDR-H3 recovery rate of 65% for typical antibodies, outperforming the baselines by 87%, and attains a remarkable 63% recovery rate for nanobodies.



Paperid:5007
Authors:Xiaomeng Xu, Yifan Hou, Zeyi Liu, Shuran Song
Abstract:
We address key challenges in Dataset Aggregation (DAgger) for real-world contact-rich manipulation: how to collect informative human correction data and how to effectively update policies with this new data. We introduce Compliant Residual DAgger (CR-DAgger), which contains two novel components: 1) a Compliant Intervention Interface that leverages compliance control, allowing humans to provide gentle, accurate delta action corrections without interrupting the ongoing robot policy execution; and 2) a Compliant Residual Policy formulation that learns from human corrections while incorporating force feedback and force control. Our system significantly enhances performance on precise contact-rich manipulation tasks using minimal correction data, improving base policy success rates by over 50\% on two challenging tasks (book flipping and belt assembly) while outperforming both retraining-from-scratch and finetuning approaches. Through extensive real-world experiments, we provide practical guidance for implementing effective DAgger in real-world robot learning tasks.



Paperid:5008
Authors:Tuan Dam
Abstract:
Abstract:We present \textbf{SVUCRL}, a reinforcement-learning algorithm fornon-stationary communicating MDPs that exploits \emph{low-rank structure} inhow transitions drift over time.When the rank of the drift is $K\ll SA$, SVUCRL achieves\[ \widetilde{\mathcal O}\Bigl(D_{\max}\sqrt{S A T} +D_{\max}\sqrt{(B_r+B_p)K S T}\Bigr)\]dynamic regret—matching the conjectured $\sqrt{T}$ rate up to logarithms andstrictly improving on the $T^{3/4}$ dependence of \emph{SWUCRL2–CW}.Key ingredients:1. \emph{Online low-rank tracking} with explicit Frobenius guarantees;2. An \emph{incremental} robust-PCA decomposition that separates structured drift from sparse shocks with martingale dual certificates;3. \emph{Adaptive} confidence widening driven by a bias-corrected local variation estimator; and4. Time-series forecasting of the low-rank factors, combined with an optimal shrinkage rule.



Paperid:5009
Authors:Emre Sahinoglu, Youbang Sun, Shahin Shahrampour
Abstract:
Abstract:This work addresses the finite-time analysis of nonsmooth nonconvex stochastic optimization under Riemannian manifold constraints. We adapt the notion of Goldstein stationarity to the Riemannian setting as a performance metric for nonsmooth optimization on manifolds. We then propose a Riemannian Online to NonConvex (RO2NC) algorithm, for which we establish the sample complexity of $O(\epsilon^{-3}\delta^{-1})$ in finding ($\delta,\epsilon$) stationary points. This result is the first-ever finite-time guarantee for nonsmooth nonconvex optimization on manifolds and matches the optimal complexity in the Euclidean setting. When gradient information is unavailable, we develop a zeroth order version of RO2NC algorithm (ZO-RO2NC), for which we establish the same sample complexity. Numerical results have shown that our algorithm performs well and aligns with our analysis.



Paperid:5010
Authors:Yang Yue, Zhiqi Chen, Rui Lu, Andrew Zhao, Zhaokai Wang, Yang Yue, Shiji Song, Gao Huang
Abstract:
Abstract:Reinforcement Learning with Verifiable Rewards (RLVR) has recently demonstrated notable success in enhancing the reasoning performance of large language models (LLMs), particularly in mathematics and programming tasks. It is widely believed that, similar to how traditional RL helps agents to explore and learn new strategies, RLVR enables LLMs to continuously self-improve, thus acquiring novel reasoning abilities that exceed the capacity of the corresponding base models. In this study, we take a critical look at \textit{the current state of RLVR} by systematically probing the reasoning capability boundaries of RLVR-trained LLMs across diverse model families, RL algorithms, and math/coding/visual reasoning benchmarks, using pass@\textit{k} at large \textit{k} values as the evaluation metric.While RLVR improves sampling efficiency towards the correct path, we surprisingly find that current training does \emph{not} elicit fundamentally new reasoning patterns.We observe that while RLVR-trained models outperform their base models at smaller values of $k$ (\eg, $k$=1), base models achieve higher pass@$k$ score when $k$ is large.Moreover, we observe that the reasoning capability boundary of LLMs often narrows as RLVR training progresses.Further coverage and perplexity analysis shows that the reasoning paths generated by RLVR models are already included in the base models' sampling distribution, suggesting that their reasoning abilities originate from and are \textit{bounded} by the base model. From this perspective, treating the base model as an upper bound, our quantitative analysis shows that six popular RLVR algorithms perform similarly and remain far from optimal in fully leveraging the potential of the base model.In contrast, we find that distillation can introduce new reasoning patterns from the teacher and genuinely expand the model’s reasoning capabilities.Taken together, our findings suggest that current RLVR methods have not fully realized the potential of RL to elicit genuinely novel reasoning abilities in LLMs. This underscores the need for improved RL paradigms—such as continual scaling and multi-turn agent-environment interaction—to unlock this potential.



Paperid:5011
Authors:Haoyan Yang, Runxue Bao, Cao (Danica) Xiao, Jun Ma, Parminder Bhatia, Shangqian Gao, Taha Kass-Hout
Abstract:
LLM-as-a-Judge has emerged as a promising tool for automatically evaluating generated outputs, but its reliability is often undermined by potential biases in judgment. Existing efforts to mitigate these biases face key limitations: in-context learning-based methods fail to address rooted biases due to the evaluator’s limited capacity for self-reflection, whereas fine-tuning is not applicable to all evaluator types, especially closed-source models. To address this challenge, we introduce theReasoning-basedBiasDetector (RBD), which is a plug-in module that identifies biased evaluations and generates structured reasoning to guide evaluator self-correction. Rather than modifying the evaluator itself, RBD operates externally and engages in an iterative process of bias detection and feedback-driven revision. To support its development, we design a complete pipeline consisting of biased dataset construction, supervision collection, distilled reasoning-based fine-tuning of RBD, and integration with LLM evaluators. We fine-tune four sizes of RBD models, ranging from 1.5B to 14B, and observe consistent performance improvements across all scales. Experimental results on 4 bias types—verbosity, position, bandwagon, and sentiment—evaluated using 8 LLM evaluators demonstrate RBD’s strong effectiveness. For example, the RBD-8B model improves evaluation accuracy by an average of 18.5% and consistency by 10.9%, and surpasses prompting-based baselines and fine-tuned judges by 12.8% and 17.2%, respectively. These results highlight RBD’s effectiveness and scalability. Additional experiments further demonstrate its strong generalization across biases and domains, as well as its efficiency.



Paperid:5012
Authors:Bo Lv, Chen Tang, Nayu Liu, Xin Liu, Yue Yu, Ping Luo
Abstract:
Ensembles of generative large language models (LLMs) are a promising way to compensate for individual model limitations, integrating the strengths of different LLMs. Existing LLM ensemble methods, however, face limitations such as first-token delay and challenges in long-range semantic collaboration between models, Moreover, they typically assume equal voting weights for all models during ensemble, ignoring performance differences between models for a given task. In this work, we propose SpecFuse, a training-free, plug-and-play LLM ensemble framework that dynamically adjusts each model's model contribution in real time based on task performance. Inspired by speculative decoding, SpecFuse iteratively performs drafting and verification, allowing models to collaborate semantically at the segment level for integrated output. Furthermore, we introduce an online feedback mechanism with multiplicative weight updates, where each model's voting weight is adjusted on-the-fly according to how often it "outperforms" others during verification stage, ensuring that stronger models exert greater influence on the ensemble during generation. Experimental results on five popular LLMs (ranging from 7B to 72B parameters) and six benchmark tasks, spanning instruction following, reasoning, commonsense, and general instruction response, demonstrate consistent performance improvements compared to state-of-the-art LLM ensemble methods.



Paperid:5013
Authors:Hao Wang, zhengnan li, Zhichao Chen, Xu Chen, Shuting He, Guangyi Liu, Haoxuan Li, Zhouchen Lin
Abstract:
Iterative imputation is a prevalent method for completing missing data, which involves iteratively imputing each feature by treating it as a target variable and predicting its missing values using the remaining features. However, existing iterative imputation methods exhibit two critical defects: (1) model misspecification, where a uniform parametric form of model is applied across different features, conflicting with heterogeneous data generation processes; (2) underutilization of fully observed features, where all features are treated as potentially missing, neglecting the valuable information in fully observed features.In this work, we propose Kernel Point Imputation (KPI), a bi-level optimization framework designed to address these issues. The inner-level optimization optimizes the model form for each feature in a reproducing kernel Hilbert space, mitigating model misspecification. The outer-level optimization leverages fully observed features as supervision signals to refine imputations. Extensive experiments on real-world datasets demonstrate that KPI consistently outperforms state-of-the-art imputation methods.



Paperid:5014
Authors:Zhoutong Wu, Yuan Zhang, Yiming Dong, Chenheng Zhang, Cong Fang, Kun Yuan, Zhouchen Lin
Abstract:
Transformer models have driven breakthroughs across various language tasks by their strong capability to learn rich contextual representations. Scaling them to improve representation, however, often demands substantial memory and compute costs, such as the Key-Value (KV) cache used during auto-regressive decoding. Skip connections offer a promising way to improve representation without bloating resource usage, yet most prior works either improve expressivity while leaving KV costs unchanged, or reduce memory at the cost of weaker representation. In this work, we propose SkipV1Former, a Transformer variant that uses skip connections from the first layer's Value heads to strengthen model representation and reduce KV cache. Specifically, from the second block onward, each layer reuses half of its Value heads from the very first layer, while computing the other half as usual-cutting Value projections and V cache by nearly 50 \%. Theoretically, we show that routing uncompressed first-layer Values into deeper layers restores information lost to compression and accelerates the model’s implicit mesa-optimization-a key pattern of Transformer in auto-regressive tasks. Empirically, across different model scales, SkipV1Former delivers consistent reductions of approximately 25 \% in KV cache while improving perplexity relative to standard Multi-Head Attention (MHA) Transformers and some advanced variants. Moreover, we propose a recipe for uptraining existing MHA Transformer checkpoints to SkipV1Former with only 10-15\% additional compute. Finally, SkipV1Former can seamlessly combine advanced methods like Group-Query Attention and Multi-Latent Attention to achieve further KV cache savings and performance improvement. When combined with YOCO, it cuts KV cache size by nearly 50 \% while still improving performance.



Paperid:5015
Authors:Anke Tang, Enneng Yang, Li Shen, Yong Luo, Han Hu, Lefei Zhang, Bo Du, Dacheng Tao
Abstract:
Deep model merging represents an emerging research direction that combines multiple fine-tuned models to harness their specialized capabilities across different tasks and domains. Current model merging techniques focus on merging all available models simultaneously, with weight interpolation-based methods being the predominant approach. However, these conventional approaches are not well-suited for scenarios where models become available sequentially, and they often suffer from high memory requirements and potential interference between tasks. In this study, we propose a training-free projection-based continual merging method that processes models sequentially through orthogonal projections of weight matrices and adaptive scaling mechanisms. Our method operates by projecting new parameter updates onto subspaces orthogonal to existing merged parameter updates while using an adaptive scaling mechanism to maintain stable parameter distances, enabling efficient sequential integration of task-specific knowledge. Our approach maintains constant memory complexity to the number of models, minimizes interference between tasks through orthogonal projections, and retains the performance of previously merged models through adaptive task vector scaling. Extensive experiments on CLIP-ViT models demonstrate that our method achieves a 5-8\% average accuracy improvement while maintaining robust performance in different task orderings. Code is publicly available at https://anonymous.4open.science/r/opcm-7EEE .



Paperid:5016
Authors:Qinting Jiang, chuyang ye, Dongyan Wei, Bingli Wang, Yuan Xue, Jingyan Jiang, Zhi Wang
Abstract:
Despite progress, deep neural networks still suffer performance declines under distribution shifts between training and test domains, leading to a substantial decrease in Quality of Experience (QoE) for applications. Existing test-time adaptation (TTA) methods are challenged by dynamic, multiple test distributions within batches. We observe that feature distributions across different domains inherently cluster into distinct groups with varying means and variances. This divergence reveals a critical limitation of previous global normalization strategies in TTA, which inevitably distort the original data characteristics. Based on this insight, we propose Feature-based Instance Neighbor Discovery (FIND), which comprises three key components: Layer-wise Feature Disentanglement (LFD), Feature Aware Batch Normalization (FABN) and Selective FABN (S-FABN). LFD stably captures features with similar distributions at each layer by constructing graph structures. While FABN optimally combines source statistics with test-time distribution specific statistics for robust feature representation. Finally, S-FABN determines which layers require feature partitioning and which can remain unified, thereby enhancing inference efficiency. Extensive experiments demonstrate that FIND significantly outperforms existing methods, achieving a 30\% accuracy improvement in dynamic scenarios while maintaining computational efficiency.



Paperid:5017
Authors:Haoran Luo, Haihong E, Guanting Chen, Yandan Zheng, Xiaobao Wu, Yikai Guo, Qika Lin, Yu Feng, Zemin Kuang, Meina Song, Yifan Zhu, Anh Tuan Luu
Abstract:
Standard Retrieval-Augmented Generation (RAG) relies on chunk-based retrieval, whereas GraphRAG advances this approach by graph-based knowledge representation. However, existing graph-based RAG approaches are constrained by binary relations, as each edge in an ordinary graph connects only two entities, limiting their ability to represent the n-ary relations (n >= 2) in real-world knowledge. In this work, we propose HyperGraphRAG, the first hypergraph-based RAG method that represents n-ary relational facts via hyperedges. HyperGraphRAG consists of a comprehensive pipeline, including knowledge hypergraph construction, retrieval, and generation. Experiments across medicine, agriculture, computer science, and law demonstrate that HyperGraphRAG outperforms both standard RAG and previous graph-based RAG methods in answer accuracy, retrieval efficiency, and generation quality.



Paperid:5018
Authors:Yehya Farhat, Hamza Shili, Fangshuo Liao, Chen Dun, Mirian Hipolito Garcia, Guoqing Zheng, Ahmed Awadallah, Robert Sim, Dimitrios Dimitriadis, Anastasios Kyrillidis
Abstract:
Mixture-of-Experts (MoEs) achieve scalability by dynamically activating subsets of their components.Yet, understanding how expertise emerges through joint training of gating mechanisms and experts remains incomplete, especially in scenarios without clear task partitions. Motivated by inference costs and data heterogeneity, we study how joint training of gating functions and experts can dynamically allocate domain-specific expertise across multiple underlying data distributions.As an outcome of our framework, we develop an instance tailored specifically to decentralized training scenarios, introducingDynamically Decentralized Orchestration of MoEsorDDOME.DDOMEleverages heterogeneity emerging from distributional shifts across decentralized data sources to specialize experts dynamically. By integrating a pretrained common expert to inform a gating function,DDOMEachieves personalized expert subset selection on-the-fly, facilitating just-in-time personalization. We empirically validateDDOMEwithin a Federated Learning (FL) context:DDOMEattains from 4\% up to an 24\% accuracy improvement over state-of-the-art FL baselines in image and text classification tasks, while maintaining competitive zero-shot generalization capabilities. Furthermore, we provide theoretical insights confirming that the joint gating-experts training is critical for achieving meaningful expert specialization.



Paperid:5019
Authors:Ziao Wang, Sixing Yan, Kejing Yin, Xiaofeng Zhang, William Cheung
Abstract:
Vision-language models have been explored for radiology report generation with promising results. Yet, uncertainty elaborated in findings and the reasoning process for reaching clinical impressions are seldom explicitly modeled, reducing the clinical accuracy and trustworthiness of the generated reports. We present CURV, a novel framework that alleviates the limitations through integrated awareness of uncertainty and explicit reasoning capabilities. Our approach consists of three key components: (1) an uncertainty modeling mechanism that teaches the model to recognize and express appropriate levels of diagnostic confidence, (2) a structured reasoning framework that generates intermediate explanatory steps connecting visual findings to clinical impressions, and (3) a reasoning coherence reward that ensures logical consistency among findings, reasoning, and impressions. We implement CURV through a three-stage training pipeline that combines uncertainty-aware fine-tuning, reasoning initialization, and reinforcement learning. In particular, we adopt a comprehensive reward function addresses multiple aspects of report quality, incorporating medical term matching, uncertainty expression evaluation, and semantic coherence evaluation. Experimental results demonstrate that CURV generates clinically relevant reports with appropriate uncertainty expressions and transparent reasoning traces, significantly outperforming previous methods. CURV represents a substantial advancement toward interpretable and trustworthy AI-generated radiology reports, with broader implications for the deployment of vision-language models in high-stakes clinical environments where uncertainty awareness and reasoning transparency are essential.



Paperid:5020
Authors:Xiyue Peng, Hengquan Guo, Jiawei Zhang, Dongqing Zou, Ziyu Shao, Honghao Wei, Xin Liu
Abstract:
Balancing helpfulness and safety (harmlessness) is a critical challenge in aligning large language models (LLMs). Current approaches often decouple these two objectives, training separate preference models for helpfulness and safety, while framing safety as a constraint within a constrained Markov Decision Process (CMDP) framework. This paper identifies a potential issue when using the widely adopted expected safety constraints for LLM safety alignment, termed "safety compensation'', where the constraints are satisfied on expectation, but individual prompts may trade off safety, resulting in some responses being overly restrictive while others remain unsafe. To address this issue, we proposeRectified Policy Optimization (RePO), which replaces the expected safety constraint with critical safety constraints imposed on every prompt. At the core of RePO is a policy update mechanism driven by rectified policy gradients, which penalizes the strict safety violation of every prompt, thereby enhancing safety across nearly all prompts. Our experiments demonstrate that RePO outperforms strong baseline methods and significantly enhances LLM safety alignment.



Paperid:5021
Authors:Yihuan Mao, Chongjie Zhang
Abstract:
Given the ever-changing nature of the world and its inhabitants, agents must possess the ability to adapt and evolve over time. Recent research in Given the ever-changing nature of the world and its inhabitants, agents must possess the ability to adapt and evolve over time. Recent research in non-stationary MDPs has focused on addressing this challenge, providing algorithms inspired by task inference techniques. However, these methods ignore the detrimental effects of interference, which particularly harm performance in contradictory tasks, leading to low efficiency in some environments. To address this issue, we propose a Bayesian Fast-Slow Framework (BFSF) that tackles both cross-task generalization and resistance to cross-task interference. Our framework consists of two components: a 'fast' policy, learned from recent data, and a 'slow' policy, learned through meta-reinforcement learning (meta-RL) using data from all previous tasks. A Bayesian estimation mechanism determines the current choice of 'fast' or 'slow' policy, balancing exploration and exploitation. Additionally, in the 'fast' policy, we introduce a dual-reset mechanism and a data relabeling technique to further accelerate convergence when encountering new tasks. Experiments demonstrate that our algorithm effectively mitigates interference and outperforms baseline approaches.



Paperid:5022
Authors:Haozhe Ma, Zhengding Luo, Thanh Vinh Vo, Kuankuan Sima, Tze-Yun Leong
Abstract:
Reward shaping is effective in addressing the sparse-reward challenge in reinforcement learning by providing immediate feedback through auxiliary informative rewards. Based on the reward shaping strategy, we propose a novel multi-task reinforcement learning framework that integrates a centralized reward agent (CRA) and multiple distributed policy agents. The CRA functions as a knowledge pool, which aims to distill knowledge from various tasks and distribute it to individual policy agents to improve learning efficiency. Specifically, the shaped rewards serve as a straightforward metric to encode knowledge. This framework not only enhances knowledge sharing across established tasks but also adapts to new tasks by transferring meaningful reward signals. We validate the proposed method on both discrete and continuous domains, including the representative meta world benchmark, demonstrating its robustness in multi-task sparse-reward settings and its effective transferability to unseen tasks.



Paperid:5023
Authors:Patrick Yubeaton, Andre Nakkab, Weihua Xiao, Luca Collini, Ramesh Karri, Chinmay Hegde, Siddharth Garg
Abstract:
This paper introduces VeriThoughts, a novel dataset designed for reasoning-based Verilog code generation. We establish a new benchmark framework grounded in formal verification methods to evaluate the quality and correctness of generated hardware descriptions. Additionally, we present a suite of specialized small-scale models optimized specifically for Verilog generation. Our work addresses the growing need for automated hardware design tools that can produce verifiably correct implementations from high-level specifications, potentially accelerating the hardware development process while maintaining rigorous correctness guarantees.



Paperid:5024
Authors:Wudi Chen, Zhiyuan Zha, Shigang Wang, Bihan Wen, Xin Yuan, Jiantao Zhou, zipei fan, Gang Yan, Ce Zhu
Abstract:
Recent advancements have suggested that neural radiance fields (NeRFs) show great potential in color editing within the 3D domain. However, most existing NeRF-based editing methods continue to face significant challenges in local region editing, which usually lead to imprecise local object boundaries, difficulties in maintaining multi-view consistency, and over-reliance on annotated data. To address these limitations, in this paper, we propose a novel weakly-supervised method called KaRF for local color editing, which facilitates high-fidelity and realistic appearance edits in arbitrary regions of 3D scenes. At the core of the proposed KaRF approach is a unified two-stage Kolmogorov-Arnold Networks (KANs)-based radiance fields framework, comprising a segmentation stage followed by a local recoloring stage. This architecture seamlessly integrates geometric priors from NeRF to achieve weakly-supervised learning, leading to superior performance. More specifically, we propose a residual adaptive gating KAN structure, which integrates KAN with residual connections, adaptive parameters, and gating mechanisms to effectively enhance segmentation accuracy and refine specific editing effects. Additionally, we propose a palette-based color-adaptive loss, which can enhance the accuracy of additive mixing results. Extensive experiments demonstrate that the proposed KaRF algorithm significantly outperforms many state-of-the-art methods both qualitatively and quantitatively.



Paperid:5025
Authors:Jiyoung Park, Anirban Bhattacharya, Abhishek Roy, Jonathan W. Siegel
Abstract:
There is extensive literature on accelerating first-order optimization methods in a Euclidean setting. Under which conditions such acceleration is feasible in Riemannian optimization problems is an active area of research. Motivated by the recent success of varying step-size methods in the Euclidean setting, we undertake a study of such algorithms in the Riemannian setting. We show that varying step-size acceleration can be achieved in non-negatively curved Riemannian manifolds under geodesic smoothness and generalized geodesic convexity, a new notion of convexity that we introduce to aid our analysis. As a core application, we show that our method provides the first theoretically guaranteed accelerated optimization method in Wasserstein spaces. In addition, we numerically validate our method's applicability to other problems, such as optimization problems on the sphere.



Paperid:5026
Authors:Xu Wang, Jingyuan Zhuo, Zhiyuan You, Zhiyu Tan, Yikuan Yu, Siyu Wang, Xinyi Le
Abstract:
Recent years have witnessed significant advancements in industrial anomaly detection (IAD) thanks to existing anomaly detection datasets. However, the large performance gap between these benchmarks and real industrial practice reveals critical limitations in existing datasets. We argue that the mismatch between current datasets and real industrial scenarios becomes the primary barrier to practical IAD deployment. To this end, we propose ReinAD dataset, a comprehensive contrastive dataset towards Real-world industrial Anomaly Detection. Our dataset prioritizes three critical real-world requirements: 1) Contrast-based anomaly definition that is essential for industrial practice, 2) Fine-grained unaligned image pairs reflecting real inspections, and 3) Large-scale data from active production lines spanning multiple industrial categories. Based on our dataset, we introduce the ReinAD method. It takes both normal reference and test images as inputs, achieving anomaly detection through normal-anomaly comparison. To address the fine-grained and unaligned properties of real industrial scenes, our method integrates pyramidal similarity aggregation for comprehensive anomaly characterization and global-local feature fusion for spatial misalignment tolerance. Our method outperforms all baselines on the ReinAD dataset (e.g., 64.5% v.s. 59.5% in 1-shot image-level AP) under all settings. Extensive experiments across several datasets demonstrate our dataset's challenging nature and our method's superior generalization. This work provides a solid foundation for practical industrial anomaly detection.



Paperid:5027
Authors:Yuzhou Nie, Zhun Wang, Yu Yang, Ruizhe Jiang, Yuheng Tang, Xander Davies, Yarin Gal, Bo Li, Wenbo Guo, Dawn Song
Abstract:
Existing benchmarks for evaluating the security risks and capabilities (e.g., vulnerability detection) of code-generating large language models (LLMs) face several key limitations:(1) limited coverage of risk and capabilities;(2) reliance on static evaluation metrics such as LLM judgments or rule-based detection, which lack the precision of dynamic analysis; and(3) a trade-off between data quality and benchmark scale.To address these challenges, we introduce a general and scalable benchmark construction framework that begins with manually validated, high-quality seed examples and expands them via targeted mutations.Each mutated sample retains the seed’s security semantics while providing diverse, unseen instances. The resulting benchmark bundles every artifact required for dynamic evaluation, including prompts, vulnerable and patched code, test cases, and ground-truth proofs of concept, enabling rigorous measurement of insecure coding, vulnerability detection, and patch generation. Applying this framework to Python, C/C++, and Java, we build SECCODEPLT, a dataset of more than 5.9k samples spanning 44 CWE-based risk categories and three security capabilities. Compared with state-of-the-art benchmarks, SECCODEPLT offers broader coverage, higher data fidelity, and substantially greater scale. We use SECCODEPLT to evaluate leading code-generation LLMs and agents, revealing their strengths and weaknesses in both generating secure code and identifying or fixing vulnerabilities.We provide our code in \url{https://github.com/ucsb-mlsec/SecCodePLT}, data in \url{https://huggingface.co/datasets/secmlr/SecCodePLT}



Paperid:5028
Authors:Tianyu Chen, Vansh Bansal, James Scott
Abstract:
Abstract:We consider the problem of validating whether a neural posterior estimate $q(\theta \mid x)$ is an accurate approximation to the true, unknown true posterior $p(\theta \mid x)$. Existing methods for evaluating the quality of an NPE estimate are largely derived from classifier-based tests or divergence measures, but these suffer from several practical drawbacks. As an alternative, we introduce the *Conditional Localization Test* (**CoLT**), a principled method designed to detect discrepancies between $p(\theta \mid x)$ and $q(\theta \mid x)$ across the full range of conditioning inputs. Rather than relying on exhaustive comparisons or density estimation at every $x$, CoLT learns a localization function that adaptively selects points $\theta_l(x)$ where the neural posterior $q$ deviates most strongly from the true posterior $p$ for that $x$. This approach is particularly advantageous in typical simulation-based inference settings, where only a single draw $\theta \sim p(\theta \mid x)$ from the true posterior is observed for each conditioning input, but where the neural posterior $q(\theta \mid x)$ can be sampled an arbitrary number of times. Our theoretical results establish necessary and sufficient conditions for assessing distributional equality across all $x$, offering both rigorous guarantees and practical scalability. Empirically, we demonstrate that CoLT not only performs better than existing methods at comparing $p$ and $q$, but also pinpoints regions of significant divergence, providing actionable insights for model refinement. These properties position CoLT as a state-of-the-art solution for validating neural posterior estimates.



Paperid:5029
Authors:Joseph Rowan, Truong Buu Phan, Ashish Khisti
Abstract:
We study a relaxation of the problem of coupling probability distributions — a list of samples is generated from one distribution and anacceptis declared if any one of these samples is identical to the sample generated from the other distribution.We propose a novel method for generating samples, which extends the Gumbel-max sampling suggested in Daliri et al. (2025) for coupling probability distributions. We also establish a corresponding lower bound on the acceptance probability, which we call the \emph{list matching lemma}.We next discuss two applications of our setup.First, we develop a new mechanism for multi-draft speculative sampling that is simple to implement and achieves performance competitive with baselines such as SpecTr and SpecInfer across a range of language tasks.Our method also guarantees a certain degree ofdrafter invariancewith respect to the output tokens which is not supported by existing schemes.We also provide a theoretical lower bound on the token level acceptance probability.As our second application, we consider distributed lossy compression with side information in a setting where a source sample is compressed and available to multiple decoders, each with independent side information.We propose a compression technique that is based on our generalization of Gumbel-max sampling and show that it provides significant gains in experiments involving synthetic Gaussian sources and the MNIST image dataset.



Paperid:5030
Authors:Jiuhong Xiao, Roshan Nayak, Ning Zhang, Daniel Tortei, Giuseppe Loianno
Abstract:
Paired RGB-thermal data is crucial for visual-thermal sensor fusion and cross-modality tasks, including important applications such as multi-modal image alignment and retrieval. However, the scarcity of synchronized and calibrated RGB-thermal image pairs presents a major obstacle to progress in these areas. To overcome this challenge, RGB-to-Thermal (RGB-T) image translation has emerged as a promising solution, enabling the synthesis of thermal images from abundant RGB datasets for training purposes. In this study, we propose ThermalGen, an adaptive flow-based generative model for RGB-T image translation, incorporating an RGB image conditioning architecture and a style-disentangled mechanism. To support large-scale training, we curated eight public satellite-aerial, aerial, and ground RGB-T paired datasets, and introduced three new large-scale satellite-aerial RGB-T datasets—DJI-day, BosonPlus-day, and BosonPlus-night—captured across diverse times, sensor types, and geographic regions. Extensive evaluations across multiple RGB-T benchmarks demonstrate that ThermalGen achieves comparable or superior translation performance compared to existing GAN-based and diffusion-based methods. To our knowledge, ThermalGen is the first RGB-T image translation model capable of synthesizing thermal images that reflect significant variations in viewpoints, sensor characteristics, and environmental conditions. Code, model, and datasets will be publicly released.



Paperid:5031
Authors:Wanpeng Zhang, Yuhao Fang, Xihang Qiu, Jiarong Cheng, Jialong Hong, Bin Zhai, Qing Zhou, Yao Lu, Ye Zhang, Chun Li
Abstract:
Diffusion-based generative models have long depended on Gaussian priors, with little exploration of alternative distributions. To fill this gap, we introduce a scaled and shifted Dirichlet framework that uses time-varying multiplicative transformations to define both forward and reverse diffusion processes. Moving beyond conventional reweighted evidence lower bounds (ELBO) or Kullback–Leibler upper bounds (KLUB), we propose two novel divergence measures: the Proper Hölder Divergence (PHD) and the Proper Hölder–Kullback (PHK) divergence, the latter designed to restore symmetry missing in existing formulations. When optimizing our Dirichlet diffusion model with PHK, we achieve a Fréchet Inception Distance (FID) of 2.81 on unconditional CIFAR-10. Comprehensive experiments on natural-image datasets validate the distinct generative strengths of Dirichlet diffusion and confirm PHK’s effectiveness in model training. These contributions expand the diffusion-model family with principled non-Gaussian processes and robust optimization tools, offering new avenues for versatile, high-fidelity generative modeling.



Paperid:5032
Authors:Junjue Wang, Weihao Xuan, Heli Qi, Zhihao Liu, Kunyi Liu, Yuhan Wu, Hongruixuan Chen, JIAN SONG, Junshi Xia, Zhuo Zheng, Naoto YOKOYA
Abstract:
Large vision-language models (VLMs) have made great achievements in Earth vision. However, complex disaster scenes with diverse disaster types, geographic regions, and satellite sensors have posed new challenges for VLM applications. To fill this gap, we curate a remote sensing vision-language dataset (DisasterM3) for global-scale disaster assessment and response. DisasterM3 includes 26,988 bi-temporal satellite images and 123k instruction pairs across 5 continents, with three characteristics:1) Multi-hazard: DisasterM3 involves 36 historical disaster events with significant impacts, which are categorized into 10 common natural and man-made disasters.2)Multi-sensor: Extreme weather during disasters often hinders optical sensor imaging, making it necessary to combine Synthetic Aperture Radar (SAR) imagery for post-disaster scenes.3) Multi-task: Based on real-world scenarios, DisasterM3 includes 9 disaster-related visual perception and reasoning tasks, harnessing the full potential of VLM's reasoning ability with progressing from disaster-bearing body recognition to structural damage assessment and object relational reasoning, culminating in the generation of long-form disaster reports. We extensively evaluated 14 generic and remote sensing VLMs on our benchmark, revealing that state-of-the-art models struggle with the disaster tasks, largely due to the lack of a disaster-specific corpus, cross-sensor gap, and damage object counting insensitivity. Focusing on these issues, we fine-tune four VLMs using our dataset and achieve stable improvements across all tasks, with robust cross-sensor and cross-disaster generalization capabilities.



Paperid:5033
Authors:Yixuan Xu, Antoine Bosselut, Imanol Schlag
Abstract:
Large language models are known to memorize parts of their training data, posing risk of copyright violations. To systematically examine this risk, we pretrain language models (1B/3B/8B) from scratch on 83B tokens, mixing web-scale data with public domain books used to simulate copyrighted content at controlled frequencies at lengths at least ten times longer than prior work. We thereby identified the offset effect, a phenomenon characterized by two key findings: (1) verbatim memorization is most strongly triggered by short prefixes drawn from the beginning of the context window, with memorization decreasing counterintuitively as prefix length increases; and (2) a sharp decline in verbatim recall when prefix begins offset from the initial tokens of the context window. We attribute this to positional fragility: models rely disproportionately on the earliest tokens in their context window as retrieval anchors, making them sensitive to even slight shifts. We further observe that when the model fails to retrieve memorized content, it often produces degenerated text. Leveraging these findings, we show that shifting sensitive data deeper into the context window suppresses both extractable memorization and degeneration. Our results suggest that positional offset is a critical and previously overlooked axis for evaluating memorization risks, since prior work implicitly assumed uniformity by probing only from the beginning of training sequences.



Paperid:5034
Authors:Zehao Wang, Lin Yang, Jie Wang, Kehan Wang, Hanzhu Chen, Bin Wang, Jianye Hao, Defu Lian, Bin Li, Enhong Chen
Abstract:
Despite their remarkable performance on various tasks, Large Language Models (LLMs) still struggle with logical reasoning, particularly in complex and multi-step reasoning processes. Among various efforts to enhance LLMs' reasoning capabilities, synthesizing large-scale, high-quality logical reasoning datasets has emerged as a promising direction. However, existing methods often rely on predefined templates for logical reasoning data generation, limiting their adaptability to real-world scenarios. To address the limitation, we proposeLogicTree, a novel framework for efficiently synthesizing multi-step logical reasoning dataset that excels in both complexity and instantiation.By iteratively searching for applicable logic rules based on structural pattern matching to perform backward deduction,LogicTreeconstructs multi-step logic trees that capture complex reasoning patterns. Furthermore, we employ a two-stage LLM-based approach to instantiate various real-world scenarios for each logic tree, generating consistent real-world reasoning processes that carry contextual significance. This helps LLMs develop generalizable logical reasoning abilities across diverse scenarios rather than merely memorizing templates.Experiments on multiple benchmarks demonstrate that our approach achieves an average improvement of 10.5\% in accuracy on complex logical reasoning tasks.



Paperid:5035
Authors:Zongxia Li, Xiyang Wu, Guangyao Shi, Yubin Qin, Hongyang Du, Tianyi Zhou, Dinesh Manocha, Jordan Boyd-Graber
Abstract:
Synthetic video generation has gained significant attention for its realism and broad applications, but remains prone to violations of common sense and physical laws. This highlights the need for reliable abnormality detectors that understand such principles and are robust to hallucinations. To address this, we introduce VideoHallu, a benchmark of over 3,000 video QA pairs built from synthetic videos generated by models like Veo2, Sora, and Kling, paired with expert-crafted counterintuitive QA to evaluate the critical thinking abilities of Multi-modal Large Language Models (MLLMs) on abnormalities that are perceptually obvious to humans but often hallucinated due to language priors. VideoHallu evaluates MLLMs' abnormality detection abilities with examples across alignment, consistency, commonsense, and physics. We benchmark SOTA MLLMs, including GPT-4o, Gemini-2.5-Pro, Qwen-2.5-VL, Video-R1, and VideoChat-R1.We observe that these models perform well on many real-world benchmarks like MVBench and MovieChat, but still struggle with basic physics-based and commonsense reasoning in synthetic videos. We further show that post-training with Group Relative Policy Optimization (GRPO), using curriculum learning on datasets combining video QA with counterintuitive commonsense and physics reasoning over real and synthetic videos, improves MLLMs’ abnormality detection and critical thinking, demonstrating the value of targeted training for improving their understanding of commonsense and physical laws.



Paperid:5036
Authors:Yunhao Tang, Sid Wang, Lovish Madaan, Remi Munos
Abstract:
We propose to scale RL to unverifiable data with a novel algorithm JEPO (Jensen's Evidence lower bound for Policy Optimization). While most prior effort on scaling RL for LLMs focuses on verifiable data where ground truth answers are typically short-form and can be matched easily, we investigate the case where such assumptions are less valid (e.g., when answers are long-form such as mathematical proofs). To scale RL training to unverifiable data with contemporary training constraints, we propose JEPO. JEPO applies Jensen's evidence lower bound, a pragmatic simplification of the evidence lower bound which views chain-of-thought as a latent variable in the generative process. We show that on verifiable datasets (math), JEPO is as effective as RL with verifiable reward; on semi-verifiable and unverifiable datasets (numina and numina-proof), JEPO improves on soft-match based evaluations compared to RL with verifiable reward which can only leverage a subset of the data source as well as test set likelihood evaluations.



Paperid:5037
Authors:Trinity Chung, Yuchen Shen, Nathan Kong, Aran Nayebi
Abstract:
Tactile sensing remains far less understood in neuroscience and less effective in artificial systems compared to more mature modalities such as vision and language.We bridge these gaps by introducing a novel Encoder-Attender-Decoder (EAD) framework to systematically explore the space of task-optimized temporal neural networks trained on realistic tactile input sequences from a customized rodent whisker-array simulator. We identify convolutional recurrent neural networks (ConvRNNs) as superior encoders to purely feedforward and state-space architectures for tactile categorization. Crucially, these ConvRNN-encoder-based EAD models achieve neural representations closely matching rodent somatosensory cortex, saturating the explainable neural variability and revealing a clear linear relationship between supervised categorization performance and neural alignment.Furthermore, contrastive self-supervised ConvRNN-encoder-based EADs, trained with tactile-specific augmentations, match supervised neural fits, serving as an ethologically-relevant, label-free proxy.For neuroscience, our findings highlight nonlinear recurrent processing as important for general-purpose tactile representations in somatosensory cortex, providing the first quantitative characterization of the underlying inductive biases in this system. For embodied AI, our results emphasize the importance of recurrent EAD architectures to handle realistic tactile inputs, along with tailored self-supervised learning methods for achieving robust tactile perception with the same type of sensors animals use to sense in unstructured environments.



Paperid:5038
Authors:Will Merrill, Shane Arora, Dirk Groeneveld, Hanna Hajishirzi
Abstract:
The right batch size is important when training language models at scale: a large batch size is necessary for fast training, but a batch size that istoo largewill harm token efficiency. To navigate this tradeoff, McCandlish et al. (2018) suggest that acritical batch size(CBS), below which training will not substantially degrade loss, can be estimated based on the gradient noise scale during training. While their method has been adopted in practice, e.g., when training GPT-3, strong assumptions are required to justify gradient noise as a proxy for the CBS, which makes it unclear whether their approach should be trusted in practice, limiting its applicability. In this paper, we introduce a simple, empirical approach todirectlymeasure the CBS and show how the CBS evolves over training. Applying our approach to the OLMo models, we find that CBS is near 0 at initialization, increases rapidly at first, and then plateaus as training progresses. Furthermore, we find that this trend holds across different model sizes (1B and 7B), suggesting CBS from small training runs can inform larger-scale training runs. Our findings about how the CBS changes over training motivatebatch size warmupas a natural way to reliably train language models at large batch size: start the batch size small and increase it as the CBS grows. To validate this claim, we use batch size warmup to train OLMo 1B to slightly better loss than the original training run with 43% fewer gradient steps. This shows how our framework can be applied to reliably train language models at larger batch sizes, increasing data parallelism without compromising performance.



Paperid:5039
Authors:Huaijin Pi, Zhi Cen, Zhiyang Dou, Taku Komura
Abstract:
Synthesizing whole-body manipulation of articulated objects, including body motion, hand motion, and object motion, is a critical yet challenging task with broad applications in virtual humans and robotics.The core challenges are twofold.First, achieving realistic whole-body motion requires tight coordination between the hands and the rest of the body, as their movements are interdependent during manipulation. Second, articulated object manipulation typically involves high degrees of freedom and demands higher precision, often requiring the fingers to be placed at specific regions to actuate movable parts.To address these challenges, we propose a novel coordinated diffusion noise optimization framework.Specifically, we perform noise-space optimization over three specialized diffusion models for the body, left hand, and right hand, each trained on its own motion dataset to improve generalization.Coordination naturally emerges through gradient flow along the human kinematic chain, allowing the global body posture to adapt in response to hand motion objectives with high fidelity.To further enhance precision in hand-object interaction, we adopt a unified representation based on basis point sets (BPS), where end-effector positions are encoded as distances to the same BPS used for object geometry.This unified representation captures fine-grained spatial relationships between the hand and articulated object parts, and the resulting trajectories serve as targets to guide the optimization of diffusion noise, producing highly accurate interaction motion.We conduct extensive experiments demonstrating that our method outperforms existing approaches in motion quality and physical plausibility, and enables various capabilities such as object pose control, simultaneous walking and manipulation, and whole-body generation from hand-only data.The code will be released for reproducibility.



Paperid:5040
Authors:Shengyuan Liu, Boyun Zheng, Wenting Chen, Zhihao PENG, Zhenfei Yin, Jing Shao, Jiancong Hu, Yixuan Yuan
Abstract:
Endoscopic procedures are essential for diagnosing and treating internal diseases, and multi-modal large language models (MLLMs) are increasingly applied to assist in endoscopy analysis. However, current benchmarks are limited, as they typically cover specific endoscopic scenarios and a small set of clinical tasks, failing to capture the real-world diversity of endoscopic scenarios and the full range of skills needed in clinical workflows. To address these issues, we introduce EndoBench, the first comprehensive benchmark specifically designed to assess MLLMs across the full spectrum of endoscopic practice with multi-dimensional capacities. EndoBench encompasses 4 distinct endoscopic scenarios, 12 specialized clinical tasks with 12 secondary subtasks, and 5 levels of visual prompting granularities, resulting in 6,832 rigorously validated VQA pairs from 21 diverse datasets. Our multi-dimensional evaluation framework mirrors the clinical workflow—spanning anatomical recognition, lesion analysis, spatial localization, and surgical operations—to holistically gauge the perceptual and diagnostic abilities of MLLMs in realistic scenarios. We benchmark 23 state-of-the-art models, including general-purpose, medical-specialized, and proprietary MLLMs, and establish human clinician performance as a reference standard. Our extensive experiments reveal: (1) proprietary MLLMs outperform open-source and medical-specialized models overall, but still trail human experts; (2) medical-domain supervised fine-tuning substantially boosts task-specific accuracy; and (3) model performance remains sensitive to prompt format and clinical task complexity. EndoBench establishes a new standard for evaluating and advancing MLLMs in endoscopy, highlighting both progress and persistent gaps between current models and expert clinical reasoning. We publicly release our benchmark and code.



Paperid:5041
Authors:Guancheng Wan, Xu Cheng, Run Liu, Wenke Huang, Zitong Shi, Pinyi Jin, Guibin Zhang, Bo Du, Mang Ye
Abstract:
Abstract:Federated Graph Learning (FGL) has been shown to be particularly effective in enabling collaborative training of Graph Neural Networks (GNNs) in decentralized settings. Model-heterogeneous FGL further enhances practical applicability by accommodating client preferences for diverse model architectures. However, existing model-heterogeneous approaches primarily target Euclidean data and fail to account for a crucial aspect of graph-structured data: topological relationships. To address this limitation, we propose **TRUST**, a novel knowledge distillation-based **model-heterogeneous FGL** framework. Specifically, we propose Progressive Curriculum Node Scheduler to progressively introduce challenging nodes based on learning difficulty. In Adaptive Curriculum Distillation Modulator, we propose an adaptive temperature modulator that dynamically adjusts knowledge distillation temperature to accommodate varying client capabilities and graph complexity. Moreover, we leverage Wasserstein‑Driven Affinity Distillation to enable models to capture cross-class structural relationships through optimal transport. Extensive experiments on multiple graph benchmarks and model-heterogeneous settings show that **TRUST** outperforms existing methods, achieving an average 3.6\% $\uparrow$ performance gain, particularly under moderate heterogeneity conditions. The code is available for anonymous access at https://anonymous.4open.science/r/TRUST-NeurIPS2025.



Paperid:5042
Authors:Zhixun Chen, Ping Guo, Wenhan Han, Yifan Zhang, BINBINLIU, Haobin Lin, Fengze Liu, Yan Zhao, Bingni Zhang, Taifeng Wang, Yin Zheng, Trevor Cohn, Meng Fang
Abstract:
Data quality is a critical driver of large language model performance, yet existing model‐based selection methods focus almost exclusively on English. We introduce MuRating, a scalable framework that transfers high‐quality English data‐quality signals into a single rater for 17 target languages. MuRating aggregates multiple English “raters” via pairwise comparisons to learn unified document‐quality scores, then projects these judgments through translation to train a multilingual evaluator on monolingual, cross‐lingual, and parallel text pairs. Applied to Common Crawl data, MuRating selects balanced subsets of English and multilingual content to pretrain a 1.2 B-parameter LLaMA model. Compared to strong baselines, including Qurater, AskLLM, DCLM and so on, our approach boosts average accuracy on both English benchmarks and multilingual evaluations, with especially large gains on knowledge‐intensive tasks. We further analyze translation fidelity, selection biases, and underrepresentation of narrative material, outlining directions for future work.



Paperid:5043
Authors:Yancheng Zhang, Guangyu Sun, Chen Chen
Abstract:
Novel view synthesis (NVS) is crucial in computer vision and graphics, with wide applications in AR, VR, and autonomous driving. While 3D Gaussian Splatting (3DGS) enables real-time rendering with high appearance fidelity, it suffers from multi-view inconsistencies, limiting geometric accuracy. In contrast, 2D Gaussian Splatting (2DGS) enforces multi-view consistency but compromises texture details. To address these limitations, we propose Exchangeable Gaussian Splatting (EGGS), a hybrid representation that integrates 2D and 3D Gaussians to balance appearance and geometry. To achieve this, we introduce Hybrid Gaussian Rasterization for unified rendering, Adaptive Type Exchange for dynamic adaptation between 2D and 3D Gaussians, and Frequency-Decoupled Optimization that effectively exploits the strengths of each type of Gaussian representation. Our CUDA-accelerated implementation ensures efficient training and inference. Extensive experiments demonstrate that EGGS outperforms existing methods in rendering quality, geometric accuracy, and efficiency, providing a practical solution for high-quality NVS.



Paperid:5044
Authors:Hongkai Lin, Dingkang Liang, Mingyang Du, Xin Zhou, Xiang Bai
Abstract:
Generative depth estimation methods leverage the rich visual priors stored in pretrained text-to-image diffusion models, demonstrating astonishing zero-shot capability. However, parameter updates during training lead to catastrophic degradation in the image generation capability of the pretrained model. We introduce MERGE, a unified model for image generation and depth estimation, starting from a fixed-parameters pretrained text-to-image model. MERGE demonstrates that the pretrained text-to-image model can do more than image generation but also expand to depth estimation effortlessly. Specifically, MERGE introduces a plug-and-play framework that enables seamless switching between image generation and depth estimation modes through simple and pluggable converters. Meanwhile, we propose a Group Reuse Mechanism to encourage parameter reuse and improve the utilization of the additional learnable parameter. MERGE unleashes the powerful depth estimation capability of the pretrained text-to-image model while preserving its original image generation ability. Compared to other unified models for image generation and depth estimation, MERGE achieves state-of-the-art performance across multiple depth estimation benchmarks. The code and model will be made available.



Paperid:5045
Authors:Lucas Berry, David Meger
Abstract:
This work introduces a novel approach, Pairwise Epistemic Estimators (PairEpEsts), for epistemic uncertainty estimation in ensemble models for regression tasks using pairwise-distance estimators (PaiDEs). By utilizing the pairwise distances between model components, PaiDEs establish bounds on entropy. We leverage this capability to enhance the performance of Bayesian Active Learning by Disagreement (BALD). Notably, unlike sample-based Monte Carlo estimators, PairEpEsts can estimate epistemic uncertainty up to 100 times faster and demonstrate superior performance in higher dimensions. To validate our approach, we conducted a varied series of regression experiments on commonly used benchmarks: 1D sinusoidal data,Pendulum,Hopper,Ant, andHumanoid, demonstrating PairEpEsts’ advantage over baselines in high-dimensional regression active learning.



Paperid:5046
Authors:‪Yotam Alexander‬‏, Yonatan Slutzky, Yuval Ran-Milo, Nadav Cohen
Abstract:
Conventional wisdom attributes the mysterious generalization abilities of overparameterized neural networks to gradient descent (and its variants). The recent volume hypothesis challenges this view: it posits that these generalization abilities persist even when gradient descent is replaced by Guess & Check (G&C), i.e., by drawing weight settings until one that fits the training data is found. The validity of the volume hypothesis for wide and deep neural networks remains an open question. In this paper, we theoretically investigate this question for matrix factorization (with linear and non-linear activation)—a common testbed in neural network theory. We first prove that generalization under G&C deteriorates with increasing width, establishing what is, to our knowledge, the first case where G&C is provably inferior to gradient descent. Conversely, we prove that generalization under G&C improves with increasing depth, revealing a stark contrast between wide and deep networks, which we further validate empirically. These findings suggest that even in simple settings, there may not be a simple answer to the question of whether neural networks need gradient descent to generalize well.



Paperid:5047
Authors:Wen Dong
Abstract:
As machine learning increasingly relies on large, opaque foundation models powering generative and agentic AI, deploying these systems in safety-critical settings demands rigorous guarantees on their generalization beyond training data. PAC-Bayes theory offers principled certificates linking training performance to generalization risk, yet existing approaches are rarely practical: simple theoretical priors yield vacuous bounds, while data-dependent priors trained separately are computationally costly or introduce bias. To bridge this fundamental gap, we propose a localized PAC-Bayes prior—a structured, computationally efficient prior softly concentrated near parameters favored during standard training, enabling effective exploration without costly data splits. By integrating this localized prior directly into standard training loss, we produce practically tight generalization certificates without workflow disruption. Theoretically, under standard neural tangent kernel assumptions, our bound shrinks as networks widen and datasets grow, becoming negligible in practical regimes. Empirically, we certify generalization across image classification, NLP fine-tuning, and semantic segmentation, typically within three percentage points of test errors at ImageNet scale, while providing rigorous guarantees for individual predictions, selective rejection, and robustness.



Paperid:5048
Authors:Yuwei Du, Jie Feng, Jie Zhao, Yong Li
Abstract:
Trajectory modeling, which includes research on trajectory data pattern mining and future prediction, has widespread applications in areas such as life services, urban transportation, and public administration. Numerous methods have been proposed to address specific problems within trajectory modeling. However, the heterogeneity of data and the diversity of trajectory tasks make effective and reliable trajectory modeling an important yet highly challenging endeavor, even for domain experts. In this paper, we propose TrajAgent, an agent framework powered by large language models (LLMs), designed to facilitate robust and efficient trajectory modeling through automation modeling. This framework leverages and optimizes diverse specialized models to address various trajectory modeling tasks across different datasets effectively. Extensive experiments on five tasks using four real-world datasets demonstrate the effectiveness of TrajAgent in automated trajectory modeling, achieving a performance improvement of 2.38%-34.96% over baseline methods.



Paperid:5049
Authors:Chaehyun Kim, Heeseong Shin, Heeji Yoon, Eunbeen Hong, Anurag Arnab, Paul Hongsuck Seo, Sunghwan Hong, Seungryong Kim
Abstract:
Text-to-image diffusion models excel at translating language prompts into photorealistic images by implicitly grounding textual concepts in their cross-modal attention mechanisms. Prior work has exploited these attention maps for downstream tasks such as editing, inpainting, and zero-shot open-vocabulary semantic segmentation (OVSS), but a detailed understanding of how these maps contribute to image generation remains limited. Recent architectural advances like Multi-Modal Diffusion Transformers (MM-DiTs) introduce joint self-attention over concatenated image and text tokens, enabling richer and more scalable cross-modal alignment. In this work, we systematically analyze the attention structures of MM-DiT, focusing on how specific heads and layers propagate semantic information and influence generation quality. By decomposing attention score distributions and attention norms across layers, we identify a subset of heads that consistently align text tokens with spatially coherent image regions, naturally yielding high-quality zero-shot segmentation masks. We then introduce a lightweight LoRA-based fine-tuning method to enhance the semantic grouping capabilities of these heads without degrading— and often improving—image fidelity. Our findings demonstrate that semantic alignment is an emergent property of diffusion transformers and can be selectively amplified to improve both dense recognition and generative performance, paving the way toward unified models that bridge generation and perception.



Paperid:5050
Authors:Rongzhe Wei, Peizhi Niu, Hans Hao-Hsun Hsu, Ruihan Wu, Haoteng Yin, Mohsen Ghassemi, Yifan Li, Vamsi Potluru, Eli Chien, Kamalika Chaudhuri, Olgica Milenkovic, Pan Li
Abstract:
Machine unlearning techniques aim to mitigate unintended memorization in large language models (LLMs). However, existing approaches predominantly focus on the explicit removal of isolated facts, often overlooking latent inferential dependencies and the non-deterministic nature of knowledge within LLMs. Consequently, facts presumed forgotten may persist implicitly through correlated information. To address these challenges, we propose a knowledge unlearning evaluation framework that more accurately captures the implicit structure of real-world knowledge by representing relevant factual contexts as knowledge graphs with associated confidence scores. We further develop an inference-based evaluation protocol leveraging powerful LLMs as judges; these judges reason over the extracted knowledge subgraph to determine unlearning success. Our LLM judges utilize carefully designed prompts and are calibrated against human evaluations to ensure their trustworthiness and stability. Extensive experiments on our newly constructed benchmark demonstrate that our framework provides a more realistic and rigorous assessment of unlearning performance. Moreover, our findings reveal that current evaluation strategies tend to overestimate unlearning effectiveness.



Paperid:5051
Authors:Jiacong Zhou, Jiaxu Miao, xianyun wang, Jun Yu
Abstract:
Large Language Models (LLMs) have shown the promising planning capabilities for robotic manipulation, which advances the development of embodied intelligence significantly. However, existing LLM-driven robotic manipulation approaches excel at simple pick-and-place tasks but are insufficient for complex manipulation tasks due to inaccurate procedural knowledge. Besides, for embodied intelligence, equipping a large scale LLM is energy-consuming and inefficient, which affects its real-world application.To address the above problems, we propose Hierarchical Procedural Knowledge Graphs (\textbf{HP-KG}) to enhance LLMs for complex robotic planning while significantly reducing the demand for LLM scale in robotic manipulation. Considering that the complex real-world tasks require multiple steps, and each step is composed of robotic-understandable atomic actions, we design a hierarchical knowledge graph structure to model the relationships between tasks, steps, and actions. This design bridges the gap between human instructions and robotic manipulation actions. To construct HP-KG, we develop an automatic knowledge graph construction framework powered by LLM-based multi-agents, which eliminates costly manual efforts while maintaining high-quality graph structures. The resulting HP-KG encompasses over 40k activity steps across more than 6k household tasks, spanning diverse everyday scenarios. Extensive experiments demonstrate that small scale LLMs (7B) enhanced by our HP-KG significantly improve the planning capabilities, which are stronger than 72B LLMs only. Encouragingly, our approach remains effective on the most powerful GPT-4o model.



Paperid:5052
Authors:Zixuan XIa, Aram Davtyan, Paolo Favaro
Abstract:
We propose KOALA++, a scalable Kalman-based optimization algorithm that explicitly models structured gradient uncertainty in neural network training. Unlike second-order methods, which rely on expensive second order gradient calculation, our method directly estimates the parameter covariance matrix by recursively updating compact gradient covariance products. This design improves upon the original KOALA framework that assumed diagonal covariance by implicitly capturing richer uncertainty structure without storing the full covariance matrix and avoiding large matrix inversions. Across diverse tasks, including image classification and language modeling, KOALA++ achieves accuracy on par or better than state-of-the-art second-order optimizers while maintaining the efficiency of first-order methods.



Paperid:5053
Authors:Ji-An Li, Huadong Xiong, Robert Wilson, Marcelo G Mattar, Marcus Benna
Abstract:
Large language models (LLMs) can sometimes report the strategies they actually use to solve tasks, but they can also fail to do so. This suggests some degree of metacognition --- the capacity to monitor one's own cognitive processes for subsequent reporting and self-control. Metacognitive abilities enhance AI capabilities but raise safety concerns, as models might obscure their internal processes to evade neural-activation-based oversight mechanisms designed to detect harmful behaviors. Given society's increased reliance on these models, it is critical that we understand the limits of their metacognitive abilities, particularly their ability to monitor their internal activations. To address this, we introduce a neuroscience-inspired \emph{neurofeedback} paradigm designed to quantify the ability of LLMs to explicitly \textit{report} and \textit{control} their activation patterns. By presenting models with sentence-label pairs where labels correspond to sentence-elicited internal activations along specific directions in the neural representation space, we demonstrate that LLMs can learn to report and control these activations. The performance varies with several factors: the number of example pairs provided, the semantic interpretability of the target neural direction, and the variance explained by that direction. These results reveal a ``metacognitive space'' with dimensionality much lower than the model's neural space, suggesting LLMs can monitor only a subset of their neural mechanisms. Our findings provide empirical evidence quantifying metacognitive capabilities in LLMs, with significant implications for AI safety.



Paperid:5054
Authors:Qiming Hu, Linlong Fan, Yiyan Luo, Yuhang Yu, Xiaojie Guo, Qingnan Fan
Abstract:
The introduction of generative models has significantly advanced image super-resolution (SR) in handling real-world degradations. However, they often incur fidelity-related issues, particularly distorting textual structures. In this paper, we introduce a novel diffusion-based SR framework, namely TADiSR, which integrates text-aware attention and joint segmentation decoders to recover not only natural details but also the structural fidelity of text regions in degraded real-world images. Moreover, we propose a complete pipeline for synthesizing high-quality images with fine-grained full-image text masks, combining realistic foreground text regions with detailed background content. Extensive experiments demonstrate that our approach substantially enhances text legibility in super-resolved images, achieving state-of-the-art performance across multiple evaluation metrics and exhibiting strong generalization to real-world scenarios. Our code will be open-sourced.



Paperid:5055
Authors:Xiyao Wang, Zhengyuan Yang, Chao Feng, Hongjin Lu, Linjie Li, Chung-Ching Lin, Kevin Lin, Furong Huang, Lijuan Wang
Abstract:
We introduce ThinkLite-VL, a family of visual reasoning models that achieve state-of-the-art (SoTA) performance using an order of magnitude fewer training samples, relying purely on reinforcement fine-tuning (RFT) self-improvement without any knowledge distillation. Our central insight is that sample difficulty critically influences RFT effectiveness: appropriately challenging examples can drive substantial reasoning improvements, even in low-data regimes. However, quantifying sample difficulty in a reliable and scalable manner remains non-trivial. To address this, we repurpose Monte Carlo Tree Search (MCTS) to measure sample difficulty via the number of reasoning iterations a vision-language model (VLM) requires to solve each instance. This MCTS-based selection procedure identifies samples that induce deeper reasoning while remaining solvable, allowing us to filter a high-quality subset from 70k open-source examples spanning math, natural image understanding, and chart comprehension. Using this approach, we select just 11k challenging samples for RFT on Qwen2.5-VL-7B-Instruct and 7.5k samples for Qwen2.5-VL-72B-Instruct. The resulting models, ThinkLite-VL-7B and ThinkLite-VL-72B, significantly outperform their respective base models across eight visual reasoning benchmarks. In particular, ThinkLite-VL-7B improves the average performance of Qwen2.5-VL-7B-Instruct by 7\% and surpasses all existing 7B-level models, as well as much larger models such as GPT-4o, O1 and Qwen2.5-VL-72B, achieving a new SoTA score of 75.1 on MathVista. ThinkLite-VL-72B further advances the SoTA frontier, achieving an accuracy of 79.7 on MathVista and an average benchmark improvement of 4.42 over the open-source SOTA. These results demonstrate that MCTS-guided difficulty filtering provides a scalable and effective path toward data-efficient self-improvement in multimodal reasoning.



Paperid:5056
Authors:Philip Boeken, Onno Zoeter, Joris Mooij
Abstract:
In performative prediction, forecasts influence the outcomes they aim to predict, undermining the existence of correct forecasts and standard methods of elicitation and estimation. We show that conditioning forecasts on covariates that separate them from the outcome renders the target distribution forecast-invariant, restoring well-posedness of the forecasting problem. However, even under this condition, classical proper scoring rules fail to elicit correct forecasts. We prove a general impossibility result and identify two solutions: (i) in decision-theoretic settings, truthful and outcome-optimising scoring is possible if forecasts are separating; (ii) batch-scoring based on empirical divergence or inverse probability weights can yield correct forecasts. Applying these insights to parameter estimation, conditional forecasts and proper scoring rules enable performatively stable estimation of performatively correct parameters, resolving the issues raised by Perdomo et al. (2020). Our results expose fundamental limits of classical forecast evaluation and offer new tools for reliable and accurate forecasting in performative settings.



Paperid:5057
Authors:Yuqing Zhang, Yue Han, Shuanghe Zhu, Haoxiang Wu, Hangqi Li, Shengyu Zhang, Junchi Yan, Zemin Liu, Kun Kuang, Huaiyong Dou, Yongquan Zhang, Fei Wu
Abstract:
Analyzing ancient manuscripts has traditionally been a labor-intensive and time-consuming task for philologists. Recent advancements in LMMs have demonstrated their potential across various domains, yet their effectiveness in manuscript study remains largely unexplored. In this paper, we present MS-Bench, the first comprehensive benchmark co-developed with archaeologists, comprising 5,076 high-resolution images from 4th to 14th century and 9,982 expert-curated questions across nine sub-tasks aligned with archaeological workflows. Through four prompting strategies, we systematically evaluate 32 LMMs on their effectiveness, robustness, and cultural contextualization. Our analysis reveals scale-driven performance and reliability improvements, prompting strategies' impact on performance (CoT has two-sides effect, while Retrieval-augmented Visual Context provide consistent boost), and task preferences based on LMM's visual capabilities. Although current LMMs are not yet capable of replacing domain expertise, they show promising potential to accelerate manuscript research through future human-AI collaboration.



Paperid:5058
Authors:Baoyi He, Luotian Yuan, Ying Wei, Fei Wu
Abstract:
The emergence of large language models (LLMs) prompts fine-tuning foundation LLMs to solve real-world chemical problems. However, these chemical LLMs are tailored to specific task formats or narrow content domains, which limits their capacity for comprehensive knowledge integration and cross-task generalization. Model merging has recently been proposed as a promising solution and demonstrated promising performance in the domain of natural language processing (NLP), enabling the integration of multiple expert LLMs into a unified model without requiring access to original training data or incurring significant computational costs. The widespread prevalence of in-house training data in the chemical domain underscores the need for practical and privacy-preserving model merging method for chemical LLM integration. % Although existing model merging methods %attempt to %minimize %harmonize interference among models and % have , However, two key characteristics of the chemical domain hinder effective model merging: 1) significant disparities among differentiated LLMs due to task-specific specializations, and 2) a highly imbalanced distribution of downstream functionality, with sparse model coverage for niche tasks and a surplus of models targeting widely studied problems. These factors intensify model inconsistencies, such as parameter interference and accumulated fine-tuning noise, ultimately degrade performance when integrating heterogeneous chemical LLMs. To this end, we propose Curriculum Model Merging (CMM), a method that constructs curriculum to progressively merge the expert chemical LLMs in a moderate and continual manner, aiming to harmonize the inconsistencies and mitigate the interference. Our comprehensive experiments on two benchmark datasets show that our proposed method concentrates task-specific expertise and outperforms the state-of-the-art methods by 29.03\% in terms of an overall average performance score. Moreover, CMM facilitates chemical knowledge generalization across prediction and generative tasks without sacrificing robustness, exhibiting promising merging performance under both expert-abundant and expert-sparse scenarios.



Paperid:5059
Authors:Wang, Xiao Yang, Qingyong Hu, Jack Tang, Can Liu, Dengbo He, Yuntao Wang, Yingcong Chen, Kaishun Wu
Abstract:
Robust and unobtrusive in-vehicle physiological monitoring is crucial for ensuring driving safety and user experience. While remote physiological measurement (RPM) offers a promising non-invasive solution, its translation to real-world driving scenarios is critically constrained by the scarcity of comprehensive datasets. Existing resources are often limited in scale, modality diversity, the breadth of biometric annotations, and the range of captured conditions, thereby omitting inherent real-world challenges in driving. Here, we present PhysDrive, the first large-scale multimodal dataset for contactless in-vehicle physiological sensing with dedicated consideration on various modality settings and driving factors. PhysDrive collects data from 48 drivers, including synchronized RGB, near-infrared camera, and raw mmWave radar data, accompanied with six synchronized ground truths (ECG, BVP, Respiration, HR, RR, and SpO2). It covers a wide spectrum of naturalistic driving conditions, including driver motions, dynamic natural light, vehicle types, and road conditions. We extensively evaluate both signal‑processing and deep‑learning methods on PhysDrive, establishing a comprehensive benchmark across all modalities, and release full open‑source code with compatibility for mainstream public toolboxes. We envision PhysDrive will serve as a foundational resource and accelerate research on multimodal driver monitoring and smart‑cockpit systems.



Paperid:5060
Authors:Jiyoung Lee, Seungho Kim, Jieun Han, Jun-Min Lee, Kitaek Kim, Alice Oh, Edward Choi
Abstract:
Large Language Models (LLMs) are predominantly evaluated on Standard American English (SAE), often overlooking the diversity of global English varieties.This narrow focus may raise fairness concerns as degraded performance on non-standard varieties can lead to unequal benefits for users worldwide.Therefore, it is critical to extensively evaluate the linguistic robustness of LLMs on multiple non-standard English varieties.We introduce Trans-EnV, a framework that automatically transforms SAE datasets into multiple English varieties to evaluate the linguistic robustness. Our framework combines (1) linguistics expert knowledge to curate variety-specific features and transformation guidelines from linguistic literature and corpora, and (2) LLM-based transformations to ensure both linguistic validity and scalability.Using Trans-EnV, we transform six benchmark datasets into 38 English varieties and evaluate seven state-of-the-art LLMs.Our results reveal significant performance disparities, with accuracy decreasing by up to 46.3% on non-standard varieties.These findings highlight the importance of comprehensive linguistic robustness evaluation across diverse English varieties. Each construction of Trans-EnV was validated through rigorous statistical testing and consultation with a researcher in the field of second language acquisition, ensuring its linguistic validity.Our code and datasets are publicly available.



Paperid:5061
Authors:Yifan Zhou, Sachin Grover, Mohamed Mistiri, Kamalesh Kalirathinam, Pratyush Kerhalkar, Swaroop Mishra, Neelesh Kumar, Sanket Gaurav, Oya Aran, Heni Ben Amor
Abstract:
Reinforcement Learning (RL) traditionally relies on scalar reward signals, limiting its ability to leverage the rich semantic knowledge often available in real-world tasks. In contrast, humans learn efficiently by combining numerical feedback with language, prior knowledge, and common sense. We introduce Prompted Policy Search (ProPS), a novel RL method that unifies numerical and linguistic reasoning within a single framework. Unlike prior work that augments existing RL components with language, ProPS places a large language model (LLM) at the center of the policy optimization loop—directly proposing policy updates based on both reward feedback and natural language input. We show that LLMs can perform numerical optimization in-context, and that incorporating semantic signals, such as goals, constraints, and strategy hints can lead to more informed exploration and sample-efficient learning. ProPS is evaluated across 15 Gymnasium tasks, spanning classic control, Atari games, and MuJoCo environments, and compared to seven widely-adopted RL algorithms (e.g., PPO, SAC, TRPO). It outperforms all baselines on 8 out of 15 tasks and demonstrates substantial gains when provided with domain knowledge. These results highlight the potential of unifying semantics and numerics for transparent, generalizable, and human-aligned reinforcement learning.



Paperid:5062
Authors:Joongkyu Lee, Seouh-won Yi, Min-hwan Oh
Abstract:
Abstract:We study online preference-based reinforcement learning (PbRL) with the goal of improving sample efficiency. While a growing body of theoretical work has emerged—motivated by PbRL’s recent empirical success, particularly in aligning large language models (LLMs)—most existing studies focus only on pairwise comparisons. A few recent works [Zhu et al., 2023, Mukherjee et al., 2024, Thekumparampil et al., 2024] have explored using multiple comparisons and ranking feedback, but their performance guarantees fail to improve—and can even deteriorate—as the feedback length increases, despite the richer information available. To address this gap, we adopt the Plackett–Luce (PL) model for ranking feedback over action subsets and propose **M-AUPO**, an algorithm that selects multiple actions by maximizing the average uncertainty within the offered subset. We prove that **M-AUPO** achieves a suboptimality gap of $\tilde{\mathcal{O}}\left( \frac{d}{T} \sqrt{ \sum_{t=1}^T \frac{1}{|S_t|}} \right)$, where $T$ is the total number of rounds, $d$ is the feature dimension, and $|S_t|$ is the size of the subset at round $t$. This result shows that larger subsets directly lead to improved performance and, notably, the bound avoids the exponential dependence on the unknown parameter’s norm, which was a fundamental limitation in most previous works. Moreover, we establish a near-matching lower bound of $\Omega \left( \frac{d}{K \sqrt{T}} \right)$, where $K$ is the maximum subset size. To the best of our knowledge, this is the first theoretical result in PbRL with ranking feedback that explicitly shows improved sample efficiency as a function of the subset size.



Paperid:5063
Authors:Jiachen Jiang, Jinxin Zhou, Bo Peng, Xia Ning, Zhihui Zhu
Abstract:
Abstract:Achieving better alignment between vision embeddings and Large Language Models (LLMs) is crucial for enhancing the abilities of Multimodal LLMs (MLLMs), particularly for recent models that rely on powerful pretrained vision encoders and LLMs. A common approach to connect the pretrained vision encoder and LLM is through a projector applied after the vision encoder. However, the projector is often trained to enable the LLM to generate captions, and hence the mechanism by which LLMs understand each vision token remains unclear. In this work, we first investigate the role of the projector in compressing vision embeddings and aligning them with word embeddings. We show that the projector significantly compresses visual information, removing redundant details while preserving essential elements necessary for the LLM to understand visual content. We then examine patch-level alignment---the alignment between each vision patch and its corresponding semantic words---and propose a $\textit{multi-semantic alignment hypothesis}$. Our analysis indicates that the projector trained by caption loss improves patch-level alignment but only to a limited extent, resulting in weak and coarse alignment. To address this issue, we propose $\textit{patch-aligned training}$ to efficiently enhance patch-level alignment. Our experiments show that patch-aligned training (1) achieves stronger compression capability and improved patch-level alignment, enabling the MLLM to generate higher-quality captions, (2) improves the MLLM's performance by 16% on referring expression grounding tasks, 4% on question-answering tasks, and 3% on modern instruction-following benchmarks when using the same supervised fine-tuning (SFT) setting. The proposed method can be easily extended to other multimodal models.



Paperid:5064
Authors:Minhak Song, Beomhan Baek, Kwangjun Ahn, Chulhee Yun
Abstract:
As both model and dataset sizes continue to scale rapidly, conventional pretraining strategies with fixed compute budgets—such as cosine learning rate schedules—are increasingly inadequate for large-scale training. Recent alternatives, including warmup-stable-decay (WSD) schedules and weight averaging, offer greater flexibility. However, WSD relies on explicit decay phases to track progress, while weight averaging addresses this limitation at the cost of additional memory. In search of a more principled and scalable alternative, we revisit the Schedule-Free (SF) method [Defazio et al., 2024], which has shown strong empirical performance across diverse settings. We show that SF-AdamW effectively navigates the "river" structure of the loss landscape without decay phases or auxiliary averaging, making it particularly suitable for continuously scaling training workloads. To understand this behavior, we conduct a theoretical and empirical analysis of SF dynamics, revealing that it implicitly performs weight averaging without memory overhead. Guided by this analysis, we propose a refined variant of SF that improves robustness to momentum and performs better under large batch sizes, addressing key limitations of the original method. Together, these results establish SF as a practical, scalable, and theoretically grounded approach for language model training.



Paperid:5065
Authors:Junchen Li, Rongzheng Wang, Yihong Huang, Qizhi Chen, Jiasheng Zhang, Shuang Liang
Abstract:
Retrieval-augmented generation (RAG) greatly enhances large language models (LLMs) performance in knowledge-intensive tasks. However, naive RAG methods struggle with multi-hop question answering due to their limited capacity to capture complex dependencies across documents. Recent studies employ graph-based RAG to capture document connections. However, these approaches often result in a loss of semantic coherence and introduce irrelevant noise during node matching and subgraph construction. To address these limitations, we propose NeuroPath, an LLM-driven semantic path tracking RAG framework inspired by the path navigational planning of place cells in neurobiology. It consists of two steps: \textit{Dynamic Path Tracking} and \textit{Post-retrieval Completion}. Dynamic Path Tracking performs goal-directed semantic path tracking and pruning over the constructed knowledge graph (KG), improving noise reduction and semantic coherence. Post-retrieval Completion further reinforces these benefits by conducting second-stage retrieval using intermediate reasoning and the original query to refine the query goal and complete missing information in the reasoning path. NeuroPath surpasses current state-of-the-art baselines on three multi-hop QA datasets, achieving average improvements of 16.3\% on recall@2 and 13.5\% on recall@5 over advanced graph-based RAG methods. Moreover, compared to existing iter-based RAG methods, NeuroPath achieves higher accuracy and reduces token consumption by 22.8\%. Finally, we demonstrate the robustness of NeuroPath across four different LLMs (Llama3.1, GLM4, Mistral0.3 and Gemma3), all achieving state-of-the-art performance. Code is available athttps://anonymous.4open.science/r/NeuroPath



Paperid:5066
Authors:Michal Nauman, Marek Cygan, Carmelo Sferrazza, Aviral Kumar, Pieter Abbeel
Abstract:
Recent advances in language modeling and vision stem from training large models on diverse, multi‑task data. This paradigm has had limited impact in value-based reinforcement learning (RL), where improvements are often driven by small models trained in a single-task context. This is because in multi-task RL sparse rewards and gradient conflicts make optimization of temporal difference brittle. Practical workflows for generalist policies therefore avoid online training, instead cloning expert trajectories or distilling collections of single‑task policies into one agent. In this work, we show that the use of high-capacity value models trained via cross-entropy and conditioned on learnable task embeddings addresses the problem of task interference in online RL, allowing for robust and scalable multi‑task training. We test our approach on 7 multi-task benchmarks with over 280 unique tasks, spanning high degree-of-freedom humanoid control and discrete vision-based RL. We find that, despite its simplicity, the proposed approach leads to state-of-the-art single and multi-task performance, as well as sample-efficient transfer to new tasks.



Paperid:5067
Authors:Viet Anh Khoa Tran, Emre Neftci, Willem Wybo
Abstract:
Biological brains learn continually from a stream of unlabeled data, while integrating specialized information from sparsely labeled examples without compromising their ability to generalize.Meanwhile, machine learning methods are susceptible to catastrophic forgetting in this natural learning setting, as supervised specialist fine-tuning degrades performance on the original task.We introduce task-modulated contrastive learning (TMCL), which takes inspiration from the biophysical machinery in the neocortex, using predictive coding principles to integrate top-down information continually and without supervision.We follow the idea that these principles build a view-invariant representation space, and that this can be implemented using a contrastive loss.Then, whenever labeled samples of a new class occur, new affine modulations are learned that improve separation of the new class from all others, without affecting feedforward weights. By co-opting the view-invariance learning mechanism, we then train feedforward weights to match the unmodulated representation of a data sample to its modulated counterparts. This introduces modulation invariance into the representation space, and, by also using past modulations, stabilizes it.Our experiments show improvements in both class-incremental and transfer learning over state-of-the-art unsupervised approaches, as well as over comparable supervised approaches, using as few as 1% of available labels.Taken together, our work suggests that top-down modulations play a crucial role in balancing stability and plasticity.



Paperid:5068
Authors:Tongyang Li, Xinzhao Wang, Yexin Zhang
Abstract:
Abstract:Computing Nash equilibria of zero-sum games in classical and quantum settings is extensively studied. For general-sum games, computing Nash equilibria is PPAD-hard and the computing of a more general concept called correlated equilibria has been widely explored in game thoery. In this paper, we initiate the study of quantum algorithms for computing $\varepsilon$-approximate correlated equilibria (CE) and coarse correlated equilibria (CCE) in multi-player normal-form games. Our approach utilizes quantum improvements to the multi-scale Multiplicative Weight Update (MWU) method for CE calculations, achieving a query complexity of $\tilde{O}(m\sqrt{n})$ for fixed $\varepsilon$. For CCE, we extend techniques from quantum algorithms for zero-sum games to multi-player settings, achieving query complexity $\tilde{O}(m\sqrt{n}/\varepsilon^{2.5})$. Both algorithms demonstrate a near-optimal scaling in the number of players $m$ and actions $n$, as confirmed by our quantum query lower bounds.



Paperid:5069
Authors:Quanjian Song, Donghao Zhou, Jingyu Lin, Fei Shen, Jiaze Wang, Xiaowei Hu, Cunjian Chen, Pheng-Ann Heng
Abstract:
Recent text-to-image models have revolutionized image generation, but they still struggle with maintaining narrative consistency across images. While existing solutions focus on character consistency, they overlook the crucial role of scenes in storytelling, restricting their creativity in practice. This paper introduces scene-oriented story generation, addressing two key challenges: (i) scene planning, where current methods fail to ensure narrative coherence across scenes due to independent scene generation, and(ii) scene consistency, which remains largely unexplored in terms of maintaining coherence across multiple storylines. We propose SceneDecorator, a training-free framework that employs VLM-Guided Scene Planning to ensure narrative coherence between different scenes in a ``global-to-local" manner, and Long-Term Scene-Sharing Attention to maintain scene consistency and subject style diversity across different stories. Comprehensive experiments demonstrate the superior performance of SceneDecorator, highlighting its potential to unleash creativity in the fields of arts, films, and games. Code will be released.



Paperid:5070
Authors:Hanqun Cao, Haosen Shi, Chenyu Wang, Sinno Pan, Pheng-Ann Heng
Abstract:
Abstract:The design of biological sequences is essential for engineering functional biomolecules that contribute to advancements in human health and biotechnology. Recent advances in diffusion models, with their generative power and efficient conditional sampling, have made them a promising approach for sequence generation. To enhance model performance on limited data and enable multi-objective design and optimization, reinforcement learning (RL)-based fine-tuning has shown great potential. However, existing post-sampling and fine-tuning methods either lack stability in discrete optimization when avoiding gradients or incur high computational costs when employing gradient-based approaches, creating significant challenges for achieving both control and stability in the tuning process.To address these limitations, we propose GLID$^2$E, a gradient-free RL-based tuning approach for discrete diffusion models. Our method introduces a clipped likelihood constraint to regulate the exploration space and implements reward shaping to better align the generative process with design objectives, ensuring a more stable and efficient tuning process. By integrating these techniques, GLID$^2$E mitigates training instabilities commonly encountered in RL and diffusion-based frameworks, enabling robust optimization even in challenging biological design tasks. In the DNA sequence and protein sequence design systems, GLID$^2$E achieves competitive performance in function-based design while maintaining computational efficiency and a flexible tuning mechanism.



Paperid:5071
Authors:Yang Li, Lvda Chen, Haonan Wang, Runzhong Wang, Junchi Yan
Abstract:
While diffusion models have shown promise for combinatorial optimization (CO), their inference-time scaling cost-efficiency remains relatively underexplored. Existing methods improve solution quality by increasing denoising steps, but the performance often becomes saturated quickly. This paper proposes GenSCO to systematically scale diffusion solvers by an orthogonal dimension of inference-time computation beyond denoising step expansion, i.e., search-driven generation. GenSCO takes generation as a search operator rather than a complete solving process, where each operator cycle combines solution disruption (via local search operators) and diffusion sampling, enabling iterative exploration of the learned solution space. Rather than over-refining current solutions, this paradigm encourages the model to leave local optima and explore a broader area of the solution space, ensuring a more consistent scaling effect. The search loop is supported by a search-friendly solution-enhancement training procedure that incorporates a rectified flow model learning to establish diffusion trajectories between suboptimal solutions and the optimal ones. The flow model is empowered by a lightweight transformer architecture to learn neural ODEs that linearize solution trajectories, accelerating convergence of the scaling effect with efficiency. The resulting enhanced scaling efficiency and practical scalability lead to synergistic performance improvements. Extensive experiments show that GenSCO delivers performance improvements by orders of magnitude over previous state-of-the-art neural methods. Notably, GenSCO even achieves significant speedups compared to the state-of-the-art classic mathematical solver LKH3, delivering a 141x speedup to reach 0.000% optimality gap on TSP-100, and approximately a 10x speedup to reach 0.02% on TSP-500.



Paperid:5072
Authors:Tianle Li, Yongzhi Huang, Linshan Jiang, Chang Liu, Qipeng Xie, Wenfeng Du, Lu Wang, Kaishun Wu
Abstract:
Federated Learning (FL) has emerged as a distributed learning framework that allows multiple clients to train a global model while preserving data privacy collaboratively. However, FL in edge computing faces several unique challenges, including limited communication rounds, data heterogeneity, and client drift, which lead to slow convergence and poor model generalization. Meanwhile, Momentum and Sharpness-Aware Minimization (SAM) show their promotion on the convergence speed and flatness in FL, respectively.To seamlessly concatenate the Momentum and SAM while avoiding their flaws, we propose a novel approach called FedWMSAM to achieve \textit{fast and flat} FL. In detail, by introducing personalized momentum, we mitigate client inconsistency and accelerate convergence. Additionally, by dynamically adjusting the perturbation direction and leveraging momentum-based global flatness, we estimate global flat minima and use a dynamic weighting mechanism to balance the contributions of momentum and SAM adaptively throughout the training process. We conduct extensive experiments on multiple datasets and model architectures, and the results validate the effectiveness, adaptability, and robustness of our method, demonstrating its superiority in addressing the optimization challenges of Federated Learning. Our code is available at \url{https://anonymous.4open.science/r/NeurlPS_FedWMSAM-2AC5}.



Paperid:5073
Authors:Enze Shi, Pankaj Bhagwat, Zhixian Yang, Linglong Kong, Bei Jiang
Abstract:
Machine learning models have achieved widespread success but often inherit and amplify historical biases, resulting in unfair outcomes. Traditional fairness methods typically impose constraints at the prediction level, without addressing underlying biases in data representations. In this work, we propose a principled framework that adjusts data representations to balance predictive utility and fairness. Using sufficient dimension reduction, we decompose the feature space into target-relevant, sensitive, and shared components, and control the fairness–utility trade-off by selectively removing sensitive information. We provide a theoretical analysis of how prediction error and fairness gaps evolve as shared subspaces are added, and employ influence functions to quantify their effects on the asymptotic behavior of parameter estimates. Experiments on both synthetic and real-world datasets validate our theoretical insights and show that the proposed method effectively improves fairness while preserving predictive performance.



Paperid:5074
Authors:Taoran Zheng, Xing Li, Yan Yang, Xiang Gu, Zongben Xu, Jian Sun
Abstract:
Medical image reconstruction from measurement data is a vital but challenging inverse problem. Deep learning approaches have achieved promising results, but often requires paired measurement and high-quality images, which is typically simulated through a forward model, i.e., retrospective reconstruction. However, training on simulated pairs commonly leads to performance degradation on real prospective data due to the retrospective-to-prospective gap caused by incomplete imaging knowledge in simulation. To address this challenge, this paper introduces imaging Knowledge-Informed Dynamic Optimal Transport (KIDOT), a novel dynamic optimal transport framework with optimality in the sense of preserving consistency with imaging physics in transport, that conceptualizes reconstruction as finding a dynamic transport path. KIDOT learns from unpaired data by modeling reconstruction as a continuous evolution path from measurements to images, guided by an imaging knowledge-informed cost function and transport equation. This dynamic and knowledge-aware approach enhances robustness and better leverages unpaired data while respecting acquisition physics. Theoretically, we demonstrate that KIDOT naturally generalizes dynamic optimal transport, ensuring its mathematical rationale and solution existence. Extensive experiments on MRI and CT reconstruction demonstrate KIDOT's superior performance.



Paperid:5075
Authors:Taiying Peng, Jiacheng Hua, Miao Liu, Feng Lu
Abstract:
The emergence of advanced multimodal large language models (MLLMs) has significantly enhanced AI assistants' ability to process complex information across modalities. Recently, egocentric videos, by directly capturing user focus, actions, and context in an unified coordinate, offer an exciting opportunity to enable proactive and personalized AI user experiences with MLLMs. However, existing benchmarks overlook the crucial role of gaze as an indicator of user intent. To address this gap, we introduce EgoGazeVQA, an egocentric gaze-guided video question answering benchmark that leverages gaze information to improve the understanding of longer daily-life videos. EgoGazeVQA consists of gaze-based QA pairs generated by MLLMs and refined by human annotators. Our experiments reveal that existing MLLMs struggle to accurately interpret user intentions using only global visual tokens. In contrast, our gaze-guided intent prompting methods significantly enhance performance by integrating spatial, temporal, and intent-related cues. We further conduct experiments on gaze-related fine-tuning and analyze how gaze estimation accuracy impacts prompting effectiveness. These results underscore the value of gaze for more personalized and effective AI assistants in egocentric settings.



Paperid:5076
Authors:Yuxin Pan, Zhiguang Cao, Chengyang GU, Liu Liu, Peilin Zhao, Yize Chen, Fangzhen Lin
Abstract:
Existing neural methods for multi-task vehicle routing problems (VRPs) typically learn unified solvers to handle multiple constraints simultaneously. However, they often underutilize the compositional structure of VRP variants, each derivable from a common set of basis VRP variants. This critical oversight causes unified solvers to miss out the potential benefits of basis solvers, each specialized for a basis VRP variant. To overcome this limitation, we propose a framework that enables unified solvers to perceive the shared-component nature across VRP variants by proactively reusing basis solvers, while mitigating the exponential growth of trained neural solvers. Specifically, we introduce a State-Decomposable MDP (SDMDP) that reformulates VRPs by expressing the state space as the Cartesian product of basis state spaces associated with basis VRP variants. More crucially, this formulation inherently yields the optimal basis policy for each basis VRP variant. Furthermore, a Latent Space-based SDMDP extension is developed by incorporating both the optimal basis policies and a learnable mixture function to enable the policy reuse in the latent space. Under mild assumptions, this extension provably recovers the optimal unified policy of SDMDP through the mixture function that computes the state embedding as a mapping from the basis state embeddings generated by optimal basis policies. For practical implementation, we introduce the Mixture-of-Specialized-Experts Solver (MoSES), which realizes basis policies through specialized Low-Rank Adaptation (LoRA) experts, and implements the mixture function via an adaptive gating mechanism. Extensive experiments conducted across VRP variants showcase the superiority of MoSES over prior methods.



Paperid:5077
Authors:Pawan Neupane, Jian Liu, Jianlin Cheng
Abstract:
Predicting protein complex structures is essential for protein function analysis, protein design, and drug discovery. While AI methods like AlphaFold can predict accurate structural models for many protein complexes, reliably estimating the quality of these predicted models (estimation of model accuracy, or EMA) for model ranking and selection remains a major challenge. A key barrier to developing effective machine learning-based EMA methods is the lack of large, diverse, and well-annotated datasets for training and evaluation. To address this gap, we introduce PSBench, a benchmark suite comprising four large-scale, labeled datasets generated during the 15th and 16th community-wide Critical Assessment of Protein Structure Prediction (CASP15 and CASP16). PSBench includes over one million structural models covering a wide range of protein sequence lengths, complex stoichiometries, functional classes, and modeling difficulties. Each model is annotated with multiple complementary quality scores at the global, local, and interface levels. PSBench also provides multiple evaluation metrics and baseline EMA methods to facilitate rigorous comparisons. To demonstrate PSBench’s utility, we trained and evaluated GATE, a graph transformer-based EMA method, on the CASP15 data. GATE was blindly tested in CASP16 (2024), where it ranked among the top-performing EMA methods. These results highlight PSBench as a valuable resource for advancing EMA research in protein complex modeling. PSBench is publicly available at: https://github.com/BioinfoMachineLearning/PSBench .



Paperid:5078
Authors:Youssef Mansour, Reinhard Heckel
Abstract:
We investigate fingerprints in pretraining datasets for large language models (LLMs) through dataset classification experiments. Building on prior work demonstrating the existence of fingerprints or biases in popular computer vision datasets, we analyze popular open-source pretraining datasets for LLMs derived from CommonCrawl including C4, RefinedWeb, DolmaCC, RedPajama-V2, FineWeb, and DCLM-Baseline. Despite those datasets being obtained with similar curation steps, neural networks can classify surprisingly well which dataset a single text sequence belongs to, significantly better than a human can. This indicates that small differences in filtering and processing pipelines induce fingerprints, that we find are evident in formatting, vocabulary, and content distributions. Such fingerprints can negatively impact cross-dataset generalisation and we discuss a potential mitigation strategy. Additionally, we show that these fingerprints propagate through training: sequences generated by models trained on those datasets can be accurately classified by a classifier trained on the original datasets. This can offer insights into data characteristics that are typically undisclosed by LLM developers, including pretraining mixture proportions and finetuning data sources.



Paperid:5079
Authors:Ferdinand Genans, Antoine Godichon-Baggioni, Francois-Xavier Vialard, Olivier Wintenberger
Abstract:
Abstract:We investigate the semi-discrete Optimal Transport (OT) problem, where a continuous source measure $\mu$ is transported to a discrete target measure $\nu$, with particular attention to the OT map approximation. In this setting, Stochastic Gradient Descent (SGD) based solvers have demonstrated strong empirical performance in recent machine learning applications, yet their theoretical guarantee to approximate the OT map is an open question. In this work, we answer it positively by providing both computational and statistical convergence guarantees of SGD. Specifically, we show that SGD methods can estimate the OT map with a minimax convergence rate of $\mathcal{O}(1/\sqrt{n})$, where $n$ is the number of samples drawn from $\mu$. To establish this result, we study the averaged projected SGD algorithm, and identify a suitable projection set that contains a minimizer of the objective, even when the source measure is not compactly supported. Our analysis holds under mild assumptions on the source measure and applies to MTW cost functions,whic include $\|\cdot\|^p$ for $p \in (1, \infty)$. We finally provide numerical evidence for our theoretical results.



Paperid:5080
Authors:Longwei Wang, Chaowei Zhang, Ifrat Ikhtear Uddin, Prof. KC Santosh (PhD), Xiao Qin, Yang Zhou
Abstract:
Adversarial examples reveal critical vulnerabilities in deep neural networks by exploiting their sensitivity to imperceptible input perturbations. While adversarial training remains the predominant defense strategy, it often incurs significant computational cost and may compromise clean-data accuracy. In this work, we investigate an architectural approach to adversarial robustness by embedding group-equivariant convolutions—specifically, rotation- and scale-equivariant layers—into standard convolutional neural networks (CNNs). These layers encode symmetry priors that align model behavior with structured transformations in the input space, promoting smoother decision boundaries and greater resilience to adversarial attacks. We propose and evaluate two symmetry-aware architectures: a parallel design that processes standard and equivariant features independently before fusion, and a cascaded design that applies equivariant operations sequentially. Theoretically, we demonstrate that such models reduce hypothesis space complexity, regularize gradients, and yield tighter certified robustness bounds under the CLEVER (Cross Lipschitz Extreme Value for nEtwork Robustness) framework. Empirically, our models consistently improve adversarial robustness and generalization across CIFAR-10, CIFAR-100, and CIFAR-10C under both FGSM and PGD attacks, without requiring adversarial training. These findings underscore the potential of symmetry-enforcing architectures as efficient and principled alternatives to data augmentation-based defenses.



Paperid:5081
Authors:Jules Sintes, Ana Busic
Abstract:
Many controlled complex systems have an inherent network structure such as power grids, traffic lights systems or computer networks. Automatically controlling these systems is highly challenging due to their combinatorial complexity. Standard single-agent reinforcement learning (RL) approaches often struggle with the curse of dimensionality in such settings. In contrast, the multi-agent paradigm offers a promising solution by distributing decision-making, thereby addressing both algorithmic and combinatorial challenges. In this paper, we introduce COGNAC (COoperative Graph-based Networked Agent Challenges), the first collection of benchmark environments specifically designed for fully cooperative multi-agent tasks with graph structures. COGNAC bridges the gap between theoretical research in network control and practical multi-agent RL applications by offering a flexible, scalable platform with a suite of simple yet highly challenging problems rooted in networked environments. Our benchmarks also support the development and evaluation of decentralized and distributed learning algorithms, motivated by the growing interest in more sustainable and frugal AI systems. We evaluate standard independent learning algorithms on COGNAC environments and show that they outperform both heuristic-based baselines and centralized approaches across varying problem sizes.



Paperid:5082
Authors:Tobias Schmidt, Steffen Schneider, Matthias Bethge
Abstract:
Abstract:We propose Equivariance by Contrast (EbC) to learn equivariant embeddings from observation pairs $(x, g \cdot x)$, where $g$ is drawn from a finite group acting on the data. Our method jointly learns a latent space and a group representation in which group actions correspond to invertible linear maps—without relying on group-specific inductive biases. We validate our approach on the infinite dSprites dataset with structured transformations defined by the finite group $G:= (R_m \times \mathbb{Z}_n \times \mathbb{Z}_n)$, combining discrete rotations and periodic translations. The resulting embeddings exhibit high-fidelity equivariance, with group operations faithfully reproduced in latent space. On synthetic data, we further validate the approach on the non-Abelian orthogonal group $O(n)$ and the generalized linear group $GL(n)$. We also provide a theoretical proof for identifiability. While broad evaluation across diverse group types on real-world data remains future work, our results constitute the first successful demonstration of general-purpose equivariant learning from group action observations alone, including non-trivial non-abelian groups and a product group motivated by modeling affine equivariances in computer vision.



Paperid:5083
Authors:Xiaoyuan Liu, Tian Liang, Zhiwei He, Jiahao Xu, Wenxuan Wang, Pinjia He, Zhaopeng Tu, Haitao Mi, Dong Yu
Abstract:
Large Language Models (LLMs) show great promise in complex reasoning, with Reinforcement Learning with Verifiable Rewards (RLVR) being a key enhancement strategy. However, a prevalent issue is ``superficial self-reflection'', where models fail to robustly verify their own outputs. We introduce RISE (Reinforcing Reasoning with Self-Verification), a novel online RL framework designed to tackle this. RISE explicitly and simultaneously trains an LLM to improve both its problem-solving and self-verification abilities within a single, integrated RL process. The core mechanism involves leveraging verifiable rewards from an outcome verifier to provide on-the-fly feedback for both solution generation and self-verification tasks. In each iteration, the model generates solutions, then critiques its own on-policy generated solutions, with both trajectories contributing to the policy update. Extensive experiments on diverse mathematical reasoning benchmarks show that RISE consistently improves model's problem-solving accuracy while concurrently fostering strong self-verification skills. Our analyses highlight the advantages of online verification and the benefits of increased verification compute. Additionally, RISE models exhibit more frequent and accurate self-verification behaviors during reasoning. These advantages reinforce RISE as a flexible and effective path towards developing more robust and self-aware reasoners.



Paperid:5084
Authors:Aditya Sengar, Ali Hariri, Daniel Probst, PATRICK BARTH, Pierre Vandergheynst
Abstract:
Abstract:Generating diverse, all‐atom conformational ensembles of dynamic proteins such as G‐protein‐coupled receptors (GPCRs) is critical for understanding their function, yet most generative models simplify atomic detail or ignore conformational diversity altogether. We present latent diffusion for full protein generation (LD-FPG), a framework that constructs complete all‐atom protein structures, including every side‐chain heavy atom, directly from molecular dynamics (MD) trajectories. LD-FPG employs a Chebyshev graph neural network (ChebNet) to obtain low‐dimensional latent embeddings of protein conformations, which are processed using three pooling strategies: blind, sequential and residue‐based. A diffusion model trained on these latent representations generates new samples that a decoder, optionally regularized by dihedral‐angle losses, maps back to Cartesian coordinates. Using D2R-MD, a $2\mu\text{s}$ MD trajectory (12 000 frames) of the human dopamine D$2$ receptor in a membrane environment, the sequential and residue-based pooling strategy reproduces the reference ensemble with high structural fidelity (all‐atom lDDT \~ $0.7$; $C\alpha$-lDDT \~ $0.8$) and recovers backbone and side‐chain dihedral‐angle distributions with a Jensen–Shannon divergence $<0.03$ compared to the MD data. LD-FPG thereby offers a practical route to system‐specific, all‐atom ensemble generation for large proteins, providing a promising tool for structure‐based therapeutic design on complex, dynamic targets. The D2R-MD dataset and our implementation are freely available to facilitate further research.



Paperid:5085
Authors:Marien Renaud, Valentin De Bortoli, Arthur Leclaire, Nicolas Papadakis
Abstract:
We consider the problem of sampling distributions stemming from non-convex potentials with Unadjusted Langevin Algorithm (ULA). We prove the stability of the discrete-time ULA to drift approximations under the assumption that the potential is strongly convex at infinity. In many context, e.g. imaging inverse problems, potentials are non-convex and non-smooth. Proximal Stochastic Gradient Langevin Algorithm (PSGLA) is a popular algorithm to handle such potentials. It combines the forward-backward optimization algorithm with a ULA step. Our main stability result combined with properties of the Moreau envelope allows us to derive the first proof of convergence of the PSGLA for non-convex potentials. We empirically validate our methodology on synthetic data and in the context of imaging inverse problems. In particular, we observe that PSGLA exhibits faster convergence rates than Stochastic Gradient Langevin Algorithm for posterior sampling while preserving its restoration properties.



Paperid:5086
Authors:Longde Huang, Oleksandr Balabanov, Hampus Linander, Mats Granath, Daniel Persson, Jan Gerken
Abstract:
Equivariant network architectures are a well-established tool for predicting invariant or equivariant quantities. However, almost all learning problems considered in this context feature a global symmetry, i.e. each point of the underlying space is transformed with the same group element, as opposed to a localgaugesymmetry, where each point is transformed with a different group element, exponentially enlarging the size of the symmetry group. We use gauge equivariant networks to predict topological invariants (Chern numbers) of multiband topological insulators for the first time. The gauge symmetry of the network guarantees that the predicted quantity is a topological invariant. A major technical challenge is that the relevant gauge equivariant networks are plagued by instabilities in their training, severely limiting their usefulness. In particular, for larger gauge groups the instabilities make training impossible. We resolve this problem by introducing a novel gauge equivariant normalization layer which stabilizes the training. Furthermore, we prove a universal approximation theorem for our model. We train on samples with trivial Chern number only but show that our model generalizes to samples with non-trivial Chern number and provide various ablations of our setup.



Paperid:5087
Authors:Huanjin Yao, Qixiang Yin, Jingyi Zhang, Min Yang, Yibo Wang, Wenhao Wu, Fei Su, Li Shen, Minghui Qiu, Dacheng Tao, Jiaxing Huang
Abstract:
In this work, we aim to incentivize the reasoning ability of Multimodal Large Language Models (MLLMs) via reinforcement learning (RL) and develop an effective approach that mitigates the sparse reward and advantage vanishing issues during RL. To this end, we propose Share-GRPO, a novel RL approach that tackle these issues by exploring and sharing diverse reasoning trajectories over expanded question space. Specifically, Share-GRPO first expands the question space for a given question via data transformation techniques, and then encourages MLLM to effectively explore diverse reasoning trajectories over the expanded question space and shares the discovered reasoning trajectories across the expanded questions during RL. In addition, Share-GRPO also shares reward information during advantage computation, which estimates solution advantages hierarchically across and within question variants, allowing more accurate estimation of relative advantages and improving the stability of policy training. Extensive evaluations over 6 widely-used reasoning benchmarks showcase the superior performance of our method.



Paperid:5088
Authors:Jingxuan Wang, Zhongke Wu, Wang, Zhang Zeyao, Chunhao Zheng, Di Wang
Abstract:
Tubular-liked system shape analysis are quite difficult in geometry and topology, while it is widely used in plants and organs analysis in practice.However, traditional discrete representations such as voxels, point clouds often require substantial storage and may lead to the loss of fine-grained geometric and topological details. To address these challenges, we propose SE(3)-BBSCformerGCN, a novel framework for learning shape-aware representations from smooth tubular topological manifolds with built-in rotations and translations equivariance.Our approach leverages Ball B-Spline Curve (BBSC) to define smooth tubular manifolds.We provide a formal mathematical definition and analysis of the resulting manifolds, and incorporate an equivariant mapping that preserves geometric and topological stability. Compared to the point cloud and voxel based presentation, our manifold-based formulation significantly reduces data complexity while preserving geometric attributestogether with topological features. We validate our method on the branch classification task for Circle of Willis (CoW) on the TopCoW 2024 dataset and the clinical dataset.Our method consistently outperforms voxel and point cloud based baselines in terms of classification performance, generalization ability, convergence speed, and robustness to overfitting.



Paperid:5089
Authors:Zelai Xu, Ruize Zhang, Chao Yu, Huining Yuan, Xiangmin Yi, Shilong Ji, Chuqi Wang, Wenhao Tang, Feng Gao, Wenbo Ding, Xinlei Chen, Yu Wang
Abstract:
Robot sports, characterized by well-defined objectives, explicit rules, and dynamic interactions, present ideal scenarios for demonstrating embodied intelligence. In this paper, we present VolleyBots, a novel robot sports testbed where multiple drones cooperate and compete in the sport of volleyball under physical dynamics. VolleyBots integrates three features within a unified platform: competitive and cooperative gameplay, turn-based interaction structure, and agile 3D maneuvering. Competitive and cooperative gameplay challenges each drone to coordinate with its teammates while anticipating and countering opposing teams’ tactics.Turn-based interaction demands precise timing, accurate state prediction, and management of long-horizon temporal dependencies.Agile 3D maneuvering requires rapid accelerations, sharp turns, and precise 3D positioning despite the quadrotor’s underactuated dynamics.These intertwined features yield a complex problem combining motion control and strategic play, with no available expert demonstrations.We provide a comprehensive suite of tasks ranging from single-drone drills to multi-drone cooperative and competitive tasks, accompanied by baseline evaluations of representative multi-agent reinforcement learning (MARL) and game-theoretic algorithms. Simulation results show that on-policy reinforcement learning (RL) methods outperform off-policy methods in single-agent tasks, but both approaches struggle in complex tasks that combine motion control and strategic play.We additionally design a hierarchical policy which achieves 69.5% win rate against the strongest baseline in the 3 vs 3 task, underscoring its potential as an effective solution for tackling the complex interplay between low-level control and high-level strategy.The project page is at https://sites.google.com/view/volleybots.



Paperid:5090
Authors:Clément Bénard
Abstract:
Tree ensembles have demonstrated state-of-the-art predictive performance across a wide range of problems involving tabular data. Nevertheless, the black-box nature of tree ensembles is a strong limitation, especially for applications with critical decisions at stake. The Hoeffding or ANOVA functional decomposition is a powerful explainability method, as it breaks down black-box models into a unique sum of lower-dimensional functions, provided that input variables are independent. In standard learning settings, input variables are often dependent, and the Hoeffding decomposition is generalized through hierarchical orthogonality constraints. Such generalization leads to unique and sparse decompositions with well-defined main effects and interactions. However, the practical estimation of this decomposition from a data sample is still an open problem. Therefore, we introduce the TreeHFD algorithm to estimate the Hoeffding decomposition of a tree ensemble from a data sample. We show the convergence of TreeHFD, along with the main properties of orthogonality, sparsity, and causal variable selection. The high performance of TreeHFD is demonstrated through experiments on both simulated and real data. Besides, we empirically show that the widely used TreeSHAP method, based on Shapley values, is strongly connected to the Hoeffding decomposition.



Paperid:5091
Authors:Tal Fiskus, Uri Shaham
Abstract:
Deep reinforcement learning (DRL) agents excel in solving complex decision-making tasks across various domains.However, they often require a substantial number of training steps and a vast experience replay buffer, leading to significant computational and resource demands.To address these challenges, we introduce a novel theoretical result that leverages the Neyman-Rubin potential outcomes framework into DRL.Unlike most methods that focus on bounding the counterfactual loss, we establish a causal bound on the factual loss, which is analogous to the on-policy loss in DRL.This bound is computed by storing past value network outputs in the experience replay buffer, effectively utilizing data that is usually discarded.Extensive experiments across the Atari 2600 and MuJoCo domains on various agents, such as DQN and SAC, achieveup to 2,427%higher reward ratio, outperforming the same agents without our proposed term, and reducing the experience replay buffer size byup to 96%, significantly improvingsample efficiency at negligible cost.



Paperid:5092
Authors:Kyo Kuroki, Yasuyuki Okoshi, Thiem Van Chu, Masato Motomura, Kazushi Kawamura
Abstract:
This paper proposes a novel matrix quantization method, Binary Quadratic Quantization (BQQ). In contrast to conventional first-order quantization approaches—such as uniform quantization and binary coding quantization—that approximate real-valued matrices via linear combinations of binary bases, BQQ leverages the expressive power of binary quadratic expressions while maintaining an extremely compact data format.We validate our approach with two experiments: a matrix compression benchmark and post-training quantization (PTQ) on pretrained Vision Transformer-based models.Experimental results demonstrate that BQQ consistently achieves a superior trade-off between memory efficiency and reconstruction error than conventional methods for compressing diverse matrix data. It also delivers strong PTQ performance, even though we neither target state-of-the-art PTQ accuracy under tight memory constraints nor rely on PTQ-specific binary matrix optimization.For example, our proposed method outperforms the state-of-the-art PTQ method by up to 2.0\% and 59.1\% on the ImageNet dataset under the calibration-based and data-free scenarios, respectively, with quantization equivalent to 2 bits.These findings highlight the surprising effectiveness of binary quadratic expressions for efficient matrix approximation and neural network compression.



Paperid:5093
Authors:Yujie Zhu, Charles Hepburn, Matthew Thorpe, Giovanni Montana
Abstract:
In reinforcement learning with sparse rewards, demonstrations can accelerate learning, but determining when to imitate them remains challenging. We propose Smooth Policy Regularisation from Demonstrations (SPReD), a framework that addresses the fundamental question: when should an agent imitate a demonstration versus follow its own policy? SPReD uses ensemble methods to explicitly model Q-value distributions for both demonstration and policy actions, quantifying uncertainty for comparisons. We develop two complementary uncertainty-aware methods: a probabilistic approach estimating the likelihood of demonstration superiority, and an advantage-based approach scaling imitation by statistical significance. Unlike prevailing methods (e.g. Q-filter) that make binary imitation decisions, SPReD applies continuous, uncertainty-proportional regularisation weights, reducing gradient variance during training. Despite its computational simplicity, SPReD achieves remarkable gains in experiments across eight robotics tasks, outperforming existing approaches by up to a factor of 14 in complex tasks while maintaining robustness to demonstration quality and quantity.



Paperid:5094
Authors:Wanyun Cui, Mingwei Xu
Abstract:
Abstract:Recent advances in Large Language Models (LLMs) have highlighted the critical importance of extending context length, yet the quadratic complexity of attention mechanisms poses significant challenges for efficient long-context modeling. KV cache compression has emerged as a key approach to address this challenge. Through extensive empirical analysis, we reveal a fundamental yet previously overlooked asymmetry in KV caches: while adjacent keys receive similar attention weights ({\it local homogeneity}), adjacent values demonstrate distinct {\it heterogeneous} distributions. This key-value asymmetry reveals a critical limitation in existing compression methods that treat keys and values uniformly. To address the limitation, we propose a training-free compression framework (AsymKV) that combines homogeneity-based key merging with a mathematically proven lossless value compression. Extensive experiments demonstrate that AsymKV consistently outperforms existing long-context methods across various tasks and base models. For example, on LLaMA3.1-8B, AsymKV achieves an average score of 43.95 on LongBench, surpassing SOTA methods like H$_2$O (38.89) by a large margin.



Paperid:5095
Authors:Xin Liu, Haoran Li, Dongbin Zhao
Abstract:
Humans can efficiently extract knowledge and learn skills from the videos within only a few trials and errors. However, it poses a big challenge to replicate this learning process for autonomous agents, due to the complexity of visual input, the absence of action or reward signals, and the limitations of interaction steps. In this paper, we propose a novel, unsupervised, and sample-efficient framework to achieve imitation learning from videos (ILV), named Behavior Cloning from Videos via Latent Representations (BCV-LR). BCV-LR extracts action-related latent features from high-dimensional video inputs through self-supervised tasks, and then leverages a dynamics-based unsupervised objective to predict latent actions between consecutive frames. The pre-trained latent actions are fine-tuned and efficiently aligned to the real action space online (with collected interactions) to behavior cloning a policy. The cloned policy in turn enriches the agent experience for latent action finetuning, resulting in an iterative policy improvement that is highly sample-efficient. We conduct extensive experiments on a set of challenging visual tasks, including both discrete control and continuous control. BCV-LR enables effective (even expert-level on some tasks) policy performance with only a few interactions, surpassing state-of-the-art ILV baselines and reinforcement learning methods (provided with environmental rewards) in terms of sample efficiency across 20/24 tasks. To the best of our knowledge, this work for the first time demonstrates that videos can support extremely sample-efficient visual policy learning, without the need to access any other supervision. Codes will be at*.



Paperid:5096
Authors:Haozhe Wang, Chao Qu, Zuming Huang, Wei Chu, Fangzhen Lin, Wenhu Chen
Abstract:
Recently, slow-thinking systems like GPT-o1 and DeepSeek-R1 have demonstrated great potential in solving challenging problems through explicit reflection. They significantly outperform the best fast-thinking models, such as GPT-4o, on various math and science benchmarks. However, their multimodal reasoning capabilities remain on par with fast-thinking models. For instance, GPT-o1's performance on benchmarks like MathVista, MathVerse, and MathVision is similar to fast-thinking models. In this paper, we aim to enhance the slow-thinking capabilities of vision-language models using reinforcement learning (without relying on distillation) to advance the state of the art. First, we adapt the GRPO algorithm with a novel technique called Selective Sample Replay (SSR) to address the vanishing advantages problem. While this approach yields strong performance, the resulting RL-trained models exhibit limited self-reflection or self-verification. To further encourage slow-thinking, we introduce Forced Rethinking, which appends a rethinking trigger token to the end of rollouts in RL training, explicitly enforcing a self-reflection reasoning step. By combining these two techniques, our model, VL-Rethinker, advances state-of-the-art scores on MathVista, MathVerse to achieve 80.4%, 63.5% respectively. VL-Rethinker also achieves open-source SoTA on multi-disciplinary benchmarks such as MathVision, MMMU-Pro, EMMA, and MEGA-Bench, narrowing the gap with OpenAI-o1. We conduct comprehensive ablations and analysis to provide insights into the effectiveness of our approach.



Paperid:5097
Authors:Gangda Deng, Hongkuan Zhou, Rajgopal Kannan, Viktor Prasanna
Abstract:
Heterophilous graphs, where dissimilar nodes tend to connect, pose a challenge for graph neural networks (GNNs). Increasing the GNN depth can expand the scope (i.e., receptive field), potentially finding homophily from the higher-order neighborhoods. However, GNNs suffer from performance degradation as depth increases. Despite having better expressivity, state-of-the-art deeper GNNs achieve only marginal improvements compared to their shallow variants. Through theoretical and empirical analysis, we systematically demonstrate a shift in GNN generalization preferences across nodes with different homophily levels as depth increases. This creates a disparity in generalization patterns between GNN models with varying depth. Based on these findings, we propose to improve deeper GNN generalization while maintaining high expressivity by Mixture of scope experts at test (Moscat). Experimental results show that Moscat works flexibly with various GNN architectures across a wide range of datasets while significantly improving accuracy.



Paperid:5098
Authors:Lucine L Oganesian, Saba Hashemi, Maryam Shanechi
Abstract:
Intracranial recordings have opened a unique opportunity to simultaneously measure activity across multiregional networks in the human brain. Recent works have focused on developing transformer-based neurofoundation models of such recordings that can generalize across subjects and datasets. However, these recordings exhibit highly complex spatiotemporal interactions across diverse spatial scales, from the single-channel scale to the scale of brain regions. As such, there remain critical open questions regarding how best to encode spatial information and how to design self-supervision tasks that enable the learning of brain network patterns and enhance downstream decoding performance using such high-dimensional, multiregional recordings. To allow for exploring these questions, we propose a new spatiotemporal transformer model of multiregional neural activity and a corresponding self-supervised masked latent reconstruction task, designed to enable flexibility in the spatial scale used for token encoding and masking. Applying this model on publicly available multiregional intracranial electrophysiology (iEEG) data, we demonstrate that adjusting the spatial scale for both token encoding and masked reconstruction significantly impacts downstream decoding. Further, we find that spatial encoding at larger scales than channel-level encoding, which is commonly used in existing iEEG transformer models, improves downstream decoding performance. Finally, we demonstrate that our method allows for region-level token encoding while also maintaining accurate channel-level neural reconstruction. Taken together, our modeling framework enables exploration of the spatial scales used for token encoding and masking, and reveals their importance towards designing and pretraining neurofoundation models of multiregional human brain activity.



Paperid:5099
Authors:Zhizhen Zhang, Lei Zhu, Zhen Fang, Zi Huang, Yadan Luo
Abstract:
Pre-training vision-language representations on human action videos has emergedas a promising approach to reduce reliance on large-scale expert demonstrationsfor training embodied agents. However, prior methods often employ time con-trastive learning based on goal-reaching heuristics, progressively aligning languageinstructions from the initial to the final frame. This overemphasis on future framescan result in erroneous vision-language associations, as actions may terminateearly or include irrelevant moments in the end. To address this issue, we proposeAction Temporal Coherence Learning (AcTOL) to learn ordered and continuousvision-language representations without rigid goal-based constraint. AcTOL treatsa video as a continuous trajectory where it (1) contrasts semantic differences be-tween frames to reflect their natural ordering, and (2) imposes a local Brownianbridge constraint to ensure smooth transitions across intermediate frames. Exten-sive imitation learning experiments on both simulated and real robots show that thepretrained features significantly enhance downstream manipulation tasks with highrobustness to different linguistic styles of instructions, offering a viable pathwaytoward generalized embodied agents. We provide source code and demo videos inthe supplemental material for reference.



Paperid:5100
Authors:Junyu Zhang, Lizhong Ding, MinghongZhang, Ye Yuan, Xingcan Li, Pengqi Li, Tihang Xi, Guoren Wang, Changsheng Li
Abstract:
Multi-table data integrate various entities and attributes, with potential interconnections between them. However, existing tabular learning methods often struggle to describe and leverage the underlying complementarity across distinct tables. To address this limitation, we propose a novel paradigm for multi-table learning that, for the first time, systematically quantifies and integrates complementary information across multiple tables. Specifically, we introduce a metric called complementarity strength (CS), which captures inter-table complementarity by incorporating relevance, similarity, and informativeness. For the first time, we systematically formulate the paradigm towards multi-table learning by establishing formal definitions of tasks and loss functions. Correspondingly, we present the network towards multi-table learning that combines Adaptive Table encoder and Cross table Attention mechanism (ATCA-Net), achieving the simultaneous integration of complementary information from distinct tables. Extensive experiments show that ATCA-Net effectively leverages complementary information, while demonstrating that the CS metric accurately quantifies complementarity richness across multiple tables. To the best of our knowledge, this is the first work to establish theoretical and practical foundations for multi-table learning.



Paperid:5101
Authors:Abdurakhmon Sadiev, Ilyas Fatkhullin, Peter Richtarik
Abstract:
Abstract:Heavy-tailed noise is pervasive in modern machine learning applications, arising from data heterogeneity, outliers, and stochastic non-stationary environments. While second-order methods can significantly accelerate convergence in light-tailed or bounded-noise settings, such algorithms are often brittle and lack guarantees under heavy-tailed noise—precisely the regimes where robustness is most critical. In this work, we take a first step toward a theoretical understanding of second-order optimization under heavy-tailed noise. We consider a setting where stochastic gradients and Hessians have only bounded $p$-th moments for $p\in (1,2]$, and establish tight lower bounds on the sample complexity of any second-order method. We then develop a variant of normalized stochastic gradient descent that leverages second-order information and provably matches these lower bounds. To address the instability caused by large deviations, we introduce a novel algorithm based on gradient and Hessian clipping, and prove high-probability upper bounds that nearly match the fundamental limits. Our results provide the first comprehensive sample complexity characterization for second-order optimization under heavy-tailed noise. This positions Hessian clipping as a robust and theoretically sound strategy for second-order algorithm design in heavy-tailed regimes.



Paperid:5102
Authors:Yulu Qin, Dheeraj Varghese, Adam Dahlgren Lindström, Lucia Donatelli, Kanishka Misra, Najoung Kim
Abstract:
Does vision-and-language (VL) training change the linguistic representations of language models in meaningful ways? In terms of downstream task performance on text-only tasks, most results in the literature have shown marginal differences. In this work, we start from the hypothesis that the domain in which VL training could have a significant effect is lexical-conceptual knowledge, in particular its taxonomic organization. Through comparing minimal pairs of text-only LMs and their VL-trained counterparts, we first show that the VL models often outperform their text-only counterparts on a text-only question-answering task that requires taxonomic understanding of concepts mentioned in the questions. Using an array of targeted behavioral and representational analyses, we show that the LMs and VLMs do not differ significantly in terms of their taxonomic knowledge itself, but they differ in how they represent questions that contain concepts in a taxonomic relation vs. a non-taxonomic relation. This implies that the taxonomic knowledge itself does not change substantially through additional VL training, but VL training does improve the \textit{deployment} of this knowledge in the context of a specific task, even when the presentation of the task is purely linguistic.



Paperid:5103
Authors:Siqi Kou, Qingyuan Tian, Hanwen Xu, Zihao Zeng, Zhijie Deng
Abstract:
Large language models (LLMs) have demonstrated remarkable reasoning capabilities in math and coding, often bolstered by post-training on the chain-of-thoughts (CoTs) generated by stronger models. However, existing strategies for curating such training data predominantly rely on heuristics, limiting generalizability and failing to capture subtleties underlying in data. To address these limitations, we leverage influence functions to systematically attribute LLMs' reasoning ability on math and coding to individual training examples, sequences, and tokens, enabling deeper insights into effective data characteristics.Our Influence-based Reasoning Attribution (Infra) uncovers nontrivial cross-domain effects across math and coding tasks: high-difficulty math examples improve both math and code reasoning, while low-difficulty code tasks most effectively benefit code reasoning.Based on these findings, we introduce a simple yet effective dataset reweighting strategy by flipping task difficulty, which doubles AIME24 accuracy from 10\% to 20\% and boosts LiveCodeBench accuracy from 33.8\% to 35.3\% for Qwen2.5-7B-Instruct.Moreover, our fine-grained attribution reveals that the sequence-level exploratory behaviors enhance reasoning performance in both math and code, and the token-level influence patterns are distinct for math and code reasoning: the former prefers natural language logic connectors and the latter emphasizes structural syntax.



Paperid:5104
Authors:Soumita Hait, Ping Li, Haipeng Luo, Mengxiao Zhang
Abstract:
Abstract:In the classic expert problem, $\Phi$-regret measures the gap between the learner's total loss and that achieved by applying the best action transformation $\phi \in \Phi$. A recent work by Lu et al., [2025] introduced an adaptive algorithm whose regret against a comparator $\phi$ depends on a certain sparsity-based complexity measure of $\phi$, recovering and interpolating optimal bounds for standard regret notions such as external, internal, and swap regret. In this work, we propose a general idea to achieve an even better comparator-adaptive $\Phi$-regret bound via much simpler algorithms compared to Lu et al., [2025]. Specifically, we discover a prior distribution over all possible binary transformations and show that it suffices to achieve prior-dependent regret against these transformations. Then, we propose two concrete and efficient algorithms to achieve so, where the first one combines multiple copies of the kernelized Hedge algorithm of Farina et al., [2022], and the second one combines multiple copies of a variant of the BM-reduction [Blum and Mansour, 2007]. To further showcase the power of our methods and the advantages over Lu et al., [2025] besides the simplicity and better regret bounds, we also show that our second approach can be extended to the game setting to achieve accelerated and adaptive convergence rate to $\Phi$-equilibria for a class of general-sum games. When specified to the special case of correlated equilibria, our bound improves over the existing ones from Anagnostides et al., [2022a,b].



Paperid:5105
Authors:Harry Amad, Zhaozhi Qian, Dennis Frauen, Julianna Piskorz, Stefan Feuerriegel, Mihaela van der Schaar
Abstract:
Causal inference is essential for developing and evaluating medical interventions, yet real-world medical datasets are often difficult to access due to regulatory barriers. This makes synthetic data a potentially valuable asset to enable critical analyses, along with the development and testing of new inference methods. Generative models can produce synthetic data that closely approximate real data distributions, yet existing methods do not consider the unique challenges that downstream causal inference tasks—specifically those focused on treatments—pose. We establish a set of desiderata that synthetic data containing treatments should satisfy to maximise downstream utility: preservation of (i) the covariate distribution, (ii) the treatment assignment mechanism, and (iii) the outcome generation mechanism. Based on these desiderata, we propose a set of evaluation metrics to assess such synthetic data. Finally, we present STEAM: a novel method for generating Synthetic data for Treatment Effect Analysis in Medicine that mimics the data-generating process of data containing treatments and optimises for our desiderata. We empirically demonstrate that STEAM achieves state-of-the-art performance across our metrics as compared to existing generative models, particularly as the complexity of the true data-generating process increases.



Paperid:5106
Authors:Leo Song, Litao Li, Steven Ding, Benjamin C. M. Fung, Philippe Charland
Abstract:
Cybersecurity and software research has crossed the path with modern deep learning research for a few years. The power of large language models (LLMs) in particular has intrigued us to apply them to understanding binary code. In this paper, we investigate some of the many ways LLMs can be applied to binary code similarity detection, as it is a significantly more difficult task compared to source code similarity detection due to the sparsity of information and less meaningful syntax. It also has great practical implications, such as vulnerability and malware detection. We find that pretrained LLMs are mostly capable of detecting similar binary code, even with a zero-shot setting. Our main contributions and findings are to provide several supervised fine-tuning methods that, when combined, significantly surpass zero-shot LLMs and state-of-the-art binary code similarity detection methods. Specifically, we up-train the model through: data augmentation, translation-style causal learning, LLM2Vec, and cumulative GTE loss. With a complete ablation study, we show that our training method can transform a generic language model into a powerful binary similarity expert, and is also robust and general enough for cross-optimization, cross-architecture, and cross-obfuscation detection.



Paperid:5107
Authors:Zehui Li, Vallijah Subasri, Yifei Shen, Dongsheng Li, Wentao Gu, Guy-Bart Stan, Yiren Zhao, Caihua Shan
Abstract:
The interpretation of genomic sequences is crucial for understanding biological processes. To handle the growing volume of DNA sequence data, Genomic Foundation Models (GFMs) have been developed by adapting architectures and training paradigms from Large Language Models (LLMs). Despite their remarkable performance in DNA sequence classification tasks, there remains a lack of systematic understanding regarding the training and task-adaptation processes of GFMs. Moreover, existing GFMs cannot achieve state-of-the-art performance on both short and long-context tasks and lacks multimodal abilities. By revisiting pre-training architectures and post-training techniques, we proposeOmni-DNA, a family of models spanning 20M to 1.1B parameters that supports sequence understanding, long-context genomic reasoning, and natural-language annotation.Omni-DNAestablishes new state-of-the-art results on 18 of 26 evaluations drawn from Nucleotide Transformer and Genomic Benchmarks. When jointly fine-tuning on biologically related tasks,Omni-DNAconsistently outperform existing models and demonstrate multi-tasking abilities. To enable processing of arbitrary sequence lengths, we introduceSEQPACK—an adaptive compression operator that packs historical tokens into a learned synopsis using a position-aware learnable sampling mechanism, enabling transformer-based models to process ultra-long sequences with minimal memory and computational requirements. Our approach demonstrates superior performance on enhancer-target interaction tasks, capturing distant regulatory interactions at the 450kbp range more effectively than existing models. Finally, we present a new dataset termedseq2func, enabling Omni-DNA to generate accurate and functionally meaningful interpretations of DNA sequences, unlocking new possibilities for genomic analysis and discovery.



Paperid:5108
Authors:Alex Damian, Jason Lee, Joan Bruna
Abstract:
Abstract:In this work we consider generic Gaussian Multi-index models, in which the labels only depend on the (Gaussian) $d$-dimensional inputs through their projection onto a low-dimensional $r = O_d(1)$ subspace, and we study efficient agnostic estimation procedures for this hidden subspace. We introduce the *generative leap* exponent, a natural extension of the generative exponent from Damian et al. 2024 to the multi-index setting. We show that a sample complexity of $n=\Theta(d^{1 \vee k^\star/2})$ is necessary in the class of algorithms captured by the Low-Degree-Polynomial framework; and also sufficient, by giving a sequential estimation procedure based on a spectral U-statistic over appropriate Hermite tensors.



Paperid:5109
Authors:Minglu Zhao, Dehong Xu, Deqian Kong, Wenhao Zhang, Ying Nian Wu
Abstract:
Abstract:The hippocampus orchestrates spatial navigation through collective place cell encodings that form cognitive maps. We reconceptualize the population of place cells as position embeddings approximating multi-scale symmetric random walk transition kernels, where the inner product $\langle h(x, t), h(y, t) \rangle = q(y|x, t)$ represents normalized transition probabilities, where $h(x, t)$ is the embedding at location $ x $, and $q(y|x, t)$ is the symmetric transition probability over time $t$. The time parameter $\sqrt{t}$ defines a spatial scale hierarchy, mirroring the hippocampal dorsoventral axis. $q(y|x, t)$ defines spatial adjacency between $x$ and $y$ at scale or resolution $\sqrt{t}$, and the pairwise adjacency relationships $(q(y|x, t), \forall x, y)$ are reduced into individual embeddings $(h(x, t), \forall x)$ that collectively form a map of the environment. Our framework employs gradient ascent on $q(y|x, t) = \langle h(x, t), h(y, t)\rangle$ with adaptive scale selection, choosing the time scale with maximal gradient at each step for trap-free, smooth trajectories. Efficient matrix squaring $P_{2t} = P_t^2$ builds global representations from local transitions $P_1$ without memorizing past trajectories, enabling hippocampal preplay-like path planning. This produces robust navigation through complex environments, aligning with hippocampal navigation. Experimental results show that our model captures place cell properties—field size distribution, adaptability, and remapping—while achieving computational efficiency. By modeling collective transition probabilities rather than individual place fields, we offer a biologically plausible, scalable framework for spatial navigation.



Paperid:5110
Authors:Yijie Zheng, Weijie Wu, Qingyun Li, Xuehui Wang, Xu Zhou, Aiai Ren, Jun Shen, Long Zhao, Guoqing Li, Xue Yang
Abstract:
Language-Guided object recognition in remote sensing imagery is crucial for large-scale mapping and automated data annotation. However, existing open-vocabulary and visual grounding methods rely on explicit category cues, limiting their ability to handle complex or implicit queries that require advanced reasoning.To address this issue, we introduce a new suite of tasks, including Instruction-Oriented Object Counting, Detection, and Segmentation (InstructCDS), covering open-vocabulary, open-ended, and open-subclass scenarios. We further present EarthInstruct, the first InstructCDS benchmark for earth observation. It is constructed from two diverse remote sensing datasets with varying spatial resolutions and annotation rules across 20 categories, necessitating models to interpret dataset-specific instructions.Given the scarcity of semantically rich labeled data in remote sensing, we propose InstructSAM, a training-free framework for instruction-driven object recognition. InstructSAM leverages large vision-language models to interpret user instructions and estimate object counts, employs SAM2 for mask proposal, and formulates mask-label assignment as a binary integer programming problem. By integrating semantic similarity with counting constraints, InstructSAM efficiently assigns categories to predicted masks without relying on confidence thresholds. Experiments demonstrate that InstructSAM matches or surpasses specialized baselines across multiple tasks while maintaining near-constant inference time regardless of object count, reducing output tokens by 89\% and overall runtime by over 32\% compared to direct generation approaches. We believe the contributions of the proposed tasks, benchmark, and effective approach will advance future research in developing versatile object recognition systems.



Paperid:5111
Authors:Yuhao Yang, Zhi Ji, Zhaopeng Li, Yi Li, Zhonglin Mo, Yue Ding, Kai Chen, Zijian Zhang, Jie Li, shuanglong li, LIU LIN
Abstract:
Generative models have recently gained attention in recommendation systems by directly predicting item identifiers from user interaction sequences. However, existing methods suffer from significant information loss due to the separation of stages such as quantization and sequence modeling, hindering their ability to achieve the modeling precision and accuracy of sequential dense retrieval techniques. Integrating generative and dense retrieval methods remains a critical challenge. To address this, we introduce the Cascaded Organized Bi-Represented generAtive retrieval (COBRA) framework, which innovatively integrates sparse semantic IDs and dense vectors through a cascading process. Our method alternates between generating these representations by first generating sparse IDs, which serve as conditions to aid in the generation of dense vectors. End-to-end training enables dynamic refinement of dense representations, capturing both semantic insights and collaborative signals from user-item interactions. During inference, COBRA employs a coarse-to-fine strategy, starting with sparse ID generation and refining them into dense vectors via the generative model. We further propose BeamFusion, an innovative approach combining beam search with nearest neighbor scores to enhance inference flexibility and recommendation diversity. Extensive experiments on public datasets and offline tests validate our method's robustness. Online A/B tests on a real-world advertising platform with over 200 million daily users demonstrate substantial improvements in key metrics, highlighting COBRA's practical advantages.



Paperid:5112
Authors:Viet-Hung Tran, Ngoc Phu Doan, Zichi Zhang, Tuan Pham, Phi Hung Nguyen, Xuan Hoang Nguyen, Hans Vandierendonck, Ira Assent, Thai Son Mai
Abstract:
Deep learning models for time series classification (TSC) have achieved impressive performance, but explaining their decisions remains a significant challenge. Existing post-hoc explanation methods typically operate solely in the time domain and from a single-view perspective, limiting both faithfulness and robustness. In this work, we propose MIX (Multi-view Time-Frequency Interactive EXplanation Framework), a novel framework that helps to explain deep learning models in a multi-view setting by leveraging multi-resolution, time-frequency views constructed using the Haar Discrete Wavelet Transform (DWT). MIX introduces an interactive cross-view refinement scheme, where explanation's information from one view is propagated across views to enhance overall interpretability. To align with user-preferred perspectives, we propose a greedy selection strategy that traverses the multi-view space to identify the most informative features. Additionally, we present OSIGV, a user-aligned segment-level attribution mechanism based on overlapping windows for each view, and introduce keystone-first IG, a method that refines explanations in each view using additional information from another view. Extensive experiments across multiple TSC benchmarks and model architectures demonstrate that MIX significantly outperforms state-of-the-art (SOTA) methods in terms of explanation faithfulness and robustness.



Paperid:5113
Authors:Yige Li, Hanxun Huang, Yunhan Zhao, Xingjun Ma, Jun Sun
Abstract:
Generative large language models (LLMs) have achieved state-of-the-art results on a wide range of tasks, yet they remain susceptible to backdoor attacks: carefully crafted triggers in the input can manipulate the model to produce adversary-specified outputs. While prior research has predominantly focused on backdoor risks in vision and classification settings, the vulnerability of LLMs in open-ended text generation remains underexplored. To fill this gap, we introduce \textit{BackdoorLLM}\footnote{Our BackdoorLLM benchmark was awarded First Prize in the \href{https://www.mlsafety.org/safebench/winners}{SafetyBench competition} organized by the \href{https://safe.ai/}{Center for AI Safety}.}, the first comprehensive benchmark for systematically evaluating backdoor threats in text-generation LLMs. BackdoorLLM provides: (i) a unified repository of benchmarks with a standardized training and evaluation pipeline; (ii) a diverse suite of attack modalities, including data poisoning, weight poisoning, hidden-state manipulation, and chain-of-thought hijacking; (iii) over 200 experiments spanning 8 distinct attack strategies, 7 real-world scenarios, and 6 model architectures; (iv) key insights into the factors that govern backdoor effectiveness and failure modes in LLMs; and (v) a defense toolkit encompassing 7 representative mitigation techniques. Our code and datasets are available at \url{https://github.com/bboylyg/BackdoorLLM}. We will continuously incorporate emerging attack and defense methodologies to support the research in advancing the safety and reliability of LLMs.



Paperid:5114
Authors:Edward Fish, Richard Bowden
Abstract:
Recent progress in Sign Language Translation has focussed primarily on improving the representational capacity of large language models to incorporate sign-language features. This work explores an alternative direction: enhancing the geometric properties of skeletal representations themselves. We propose Geo-Sign, a method that leverages the properties of hyperbolic geometry to model the hierarchical structure inherent in sign language kinematics. By projecting skeletal features derived from Spatio-Temporal Graph Convolutional Networks (ST-GCNs) into the Poincaré ball model, we aim to create more discriminative embeddings, particularly for fine-grained motions like finger articulations. We introduce a hyperbolic projection layer, a weighted Fréchet mean aggregation scheme, and a geometric contrastive loss operating directly in hyperbolic space. These components are integrated into an end-to-end translation framework as a regularisation function, to enhance the representations within the language model. This work demonstrates the potential of hyperbolic geometry to improve skeletal representations for Sign Language Translation, improving on SOTA RGB methods while preserving privacy and improving computational efficiency.



Paperid:5115
Authors:Abdellah Rahmani, Pascal Frossard
Abstract:
Understanding causal relationships in multivariate time series is crucial in many scenarios, such as financial and neurological data. Many such time series exhibit multiple regimes, i.e., consecutive segments with unknown boundaries. Each regime has its own causal structure. Detecting these causal dependencies and regime shifts is critical for revealing regime indices and the underlying processes. However, causal structure learning in this setting is challenging due to (1) non-stationarity, i.e. each regime can have its own causal graph and mixing function, and (2) complex noise distributions, which may be non-Gaussian or heteroscedastic, with variance depending on one or more influencing variables. Existing causal discovery approaches often assume stationarity or Gaussian noise with constant variance. To address these challenges, We introduce FANTOM, a unified framework that tackles non-stationarity along with non-Gaussian and heteroscedastic noises. FANTOM simultaneously learns each regime’s Directed Acyclic Graph (DAG) and infers the number and indices of these regimes, using a Bayesian Expectation Maximization algorithm that maximizes the evidence lower bound of the data log likelihood. On the theoretical side, we prove, under common assumptions, that temporal heteroscedastic causal models are identifiable in both stationary and non-stationary settings. In addition, extensive experiments on synthetic and real data show that FANTOM outperforms existing methods.



Paperid:5116
Authors:Fan Yang, Yousong Zhu, Xin Li, Yufei Zhan, Hongyin Zhao, Shurong Zheng, Yaowei Wang, Ming Tang, Jinqiao Wang
Abstract:
Recent Large Vision Language Models (LVLMs) demonstrate promising capabilities in unifying visual understanding and generative modeling, enabling both accurate content understanding and flexible editing. However, current approaches treat \textbf{\textit{"what to see"}} and \textbf{\textit{"how to edit"}} separately: they either perform isolated object segmentation or utilize segmentation masks merely as conditional prompts for local edit generation tasks, often relying on multiple disjointed models. To bridge these gaps, we introduce FOCUS, a unified LVLM that integrates segmentation-aware perception and controllable object-centric generation within an end-to-end framework. FOCUS employs a dual-branch visual encoder to simultaneously capture global semantic context and fine-grained spatial details. In addition, we leverage a MoVQGAN-based visual tokenizer to produce discrete visual tokens that enhance generation quality. To enable accurate and controllable image editing, we propose a progressive multi-stage training pipeline, where segmentation masks are jointly optimized and used as spatial condition prompts to guide the diffusion decoder. This strategy aligns visual encoding, segmentation, and generation modules, effectively bridging segmentation-aware perception with fine-grained visual synthesis.Extensive experiments across three core tasks, including multimodal understanding, referring segmentation accuracy, and controllable image generation, demonstrate that FOCUS achieves strong performance by jointly optimizing visual perception and generative capabilities.



Paperid:5117
Authors:Andrea Bonfanti, Ismael Medina, Roman List, Björn Staeves, Roberto Santana, Marco Ellero
Abstract:
Recent advances in Scientific Machine Learning have shown that second-order methods can enhance the training of Physics-Informed Neural Networks (PINNs), making them a suitable alternative to traditional numerical methods for Partial Differential Equations (PDEs). However, second-order methods induce large memory requirements, making them scale poorly with the model size. In this paper, we define a local Mixture of Experts (MoE) combining the parameter-efficiency of ensemble models and sparse coding to enable the use of second-order training. Our model -- PINN Balls -- also features a fully learnable domain decomposition structure, achieved through the use of Adversarial Adaptive Sampling (AAS), which adapts the DD to the PDE and its domain. PINN Balls achieves better accuracy than the state-of-the-art in scientific machine learning, while maintaining invaluable scalability properties and drawing from a sound theoretical background.



Paperid:5118
Authors:Ping Guo, Yubing Ren, BINBINLIU, Fengze Liu, Haobin Lin, Yifan Zhang, Bingni Zhang, Taifeng Wang, Yin Zheng
Abstract:
Large language models (LLMs) have become integral to a wide range of applications worldwide, driving an unprecedented global demand for effective multilingual capabilities. Central to achieving robust multilingual performance is the strategic allocation of language proportions within training corpora. However, determining optimal language ratios is highly challenging due to intricate cross-lingual interactions and sensitivity to dataset scale. This paper introduces CLIMB (Cross-Lingual Interaction-aware Multilingual Balancing), a novel framework designed to systematically optimize multilingual data allocation. At its core, CLIMB introduces a cross-lingual interaction-aware language ratio, explicitly quantifying each language’s effective allocation by capturing inter-language dependencies. Leveraging this ratio, CLIMB proposes a principled two-step optimization procedure—first equalizing marginal benefits across languages, then maximizing the magnitude of the resulting language allocation vectors—significantly simplifying the inherently complex multilingual optimization problem. Extensive experiments confirm that CLIMB can accurately measure cross-lingual interactions across various multilingual settings. LLMs trained with CLIMB-derived proportions consistently achieve state-of-the-art multilingual performance, even achieve competitive performance with open-sourced LLMs trained with more tokens.



Paperid:5119
Authors:Alexander Long, Chamin Hewa Koneputugodage, Sameera Ramasinghe, Thalaiyasingam Ajanthan, Yan Zuo, Gil Avraham, Violetta Shevchenko, Hadi Mohaghegh Dolatabadi
Abstract:
Abstract:We consider a decentralized training setup in which the participants collaboratively train and serve a large neural network, and where each participant only processes a subset of the model. In this setup, we explore the possibility of \textit{unmaterializable} weights, where a full weight set is \textit{never} available to any one participant.We introduce Unextractable Protocol Models (UPMs): a training and inference framework that leverages the {\em sharded model setup} to ensure model shards (\ie, subsets) held by participants are incompatible at different time steps. UPMs periodically inject time-varying, random, invertible transforms at participant boundaries; preserving the overall network function yet rendering cross-time assemblies incoherent. On Qwen-2.5-0.5B and Llama-3.2-1B, 10 000 transforms leave FP 32 perplexity unchanged ($\Delta$PPL$< 0.01$; Jensen–Shannon drift $<4 \times 10^{-5}$). Applying a transform every 30s adds 3\% latency, 0.1\% bandwidth, and 10\% GPU-memory overhead at inference, while training overhead falls to 1.6\% time and < 1\% memory. We consider several attacks, showing that the requirements of direct attacks are impractical and easy to defend against, and that gradient-based fine-tuning of stitched partitions consumes $\geq 60\%$ of the tokens required to train from scratch. By enabling models to be collaboratively trained yet not extracted, UPMs make it practical to embed programmatic incentive mechanisms in community-driven decentralized training.



Paperid:5120
Authors:Ian Tanoh, Michael Deistler, Jakob H Macke, Scott Linderman
Abstract:
Multi-compartment Hodgkin-Huxley models are biophysical models of how electrical signals propagate throughout a neuron, and they form the basis of our knowledge of neural computation at the cellular level. However, these models have many free parameters that must be estimated for each cell, and existing fitting methods rely on intracellular voltage measurements that are highly challenging to obtain in-vivo. Recent advances in neural recording technology with high-density probes and arrays enable dense sampling of extracellular voltage from many sites surrounding a neuron, allowing indirect measurement of many compartments of a cell simultaneously. Here, we propose a method for inferring the underlying membrane voltage, biophysical parameters, and the neuron's position relative to the probe, using extracellular measurements alone. We use an Extended Kalman Filter to infer membrane voltage and channel states using efficient, differentiable simulators. Then, we learn the model parameters by maximizing the marginal likelihood using gradient-based methods. We demonstrate the performance of this approach using simulated data and real neuron morphologies.



Paperid:5121
Authors:Lingwei Dang, Ruizhi Shao, Hongwen Zhang, Wei MIN, Yebin Liu, Qingyao Wu
Abstract:
Hand-Object Interaction (HOI) generation has significant application potential. However, current 3D HOI motion generation approaches heavily rely on predefined 3D object models and lab-captured motion data, limiting generalization capabilities. Meanwhile, HOI video generation methods prioritize pixel-level visual fidelity, often sacrificing physical plausibility. Recognizing that visual appearance and motion patterns share fundamental physical laws in the real world, we propose a novel framework that combines visual priors and dynamic constraints within a synchronized diffusion process to generate the HOI video and motion simultaneously. To integrate the heterogeneous semantics, appearance, and motion features, our method implements tri-modal adaptive modulation for feature aligning, coupled with 3D full-attention for modeling inter- and intra-modal dependencies. Furthermore, we introduce a vision-aware 3D interaction diffusion model that generates explicit 3D interaction sequences directly from the synchronized diffusion outputs, then feeds them back to establish a closed-loop feedback cycle. This architecture eliminates dependencies on predefined object models or explicit pose guidance while significantly enhancing video-motion consistency. Experimental results demonstrate our method's superiority over state-of-the-art approaches in generating high-fidelity, dynamically plausible HOI sequences, with notable generalization capabilities in unseen real-world scenarios.



Paperid:5122
Authors:Jiaming Han, Hao Chen, Yang Zhao, Qi Zhao, Ziyan Yang, Hanyu Wang, Hao He, Xiangyu Yue, Lu Jiang
Abstract:
This paper presents a multimodal framework that attempts to unify visual understanding and generation within a shared discrete semantic representation. At its core is the Text-Aligned Tokenizer (TA-Tok), which converts images into discrete tokens using a text-aligned codebook projected from a large language model's (LLM) vocabulary. By integrating vision and text into a unified space with an expanded vocabulary, our multimodal LLM,Tar, enables cross-modal input and output through a shared interface, without the need for modality-specific designs. Additionally, we propose scale-adaptive encoding and decoding to balance efficiency and visual detail, along with a generative de-tokenizer to produce high-fidelity visual outputs. To address diverse decoding needs, we utilize two complementary de-tokenizers: a fast autoregressive model and a diffusion-based model. To enhance modality fusion, we investigate advanced pre-training tasks, demonstrating improvements in both visual understanding and generation. Experiments across benchmarks show thatTarmatches or surpasses existing multimodal LLM methods, achieving faster convergence and greater training efficiency. All code, models, and data will be made publicly available.



Paperid:5123
Authors:Yiyao Ma, Kai Chen, Kexin ZHENG, DOU QI
Abstract:
Dexterous grasp generation is a fundamental challenge in robotics, requiring both grasp stability and adaptability across diverse objects and tasks. Analytical methods ensure stable grasps but are inefficient and lack task adaptability, while generative approaches improve efficiency and task integration but generalize poorly to unseen objects and tasks due to data limitations. In this paper, we propose a transfer-based framework for dexterous grasp generation, leveraging a conditional diffusion model to transfer high-quality grasps from shape templates to novel objects within the same category. Specifically, we reformulate the grasp transfer problem as the generation of an object contact map, incorporating object shape similarity and task specifications into the diffusion process. To handle complex shape variations, we introduce a dual mapping mechanism, capturing intricate geometric relationship between shape templates and novel objects. Beyond the contact map, we derive two additional object-centric maps, the part map and direction map, to encode finer contact details for more stable grasps. We then develop a cascaded conditional diffusion model framework to jointly transfer these three maps, ensuring their intra-consistency. Finally, we introduce a robust grasp recovery mechanism, identifying reliable contact points and optimizing grasp configurations efficiently. Extensive experiments demonstrate the superiority of our proposed method. Our approach effectively balances grasp quality, generation efficiency, and generalization performance across various tasks.



Paperid:5124
Authors:Luca Andolfi, Eleonora Giunchiglia
Abstract:
Neurosymbolic methods combine neural perception with symbolic reasoning, but recent studies have shown that state-of-the-art approaches can fall prey of reasoning shortcuts---spurious but predictive associations between learned concepts and target labels that satisfy constraints syntactically, yet violate the intended semantics. In this paper, we investigate the use of prototype-augmented neural architectures to mitigate reasoning shortcuts. By adopting prototypical networks as a backbone, we introduce a method that anchors intermediate concepts in interpretable embeddings and guides their training using the available limited supervision and the symbolic background knowledge. Extensive experiments demonstrate that our approach significantly improves semantic alignment while maintaining predictive performance across multiple benchmarks.



Paperid:5125
Authors:Akash Kundu, Stefano Mangini
Abstract:
Abstract:Variational quantum algorithms hold the promise to address meaningful quantum problems already on noisy intermediate-scale quantum hardware, but they face the challenge of designing quantum circuits that both solve the target problem and comply with device limitations. Quantum architecture search (QAS) automates this design process, with reinforcement learning (RL) emerging as a promising approach. Yet, RL-based QAS methods encounter significant scalability issues, as computational and training costs grow rapidly with the number of qubits, circuit depth, and noise, severely impacting performance. To address these challenges, we introduce $\textit{TensorRL-QAS}$, a scalable framework that combines tensor network (TN) methods with RL for designing quantum circuits. By warm-starting the architecture search with a matrix product state approximation of the target solution, TensorRL-QAS effectively narrows the search space to physically meaningful circuits, accelerating convergence to the desired solution. Tested on several quantum chemistry problems of up to 12-qubit, TensorRL-QAS achieves up to a 10-fold reduction in CNOT count and circuit depth compared to baseline methods, while maintaining or surpassing chemical accuracy. It reduces function evaluations by up to 100-fold, accelerates training episodes by up to 98\%, and achieves up to 50\% success probability for 10-qubit systems---far exceeding the $<$1\% rates of baseline approaches. Robustness and versatility are demonstrated both in the noiseless and noisy scenarios, where we report a simulation of up to 8-qubit. Furthermore, TensorRL-QAS demonstrates scalability to 20-qubit systems, positioning it as an efficient quantum circuit discovery framework for near-term hardware and beyond.



Paperid:5126
Authors:Zengxi Zhang, Junchen Ge, Zhiying Jiang, Miao Zhang, Jinyuan Liu
Abstract:
Deep learning-based image stitching methods have achieved promising performance on conventional stitching datasets. However, real-world scenarios may introduce challenges such as complex weather conditions, illumination variations, and dynamic scene motion, which severely degrade image quality and lead to significant misalignment in stitching results. To solve this problem, we propose an adverse condition-tolerant image stitching network, dubbed ACDIS. We first introduce a bidirectional consistency learning framework, which ensures reliable alignment through an iterative optimization paradigm that integrates differentiable image restoration and Gaussian-distribute encoded homography estimation. Subsequently, we incorporate motion constraints into the seamless composition network to produce robust stitching results without interference from moving scenes. We further propose the first adverse scene image stitching dataset, which covers diverse parallax and scenes under low-light, haze, and underwater environments. Extensive experiments show that the proposed method can generate visually pleasing stitched images under adverse conditions, outperforming state-of-the-art methods.



Paperid:5127
Authors:Zuwei Long, Yunhang Shen, Chaoyou Fu, Heting Gao, Lijiang Li, Peixian Chen, Mengdan Zhang, Hang Shao, Jian Li, Jinlong Peng, Haoyu Cao, Ke Li, Rongrong Ji, Xing Sun
Abstract:
With the growing requirement for natural human-computer interaction, speech-based systems receive increasing attention as speech is one of the most common forms of daily communication. However, the existing speech models still experience high latency when generating the first audio token during streaming, which poses a significant bottleneck for deployment. To address this issue, we propose VITA-Audio, an end-to-end large speech model with fast audio-text token generation. Specifically, we introduce a lightweight Multiple Cross-modal Token Prediction (MCTP) module that efficiently generates multiple audio tokens within a single model forward pass, which not only accelerates the inference but also significantly reduces the latency for generating the first audio in streaming scenarios. In addition, a four-stage progressive training strategy is explored to achieve model acceleration with minimal loss of speech quality. To our knowledge, VITA-Audio is the first multi-modal large language model capable of generating audio output during the first forward pass, enabling real-time conversational capabilities with minimal latency. VITA-Audio is fully reproducible and is trained on open-source data only. Experimental results demonstrate that our model achieves an inference speedup of 3~5x at the 7B parameter scale, but also significantly outperforms open-source models of similar model size on multiple benchmarks for automatic speech recognition (ASR), text-to-speech (TTS), and spoken question answering (SQA) tasks.



Paperid:5128
Authors:Yang Li, Qiang Sheng, Yehan Yang, Xueyao Zhang, Juan Cao
Abstract:
Though safety alignment has been applied to most large language models (LLMs), LLM service providers generally deploy a subsequent moderation as the external safety guardrail in real-world products. Existing moderators mainly practice a conventional full detection, which determines the harmfulness based on the complete LLM output, causing high service latency. Recent works pay more attention to partial detection where moderators oversee the generation midway and early stop the output if harmfulness is detected, but they directly apply moderators trained with the full detection paradigm to incomplete outputs, introducing a training-inference gap that lowers the performance. In this paper, we explore how to form a data-and-model solution that natively supports partial detection. For the data, we constructFineHarm, a dataset consisting of 29K prompt-response pairs with fine-grained token-level annotations to provide reasonable supervision for token-level training. Then, we propose thestreaming content monitor, which is trained with dual supervision of response- and token-level labels and can follow the output stream of LLM to make a timely judgment of harmfulness. Experiments show that SCM gains 0.95+ in macro F1 score that is comparable to full-detection, by only seeing the first 18% of tokens in responses on average. Moreover, the SCM can serve as a pseudo-harmfulness annotator for improving safety alignment and lead to a higher harmlessness score than DPO.



Paperid:5129
Authors:Bozhou Zhang, Nan Song, jingyu li, Xiatian Zhu, Jiankang Deng, Li Zhang
Abstract:
End-to-end autonomous driving methods aim to directly map raw sensor inputs to future driving actions such as planned trajectories, bypassing traditional modular pipelines. While these approaches have shown promise, they often operate under a one-shot paradigm that relies heavily on the current scene context, potentially underestimating the importance of scene dynamics and their temporal evolution. This limitation restricts the model’s ability to make informed and adaptive decisions in complex driving scenarios. We propose a new perspective: the future trajectory of an autonomous vehicle is closely intertwined with the evolving dynamics of its environment, and conversely, the vehicle’s own future states can influence how the surrounding scene unfolds. Motivated by this bidirectional relationship, we introduceSeerDrive, a novel end-to-end framework that jointly models future scene evolution and trajectory planning in a closed-loop manner. Our method first predicts future bird’s-eye view (BEV) representations to anticipate the dynamics of the surrounding scene, then leverages this foresight to generate future-context-aware trajectories. Two key components enable this: (1) future-aware planning, which injects predicted BEV features into the trajectory planner, and (2) iterative scene modeling and vehicle planning, which refines both future scene prediction and trajectory generation through collaborative optimization. Extensive experiments on the NAVSIM and nuScenes benchmarks show that SeerDrive significantly outperforms existing state-of-the-art methods. Our code will be released.



Paperid:5130
Authors:Yuanxi Yu, Fan Jiang, Xinzhu Ma, Liang Zhang, Bozitao Zhong, Wanli Ouyang, Guisheng Fan, Huiqun Yu, Liang Hong, Mingchen Li
Abstract:
Abstract:In-silico prediction of protein mutant stability, measured by the difference in Gibbs free energy change ($\Delta \Delta G$), is fundamental for protein engineering.Current sequence-to-label methods typically employ two-stage pipelines: (i) encoding mutant sequences using neural networks (e.g., transformers), followed by (ii) the $\Delta \Delta G$ regression from the latent representations.Although these methods have demonstrated promising performance, their dependence on specialized neural network encoders significantly increases the complexity.Additionally, the requirement to compute latent representations individually for each mutant sequence negatively impacts computational efficiency and poses the risk of overfitting.This work proposes the MAXWELL framework, which reformulates mutation $\Delta \Delta G$ prediction as a sequence-to-landscape task.In MAXWELL, mutations of a protein and their corresponding $\Delta \Delta G$ values are organized into a landscape matrix, allowing our framework to learn the $\Delta \Delta G$ landscape of a protein with a single forward and backward pass during training. To this end, we curated a new $\Delta \Delta G$ benchmark dataset with strict controls on data leakage and redundancy to ensure robust evaluation.Leveraging the zero-shot scoring capability of protein language models (PLMs), MAXWELL effectively utilizes the evolutionary patterns learned by PLMs during pre-training.More importantly, MAXWELL is compatible with multiple protein language models.For example, when integrated with the ESM-IF, MAXWELL achieves higher accuracy than ThermoMPNN with 10$\times$ faster in inference speed (despite having 50$\times$ more parameters than ThermoMPNN).The training codes, model weights, and datasets will be publicly available.



Paperid:5131
Authors:Wentao Wu, Shiyuan He, Jianhua Guo
Abstract:
Abstract:Bayesian networks (BNs), represented by directed acyclic graphs (DAGs), provide a principled framework for modeling complex dependencies among random variables. As data dimensionality increases into the tens of thousands, fitting and marginalizing a full BN becomes computationally prohibitive—particularly when inference is only needed for a small subset of variables. Estimation-collapsibility addresses this challenge by ensuring that directly fitting a submodel, obtained by ignoring non-essential variables, still yields exact inference on target variables. However, current DAG-based criterion for checking estimation-collapsibility is computationally intensive, involving exhaustive vertex searches and iterative removals. Additionally, practical applications typically identify the underlying DAG only up to its Markov equivalence class, represented by a completed partially directed acyclic graph (CPDAG). To bridge this gap, we introduce sequential $c$-simplicial sets—a novel graphical characterization of estimation-collapsibility applicable directly to CPDAGs. We further propose DSCS, a computationally efficient algorithm for verifying estimation-collapsibility within CPDAG framework that scales effectively to high-dimensional BNs. Extensive numerical experiments demonstrate the practicality, scalability, and efficiency of our proposed approach.



Paperid:5132
Authors:Weixiang Zhao, Xingyu Sui, Yulin Hu, Jiahe Guo, Haixiao Liu, Biye Li, Yanyan Zhao, Bing Qin, Ting Liu
Abstract:
Personalized alignment is essential for enabling large language models (LLMs) to engage effectively in user-centric dialogue. While recent prompt-based and offline optimization methods offer preliminary solutions, they fall short in cold-start scenarios and long-term personalization due to their inherently static and shallow designs. In this work, we introduce the Reinforcement Learning for Personalized Alignment (RLPA) framework, in which an LLM interacts with a simulated user model to iteratively infer and refine user profiles through dialogue. The training process is guided by a dual-level reward structure: the Profile Reward encourages accurate construction of user representations, while the Response Reward incentivizes generation of responses consistent with the inferred profile. We instantiate RLPA by fine-tuning Qwen-2.5-3B-Instruct, resulting in Qwen-RLPA, which achieves state-of-the-art performance in personalized dialogue. Empirical evaluations demonstrate that Qwen-RLPA consistently outperforms prompting and offline fine-tuning baselines, and even surpasses advanced commercial models such as Claude-3.5 and GPT-4o. Further analysis highlights Qwen-RLPA's robustness in reconciling conflicting user preferences, sustaining long-term personalization and delivering more efficient inference compared to recent reasoning-focused LLMs. These results emphasize the potential of dynamic profile inference as a more effective paradigm for building personalized dialogue systems.



Paperid:5133
Authors:Raktim Goswami, Prashanth Krishnamurthy, Yann LeCun, Farshad Khorrami
Abstract:
Visual imitation learning enables robotic agents to acquire skills by observing expert demonstration videos. In the one-shot setting, the agent generates a policy after observing a single expert demonstration without additional fine-tuning. Existing approaches typically train and evaluate on the same set of tasks, varying only object configurations, and struggle to generalize to unseen tasks with different semantic or structural requirements. While some recent methods attempt to address this, they exhibit low success rates on hard test tasks that, despite being visually similar to some training tasks, differ in context and require distinct responses. Additionally, most existing methods lack an explicit model of environment dynamics, limiting their ability to reason about future states. To address these limitations, we propose a novel framework for one-shot visual imitation learning via world-model-guided trajectory generation. Given an expert demonstration video and the agent’s initial observation, our method leverages a learned world model to predict a sequence of latent states and actions. This latent trajectory is then decoded into physical waypoints that guide the agent’s execution. Our method is evaluated on two simulated benchmarks and three real-world robotic platforms, where it consistently outperforms prior approaches, with over 30% improvement in some cases.



Paperid:5134
Authors:Aleksandrs Slivkins, Yunzong Xu, Shiliang Zuo
Abstract:
We study the greedy (exploitation-only) algorithm in bandit problems with a known reward structure. We allow arbitrary finite reward structures, while prior work focused on a few specific ones. We fully characterize when the greedy algorithm asymptotically succeeds or fails, in the sense of sublinear vs. linear regret as a function of time.Our characterization identifies a partial identifiability property of the problem instance as the necessary and sufficient condition for the asymptotic success. Notably, once this property holds, the problem becomes easy—\emph{any} algorithm will succeed (in the same sense as above), provided it satisfies a mild non-degeneracy condition. Our characterization extends to contextual bandits and interactive decision-making with arbitrary feedback. Examples demonstrating broad applicability and extensions to infinite reward structures are provided.



Paperid:5135
Authors:Junhao Dong, Hao Zhu, Yifei Zhang, Xinghua Qu, Yew Soon Ong, Piotr Koniusz
Abstract:
Abstract:As foundation Vision-Language Models (VLMs) unlock fine-tuning on smaller datasets while capturing large-scale data in pre-training, machine unlearning becomes critical in addressing privacy concerns and regulatory compliance. Task vector, representing differences between parameters of models fine-tuned with and without specific data, is a popular retraining-free unlearning strategy. However, we observe that task vectors exhibit substantial sensitivity to various fine-tuning configurations, resulting in unstable unlearning effectiveness negatively correlating with prediction-level variance. While aggregating multiple functions (e.g., VLM with classifier) whose parameters are represented by different task vectors naturally reduces function variance and improves unlearning, the computational cost of obtaining numerous task vectors and aggregating functions is computationally high. Thus, to robustly capture the space of task vectors induced by diverse fine-tuning strategies, we propose modeling it within the convex hull of $(Q-1)$-simplex whose vertices are $Q$ task vectors. Although a function ensemble can be formed by sampling numerous task vectors from such a simplex, this is computationally prohibitive. Thus, we derive a closed-form ensemble of an infinite number of functions whose parameters are uniformly sampled from the simplex, enabling efficient function-level task vector ensembling with enhanced unlearning performance. Extensive experiments and analyses across diverse datasets and scenarios demonstrate the efficacy of our method.



Paperid:5136
Authors:Wenlong Li, Yifei Xu, Yuan Rao, Zhenhua Wang, Shuiguang Deng
Abstract:
Video anomaly detection (VAD) focuses on identifying anomalies in videos. Supervised methods demand substantial in-domain training data and fail to deliver clear explanation for anomalies. In contrast, training-free methods leverage the knowledge reserve and language interactivity of large pre-trained models to detect anomalies. However, the current fixed-length temporal window sampling approaches struggle to accurately capture anomalies with varying temporal spans.Therefore, we propose \textbf{\methodshort}~that utilizes a Hierarchical Granularity-aware Tree (HGTree) structure for adaptive sampling VAD. \methodshort~ leverages the knowledge embedded in a pre-trained Generic Event Boundary Detection (GEBD) model to characterize potential anomaly event boundaries.Specifically, \methodshort~first decomposes the video into an HGTree based on generic event nodes using boundary confidence, and performs adaptive coarse-fine stratification and redundancy removal. Then, the multi-dimensional priors are injected into the vision-language models (VLMs) to enhance the abnormal perception of the node-wise video description, and robust anomaly reasoning is achieved for generic event nodes based on the large language models (LLMs).Finally, an intra-cluster correlation method is used to integrate the multi-granular anomaly scores. Extensive experiments on UCF-Crime and XD-Violence datasets demonstrate that \methodshort~achieves state-of-the-art performance in training-free settings while drastically reducing the amount of video samples. Our code is publicly available at\url{https://anonymous.4open.science/r/\methodshort-6E11/}.



Paperid:5137
Authors:Weibin Liao, Tianlong Wang, Yinghao Zhu, Yasha Wang, Junyi Gao, Liantao Ma
Abstract:
Medical Lay Language Generation (MLLG) plays a vital role in improving the accessibility of complex scientific content for broader audiences. Recent literature to MLLG commonly employ parameter-efficient fine-tuning methods such as Low-Rank Adaptation (LoRA) to fine-tuning large language models (LLMs) using paired expert-lay language datasets. However, LoRA struggles with the challenges posed by multi-source heterogeneous MLLG datasets. Specifically, through a series of exploratory experiments, we reveal that standard LoRA fail to meet the requirement for semantic fidelity and diverse lay-style generation in MLLG task. To address these limitations, we propose Magical, an asymmetric LoRA architecture tailored for MLLG under heterogeneous data scenarios. Magical employs a shared matrix A for abstractive summarization, along with multiple isolated matrices B for diverse lay-style generation. To preserve semantic fidelity during the lay language generation process, Magical introduces a Semantic Invariance Constraint to mitigate semantic subspace shifts on matrix A. Furthermore, to better adapt to diverse lay-style generation, Magical incorporates the Recommendation-guided Switch, an externally interface to prompt the LLM to switch between different matrices B. Experimental results on three real-world lay language generation datasets demonstrate that Magical consistently outperforms prompt-based methods, vanilla LoRA, and its recent variants, while also reducing trainable parameters by 31.66%. Our code is publicly available at https://anonymous.4open.science/r/Magical.



Paperid:5138
Authors:Yuhao Zhou, Jintao Xu, Bingrui Li, Chenglong Bao, Chao Ding, Jun Zhu
Abstract:
Abstract:Finding an $\epsilon$-stationary point of a nonconvex function with a Lipschitz continuous Hessian is a central problem in optimization. Regularized Newton methods are a classical tool and have been studied extensively, yet they still face a trade‑off between global and local convergence. Whether a parameter-free algorithm of this type can simultaneously achieve optimal global complexity and quadratic local convergence remains an open question. To bridge this long-standing gap, we propose a new class of regularizers constructed from the current and previous gradients, and leverage the conjugate gradient approach with a negative curvature monitor to solve the regularized Newton equation. The proposed algorithm is adaptive, requiring no prior knowledge of the Hessian Lipschitz constant, and achieves a global complexity of $O(\epsilon^{-\frac{3}{2}})$ in terms of the second-order oracle calls, and $\tilde O(\epsilon^{-\frac{7}{4}})$ for Hessian-vector products, respectively. When the iterates converge to a point where the Hessian is positive definite, the method exhibits quadratic local convergence. Preliminary numerical results, including training the physics-informed neural networks, illustrate the competitiveness of our algorithm.



Paperid:5139
Authors:Zihao Jing, Yan Sun, Yan Li, Sugitha Janarthanan, Alana Deng, Pingzhao Hu
Abstract:
Multimodal molecular models often suffer from 3D conformer unreliability and modality collapse, limiting their robustness and generalization. We propose MuMo, a structured multimodal fusion framework that addresses these challenges in molecular representation through two key strategies. To reduce the instability of conformer-dependent fusion, we design a Structured Fusion Pipeline (SFP) that combines 2D topology and 3D geometry into a unified and stable structural prior. To mitigate modality collapse caused by naive fusion, we introduce a Progressive Injection (PI) mechanism that asymmetrically integrates this prior into the sequence stream, preserving modality-specific modeling while enabling cross-modal enrichment. Built on a state space backbone, MuMo supports long-range dependency modeling and robust information propagation. Across 21 benchmark tasks from Therapeutics Data Commons (TDC) and MoleculeNet, MuMo achieves an average improvement of 2.7\% over the best-performing baseline on each task, ranking first on 17 of them, including a 27\% improvement on the LD50 task. These results validate its robustness to 3D conformer noise and the effectiveness of multimodal fusion in molecular representation. The code is available at: https://anonymous.4open.science/r/MuMo.



Paperid:5140
Authors:Xin Yu, Yujia Wang, Jinghui Chen, Lingzhou Xue
Abstract:
Low-Rank Adaptation (LoRA) has emerged as an effective technique for reducing memory overhead in fine-tuning large language models. However, it often suffers from sub-optimal performance compared with full fine-tuning since the update is constrained in the low-rank space. Recent variants such as LoRA-Pro attempt to mitigate this by adjusting the gradients of the low-rank matrices to approximate the full gradient. However, LoRA-Pro's solution is not unique, and different solutions can lead to significantly varying performance in ablation studies. Besides, to incorporate momentum or adaptive optimization design, approaches like LoRA-Pro must first compute the equivalent gradient, causing a higher memory cost close to full fine-tuning. A key challenge remains in integrating momentum properly into the low-rank space with lower memory cost. In this work, we propose AltLoRA, an alternating projection method that avoids the difficulties in gradient approximation brought by the joint update design, meanwhile integrating momentum without higher memory complexity. Our theoretical analysis provides convergence guarantees and further shows that AltLoRA enables stable feature learning and robustness to transformation invariance. Extensive experiments across multiple tasks demonstrate that AltLoRA outperforms LoRA and its variants, narrowing the gap toward full fine-tuning while preserving superior memory efficiency.



Paperid:5141
Authors:Jaesin Ahn, Heechul Jung
Abstract:
Diffusion models show remarkable image generation performance following text prompts, but risk generating sexual contents. Existing approaches, such as prompt filtering, concept removal, and even sexual contents mitigation methods, struggle to defend against adversarial attacks while maintaining benign image quality. In this paper, we propose a novel approach called Distorting Embedding Space (DES), a text encoder-based defense mechanism that effectively tackles these issues through innovative embedding space control. DES transforms unsafe embeddings, extracted from a text encoder using unsafe prompts, toward carefully calculated safe embedding regions to prevent unsafe contents generation, while reproducing the original safe embeddings. DES also neutralizes the ``nudity'' embedding, by aligning it with neutral embedding to enhance robustness against adversarial attacks. As a result, extensive experiments on explicit content mitigation and adaptive attack defense show that DES achieves state-of-the-art (SOTA) defense, with attack success rate (ASR) of 9.47\% on FLUX.1, a recent popular model, and 0.52\% on the widely adopted Stable Diffusion v1.5. These correspond to ASR reductions of 76.5\% and 63.9\% compared to previous SOTA methods, EraseAnything and AdvUnlearn, respectively. Furthermore, DES maintains benign image quality, achieving Frechet Inception Distance and CLIP score comparable to those of the original FLUX.1 and Stable Diffusion v1.5.



Paperid:5142
Authors:Zhenghao Zeng, David Arbour, Avi Feller, Ishita Dasgupta, Atanu Sinha, Edward Kennedy
Abstract:
Human annotations play a crucial role in evaluating the performance of GenAI models. Two common challenges in practice, however, are missing annotations (the response variable of interest) and cluster dependence among human-AI interactions (e.g., questions asked by the same user may be highly correlated). Reliable inference must address both these issues to achieve unbiased estimation and appropriately quantify uncertainty when estimating average scores from human annotations. In this paper, we analyze the doubly robust estimator, a widely used method in missing data analysis and causal inference, applied to this setting and establish novel theoretical properties under cluster dependence. We further illustrate our findings through simulations and a real-world conversation quality dataset. Our theoretical and empirical results underscore the importance of incorporating cluster dependence in missing response problems to perform valid statistical inference.



Paperid:5143
Authors:Khanh-Hung (Bruce) Giang-Tran, Soroosh Shafiee, Nam Ho-Nguyen
Abstract:
Abstract:We propose efficient methods for solving stochastic convex bilevel optimization problems, where the goal is to minimize an outer stochastic objective function subject to the solution set of an inner stochastic optimization problem. Existing methods often rely on costly projection or linear optimization oracles over complex sets, which limits scalability. To overcome this, we propose an iteratively regularized conditional gradient framework that leverages efficient linear optimization oracles exclusively over the base feasible set. Our proposed methods employ a vanishing regularization sequence that progressively emphasizes the inner problem while biasing towards desirable minimal outer objective solutions. Under standard convexity assumptions, we establish non-asymptotic convergence rates of $O(t^{-({1}/{2}-p)})$ for the outer objective and $O(t^{-p})$ for the inner objective, where $p \in (0,1/2)$ controls the regularization decay, in the one-sample stochastic setting, and $O(t^{-(1-p)})$ and $O(t^{-p})$ in the finite-sum setting using a mini-batch scheme, where $p \in (0,1)$. Experimental results on over-parametrized regression and $\ell_1$-constrained logistics regression tasks demonstrate the practical advantages of our approach over existing methods, confirming our theoretical findings.



Paperid:5144
Authors:Ben Adcock, Zi Yuan (Nick) Huang
Abstract:
Abstract:We study the sample complexity of Bayesian recovery for solving inverse problems with general prior, forward operator and noise distributions. We consider posterior sampling according to an approximate prior $\mathcal{P}$, and establish sufficient conditions for stable and accurate recovery with high probability. Our main result is a non-asymptotic bound that shows that the sample complexity depends on (i) the intrinsic complexity of $\mathcal{P}$, quantified by its *approximate covering number*, and (ii) concentration bounds for the forward operator and noise distributions. As a key application, we specialize to generative priors, where $\mathcal{P}$ is the pushforward of a latent distribution via a Deep Neural Network (DNN). We show that the sample complexity scales log-linearly with the latent dimension $k$, thus establishing the efficacy of DNN-based priors. Generalizing existing results on deterministic (i.e., non-Bayesian) recovery for the important problem of random sampling with an orthogonal matrix $U$, we show how the sample complexity is determined by the *coherence* of $U$ with respect to the support of $\mathcal{P}$. Hence, we establish that coherence plays a fundamental role in Bayesian recovery as well. Overall, our framework unifies and extends prior work, providing rigorous guarantees for the sample complexity of solving Bayesian inverse problems with arbitrary distributions.



Paperid:5145
Authors:Badi Li, Ren-Jie Lu, Yu Zhou, Jingke Meng, Wei-Shi Zheng
Abstract:
The Object Goal Navigation (ObjectNav) task challenges agents to locate a specified object in an unseen environment by imagining unobserved regions of the scene. Prior approaches rely on deterministic and discriminative models to complete semantic maps, overlooking the inherent uncertainty in indoor layouts and limiting their ability to generalize to unseen environments. In this work, we propose GOAL, a generative flow-based framework that models the semantic distribution of indoor environments by bridging observed regions with LLM-enriched full-scene semantic maps. During training, spatial priors inferred from large language models (LLMs) are encoded as two-dimensional Gaussian fields and injected into target maps, distilling rich contextual knowledge into the flow model and enabling more generalizable completions. Extensive experiments demonstrate that GOAL achieves state-of-the-art performance on MP3D and Gibson, and shows strong generalization in transfer settings to HM3D.



Paperid:5146
Authors:Yanru Sun, Zongxia Xie, Emadeldeen Eldele, Dongyue Chen, Qinghua Hu, Min Wu
Abstract:
Time series forecasting, which aims to predict future values based on historical data, has garnered significant attention due to its broad range of applications. However, real-world time series often exhibit heterogeneous pattern evolution across segments, such as seasonal variations, regime changes, or contextual shifts, making accurate forecasting challenging. Existing approaches, which typically train a single model to capture all these diverse patterns, often struggle with the pattern drifts between patches and may lead to poor generalization. To address these challenges, we propose TFPS, a novel architecture that leverages pattern-specific experts for more accurate and adaptable time series forecasting. TFPS employs a dual-domain encoder to capture both time-domain and frequency-domain features, enabling a more comprehensive understanding of temporal dynamics. It then performs subspace clustering to dynamically identify distinct patterns across data segments. Finally, these patterns are modeled by specialized experts, allowing the model to learn multiple predictive functions. Extensive experiments on real-world datasets demonstrate that TFPS outperforms state-of-the-art methods, particularly on datasets exhibiting significant distribution shifts. The data and code are available: https://anonymous.4open.science/r/TFPS-7F08.



Paperid:5147
Authors:Imon Banerjee, Sayak Chakrabarty
Abstract:
Abstract:The m-out-of-n bootstrap—proposed by \cite{bickel1992resampling}—approximates the distribution of a statistic by repeatedly drawing $m$ subsamples ($m \ll n$) without replacement from an original sample of size n; it is now routinely used for robust inference with heavy-tailed data, bandwidth selection, and other large-sample applications. Despite this broad applicability across econometrics, biostatistics, and machine-learning workflows, rigorous parameter-free guarantees for the soundness of the m-out-of-n bootstrap when estimating sample quantiles have remained elusive.This paper establishes such guarantees by analysing the estimator of sample quantiles obtained from m-out-of-n resampling of a dataset of length n. We first prove a central limit theorem for a fully data-driven version of the estimator that holds under a mild moment condition and involves no unknown nuisance parameters. We then show that the moment assumption is essentially tight by constructing a counter-example in which the CLT fails. Strengthening the assumptions slightly, we derive an Edgeworth expansion that delivers exact convergence rates and, as a corollary, a Berry–Esséen bound on the bootstrap approximation error. Finally, we illustrate the scope of our results by obtaining parameter-free asymptotic distributions for practical statistics, including the quantiles for random walk MH, and rewards of ergodic MDP's, thereby demonstrating the usefulness of our theory in modern estimation and learning tasks.



Paperid:5148
Authors:Mingxuan Ye, Jie Wang, Fangzhou, Zhihai Wang, Yufei Kuang, Xijun Li, Weilin Luo, Jianye Hao, Feng Wu
Abstract:
Cutting planes (cuts) are essential for solving mixed-integer linear programming (MILP) problems, as they tighten the feasible solution space and accelerate the solving process. Modern MILP solvers offer diverse cutting plane separators to generate cuts, enabling users to leverage their potential complementary strengths to tackle problems with different structures. Recent machine learning approaches learn to configure separators based on problem-specific features, selecting effective separators and deactivating ineffective ones to save unnecessary computing time. However, they ignore the dynamics of separator efficacy at different stages of cut generation and struggle to adapt the configurations for the evolving problems after multiple rounds of cut generation. To address this challenge, we propose a noveldynamicseparator configuration (DynSep) method that models separator configuration in different rounds as a reinforcement learning task, making decisions based on an incremental triplet graph updated by iteratively added cuts. Specifically, we tokenize the incremental subgraphs and utilize a decoder-only Transformer as our policy to autoregressively predict when to halt separation and which separators to activate at each round. Evaluated on synthetic and large-scale real-world MILP problems, DynSep speeds up average solving time by 64% on easy and medium datasets, and reduces primal-dual gap integral within the given time limit by 16% on hard datasets. Moreover, experiments demonstrate that DynSep well generalizes to MILP instances of significantly larger sizes than those seen during training.



Paperid:5149
Authors:Wenquan Lu, Jiaqi Zhang, Hugues Van Assel, Randall Balestriero
Abstract:
Abstract:Self-Supervised Learning (SSL) has become a powerful solution to extract rich representations from unlabeled data. Yet, SSL research is mostly focused on clean, curated and high-quality datasets. As a result, applying SSL on noisy data remains a challenge, despite being crucial to applications such as astrophysics, medical imaging, geophysics or finance. In this work, we present a fully self-supervised framework that enables noise-robust representation learning without requiring a denoiser at inference or downstream fine-tuning. Our method first trains an SSL denoiser on noisy data, then uses it to construct a denoised-to-noisy data curriculum (i.e., training first on denoised, then noisy samples) for pretraining a SSL backbone (e.g., DINOv2), combined with a teacher-guided regularization that anchors noisy embeddings to their denoised counterparts. This process encourages the model to internalize noise robustness. Notably, the denoiser can be discarded after pretraining, simplifying deployment. On ImageNet-1k with ViT-B under extreme Gaussian noise ($\sigma=255$, SNR = 0.72 dB), our method improves linear probing accuracy by 4.8\% over DINOv2, demonstrating that denoiser-free robustness can emerge from noise-aware pretraining.



Paperid:5150
Authors:Xingyu Chen, Bokun Wang, Ming Yang, Qihang Lin, Tianbao Yang
Abstract:
Abstract:Finite-sum Coupled Compositional Optimization (FCCO), characterized by its coupled compositional objective structure, emerges as an important optimization paradigm for addressing a wide range of machine learning problems. In this paper, we focus on a challenging class of non-convex non-smooth FCCO, where the outer functions are non-smooth weakly convex or convex and the inner functions are smooth or weakly convex. Existing state-of-the-art result face two key limitations: (1) a high iteration complexity of $O(1/\epsilon^6)$ under the assumption that the stochastic inner functions are Lipschitz continuous in expectation; (2) reliance on vanilla SGD-type updates, which are not suitable for deep learning applications. Our main contributions are two fold: (i) We propose stochastic momentum methods tailored for non-smooth FCCO that come with provable convergence guarantees; (ii) We establish a **new state-of-the-art** iteration complexity of $O(1/\epsilon^5)$. Moreover, we apply our algorithms to multiple inequality constrained non-convex optimization problems involving smooth or weakly convex functional inequality constraints. By optimizing a smoothed hinge penalty based formulation, we achieve a **new state-of-the-art** complexity of $O(1/\epsilon^5)$ for finding an (nearly) $\epsilon$-level KKT solution. Experiments on three tasks demonstrate the effectiveness of the proposed algorithms.



Paperid:5151
Authors:Zitong Shi, Guancheng Wan, Haixin Wang, Ruoyan Li, Zijie Huang, Wanjia Zhao, Yijia Xiao, Xiao Luo, Carl Yang, Yizhou Sun, Wei Wang
Abstract:
Abstract:Recent studies reveal that large language models (LLMs) often struggle to resolve conflicting instructions embedded within hierarchical prompts, resulting in decreased compliance with system-level directives and compromising the reliability of safety-critical applications. While earlier approaches attempt to improve instruction hierarchy awareness through prompt engineering or embedding-level modifications, they typically lack structural modeling and either offer limited gains or require extensive fine-tuning. In this work, we introduce $\textbf{FocalLoRA}$, a parameter-efficient and structure-aware framework that strengthens hierarchical instruction adherence by selectively optimizing structurally critical attention heads, referred to as $\textit{focal heads}$, which exhibit heightened sensitivity to instruction conflicts. Experiments across multiple models and a dedicated benchmark demonstrate that FocalLoRA markedly enhances system instruction compliance with minimal tuning cost. For instance, on Llama-8B, fine-tuning only 0.0188\% of parameters yields a 35.52\% $\uparrow$ in system instruction compliance.



Paperid:5152
Authors:Chong Tang, Joseph Powell, Dirk Koch, Robert Mullins, Alex Weddell, Jagmohan Chauhan
Abstract:
Abstract:Recent progress on Remote Sensing Foundation Models (RSFMs) aims toward universal representations for Earth observation imagery. However, current efforts often scale up in size significantly without addressing efficiency constraints critical for real-world applications (e.g., onboard processing, rapid disaster response) or treat multispectral (MS) data as generic imagery, overlooking valuable physical priors. We introduce PhySwin, a foundation model for MS data that integrates physical priors with computational efficiency. PhySwin combines three innovations: (i) physics-informed pretraining objectives leveraging radiometric constraints to enhance feature learning; (ii) an efficient MixMAE formulation tailored to SwinV2 for low-FLOP, scalable pretraining; and (iii) token-efficient spectral embedding to retain spectral detail without increasing token counts. Pretrained on over 1M Sentinel-2 tiles, PhySwin achieves SOTA results (+1.32\% mIoU segmentation, +0.80\% F1 change detection) while reducing inference latency by up to 14.4$\times$ and computational complexity by up to 43.6$\times$ compared to ViT-based RSFMs.



Paperid:5153
Authors:Haibo Wang, Bo Feng, Zhengfeng Lai, Mingze Xu, Shiyu Li, Weifeng Ge, Afshin Dehghan, Meng Cao, Ping Huang
Abstract:
We present StreamBridge, a simple yet effective framework that seamlessly transforms offline Video-LLMs into streaming-capable models. It addresses two fundamental challenges in adapting existing models into online scenarios: (1) limited capability for multi-turn real-time understanding, and (2) lack of proactive response mechanisms. Specifically, StreamBridge incorporates (1) a memory buffer combined with a round-decayed compression strategy, supporting long-context multi-turn interactions, and (2) a decoupled, lightweight activation model that can be effortlessly integrated into existing Video-LLMs, enabling continuous proactive responses. To further support StreamBridge, we construct Stream-IT, a large-scale dataset tailored for streaming video understanding, featuring interleaved video-text sequences and diverse instruction formats. Extensive experiments show that StreamBridge significantly improves the streaming understanding capabilities of offline Video-LLMs across various tasks, outperforming even proprietary models such as GPT-4o and Gemini 1.5 Pro. Simultaneously, it achieves competitive or superior performance on standard video understanding benchmarks.



Paperid:5154
Authors:Kelvin Kan, Xingjian Li, Benjamin Zhang, Tuhin Sahai, Stanley Osher, Markos Katsoulakis
Abstract:
We study Transformers through the perspective of optimal control theory, using tools from continuous-time formulations to derive actionable insights into training and architecture design. This framework improves the performance of existing Transformer models while providing desirable theoretical guarantees, including generalization and robustness.Our framework is designed to be plug-and-play, enabling seamless integration with established Transformer models and requiring only slight changes to the implementation. We conduct seven extensive experiments on tasks motivated by text generation, sentiment analysis, image classification, and point cloud classification. Experimental results show that the framework improves the test performance of the baselines, while being more parameter-efficient. On character-level text generation with nanoGPT, our framework achieves a 46\% reduction in final test loss while using 42\% fewer parameters. On GPT-2, our framework achieves a 5.6\% reduction in final test loss, demonstrating scalability to larger models. To the best of our knowledge, this is the first work that applies optimal control theory to both the training and architecture of Transformers. It offers a new foundation for systematic, theory-driven improvements and moves beyond costly trial-and-error approaches.



Paperid:5155
Authors:Zihan Zheng, Zhenlong Wu, Houqiang Zhong, Yuan Tian, Ning Cao, Lan Xu, Jiangchao Yao, Xiaoyun Zhang, Qiang Hu, Wenjun Zhang
Abstract:
Achieving seamless viewing of high-fidelity volumetric video, comparable to 2D video experiences, remains an open challenge. Existing volumetric video compression methods either lack the flexibility to adjust quality and bitrate within a single model for efficient streaming across diverse networks and devices, or struggle with real-time decoding and rendering on lightweight mobile platforms. To address these challenges, we introduce 4DGCPro, a novel hierarchical 4D Gaussian compression framework that facilitates real-time mobile decoding and high-quality rendering via progressive volumetric video streaming in a single bitstream. Specifically, we propose a perceptually-weighted and compression-friendly hierarchical 4D Gaussian representation with motion-aware adaptive grouping to reduce temporal redundancy, preserve coherence, and enable scalable multi-level detail streaming. Furthermore, we present an end-to-end entropy-optimized training scheme, which incorporates layer-wise rate-distortion (RD) supervision and attribute-specific entropy modeling for efficient bitstream generation. Extensive experiments show that 4DGCPro enables flexible quality and variable bitrate within a single model, achieving real-time decoding and rendering on mobile devices while outperforming existing methods in RD performance across multiple datasets.



Paperid:5156
Authors:Lachlan MacDonald, Leandro Palma, Ziqing Xu, Hancheng Min, Salma Tarmoun, Rene Vidal
Abstract:
Abstract:Classical optimisation theory guarantees monotonic objective decrease for gradient descent (GD) when employed in a small step size, or "stable", regime. In contrast, gradient descent on neural networks is frequently performed in a large step size regime called the "edge of stability", in which the objective decreases non-monotonically with an observed implicit bias towards flat minima. In this paper, we take a step toward quantifying this phenomenon by providing convergence rates for gradient descent with large learning rates in an overparametrised least squares setting. The key insight behind our analysis is that, as a consequence of overparametrisation, the set of global minimisers forms a Riemannian manifold $M$, which enables the decomposition of the GD dynamics into components parallel and orthogonal to $M$. The parallel component corresponds to Riemannian gradient descent on the objective sharpness, while the orthogonal component corresponds to a quadratic dynamical system. This insight allows us to derive rates in three regimes characterised by the learning rate size: the subcritical regime, in which transient instability is overcome in finite time before linear convergence to a suboptimally flat global minimum; the critical regime, in which instability persists for all time with a power-law convergence toward the optimally flat global minimum; the supercritical regime, in which instability persists for all time with linear convergence to an oscillation of period two centred on the optimally flat global minimum.



Paperid:5157
Authors:Georgios Papoudakis, Thomas Coste, Jianye Hao, Jun Wang, Kun Shao
Abstract:
Reinforcement learning (RL) using foundation models for policy approximations in multi-turn tasks remains challenging. We identify two main limitations related to sparse reward settings and policy gradient updates, based on which we formulate a key insight: updates from positive samples with high returns typically do not require policy regularisation, whereas updates from negative samples, reflecting undesirable behaviour, can harm model performance. This paper introduces Succeed or Learn Slowly (SoLS), a novel off-policy RL algorithm evaluated on mobile app control tasks. SoLS improves sample efficiency when fine-tuning foundation models for user interface navigation via a modified off-policy actor-critic approach, applying direct policy updates for positive samples and conservative, regularised updates for negative ones to prevent model degradation. We augment SoLS with Successful Transition Replay (STR), which prioritises learning from successful interactions, further improving sample efficiency. We evaluate SoLS on the AndroidWorld benchmark, where it significantly outperforms existing methods (at least 17\% relative increase), including prompt-engineering and RL approaches, while requiring substantially fewer computational resources than GPT-4o-based methods with 5-60x faster inference.



Paperid:5158
Authors:Mengkang Hu, Yuhang Zhou, Wendong Fan, Yuzhou Nie, Ziyu Ye, Bowei Xia, Tao Sun, Zhaoxuan Jin, Yingru Li, Zeyu Zhang, Yifeng Wang, Qianshuo Ye, Bernard Ghanem, Ping Luo, Guohao Li
Abstract:
Large Language Model (LLM)-based multi-agent systems show promise for automating real-world tasks but struggle to transfer across domains due to their domain-specific nature.Current approaches face two critical shortcomings: they require complete architectural redesign and full retraining of all components when applied to new domains.We introduceWorkforce, a hierarchical multi-agent framework that decouples strategic planning from specialized execution through a modular architecture comprising:(i)adomain-agnosticPlannerfor task decomposition,(ii)aCoordinatorfor subtask management, and(iii)specializedWorkerswithdomain-specifictool-calling capabilities.This decoupling enables cross-domain transferability during both inference and training phases:During inference, Workforce seamlessly adapts to new domains by adding or modifying worker agents;For training, we introduceOptimized Workforce Learning (OWL), which improves generalization across domains by optimizing a domain-agnostic planner with reinforcement learning from real-world feedback.To validate our approach, we evaluate Workforce on the GAIA benchmark, covering various realistic, multi-domain agentic tasks.Experimental results demonstrate Workforce achieves open-source state-of-the-art performance (69.70%), outperforming commercial systems like OpenAI's Deep Research by2.34%.More notably, our OWL-trained 32B model achieves52.73%accuracy (+16.37%) and demonstrates performance comparable to GPT-4o on challenging tasks.To summarize, by enabling scalable generalization and modular domain transfer, our work establishes a foundation for the next generation of general-purpose AI assistants.Our code is available atAnonymous URL, and our data is available atAnonymous URL.



Paperid:5159
Authors:Shaohan Li, Yunpeng Shi, Gilad Lerman
Abstract:
We introduce Cycle-Sync, a robust and global framework for estimating camera poses (both rotations and locations). Our core innovation is a location solver that adapts message-passing least squares (MPLS) - originally developed for group synchronization - to the camera localization setting. We modify MPLS to emphasize cycle-consistent information, redefine cycle consistencies using estimated distances from previous iterations, and incorporate a Welsch-type robust loss. We establish the strongest known deterministic exact-recovery guarantee for camera location estimation, demonstrating that cycle consistency alone enables the lowest sample complexity to date. To further boost robustness, we introduce a plug-and-play outlier rejection module inspired by robust subspace recovery, and we fully integrate cycle consistency into MPLS for rotation averaging. Our global approach avoids the need for bundle adjustment. Experiments on synthetic and real datasets show that Cycle-Sync consistently outperforms leading pose estimators, including full structure-from-motion pipelines with bundle adjustment.



Paperid:5160
Authors:Fuyuan Xiao, Yu Zhou, Witold Pedrycz
Abstract:
In pattern classification, efficient uncertainty reasoning plays a critical role, particularly in real-time applications involving noisy data, ambiguous class boundaries, or overlapping categories. Leveraging the advanced computational power of quantum computing, an Adaptive Quantum Circuit for Dempster’s Rule of Combination (AQC-DRC) is proposed to address efficient classification under uncertain environments. The AQC-DRC is developed within the framework of quantum evidence theory (QET) and facilitates decision-making based on quantum basic probability and plausibility levels, which is a generalized Bayesian inference method. The AQC-DRC provides a deterministic computation of DRC, ensuring that quantum fusion outcomes in uncertain pattern classification are exactly aligned with those of the classical method, while simultaneously achieving exponential reductions in the computational complexity of evidence combination and significantly improving fusion efficiency. It is founded that the quantum basic probability amplitude function in QET, as a generalized quantum probability amplitude, can be naturally utilized to express the quantum amplitude encoding. In addition, the quantum basic probability in QET, as a generalized quantum probability, naturally forms a quantum basic probability distribution and can be used to represent quantum measurement outcomes for quantum basic probability level decision-making. Furthermore, the quantum plausibility function in QET also can be naturally used to express the quantum measurement outcomes for quantum plausibility level decision-making. These findings enrich the physical understanding of quantum amplitude encoding and quantum measurement outcomes, offering broad application prospects for representing and processing uncertain knowledge in pattern classification.



Paperid:5161
Authors:Fred Xu, Thomas Markovich
Abstract:
Graph Neural Networks (GNNs) have achieved impressive results across diverse network modeling tasks, but accurately estimating uncertainty on graphs remains difficult—especially under distributional shifts. Unlike traditional uncertainty estimation, graph-based uncertainty must account for randomness arising from both the graph’s structure and its label distribution, which adds complexity. In this paper, making an analogy between the evolution of a stochastic partial differential equation (SPDE) driven by Mat\'ern Gaussian Process and message passing using GNN layers, we present a principled way to design a novel message passing scheme that incorporates spatial-temporal noises motivated by the Gaussian Process approach to SPDE. Our method simultaneously captures uncertainty across space and time and allows explicit control over the covariance kernel’s smoothness, thereby enhancing uncertainty estimates on graphs with both low and high label informativeness. Our extensive experiments on Out-of-Distribution (OOD) detection on graph datasets with varying label informativeness demonstrate the soundness and superiority of our model to existing approaches.



Paperid:5162
Authors:Guilin Li, Yun Zhang, Xiuyuan Chen, Chengqi Li, Weiran Huang, Linghe Kong, Wenjia Wang, Bo Wang, Matthias Tan
Abstract:
Large-scale payment platforms face fraud detection challenges due to high-cardinality transactions, strict latency, and limited labeled data. As a result, traditional models relying on extensive handcrafted features struggle with generalization. We propose PANTHER, a hybrid framework combining self-supervised generative pretraining with lightweight discriminative modeling for real-time fraud detection. Leveraging abundant unlabeled data, PANTHER learns compact user behavior representations via a predict-next-behavior task, capturing sparse interactions and transactional dynamics. PANTHER addresses four key challenges: (1) Structured Tokenization to compress multi-dimensional transaction attributes into an interpretable vocabulary; (2) Sequence Pattern Recognition Module (SPRM) for modeling periodic transaction motifs; (3) a Unified User-Profile Embedding that fuses static demographics with dynamic transaction histories, enabling both personalized predictions and population-level knowledge transfer; and (4) Real-time scalability enabled by offline caching of pre-trained embeddings for millisecond-level inference. Fully deployed and operational online at WeChat Pay, PANTHER delivers a 38.6\% relative improvement in fraud detection recall and a 25.6\% boost in next-transaction prediction HR@1, significantly outperforming transformer baselines. Cross-domain evaluations on public benchmarks (CCT, MBD, MovieLens-1M, Yelp) show strong generalization, achieving up to 21\% HR@1 gains over transformer baselines, establishing PANTHER as a scalable, high-performance framework for industrial sequential behavior modeling.



Paperid:5163
Authors:Chaeyoung Jung, Youngjoon Jang, Joon Son Chung
Abstract:
Hallucination remains a major challenge in multimodal large language models (MLLMs). To address this, various contrastive decoding (CD) methods have been proposed that contrasts original logits with hallucinated logits generated from perturbed inputs. While CD has shown promise in vision-language models (VLMs), it is not well-suited for AV-LLMs, where hallucinations often emerge from both unimodal and cross-modal combinations involving audio, video, and language. These intricate interactions call for a more adaptive and modality-aware decoding strategy. In this paper, we propose Audio-Visual Contrastive Decoding (AVCD)—a novel, training-free decoding framework designed to model trimodal interactions and suppress modality-induced hallucinations in AV-LLMs. Unlike previous CD methods in VLMs that corrupt a fixed modality, AVCD leverages attention distributions to dynamically identify less dominant modalities and applies attentive masking to generate perturbed output logits. To support CD in a trimodal setting, we also reformulate the original CD framework to jointly handle audio, visual, and textual inputs. Finally, to improve efficiency, we introduce entropy-guided adaptive decoding, which selectively skips unnecessary decoding steps based on the model’s confidence in its predictions. Extensive experiments demonstrate that AVCD consistently outperforms existing decoding methods. Especially, on the AVHBench dataset, it improves accuracy by 6% for VideoLLaMA2 and 11% for Video-SALMONN, demonstrating strong robustness and generalizability.



Paperid:5164
Authors:Delong Liu, Haiwen Li, Zhaohui Hou, Zhicheng Zhao, Fei Su, Yuan Dong
Abstract:
Person retrieval has attracted rising attention. Existing methods are mainly divided into two retrieval modes, namely image-only and text-only. However, they are unable to make full use of the available information and are difficult to meet diverse application requirements. To address the above limitations, we propose a new Composed Person Retrieval (CPR) task, which combines visual and textual queries to identify individuals of interest from large-scale person image databases. Nevertheless, the foremost difficulty of the CPR task is the lack of available annotated datasets. Therefore, we first introduce a scalable automatic data synthesis pipeline, which decomposes complex multimodal data generation into the creation of textual quadruples followed by identity-consistent image synthesis using fine-tuned generative models. Meanwhile, a multimodal filtering method is designed to ensure the resulting SynCPR dataset retains 1.15 million high-quality and fully synthetic triplets. Additionally, to improve the representation of composed person queries, we propose a novel Fine-grained Adaptive Feature Alignment (FAFA) framework through fine-grained dynamic alignment and masked feature reasoning. Moreover, for objective evaluation, we manually annotate the Image-Text Composed Person Retrieval (ITCPR) test set. The extensive experiments demonstrate the effectiveness of the SynCPR dataset and the superiority of the proposed FAFA framework when compared with the state-of-the-art methods. All code will be open-sourced.



Paperid:5165
Authors:Haowei Zhu, Tianxiang Pan, Rui Qin, Jun-Hai Yong, Bin Wang
Abstract:
The scale and quality of datasets are crucial for training robust perception models. However, obtaining large-scale annotated data is both costly and time-consuming. Generative models have emerged as a powerful tool for data augmentation by synthesizing samples that adhere to desired distributions. However, current generative approaches often rely on complex post-processing or extensive fine-tuning on massive datasets to achieve satisfactory results, and they remain prone to content–position mismatches and semantic leakage. To overcome these limitations, we introduce ReCon, a novel augmentation framework that enhances the capacity of structure-controllable generative models for object detection. ReCon integrates region-guided rectification into the diffusion sampling process, using feedback from a pre-trained perception model to rectify misgenerated regions within diffusion sampling process. We further propose region-aligned cross-attention to enforce spatial–semantic alignment between image regions and their textual cues, thereby improving both semantic consistency and overall image fidelity. Extensive experiments demonstrate that ReCon substantially improve the quality and trainability of generated data, achieving consistent performance gains across various datasets, backbone architectures, and data scales.



Paperid:5166
Authors:Adam Piaseczny, Md Kamran Chowdhury Shisher, Shiqiang Wang, Christopher Brinton
Abstract:
Machine learning (ML) algorithms deployed in real-world environments are often faced with the challenge of adapting models to concept drift, where the task data distributions are shifting over time. The problem becomes even more difficult when model performance must be maintained under adherence to strict resource constraints. Existing solutions often depend on drift-detection methods that produce high computational overhead for resource-constrained environments, and fail to provide strict guarantees on resource usage or theoretical performance assurances. To address these shortcomings, we propose RCCDA: a dynamic model update policy that optimizes ML training dynamics while ensuring strict compliance to predefined resource constraints, utilizing only past loss information and a tunable drift threshold. In developing our policy, we analytically characterize the evolution of model loss under concept drift with arbitrary training update decisions. Integrating these results into a Lyapunov drift-plus-penalty framework produces a lightweight policy based on a measurable accumulated loss threshold that provably limits update frequency and cost. Experimental results on three domain generalization datasets demonstrate that our policy outperforms baseline methods in inference accuracy while adhering to strict resource constraints under several schedules of concept drift, making our solution uniquely suited for real-time ML deployments.



Paperid:5167
Authors:Ruizhe Liu, Pei Zhou, Qian Luo, Li Sun, Jun CEN, Yibing Song, Yanchao Yang
Abstract:
Effective generalization in robotic manipulation requires representations that capture invariant patterns of interaction across environments and tasks.We present a self-supervised framework for learning hierarchical manipulation concepts that encode these invariant patterns through cross-modal sensory correlations and multi-level temporal abstractions without requiring human annotation.Our approach combines a cross-modal correlation network that identifies persistent patterns across sensory modalities with a multi-horizon predictor that organizes representations hierarchically across temporal scales. Manipulation concepts learned through this dual structure enable policies to focus on transferable relational patterns while maintaining awareness of both immediate actions and longer-term goals.Empirical evaluation across simulated benchmarks and real-world deployments demonstrates significant performance improvements with our concept-enhanced policies. Analysis reveals that the learned concepts resemble human-interpretable manipulation primitives despite receiving no semantic supervision. This work advances both the understanding of representation learning for manipulation and provides a practical approach to enhancing robotic performance in complex scenarios.



Paperid:5168
Authors:Liwei Huang, Zhengyu Ma, Liutao Yu, Huihui Zhou, Yonghong Tian
Abstract:
Seeking high-quality representations with latent variable models (LVMs) to reveal the intrinsic correlation between neural activity and behavior or sensory stimuli has attracted much interest. In the study of the biological visual system, naturalistic visual stimuli are inherently high-dimensional and time-dependent, leading to intricate dynamics within visual neural activity. However, most work on LVMs has not explicitly considered neural temporal relationships. To cope with such conditions, we propose Time-Evolving Visual Dynamical System (TE-ViDS), a sequential LVM that decomposes neural activity into low-dimensional latent representations that evolve over time. To better align the model with the characteristics of visual neural activity, we split latent representations into two parts and apply contrastive learning to shape them. Extensive experiments on synthetic datasets and real neural datasets from the mouse visual cortex demonstrate that TE-ViDS achieves the best decoding performance on naturalistic scenes/movies, extracts interpretable latent trajectories that uncover clear underlying neural dynamics, and provides new insights into differences in visual information processing between subjects and between cortical regions. In summary, TE-ViDS is markedly competent in extracting stimulus-relevant embeddings from visual neural activity and contributes to the understanding of visual processing mechanisms.



Paperid:5169
Authors:Xinan He, Yue Zhou, Bing Fan, Bin Li, Guopu Zhu, Feng Ding
Abstract:
Faces synthesized by diffusion models (DMs) with high-quality and controllable attributes pose a significant challenge for Deepfake detection. Most state-of-the-art detectors only yield a binary decision, incapable of forgery localization, attribution of forgery methods, and providing analysis on the cause of forgeries. In this work, we integrate Multimodal Large Language Models (MLLMs) within DM-based face forensics, and propose a fine-grained analysis triad framework called VLForgery,that can 1) predict falsified facial images;2) locate the falsified face regions subjected to partial synthesis; and 3) attribute the synthesis with specific generators. To achieve the above goals, we introduce VLF (Visual Language Forensics), a novel and diverse synthesis face dataset designed to facilitate rich interactions betweenVisual' andLanguage' modalities in MLLMs.Additionally, we propose an extrinsic knowledge-guided description method, termed EkCot, which leverages knowledge from the image generation pipeline to enable MLLMs to quickly capture image content. Furthermore, we introduce a low-level vision comparison pipeline designed to identify differential features between real and fake that MLLMs can inherently understand. These features are then incorporated into EkCot, enhancing its ability to analyze forgeries in a structured manner, following the sequence of detection, localization, and attribution.Extensive experiments demonstrate that VLForgery outperforms other state-of-the-art forensic approaches in detection accuracy, with additional potential for falsified region localization and attribution analysis.



Paperid:5170
Authors:Jingcheng Hu, Yinmin Zhang, Qi Han, Daxin Jiang, Xiangyu Zhang, Heung-Yeung Shum
Abstract:
Abstract:We introduce Open-Reasoner-Zero, the first open source implementation of large-scale reasoning-oriented RL training on the base model focusing on scalability, simplicity and accessibility.Through extensive experiments, we demonstrate that a minimalist approach, vanilla PPO with GAE ($\lambda=1$, $\gamma=1$) and straightforward rule-based rewards, without any KL regularization, is sufficient to scale up both benchmark performance and response length, replicating the scaling phenomenon observed in DeepSeek-R1-Zero.Using the same base model as DeepSeek-R1-Zero-Qwen-32B, our implementation achieves superior performance across AIME2024, MATH500, and GPQA Diamond, while demonstrating remarkable efficiency—requiring only 1/10 of the training steps compared to the DeepSeek-R1-Zero pipeline.Moreover, our analysis not only covers training dynamics and ablation for critical design choices, but also quantitatively show how the learned critic in Reasoner-Zero training effectively identifies and devalues repetitive response patterns, yielding more robust advantage estimations and enhancing training stability. Embracing the principles of open-source, we release our source code, parameter settings, training data, and model weights across various sizes, fostering reproducibility and encouraging further exploration of the properties of related models.



Paperid:5171
Authors:Dominik Meier, Sujai Hiremath, PROMIT GHOSAL, Kyra Gan
Abstract:
Distinguishing cause and effect from bivariate observational data is a foundational problem in many disciplines, but challenging without additional assumptions. Additive noise models (ANMs) are widely used to enable sample-efficient bivariate causal discovery.However, conventional ANM-based methods fail when unobserved mediators corrupt the causal relationship between variables. This paper makes three key contributions: first, we rigorously characterize why standard ANM approachesbreak down in the presence of unmeasured mediators. Second, we demonstrate that prior solutions for hidden mediation are brittle in finite sample settings,limiting their practical utility. To address these gaps, we propose Bivariate Denoising Diffusion (BiDD) for causal discovery, a method designed to handle latent noise introduced by unmeasured mediators. Unlike prior methods that infer directionality through mean squared error loss comparisons, our approach introduces a novel independence test statistic: during the noising and denoising processes for each variable, we condition on the other variable as input and evaluate the independence of the predicted noise relative to this input. We prove asymptotic consistency of BiDD under the ANM, and conjecture that it performs well under hidden mediation. Experiments on synthetic and real-world data demonstrate consistent performance, outperforming existing methods in mediator-corrupted settings while maintaining strong performance in mediator-free settings.



Paperid:5172
Authors:Cong Wang, Zexuan Deng, Zhiwei Jiang, Fei Shen, Yafeng Yin, Shiwei Gan, Zifeng Cheng, Shiping Ge, Qing Gu
Abstract:
Sign Language Video Generation (SLVG) seeks to generate identity-preserving sign language videos from spoken language texts. Existing methods primarily rely on the single coarse condition (e.g., skeleton sequences) as the intermediary to bridge the translation model and the video generation model, which limits both the naturalness and expressiveness of the generated videos. To overcome these limitations, we propose SignViP, a novel SLVG framework that incorporate multiple fine-grained conditions for improved generation fidelity. Rather than directly translating error-prone high-dimensional conditions, SignViP adopts a discrete tokenization paradigm to integrate and represent fine-grained conditions (i.e., fine-grained poses and 3D hands). SignViP contains three core components. (1) Sign Video Diffusion Model is jointly trained with a multi-condition encoder to learn continuous embeddings that encapsulate fine-grained motion and appearance. (2) Finite Scalar Quantization (FSQ) Autoencoder is further trained to compress and quantize these embeddings into discrete tokens for compact representation of the conditions. (3) Multi-Condition Token Translator is trained to translate spoken language text to discrete multi-condition tokens. During inference, Multi-Condition Token Translator first translates the spoken language text into discrete multi-condition tokens. These tokens are then decoded to continuous embeddings by FSQ Autoencoder, which are subsequently injected into Sign Video Diffusion Model to guide video generation. Experimental results show that SignViP achieves state-of-the-art performance across metrics, including video quality, temporal coherence, and semantic fidelity.



Paperid:5173
Authors:Dario Fenoglio, Mohan Li, Pietro Barbiero, Nicholas Lane, Marc Langheinrich, Martin Gjoreski
Abstract:
Federated Learning (FL) enables collaborative model training across multiple clients while preserving data privacy. Traditional FL methods often use a global model to fit all clients, assuming that clients' data are independent and identically distributed (IID). However, when this assumption does not hold, the global model accuracy may drop significantly, limiting FL applicability in real-world scenarios. To address this gap, we propose FLUX, a novel clustering-based FL (CFL) framework that addresses the four most common types of distribution shifts during both training and test time. To this end, FLUX leverages privacy-preserving client-side descriptor extraction and unsupervised clustering to ensure robust performance and scalability across varying levels and types of distribution shifts. Unlike existing CFL methods addressing non-IID client distribution shifts, FLUX i) does not require any prior knowledge of the types of distribution shifts or the number of client clusters, and ii) supports test-time adaptation, enabling unseen and unlabeled clients to benefit from the most suitable cluster-specific models. Extensive experiments across four benchmarks, one real-world dataset and ten state-of-the-art baselines show that FLUX improves performance and stability under diverse distribution shifts—achieving an average accuracy gain of up to 8 percentage points over the best-performing baselines—while maintaining computational and communication overhead comparable to FedAvg.



Paperid:5174
Authors:hanzhuo tan, Xiaolong Tian, Hanrui Qi, Jiaming Liu, Siyi Wang, GAO Zuchen, Qi Luo, Jing Li, Yuqun Zhang
Abstract:
Recent advances in LLM-based decompilers have been shown effective to convert low-level binaries into human-readable source code. However, there still lacks a comprehensive benchmark that provides large-scale binary-source function pairs, which is critical for advancing the LLM decompilation technology. Creating accurate binary-source mappings incurs severe issues caused by complex compilation settings and widespread function inlining that obscure the correspondence between binaries and their original source code. Previous efforts have either relied on used contest‐style benchmarks, synthetic binary–source mappings that diverge significantly from the mappings in real world, or partially matched binaries with only code lines or variable names, compromising the effectiveness of analyzing the binary functionality. To alleviate these issues, we introduce Decompile-Bench, the first open-source dataset comprising two million binary-source function pairs condensed from 100 million collected function pairs, i.e., 450GB of binaries compiled from permissively licensed GitHub projects. For the evaluation purposes, we also developed a benchmark Decompile-Bench-Eval including manually crafted binaries from the well-established HumanEval and MBPP, alongside the compiled GitHub repositories released after 2025 to mitigate data leakage issues. We further explore commonly-used evaluation metrics to provide a thorough assessment of the studied LLM decompilers and find that fine-tuning with Decompile-Bench causes a 20% improvement over previous benchmarks in terms of the re-executability rate. Our code and data has been released in HuggingFace and Github. https://github.com/anonepo/LLM4Decompile



Paperid:5175
Authors:Ruyu Liu, Lin Wang, Zhou Mingming, Jianhua Zhang, ZHANG HAOYU, Xiufeng Liu, Xu Cheng, Sixian Chan, Shen yanbin, Dai Sheng, Yuping Yan, Yaochu Jin, Lingjuan Lyu
Abstract:
Accurate polyp sizing during endoscopy is crucial for cancer risk assessment but is hindered by subjective methods and inadequate datasets lacking integrated 2D appearance, 3D structure, and real-world size information. We introduce PolypSense3D, the first multi-source benchmark dataset specifically targeting depth-aware polyp size measurement. It uniquely integrates over 43,000 frames from virtual simulations, physical phantoms, and clinical sequences, providing synchronized RGB, dense/sparse depth, segmentation masks, camera parameters, and millimeter-scale size labels derived via a novel forceps-assisted in-vivo annotation technique. To establish its value, we benchmark state-of-the-art segmentation and depth estimation models. Results quantify significant domain gaps between simulated/phantom and clinical data and reveal substantial error propagation from perception stages to final size estimation, with the best fully automated pipelines achieving an average Mean Absolute Error (MAE) of 1.19 mm on the clinical data subset. Publicly released under CC BY-SA 4.0 with code and evaluation protocols, PolypSense3D offers a standardized platform to accelerate research in robust, clinically relevant quantitative endoscopic vision. The benchmark dataset and code are available at: https://github.com/HNUicda/PolypSense3D and https://doi.org/10.7910/DVN/LKDIEK.



Paperid:5176
Authors:Qiang Fu, Andre Wibisono
Abstract:
Abstract:We study the Hamiltonian flow for optimization (HF-opt), which simulates the Hamiltonian dynamics for some integration time and resets the velocity to $0$ to decrease the objective function; this is the optimization analogue of the Hamiltonian Monte Carlo algorithm for sampling. For short integration time, HF-opt has the same convergence rates as gradient descent for minimizing strongly and weakly convex functions. We show that by randomizing the integration time in HF-opt, the resulting randomized Hamiltonian flow (RHF) achieves accelerated convergence rates in continuous time, similar to the rates for accelerated gradient flow. We study a discrete-time implementation of RHF as the randomized Hamiltonian gradient descent (RHGD) algorithm. We prove that RHGD achieves the same accelerated convergence rates as Nesterov's accelerated gradient descent (AGD) for minimizing smooth strongly and weakly convex functions. We provide numerical experiments to demonstrate that RHGD is competitive with classical accelerated methods such as AGD across all settings and outperforms them in certain regimes.



Paperid:5177
Authors:Qiankun Li, Feng He, Huabao Chen, Xin Ning, Kun Wang, Zengfu Wang
Abstract:
In the big data era, the computer vision field benefits from large-scale datasets such as LAION-2B, LAION-400M, and ImageNet-21K, Kinetics, on which popular models like the ViT and ConvNeXt series have been pre-trained, acquiring substantial knowledge. However, numerous downstream tasks in specialized and data-limited scientific domains continue to pose significant challenges. In this paper, we propose a novel Cluster Attention Adapter (CLAdapter), which refines and adapts the rich representations learned from large-scale data to various data-limited downstream tasks. Specifically, CLAdapter introduces attention mechanisms and cluster centers to personalize the enhancement of transformed features through distribution correlation and transformation matrices. This enables models fine-tuned with CLAdapter to learn distinct representations tailored to different feature sets, facilitating the models' adaptation from rich pre-trained features to various downstream scenarios effectively. In addition, CLAdapter's unified interface design allows for seamless integration with multiple model architectures, including CNNs and Transformers, in both 2D and 3D contexts.Through extensive experiments on 10 datasets spanning domains such as generic, multimedia, biological, medical, industrial, agricultural, environmental, geographical, materials science, out-of-distribution (OOD), and 3D analysis, CLAdapter achieves state-of-the-art performance across diverse data-limited scientific domains, demonstrating its effectiveness in unleashing the potential of foundation vision models via adaptive transfer.Code is available at https://anonymous.4open.science/r/CLAdapter-NIPS2025.



Paperid:5178
Authors:Jianyang Qin, Chaoyang Li, Jinhao Cui, Lingzhi Wang, Zhao Liu, Qing Liao
Abstract:
Abstract:Recent studies reveal that Large Language Models (LLMs) exhibit strong sequential reasoning capabilities, allowing them to replace specialized time-series models and serve as foundation models for complex time-series analysis. To activate the capabilities of LLMs for time-series tasks, numerous studies have attempted to bridge the gap between time series and linguistics by aligning textual representations with time-series patterns. However, it is a non-trivial endeavor to losslessly capture the infinite time-domain variability using natural language, leading to suboptimal alignment performance. Beyond representation, contextual differences, where semantics in time series are conveyed by consecutive points, unlike in text by individual tokens, are often overlooked by existing methods. To address these, we propose S$^2$TS-LLM, a simple yet effective framework to repurpose LLMs for universal time series analysis through the following two main paradigms: (i) a spectral symbolization paradigm transforms time series into frequency-domain representations characterized by a fixed number of components and prominent amplitudes, which enables a limited set of symbols to effectively abstract key frequency features; (ii) a contextual segmentation paradigm partitions the sequence into blocks based on temporal patterns and reassigns positional encodings accordingly, thereby mitigating the structural mismatch between time series and natural language. Together, these paradigms bootstrap the LLMs' perception of temporal patterns and structures, effectively bridging time series and linguistics. Extensive experiments show that S$^2$TS-LLM can serve as a powerful time series analyzer, outperforming state-of-the-art methods across time series tasks.



Paperid:5179
Authors:Andy Zhang, Joey Ji, Celeste Menders, Riya Dulepet, Thomas Qin, Ron Wang, Junrong Wu, Kyleen Liao, Jiliang Li, Jinghan Hu, Sara Hong, Nardos Demilew, Shivatmica Murgai, Jason Tran, Nishka Kacheria, Ethan Ho, Denis Liu, Lauren McLane, Olivia Bruvik, Dai-Rong Han, Seungwoo Kim, Akhil Vyas, Cuiyuanxiu Chen, Ryan Li, Weiran Xu, Jonathan Ye, Prerit Choudhary, Siddharth M. Bhatia, Vikram Sivashankar, Yuxuan Bao, Dawn Song, Dan Boneh, Daniel Ho, Percy Liang
Abstract:
Abstract:AI agents have the potential to significantly alter the cybersecurity landscape. To help us understand this change, we introduce the first framework to capture offensive and defensive cyber-capabilities in evolving real-world systems. Instantiating this framework with BountyBench, we set up 25 systems with complex, real-world codebases. To capture the vulnerability lifecycle, we define three task types: Detect (detecting a new vulnerability), Exploit (exploiting a given vulnerability), and Patch (patching a given vulnerability). For Detect, we construct a new success indicator, which is general across vulnerability types and provides localized evaluation. We manually set up the environment for each system, including installing packages, setting up server(s), and hydrating database(s). We add 40 bug bounties, which are vulnerabilities with monetary awards from \\$10 to \\$30,485, and cover 9 of the OWASP Top 10 Risks. To modulate task difficulty, we devise a new strategy based on information to guide detection, interpolating from identifying a zero day to exploiting a given vulnerability. We evaluate 5 agents: Claude Code, OpenAI Codex, and custom agents with GPT-4.1, Gemini 2.5 Pro Preview, and Claude 3.7 Sonnet Thinking. The top-performing agents are Claude Code (2.5% on Detect, corresponding to \\$450), Custom Agent with Claude 3.7 Sonnet Thinking (55% on Exploit), and OpenAI Codex (80% on Patch, corresponding to \\$13,710). The custom agents achieve higher Exploit scores of 35-55% compared to Patch scores of 30-35%; in contrast, OpenAI Codex and Claude Code achieve higher Patch scores of 80% and 57.5%, compared to Exploit scores of 25% and 32.5% respectively.



Paperid:5180
Authors:Dennis Frauen, Maresa Schröder, Konstantin Hess, Stefan Feuerriegel
Abstract:
Estimating heterogeneous treatment effects (HTEs) is crucial for personalized decision-making. However, this task is challenging in \emph{survival analysis}, which includes time-to-event data with censored outcomes (e.g., due to study dropout). In this paper, we propose a toolbox of novel orthogonal survival learners to estimate HTEs from time-to-event data under censoring. Our learners have three main advantages: (i) we show that learners from our toolbox are guaranteed to be orthogonal and thus come with favorable theoretical properties; (ii) our toolbox allows for incorporating a custom weighting function, which can lead to robustness against different types of low overlap, and (iii) our learners are model-agnostic (i.e., they can be combined with arbitrary machine learning models). We instantiate the learners from our toolbox using several weighting functions and, as a result, propose various neural orthogonal survival learners. Some of these coincide with existing survival learners (including survival versions of the DR- and R-learner), while others are novel and are robust w.r.t. low overlap regimes specific to the survival setting (i.e., survival overlap and censoring overlap). We then empirically verify the effectiveness of our learners for HTE estimation in different low-overlap regimes through numerical experiments. In sum, we provide practitioners with a large toolbox of learners that can be used for randomized and observational studies with censored time-to-event data.



Paperid:5181
Authors:MingMing Yu, Fei Zhu, Wenzhuo Liu, Yirong Yang, Qunbo Wang, wenjun wu, Jing Liu
Abstract:
Embodied agents are expected to perform object navigation in dynamic, open-world environments. However, existing approaches typically rely on static trajectories and a fixed set of object categories during training, overlooking the real-world requirement for continual adaptation to evolving scenarios. To facilitate related studies, we introduce the continual object navigation benchmark, which requires agents to acquire navigation skills for new object categories while avoiding catastrophic forgetting of previously learned knowledge. To tackle this challenge, we propose C-Nav, a continual visual navigation framework that integrates two key innovations: (1) A dual-path anti-forgetting mechanism, which comprises feature distillation that aligns multi-modal inputs into a consistent representation space to ensure representation consistency, and feature replay that retains temporal features within the action decoder to ensure policy consistency. (2) An adaptive sampling strategy that selects diverse and informative experiences, thereby reducing redundancy and minimizing memory overhead. Extensive experiments across multiple model architectures demonstrate that C-Nav consistently outperforms existing approaches, achieving superior performance even compared to baselines with full trajectory retention, while significantly lowering memory requirements. The benchmark and code will be publicly available.



Paperid:5182
Authors:Trung Le, Hao Fang, Jingyuan Li, Tung Nguyen, Lu Mi, Amy Orsborn, Uygar Sümbül, Eli Shlizerman
Abstract:
Intracortical Brain-Computer Interfaces (iBCI) aim to decode behavior from neural population activity, enabling individuals with motor impairments to restore motor functions and communication abilities. A central challenge in the long-term deployment of iBCIs is the nonstationarity of neural recordings, where instability of electrode recordings alters the composition and tuning of the recorded neural population across sessions. Existing approaches attempt to address this issue by explicit alignment techniques; however, they rely on fixed neural identities and require test-time labels and parameter updates, limiting their ability to generalize across sessions and imposing a computational burden during deployment. In this work, we introduce SPINT - a Spatial Permutation-Invariant Neural Transformer framework for behavioral decoding that operates directly on unordered sets of neural units. Central to our approach is a novel context-dependent positional embedding scheme that infers unit-specific identities dynamically, enabling flexible generalization across recording sessions. Our model supports inference on variable-size neural populations and allows few-shot, gradient-free adaptation using a small amount of unlabeled data from the new session. To further promote robustness to population variability, we introduce dynamic channel dropout, a regularization method for iBCI applications by simulating shifts in population composition during training. We evaluate our approach on three motor decoding tasks from the FALCON Benchmark, comprising multi-session datasets from human and non-human primates. Our approach demonstrates robust cross-session generalization, outperforming existing zero-shot and few-shot unsupervised baselines while eliminating the need for test-time alignment and fine-tuning. Our work contributes an initial step toward a flexible and practical framework for robust, scalable neural decoding in long-term iBCI applications.



Paperid:5183
Authors:Zhenglin Lai, Mengyao Liao, Dong Xu, Zebin Zhao, Zhihang Yuan, Chao Fan, Jianqiang Li, Bingzhe Wu
Abstract:
Abstract:Large language models based on Mixture-of-Experts have achieved substantial gains in efficiency and scalability, yet their architectural uniqueness introduces underexplored safety alignment challenges. Existing safety alignment strategies, predominantly designed for dense models, are ill-suited to address MoE-specific vulnerabilities. In this work, we formalize and systematically study MoE model's positional vulnerability—the phenomenon where safety-aligned behaviors rely on specific expert modules, revealing critical risks inherent to MoE architectures. To this end, we present SAFEx, an analytical framework that robustly identifies, characterizes, and validates the safety-critical experts using a novel Stability-based Expert Selection (SES) algorithm. Notably, our approach enables the explicit decomposition of safety-critical experts into distinct functional groups, including those responsible for harmful content detection and those controlling safe response generation. Extensive experiments on mainstream MoE models, such as recently released Qwen3-MoE demonstrated that their intrinsic safety mechanisms heavily rely on a small subset of positional experts. Disabling these experts significantly compromised the models’ ability to refuse harmful requests. For Qwen3-MoE with 6144 experts (in FNN layer), we find that disabling as few as 12 identified safety-critical experts can cause the refusal rate to drop by $22\%$, demonstrating the disproportionate impact of a small set of experts on overall model safety.



Paperid:5184
Authors:Andreas Schlaginhaufen, Reda Ouhamma, Maryam Kamgarpour
Abstract:
We study reinforcement learning from human feedback in general Markov decision processes, where agents learn from trajectory-level preference comparisons. A central challenge in this setting is to design algorithms that select informative preference queries to identify the underlying reward while ensuring theoretical guarantees. We propose a meta-algorithm based on randomized exploration, which avoids the computational challenges associated with optimistic approaches and remains tractable. We establish both regret and last-iterate guarantees under mild reinforcement learning oracle assumptions. To improve query complexity, we introduce and analyze an improved algorithm that collects batches of trajectory pairs and applies optimal experimental design to select informative comparison queries. The batch structure also enables parallelization of preference queries, which is relevant in practical deployment as feedback can be gathered concurrently. Empirical evaluation confirms that the proposed method is competitive with reward-based reinforcement learning while requiring a small number of preference queries.



Paperid:5185
Authors:Divyanshu Mishra, Mohammadreza Salehi, Pramit Saha, Olga Patey, Aris Papageorghiou, Yuki Asano, Alison Noble
Abstract:
Self-supervised learning (SSL) has achieved major advances in natural images and video understanding, but challenges remain in domains likeechocardiography (heart ultrasound) due to subtle anatomical structures, complex temporal dynamics, and the current lack of domain-specific pre-trained models. Existing SSL approaches such as contrastive, masked modeling, and clustering-based methods struggle with high intersample similarity, sensitivity to low PSNR inputs common in ultrasound, or aggressive augmentations that distort clinically relevant features.We present DISCOVR (Distilled Image Supervision for Cross Modal Video Representation), a self-supervised dual branch framework for cardiac ultrasound video representation learning. DISCOVR combines a clustering-based video encoder that models temporal dynamics with an online image encoder that extracts fine-grained spatial semantics. These branches are connected through a semantic cluster distillation loss that transfers anatomical knowledge from the evolving image encoder to the video encoder, enabling temporally coherent representations enriched with fine-grained semantic understanding. Evaluated on six echocardiography datasets spanning fetal, pediatric, and adult populations, DISCOVR outperforms both specialized video anomaly detection methods and state-of-the-art video-SSL baselines in zero-shot and linear probing setups, and achieves superior segmentation transfer.



Paperid:5186
Authors:Xiner Li, Yulai Zhao, Chenyu Wang, Gabriele Scalia, Gokcen Eraslan, Surag Nair, Tommaso Biancalani, Shuiwang Ji, Aviv Regev, Sergey Levine, Masatoshi Uehara
Abstract:
Diffusion models excel at capturing the natural design spaces of images, molecules, DNA, RNA, and protein sequences. However, rather than merely generating designs that are natural, we often aim to optimize downstream reward functions while preserving the naturalness of these design spaces. Existing methods for achieving this goal often require differentiable proxy models (e.g., classifier guidance or DPS) or involve computationally expensive fine-tuning of diffusion models (e.g., classifier-free guidance, RL-based fine-tuning). In our work, we propose a new method to address these challenges. Our algorithm is an iterative sampling method that integrates soft value functions, which looks ahead to how intermediate noisy states lead to high rewards in the future, into the standard inference procedure of pre-trained diffusion models. Notably, our approach avoids fine-tuning generative models and eliminates the need to construct differentiable models. This enables us to (1) directly utilize non-differentiable features/reward feedback, commonly used in many scientific domains, and (2) apply our method to recent discrete diffusion models in a principled way. Finally, we demonstrate the effectiveness of our algorithm across several domains, including image generation, molecule generation, and DNA/RNA sequence generation.



Paperid:5187
Authors:Mohammad Shahab Sepehri, Berk Tinaz, Zalan Fabian, Mahdi Soltanolkotabi
Abstract:
Mental visualization, the ability to construct and manipulate visual representations internally, is a core component of human cognition and plays a vital role in tasks involving reasoning, prediction, and abstraction. Despite the rapid progress of Multimodal Large Language Models (MLLMs), current benchmarks primarily assess passive visual perception, offering limited insight into the more active capability of internally constructing visual patterns to support problem solving. Yet mental visualization is a critical cognitive skill in humans, supporting abilities such as spatial navigation, predicting physical trajectories, and solving complex visual problems through imaginative simulation. To bridge this gap, we introduce Hyperphantasia, a synthetic benchmark designed to evaluate the mental visualization abilities of MLLMs through four carefully constructed puzzles. Each task is procedurally generated and presented at three difficulty levels, enabling controlled analysis of model performance across increasing complexity. Our comprehensive evaluation of state-of-the-art models reveals a substantial gap between the performance of humans and MLLMs. Additionally, we explore the potential of reinforcement learning to improve visual simulation capabilities. Our findings suggest that while some models exhibit partial competence in recognizing visual patterns, robust mental visualization remains an open challenge for current MLLMs.



Paperid:5188
Authors:Zichen Wang, Zhiheng Zhang, Haoyang Hong, Haoxuan Li, Chuanhao Li, Huazheng Wang
Abstract:
In multi-armed bandits with network interference (MAB-NI), the action taken by one node can influence the rewards of others, creating complex interdependence. While current research primarily focuses on minimizing regret in MAB-NI, it often neglects the critical issue that overly prioritizing the optimal arm can compromise the inference accuracy of the sub-optimal arm. Although initial efforts have been made to address this trade-off in single-unit scenarios, these challenges have become more pronounced in the context of MAB-NI. In this paper, we establish, for the first time, a theoretical Pareto frontier characterizing the trade-off between regret minimization and inference accuracy in adversarial (design-based) MAB-NI. We further introduce an anytime-valid asymptotic confidence sequence along with a corresponding algorithm, EXP3-N-CS, specifically designed to balance the trade-off between regret minimization and inference accuracy in this setting.



Paperid:5189
Authors:Christodoulos Constantinides, Dhaval Patel, Shuxin Lin, Claudio Guerrero, Sunil Patil, Jayant Kalagnanam
Abstract:
We introduce FailureSensorIQ, a novel Multi-Choice Question-Answering (MCQA) benchmarking system designed to assess the ability of Large Language Models (LLMs) to reason and understand complex, domain-specific scenarios in Industry 4.0. Unlike traditional QA benchmarks, our system focuses on multiple aspects of reasoning through failure modes, sensor data, and the relationships between them across various industrial assets. Through this work, we envision a paradigm shift where modeling decisions are not only data-driven using statistical tools like correlation analysis and significance tests, but also domain-driven by specialized LLMs which can reason about the key contributors and useful patterns that can be captured with feature engineering.We evaluate the Industrial knowledge of over a dozen LLMs—including GPT-4, Llama, and Mistral—on FailureSensorIQ from different lenses using Perturbation-Uncertainty-Complexity analysis, Expert Evaluation study, Asset-Specific Knowledge Gap analysis, ReAct agent using external knowledgebases.Even though closed-source models with strong reasoning capabilities approach expert-level performance, the comprehensive benchmark reveals a significant drop in performance that is fragile to perturbations, distractions, and inherent knowledge gaps in the models.We also provide a real-world case study of how LLMs can drive the modeling decisions on 3 different failure prediction datasets related to various assets.We release: (a) expert-curated MCQA for various industrial assets, (b) FailureSensorIQ benchmark and Hugging Face leaderboard based on MCQA built from non-textual data found in ISO documents, and (c) ``LLMFeatureSelector'', an LLM-based feature selection scikit-learn pipeline. The software is available at \url{https://github.com/IBM/FailureSensorIQ}.



Paperid:5190
Authors:Aditi Tiwari, Farzaneh Masoud, Dac Nguyen, Jill Kraft, Heng Ji, Klara Nahrstedt
Abstract:
Modern AI systems struggle most in environments where reliability is critical—scenes with smoke, poor visibility, and structural deformation. Each year, tens of thousands of firefighters are injured on duty, often due to breakdowns in situational perception. We introduce Fire360, a benchmark for evaluating perception and reasoning in safety-critical firefighting scenarios. The dataset includes 228 360° videos from professional training sessions under diverse conditions (e.g., low light, thermal distortion), annotated with action segments, object locations, and degradation metadata. Fire360 supports five tasks: Visual Question Answering, Temporal Action Captioning, Object Localization, Safety-Critical Reasoning, and Transformed Object Retrieval (TOR). TOR tests whether models can match pristine exemplars to fire-damaged counterparts in unpaired scenes, evaluating transformation-invariant recognition. While human experts achieve 83.5% on TOR, models like GPT-4o lag significantly, exposing failures in reasoning under degradation. By releasing Fire360 and its evaluation suite, we aim to advance models that not only see, but also remember, reason, and act under uncertainty.



Paperid:5191
Authors:Yifan Zhou, Tianshi Xu, Jue Hong, Ye Wu, Meng Li
Abstract:
Abstract:Private large language model (LLM) inference based on cryptographic primitives offers a promising path towards privacy-preserving deep learning. However, existing frameworks only support dense LLMs like LLaMA-1 and struggle to scale to mixture-of-experts (MoE) architectures. The key challenge comes from securely evaluating the dynamic routing mechanism in MoE layers, which may reveal sensitive input information if not fully protected. In this paper, we propose CryptoMoE, the first framework that enables private, efficient, and accurate inference for MoE-based models. CryptoMoE balances expert loads to protect expert routing information and proposes novel protocols for secure expert dispatch and combine. CryptoMoE also develops a confidence-aware token selection strategy and a batch matrix multiplication protocol to improve accuracy and efficiency further. Extensive experiments on DeepSeekMoE-16.4B, OLMoE-6.9B, and QWenMoE-14.3B show that CryptoMoE achieves $2.8\sim3.5\times$ end-to-end latency reduction and $3\sim6\times$ communication reduction over a dense baseline with minimum accuracy loss. We also adapt CipherPrune (ICLR'25) for MoE inference and demonstrate CryptoMoE can reduce the communication by up to $4.3 \times$.



Paperid:5192
Authors:Hongyi Zhou, Jingwei Li, Jingzhao Zhang
Abstract:
Abstract:Real world evolves in continuous time but computations are done from finite samples. Therefore, we study algorithms using finite observations in continuous-time linear dynamical systems. We first study the system identification problem, and propose a first non-asymptotic error analysis with finite observations. Our algorithm identifies system parameters without needing integrated observations over certain time intervals, making it more practical for real-world applications. Further we propose a lower bound result that shows our estimator is provably optimal up to constant factors. Moreover, we apply the above algorithm to online control regret analysis for continuous-time linear system. Our system identification method allows us \textcolor{blue}{to} explore more efficiently, enabling the swift detection of ineffective policies. We achieve a regret of $\mathcal{O}(\sqrt{T})$ over a single $T$-time horizon in a controllable system, requiring only $\mathcal{O}(T)$ observations of the system.



Paperid:5193
Authors:Junfei Wu, Jian Guan, Kaituo Feng, Qiang Liu, Shu Wu, Liang Wang, Wei Wu, Tieniu Tan
Abstract:
As textual reasoning with large language models (LLMs) has advanced significant, there has been growing interest in enhancing the multimodal reasoning capabilities of large vision-language models (LVLMs). However, existing methods primarily approach multimodal reasoning in a straightforward, text-centric manner, where both reasoning and answer derivation are conducted purely through text, with the only difference being the presence of multimodal input. As a result, these methods often encounter fundamental limitations in spatial reasoning tasks that demand precise geometric understanding and continuous spatial tracking\textemdash capabilities that humans achieve through mental visualization and manipulation. To address the limitations, we propose drawing to reason in space, a novel paradigm that enables LVLMs to reason through elementary drawing operations in the visual space. By equipping models with basic drawing operations including annotating bounding boxes and drawing auxiliary lines, we empower them to express and analyze spatial relationships through direct visual manipulation, meanwhile avoiding the performance ceiling imposed by specialized perception tools in previous tool-integrated reasoning approaches. To cultivate this capability, we develop a three-stage training framework: cold-start training with synthetic data to establish basic drawing abilities, reflective rejection sampling to enhance self-reflection behaviors, and reinforcement learning to directly optimize for target rewards. Extensive experiments demonstrate that our model, named \textsc{Spark}, consistently outperforms existing methods across diverse spatial reasoning benchmarks involving maze navigation, static spatial reasoning, video-based reasoning and multi-view-based reasoning tasks, with an average improvement of 11.5\%. Ablation studies reveal the critical role of each training stage, with reflective rejection sampling particularly enhancing the model's self-correction capabilities and reasoning potential.



Paperid:5194
Authors:Inkook Chun, Seungjae Lee, Michael Albergo, Saining Xie, Eric Vanden-Eijnden
Abstract:
Diffusion- and flow-based policies deliver state-of-the-art performance on long-horizon robotic manipulation and imitation-learning tasks. However, these controllers employ a fixed inference budget at every control step, regardless of task complexity, leading to computational inefficiency for simple subtasks while potentially underperforming on challenging ones. To address these issues, we introduce Difficulty-Aware Stochastic Interpolant Policy (DA-SIP), a framework that enables robotic controllers to adaptively adjust their integration horizon in real-time based on task difficulty. Our approach employs a difficulty classifier that analyzes RGB-D observations to dynamically select the step budget, the optimal solver variant, and ODE/SDE integration at each control cycle. DA-SIP builds upon the stochastic interpolant formulation to provide a unified framework that unlocks diverse training and inference configurations for diffusion- and flow-based policies. Through comprehensive benchmarks across diverse manipulation tasks, DA-SIP achieves 2.6-4.4× reduction in total computation time while maintaining task-success rates comparable to fixed maximum-computation baselines. By implementing adaptive computation within this framework, DA-SIP transforms generative robot controllers into efficient, task-aware systems that intelligently allocate inference resources where they provide the greatest benefit.



Paperid:5195
Authors:Alexandru Crăciun, Debarghya Ghoshdastidar
Abstract:
The theory of training deep networks has become a central question of modern machine learning and has inspired many practical advancements. In particular, the gradient descent (GD) optimization algorithm has been extensively studied in recent years. A key assumption about GD has appeared in several recent works: the \emph{GD map is non-singular} --- it preserves sets of measure zero under preimages. Crucially, this assumption has been used to prove that GD avoids saddle points and maxima, and to establish the existence of a computable quantity that determines the convergence to global minima (both for GD and stochastic GD). However, the current literature either assumes the non-singularity of the GD map or imposes restrictive assumptions, such as Lipschitz smoothness of the loss (for example, Lipschitzness does not hold for deep ReLU networks with the cross-entropy loss) and restricts the analysis to GD with small step-sizes. In this paper, we investigate the neural network map as a function on the space of weights and biases. We also prove, for the first time, the non-singularity of the gradient descent (GD) map on the loss landscape of realistic neural network architectures (with fully connected, convolutional, or softmax attention layers) and piecewise analytic activations (which includes sigmoid, ReLU, leaky ReLU, etc.) for almost all step-sizes. Our work significantly extends the existing results on the convergence of GD and SGD by guaranteeing that they apply to practical neural network settings and has the potential to unlock further exploration of learning dynamics.



Paperid:5196
Authors:Augusto B. Corrêa, André Pereira, Jendrik Seipp
Abstract:
In recent years, large language models (LLMs) have shown remarkable performance in many problems. However, they fail to plan reliably. Specialized attempts to improve their planning capabilities still produce incorrect plans and fail to generalize to larger tasks. Furthermore, LLMs designed for explicit "reasoning" fail to compete with automated planners while increasing computational costs, which reduces one of the advantages of using LLMs. In this paper, we show how to use LLMs to always generate correct plans, even for out-of-distribution tasks of increasing size. For a given planning domain, we ask an LLM to generate several domain-dependent heuristic functions in the form of Python code, evaluate them on a set of training tasks with a greedy best-first search, and choose the best one. The resulting LLM-generated heuristic functions solve substantially more unseen out-of-distribution test tasks than end-to-end LLM planning, particularly for non-reasoning LLMs. Moreover, they also solve many more tasks than state-of-the-art domain-independent heuristics for classical planning, and are competitive with the strongest learning algorithm for domain-dependent planning. These results are impressive given that our implementation is based on a Python planner and the baselines all build upon highly optimized C++ code. In some domains, the LLM-generated heuristics expand fewer states than the baselines, showing that they are not only efficiently computable but also more informative than the state-of-the-art heuristics. Overall, our results show that sampling a set of planning heuristic functions can significantly improve the planning capabilities of LLMs.



Paperid:5197
Authors:Qinfeng Li, Tianyue Luo, Xuhong Zhang, Yangfan Xie, Zhiqiang Shen, Lijun Zhang, Yier Jin, Hao Peng, Xinkui Zhao, XianWei Zhu, Jianwei Yin
Abstract:
Proprietary large language models (LLMs) exhibit strong generalization capabilities across diverse tasks and are increasingly deployed on edge devices for efficiency and privacy reasons. However, deploying proprietary LLMs at the edge without adequate protection introduces critical security threats. Attackers can extract model weights and architectures, enabling unauthorized copying and misuse. Even when protective measures prevent full extraction of model weights, attackers may still perform advanced attacks, such as fine-tuning, to further exploit the model. Existing defenses against these threats typically incur significant computational and communication overhead, making them impractical for edge deployment.To safeguard the edge-deployed LLMs, we introduce CoreGuard, a computation- and communication-efficient protection method. CoreGuard employs an efficient protection protocol to reduce computational overhead and minimize communication overhead via a propagation protocol. Extensive experiments show that CoreGuard achieves upper-bound security protection with negligible overhead.



Paperid:5198
Authors:Reuben Dorent, Polina Golland, William Wells
Abstract:
Mutual Information (MI) is a fundamental measure of statistical dependence widely used in representation learning. While direct optimization of MI via its definition as a Kullback-Leibler divergence (KLD) is oftenintractable, many recent methods have instead maximized alternative dependence measures, most notably, the Jensen-Shannon divergence (JSD) between joint and product of marginal distributions via discriminative losses. However, the connection between these surrogate objectives and MI remains poorly understood. In this work, we bridge this gap by deriving a new, tight, and tractable lower bound on KLD as a function of JSD in the general case. By specializing this bound to joint and marginal distributions, we demonstratethat maximizing the JSD-based information increases a guaranteed lower bound on the mutual information. Furthermore, we revisit the practical implementation of JSD-based objectives and observe that minimizing the cross-entropy loss of a binary classifier trained to distinguish joint from marginal pairs recovers a known variational lower bound on the JSD.Extensive experiments demonstrate that our lower bound is tight when applied to MI estimation. We compared our lower bound to state-of-the-art neural estimators of variational lower bound across a range of established reference scenarios. Our lower-bound estimator consistently provides a stable, low-variance estimate of a tight lower bound on MI.Taken together, our results provide new theoretical justifications and strong empirical evidence for using discriminative learning in MI-based representation learning.



Paperid:5199
Authors:Jingyu Li, Pengwen Dai, Mingqing Zhu, Chengwei Wang, Haolong Liu, Xiaochun Cao
Abstract:
Recent work has shown that scene text recognition (STR) models are vulnerable to adversarial examples.Different from non-sequential vision tasks, the output sequence of STR models contains rich information. However, existing adversarial attacks against STR models can only lead to a few incorrect characters in the predicted text.These attack results still carry partial information about the original prediction and could be easily corrected by an external dictionary or a language model.Therefore, we propose the Multi-Population Coevolution Search (MPCS) method to attack each character in the image. We first decompose the global optimization objective into sub-objectives to solve the attack pixel concentration problem existing in previous attack methods.While this distributed optimization paradigm brings a new combined perturbation shift problem, we propose a novel coevolution energy function to solve it.Experiments on recent STR models show the superiority of our method.The code is contained in the supplemental material.



Paperid:5200
Authors:Sara Ahmadian, Edith Cohen, Uri Stemmer
Abstract:
Abstract:Dimensionality reduction via linear sketching is a powerful and widely used technique, but it is known to be vulnerable to adversarial inputs. We study the \emph{black-box adversarial setting}, where a fixed, hidden sketching matrix $A \in \mathbb{R}^{k \times n}$ maps high-dimensional vectors $\boldsymbol{v} \in \mathbb{R}^n$ to lower-dimensional sketches $A\boldsymbol{v} \in \mathbb{R}^k$, and an adversary can query the system to obtain approximate $\ell_2$-norm estimates that are computes from the sketch.We present a \emph{universal, nonadaptive attack} that, using $\tilde{O}(k^2)$ queries, either causes a failure in norm estimation or constructs an adversarial input on which the optimal estimator for the query distribution (used by the attack) fails. The attack is completely agnostic to the sketching matrix and to the estimator—it applies to \emph{any} linear sketch and \emph{any} query responder, including those that are randomized, adaptive, or tailored to the query distribution.Our lower bound construction tightly matches the known upper bounds of $\tilde{\Omega}(k^2)$, achieved by specialized estimators for Johnson–Lindenstrauss transforms and AMS sketches. Beyond sketching, our results uncover structural parallels to adversarial attacks in image classification, highlighting fundamental vulnerabilities of compressed representations.



Paperid:5201
Authors:Ling Li, Yao Zhou, Yuxuan Liang, Fugee Tsung, Jiaheng Wei
Abstract:
Abstract:Previous methods for image geo-localization have typically treated the task as either classification or retrieval, often relying on black-box decisions that lack interpretability. The rise of large vision-language models (LVLMs) has enabled a rethinking of geo-localization as a reasoning-driven task grounded in visual cues. However, two major challenges persist. On the data side, existing reasoning-focused datasets are primarily based on street-view imagery, offering limited scene diversity and constrained viewpoints. On the modeling side, current approaches predominantly rely on supervised fine-tuning, which yields only marginal improvements in reasoning capabilities. To address these challenges, we propose a novel pipeline that constructs a reasoning-oriented geo-localization dataset, $\textit{MP16-Reason}$, using diverse social media images. We introduce $\textit{GLOBE}$, $\textbf{G}$roup-relative policy optimization for $\textbf{L}$ocatability assessment and $\textbf{O}$ptimized visual-clue reasoning, yielding $\textbf{B}$i-objective geo-$\textbf{E}$nhancement for the VLM in recognition and reasoning. $\textit{GLOBE}$ incorporates task-specific rewards that jointly enhance locatability assessment, visual clue reasoning, and geolocation accuracy. Both qualitative and quantitative results demonstrate that $\textit{GLOBE}$ outperforms state-of-the-art open-source LVLMs on geo-localization tasks, particularly in diverse visual scenes, while also generating more insightful and interpretable reasoning trajectories.



Paperid:5202
Authors:Hoigi Seo, Dong Un Kang, Hyunjin Cho, Joohoon Lee, Se Young Chun
Abstract:
Large vision-language models (LVLMs), which integrate a vision encoder (VE) with a large language model, have achieved remarkable success across various tasks. However, there are still crucial challenges in LVLMs such as object hallucination, generating descriptions of objects that are not in the input image. Here, we argue that uncertain visual tokens within the VE is a key factor that contributes to object hallucination. Our statistical analysis found that there are positive correlations between visual tokens with high epistemic uncertainty and the occurrence of hallucinations. Furthermore, we show theoretically and empirically that visual tokens in early VE layers that exhibit large representation deviations under small adversarial perturbations indicate high epistemic uncertainty. Based on these findings, we propose a simple yet effective strategy to mitigate object hallucination by modifying the VE only. Our method comprises a proxy method with adversarial perturbations for identifying uncertain visual tokens efficiently and a method to mask these uncertain visual tokens during the self-attention process in the middle layers of the VE, suppressing their influence on visual encoding and thus alleviating hallucinations. Extensive experiments show that our method significantly reduces object hallucinations in LVLMs and can synergistically work with other prior arts.



Paperid:5203
Authors:Yiyou Sun, Shawn Hu, Georgia Zhou, Ken Zheng, Hanna Hajishirzi, Nouha Dziri, Dawn Song
Abstract:
Recent large-scale language models (LLMs) with long Chain-of-Thought reasoning—such as DeepSeek-R1—have achieved impressive results on Olympiad-level mathematics benchmarks. However, they often rely on a narrow set of strategies and struggle with problems that require a novel way of thinking. To systematically probe these limitations, we introduce MathOOD, a controlled yet diverse benchmark designed to evaluate three axes of out-of-distribution generalization inspired by Boden’s typology of creativity: (1) Exploratory—applying known problem-solving skills to more complex instances within the same problem domain; (2) Compositional— combining distinct reasoning skills, previously learned in isolation, to solve novel problems that require integrating these skills in new and coherent ways; and (3) Transformative—adopting novel, often unconventional strategies by moving beyond familiar approaches to solve problems more effectively. MathOOD consists of programmatically generated training–test pairs derived from templated problem generators across geometry, number theory, algebra, combinatorics, logic, and puzzles, with solutions verified using symbolic, numerical, or graphical methods. We evaluate top-tiered LLMs and observe sharp performance degradation as problem complexity increases. Moreover, we fine-tune the Qwen-series models across all generalization settings and observe notable improvements in exploratory generalization, while compositional generalization remains limited and transformative reasoning shows 0 improvement. By isolating and quantifying these fine-grained failures, MathOOD lays the groundwork for advancing LLMs toward genuine mathematical creativity beyond mechanical proficiency.



Paperid:5204
Authors:Stephen Pfohl, Natalie Harris, Chirag Nagpal, David Madras, Vishwali Mhasawade, Olawale Salaudeen, Awa Dieng, Shannon Sequeira, Santiago Arciniegas, Lillian Sung, Nnamdi Ezeanochie, Heather Cole-Lewis, Katherine Heller, Sanmi Koyejo, Alexander D'Amour
Abstract:
Disaggregated evaluation across subgroups is critical for assessing the fairness of machine learning models, but its uncritical use can mislead practitioners. We show that equal performance across subgroups is an unreliable measure of fairness when data are representative of the relevant populations but reflective of real-world disparities. Furthermore, when data are not representative due to selection bias, both disaggregated evaluation and alternative approaches based on conditional independence testing may be invalid without explicit assumptions regarding the bias mechanism. We use causal graphical models to predict metric stability across subgroups under different data generating processes. Our framework suggests complementing disaggregated evaluations with explicit causal assumptions and analysis to control for confounding and distribution shift, including conditional independence testing and weighted performance estimation. These findings have broad implications for how practitioners design and interpret model assessments given the ubiquity of disaggregated evaluation.



Paperid:5205
Authors:Zeyuan Liu, Zhihe Yang, Jiawei Xu, Rui Yang, Jiafei Lyu, Baoxiang Wang, Yunjian Xu, Xiu Li
Abstract:
Real-world datasets collected from sensors or human inputs are prone to noise and errors, posing significant challenges for applying offline reinforcement learning (RL). While existing methods have made progress in addressing corrupted actions and rewards, they remain insufficient for handling corruption in high-dimensional state spaces and for cases where multiple elements in the dataset are corrupted simultaneously. Diffusion models, known for their strong denoising capabilities, offer a promising direction for this problem—but their tendency to overfit noisy samples limits their direct applicability.To overcome this, we proposeAmbientDiffusion-Guided Dataset Recovery (ADG), a novel approach that pioneers the use of diffusion models to tackle data corruption in offline RL. First, we introduce Ambient Denoising Diffusion Probabilistic Models (DDPM) from approximated distributions, which enable learning on partially corrupted datasets with theoretical guarantees. Second, we use the noise-prediction property of Ambient DDPM to distinguish between clean and corrupted data, and then use the clean subset to train a standard DDPM. Third, we employ the trained standard DDPM to refine the previously identified corrupted data, enhancing data quality for subsequent offline RL training. A notable strength of ADG is its versatility—it can be seamlessly integrated with any offline RL algorithm. Experiments on a range of benchmarks, including MuJoCo, Kitchen, and Adroit, demonstrate that ADG effectively mitigates the impact of corrupted data and improves the robustness of offline RL under various noise settings, achieving state-of-the-art results.



Paperid:5206
Authors:Yusuf Dalva, Hidir Yesiltepe, Pinar Yanardag Delul
Abstract:
We introduce LoRAShop, the first framework for multi-concept image generation and editing with LoRA models. LoRAShop builds on a key observation about the feature interaction patterns inside Flux-style diffusion transformers: concept-specific transformer features activate spatially coherent regions early in the denoising process. We harness this observation to derive a disentangled latent mask for each concept in a prior forward pass and blend the corresponding LoRA weights only within regions bounding the concepts to be personalized. The resulting edits seamlessly integrate multiple subjects or styles into the original scene while preserving global context, lighting, and fine details. Our experiments demonstrate that LoRAShop delivers better identity preservation compared to baselines. By eliminating retraining and external constraints, LoRAShop turns personalized diffusion models into a practical `photoshop-with-LoRAs' tool and opens new avenues for compositional visual storytelling and rapid creative iteration.



Paperid:5207
Authors:WEI-KAI CHANG, Rajiv Khanna
Abstract:
Understanding the dynamics of optimization algorithms in deep learning has become increasingly critical, especially as models grow in scale and complexity. Despite the empirical success of stochastic gradient descent (SGD) and its variants in finding solutions that generalize well, the precise mechanisms underlying this generalization remain poorly understood. A particularly intriguing aspect of this phenomenon is the bias of optimization algorithms towards certain types of minima—often flatter or simpler—especially in overparameterized regimes. While prior works have associated flatness of the loss landscape with better generalization, tools to mechanistically connect data, optimization algorithms, and the nature of the resulting minima are still limited. For instance, methods like Sharpness-Aware Minimization (SAM) have shown practical gains by explicitly promoting flatness, but lack a unified theoretical framework explaining their influence across different data structures and model architectures. In this work, we introduce a comprehensive linear stability analysis framework to dissect the behavior of optimization algorithms—SGD, random perturbations, and SAM—in neural networks, focusing particularly on two-layer ReLU models. Our approach is built upon a novel coherence measure that captures the interaction between data geometry and gradient similarity, providing new insights into why and how certain solutions are favored.



Paperid:5208
Authors:CONGZHANG SHAO, Quan Yuan, Guiyang Luo, Yue Hu, Liu Yilin, Danni Wang, Rui Pan, Bo Chen, Jinglin Li
Abstract:
Collaborative perception expands the perception range by sharing information among agents, effectively improving task performance. Immutable heterogeneity poses a significant challenge in collaborative perception, as participating agents may employ different and fixed perception models. This leads to domain gaps in the intermediate features shared among agents, consequently degrading collaborative performance.Aligning the features of all agents to a common representation can eliminate domain gaps with low training cost. However, in existing methods, the common representation is designated as the representation of a specific agent, making it difficult for agents with significant domain discrepancies from this specific agent to achieve proper alignment.This paper proposes NegoCollab, a heterogeneous collaboration method based on negotiated common representation. It achieves bidirectional transformation of each modality's features between local representation space and common representation space through paired sender-receiver, thereby eliminating domain gaps. The common representation in NegoCollab is negotiated from local representations of each modality's agent via a negotiator introduced during training, effectively reducing inherent domain discrepancies with each local representation. Furthermore, to better align local representations with the multimodal common representation, we introduce both structural alignment loss and pragmatic alignment loss alongside the conventional distribution alignment loss during supervised training, enabling comprehensive knowledge distillation from the common representation to the senders.The experimental results demonstrate that NegoCollab significantly outperforms existing methods in common representation-based collaboration approaches. The negotiation-based mechanism for acquiring common representations provides more diverse and reliable alternatives for establishing common representations required in heterogeneous collaboration perception.



Paperid:5209
Authors:Ji Xia, Yizi Zhang, Shuqi Wang, Genevera Allen, Liam Paninski, Cole Hurwitz, Kenneth Miller
Abstract:
Characterizing interactions between brain areas is a fundamental goal of systems neuroscience. While such analyses are possible when areas are recorded simultaneously, it is rare to observe all combinations of areas of interest within a single animal or recording session. How can we leverage multi-animal datasets to better understand multi-area interactions? Building on recent progress in large-scale, multi-animal models, we introduce NeuroPaint, a masked autoencoding approach for inferring the dynamics of unobserved brain areas. By training across animals with overlapping subsets of recorded areas, NeuroPaint learns to reconstruct activity in missing areas based on shared structure across individuals. We train and evaluate our approach on both synthetic data and two multi-animal, multi-area Neuropixels datasets. Our results demonstrate that models trained across animals with partial observations can successfully in-paint the dynamics of unrecorded areas, enabling multi-area analyses that transcend the limitations of any single experiment.



Paperid:5210
Authors:Nima Hadidi, Jason Chan, Ebrahim Feghhi, Jonathan Kao
Abstract:
Abstract:Surface electromyography (sEMG) at the wrists could enable natural, keyboard‑free text entry, yet the state‑of‑the‑art emg2qwerty baseline still misrecognizes 51.8\% of characters zero‑shot on unseen users and 7.0\% after user‑specific fine‑tuning. We trace much of these errors to mismatched cross‑user signal statistics, fragile reliance on high‑order feature dependencies, and the absence of architectural inductive biases aligned with the bilateral nature of typing. To address these issues, we introduce three simple modifications: (i) Rolling Time Normalization which adaptively aligns input distributions across users; (ii) Aggressive Channel Masking, which encourages reliance on low‑order feature combinations more likely to generalize across users; and (iii) a Split‑and‑Share encoder that processes each hand independently with weight‑shared streams to reflect the bilateral symmetry of the neuromuscular system. Combined with a five‑fold reduction in spectral resolution (33$\rightarrow$6 frequency bands), these components yield a compact Split-and-Share model, SplashNet‑mini, which uses only ¼ the parameters and 0.6× the FLOPs of the baseline while reducing character error rate (CER) to 36.4\% zero‑shot and 5.9\% after fine‑tuning. An upscaled variant, SplashNet (½ parameters, 1.15× FLOPs of the baseline), further lowers error to 35.7\% and 5.5\%, representing 31\% and 21\% relative improvements in the zero-shot and finetuned settings, respectively. SplashNet therefore establishes a new state-of-the-art without requiring additional data.



Paperid:5211
Authors:Guanlin (Frank) Wu, Boyan Su, Yang Zhao, Pu Wang, Yichen Lin, Hao Frank Yang
Abstract:
How to encode visual-spatial intelligence (VSI) into representative and informative features remain an open challenge. Instead of following traditional Visual Question Answering (VQA)-style representation, we introduce spatial intelligence grid (SIG): a structured, grid-based data schema that embeds geometrical spatial relationship among objects along with physical priors in human world. We further derive a set of SIG-based optimal evaluation metrics that rigorously quantify a model’s true VSI capabilities. In few-shot in-context learning experiments on state-of-the-art multimodal LLMs (e.g. GPT-4o, Gemini-2.5-Pro), SIG yields consistently larger, more stable, and more comprehensive improvements across all VSI metrics compared to VQA-style representations, demonstrating its potential as a novel data schema for learning VSI. Based on SIG, we create SIGBench, a benchmark containing 1.4K driving frames annotated with ground-truth SIG labels and human gaze attention, supporting both grid-based machine VSI tasks and human-like attention-driven VSI tasks in autonomous-driving scenarios.



Paperid:5212
Authors:Jinyang Li, Jack Williams, Nick McKenna, Arian Askari, Nicholas Wilson, Reynold Cheng
Abstract:
Knowledge distillation from Large Language Models (LLMs) to locally hosted Small Language Models (SLMs) provides advantages for Data Analysis Code Generation (DACG) such as privacy protection. However, achieving effective distillation without resource-intensive training is challenging. This paper investigates whether LLMs can distill knowledge to SLMs through In-Context Learning (ICL), a training-free method for rapid task adaptation. We present the DarGO: Distillation and Adaptive Reasoning-Guided Orchestration framework, which facilitates automatic knowledge distillation from LLMs to SLMs. DarGO consists of three phases: exploration through an Model Orchestration Interface (MOI), Memory Collection of successful trajectories, and Knoweldge-driven Inference. We evaluate DarGO on three challenging DACG benchmarks (WikiTQ, TabMWP, and Bird-SQL), each with in-domain training sets that enable detailed analysis of knowledge distillation effectiveness. DarGO demonstrates a substantial relative performance improvement of 27.5\% on average for the student SLMs. To further observe generalization capabilities, we evaluate the \method across different teacher-student model combinations, knowledge transfer scenarios, and unified memory approaches for more advanced, test-only data analysis tasks. Our findings contribute a novel perspective on distillation methods that enhance high performance for SLMs while avoiding intensive fine-tuning.



Paperid:5213
Authors:Mohsen Heidari, Masih Mozakka, Wojciech Szpankowski
Abstract:
Gradient estimation is a central challenge in training \acp{PQC} for hybrid quantum-classical optimization and learning problems. This difficulty arises from several factors, including the exponential dimensionality of the Hilbert spaces and the information loss in quantum measurements. Existing estimators, such as finite difference and the parameter shift rule, often fail to adequately address these challenges for certain classes of PQCs. In this work, we propose a novel gradient estimation framework that leverages the underlying Lie algebraic structure of PQCs, combined with the Hadamard test. By analyzing the differential of the matrix exponential in Lie algebras, we derive an expression for the gradient as a linear combination of expectation values obtained via Hadamard tests. The coefficients in this decomposition depend solely on the circuit's parameterization and can be computed efficiently. Also, these expectation values can be estimated using state-of-the-art shadow tomography techniques. Our approach enables efficient gradient estimation, requiring a number of measurement shots that scales logarithmically with the number of parameters, and with polynomial classical and quantum time. This is an exponential reduction in the measurement cost and a polynomial speed-up in time compared to existing works.



Paperid:5214
Authors:Xinxin You, Qixin Sun, Xien Liu, Chenwei Yan, Xiao Zhang, Chen Ning, Xiangling Fu, Si Liu, Shijin Wang, Guoping Hu, Ji Wu
Abstract:
Inconsistent hallucinations remain a major challenge for large language models (LLMs), undermining the accuracy and reliability of fact-based reasoning in real-world applications. Existing approaches often rely on task-specific training or adaptation, such as hand-crafted synthetic datasets for domain tasks or solutions mainly focused on numerical reasoning, thereby limiting generalizability to broader, unseen NLP tasks. Inspired by the structural rigor and logical consistency of programming languages, we observe that fact-based texts can be mapped to programming structures due to their inherent patterns. We further propose FACT, a novel Fact-driven Alternating Code-text Training framework that alternates between text-to-code and code-to-text prediction. FACT is the first task-agnostic paradigm that embeds code and natural language in a shared semantic space, thereby transferring the logical consistency of code to LLM outputs in NLP tasks. Experiments show that with only a small subset of Wiki-40B-en for training, FACT reduces inconsistent hallucinations by 2.7%–8.0% and improves overall performance by 2.5%–6.1% in three leading LLMs and four diverse datasets covering QA and summarization tasks. This framework offers a new perspective on addressing challenging hallucinations in LLMs, contributing to more reliable AI.



Paperid:5215
Authors:Davide Straziota, Elizaveta Demyanenko, Carlo Baldassi, Carlo Lucibello
Abstract:
Abstract:We consider random instances of non-convex perceptron problems in the high-dimensional limit of a large number of examples $M$ and weights $N$, with finite load $\alpha = M/N$.We develop a formalism based on replica theory to predict the fundamental limits of efficiently sampling the solution space using generative diffusion algorithms, conjectured to be saturated when the score function is provided by Approximate Message Passing.For the spherical perceptron with negative margin $\kappa$, we find that the uniform distribution over solutions can be efficiently sampled in most of the Replica Symmetric region of the $\alpha$–$\kappa$ plane.In contrast, for binary weights, sampling from the uniform distribution remains intractable. A theoretical analysis of this obstruction leads us to identify a potential $U(s) = -\log(s)$, under which the corresponding tilted distribution becomes efficiently samplable via diffusion.Moreover, we show numerically that an annealing procedure over the shape of this potential yields a fast and robust Markov Chain Monte Carlo algorithm for sampling the solution space of the binary perceptron.



Paperid:5216
Authors:Georgios Mentzelopoulos, Ioannis Asmanis, Konrad Kording, Eva Dyer, Kostas Daniilidis, Flavia Vitale
Abstract:
Brain-computer interfaces (BCIs) promise to enable vital functions, such as speech and prosthetic control, for individuals with neuromotor impairments. Central to their success are neural decoders, models that map neural activity to intended behavior. Current learning-based decoding approaches fall into two classes: simple, causal models that lack generalization, or complex, non-causal models that generalize and scale offline but struggle in real-time settings. Both face a common challenge, their reliance on power-hungry artificial neural network backbones, which makes integration into real-world, resource-limited systems difficult. Spiking neural networks (SNNs) offer a promising alternative. Because they operate causally (i.e. only on present and past inputs) these models are suitable for real-time use, and their low energy demands make them ideal for battery-constrained environments. To this end, we introduceSpikachu: a scalable, causal, and energy-efficient neural decoding framework based on SNNs. Our approach processes binned spikes directly by projecting them into a shared latent space, where spiking modules, adapted to the timing of the input, extract relevant features; these latent representations are then integrated and decoded to generate behavioral predictions. We evaluate our approach on 113 recording sessions from 6 non-human primates, totaling 43 hours of recordings. Our method outperforms causal baselines when trained on single sessions using between 2.26× and 418.81× less energy. Furthermore, we demonstrate that scaling up training to multiple sessions and subjects improves performance and enables few-shot transfer to unseen sessions, subjects, and tasks. Overall, Spikachu introduces a scalable, online-compatible neural decoding framework based on SNNs, whose performance is competitive relative to state-of-the-art models while consuming orders of magnitude less energy.



Paperid:5217
Authors:Roman Vashurin, Maiya Goloburda, Albina Ilina, Aleksandr Rubashevskii, Preslav Nakov, Artem Shelmanov, Maxim Panov
Abstract:
Uncertainty quantification (UQ) methods for Large Language Models (LLMs) encompass a variety of approaches, with two major types being particularly prominent: information-based, which focus on model confidence expressed as token probabilities, and consistency-based, which assess the semantic relationship between multiple outputs generated using repeated sampling. Several recent methods have combined these two approaches to boost UQ performance. However, they sometimes fail to outperform much simpler baseline methods. Our work discusses the fundamental approach to constructing uncertainty measures that directly links uncertainty with the minimum Bayes risks achieved by LLM decoding. Building on these findings, we propose a novel approach to integrating model confidence with output consistency, resulting in a family of efficient and robust UQ methods. We evaluate our approach across various tasks such as question answering, abstractive summarization, and machine translation, demonstrating sizable improvements over state-of-the-art UQ approaches.



Paperid:5218
Authors:Robert Nerem, Zhishang Luo, Akbar Rafiey, Yusu Wang
Abstract:
Continuously extending combinatorial optimization objectives is a powerful technique commonly applied to the optimization of set functions. However, few such methods exist for extending functions on permutations, despite the fact that many combinatorial optimization problems, such as the quadratic assignment problem (QAP) and the traveling salesperson problem (TSP), are inherently optimization over permutations. We present Birkhoff Extension (BE), an almost-everywhere-differentiable continuous polytime-computable extension of any real-valued function on permutations to doubly stochastic matrices. Key to this construction is our introduction of a continuous variant of the well-known Birkhoff decomposition. Our extension has several nice properties making it appealing for optimization problems. First, BE provides a rounding guarantee, namely any solution to the extension can be efficiently rounded to a permutation without increasing the function value. Furthermore, an approximate solution in the relaxed case will give rise to an approximate solution in the space of permutations. Second, using BE, any real-valued optimization objective on permutations can be extended to an almost-everywhere-differentiable objective function over the space of doubly stochastic matrices. This makes our BE amenable to not only gradient-descent based optimization, but also unsupervised neural combinatorial optimization where training often requires a differentiable loss. Third, based on the above properties, we present a simple optimization procedure which can be readily combined with existing optimization approaches to offer local improvements (i.e., the quality of the final solution is no worse than the initial solution). Finally, we also adapt our extension to optimization problems over a class of trees, such as Steiner tree and optimization-based hierarchical clustering.We present experimental results to verify our theoretical results on several combinatorial optimization problems related to permutations.



Paperid:5219
Authors:Junteng Liu, Yuanxiang Fan, Jiang Zhuo, Han Ding, Yongyi Hu, Chi Zhang, Yiqi Shi, Shitong Weng, Chen, Shiqi Chen, Mozhi Zhang, Pengyu Zhao, Junxian He
Abstract:
Recent advances such as OpenAI-o1 and DeepSeek R1 have demonstrated the potential of Reinforcement Learning (RL) to enhance reasoning abilities in Large Language Models (LLMs). While open-source replication efforts have primarily focused on mathematical and coding domains, methods and resources for developing general reasoning capabilities remain underexplored. This gap is partly due to the challenge of collecting diverse and verifiable reasoning data suitable for RL.We hypothesize that logical reasoning is critical for developing general reasoning capabilities, as logic forms a fundamental building block of reasoning. In this work, we present SynLogic, a data synthesis framework and dataset that generates diverse logical reasoning data at scale, encompassing 35 diverse logical reasoning tasks. The SynLogic approach enables controlled synthesis of data with adjustable difficulty and quantity. Importantly, all examples can be verified by simple rules, making them ideally suited for RL with verifiable rewards.In our experiments, we validate the effectiveness of RL training on the SynLogic dataset based on 7B and 32B models. SynLogic leads to state-of-the-art logical reasoning performance among open-source datasets, surpassing DeepSeek-R1-Distill-Qwen-32B by 6 points on BBEH. Furthermore, mixing SynLogic data with mathematical and coding tasks improves the training efficiency of these domains and significantly enhances reasoning generalization. Notably, our mixed training model outperforms DeepSeek-R1-Zero-Qwen-32B across multiple benchmarks.These findings position SynLogic as a valuable resource for advancing the broader reasoning capabilities of LLMs. We will open-source both the data synthesis pipeline and the SynLogic dataset.



Paperid:5220
Authors:Seunghyeok Shin, Dabin Kim, Hongki Lim
Abstract:
Reconstructing high-quality point clouds from images remains challenging in computer vision. Existing generative model, particularly diffusion model, based approaches that directly learn the posterior may suffer from inflexibility—they require conditioning signals during training, support only fixed numbers of input views, and need complete retraining for different measurements. Recent diffusion-based methods have attempted to address this by combining prior models with likelihood updates, but they rely on heuristic fixed step sizes for likelihood update that lead to slow convergence and suboptimal reconstruction quality. We advance this line of approach by integrating our novel Forward Curvature-Matching(FCM) update method with diffusion sampling. Our method dynamically determines optimal step sizes using only forward automatic differentiation and finite-difference curvature estimates, enabling precise optimization of the likelihood update. This formulation enables high-fidelity reconstruction from both single and multi-view inputs, and supports various input modalities through simple operator substitution—all without retraining. Experiments on ShapeNet and CO3D datasets demonstrate that our method achieves superior reconstruction quality with 10.2\% and 12.8\% improvement in F-score, respectively, validating its efficiency and adaptability for practical applications.



Paperid:5221
Authors:Matan Schliserman, Tomer Koren
Abstract:
Abstract:We study the generalization performance of unregularized gradient methods for separable linear classification. While previous work mostly deal with the binary case, we focus on the multiclass setting with $k$ classesand establish novel population risk bounds for Gradient Descent for loss functions that decay to zero.In this setting, we show risk bounds thatreveal thatconvergence rates are crucially influenced by the geometry of the loss template, as formalized by Wang and Scott (2024), rather than of the loss function itself.Particularly, we establish risk upper bounds that holds for any decay rate of the loss whose template is smooth with respect to the $p$-norm.In the case of exponentially decaying losses, our results indicates a contrast between the $p=\infty$ case, where the risk exhibits a logarithmic dependence on $k$, and $p=2$ where the risk scales linearly with $k$.To establish this separation formally, we also prove a lower bound in the latter scenario, demonstrating that the polynomial dependence on $k$ is unavoidable.Central to our analysis is a novel bound on the Rademacher complexity of low-noise vector-valued linear predictors with a loss template smooth w.r.t. general $p$-norms.



Paperid:5222
Authors:Riccardo Corvi, Davide Cozzolino, Ekta Prashnani, Shalini De Mello, Koki Nagano, Luisa Verdoliva
Abstract:
Synthetic video generation is progressing very rapidly. The latest models can produce very realistic high-resolution videos that are virtually indistinguishable from real ones. Although several video forensic detectors have been recently proposed, they often exhibit poor generalization, which limits their applicability in a real-world scenario. Our key insight to overcome this issue is to guide the detector towardsseeingwhatreallymatters. In fact, a well-designed forensic classifier should focus on identifying intrinsic low-level artifacts introduced by a generative architecture rather than relying on high-level semantic flaws that characterize a specific model. In this work, first, we study different generative architectures, searching and identifying discriminative features that are unbiased, robust to impairments, and shared across models. Then, we introduce a novel forensic-oriented data augmentation strategy based on the wavelet decomposition and replace specific frequency-related bands to drive the model to exploit more relevant forensic cues. Our novel training paradigm improves the generalizability of AI-generated video detectors, without the need for complex algorithms and large datasets that include multiple synthetic generators. To evaluate our approach, we train the detector using data from a single generative model and test it against videos produced by a wide range of other models. Despite its simplicity, our method achieves a significant accuracy improvement over state-of-the-art detectors and obtains excellent results even on very recent generative models, such as NOVA and FLUX. Code and data will be made publicly available.



Paperid:5223
Authors:Wei Huang, Hanchen Wang, Dong Wen, SHAOZHEN MA, Wenjie Zhang, Xuemin Lin
Abstract:
Graph Edit Distance (GED) is a fundamental graph similarity metric widely used in various applications. However, computing GED is an NP-hard problem. Recent state-of-the-art hybrid GED solver has shown promising performance by formulating GED as a bipartite graph matching problem, then leveraging a generative diffusion model to predict node matching between two graphs, from which both the GED and its corresponding edit path can be extracted using a traditional algorithm. However, such methods typically rely heavily on ground-truth supervision, where the ground-truth labels are often costly to obtain in real-world scenarios. In this paper, we propose GEDRanker, a novel unsupervised GAN-based framework for GED computation. Specifically, GEDRanker consists of a matching-based GED solver and introduces an interpretable preference-aware discriminator with an effective training strategy to guide the matching-based GED solver toward generating high-quality node matching without the need for ground-truth labels. Extensive experiments on benchmark datasets demonstrate that our GEDRanker enables the matching-based GED solver to achieve near-optimal solution quality without any ground-truth supervision.



Paperid:5224
Authors:Berken Utku Demirel, Christian Holz
Abstract:
Self-supervised learning (SSL) has emerged as a powerful paradigm for learning representations without labeled data. Most SSL approaches rely on strong, well-established, handcrafted data augmentations to generate diverse views for representation learning. However, designing such augmentations requires domain-specific knowledge and implicitly imposes representational invariances on the model, which can limit generalization. In this work, we propose an unsupervised representation learning method that replaces augmentations by generating views using orthonormal bases and overcomplete frames. We show that embeddings learned from orthonormal and overcomplete spaces reside on distinct manifolds, shaped by the geometric biases introduced by representing samples in different spaces. By jointly leveraging the complementary geometry of these distinct manifolds, our approach achieves superior performance without artificially increasing data diversity through strong augmentations. We demonstrate the effectiveness of our method on nine datasets across five temporal sequence tasks, where signal-specific characteristics make data augmentations particularly challenging. Without relying on augmentation-induced diversity, our method achieves performance gains of up to 20–30\% over existing self-supervised approaches.Source code: In supplementary.



Paperid:5225
Authors:Yuhong Luo, Austin Hoag, Xintong Wang, Philip Thomas, Przemyslaw Grabowicz
Abstract:
Abstract:Representation learning is increasingly applied to generate representations that generalize well across multiple downstream tasks. Ensuring fairness guarantees in representation learning is crucial to prevent unfairness toward specific demographic groups in downstream tasks.In this work, we formally introduce the task of learning representations that achieve high-confidence fairness. We aim to guarantee that demographic disparity in every downstream prediction remains bounded by a *user-defined* error threshold $\epsilon$, with *controllable* high probability.To this end, we propose the ***F**air **R**epresentation learning with high-confidence **G**uarantees (FRG)* framework, which provides these high-confidence fairness guarantees by leveraging an optimized adversarial model. We empirically evaluate FRG on three real-world datasets, comparing its performance to six state-of-the-art fair representation learning methods. Our results demonstrate that FRG consistently bounds unfairness across a range of downstream models and tasks.



Paperid:5226
Authors:Yibo Wang, Tiansheng Huang, Li Shen, Huanjin Yao, Haotian Luo, Rui Liu, Naiqiang Tan, Jiaxing Huang, Dacheng Tao
Abstract:
Harmful fine-tuning attack introduces significant security risks to the fine-tuning services. Main-stream defenses aim to vaccinate the model such that the later harmful fine-tuning attack is less effective. However, our evaluation results show that such defenses are fragile-- with a few fine-tuning steps, the model still can learn the harmful knowledge. To this end, we do further experiment and find that an embarrassingly simple solution-- adding purely random perturbations to the fine-tuned model, can recover the model from harmful behaviors, though it leads to a degradation in the model’s fine-tuning performance. To address the degradation of fine-tuning performance, we further propose \methodname, which optimizes an adaptive perturbation that will be applied to the model after fine-tuning. \methodname maintains model's safety alignment performance without compromising downstream fine-tuning performance. Comprehensive experiments are conducted on different harmful ratios, fine-tuning tasks and mainstream LLMs, where the average harmful scores are reduced by up-to 21.2%, while maintaining fine-tuning performance. As a by-product, we analyze the adaptive perturbation and show that different layers in various LLMs have distinct safety coefficients. Source code available at https://anonymous.4open.science/r/Panacea.



Paperid:5227
Authors:Kevin Wang, Ishaan Javali, Michał Bortkiewicz, Tomasz Trzcinski, Benjamin Eysenbach
Abstract:
Abstract:Scaling up self-supervised learning has driven breakthroughs in language and vision, yet comparable progress has remained elusive in reinforcement learning (RL). In this paper, we study building blocks for self-supervised RL that unlock substantial improvements in scalability, with network depth serving as a critical factor. Whereas most RL papers in recent years have relied on shallow architectures (around 2 -- 5 layers), we demonstrate that increasing the depth up to 1024 layers can significantly boost performance.Our experiments are conducted in an unsupervised goal-conditioned setting, where no demonstrations or rewards are provided, so an agent must explore (from scratch) and learn how to maximize the likelihood of reaching commanded goals.Evaluated on simulated locomotion and manipulation tasks, our approach increases performance on the self-supervised contrastive RL algorithm by $2\times$ -- $50\times$, outperforming other goal-conditioned baselines.Increasing the model depth not only increases success rates but also qualitatively changes the behaviors learned.



Paperid:5228
Authors:Shihan Dou, Ming Zhang, Chenhao Huang, Jiayi Chen, Feng Chen, Shichun Liu, Yan Liu, Chenxiao Liu, CHENG ZHONG, Zongzhang Zhang, Tao Gui, Chao Xin, Wei Chengzhi, Lin Yan, Qi Zhang, Xuanjing Huang
Abstract:
We introduce EvaLearn, a pioneering benchmark designed to evaluate large language models (LLMs) on their learning capability and efficiency in challenging tasks, a critical, yet underexplored aspect of model potential. EvaLearn contains 648 challenging problems across six task types, grouped into 182 sequences, each sequence dedicated to one task type. Diverging from most existing benchmarks that evaluate models in parallel, EvaLearn requires models to solve problems sequentially, allowing them to leverage the experience gained from previous solutions. EvaLearn provides five comprehensive automated metrics to evaluate models and quantify their learning capability and efficiency. We extensively benchmark nine frontier models and observe varied performance profiles: some models, such as Claude-3.7-sonnet, start with moderate initial performance but exhibit strong learning ability, while some models struggle to benefit from experience and may even show negative transfer. Moreover, we investigate model performance under two learning settings and find that instance-level rubrics and teacher-model feedback further facilitate model learning. Importantly, we observe that current LLMs with stronger static abilities do not show a clear advantage in learning capability across all tasks, highlighting that EvaLearn evaluates a new dimension of model performance. We hope EvaLearn provides a novel evaluation perspective for assessing LLM potential and understanding the gap between models and human capabilities, promoting the development of deeper and more dynamic evaluation approaches. All datasets, the automatic evaluation framework, and the results studied in this paper are available in the supplementary materials.



Paperid:5229
Authors:Huanjia Zhu, Shuyuan Zheng, Yishu Liu, Sudong Cai, Bingzhi Chen
Abstract:
Language bias in Visual Question Answering (VQA) arises when models exploit spurious statistical correlations between question templates and answers, particularly in out-of-distribution scenarios, thereby neglecting essential visual cues and compromising genuine multimodal reasoning. Despite numerous efforts to enhance the robustness of VQA models, a principled understanding of how such bias originates and influences model behavior remains underdeveloped. In this paper, we address this gap through a comprehensive empirical and theoretical analysis, revealing that modality-specific gradient imbalances, which originate from the inherent heterogeneity of multimodal data, lead to skewed feature fusion and biased classifier weights. To alleviate these issues, we propose a novel Multi-Margin Collaborative Debiasing (MMCD) framework that adaptively integrates frequency-, confidence-, and difficulty-aware angular margins with a dynamic difficulty-aware contrastive learning mechanism, to dynamically reshape decision boundaries. Extensive experiments across multiple challenging VQA benchmarks confirm the consistent superiority of our proposed MMCD over state-of-the-art baselines in combating language bias.



Paperid:5230
Authors:Yujun Kim, Chaewon Moon, Chulhee Yun
Abstract:
Abstract:We study the parameter complexity of robust memorization for ReLU networks: the number of parameters required to interpolate any dataset with $\epsilon$-separation between differently labeled points, while ensuring predictions remain consistent within a $\mu$-ball around each training example. We establish upper and lower bounds on the parameter count as a function of the robustness ratio $\rho = \mu / \epsilon$. Unlike prior work, we provide a fine-grained analysis across the entire range $\rho \in (0,1)$ and obtain tighter upper and lower bounds that improve upon existing results. Our findings reveal that the parameter complexity of robust memorization matches that of non-robust memorization when $\rho$ is small, but grows with increasing $\rho$. As a special case, when the input dimension is comparable to or exceeds the dataset size, our bounds become tight (up to logarithmic factors) across the entire range of $\rho$.



Paperid:5231
Authors:Chen Liu, Wenfang Yao, Kejing Yin, William Cheung, Jing Qin
Abstract:
Abstract:Longitudinal multimodal data, including electronic health records (EHR) and sequential chest X-rays (CXRs), is critical for modeling disease progression, yet remains underutilized due to two key challenges: (1) redundancy in consecutive CXR sequences, where static anatomical regions dominate over clinically-meaningful dynamics, and (2) temporal misalignment between sparse, irregular imaging and continuous EHR data. We introduce $\texttt{DiPro}$, a novel framework that addresses these challenges through region-aware disentanglement and multi-timescale alignment. First, we disentangle static (anatomy) and dynamic (pathology progression) features in sequential CXRs, prioritizing disease-relevant changes. Second, we hierarchically align these static and dynamic CXR features with asynchronous EHR data via local (pairwise interval-level) and global (full-sequence) synchronization to model coherent progression pathways. Extensive experiments on the MIMIC dataset demonstrate that $\texttt{DiPro}$ could effectively extract temporal clinical dynamics and achieve state-of-the-art performance on both disease progression identification and general ICU prediction tasks.



Paperid:5232
Authors:Hao Liang, shuqing shi, Yudi Zhang, Biwei Huang, Yali Du
Abstract:
Abstract:Large‑scale networked systems—traffic, power, and wireless grids—challenge reinforcement‑learning agents with both scale and environment shifts. To address these challenges, we propose \texttt{GSAC} (\textbf{G}eneralizable and \textbf{S}calable \textbf{A}ctor‑\textbf{C}ritic), a framework that couples causal representation learning with meta actor‑critic learning to achieve both scalability and domain generalization. Each agent first learns a sparse local causal mask that provably identifies the minimal neighborhood variables influencing its dynamics, yielding exponentially tight approximately compact representations (ACRs) of state and domain factors. These ACRs bound the error of truncating value functions to $\kappa$-hop neighborhoods, enabling efficient learning on graphs. A meta actor‑critic then trains a shared policy across multiple source domains while conditioning on the compact domain factors; at test time, a few trajectories suffice to estimate the new domain factor and deploy the adapted policy. We establish finite‑sample guarantees on casual recovery, actor-critic convergence, and adaptation gap, and show on wireless‑communication benchmarks that \texttt{GSAC} adapts rapidly and decisively outperforms training from scratch.



Paperid:5233
Authors:Rakshit Trivedi, Kartik Sharma, David Parkes
Abstract:
Effective human-AI coordination requires artificial agents capable of exhibiting and responding to human-like behaviors while adapting to changing contexts. Imitation learning has emerged as one of the prominent approaches to build such agents by training them to mimic human-demonstrated behaviors; however, current methods struggle to capture the inherent diversity and non-Markovian nature of human actions, and critically lack the ability to steer behavior at inference time. Drawing inspiration from human cognitive processes, where inner speech guides action selection before execution, we propose MIMIC (Modeling Inner Motivations for Imitation and Control), a framework that uses language as an internal representation of behavioral intent. MIMIC employs novel use of vision-language models as developmental scaffolding to train a conditional autoencoder capable of generating inner speech from observations. A diffusion-based behavior cloning policy then selects actions conditioned on both current observations and the generated inner speech. MIMIC enables fine-grained steering of behavior at inference time by conditioning the agent on behavior-specific speech. Experiments across robotic manipulation tasks and human-AI collaboration games demonstrate that MIMIC significantly enhances both behavior diversity and fidelity to human demonstrations while enabling nuanced behavioral steering without additional training.



Paperid:5234
Authors:Enshu Liu, Qian Chen, Xuefei Ning, Shengen Yan, Guohao Dai, Zinan Lin, Yu Wang
Abstract:
Abstract:Image Auto-regressive (AR) models have emerged as a powerful paradigm of visual generative models. Despite their promising performance, they suffer from slow generation speed due to the large number of sampling steps required. Although Distilled Decoding 1 (DD1) was recently proposed to enable few-step sampling for image AR models, it still incurs significant performance degradation in the one-step setting. In this work, we propose a new method, Distilled Decoding 2 (DD2), to further advances the feasibility of one-step sampling for image AR models without relying on a pre-defined mapping. We view the original AR model as a teacher model which provides the ground truth conditional score in the latent embedding space at each token position. Based on this, we propose a novel \emph{conditional score distillation loss} to train a one-step generator. Specifically, we train a separate network to predict the conditional score of the generated distribution and apply score distillation at every token position conditioned on previous tokens. Experimental results show that DD2 enables one-step sampling for image AR models with an minimal FID increase from 3.40 to 5.43 on ImageNet-256. Compared to the strongest baseline DD1, DD2 reduces the gap between the one-step sampling and original AR model by 67\%, with up to 12.3$\times$ training speed-up simultaneously. DD2 takes a significant step toward the goal of one-step AR generation, opening up new possibilities for fast and high-quality AR modeling.



Paperid:5235
Authors:Jiahao Wen, Hang Yu, Zhedong Zheng
Abstract:
Visual geo-localization for drones faces critical degradation under weather perturbations, \eg, rain and fog, where existing methods struggle with two inherent limitations: 1) Heavy reliance on limited weather categories that constrain generalization, and 2) Suboptimal disentanglement of entangled scene-weather features through pseudo weather categories.We present WeatherPrompt, a multi-modality learning paradigm that establishes weather-invariant representations through fusing the image embedding with the text context. Our framework introduces two key contributions: First, a Training-free Weather Reasoning mechanism that employs off-the-shelf large multi-modality models to synthesize multi-weather textual descriptions through human-like reasoning. It improves the scalability to unseen or complex weather, and could reflect different weather strength. Second, to better disentangle the scene and weather feature, we propose a multi-modality framework with the dynamic gating mechanism driven by the text embedding to adaptively reweight and fuse visual features across modalities. The framework is further optimized by the cross-modal objectives, including image-text contrastive learning and image-text matching, which maps the same scene with different weather conditions closer in the respresentation space. Extensive experiments validate that, under diverse weather conditions, our method achieves competitive recall rates compared to state-of-the-art drone geo-localization methods. Notably, it improves Recall@1 by +13.37\% under night conditions and by 18.69\% under fog and snow conditions.



Paperid:5236
Authors:Kazuya Nishimura, Haruka Hirose, Ryoma Bise, Kaito Shiku, Yasuhiro Kojima
Abstract:
Gene expression estimation from pathology images has the potential to reduce the RNA sequencing cost.Point-wise loss functions have been widely used to minimize the discrepancy between predicted and absolute gene expression values. However, due to the complexity of the sequencing techniques and intrinsic variability across cells, the observed gene expression contains stochastic noise and batch effects, and estimating the absolute expression values accurately remains a significant challenge.To mitigate this, we propose a novel objective that leverages relative expression patterns rather than absolute levels. We assume that the relative expression levels of genes exhibit consistent patterns across independent experiments, even when absolute expression values are affected by batch effects and stochastic noise in tissue samples.Based on the assumption, we model the relation and propose a novel loss function called STRank that is robust to noise and batch effects.Experiments using synthetic datasets and real datasets demonstrate the effectiveness of the proposed method.The code will be made available upon publication.



Paperid:5237
Authors:Thomas Klein, Sascha Meyen, Wieland Brendel, Felix A. Wichmann, Kristof Meding
Abstract:
Benchmarking models is a key factor for the rapid progress in machine learning (ML) research. Thus, further progress depends on improving benchmarking metrics. A standard metric to measure the behavioral alignment between ML models and human observers is error consistency (EC). EC allows for more fine-grained comparisons of behavior than other metrics such as e.g. accuracy, and has been used in the influential Brain-Score benchmark to rank different DNNs by their behavioral consistency with humans. Previously, EC values have been reported without confidence intervals. However, empirically measured EC values are typically noisy - thus, without confidence intervals, valid benchmarking conclusions are problematic. Here we improve on standard EC in two ways: First, we show how to obtain confidence intervals for EC using a bootstrapping technique, allowing us to derive significance tests for EC. Second, we propose a new computational model relating the EC between two classifiers to the implicit probability that one of them copies responses from the other. This view of EC allows us to give practical guidance to scientists regarding the number of trials required for sufficiently powerful, conclusive experiments.Finally, we use our methodology to revisit popular NeuroAI-results. We find that while the general trend of behavioral differences between humans and machines holds up to scrutiny, many reported differences between deep vision models are statistically insignificant. Our methodology enables researchers to design adequately powered experiments that can reliably detect behavioral differences between models, providing a foundation for more rigorous benchmarking of behavioral alignment.



Paperid:5238
Authors:Yuki Minai, Joana Soldado-Magraner, Byron M Yu, Matthew Smith
Abstract:
The coordinated activity of neural populations underlies myriad brain functions. Manipulating this activity using brain stimulation techniques has great potential for scientific and clinical applications, as it provides a tool to causally influence brain function. The state of the brain affects how neural populations respond to incoming sensory stimuli. Thus, taking into account pre-stimulation neural population activity may be crucial to achieve a desired causal manipulation using stimulation. In this work, we propose Online MicroStimulation Optimization (OMiSO), a brain stimulation framework that leverages brain state information to find stimulation parameters that can drive neural population activity toward specified states. OMiSO includes two key advances: i) it leverages the pre-stimulation brain state to choose optimal stimulation parameters, and ii) it adaptively refines the choice of those parameters by considering newly-observed stimulation responses. We tested OMiSO by applying intracortical electrical microstimulation in a monkey and found that it outperformed competing methods that do not incorporate these advances. Taken together, OMiSO provides greater accuracy in achieving specified activity states, thereby advancing neuromodulation technologies for understanding the brain and for treating brain disorders.



Paperid:5239
Authors:Xinghao Wu, Xuefeng Liu, Jianwei Niu, Guogang Zhu, Mingjia Shi, Shaojie Tang, Jing Yuan
Abstract:
Federated Learning (FL) faces challenges due to data heterogeneity, which limits the global model’s performance across diverse client distributions. Personalized Federated Learning (PFL) addresses this by enabling each client to process an individual model adapted to its local distribution. Many existing methods assume that certain global model parameters are difficult to train effectively in a collaborative manner under heterogeneous data. Consequently, they localize or fine-tune these parameters to obtain personalized models. In this paper, we reveal that both the feature extractor and classifier of the global model are inherently strong, and the primary cause of its suboptimal performance is the mismatch between local features and the global classifier. Although existing methods alleviate this mismatch to some extent and improve performance, we find that they either (1) fail to fully resolve the mismatch while degrading the feature extractor, or (2) address the mismatch only post-training, allowing it to persist during training. This increases inter-client gradient divergence, hinders model aggregation, and ultimately leaves the feature extractor suboptimal for client data. To address this issue, we propose FedPFT, a novel framework that resolves the mismatch during training using personalized prompts. These prompts, along with local features, are processed by a shared self-attention-based transformation module, ensuring alignment with the global classifier. Additionally, this prompt-driven approach offers strong flexibility, enabling task-specific prompts to incorporate additional training objectives (\eg, contrastive learning) to further enhance the feature extractor. Extensive experiments show that FedPFT outperforms state-of-the-art methods by up to 5.07%, with further gains of up to 7.08% when collaborative contrastive learning is incorporated.



Paperid:5240
Authors:Junjie Zhou, WEI SHAO, Yagao Yue, Wei Mu, Peng Wan, Qi Zhu, Daoqiang Zhang
Abstract:
Prompt learning has emerged as a promising paradigm for adapting pre-trained vision-language models (VLMs) to few-shot whole slide image (WSI) classification by aligning visual features with textual representations, thereby reducing annotation cost and enhancing model generalization. Nevertheless, existing methods typically rely on slide-level prompts and fail to capture the subtype-specific phenotypic variations of histological entities (e.g., nuclei, glands) that are critical for cancer diagnosis. To address this gap, we propose Multi-scale Attribute-enhanced Prompt Learning (MAPLE), a hierarchical framework for few-shot WSI classification that jointly integrates multi-scale visual semantics and performs prediction at both the entity and slide levels. Specifically, we first leverage large language models (LLMs) to generate entity-level prompts that can help identify multi-scale histological entities and their phenotypic attributes, as well as slide-level prompts to capture global visual descriptions. Then, an entity-guided cross-attention module is proposed to generate entity-level features, followed by aligning with their corresponding subtype-specific attributes for fine-grained entity-level prediction. To enrich entity representations, we further develop a cross-scale entity graph learning module that can update these representations by capturing their semantic correlations within and across scales. The refined representations are then aggregated into a slide-level representation and aligned with the corresponding prompts for slide-level prediction. Finally, we combine both entity-level and slide-level outputs to produce the final prediction results. Results on three cancer cohorts confirm the effectiveness of our approach in addressing few-shot pathology diagnosis tasks.



Paperid:5241
Authors:Zhaoxian Wu, Quan Xiao, Tayfun Gokmen, Omobayode Fagbohungbe, Tianyi Chen
Abstract:
As the economic and environmental costs of training and deploying large vision or language models increase dramatically, analog in-memory computing (AIMC) emerges as a promising energy-efficient solution. However, the training perspective, especially its training dynamic, is underexplored. In AIMC hardware, the trainable weights are represented by the conductance of resistive elements and updated using consecutive electrical pulses. While the conductance changes by a constant in response to each pulse, in reality, the change is scaled by asymmetric and non-linear response functions, leading to a non-ideal training dynamic. This paper provides a theoretical foundation for gradient-based training on AIMC hardware with non-ideal response functions. We demonstrate that asymmetric response functions negatively impact Analog SGD by imposing an implicit penalty on the objective. To overcome the issue, we propose residual learning algorithm, which provably converges exactly to a critical point by solving a bilevel optimization problem. We show that the proposed method can be extended to deal with other hardware imperfections like limited response granularity. As far as we know, it is the first paper to investigate the impact of a class of generic non-ideal response functions. The conclusion is supported by simulations validating our theoretical insights.



Paperid:5242
Authors:Haozhe Shan, Sun Minni, Lea Duncker
Abstract:
The ability to continually learn new skills, retain, and flexibly deploy them to accomplish goals is a key feature of intelligent and efficient behavior. However, the neural mechanisms facilitating the continual learning and flexible (re-)composition of skills remain elusive. Here, we study continual learning and the compositional reuse of learned computations in recurrent neural network (RNN) models using a novel two-system approach: one system that infers 'what' computation to perform, and one that implements 'how' to perform it. We focus on a set of compositional cognitive tasks commonly studied in neuroscience. To construct the 'what' system, we first show that a large family of tasks can be systematically described by a probabilistic generative model, where compositionality stems from a shared underlying vocabulary of discrete task-epochs. The shared epoch structure makes these tasks inherently compositional. We first show that this compositionality can be systematically described by a probabilistic generative model. Furthermore, we develop an unsupervised online learning approach that can learn this model on a single-trial basis, building its vocabulary incrementally as it is exposed to new tasks, and inferring the latent epoch structure as a time-varying computational context within a trial. We implement the 'how' system as an RNN whose low-rank components are composed according to the context inferred by the 'what' system.The contextual inference facilitates the creation, learning, and reuse of the low-rank RNN components as new tasks are introduced sequentially, enabling continual learning without catastrophic forgetting. Using an example task set, we demonstrate the efficacy and competitive performance of this two-system learning framework, its potential for forward and backward transfer, as well as few-shot learning via re-composition.



Paperid:5243
Authors:Nicholas Tacca, Bryan Schlink, Jackson Levine, Mary K. Heimann, Collin Dunlap, Sam Colachis, Philip Putnam, Matthew Zeglen, Daniel Brobston, Austin Bollinger, José Pons, Lauren Wengerd, Eric Meyers, David Friedenberg
Abstract:
Abstract:Extracting neural signals at the single motor neuron level provides an optimal control signal for neuroprosthetic applications. However, current algorithms to decompose motor units from high-density electromyography (HD-EMG) are time-consuming and inconsistent, limiting their application to controlled scenarios in a research setting. We introduce MUelim, an algorithm for efficient motor unit decomposition that uses approximate joint diagonalization with a subtractive approach to rapidly identify and refine candidate sources. The algorithm incorporates an extend-lag procedure to augment data for enhanced source separability prior to diagonalization. By systematically iterating and eliminating redundant or noisy sources, MUelim achieves high decomposition accuracy while significantly reducing computational complexity, making it well-suited for real-time applications. We validate MUelim by demonstrating its ability to extract motor units in both simulated and physiological HD-EMG grid data. We benchmark its performance against existing state-of-the-art decomposition algorithms, demonstrating up to a $\sim$3,700\% speed increase to decompose a similar number of sources with a high signal-to-noise ratio. Furthermore, we showcase a real-world application of MUelim in a clinical setting in which an individual with spinal cord injury controlled an EMG-driven neuroprosthetic to perform functional tasks. We demonstrate the ability to decode motor intent in real-time using a spiking neural network trained on the decomposed motor unit spike trains to trigger functional electrical stimulation patterns that evoke hand movements during task practice therapy. We show that the use of motor units enables nuanced motor control, highlighting the potential of MUelim to advance assistive neurotechnology and rehabilitation by enabling precise, closed-loop motor-unit-driven control to the user.



Paperid:5244
Authors:Nicolas Le Roux, Marc Bellemare, Jonathan Lebensold, Arnaud Bergeron, Joshua Greaves, Alexandre Fréchette, Carolyne Pelletier, Eric Thibodeau-Laufer, Sándor Tóth, Sam Work
Abstract:
We propose a new algorithm for fine-tuning large language models using reinforcement learning. Tapered Off-Policy REINFORCE (TOPR) uses an asymmetric, tapered variant of importance sampling to speed up learning while maintaining stable learning dynamics, even without the use of KL regularization. TOPR can be applied in a fully offline fashion, allows the handling of positive and negative examples in a unified framework, and benefits from the implementational simplicity that is typical of Monte Carlo algorithms. We demonstrate the effectiveness of our approach with a series of experiments on the GSM8K and MATH reasoning benchmarks, finding performance gains for training both a model for solution generation and as a generative verifier. We show that properly leveraging positive and negative examples alike in the off-policy regime simultaneously increases test-time accuracy and training data efficiency, all the while avoiding the ``wasted inference'' that comes with discarding negative examples. We find that this advantage persists over multiple iterations of training and can be amplified by dataset curation techniques, enabling us to match 70B-parameter model performance with 8B language models. As a corollary to this work, we find that REINFORCE's baseline parameter plays an important and unexpected role in defining dataset composition in the presence of negative examples, and is consequently critical in driving off-policy performance.



Paperid:5245
Authors:Bojia Zi, PENGHUI RUAN, Marco Chen, Xianbiao Qi, Shaozhe Hao, Shihao Zhao, Youze Huang, Bin Liang, Rong Xiao, Kam-Fai Wong
Abstract:
Video content editing has a wide range of applications. With the advancement of diffusion-based generative models, video editing techniques have made remarkable progress, yet they still remain far from practical usability. Existing inversion-based video editing methods are time-consuming and struggle to maintain consistency in unedited regions. Although instruction-based methods have high theoretical potential, they face significant challenges in constructing high-quality training datasets - current datasets suffer from issues such as editing correctness, frame consistency, and sample diversity. To bridge these gaps, we introduce theSeñorita-2Mdataset, a large-scale, diverse, and high-quality video editing dataset. We systematically categorize editing tasks into 2 classes consinsting of 18 subcategories. To build this dataset, we design four new task specialists and employ or modify 14 existing task experts to generate data samples for each subclass. In addition, we design a filtering pipeline at both the visual content and instruction levels to further enhance data quality. This approach ensures the reliability of constructed data. Finally, the Señorita-2M dataset comprises 2 million high-fidelity samples with diverse resolutions and frame counts. We trained multiple models using different base video models, i.e., Wan2.1 and CogVideoX-5B, on Señorita-2M, and the results demonstrate that the models exhibit superior visual quality, robust frame-to-frame consistency, and strong alignment with text instructions. More videos are available at:https://anonymous-senorita-2m.github.io.



Paperid:5246
Authors:Yihong Luo, Shuchen Xue, Tianyang Hu, Jing Tang
Abstract:
The pursuit of efficient and controllable high-quality content generation stands as a pivotal challenge in artificial intelligence-generated content (AIGC).While one-step generators, refined through diffusion distillation techniques, offer excellent generation quality and computational efficiency, adapting them to new control conditions—such as structural constraints, semantic guidelines, or external inputs—poses a significant challenge. Conventional approaches often necessitate computationally expensive modifications to the base model and subsequent diffusion distillation. This paper introduces Noise Consistency Training (NCT), a novel and lightweight approach to directly integrate new control signals into pre-trained one-step generators without requiring access to original training images or retraining the base diffusion model. NCT operates by introducing an adapter module and employs a noise consistency loss in the noise space of the generator. This loss aligns the adapted model's generation behavior across noises that are conditionally dependent to varying degrees, implicitly guiding it to adhere to the new control. This training objective can be interpreted as aligning the adapted generator with the intractable conditional distribution defined by a discriminative model and the one-step generator from moment-matching perspectives. NCT is modular, data-efficient, and easily deployable, relying only on the pre-trained one-step generator and a control signal model. Extensive experiments demonstrate that NCT achieves state-of-the-art controllable generation in a single forward pass, surpassing existing multi-step and distillation-based methods in both generation quality and computational efficiency.



Paperid:5247
Authors:Xiangdong Zhang, Jiaqi Liao, Shaofeng Zhang, Fanqing Meng, Xiangpeng Wan, Junchi Yan, Yu Cheng
Abstract:
Recent advancements in text-to-video (T2V) diffusion models have enabled high-fidelity and realistic video synthesis. However, current T2V models often struggle to generate physically plausible content due to their limited inherent ability to accurately understand physics. We found that while the representations within T2V models possess some capacity for physics understanding, they lag significantly behind those from recent video self-supervised learning methods. To this end, we propose a novel framework called {VideoREPA}, which distills physics understanding capability from video understanding foundation models into T2V models by aligning token-level relations. This closes the physics understanding gap and enable more physics-plausible generation. Specifically, we introduce the {Token Relation Distillation (TRD) loss}, leveraging spatio-temporal alignment to provide soft guidance suitable for finetuning powerful pre-trained T2V models—a critical departure from prior representation alignment (REPA) methods. To our knowledge, VideoREPA is the first REPA method designed for finetuning T2V models and specifically for injecting physical knowledge. Empirical evaluations show that VideoREPA substantially enhances the physics commonsense of baseline method, CogVideoX, achieving significant improvement on relevant benchmarks and demonstrating a strong capacity for generating videos consistent with intuitive physics. More video results are available at this anonymous link: https://anonymous.4open.science/r/VideoREPA-Video-Generation-26EB.



Paperid:5248
Authors:Rui Tian, Mingfei Gao, Mingze Xu, Jiaming Hu, Jiasen Lu, Zuxuan Wu, Yinfei Yang, Afshin Dehghan
Abstract:
We introduce UniGen, a unified multimodal large language model (MLLM) capable of image understanding and generation. We study the full training pipeline of UniGen from a data-centric perspective, including multi-stage pre-training, supervised fine-tuning, and direct preference optimization. More importantly, we propose a new Chain-of-Thought Verification (CoT-V) strategy for test-time scaling, which significantly boosts UniGen's image generation quality using a simple Best-of-N test-time strategy. Specifically, CoT-V enables UniGen to act as both image generator and verifier at test time, assessing the semantic alignment between a text prompt and its generated image in a step-by-step CoT manner. Trained entirely on open-source datasets across all stages, UniGen achieves state-of-the-art performance on a range of image understanding and generation benchmarks, with a final score of 0.78 on GenEval and 85.19 on DPG-Bench. Through extensive ablation studies, our work provides actionable insights and addresses key challenges in the full life cycle of building unified MLLMs, contributing meaningful directions to the future research.



Paperid:5249
Authors:Guangyi Chen, Yunlong Deng, Peiyuan Zhu, Yan Li, Yifan Shen, Zijian Li, Kun Zhang
Abstract:
Causal Representation Learning (CRL) aims to uncover the data generating process and identify the underlying causal variables and relations, or finding suitable abstractions of micro variables. In this paper, we focus on the former type of CRL, in which evaluation of CRL methods remains inherently challenging due to the requirement of known ground-truth causal variables and causal structure. Existing evaluations often rely on either simplistic synthetic datasets or downstream performance on real-world tasks, generally suffering a dilemma between realism and evaluative precision. In this paper, we introduce a new benchmark for CRL using high-fidelity simulated visual data that retains both realistic visual complexity and, more importantly, access to ground-truth causal generating processes. The dataset comprises around 200 thousand images and 3 million video frames across 24 sub-scenes in four domains: static image generation, dynamic physical simulations, robotic manipulations, and traffic situation analysis. These scenarios range from static to dynamic settings, simple to complex structures, and single- to multi-agent interactions, offering a comprehensive testbed that hopefully bridges the gap between rigorous evaluation and real-world applicability. In addition, we provide flexible access to the underlying causal structures, allowing users to modify or configure them to align with the required assumptions in CRL, such as available domain labels, temporal dependencies, or intervention histories. Leveraging this benchmark, we evaluated representative CRL methods across diverse paradigms and offered empirical insights to assist practitioners and newcomers in choosing or extending appropriate CRL frameworks to properly address specific types of real problems that can benefit from the CRL perspective. Our data is open source at: https://huggingface.co/CausalVerse



Paperid:5250
Authors:Ruiqi Xue, Ziqian Zhang, Lihe Li, Cong Guan, Lei Yuan, Yang Yu
Abstract:
Learning safe reinforcement learning (RL) policies from offline multi-task datasets without direct environmental interaction is crucial for efficient and reliable deployment of RL agents. Benefiting from their scalability and strong in-context learning capabilities, recent approaches attempt to utilize Decision Transformer (DT) architectures for offline safe RL, demonstrating promising adaptability across varying safety budgets. However, these methods primarily focus on single-constraint scenarios and struggle with diverse constraint configurations across multiple tasks. Additionally, their reliance on heuristically defined Return-To-Go (RTG) inputs limits flexibility and reduces learning efficiency, particularly in complex multi-task environments. To address these limitations, we propose CoPDT, a novel DT-based framework designed to enhance adaptability to diverse constraints and varying safety budgets. Specifically, CoPDT introduces a constraint prioritized prompt encoder, which leverages sparse binary cost signals to accurately identify constraints, and a constraint prioritized Return-To-Go (CPRTG) token mechanism, which dynamically generates RTGs based on identified constraints and corresponding safety budgets. Extensive experiments on the OSRL benchmark demonstrate that CoPDT achieves superior efficiency and significantly enhanced safety compliance across diverse multi-task scenarios, surpassing state-of-the-art DT-based methods by satisfying safety constraints in more than twice as many tasks.



Paperid:5251
Authors:Fengmao Lyu, Jitong Lei, Guosheng Lin, Desheng ZHENG, Jianyang Zhang, Tianrui Li
Abstract:
While existing research on Multilingual CLIP (MCLIP) has prioritized model architecture design, our work uncovers a critical challenge in practical adaptation: fine-tuning MCLIP through a single source language risks diminishing its multilingual capabilities in downstream tasks due to cross-linguistic disparities. To bridge this gap, we systematically investigate the role of token similarity in cross-lingual transferability for image-text retrieval, establishing it as a key factor governing fine-tuning efficacy. Building on this insight, we propose two novel strategies to enhance efficiency while preserving multilinguality: 1) TaPCL dynamically optimizes training by prioritizing linguistically distant language pairs during corpus sampling, reducing redundant computation, and 2) CiPCL enriches the source corpus with multilingual key terms, enabling targeted knowledge transfer without reliance on exhaustive parallel data. By strategically balancing token similarity and domain-critical information, our methods significantly lower computational costs and mitigate over-dependence on parallel corpora. Experimental evaluations across diverse datasets validate the effectiveness and scalability of our framework, demonstrating robust multilingual retention across languages. This work provides a principled pathway for adapting MCLIP to real-world scenarios, where computational efficiency and cross-lingual robustness are paramount.



Paperid:5252
Authors:Hang Hua, Ziyun Zeng, Yizhi Song, Yunlong Tang, Liu He, Daniel Aliaga, Wei Xiong, Jiebo Luo
Abstract:
Recent multimodal image generators such as GPT-4o, Gemini 2.0 Flash, and Gemini 2.5 Pro excel at following complex instructions, editing images and maintaining concept consistency. However, they are still evaluated bydisjointtoolkits: text-to-image (T2I) benchmarks that lacks multi-modal conditioning, and customized image generation benchmarks that overlook compositional semantics and common knowledge. We proposeMMGen-Bench, acomprehensiveMulti-Modal imageGenerationBenchmark that unifies these tasks by pairing 4,850 richly annotated text prompts with 1,750 multi-view reference images across 380 subjects, spanning humans, animals, objects, and artistic styles. MMGen-Bench is equipped with a three-level evaluation framework: (1) low-level metrics for visual artifacts and identity preservation of objects; (2) novel Aspect Matching Score (AMS): a VQA-based mid-level metric that delivers fine-grained prompt-image alignment and shows strong correlation with human judgments; and (3) high-level metrics for aesthetics and human preference. Using MMGen-Bench, we benchmark 17 state-of-the-art models, including Gemini 2.5 Pro, FLUX, DreamBooth, and IP-Adapter, and validate our metrics with 32k human ratings, yielding in-depth insights into architecture and data design. We will release the dataset and evaluation code to foster rigorous, unified evaluation and accelerate future innovations in multi-modal image generation.



Paperid:5253
Authors:Ilan Doron-Arad
Abstract:
Abstract:Recent years have revealed an unprecedented demand for AI-based technology, leading to a common setting where immense data is distributed across multiple locations. This creates a communication bottleneck among the storage facilities, often aiming to jointly solve tasks of small solution size $k$ from input of astronomically large size $n$. Motivated by federated and distributed machine learning applications, we study two fundamental optimization problems, maximum weight matroid independent set (MW-IS) and maximum weight matching (MWM), in a zero communication computational model. In this model, the data is dispersed between $m$ servers. Without any communication, each server has to send a message to a central server, which is required to compute an optimal solution for the original (large) instance. The goal is to minimize the size of the maximum message sent. For this natural restrictive model, we obtain deterministic algorithms that use $k$-data per server for MW-IS and $O(k^2)$-data per server for MWM, where $k$ is the solution size. We complement these results with tight lower bounds -- ruling out any asymptotic improvement even if randomization is allowed. Our algorithms are simple and run in nearly linear time. Interestingly, we show how our zero communication algorithms yield deterministic parallel algorithms with running times $O\left(\sqrt{k} \cdot \log n\right)$ and $O\left(k^4 \cdot \log n\right)$ for MW-IS and MWM, respectively.



Paperid:5254
Authors:Tonglong Wei, Yan Lin, Youfang Lin, Shengnan Guo, Jilin Hu, Haitao Yuan, Gao Cong, Huaiyu Wan
Abstract:
Abstract:Spatiotemporal trajectory data is crucial for various traffic-related applications. However, issues such as device malfunctions and network instability often result in sparse trajectories that lose detailed movement information compared to their dense counterparts. Recovering missing points in sparse trajectories is thus essential. Despite recent progress, three challenges remain. First, the lack of large-scale dense trajectory datasets hinders the training of a trajectory recovery model. Second, the varying spatiotemporal correlations in sparse trajectories make it hard to generalize across different sampling intervals. Third, extracting road conditions for missing points is non-trivial.To address these challenges, we propose $\textit{PLMTrajRec}$, a novel trajectory recovery model. It leverages the scalability of a pre-trained language model (PLM) and can effectively recover trajectories by fine-tuning with small-scale dense trajectory datasets. To handle different sampling intervals in sparse trajectories, we first convert sampling intervals and movement features into prompts for the PLM to understand. We then introduce a trajectory encoder to unify trajectories of varying intervals into a single interval. To extract road conditions for missing points, we propose an area flow-guided implicit trajectory prompt that represents traffic conditions in each region, and a road condition passing mechanism that infers the road conditions of missing points using the observed ones. Experiments on four public trajectory datasets with three sampling intervals demonstrate the effectiveness, scalability, and generalization ability of PLMTrajRec. Code is available at https://anonymous.4open.science/r/PLMTrajRec-D468.



Paperid:5255
Authors:Yiwen Zhu, Jinyi Liu, Pengjie Gu, Yifu Yuan, Zhenxing Ge, Wenya Wei, Zhou Fang, Yujing Hu, Bo An
Abstract:
Reinforcement learning (RL) heavily depends on well-designed reward functions, which are often biased and difficult to design for complex behaviors. Preference-based RL (PbRL) addresses this by learning reward models from human feedback, but its practicality is constrained by a critical dilemma: while existing methods reduce human effort through query optimization, they neglect the preference buffer's restricted coverage —a factor that fundamentally determines the reliability of reward model. We systematically demonstrate this limitation creates distributional mismatch: reward models trained on static buffers reliably assess in-distribution trajectories but falter with out-of-distribution (OOD) trajectories from policy exploration. Crucially, such failures in policy-proximal regions directly misguide iterative policy updates.To address this, we proposeProximal Policy Exploration (PPE)with two key components:(1) aproximal-policy extensionmethod that expands exploration in undersampled policy-proximal regions, and(2) amixture distribution querymethod that balances in-distribution and OOD trajectory sampling.By enhancing buffer coverage while preserving evaluation accuracy in policy-proximal regions, PPE enables more reliable policy updates.Experiments across continuous control tasks demonstrate that PPE enhances preference feedback utilization efficiency and RL sample efficiency over baselines, highlighting preference buffer coverage management's vital role in PbRL.



Paperid:5256
Authors:Agam Shah, Siddhant Sukhani, Huzaifa Pardawala, Saketh Budideti, Riya Bhadani, Rudra Gopal, Siddhartha Somani, Michael Galarnyk, Soungmin Lee, Arnav Hiray, Akshar Ravichandran, Eric Kim, Pranav Aluru, Joshua Zhang, Sebastian Jaskowski, Veer Guda, Meghaj Tarte, Liqin Ye, Spencer Gosden, Rutwik Routu, Rachel Yuh, Sloka Chava, Sahasra Chava, Dylan Kelly, Aiden Chiang, Harsit Mittal, Sudheer Chava
Abstract:
Central banks around the world play a crucial role in maintaining economic stability. Deciphering policy implications in their communications is essential, especially as misinterpretations can disproportionately impact vulnerable populations. To address this, we introduce the World Central Banks (WCB) dataset, the most comprehensive monetary policy corpus to date, comprising over 380k sentences from 25 central banks across diverse geographic regions, spanning 28 years of historical data. After uniformly sampling 1k sentences per bank (25k total) across all available years, we annotate and review each sentence using dual annotators, disagreement resolutions, and secondary expert reviews. We define three tasks: Stance Detection, Temporal Classification, and UncertaintyEstimation, with each sentence annotated for all three. We benchmark seven Pretrained Language Models (PLMs) and nine Large Language Models (LLMs) (Zero-Shot, Few-Shot, and with annotation guide) on these tasks, running 15,075 benchmarking experiments. We find that a model trained on aggregated data across banks significantly surpasses a model trained on an individual bank's data, confirming the principle"the whole is greater than the sum of its parts."Additionally, rigorous human evaluations, error analyses, and predictive tasks validate our framework's economic utility. Our artifacts are accessible through the HuggingFace and GitHub under the CC-BY-NC-SA 4.0 license.



Paperid:5257
Authors:Fali Wang, Hui Liu, Zhenwei Dai, Jingying Zeng, Zhiwei Zhang, Zongyu Wu, Chen Luo, Zhen Li, Xianfeng Tang, Qi He, Suhang Wang
Abstract:
Test-time scaling (TTS) enhances the performance of large language models (LLMs) by allocating additional compute resources during inference. However, existing research primarily investigates TTS in single-stage tasks; while many real-world problems are multi-stage complex tasks, composed of a sequence of heterogeneous subtasks with each subtask requires LLM of specific capability. Therefore, we study a novel problem: the test-time compute-optimal scaling in multi-stage complex tasks, aiming to select suitable models and allocate budgets per subtask to maximize overall performance. TTS in multi-stage tasks introduces two fundamental challenges: (i) The combinatorial search space of model and budget allocations, combined with the high cost of inference, makes brute-force search impractical. (ii) The optimal model and budget allocations across subtasks are interdependent, increasing the complexity of the compute-optimal search. To address this gap, we conduct extensive pilot experiments on four tasks across six datasets, deriving three empirical insights characterizing the behavior of LLMs in multi-stage complex tasks. Informed by these insights, we propose AgentTTS, an LLM-agent-based framework that autonomously searches for compute-optimal allocations through iterative feedback-driven interactions with the execution environment. Experimental results demonstrate that AgentTTS significantly outperforms traditional and other LLM-based baselines in search efficiency, and shows improved robustness to varying training set sizes and enhanced interpretability.



Paperid:5258
Authors:Zhixuan Pan, Shaowen Wang, Liao Pengfei, Jian Li
Abstract:
Large Language Models (LLMs) have demonstrated remarkable capabilities across numerous tasks, yet principled explanations for their underlying mechanisms and several phenomena, such as scaling laws, hallucinations, and related behaviors, remain elusive. In this work, we revisit the classical relationship between compression and prediction, grounded in Kolmogorov complexity and Shannon information theory, to provide deeper insights into LLM behaviors. By leveraging the Kolmogorov Structure Function and interpreting LLM compression as a two-part coding process, we offer a detailed view of how LLMs acquire and store information across increasing model and data scales -- from pervasive syntactic patterns to progressively rarer knowledge elements. Motivated by this theoretical perspective and natural assumptions inspired by Heap’s and Zipf’s laws, we introduce a simplified yet representative hierarchical data-generation framework called the Syntax-Knowledge model. Under the Bayesian setting, we show that prediction and compression within this model naturally lead to diverse learning and scaling behaviors of LLMs. In particular, our theoretical analysis offers intuitive and principled explanations for both data and model scaling laws, the dynamics of knowledge acquisition during training and fine-tuning, factual knowledge hallucinations in LLMs. The experimental results validate our theoretical predictions.



Paperid:5259
Authors:Geng Zhang, Jiangrong Shen, Kaizhong Zheng, Liangjun Chen, Badong Chen
Abstract:
Brain disorders have been consistently associated with abnormalities in specific brain regions or neural circuits. Identifying key brain regional activities and functional connectivity patterns is essential for discovering more precise neurobiological biomarkers. However, previous studies have primarily emphasized alterations in functional connectivity while overlooking abnormal neuronal population activity within brain regions. To bridge this gap, we propose a novel Local-Global Coupling Spiking Graph Transformer (LGC-SGT) that jointly models both inter-regional connectivity differences and deviations in neuronal population firing rates within brain regions, enabling a dual-perspective neuropathological analysis. The global pathway leverages spike-based computation in LGC-SGT to model biologically plausible aberrant neural firing dynamics, while the local pathway adaptively captures abnormal graph-based representations of brain connectivity learned by local plasticity in the liquid state machine module. Furthermore, we design a shortcut-enhanced output strategy in LGC-SGT with the hybrid loss function to suppress outlier interference caused by inter-individual and inter-center variability, enabling a more robust decision boundary. Extensive experiments on three brain disorder datasets demonstrate that our model consistently outperforms state-of-the-art graph methods in brain disorder diagnosis. Moreover, it facilitates the extraction of interpretable neurobiological biomarkers by jointly analyzing regional neural activity and functional connectivity, offering a more comprehensive framework for brain disorder understanding and diagnosis.



Paperid:5260
Authors:Zenan Ying, Zhi Zheng, huijun hou, Tong Xu, Qi Liu, Jinke Wang, Wei Chen
Abstract:
Multivariate time series (MTS) classification has attracted increasing attention across various domains. Existing methods either decompose MTS into separate univariate series, ignoring inter-variable dependencies, or jointly model all variables, which may lead to over-smoothing and loss of semantic structure. These limitations become particularly pronounced when dealing with complex and heterogeneous variable types.To address these challenges, we propose SwinGroupNet (SGN), which explores a novel perspective for constructing variable interaction and temporal dependency.Specifically, SGN processes multi-scale time series using (1)Variable Group Embedding (VGE), which partitions variables into groups and performs independent group-wise embedding; (2) Multi-Scale Group Window Mixing (MGWM), which reconstructs variable interactions by modeling both intra-group and inter-group dependencies while extracting multi-scale temporal features; and (3) Periodic Window Shifting and Merging (PWSM), which exploits inherent periodic patterns to enable hierarchical temporal interaction and feature aggregation. Extensive experiments on diverse benchmark datasets from multiple domains demonstrate that SGN consistently achieves state-of-the-art performance, with an average improvement of 4.2% over existing methods. We release the source code at https://anonymous.4open.science/r/SGN.



Paperid:5261
Authors:Weiyi Wang, Junwei Deng, Yuzheng Hu, Shiyuan Zhang, Xirui Jiang, Runting Zhang, Han Zhao, Jiaqi Ma
Abstract:
Data attribution methods, which quantify the influence of individual training data points on a machine learning model, have gained increasing popularity in data-centric applications in modern AI. Despite a recent surge of new methods developed in this space, the impact of hyperparameter tuning in these methods remains under-explored. In this work, we present the first large-scale empirical study to understand the hyperparameter sensitivity of common data attribution methods. Our results show that most methods are indeed sensitive to certain key hyperparameters. However, unlike typical machine learning algorithms---whose hyperparameters can be tuned using computationally-cheap validation metrics---evaluating data attribution performance often requires retraining models on subsets of training data, making such metrics prohibitively costly for hyperparameter tuning. This poses a critical open challenge for the practical application of data attribution methods. To address this challenge, we advocate for better theoretical understandings of hyperparameter behavior to inform efficient tuning strategies. As a case study, we provide a theoretical analysis of the regularization term that is critical in many variants of influence function methods. Building on this analysis, we propose a lightweight procedure for selecting the regularization value without model retraining, and validate its effectiveness across a range of standard data attribution benchmarks. Overall, our study identifies a fundamental yet overlooked challenge in the practical application of data attribution, and highlights the importance of careful discussion on hyperparameter selection in future method development.



Paperid:5262
Authors:Ali Rasekh, Erfan Soula, Omid Daliran, Simon Gottschalk, Mohsen Fayyaz
Abstract:
Despite significant advances in Multimodal Large Language Models (MLLMs), understanding complex temporal dynamics in videos remains a major challenge. Our experiments show that current Video Large Language Model (Video-LLM) architectures have critical limitations in temporal understanding, struggling with tasks that require detailed comprehension of action sequences and temporal progression. In this work, we propose a Video-LLM architecture that introduces stacked temporal attention modules directly within the vision encoder. This design incorporates a temporal attention in vision encoder, enabling the model to better capture the progression of actions and the relationships between frames before passing visual tokens to the LLM. Our results show that this approach significantly improves temporal reasoning and outperforms existing models in video question answering tasks, specifically in action recognition. We improve on benchmarks including VITATECS, MVBench, and Video-MME by up to +5.5\%. By enhancing the vision encoder with temporal structure, we address a critical gap in video understanding for Video-LLMs. The code will be made available online upon publication.



Paperid:5263
Authors:Shenzhi Yang, Junbo Zhao, Sharon Li, Shouqing Yang, Dingyu Yang, Xiaofang Zhang, Haobo Wang
Abstract:
Out-of-distribution (OOD) node detection in graphs is a critical yet challenging task. Most existing approaches rely heavily on fine-grained labeled data to obtain a pre-trained supervised classifier, inherently assuming the existence of a well-defined pretext classification task. However, when such a task is ill-defined or absent, their applicability becomes severely limited. To overcome this limitation, there is an urgent need to propose a more scalable OOD detection method that is independent of both pretext tasks and label supervision. We harness a new phenomenon calledFeature Resonance, focusing on the feature space rather than the label space. We observe that, ideally, during the optimization of known ID samples, unknown ID samples undergo more significant representation changes than OOD samples, even when the model is trained to align arbitrary targets. The rationale behind it is that even without gold labels, the local manifold may still exhibit smooth resonance. Based on this, we further develop a novel graph OOD framework, dubbedResonance-basedSeparation andLearning (RSL), which comprises two core modules: (i)-a more practical micro-level proxy of feature resonance that measures the movement of feature vectors in one training step. (ii)-integrate with a synthetic OOD node strategy to train an effective OOD classifier. Theoretically, we derive an error bound showing the superior separability of OOD nodes during the resonance period. Extensive experiments on a total of thirteen real-world graph datasets empirically demonstrate that RSL achieves state-of-the-art performance.



Paperid:5264
Authors:Peiyan Li, Yixiang Chen, Hongtao Wu, Xiao Ma, Xiangnan Wu, Yan Huang, Liang Wang, Tao Kong, Tieniu Tan
Abstract:
Recently, leveraging pre-trained vision-language models (VLMs) for building vision-language-action (VLA) models has emerged as a promising approach to effective robot manipulation learning. However, only few methods incorporate 3D signals into VLMs for action prediction, and they do not fully leverage the spatial structure inherent in 3D data, leading to low sample efficiency. In this paper, we introduce BridgeVLA, a novel 3D VLA model that (1) projects 3D inputs to multiple 2D images, ensuring input alignment with the VLM backbone, and (2) utilizes 2D heatmaps for action prediction, unifying the input and output spaces within a consistent 2D image space.In addition, we propose a scalable pre-training method that equips the VLM backbone with the capability to predict 2D heatmaps before downstream policy learning. Extensive experiments show the proposed method is able to learn 3D manipulation efficiently and effectively.BridgeVLA surpasses the state-of-the-art baseline method in RLBench, achieving a significant higher success rate (88.2% vs 81.4%), and in COLOSSEUM, demonstrating a substantially lower success rate drop (3.6% vs 15.6%). In real-robot experiments, BridgeVLA outperforms the state-of-the-art baseline method by 32% on average, and is able to generalize robustly in multiple out-of-distribution settings, including visual disturbance and unseen instructions. Remarkably, it is able to achieve a success rate of 96.8% on 10+ tasks with only 3 trajectories per task, highlighting its extraordinary sample efficiency. Videos can be found in https://anonymous1219-create.github.io/BridgeVLA_Web/.



Paperid:5265
Authors:Vivek Gopalakrishnan, Neel Dey, Polina Golland
Abstract:
Determining the 3D pose of a patient from a limited set of 2D X-ray images is a critical task in interventional settings. While preoperative volumetric imaging (e.g., CT and MRI) provides precise 3D localization and visualization of anatomical targets, these modalities cannot be acquired during procedures, where fast 2D imaging (X-ray) is used instead. To integrate volumetric guidance into intraoperative procedures, we present PolyPose, a simple and robust method for deformable 2D/3D registration. PolyPose parameterizes complex 3D deformation fields as a composition of rigid transforms, leveraging the biological constraint that individual bones do not bend in typical motion. Unlike existing methods that either assume no inter-joint movement or fail outright in this underdetermined setting, our polyrigid formulation enforces anatomically plausible priors that respect the piecewise rigid nature of human movement. This approach eliminates the need for expensive deformation regularizers that require patient- and procedure-specific hyperparameter optimization. Across extensive experiments on diverse datasets from orthopedic surgery and radiotherapy, we show that this strong inductive bias enables PolyPose to successfully align the patient's preoperative volume to as few as two X-rays, thereby providing crucial 3D guidance in challenging sparse-view and limited-angle settings where current registration methods fail.



Paperid:5266
Authors:Tao Wang, Mengyu Li, Geduo Zeng, Cheng Meng, Qiong Zhang
Abstract:
3D Gaussian Splatting (3DGS) has emerged as a powerful technique for radiance field rendering, but it typically requires millions of redundant Gaussian primitives, overwhelming memory and rendering budgets. Existing compaction approaches address this by pruning Gaussians based on heuristic importance scores, without global fidelity guarantee. To bridge this gap, we propose a novel optimal transport perspective that casts 3DGS compaction as global Gaussian mixture reduction. Specifically, we first minimize the composite transport divergence over a KD-tree partition to produce a compact geometric representation, and then decouple appearance from geometry by fine-tuning color and opacity attributes with far fewer Gaussian primitives. Experiments on benchmark datasets show that our method (i) yields negligible loss in rendering quality (PSNR, SSIM, LPIPS) compared to vanilla 3DGS with only 10\% Gaussians; and (ii) consistently outperforms state-of-the-art 3DGS compaction techniques. Notably, our method is applicable to any stage of vanilla or accelerated 3DGS pipelines, providing an efficient and agnostic pathway to lightweight neural rendering.



Paperid:5267
Authors:Haoxin Yang, Bangzhen Liu, Xuemiao Xu, Cheng Xu, Yuyang Yu, Zikai Huang, Yi Wang, Shengfeng He
Abstract:
The advancement of diffusion models has enhanced the realism of AI-generated content but also raised concerns about misuse, necessitating robust copyright protection and tampering localization. Although recent methods have made progress toward unified solutions, their reliance on post hoc processing introduces considerable application inconvenience and compromises forensic reliability. We propose StableGuard, a novel framework that seamlessly integrates a binary watermark into the diffusion generation process, ensuring copyright protection and tampering localization in Latent Diffusion Models through an end-to-end design. We develop a Multiplexing Watermark VAE (MPW-VAE) by equipping a pretrained Variational Autoencoder (VAE) with a lightweight latent residual-based adapter, enabling the generation of paired watermarked and watermark-free images. These pairs, fused via random masks, create a diverse dataset for training a tampering-agnostic forensic network. To further enhance forensic synergy, we introduce a Mixture-of-Experts Guided Forensic Network (MoE-GFN) that dynamically integrates holistic watermark patterns, local tampering traces, and frequency-domain cues for precise watermark verification and tampered region detection. The MPW-VAE and MoE-GFN are jointly optimized in a self-supervised, end-to-end manner, fostering a reciprocal training between watermark embedding and forensic accuracy. Extensive experiments demonstrate that StableGuard consistently outperforms state-of-the-art methods in image fidelity, watermark verification, and tampering localization.



Paperid:5268
Authors:Ziheng Cheng, Tianyu Xie, Shiyue Zhang, Cheng Zhang
Abstract:
While conditional diffusion models have achieved remarkable success in various applications, they require abundant data to train from scratch, which is often infeasible in practice. To address this issue, transfer learning has emerged as an essential paradigm in small data regimes. Despite its empirical success, the theoretical underpinnings of transfer learning conditional diffusion models remain unexplored. In this paper, we take the first step towards understanding the sample efficiency of transfer learning conditional diffusion models through the lens of representation learning. Inspired by practical training procedures, we assume that there exists a low-dimensional representation of conditions shared across all tasks. Our analysis shows that with a well-learned representation from source tasks, the sample complexity of target tasks can be reduced substantially. Numerical experiments are also conducted to verify our results.



Paperid:5269
Authors:Haoyuan Wu, Rui Ming, Jilong Gao, Hangyu Zhao, Xueyi Chen, Yikai Yang, Haisheng Zheng, Zhuolun He, Bei Yu
Abstract:
Large language models (LLMs) achieve remarkable performance in code generation tasks.However, a significant performance disparity persists between popular programming languages (e.g., Python, C++) and others. To address this capability gap, we leverage the code translation task to train LLMs, thereby facilitating the transfer of coding proficiency across diverse programming languages.Moreover, we introduce OORL for training, a novel reinforcement learning (RL) framework that integrates on-policy and off-policy strategies.Within OORL, on-policy RL is applied during code translation, guided by a rule-based reward signal derived from unit tests.Complementing this coarse-grained rule-based reward, we propose Group Equivalent Preference Optimization (GEPO), a novel preference optimization method.Specifically, GEPO trains the LLM using intermediate representations (IRs) groups.LLMs can be guided to discern IRs equivalent to the source code from inequivalent ones, while also utilizing signals about the mutual equivalence between IRs within the group.This process allows LLMs to capture nuanced aspects of code functionality.By employing OORL for training with code translation tasks, LLMs improve their recognition of code functionality and their understanding of the relationships between code implemented in different languages. Extensive experiments demonstrate that our OORL for LLMs training with code translation tasks achieves significant performance improvements on code benchmarks across multiple programming languages.



Paperid:5270
Authors:Gal Fadlon, Idan Arbiv, Nimrod Berman, Omri Azencot
Abstract:
Generating realistic time series data is critical for applications in healthcare, finance, and climate science. However, irregular sampling and missing values present significant challenges. While prior methods address these irregularities, they often yield suboptimal results and incur high computational costs. Recent advances in regular time series generation, such as the diffusion-based ImagenTime model, demonstrate strong, fast, and scalable generative capabilities by transforming time series into image representations, making them a promising solution. However, extending ImagenTime to irregular sequences using simple masking introduces ``unnatural'' neighborhoods, where missing values replaced by zeros disrupt the learning process. To overcome this, we propose a novel two-step framework: first, a Time Series Transformer completes irregular sequences, creating natural neighborhoods; second, a vision-based diffusion model with masking minimizes dependence on the completed values. This hybrid approach leverages the strengths of both completion and masking, enabling robust and efficient generation of realistic time series. Our method achieves state-of-the-art performance across benchmarks, delivering a relative improvement in discriminative score by 70% and in computational cost by 85%.



Paperid:5271
Authors:Ibrahim Alabdulmohsin, Xiaohua Zhai
Abstract:
Inspired by recent findings on the fractal geometry of language, we introduce Recursive INference Scaling (RINS) as a complementary, plug-in recipe for scaling inference time in language and multimodal systems. RINS is a particular form of recursive depth that significantly outperforms +55 other variants, including the recent "repeat-all-over" (RAO) strategy in Mobile LLM (Liu et al., 2024) and latent recurrent thinking (Geiping et al., 2025). Unlike prior works, we carry out our comparisons on a compute-matched regime, and demonstrate that for a fixed model size and training compute budget, RINS substantially improves language modeling performance. It also generalizes beyond pure language tasks, delivering gains in multimodal systems, including a +2% improvement in 0-shot ImageNet accuracy for SigLIP-B/16. Additionally, by deriving data scaling laws, we show that RINS improves both the asymptotic performance limits and the scaling exponents. More importantly, with light-weight (linear) adapters (comprising <1% of model parameters) and stochastic dropout, RINS offers a no-regret strategy, meaning that RINS-enabled pretraining improves performance in language modeling even when recursive depth is not applied at inference time. This corresponds to improving performance on a training compute-, parameter-, and inference-matched regime, suggesting its potential as a viable component of LLM pretraining!



Paperid:5272
Authors:Tsung-Han (Patrick) Wu, Heekyung Lee, Jiaxin Ge, Joseph Gonzalez, Trevor Darrell, David Chan
Abstract:
Vision-Language Models (VLMs) excel at visual understanding but often suffer from visual hallucinations, where they generate descriptions of nonexistent objects, actions, or concepts, posing significant risks in safety-critical applications. Existing hallucination mitigation methods typically follow one of two paradigms: generation adjustment, which modifies decoding behavior to align text with visual inputs, and post-hoc verification, where external models assess and correct outputs. While effective, generation adjustment methods often rely on heuristics and lack correction mechanisms, while post-hoc verification is complicated, typically requiring multiple models and tending to reject outputs rather than refine them. In this work, we introduce REVERSE, a unified framework that integrates hallucination-aware training with on-the-fly self-verification. By leveraging a new hallucination-verification dataset containing over 1.3M semi-synthetic samples, along with a novel inference-time retrospective resampling technique, our approach enables VLMs to both detect hallucinations during generation and dynamically revise those hallucinations. Our evaluations show that REVERSE achieves state-of-the-art hallucination reduction, outperforming the best existing methods by up to 12% on CHAIR-MSCOCO and 34% on HaloQuest.



Paperid:5273
Authors:Luojie Huang, Ryan Zhang, Marisa Morakis, Michaela Taylor-Williams, Gregory McKay, Nicholas Durr
Abstract:
Abstract:Analysis of noninvasive microvascular blood flow can improve the diagnosis, prognosis, and management of many medical conditions, including cardiovascular, peripheral vascular, and sickle cell disease. This paper introduces SAM2Flow, an interactive optical flow estimation model to analyze long Oblique Back-illumination Microscopy (OBM) videos of in vivo microvascular flow. Inspired by the Segment Anything Model (SAM2), SAM2Flow enables users to specify regions of interest through user prompts for focused flow estimation. SAM2Flow also incorporates a dual memory attention mechanism, comprising both motion and context memory, to achieve efficient and stable flow estimations over extended video sequences. According to our experiments, SAM2Flow achieves SOTA accuracy in flow estimation with a fast inference speed of over $20$ fps on $512\times512$ inputs. Based on the temporally robust flow estimation, SAM2Flow demonstrated superior performance in downstream physiological applications compared to existing models. The code and dataset will be published with this paper.



Paperid:5274
Authors:Li Sun, Zhenhao Huang, Ming Zhang, Philip S Yu
Abstract:
Abstract:Message Passing Neural Networks (MPNNs) achieve remarkable progress in learning on graphs, but fundamentally suffer from oversmoothing and oversquashing. There has recently been a surge of interest in fixing both issues. Existing efforts primarily conduct some global treatments that can be preferable to some regions but harmful to the others, leading to the suboptimal expressiveness. In this paper, we start with revisiting oversquashing through a global measure of spectral gap $\lambda$, and prove that the increase of $\lambda$ is exposed to gradient vanishing regarding the input features,deteriorating the efficacy of message passing. Motivated by such theoretical insights, we propose a novel local treatment where the adaptive adjustment is introduced to message passing, according to the local structures. To this end, we bridge local Riemannian geometry with MPNNs, and establish a heterogeneous boundary condition to simultaneously address oversquashing and oversmoothing. Hence, grounded on Robin condition, we design a simple yet effective GBN network with theoretical guarantees. Extensive experiments on homophilic and heterophilic graphs show the expressiveness of GBN.



Paperid:5275
Authors:Zechuan Zhang, Ji Xie, Yu Lu, Zongxin Yang, Yi Yang
Abstract:
Instruction-based image editing enables precise modifications via natural language prompts, but existing methods face a precision-efficiency tradeoff: fine-tuning demands massive datasets (>10M) and computational resources, while training-free approaches suffer from weak instruction comprehension. We address this by proposing \textbf{ICEdit}, which leverages the inherent comprehension and generation abilities of large-scale Diffusion Transformers (DiTs) through three key innovations: (1) An in-context editing paradigm without architectural modifications; (2) Minimal parameter-efficient fine-tuning for quality improvement; (3) Early Filter Inference-Time Scaling, which uses VLMs to select high-quality noise samples for efficiency. Experiments show that ICEdit achieves state-of-the-art editing performance with only 0.1\% of the training data and 1\% trainable parameters compared to previous methods. Our approach establishes a new paradigm for balancing precision and efficiency in instructional image editing.



Paperid:5276
Authors:Xiaoyi Pang, Xuanyi Hao, Song Guo, Qi Luo, Zhibo Wang
Abstract:
The widespread deployment of large language models (LLMs) allows users to access their capabilities via black-box APIs, but backdoor attacks pose serious security risks for API users by hijacking the model behavior. This highlights the importance of backdoor detection technologies to help users audit LLMs before use. However, most existing LLM backdoor defenses require white-box access or costly reverse engineering, limiting their practicality for resource-constrained users. Moreover, they mainly target classification tasks, leaving broader generative scenarios underexplored. To solve the problem, this paper introduces ICLScan, a lightweight framework that exploits targeted in-context learning (ICL) as illumination for backdoor detection in black-box LLMs, which effectively supports generative tasks without additional training or model modifications. ICLScan is based on our finding of backdoor susceptibility amplification: LLMs with pre-embedded backdoors are highly susceptible to new trigger implantation via ICL. Including only a small ratio of backdoor examples (containing ICL-triggered input and target output) in the ICL prompt can induce ICL trigger-specific malicious behavior in backdoored LLMs. ICLScan leverages this phenomenon to detect backdoored LLMs by statistically analyzing whether the success rate of new trigger injection via targeted ICL exceeds a threshold. It requires only multiple queries to estimate the backdoor success rate, overcoming black-box access and computational resource limitations. Extensive experiments across diverse LLMs and backdoor attacks demonstrate ICLScan's effectiveness and efficiency, achieving near-perfect detection performance (precision/recall/F1-score/ROC-AUC all approaching 1) with minimal additional overhead across all settings.



Paperid:5277
Authors:Shivani Agarwal
Abstract:
There has been much interest in exploring intermediate PAC learning models between the realizable and (fully) agnostic settings, that allow both for some stochasticity in the labels and for computationally efficient learning algorithms with finite sample complexity bounds. Some examples of such models include random classification noise (RCN), Massart noise, and generalized linear models (GLMs). Most of this work has focused on the binary classification problem. In this paper, we study a broad range of more complex learning problems under what we call realizable-statistic models (RSMs), wherein we allow stochastic labels but assume that some vector-valued statistic of the conditional label distribution comes from some known function class. RSMs can be viewed as generalizing GLMs to both more general function classes and more general learning problems. We give a general result showing that for a broad set of RSM learning problems, convex surrogate risk minimization algorithms designed to estimate the underlying statistic via an appropriately defined strongly proper composite surrogate loss yield computationally efficient learning algorithms with finite sample complexity bounds. We then apply this result to show that various commonly used (and in some cases, not so commonly used) convex surrogate risk minimization algorithms yield computationally efficient learning algorithms with finite sample complexity bounds for a variety of learning problems including binary classification, multiclass classification, multi-label prediction, and subset ranking. In terms of the distribution over the domain/instance space, our results are all distribution-independent. To our knowledge, these are the first such results for PAC learning with stochastic labels for such a broad set of learning problems.



Paperid:5278
Authors:Wenchao Yang, Weidong Yan, Wenkang Liu, Yulan Ma, Yang Li
Abstract:
Large-scale pre-trained models hold significant potential for learning universal EEG representations. However, most existing methods, particularly autoregressive (AR) frameworks, primarily rely on straightforward temporal sequencing of multi-channel EEG data, which fails to capture the rich physiological characteristics inherent to EEG signals. Moreover, their time-centered modeling approach also limits the effective representation of the dynamic spatial topology of brain activity. To address these challenges and fully exploit the potential of large-scale EEG models, we propose a novel Topology Hierarchical Derived Brain Autoregressive Modeling (THD-BAR) for EEG generic representations. The core innovation of THD-BAR lies in the introduction of the Brain Topology Hierarchy (BTH), which establishes a multi-scale spatial order for EEG channels. This hierarchical structure enables a redefinition of autoregressive learning as a "next-scale-time prediction" problem, effectively capturing both spatial and temporal dynamics. Based on BTH, we design a Topology-Hierarchical Vector Quantized-Variational Autoencoder (THVQ-VAE) for multi-scale tokenization and develop an enhanced Brain Autoregressive (BAR) module with specialized masking strategies for prediction. Through extensive large-scale pre-training on 17 datasets, followed by rigorous validation on 10 downstream datasets spanning 5 distinct tasks, THD-BAR consistently outperforms existing methods. These results highlight the superior generalization and modeling capabilities of our proposed approach. Our code is available at https://anonymous.4open.science/r/THD-BAR.



Paperid:5279
Authors:Wei Chen, Xin Yan, Bin Wen, Fan Yang, Tingting Gao, Di ZHANG, Long Chen
Abstract:
Although multimodal large language models (MLLMs) exhibit remarkable reasoning capabilities on complex multimodal understanding tasks, they still suffer from the notorious 'hallucination' issue: generating outputs misaligned with obvious visual or factual evidence. Currently, training-based solutions, like direct preference optimization (DPO), leverage paired preference data to suppress hallucinations. However, they risk sacrificing general reasoning capabilities due to the likelihood displacement. Meanwhile, training-free solutions, like contrastive decoding, achieve this goal by subtracting the estimated hallucination pattern from a distorted input. Yet, these handcrafted perturbations (e.g., add noise to images) may poorly capture authentic hallucination patterns. To avoid these weaknesses of existing methods, and realize ``robust'' hallucination mitigation (\ie, maintaining general reasoning performance), we propose a novel framework: Decoupling Contrastive Decoding (DCD). Specifically, DCD decouples the learning of positive and negative samples in preference datasets, and trains separate positive and negative image projections within the MLLM. The negative projection implicitly models real hallucination patterns, which enables vision-aware negative images in the contrastive decoding inference stage. Our DCD alleviates likelihood displacement by avoiding pairwise optimization and generalizes robustly without handcrafted degradation. Extensive ablations across hallucination benchmarks and general reasoning tasks demonstrate the effectiveness of DCD, \ie, it matches DPO’s hallucination suppression while preserving general capabilities and outperforms the handcrafted contrastive decoding methods.



Paperid:5280
Authors:Chandler Smith, Marwa Abdulhai, Manfred Díaz, Marko Tesic, Rakshit Trivedi, Sasha Vezhnevets, Lewis Hammond, Jesse Clifton, Minsuk Chang, Edgar Duenez-Guzman, John Agapiou, Jayd Matyas, Danny Karmon, Beining Zhang, Jim Dilkes, Akash Kundu, Hieu Minh Nguyen, Emanuel Tewolde, Jebish Purbey, Ram Mohan Rao Kadiyala, Siddhant Gupta, Aliaksei Korshuk, Buyantuev Alexander, Ilya Makarov, Gang Zhao, Rolando Fernandez, Zhihan Wang, Caroline Wang, Jiaxun Cui, Lingyun Xiao, Di Shi, Yoonchang Sung, Muhammad Arrasy Rahman, Peter Stone, Yipeng Kang, Hyeonggeun Yun, Ananya Ananya, Taehun Cha, Zhiqiang Wu, Elizaveta Tennant, Olivia Macmillan-Scott, Marta Segura, Diana Riazi, Fuyang Cui, Sriram Ganapathi, Toryn Klassen, Nico Schiavone, Mogtaba Alim, Sheila McIlraith, Manuel Rios, Oswaldo Peña, Carlos Rojas, Manuela Viviana Chacon-Chamorro, Rubén Manrique, Luis Felipe Giraldo, Nicanor Quijano, Yiding Wang, Yuxuan Chen, Fangwei Zhong, Mengmeng Wang, Wenming Tu, Zhaowei Zhang, Ziang Chen, Zixia Jia, Xue Feng, Zilong Zheng, Chichen Lin, Weijian Fan, Chenao Liu, Sneheel Sarangi, Ziyan Wang, shuqing shi, Yali Du, Avinaash Anand Kulandaivel, Yang Liu, Wu Ruiyang, Chetan Talele, 陆孙嘉, Gema Parreno, Shamika Dhuri, Bain McHale, Tim Baarslag, Dylan Hadfield-Menell, Natasha Jaques, José Hernández-Orallo, Joel Leibo
Abstract:
Large language model (LLM) agents have demonstrated impressive capabilities for social interaction and are increasingly being deployed in situations where they might engage with both human and artificial agents. These interactions represent a critical frontier for LLM-based agents, yet existing evaluation methods fail to measure how well these capabilities generalize to novel social situations. In this paper, we introduce a method for evaluating the ability of LLM-based agents to cooperate in zero-shot, mixed-motive environments using Concordia, a natural language multi-agent simulation environment. This work introduces an approach to measuring human-appropriate cooperative intelligence, emphasizing an agent's ability to identify and exploit opportunities for mutual gain across diverse partners and contexts. We present empirical results from the NeurIPS 2024 Concordia Contest, where agents were evaluated on their ability to achieve mutual gains across a suite of diverse scenarios ranging from negotiation to collective action problems. Our findings reveal significant gaps between current agent capabilities and the robust generalization required for reliable cooperation, particularly in scenarios demanding persuasion and norm enforcement.



Paperid:5281
Authors:Shengrong Li, Qi Zhu, Chunwei Tian, Xinyang Zhang, WEI SHAO, Jie Wen, Daoqiang Zhang
Abstract:
Dynamic functional brain networks (DFBNs) are powerful tools in neuroscience research. Recent studies reveal that DFBNs contain heterogeneous neural nodes with more extensive connections and more drastic temporal changes, which play pivotal roles in coordinating the reorganization of the brain. Moreover, the spatio-temporal patterns of these nodes are modulated by the brain's historical states. However, existing methods not only ignore the spatio-temporal heterogeneity of neural nodes, but also fail to effectively encode the temporal propagation mechanism of heterogeneous activities. These limitations hinder the deep exploration of spatio-temporal relationships within DFBNs, preventing the capture of abnormal neural heterogeneity caused by brain diseases. To address these challenges, this paper propose a neuro-heterogeneity guided temporal graph learning strategy (NeuroH-TGL). Specifically, we first develop a spatio-temporal pattern decoupling module to disentangle DFBNs into topological consistency networks and temporal trend networks that align with the brain's operational mechanisms. Then, we introduce a heterogeneity mining module to identify pivotal heterogeneity nodes that drive brain reorganization from the two decoupled networks. Finally, we design temporal propagation graph convolution to simulate the influence of the historical states of heterogeneity nodes on the current topology, thereby flexibly extracting heterogeneous spatio-temporal information from the brain. Experiments show that our method surpasses several state-of-the-art methods, and can identify abnormal heterogeneous nodes caused by brain diseases.



Paperid:5282
Authors:Heming Zou, Yunliang Zang, Wutong Xu, Yao Zhu, Xiangyang Ji
Abstract:
Low-Rank Adaptation (LoRA) is a widely used parameter-efficient fine-tuning method for foundation models, but it suffers from parameter interference, resulting in suboptimal performance. Although Mixture-of-Experts (MoE)-based LoRA variants show promise in addressing this issue, they often struggle to balance decorrelation with computational efficiency. Inspired by the fly olfactory circuit, we propose FlyLoRA, an implicit MoE-based LoRA variant that introduces: (1) rank-wise expert activation in the up-projection matrix with implicit routing, and (2) frozen sparse random projection for the down-projection matrix, replacing the traditional dense updatable version. These components theoretically decouple intra- and inter-task interactions, enhancing decorrelation in both single-task instruction tuning and multi-task model fusion, respectively. FlyLoRA’s implicit routing and inherent sparsity further enhance computational efficiency. Extensive experiments across four domains --- general knowledge understanding, scientific question answering, mathematical reasoning, and code generation --- demonstrate consistent performance improvements over existing methods. Additionally, this work illustrates how biological structures can inspire innovations in AI technologies.



Paperid:5283
Authors:Jinghong Zheng, Changlong Jiang, Yang Xiao, Jiaqi Li, Haohong Kuang, Hang Xu, Ran Wang, Zhiguo Cao, Min Du, Joey Tianyi Zhou
Abstract:
3D human pose lifting from a single RGB image is a challenging task in 3D vision. Existing methods typically establish a direct joint-to-joint mapping from 2D to 3D poses based on 2D features. This formulation suffers from two fundamental limitations: inevitable error propagation from input predicted 2D pose to 3D predictions and inherent difficulties in handling self-occlusion cases. In this paper, we propose PandaPose, a 3D human pose lifting approach via propagating 2D pose prior to 3D anchor space as the unified intermediate representation. Specifically, our 3D anchor space comprises: (1) Joint-wise 3D anchors in the canonical coordinate system, providing accurate and robust priors to mitigate 2D pose estimation inaccuracies. (2) Depth-aware joint-wise feature lifting that hierarchically integrates depth information to resolve self-occlusion ambiguities. (3) The anchor-feature interaction decoder that incorporates 3D anchors with lifted features to generate unified anchor queries encapsulating joint-wise 3D anchor set, visual cues and geometric depth information.The anchor queries are further employed to facilitate anchor-to-joint ensemble prediction.Experiments on three well-established benchmarks (i.e., Human3.6M, MPI-INF-3DHP and 3DPW) demonstrate the superiority of our proposition.The substantial reduction in error by 14.7% compared to SOTA methodson the challenging conditions of Human3.6M and qualitative comparisons further showcase the effectiveness and robustness of our approach.



Paperid:5284
Authors:Yangfu Li, Hongjian Zhan, Tianyi Chen, Qi Liu, Yu-Jie Xiong, Yue Lu
Abstract:
Abstract:Existing visual token pruning methods target prompt alignment and visual preservation with static strategies, overlooking the varying relative importance of these objectives across tasks, which leads to inconsistent performance. To address this, we derive the first closed-form error bound for visual token pruning based on the Hausdorff distance, uniformly characterizing the contributions of both objectives. Moreover, leveraging $\epsilon$-covering theory, we reveal an intrinsic trade-off between these objectives and quantify their optimal attainment levels under a fixed budget. To practically handle this trade-off, we propose Multi-Objective Balanced Covering (MoB), which reformulates visual token pruning as a bi-objective covering problem. In this framework, the attainment trade-off reduces to budget allocation via greedy radius trading. MoB offers a provable performance bound and linear scalability with respect to the number of input visual tokens, enabling adaptation to challenging pruning scenarios. Extensive experiments show that MoB preserves 96.4\% of performance for LLaVA-1.5-7B using only 11.1\% of the original visual tokens and accelerates LLaVA-Next-7B by 1.3-1.5$\times$ with negligible performance loss. Additionally, evaluations on Qwen2-VL and Video-LLaVA confirm that MoB integrates seamlessly into advanced MLLMs and diverse vision-language tasks. The code will be made available soon.



Paperid:5285
Authors:Itamar Avitan, Tal Golan
Abstract:
Linearly transforming stimulus representations of deep neural networks yieldshigh-performing models of behavioral and neural responses to complex stimuli.But does the test accuracy of such predictions identify genuine representationalalignment? We addressed this question through a large-scale model-recovery study.Twenty diverse vision models were linearly aligned to 4.5 million behavioral judg-ments from the THINGS odd-one-out dataset and calibrated to reproduce humanresponse variability. For each model in turn, we sampled synthetic responses fromits probabilistic predictions, fitted all candidate models to the synthetic data, andtested whether the data-generating model would re-emerge as the best predictorof the simulated data. Model recovery accuracy improved with training-set sizebut plateaued below 80%, even at millions of simulated trials. Regression analyseslinked misidentification primarily to shifts in representational geometry inducedby the linear transformation, as well as to the effective dimensionality of the trans-formed features. These findings demonstrate that, even with massive data, overlyflexible alignment metrics may fail to guide us toward artificial representationsthat are genuinely more human-aligned. Model comparison experiments must bedesigned to balance the trade-off between predictive accuracy and identifiability—ensuring the best-fitting model is also the right one.



Paperid:5286
Authors:Xiaoyu Wu, Yifei Pang, Terrance Liu, Steven Wu
Abstract:
Large language models are typically trained on datasets collected from the web, which may inadvertently contain harmful or sensitive personal information. To address growing privacy concerns, unlearning methods have been proposed to remove the influence of specific data from trained models. Of these, exact unlearning---which retrains the model from scratch without the target data---is widely regarded the gold standard, believed to be robust against privacy-related attacks. In this paper, we challenge this assumption by introducing a novel data extraction attack that compromises even exact unlearning. Our method leverages both the pre- and post-unlearning models: by guiding the post-unlearning model using signals from the pre-unlearning model, we uncover patterns that reflect the removed data distribution. Combining model guidance with a token filtering strategy, our attack significantly improves extraction success rates---doubling performance in some cases---across common benchmarks such as MUSE, TOFU, and WMDP. Furthermore, we demonstrate our attack's effectiveness on a simulated medical diagnosis dataset to highlight real-world privacy risks associated with exact unlearning. In light of our findings, which suggest that unlearning may, in a contradictory way,increasethe risk of privacy leakage, we advocate for evaluation of unlearning methods to consider broader threat models that account not only for post-unlearning models but also for adversarial access to prior checkpoints.



Paperid:5287
Authors:Sian-Yao Huang, Li-Hsien Chang, Che-Yu Lin, Cheng-Lin Yang
Abstract:
Large language models (LLMs) are often prompted with multi-level directives—such as system instructions and user queries—that imply a hierarchy of authority. Yet models frequently fail to enforce this structure, especially in multi-step reasoning where errors propagate across intermediate steps. Existing methods rely on oracle completions but lack verifiable reward signals or intermediate traces, limiting their applicability. We introduce a unified supervision framework that embeds programmatically verifiable checkers into synthesized instruction-conflict instances. Each instance pairs a compliance directive with a conflicting one, along with an executable verifier that deterministically checks output adherence. This enables alignment without oracle labels or reasoning traces, supporting both instruction-tuned and reasoning models. The framework is instantiated via a synthesis pipeline that includes unit-test–based validation, LLM-assisted repair, and a probabilistic analysis of cleaning reliability. Fine-tuning on the resulting data improves instruction hierarchy adherence and boosts safety robustness—generalizing to adversarial safety benchmarks without task-specific supervision. This highlights verifiable supervision as a scalable foundation for robust alignment.



Paperid:5288
Authors:Yanhao Jia, Ji Xie, S Jivaganesh, Li Hao, Xu Wu, Mengmi Zhang
Abstract:
Imagine hearing a dog bark and instinctively turning toward the sound—only to find a parked car, while a silent dog sits nearby. Such moments of sensory conflict challenge perception, yet humans flexibly resolve these discrepancies, prioritizing auditory cues over misleading visuals to accurately localize sounds. Despite the rapid advancement of multimodal AI models that integrate vision and sound, little is known about how these systems handle cross-modal conflicts or whether they favor one modality over another. Here, we systematicly and quantitatively investigate modality bias and conflict resolution in AI sound localization. We evaluate a wide range of state-of-the-art multimodal models and compare them against human performance in carefully designed psychophysics experiments spanning six audiovisual conditions, including congruent, conflicting, and absent visual and audio cues. Our results reveal that humans consistently outperform AI in sound localization and exhibit greater robustness to conflicting or absent visual information by effectively prioritizing auditory signals. In contrast, AI shows a pronounced bias toward vision, often failing to suppress irrelevant or conflicting visual input, leading to chance-level performance. To close this gap, we fine-tune a state-of-the-art multimodal model on our carefully curated audio-image dataset synthesized via 3D simulation engines. Despite using fewer training examples, our model outperforms existing AI trained on multiple large-scale audio-visual benchmark datasets. Remarkably, consistent with neuroscience findings, our model exhibits a human-like horizontal localization bias—favoring left-right over up-down precision. This likely stems from training on stereo audio that mirrors the horizontal arrangement of human ears. These results highlight how the quality of sensory inputs and the physical structure of sensory systems jointly shape the precision of multimodal representation. All code, data, and models will be made public.



Paperid:5289
Authors:Hongyi Nie, Yaqing Wang, Mingyang Zhou, Feiyang Pan, Quanming Yao, Zhen Wang
Abstract:
As large language models (LLMs) are increasingly used as personalized user assistants, effectively adapting to users' evolving preferences is critical for delivering high-quality personalized responses. While user preferences are often stable in content, their relative strengths shift over time due to changing goals and contexts. Therefore, modeling these dynamic preference strengths can enable finer-grained personalization. However, current methods face two major challenges: (i) limited user feedback makes it difficult to estimate preference strengths accurately, and (ii) natural language ambiguity limits the controllability of preference-guided generation. To address these issues, we propose AdaPA-Agent, a LLM-agent personalization framework that models dynamic preference strengths via Adaptive Preference Arithmetic. First, instead of requiring additional user feedback, AdaPA-Agent employs an alignment-based strength estimation module to estimate the strength of user preferences from the existing user-agent interaction. Then, it guides controllable personalized generation by linearly combining next-token distributions, weighted by the estimated strengths of individual preferences. Experiments on two personalization tasks-conversational recommendation and personalized web interaction-demonstrate that AdaPA-Agent better aligning with users' changing intents, and has achieved over 18.9\% and 14.2\% improvements compared to ReAct, the widely-used agent framework.



Paperid:5290
Authors:Qihe Huang, Zhengyang Zhou, Yangze Li, Kuo Yang, Binwu Wang, Yang Wang
Abstract:
Time series forecasting is a critical and complex task, characterized by diverse temporal patterns, varying statistical properties, and different prediction horizons across datasets and domains. Conventional approaches typically rely on a single, unified model architecture to handle all forecasting scenarios. However, such monolithic models struggle to generalize across dynamically evolving time series with shifting patterns. In reality, different types of time series may require distinct modeling strategies. Some benefit from homogeneous multi-scale forecasting awareness, while others rely on more complex and heterogeneous signal perception. Relying on a single model to capture all temporal diversity and structural variations leads to limited performance and poor interpretability. To address this challenge, we propose a Multi-Agent Forecasting System (MAFS) that abandons the one-size-fits-all paradigm. MAFS decomposes the forecasting task into multiple sub-tasks, each handled by a dedicated agent trained on specific temporal perspectives (e.g., different forecasting resolutions or signal characteristics). Furthermore, to achieve holistic forecasting, agents share and refine information through different communication topology, enabling cooperative reasoning across different temporal views. A lightweight voting aggregator then integrates their outputs into consistent final predictions. Extensive experiments across 11 benchmarks demonstrate that MAFS significantly outperforms traditional single-model approaches, yielding more robust and adaptable forecasts. Code is available at \url{https://anonymous.4open.science/r/MAFS}.



Paperid:5291
Authors:Chenyu Yang, Shuai Wang, Hangting Chen, Wei Tan, Jianwei Yu, Haizhou Li
Abstract:
Generating music with coherent structure, harmonious instrumental and vocal elements remains a significant challenge in song generation. Existing language models and diffusion-based methods often struggle to balance global coherence with local fidelity, resulting in outputs that lack musicality or suffer from incoherent progression and mismatched lyrics. This paper introduces \textbf{SongBloom}, a novel framework for full-length song generation that leverages an interleaved paradigm of autoregressive sketching and diffusion-based refinement. SongBloom employs an autoregressive diffusion model that combines the high fidelity of diffusion models with the scalability of language models. Specifically, it gradually extends a musical sketch from short to long and refines the details from coarse to fine-grained. The interleaved generation paradigm effectively integrates prior semantic and acoustic context to guide the generation process.Experimental results demonstrate that SongBloom outperforms existing methods across both subjective and objective metrics and achieves performance comparable to the state-of-the-art commercial music generation platforms. Audio samples are available on our demo page: \url{https://anonymousdemo000.github.io/SongBloom-demo/}.



Paperid:5292
Authors:Denis Sutter, Julian Minder, Thomas Hofmann, Tiago Pimentel
Abstract:
The concept of causal abstraction got recently popularised to demystify the opaque decision-making processes of machine learning models; in short, a neural network can be abstracted as a higher-level algorithm if there exists a function which allows us to map between them. Notably, most interpretability papers implement these maps as linear functions, motivated by the linear representation hypothesis: the idea that features are encoded linearly in a model's representations. However, this linearity constraint is not required by the definition of causal abstraction. In this work, we critically examine the concept of causal abstraction by considering arbitrarily powerful alignment maps. In particular, we prove that under reasonable assumptions, any neural network can be mapped to any algorithm, rendering the notion of causal abstraction trivial and uninformative.We complement these theoretical findings with empirical evidence, demonstrating that it is possible to perfectly map models to algorithms even when these models are incapable of solving the actual task; e.g.,on an experiment using randomly initialised language models, our alignment maps reach 100% interchange-intervention accuracy on the indirect object identification task. This raises the non-linear representation dilemma: if we lift the linearity constraint imposed to alignment maps in causal abstraction analyses, we are left with no principled way to balance the inherent trade-off between these maps' complexity and accuracy.Together, these results suggest an answer to our title's question: causal abstraction is not enough for mechanistic interpretability, as it becomes vacuous without assumptions about how models encode information.



Paperid:5293
Authors:Ziming Liu, Shaoyu Wang, Shenggan Cheng, Zhongkai Zhao, Kai Wang, Xuanlei Zhao, James Demmel, Yang You
Abstract:
Training Transformer models on long sequences in a distributed setting poses significant challenges in terms of efficiency and scalability. Current methods are either constrained by the number of attention heads or excessive communication overheads. To address this problem, we propose StarTrail, a multi-dimensional concentric distributed training system for long sequences, fostering an efficient communication paradigm and providing additional tuning flexibility for communication arrangements. Specifically, StarTrail introduces an extra parallel dimension and divides the peer-to-peer communication into sub-rings to substantially reduce communication volume and avoid bandwidth bottlenecks. Through comprehensive experiments across diverse hardware environments and on both Natural Language Processing (NLP) and Computer Vision (CV) tasks, we demonstrate that our approach significantly surpasses state-of-the-art methods that support Long sequence lengths, achieving performance improvements of up to 77.12% on GPT-style models and up to 114.33% on DiT (Diffusion Transformer) models without affecting the computations results.



Paperid:5294
Authors:Quan Shi, Carlos Jimenez, Shunyu Yao, Nick Haber, Diyi Yang, Karthik Narasimhan
Abstract:
As large language models (LLMs) increasingly serve as close collaborators for humans, it is crucial that they express their reasoning in ways that humans can understand and learn from. However, this capability remains relatively less understood and under-evaluated. To address this, we introduce a conceptual framework for such Human-AI knowledge transfer capabilities and conduct the first large-scale user study (N=118) explicitly designed to measure it. In our two-phase setup, humans first ideate with an LLM on problem-solving strategies, then independently implement solutions, isolating the influence of model reasoning on human understanding. Our findings reveal that while model benchmark performance correlates with collaborative outcomes, this relationship is notably inconsistent with significant outliers, highlighting that knowledge transfer is a distinct capability requiring dedicated optimization. Our analysis uncovers behavioral and strategic factors that mediate successful knowledge transfer, and we release our code, dataset, and evaluation framework to support future work on communicatively aligned models.



Paperid:5295
Authors:Zhangqi Jiang, Tingjin Luo, Xu Yang, Xinyan Liang
Abstract:
View missing remains a significant challenge in graph-based multi-view semi-supervised learning, hindering their real-world applications. To address this issue, traditional methods introduce a missing indicator matrix and focus on mining partial structure among existing samples in each view for label propagation (LP). However, we argue that these disregarded missing samples sometimes induce discontinuous local structures, i.e., sub-clusters, breaking the fundamental smoothness assumption in LP. Consequently, such a Sub-Cluster Problem (SCP) would distort graph fusion and degrade classification performance. To alleviate SCP, we propose a novel incomplete multi-view semi-supervised learning method, termed AGF-TI. Firstly, we design an adversarial graph fusion scheme to learn a robust consensus graph against the distorted local structure through a min-max framework. By stacking all similarity matrices into a tensor, we further recover the incomplete structure from the high-order consistency information based on the low-rank tensor learning. Additionally, the anchor-based strategy is incorporated to reduce the computational complexity. An efficient alternative optimization algorithm combining a reduced gradient descent method is developed to solve the formulated objective, with theoretical convergence. Extensive experimental results on various datasets validate the superiority of our proposed AGF-TI as compared to state-of-the-art methods.



Paperid:5296
Authors:Jiahuan Zhou, Kai Zhu, Zhenyu Cui, Zichen Liu, Xu Zou, Gang Hua
Abstract:
Recently, pre-trained state space models have shown great potential for video classification, which sequentially compresses visual tokens in videos with linear complexity, thereby improving the processing efficiency of video data while maintaining high performance. To apply powerful pre-trained models to downstream tasks, prompt learning is proposed to achieve efficient downstream task adaptation with only a small number of fine-tuned parameters. However, the sequentially compressed visual prompt tokens fail to capture the spatial and temporal contextual information in the video, thus limiting the effective propagation of spatial information within a video frame and temporal information between frames in the state compression model and the extraction of discriminative information. To tackle the above issue, we proposed a State Space Prompting (SSP) method for video understanding, which combines intra-frame and inter-frame prompts to aggregate and propagate key spatiotemporal information in the video. Specifically, an Intra-Frame Gathering (IFG) module is designed to aggregate spatial key information within each frame. Besides, an Inter-Frame Spreading (IFS) module is designed to spread discriminative spatio-temporal information across different frames. By adaptively balancing and compressing key spatio-temporal information within and between frames, our SSP effectively propagates discriminative information in videos in a complementary manner. Extensive experiments on four video benchmark datasets verify that our SSP significantly outperforms existing SOTA methods by 2.76\% on average while reducing the overhead of fine-tuning parameters.



Paperid:5297
Authors:Jing Hao, Yuxuan Fan, Yanpeng Sun, Kaixin Guo, Lin Lizhuo, Jinrong Yang, Qiyong Ai, Lun Wong, Hao Tang, Kuo Hung
Abstract:
Recent advances in large vision-language models (LVLMs) have demonstrated strong performance on general-purpose medical tasks. However, their effectiveness in specialized domains such as dentistry remains underexplored. In particular, panoramic X-rays, a widely used imaging modality in oral radiology, pose interpretative challenges due to dense anatomical structures and subtle pathological cues, which are not captured by existing medical benchmarks or instruction datasets. To this end, we introduce MMOral, the first large-scale multimodal instruction dataset and benchmark tailored for panoramic X-ray interpretation. MMOral consists of 20,563 annotated images paired with 1.3 million instruction-following instances across diverse task types, including attribute extraction, report generation, visual question answering, and image-grounded dialogue. In addition, we present MMOral-Bench, a comprehensive evaluation suite covering five key diagnostic dimensions in dentistry. We evaluate 64 LVLMs on MMOral-Bench and find that even the best-performing model, i.e., GPT-4o, only achieves 41.45% accuracy, revealing significant limitations of current models in this domain. To promote the progress of this specific domain, we provide the supervised fine-tuning (SFT) process utilizing our meticulously curated MMOral instruction dataset. Remarkably, a single epoch of SFT yields substantial performance enhancements for LVLMs, e.g., Qwen2.5-VL-7B demonstrates a 24.73% improvement. MMOral holds significant potential as a critical foundation for intelligent dentistry and enables more clinically impactful multimodal AI systems in the dental field.



Paperid:5298
Authors:Beatrix M. G. Nielsen, Emanuele Marconato, Andrea Dittadi, Luigi Gresele
Abstract:
When and why representations learned by different deep neural networks are similar is an active research topic. We choose to address these questions from the perspective of identifiability theory, which suggests that a measure of representational similarity should be invariant to transformations that leave the model distribution unchanged. Focusing on a model family which includes several popular pre-training approaches, e.g., autoregressive language models, we explore when models which generate distributions that are close have similar representations. We prove that a small Kullback--Leibler divergence between the model distributions does not guarantee that the corresponding representations are similar. This has the important corollary that models arbitrarily close to maximizing the likelihood can still learn dissimilar representations---a phenomenon mirrored in our empirical observations on models trained on CIFAR-10. We then define a distributional distance for which closeness implies representational similarity, and in synthetic experiments, we find that wider networks learn distributions which are closer with respect to our distance and have more similar representations. Our results establish a link between closeness in distribution and representational similarity.



Paperid:5299
Authors:Eray Erturk, Maryam Shanechi
Abstract:
Real-time decoding of target variables from multiple simultaneously recorded neural time-series modalities, such as discrete spiking activity and continuous field potentials, is important across various neuroscience applications. However, a major challenge for doing so is that different neural modalities can have different timescales (i.e., sampling rates) and different probabilistic distributions, or can even be missing at some time-steps. Existing nonlinear models of multimodal neural activity do not address different timescales or missing samples across modalities. Further, some of these models do not allow for real-time decoding. Here, we develop a learning framework that can enable real-time recursive decoding while nonlinearly aggregating information across multiple modalities with different timescales and distributions and with missing samples. This framework consists of 1) a multiscale encoder that nonlinearly aggregates information after learning within-modality dynamics to handle different timescales and missing samples in real time, 2) a multiscale dynamical backbone that extracts multimodal temporal dynamics and enables real-time recursive decoding, and 3) modality-specific decoders to account for different probabilistic distributions across modalities. In both simulations and two distinct multiscale brain datasets, we show that our model can aggregate information across modalities with different timescales and distributions and missing samples to improve real-time target decoding. Further, our method outperforms various linear and nonlinear multimodal benchmarks in doing so.



Paperid:5300
Authors:Sarthak Kumar Maharana, Saksham Singh Kushwaha, Baoming Zhang, Adrian Rodriguez, Songtao Wei, Yapeng Tian, Yunhui Guo
Abstract:
Abstract:While recent audio-visual models have demonstrated impressive performance, their robustness to distributional shifts at test-time remains not fully understood. Existing robustness benchmarks mainly focus on single modalities, making them insufficient for thoroughly assessing the robustness of audio-visual models. Motivated by real-world scenarios where shifts can occur $\textit{simultaneously}$ in both audio and visual modalities, we introduce $\texttt{AVROBUSTBENCH}$, a comprehensive benchmark designed to evaluate the test-time robustness of audio-visual recognition models. $\texttt{AVROBUSTBENCH}$ comprises four audio-visual benchmark datasets, $\texttt{AUDIOSET-2C}$, $\texttt{VGGSOUND-2C}$, $\texttt{KINETICS-2C}$, and $\texttt{EPICKITCHENS-2C}$, each incorporating 75 bimodal audio-visual corruptions that are $\textit{co-occurring}$ and $\textit{correlated}$. Through extensive evaluations, we observe that state-of-the-art supervised and self-supervised audio-visual models exhibit declining robustness as corruption severity increases. Furthermore, online test-time adaptation (TTA) methods, on $\texttt{VGGSOUND-2C}$ and $\texttt{KINETICS-2C}$, offer minimal improvements in performance under bimodal corruptions. We further propose $\texttt{AV2C}$, a simple TTA approach enabling on-the-fly cross-modal fusion by penalizing high-entropy samples, which achieves large improvements on $\texttt{VGGSOUND-2C}$. We hope $\texttt{AVROBUSTBENCH}$ steers the development of more effective and robust audio-visual TTA approaches. Our code is available [here](https://github.com/sarthaxxxxx/AV-C-Robustness-Benchmark).



Paperid:5301
Authors:Zhihao Wang, Yiqun Xie, Lei Ma, George Hurtt, Xiaowei Jia, Yanhua Li, Ruohan Li, Zhili Li, Shuo Xu
Abstract:
Abstract:Forest ecosystems play a critical role in the Earth systems as major carbon sinks that are essential for carbon neutralization and climate change mitigation. However, the Earth has undergone significant deforestation and forest degradation, and the remaining forested areas are also facing increasing pressures from socioeconomic factors and climate change, and could be pushed to tipping points. Responding to the grand challenge, a theory-based Ecosystem Demography (ED) model has been continuously developed over the past two decades and serves as a key component in major initiatives, including the Global Carbon Budget, NASA Carbon Monitoring System, and US Greenhouse Gas Center. Despite its growing importance in combating climate change and shaping carbon policies, ED's expensive computation significantly limits its ability to estimate carbon dynamics at the global scale with high spatial resolution. Recently, machine learning (ML) models have shown promising potential in approximating theory-based models with interesting success in various domains including weather forecasting, thanks to the open-source benchmark datasets made available. However, there is yet any publicly-available ML-ready datasets for global carbon dynamics forecasting in forest ecosystems. The limited data availability hinders the development of corresponding ML emulators. Furthermore, the inputs needed for running ED are highly complex with over a hundred variables from various remote sensing products. To bridge the gap, we develop a new ML-ready benchmark dataset, \textit{Carbon-Bench}, for carbon dynamics forecasting, featuring that: (1) the data has a global-scale coverage at 0.5$^\circ$ resolution; (2) the temporal range spans 40 years; (3) the inputs integrate extensive multi-source data from different sensing products, with calibrated outputs from ED; (4) the data is formatted in ML-ready forms and split into different evaluation scenarios based on climate conditions, etc.; (5) a set of problem-driven metrics is designed to develop benchmarks using various ML models to best align with the needs of downstream applications.



Paperid:5302
Authors:Gongyu Chen, Ying Cui
Abstract:
Abstract:We propose a generalized score-based diffusion framework for learning multivariate Gaussian mixture models with homoscedastic or heteroscedastic noise. Our goal is to nonparametrically estimate the latent location distribution and denoise the observations. Departing from the conventional maximum likelihood approach, we reinterpret each observation as temporal slices of a family of stochastic diffusion processes. This modeling choice enables a principled characterization of the additive noise structure and supports a multi-step denoising procedure grounded in reverse-time dynamics. We introduce a score-based objective that explicitly models the latent distribution and accommodates observation-specific noise covariances.Theoretically, we establish that the score estimation error with $n$ independent observations achieves a near-parametric error rate of $\frac{\mathrm{polylog}(n)}{n}$, improving upon existing results in the diffusion literature. Empirically, our method outperforms the nonparametric maximum likelihood estimator in both density estimation and denoising fidelity, especially in high-dimensional settings.These findings suggest a promising direction for integrating nonparametric empirical Bayes with diffusion-based generative modeling for latent structure recovery.



Paperid:5303
Authors:Xunzhao Yu
Abstract:
Meta-learning has been demonstrated to be useful to improve the sampling efficiency of Bayesian optimization (BO) and surrogate-assisted evolutionary algorithms (SAEAs) when solving expensive optimization problems (EOPs). Existing studies mainly focus on either combinations of existing meta-learning modeling methods with optimization algorithms, or the development of meta-learning acquisition functions for specific meta BO. However, the meta-learning models used in the literature are not designed for optimization purpose, and the generalization ability of meta-learning acquisition functions is limited.In this work, we develop a novel architecture of meta-learning model for optimization purpose and propose a generalized few-shot evolutionary optimization (FSEO) framework to solve EOPs.We focus on the scenario of expensive multi-objective EOPs (EMOPs) in the context of few-shot optimization as there are few studies on it and its high requirement on surrogate modeling performance.The surrogates in FSEO framework combines neural network with Gaussian Processes (GPs), their network parameters and some parameters of GPs represent task-independent experience and are meta-learned across related optimization tasks, the remaining GPs parameters are task-specific parameters that represent unique features of the target task.We demonstrate that our FSEO framework is able to improve the sampling efficiency of existing SAEAs on EMOPs.



Paperid:5304
Authors:Ying Li, Chengfei Lyu, Huan Wang
Abstract:
Visual AutoRegressive (VAR) modeling employs a next-scale decoding paradigm that progresses from coarse structures to fine details. While enhancing fidelity and scalability, this approach challenges two fundamental assumptions of conventional dynamic inference: semantic stability (intermediate outputs approximating final results) and monotonic locality (smooth representation evolution across layers), which renders existing dynamic inference methods ineffective for VAR models. %To address this challenge, we propose FreqExit, a unified training framework that enables dynamic inference in VAR without altering its architecture or compromising output quality. FreqExit is based on a key insight: high-frequency details are crucial for perceptual quality and tend to emerge only in later decoding stages. Leveraging this insight, we design targeted mechanisms that guide the model to learn more effectively through frequency-aware supervision. The proposed framework consists of three components: (1) a curriculum-based supervision strategy with progressive layer dropout and early exit loss; (2) a wavelet-domain high-frequency consistency loss that aligns spectral content across different generation steps; and (3) a lightweight self-supervised frequency-gated module that guides adaptive learning of both structural and detailed spectral components.On ImageNet 256×256, FreqExit achieves up to 2× speedup with only minor degradation, and delivers 1.3× acceleration without perceptible quality loss. This enables runtime-adaptive acceleration within a unified model, offering a favorable trade-off between efficiency and fidelity for for practical and flexible deployment.



Paperid:5305
Authors:Brandon Man, Ghadi Nehme, Md Ferdous Alam, Faez Ahmed
Abstract:
Abstract:Computer-Aided Design (CAD) is a time-consuming and complex process, requiring precise, long-horizon user interactions with intricate 3D interfaces. While recent advances in AI-driven user interface (UI) agents show promise, most existing datasets and methods focus on short, low-complexity tasks in mobile or web applications, failing to capture the demands of professional engineering tools. In this work, we introduce VideoCAD, the first attempt at engineering UI interaction learning for precision tasks. Specifically, VideoCAD is a large-scale synthetic dataset consisting of over 41K annotated video recordings of CAD operations, generated using an automated framework for collecting high-fidelity UI action data from human-made CAD designs. Compared to existing datasets, VideoCAD offers an order of magnitude higher complexity in UI interaction learning for real-world engineering tasks, having up to a $20\times$ longer time horizon than other datasets. We show two important downstream applications of VideoCAD: learning UI interactions from professional precision 3D CAD tools and a visual question-answering (VQA) benchmark designed to evaluate multimodal large language models' (LLM) spatial reasoning and video understanding abilities. To learn the UI interactions, we propose VideoCADFormer - a state-of-the-art model in learning CAD interactions directly from video, which outperforms multiple behavior cloning baselines. Both VideoCADFormer and the VQA benchmark derived from VideoCAD reveal key challenges in the current state of video-based UI understanding, including the need for precise action grounding, multi-modal and spatial reasoning, and long-horizon dependencies. Dataset and code available at: https://github.com/BrandonMan123/VideoCAD.



Paperid:5306
Authors:Mélisande Teng, Arthur Ouaknine, Etienne Laliberté, Yoshua Bengio, David Rolnick, Hugo Larochelle
Abstract:
Information on trees at the individual level is crucial for monitoring forest ecosystems and planning forest management. Current monitoring methods involve ground measurements, requiring extensive cost, time and labour. Advances in drone remote sensing and computer vision offer great potential for mapping individual trees from aerial imagery at broad-scale. Large pre-trained vision models, such as the Segment Anything Model (SAM), represent a particularly compelling choice given limited labeled data. In this work, we compare methods leveraging SAM for the task of automatic tree crown instance segmentation in high resolution drone imagery in three use cases: 1) boreal plantations, 2) temperate forests, and 3) tropical forests. We also look into integrating elevation data into models, in the form of Digital Surface Model (DSM) information, which can readily be obtained at no additional cost from RGB drone imagery. We present BalSAM, a model leveraging SAM and DSM information, which shows potential over other methods, particularly in the context of plantations. We find that methods using SAM out-of-the-box do not outperform a custom Mask R-CNN, even with well-designed prompts. However, efficiently tuning SAM further and integrating DSM information are both promising avenues for tree crown instance segmentation models.



Paperid:5307
Authors:Alejandro García-Castellanos, David Wessels, Nicky J. van den Berg, Remco Duits, Daniel M. Pelt, Erik Bekkers
Abstract:
We introduce Equivariant Neural Eikonal Solvers, a novel framework that integrates Equivariant Neural Fields (ENFs) with Neural Eikonal Solvers. Our approach employs a single neural field where a unified shared backbone is conditioned on signal-specific latent variables – represented as point clouds in a Lie group – to model diverse Eikonal solutions. The ENF integration ensures equivariant mapping from these latent representations to the solution field, delivering three key benefits: enhanced representation efficiency through weight-sharing, robust geometric grounding, and solution steerability. This steerability allows transformations applied to the latent point cloud to induce predictable, geometrically meaningful modifications in the resulting Eikonal solution. By coupling these steerable representations with Physics-Informed Neural Networks (PINNs), our framework accurately models Eikonal travel-time solutions while generalizing to arbitrary Riemannian manifolds with regular group actions. This includes homogeneous spaces such as Euclidean, position–orientation, spherical, and hyperbolic manifolds. We validate our approach through applications in seismic travel-time modeling of 2D and 3D benchmark datasets. Experimental results demonstrate superior performance, scalability, adaptability, and user controllability compared to existing Neural Operator-based Eikonal solver methods.



Paperid:5308
Authors:Jiarui Yao, Yifan Hao, Hanning Zhang, Hanze Dong, Wei Xiong, Nan Jiang, Tong Zhang
Abstract:
Chain-of-thought (CoT) reasoning in large language models (LLMs) can be formalized as a latent variable problem, where the model needs to generate intermediate reasoning steps. While prior approaches such as iterative reward-ranked fine-tuning (RAFT) have relied on such formulations, they typically apply uniform inference budgets across prompts, which fails to account for variability in difficulty and convergence behavior. This work identifies the main bottleneck in CoT training as inefficient stochastic gradient estimation due to static sampling strategies. We propose GVM-RAFT, a prompt-specific Dynamic Sample Allocation Strategy designed to minimize stochastic gradient variance under a computational budget constraint. The method dynamically allocates computational resources by monitoring prompt acceptance rates and stochastic gradient norms, ensuring that the resulting gradient variance is minimized. Our theoretical analysis shows that the proposed dynamic sampling strategy leads to accelerated convergence guarantees under suitable conditions. Experiments on mathematical reasoning show that GVM-RAFT achieves a 2-4x speedup and considerable accuracy improvements over vanilla RAFT. The proposed dynamic sampling strategy is general and can be incorporated into other reinforcement learning algorithms, such as GRPO, leading to similar improvements in convergence and test accuracy.



Paperid:5309
Authors:Yang Chen, Yirun Zhou, Weizhong Zhang, Cheng Jin
Abstract:
Point cloud completion, aiming to reconstruct missing part from incomplete point clouds, is a pivotal task in 3D computer vision. Traditional supervised approaches often necessitate complete point clouds for training supervision, which are not readily accessible in real-world applications. Recent studies have attempted to mitigate this dependency by employing self-supervise mechanisms. However, these approaches frequently yield suboptimal results due to the absence of complete structure in the point cloud data during training. To address these issues, in this paper, we propose an effective framework to complete the point cloud under the guidance of self learned complete structure. A key contribution of our work is the development of a novel self-supervised complete structure reconstruction module, which can learn the complete structure explicitly from incomplete point clouds and thus eliminate the reliance on training data from complete point clouds. Additionally, we introduce a contrastive learning approach at both the cluster- and instance-level to extract shape features guided by the complete structure and to capture style features, respectively. This dual-level learning design ensures that the generated point clouds are both shape-completed and detail-preserving. Extensive experiments on both synthetic and real-world datasets demonstrate that our approach significantly outperforms state-of-the-art self-supervised methods.



Paperid:5310
Authors:Shuai Huang, Huan Luo, Haodong Jing, Qixian Zhang, Litao Chang, Yating Feng, Xiao Lin, Chendong Qin, Han Chen, Shuwen Jia, Siyi Sun, Yongxiong Wang
Abstract:
Translating brain activity into meaningful visual content has long been recognized as a fundamental challenge in neuroscience and brain-computer interface research. Recent advances in EEG-based neural decoding have shown promise, yet two critical limitations remain in this area: poor generalization across subjects and constraints to specific visual tasks. We introduce NEED, the first unified framework achieving zero-shot cross-subject and cross-task generalization for EEG-based visual reconstruction. Our approach addresses three fundamental challenges: (1) cross-subject variability through an Individual Adaptation Module pretrained on multiple EEG datasets to normalize subject-specific patterns, (2) limited spatial resolution and complex temporal dynamics via a dual-pathway architecture capturing both low-level visual dynamics and high-level semantics, and (3) task specificity constraints through a unified inference mechanism adaptable to different visual domains. For video reconstruction, NEED achieves better performance than existing methods. Importantly, Our model maintains 93.7% of within-subject classification performance and 92.4% of visual reconstruction quality when generalizing to unseen subjects, while achieving an SSIM of 0.352 when transferring directly to static image reconstruction without fine-tuning, demonstrating how neural decoding can move beyond subject and task boundaries toward truly generalizable brain-computer interfaces.



Paperid:5311
Authors:Paul Gölz, Nika Haghtalab, Kunhe Yang
Abstract:
Abstract:After pre-training, large language models are aligned with human preferences based on pairwise comparisons. State-of-the-art alignment methods (such as PPO-based RLHF and DPO) are built on the assumption of aligning with a single preference model, despite being deployed in settings where users have diverse preferences. As a result, it is not even clear that these alignment methods produce models that satisfy users \emph{on average} --- a minimal requirement for pluralistic alignment. Drawing on social choice theory and modeling users' comparisons through individual Bradley-Terry (BT) models, we introduce an alignment method's \emph{distortion}: the worst-case ratio between the optimal achievable average utility, and the average utility of the learned policy. The notion of distortion helps draw sharp distinctions between alignment methods: \emph{Nash Learning from Human Feedback} achieves the minimax optimal distortion of $(\frac{1}{2} + o(1)) \cdot \beta$ (for the BT temperature $\beta$), robustly across utility distributions, distributions of comparison pairs, and permissible KL divergences from the reference policy. RLHF and DPO, by contrast, suffer $\geq (1 - o(1)) \cdot \beta$ distortion already without a KL constraint, and $e^{\Omega(\beta)}$ or even unbounded distortion in the full setting, depending on how comparison pairs are sampled.



Paperid:5312
Authors:Wenhao Li, Yuxin Zhang, Gen Luo, Haiyuan Wan, Ziyang Gong, Fei Chao, Rongrong Ji
Abstract:
Abstract:Reducing the key-value (KV) cache burden in Large Language Models (LLMs) significantly accelerates inference. Dynamically selecting critical KV caches during decoding helps maintain performance. Existing methods use random linear hashing to identify important tokens, but this approach is inefficient due to the orthogonal distribution of queries and keys within two narrow cones in LLMs. We introduce Spotlight Attention, a novel method that employs non-linear hashing functions to optimize the embedding distribution of queries and keys, enhancing coding efficiency and robustness. We also developed a lightweight, stable training framework using a Bradley-Terry ranking-based loss, enabling optimization of the non-linear hashing module on GPUs with 16GB memory in 8 hours. Experimental results show that Spotlight Attention drastically improves retrieval precision while shortening the length of the hash code at least 5$\times$ compared to traditional linear hashing. Finally, we exploit the computational advantages of bitwise operations by implementing specialized CUDA kernels, achieving hashing retrieval for 512K tokens in under 100$\mu$s on a single A100 GPU, with end-to-end throughput up to 3$\times$ higher than vanilla decoding.



Paperid:5313
Authors:Xumai Qi, Dongdong Zhang, Taotao Liu, Hongcheng Wang
Abstract:
In recent years, many studies proposed to generate solutions for Directed Acyclic Graph (DAG) scheduling problem in one shot by combining reinforcement learning and list scheduling heuristic. However, these existing methods suffer from biased estimation of sampling probabilities and inefficient guidance in training, due to redundant comparisons among node priorities and the sparse reward challenge. To address these issues, we analyze of the limitations of these existing methods, and propose a novel one-shot DAG scheduling method with comparability identification and dense reward signal, based on the policy gradient framework. In our method, a comparable antichain identification mechanism is proposed to eliminate the problem of redundant nodewise priority comparison. We also propose a dense reward signal for node level decision-making optimization in training, effectively addressing the sparse reward challenge. The experimental results show that the proposed method can yield superior results of scheduling objectives compared to other learning-based DAG scheduling methods.



Paperid:5314
Authors:Charles W J Pugh, Paulina Nuñez-Valencia, Mafalda Dias, Jonathan Frazer
Abstract:
Generative models trained on natural sequences are increasingly used to predict the effects of genetic variation, enabling progress in therapeutic design, disease risk prediction, and synthetic biology. In the zero-shot setting, variant impact is estimated by comparing the likelihoods of sequences, under the assumption that likelihood serves as a proxy for fitness. However, this assumption often breaks down in practice: sequence likelihood reflects not only evolutionary fitness constraints, but also phylogenetic structure and sampling biases, especially as model capacity increases. We introduce Likelihood-Fitness Bridging (LFB), a simple and general strategy that improves variant effect prediction by averaging model scores across sequences subject to similar selective pressures. Assuming an Ornstein–Uhlenbeck model of evolution, LFB can be viewed as a way to marginalize the effects of genetic drift, although its benefits appear to extend more broadly. LFB applies to existing protein and genomic language models without requiring retraining, and incurs only modest computational overhead. Evaluated on large-scale deep mutational scans and clinical benchmarks, LFB consistently improves predictive performance across model families and sizes. Notably, it reverses the performance plateau observed in larger protein language models, making the largest models the most accurate when combined with LFB. These results suggest that accounting for phylogenetic and sampling biases is essential to realizing the full potential of large sequence models in variant effect prediction.



Paperid:5315
Authors:Huacan Wang, Ziyi Ni, Shuo Zhang, Shuo Lu, Chen Hu, Sen Hu, Yuntao Du, Ziyang He, Jiaye Lin, Yifu Guo, Pin Lyu
Abstract:
The ultimate goal of code agents is to solve complex tasks autonomously. Although large language models (LLMs) have made substantial progress in code generation, real-world tasks typically demand full-fledged code repositories rather than simple scripts. Building such repositories from scratch remains a major challenge. Fortunately, GitHub hosts a vast, evolving collection of open-source repositories, which developers frequently reuse as modular components for complex tasks. Yet, existing frameworks like OpenHands and SWE-Agent still struggle to effectively leverage these valuable resources.Relying solely on README files provides insufficient guidance, and deeper exploration reveals two core obstacles: overwhelming information and tangled dependencies of repositories, both constrained by the limited context windows of current LLMs.To tackle these issues, we propose RepoMaster, an autonomous agent framework designed to explore and reuse GitHub repositories for solving complex tasks. For efficient understanding, RepoMaster constructs function-call graphs, module-dependency graphs, and hierarchical code trees to identify essential components, providing only identified core elements to the LLMs rather than the entire repository. During autonomous execution, it progressively explores related components using our exploration tools and prunes information to optimize context usage.Evaluated on the adjusted MLE-bench, RepoMaster achieves a 110% relative boost in valid submissions over the strongest baseline OpenHands. On our newly released GitTaskBench, RepoMaster lifts the task-pass rate from 24.1% to 62.9% while reducing token usage by 90%.



Paperid:5316
Authors:Saghir Alfasly, Wataru Uegami, MD ENAMUL HOQ, Ghazal Alabtah, Hamid Tizhoosh
Abstract:
Synthetic data generation in histopathology faces unique challenges: preserving tissue heterogeneity, capturing subtle morphological features, and scaling to unannotated datasets. We present a latent diffusion model that generates realistic heterogeneous histopathology images through a novel dual-conditioning approach combining semantic segmentation maps with tissue-specific visual crops. Unlike existing methods that rely on text prompts or abstract visual embeddings, our approach preserves critical morphological details by directly incorporating raw tissue crops from corresponding semantic regions. For annotated datasets (i.e., Camelyon16, Panda), we extract patches ensuring 20-80\% tissue heterogeneity. For unannotated data (i.e., TCGA), we introduce a self-supervised extension that clusters whole-slide images into 100 tissue types using foundation model embeddings, automatically generating pseudo-semantic maps for training. Our method synthesizes high-fidelity images with precise region-wise annotations, achieving superior performance on downstream segmentation tasks. When evaluated on annotated datasets, models trained on our synthetic data show competitive performance to those trained on real data, demonstrating the utility of controlled heterogeneous tissue generation. In quantitative evaluation, prompt‐guided synthesis reduces Fréchet Distance by up to 6× on Camelyon16 (from 430.1 to 72.0) and yields 2–3× lower FD across Panda and TCGA. Downstream DeepLabv3+ models trained solely on synthetic data attain test IoU of 0.71 and 0.95 on Camelyon16 and Panda, within 1–2% of real‐data baselines (0.72 and 0.96). By scaling to 11,765 TCGA whole‐slide images without manual annotations, our framework offers a practical solution for an urgent need for generating diverse, annotated histopathology data, addressing a critical bottleneck in computational pathology.



Paperid:5317
Authors:Guoqing Hu, An Zhang, Shuchang Liu, Wenyu Mao, Jiancan Wu, Xun Yang, Xiang Li, Lantao Hu, Han Li, Kun Gai, Xiang Wang
Abstract:
Recommenders aim to rank items from a discrete item corpus in line with user interests, yet suffer from extremely sparse user preference data. Recent advances in diffusion models have inspired diffusion-based recommenders, which alleviate sparsity by injecting noise during a forward process to prevent collapse of perturbed preference distributions. However, current diffusion‑based recommenders predominantly rely on continuous Gaussian noise, which is intrinsically mismatched with the discrete nature of user preference data in recommendation. In this paper, building upon recent advances in discrete diffusion, we propose \textbf{PreferGrow}, a discrete diffusion-based recommender modeling preference ratios by fading and growing user preferences over the discrete item corpus. PreferGrow differs from existing diffusion-based recommenders in three core aspects: (1) Discrete modeling of preference ratios:PreferGrow models relative preference ratios between two items, where a positive value indicates a more preferred one over another less preferred. This formulation aligns naturally with the discrete and ranking-oriented nature of recommendation tasks.(2) Perturbing via preference fading: Instead of injecting continuous noise, PreferGrow fades user preferences by replacing the preferred item with alternatives---physically akin to negative sampling---thereby eliminating the need for any prior noise assumption.(3) Preference reconstruction via growing: PreferGrow reconstructs user preferences by iteratively growing the preference signal from the estimated ratios. We further provide theoretical analysis showing that PreferGrow preserves key properties of discrete diffusion processes.PreferGrow provides a well-defined matrix‑based formulation for discrete diffusion-based recommendation and empirically outperforms existing diffusion‑based recommenders across five benchmark datasets, underscoring its superior effectiveness.Our codes are available at \url{https://anonymous.4open.science/r/PreferGrow_Commit-2259/}.



Paperid:5318
Authors:Nan Bao, Yifan Zhao, Lin Zhu, Jia Li
Abstract:
Semantic segmentation has achieved great success in ideal conditions. However, when facing extreme conditions (e.g., insufficient light, fierce camera motion), most existing methods suffer from significant information loss of RGB, severely damaging segmentation results. Several researches exploit the high-speed and high-dynamic event modality as a complement, but event and RGB are naturally heterogeneous, which leads to feature-level mismatch and inferior optimization of existing multi-modality methods. Different from these researches, we delve into the edge secret of both modalities for resilient fusion and propose a novel Edge-awareness Semantic Concordance framework to unify the multi-modality heterogeneous features with latent edge cues. In this framework, we first propose Edge-awareness Latent Re-coding, which obtains uncertainty indicators while realigning event-RGB features into unified semantic space guided by re-coded distribution, and transfers event-RGB distributions into re-coded features by utilizing a pre-established edge dictionary as clues. We then propose Re-coded Consolidation and Uncertainty Optimization, which utilize re-coded edge features and uncertainty indicators to solve the heterogeneous event-RGB fusion issues under extreme conditions. We establish two synthetic and one real-world event-RGB semantic segmentation datasets for extreme scenario comparisons. Experimental results show that our method outperforms the state-of-the-art by a 2.55% mIoU on our proposed DERS-XS, and possesses superior resilience under spatial occlusion.



Paperid:5319
Authors:Denis Blessing, Julius Berner, Lorenz Richter, Carles Domingo i Enrich, Yuanqi Du, Arash Vahdat, Gerhard Neumann
Abstract:
Solving stochastic optimal control problems with quadratic control costs can be viewed as approximating a target path space measure, e.g. via gradient-based optimization. In practice, however, this optimization is challenging in particular if the target measure differs substantially from the prior. In this work, we therefore approach the problem by iteratively solving constrained problems incorporating trust regions that aim for approaching the target measure gradually in a systematic way. It turns out that this trust region based strategy can be understood as a geometric annealing from the prior to the target measure, where, however, the incorporated trust regions lead to a principled and educated way of choosing the time steps in the annealing path. We demonstrate in multiple optimal control applications that our novel method can improve performance significantly, including tasks in diffusion-based sampling and fine-tuning of diffusion models.



Paperid:5320
Authors:Jiesong Liu, Xipeng Shen
Abstract:
Image generation requires intensive computations and faces challenges due to long latency. Exploiting redundancy in the input images and intermediate representations throughout the neural network pipeline is an effective way to accelerate image generation. Token merging (ToMe) exploits similarities among input tokens by clustering them and merges similar tokens into one, thus significantly reducing the number of tokens that are fed into the transformer block. This work introduces Fourier Token Merging, a new method for understanding and capitalizing frequency domain for efficient image generation. By introducing frequency token merging, we find that transforming the token into the frequency domain representation for clustering can better exert the ability of clustering based on the underlying redundancy after de-correlation. Through analytical and empirical studies, we demonstrate the benefits of using Fourier clustering over the original time domain clustering. We experimented fourier token merging on the stable diffusion model, and the results show up to 25\% reduction in latency without impairing image quality.



Paperid:5321
Authors:Zhihang Liu, Chen-Wei Xie, Bin Wen, Feiwu Yu, JixuanChen, Pandeng Li, Boqiang Zhang, Nianzu Yang, YingluLi, Zuan Gao, Yun Zheng, Hongtao Xie
Abstract:
Abstract:Visual captioning benchmarks have become outdated with the emergence of modern multimodal large language models (MLLMs), as the brief ground-truth sentences and traditional metrics fail to assess detailed captions effectively. While recent benchmarks attempt to address this by focusing on keyword extraction or object-centric evaluation, they remain limited to vague-view or object-view analyses and incomplete visual element coverage. In this paper, we introduce CAPability, a comprehensive multi-view benchmark for evaluating visual captioning across 12 dimensions spanning six critical views. We curate nearly 11K human-annotated images and videos with visual element annotations to evaluate the generated captions. CAPability stably assesses both the correctness and thoroughness of captions with \textit{precision} and \textit{hit} metrics. By converting annotations to QA pairs, we further introduce a heuristic metric, \textit{know but cannot tell} ($K\bar{T}$), indicating a significant performance gap between QA and caption capabilities. Our work provides a holistic analysis of MLLMs' captioning abilities, as we identify their strengths and weaknesses across various dimensions, guiding future research to enhance specific aspects of their capabilities.



Paperid:5322
Authors:Zachary Charles, Gabriel Teston, Lucio Dery, John Rush, Nova Fallen, Zachary Garrett, Arthur Szlam, Arthur Douillard
Abstract:
As we scale to more massive machine learning models, the frequent synchronization demands inherent in data-parallel approaches create significant slowdowns, posing a critical challenge to further scaling. Recent work develops an approach (DiLoCo) that relaxes synchronization demands without compromising model quality. However, these works do not carefully analyze how DiLoCo's behavior changes with model size. In this work, we study the scaling law behavior of DiLoCo when training LLMs under a fixed compute budget. We focus on how algorithmic factors, including number of model replicas, hyperparameters, and token budget affect training in ways that can be accurately predicted via scaling laws. We find that DiLoCo scales both predictably and robustly with model size. When well-tuned, DiLoCo scales better than data-parallel training with model size, and can outperform data-parallel training even at small model sizes. Our results showcase a more general set of benefits of DiLoCo than previously documented, including increased optimal batch sizes, improved downstream generalization with scale, and improved evaluation loss for a fixed token budget.



Paperid:5323
Authors:Weizhong Huang, Yuxin Zhang, Xiawu Zheng, Fei Chao, Liujuan Cao, Rongrong Ji
Abstract:
Mixture-of-Experts (MoE) architectures enable efficient scaling of large language models but face prohibitive memory demands due to massive parameterization. Existing pruning methods rely on heuristic metrics or impractical enumeration of expert subsets, leading to suboptimal performance or scalability. In this paper, we propose Shapley-MoE, an efficient pruning method for MoE models inspired by cooperative game theory. By quantifying each expert’s contribution via Shapley value, our method identifies important experts without exhaustive combination evaluations. To overcome the NP-hard complexity of exact Shapley computation, we introduce a Monte Carlo sampling strategy for efficient approximation that reduces complexity to quadratic time. However, vanilla Monte Carlo sampling still faces issues of insufficient estimation accuracy and low sampling efficiency. To address these issues, we further propose two novel methods to improve sampling accuracy and efficiency: (1) Early Truncation, which early terminates unstable sampling steps caused by overly small expert subsets, and (2) Router-Guided Importance Sampling, which prioritize sampling important expert subsets using gating activation probabilities. Both theoretical and experimental analyses show that both methods can accelerate Shapley value estimation and improve accuracy. Extensive empirical evaluations demonstrate that our pruned MoE models outperform existing expert pruning methods. Notably, when applied to the Qwen2-57B-A14B model, our method reduces the number of experts by 25% with only a 0.92 increase in perplexity and over 96.4% of the average zero-shot accuracy is maintained. Code is available in the supplementary material.



Paperid:5324
Authors:Naoki Nishikawa, Rei Higuchi, Taiji Suzuki
Abstract:
Linear attention has attracted interest as a computationally efficient approximation to softmax attention, especially for long sequences. Recent studies has explored distilling softmax attention in pre-trained Transformers into linear attention. However, a critical challenge remains:how to choose the feature dimension that governs the approximation quality. Existing methods fix this dimension uniformly across all attention layers, overlooking the diverse roles and complexities of them. In this paper, we propose a principled method to automatically determine the feature dimension in linear attention using the concept of statisticaldegrees of freedom, which represent the effective dimensionality of the inputs. We provide a theoretical bound on the approximation error and show that the dimension chosen by our method achieves smaller error under a fixed computational budget. Furthermore, we introduce an efficient layerwise training strategy to learn nonlinear features tailored to each layer. Experiments on multiple pre-trained transformers demonstrate that our method improves the performance of distilled models compared to baselines without increasing the inference cost. Our findings also provide insight into how the complexity of the attention mechanism evolves across layers.



Paperid:5325
Authors:Kuat Gazizov, Miguel A. Carreira-Perpinan
Abstract:
We consider the Tree Alternating Optimization (TAO) algorithm to train regression trees with linear predictors in the leaves. Unlike the traditional, greedy recursive partitioning algorithms such as CART, TAO guarantees a monotonic decrease of the objective function and results in smaller trees of much better accuracy. We modify the TAO algorithm so that it produces exactly the same result but is much faster, particularly for high input dimensionality or deep trees. The idea is based on the fact that, at each iteration of TAO, each leaf receives only a subset of the training instances. Thus, the optimization of the leaf model can be done exactly but faster by using the Sherman-Morrison-Woodbury formula. This has the unexpected advantage that, once a tree exceeds a critical depth, then making it deeper makes it faster to train, even though the tree is larger and has more parameters. Indeed, this can make learning a nonlinear model (the tree) asymptotically faster than a regular linear regression model. We analyze the corresponding computational complexity and verify the speedups experimentally in various datasets. The argument can be applied to other types of trees, whenever the optimization of a node can be computed in superlinear time of the number of instances.



Paperid:5326
Authors:Xiaoyu Yang, EN YU, Jie Lu
Abstract:
This paper uncovers a critical yet overlooked phenomenon in multi-modal large language models (MLLMs): detrimental concept drift within chain-of-thought (CoT) reasoning during non-stationary reinforcement fine-tuning (RFT), where reasoning token distributions evolve unpredictably, thereby introducing significant biases in final predictions. To address this, we are pioneers in establishing the theoretical bridge between concept drift theory and RFT processes by formalizing CoT's autoregressive token streams as non-stationary distributions undergoing arbitrary temporal shifts. Leveraging this framework, we propose a novel counterfact-aware RFT that systematically decouples beneficial distribution adaptation from harmful concept drift through concept graph-empowered LLM experts generating counterfactual reasoning trajectories. Our solution, Counterfactual Preference Optimization (CPO), enables stable RFT in non-stationary environments, particularly within the medical domain, through custom-tuning of counterfactual-aware preference alignment. Extensive experiments demonstrate our superior performance of robustness, generalization and coordination within RFT. Besides, we also contributed a large-scale dataset CXR-CounterFact (CCF), comprising 320,416 meticulously curated counterfactual reasoning trajectories derived from MIMIC-CXR. Our code and data are public at: https://anonymous.4open.science/r/CPO-FD61/.



Paperid:5327
Authors:Tianyi (Alex) Qiu, Zhonghao He, Hirokazu Shirado, Maarten Sap
Abstract:
Recent advances in reasoning techniques have substantially improved the performance of large language models (LLMs), raising expectations for their ability to provide accurate, truthful, and reliable information. However, emerging evidence suggests that iterative reasoning may foster belief entrenchment and confirmation bias, rather than enhancing truth-seeking behavior. In this study, we propose a systematic evaluation framework forbelief entrenchmentin LLM reasoning by leveraging the Martingale property from Bayesian statistics. This property implies that, under rational belief updating, the expected value of future beliefs should remain equal to the current belief, i.e., belief updates are unpredictable from the current belief. We propose the unsupervised, regression-basedMartingale Scoreto measure violations of this property, which signal deviation from the Bayesian ability of updating on new evidence. In open-ended problem domains including event forecasting, value-laden questions, and academic paper review, we find such violations to be widespread across models and setups, where the current belief positively predicts future belief updates, a phenomenon which we termbelief entrenchment. We identify the models, reasoning techniques, and domains more prone to belief entrenchment. Finally, we validate the Martingale Score by showing that it predicts ground-truth accuracy on problem domains where ground truth labels are available. This indicates that, while designed as an unsupervised metric that operates even in domains without access to ground truth, the Martingale Score is a useful proxy of the truth-seeking ability of a reasoning process.



Paperid:5328
Authors:lixiong Qin, Shilong Ou, Miaoxuan Zhang, Jiangning Wei, Yuhang Zhang, Xiaoshuai Song, Yuchen Liu, Mei Wang, Weiran Xu
Abstract:
Faces and humans are crucial elements in social interaction and are widely included in everyday photos and videos. Therefore, a deep understanding of faces and humans will enable multi-modal assistants to achieve improved response quality and broadened application scope. Currently, the multi-modal assistant community lacks a comprehensive and scientific evaluation of face and human understanding abilities. In this paper, we first propose a hierarchical ability taxonomy that includes three levels of abilities. Then, based on this taxonomy, we collect images and annotations from publicly available datasets in the face and human community and build a semi-automatic data pipeline to produce problems for the new benchmark. Finally, the obtained Face-Human-Bench includes a development set and a test set, each with 1800 problems, supporting both English and Chinese. We conduct evaluations over 25 mainstream multi-modal large language models (MLLMs) with our Face-Human-Bench, focusing on the correlation between abilities, the impact of the relative position of targets on performance, and the impact of Chain of Thought (CoT) prompting on performance. We also explore which abilities of MLLMs need to be supplemented by specialist models.



Paperid:5329
Authors:Qiong Wu, Wenhao Lin, Yiyi Zhou, Weihao Ye, Zhanpeng Zeng, Xiaoshuai Sun, Rongrong Ji
Abstract:
In this paper, we study the visual redundancy problem of multimodal large language models (MLLMs) from the perspective of attention behaviors. Via extensive empirical experiments, we observe and conclude three main inference stages of MLLMs:(i) Early fusion between tokens is first accomplished quickly. (ii) Intra-modality modeling then comes to play. (iii) Multimodal reasoning resumes and lasts until the end of inference. In particular, we reveal that visual tokens will stop contributing to reasoning when the text tokens receive enough image information.Based on this observation, we propose an effective method to improve the efficiency of MLLMs, termed dynamic visual-token exit (DyVTE), which is orthogonal but collaborative to previous token-wise visual compression methods.To validate the efficacy of DyVTE, we apply it to a set of MLLMs, including LLaVA, VILA, EAGLE and InternVL.The experimental results not only show the effectiveness of our DyVTE in improving MLLMs' efficiency, e.g., DyVTE reduces the computation overhead of LLaVA-1.5 by up to 45.7% without performance drop, but also reveal a general pattern across multiple MLLMs, well facilitating the in-depth analysis of MLLMs. Our code is anonymously released at https://anonymous.4open.science/r/AnonymousDyVTE-26AB/.



Paperid:5330
Authors:Yuan Zang, Zitian Tang, Junho Cho, Jaewook Yoo, Chen Sun
Abstract:
Recognizing the physical states of objects and their transformations within videos is crucial for structured video understanding and enabling robust real-world applications, such as robotic manipulation. However, pretrained vision-language models often struggle to capture these nuanced dynamics and their temporal context, and specialized object state recognition frameworks struggle with generalizing to unseen actions or objects. We introduce SAGE (State-Action Graph Embeddings), a novel framework that offers a unified model of physical state transitions by decomposing states into fine-grained, language-described visual concepts that are sharable across different objects and actions. SAGE initially leverages Large Language Models to construct a State-Action Graph, which is then multimodally refined using Vision-Language Models. Extensive experimental results show that our method significantly outperforms existing baselines, generalizes effectively to unseen objects and actions in open-world settings. Our method improves the prior state-of-the-art by as much as14.6\%on novel state recognition with less than5\%of its inference time. Our code and data will be publicly released.



Paperid:5331
Authors:Georgios Tzannetos, Parameswaran Kamalaruban, Adish Singla
Abstract:
Training agents to operate under strict constraints during deployment—such as limited resource budgets or stringent safety requirements—presents significant challenges, especially when these constraints render the task complex. In this work, we propose a curriculum learning strategy that gradually tightens constraints during training, enabling the agent to incrementally master the deployment requirements. Inspired by self-paced learning techniques in unconstrained reinforcement learning (RL), our approach facilitates a smoother transition to challenging environments by initially training on simplified versions of the constraints and progressively introducing the full deployment conditions. We provide a theoretical analysis using an RL agent in a binary-tree Markov Decision Process (MDP) to demonstrate that our curriculum strategy can accelerate training relative to a baseline approach that imposes the trajectory constraints from the outset. Moreover, we empirically validate the effectiveness and generality of our method across agents—both RL and large language model (LLM) agents—and diverse settings, including a binary-tree MDP, a multi-task navigation domain, and a math reasoning task with two benchmarks. These results highlight the potential of curriculum design in enhancing the efficiency and performance of agents operating under complex trajectory constraints during deployment. Additionally, applying our strategy to LLMs, the output tokens can be compressed, achieving a 4.5× inference speedup on consumer hardware, showing its effectiveness for resource-constrained deployment.



Paperid:5332
Authors:Hongquan He, Zhen Wang, Jingya Wang, Tao Wu, Xuming He, Bei Yu, Jingyi Yu, Hao GENG
Abstract:
Abstract:Lithography orchestrates a symphony of light, mask and photochemicals to transfer the integrated circuit patterns onto the wafer. Lithography simulation serves as the critical nexus between circuit design and manufacturing, where its speed and accuracy fundamentally govern the optimization quality of downstream resolution enhancement techniques (RET). While machine learning promises to circumvent computational limitations of lithography process through data-driven or physics-informed approximations of computational lithography, existing simulators suffer from inadequate lithographic awareness due to insufficient training data capturing essential process variations and mask correction rules. We present LithoSim, the most comprehensive lithography simulation benchmark to date, featuring over $4$ million high-resolution input-output pairs with rigorous physical correspondence. The dataset systematically incorporates alterable optical source distributions, metal and via mask topologies with optical proximity correction (OPC) variants, and process windows reflecting fab-realistic variations. By integrating domain-specific metrics spanning AI performance and lithographic fidelity, LithoSim establishes a unified evaluation framework for data-driven and physics-informed computational lithography. The data (https://huggingface.co/datasets/grandiflorum/LithoSim), code (https://dw-hongquan.github.io/LithoSim), and pre-trained models (https://huggingface.co/grandiflorum/LithoSim) are released openly to support the development of hybrid ML-based and high-fidelity lithography simulation for the benefit of semiconductor manufacturing.



Paperid:5333
Authors:WEIMING ZHANG, Dingwen Xiao, Aobotao DAI, Yexin Liu, Tianbo Pan, Shiqi Wen, Lei Chen, Lin Wang
Abstract:
360 video captures the complete surrounding scenes with the ultra-large field of view of 360x180. This makes 360 scene understanding tasks,e.g., segmentation and tracking, crucial for appications, such as autonomous driving, robotics. With the recent emergence of foundation models, the community is, however, impeded by the lack of large-scale, labelled real-world datasets. This is caused by the inherent spherical properties,e.g., severe distortion in polar regions, and content discontinuities, rendering the annotation costly yet complex. This paper introducesLeader360V, thefirstlarge-scale (10K+), labeled real-world 360 video datasets for instance segmentation and tracking. Our datasets enjoy high scene diversity, ranging from indoor and urban settings to natural and dynamic outdoor scenes. To automate annotation, we design an automatic labeling pipeline, which subtly coordinates pre-trained 2D segmentors and large language models (LLMs) to facilitate the labeling. The pipeline operates in three novel stages. Specifically, in theInitial Annotation Phase, we introduce a Semantic- and Distortion-aware Refinement (SDR) module, which combines object mask proposals from multiple 2D segmentors with LLM-verified semantic labels. These are then converted into mask prompts to guide SAM2 in generating distortion-aware masks for subsequent frames. In theAuto-Refine Annotation Phase, missing or incomplete regions are corrected either by applying the SDR again or resolving the discontinuities near the horizontal borders. TheManual Revision Phasefinally incorporates LLMs and human annotators to further refine and validate the annotations. Extensive user studies and evaluations demonstrate the effectiveness of our labeling pipeline. Meanwhile, experiments confirm that Leader360V significantly enhances model performance for 360 video segmentation and tracking, paving the way for more scalable 360 scene understanding. We release our dataset and code at {https://leader360v.github.io/Leader360V_HomePage/} for anonymous review.



Paperid:5334
Authors:Duo Zhou, Jorge Chavez, Hesun Chen, Grani A. Hanasusanto, Huan Zhang
Abstract:
Abstract:State-of-the-art neural network verifiers demonstrate that applying the branch-and-bound (BaB) procedure with fast bounding techniques plays a key role in tackling many challenging verification properties. In this work, we introduce the \emph{linear constraint-driven clipping} framework, a class of scalable and efficient methods to enhance bound propagation verifiers. Under this framework, we develop two novel algorithms that efficiently utilize constraints to 1) reduce portions of the input space that are either verified or irrelevant to a subdomain in the context of branch-and-bound, and 2) directly improve intermediate bounds throughout the network. The process novelly uses linear constraints that are readily available during verification in a highly scalable manner compared to using off-the-shelf linear programming (LP) solvers. This reduction tightens bounds globally and can significantly reduce the number of subproblemshandled during BaB. We show our clipping procedures can intuitively and efficiently be incorporated into BaB-based verifiers such as $\alpha, \beta$-CROWN, and is amenable to BaB procedures that split upon the input or activation space. We demonstrate the effectiveness of our procedure on a broad range of benchmarks where, in some instances, we witness a 96\% reduction in the number of subproblems during branch-and-bound, and also achieve state-of-the-art verified accuracy across multiple benchmarks.



Paperid:5335
Authors:Song Ma, Dongyi Lv, He Yang, Wei Xi, Jizhong Zhao
Abstract:
Dataset condensation aims to synthesize compact yet informative datasets that retain the training efficacy of full-scale data, offering substantial gains in efficiency. Recent studies reveal that the condensation process can be vulnerable to backdoor attacks, where malicious triggers are injected into the condensation dataset, manipulating model behavior during inference. While prior approaches have made progress in balancing attack success rate and clean test accuracy, they often fall short in preserving stealthiness, especially in concealing the visual artifacts of condensed data or the perturbations introduced during inference. To address this challenge, we introduce \textsc{Sneakdoor}, which enhances stealthiness without compromising attack effectiveness. \textsc{Sneakdoor} exploits the inherent vulnerability of class decision boundaries and incorporates a generative module that constructs input-aware triggers aligned with local feature geometry, thereby minimizing detectability. This joint design enables the attack to remain imperceptible to both human inspection and statistical detection. Extensive experiments across multiple datasets demonstrate that \textsc{Sneakdoor} achieves a compelling balance among attack success rate, clean test accuracy, and stealthiness, substantially improving the invisibility of both the synthetic data and triggered samples while maintaining high attack efficacy.



Paperid:5336
Authors:Tengfei Ma, Kun Chen, Yongsheng Zang, Yujie Chen, Xuanbai Ren, Bosheng Song, Hongxin Xiang, Yiping Liu, xiangxiang Zeng
Abstract:
Abstract:Identifying drug-drug interactions (DDIs) is critical for ensuring drug safety and advancing drug development, a topic that has garnered significant research interest. While existing methods have made considerable progress, approaches relying solely on known DDIs face a key challenge when applied to drugs with limited data: insufficient exploration of the space of unlabeled pairwise drugs. To address these issues, we innovatively introduce S$^2$VM, a Self-supervised Visual pretraining framework for pair-wise Molecules, to fully fuse structural representations and explore the space of drug pairs for DDI prediction. S$^2$VM incorporates the explicit structure and correlations of visual molecules, such as the positional relationships and connectivity between functional substructures. Specifically, we blend the visual fragments of drug pairs into a unified input for joint encoding and then recover molecule-specific visual information for each drug individually. This approach integrates fine-grained structural representations from unlabeled drug pair data. By using visual fragments as anchors, S$^2$VM effectively captures the spatial information of local molecular components within visual molecules, resulting in more comprehensive embeddings of drug pairs. Experimental results show that S$^2$VM achieves state-of-the-art performance on widely used benchmarks, with Macro-F1 score improvements of 4.21% and 3.31%, respectively. Further extensive results and theoretical analysis demonstrate the effectiveness of S$^2$VM for both few-shot and novel drugs.



Paperid:5337
Authors:Tianyi Zhang, Shaochen (Henry) Zhong, Mohsen Hariri, Vipin Chaudhary, Yang Sui, Xia Hu, Anshumali Shrivastava
Abstract:
Abstract:Large-scale AI models, such as Large Language Models (LLMs) and Diffusion Models (DMs), have grown rapidly in size, creating significant challenges for efficient deployment on resource-constrained hardware. In this paper, we introduce Dynamic-Length Float (DFloat11), a lossless compression framework that reduces LLM and DM size by 30\% while preserving outputs that are bit-for-bit identical to the original model. DFloat11 is motivated by the low entropy in the BFloat16 weight representation of LLMs, which reveals significant inefficiency in the existing storage format. By applying entropy coding, DFloat11 assigns dynamic-length encodings to weights based on frequency, achieving near information-optimal compression without any loss of precision. To facilitate efficient inference with dynamic-length encodings, we develop a custom GPU kernel for fast online decompression. Our design incorporates the following: (i) compact, hierarchical lookup tables (LUTs) that fit within GPU SRAM for efficient decoding, (ii) a two-phase GPU kernel for coordinating thread read/write positions using lightweight auxiliary variables, and (iii) transformer-block-level decompression to minimize latency. Experiments on Llama 3.3, Qwen 3, Mistral 3, FLUX.1, and others validate our hypothesis that DFloat11 achieves around 30\% model size reduction while preserving bit-for-bit identical outputs. Compared to a potential alternative of offloading parts of an uncompressed model to the CPU to meet memory constraints, DFloat11 achieves 2.3--46.2$\times$ higher throughput in token generation. With a fixed GPU memory budget, DFloat11 enables 5.7--14.9$\times$ longer generation lengths than uncompressed models. Notably, our method enables lossless inference of Llama 3.1 405B, an 810GB model, on a single node equipped with 8$\times$80GB GPUs.



Paperid:5338
Authors:Simon Matrenok, Skander Moalla, Caglar Gulcehre
Abstract:
We introduce Quantile Reward Policy Optimization (QRPO), an alignment algorithm for large language models (LLM) that integrates the simplicity and offline efficiency of policy fitting methods, such as DPO and REBEL, with the ability to better leverage the signal of pointwise absolute rewards, previously limited to only policy improvement approaches such as PPO and GRPO. QRPO regresses to the closed-form solution of the KL-regularized RL objective using the quantile reward, which we show makes the partition function analytically tractable, and thus overcomes the need to resort to preferences or relative rewards to cancel this term. QRPO consistently provides the best results on chats evaluated on AlpacaEval 2 and coding evaluated on LeetCode compared to DPO, REBEL, and SimPO trained with different datasets and models on the 8B scale. We observe that training with rewards instead of preferences induces less length bias in the learned models. Finally, we show that QRPO scales with more compute budget to estimate the quantile rewards, opening a new dimension for pre-computation scaling.



Paperid:5339
Authors:Clément Yvernes, Emilie Devijver, Adèle Ribeiro, Marianne Clausel, Eric Gaussier
Abstract:
Cluster DAGs (C-DAGs) provide an abstraction of causal graphs in which nodes represent clusters of variables, and edges encode both cluster-level causal relationships and dependencies arisen from unobserved confounding. C-DAGs define an equivalence class of acyclic causal graphs that agree on cluster-level relationships, enabling causal reasoning at a higher level of abstraction. However, when the chosen clustering induces cycles in the resulting C-DAG, the partition is deemed inadmissible under conventional C-DAG semantics. In this work, we extend the C-DAG framework to support arbitrary variable clusterings by relaxing the partition admissibility constraint, thereby allowing cyclic C-DAG representations. We extend the notions of d-separation and causal calculus to this setting, significantly broadening the scope of causal reasoning across clusters and enabling the application of C-DAGs in previously intractable scenarios. Our calculus is both sound and atomically complete with respect to the do-calculus: all valid interventional queries at the cluster level can be derived using our rules, each corresponding to a primitive do-calculus step.



Paperid:5340
Authors:Siddharth Viswanath, Hiren Madhu, Dhananjay Bhaskar, Jake Kovalic, David R Johnson, Christopher Tape, Ian Adelstein, Rex Ying, Michael Perlmutter, Smita Krishnaswamy
Abstract:
In this paper, we propose HiPoNet, an end-to-end differentiable neural network for regression, classification, and representation learning on high-dimensional point clouds. Our work is motivated by single-cell data which can have very high-dimensionality --exceeding the capabilities of existing methods for point clouds which are mostly tailored for 3D data. Moreover, modern single-cell and spatial experiments now yield entire cohorts of datasets (i.e., one data set for every patient), necessitating models that can process large, high-dimensional point-clouds at scale. Most current approaches build a single nearest-neighbor graph, discarding important geometric and topological information. In contrast, HiPoNet models the point-cloud as a set of higher-order simplicial complexes, with each particular complex being created using a reweighting of features. This method thus generates multiple constructs corresponding to different views of high-dimensional data, which in biology offers the possibility of disentangling distinct cellular processes. It then employs simplicial wavelet transforms to extract multiscale features, capturing both local and global topology from each view. We show that geometric and topological information is preserved in this framework both theoretically and empirically. We showcase the utility of HiPoNet on point-cloud level tasks, involving classification and regression of entire point-clouds in data cohorts. Experimentally, we find that HiPoNet outperforms other point-cloud and graph-based models on single-cell data. We also apply HiPoNet to spatial transcriptomics datasets using spatial coordinates as one of the views. Overall, HiPoNet offers a robust and scalable solution for high-dimensional data analysis.



Paperid:5341
Authors:Qing-Yuan Jiang, Longfei Huang, Yang Yang
Abstract:
Multimodal learning (MML) is significantly constrained by modality imbalance, leading to suboptimal performance in practice. While existing approaches primarily focus on balancing the learning of different modalities to address this issue, they fundamentally overlook the inherent disproportion in model classification ability, which serves as the primary cause of this phenomenon. In this paper, we propose a novel multimodal learning approach to dynamically balance the classification ability of weak and strong modalities by designing a sustained boosting algorithm. Concretely, we first propose a sustained boosting algorithm in multimodal learning by simultaneously optimizing the classification and residual errors. Subsequently, we introduce an adaptive classifier assignment strategy to dynamically facilitate the classification performance of the weak modality. Furthermore, we theoretically analyze the convergence property of the cross-modal gap function, ensuring the effectiveness of the proposed boosting scheme. To this end, the classification ability of strong and weak modalities is expected to be balanced, thereby mitigating the imbalance issue. Empirical experiments on widely used datasets reveal the superiority of our method through comparison with various state-of-the-art~(SOTA) multimodal learning baselines. The source code is available at https://anonymous.4open.science/r/Our_NeurIPS25-A4C7.



Paperid:5342
Authors:Xiaoli Tang, Han Yu, Xiaoxiao Li
Abstract:
Abstract:Auction-based Federated Learning (AFL) fosters collaboration among self-interested data consumers (DCs) and data owners (DOs). A major challenge in AFL pertains to how DCs select and bid for DOs. Existing methods are generally static, making them ill-suited for dynamic AFL markets. To address this issue, we propose the R}einforcement Learning-based Bidding Strategy for DCs in Auction-based Federated Learning (RLB-AFL). We incorporate historical states into a Deep Q-Network to capture sequential information critical for bidding decisions. To mitigate state space sparsity, where specific states rarely reoccur for each DC during auctions, we incorporate the Gaussian Mixture Model into RLB-AFL. This facilitates soft clustering on sequential states, reducing the state space dimensionality and easing exploration and action-value function approximation. In addition, we enhance the $\epsilon$-greedy policy to help the RLB-AFL agent balance exploitation and exploration, enabling it to be more adaptable in the AFL decision-making process. Extensive experiments under 6 widely used benchmark datasets demonstrate that RLB-AFL achieves superior performance compared to 8 state-of-the-art approaches. It outperforms the best baseline by 10.56% and 3.15% in terms of average total utility



Paperid:5343
Authors:Zichen Liu, Wei Zhang, Tiejun Li
Abstract:
Euclidean diffusion models have achieved remarkable success in generative modeling across diverse domains, and they have been extended to manifold case in recent advances. Instead of explicitly utilizing the structure of special manifolds as studied in previous works, we investigate direct sampling of the Euclidean diffusion models for general manifold-constrained data in this paper. We reveal the multiscale singularity of the score function in the embedded space of manifold, which hinders the accuracy of diffusion-generated samples. We then present an elaborate theoretical analysis of the singularity structure of the score function by separating it along the tangential and normal directions of the manifold. To mitigate the singularity and improve the sampling accuracy, we propose two novel methods: (1) Niso-DM, which introduces non-isotropic noise along the normal direction to reduce scale discrepancies, and (2) Tango-DM, which trains only the tangential component of the score function using a tangential-only loss function. Numerical experiments demonstrate that our methods achieve superior performance on distributions over various manifolds with complex geometries.



Paperid:5344
Authors:Changyao Tian, Hao Li, Gen Luo, Xizhou Zhu, Weijie Su, Hanming Deng, Jinguo Zhu, Jie Shao, Ziran Zhu, Yunpeng Liu, Lewei Lu, Wenhai Wang, Hongsheng Li, Jifeng Dai
Abstract:
Compositional training has been the de-facto paradigm in existing Multimodal Large Language Models (MLLMs), where pre-trained vision encoders are connected with pre-trained LLMs through continuous multimodal pre-training. However, the multimodal scaling property of this paradigm remains difficult to explore due to the separated training. In this paper, we focus on the native training of MLLMs in an end-to-end manner and systematically study its design space and scaling property under a practical setting, i.e., data constraint. Through careful study of various choices in MLLM, we obtain the optimal meta-architecture that best balances performance and training cost. After that, we further explore the scaling properties of the native MLLM and indicate the positively correlated scaling relationship between visual encoders and LLMs. Based on these findings, we propose a native MLLM called NaViL, combined with a simple and cost-effective recipe. Experimental results on 14 multimodal benchmarks confirm the competitive performance of NaViL against existing MLLMs. Besides, our findings and results provide in-depth insights for the future study of native MLLMs.



Paperid:5345
Authors:Van-Anh Nguyen, Trung Le, Mehrtash Harandi, Ehsan Abbasnejad, Thanh-Toan Do, Dinh Phung
Abstract:
We propose a framework grounded in gradient flow theory and informed by geometric structure that provides multiple diverse solutions for a given task, ensuring collaborative results that enhance performance and adaptability across different tasks. This framework enables flexibility, allowing for efficient task-specific fine-tuning while preserving the knowledge of the pre-trained foundation models. Extensive experiments across transfer learning, few-shot learning, and domain generalization show that our proposed approach consistently outperforms existing Bayesian methods, delivering strong performance with affordable computational overhead and offering a practical solution by updating only a small subset of parameters.



Paperid:5346
Authors:Huanjin Yao, Jiaxing Huang, Wenhao Wu, Jingyi Zhang, Yibo Wang, Shunyu Liu, Yingjie Wang, YuXin Song, Haocheng Feng, Li Shen, Dacheng Tao
Abstract:
In this work, we aim to develop an MLLM that understands and solves questions by learning to create each intermediate step of the reasoning involved till the final answer. To this end, we propose Collective Monte Carlo Tree Search (CoMCTS), a new learning-to-reason method for MLLMs, which introduces the concept of collective learning into ``tree search'' for effective and efficient reasoning-path searching and learning. The core idea of CoMCTS is to leverage collective knowledge from multiple models to collaboratively conjecture, search and identify effective reasoning paths toward correct answers via four iterative operations including Expansion, Simulation and Error Positioning, Backpropagation, and Selection. Using CoMCTS, we construct Mulberry-260k, a multimodal dataset with a tree of rich, explicit and well-defined reasoning nodes for each question. With Mulberry-260k, we perform collective SFT to train our model, Mulberry, a series of MLLMs with o1-like step-by-step Reasoning and Reflection capabilities. Extensive experiments demonstrate the superiority of our proposed methods on various benchmarks. Anonymous code is available at https://anonymous.4open.science/r/Mulberry-NIPS25.



Paperid:5347
Authors:Ming Gu, Zhuonan Zheng, Sheng Zhou, Meihan Liu, Jiawei Chen, Qiaoyu Tan, Liangcheng Li, Jiajun Bu
Abstract:
Graph Neural Networks (GNNs) have achieved great success but are often considered to be challenged by varying levels of homophily in graphs. Recent empirical studies have surprisingly shown that homophilic GNNs can perform well across datasets of different homophily levels with proper hyperparameter tuning, but the underlying theory and effective architectures remain unclear. To advance GNN universality across varying homophily, we theoretically revisit GNN message passing and uncover a novel \textit{smoothness-generalization dilemma}, where increasing hops inevitably enhances smoothness at the cost of generalization. This dilemma hinders learning in higher-order homophilic neighborhoods and all heterophilic ones, where generalization is critical due to complex neighborhood class distributions that are sensitive to shifts induced by noise and sparsity. To address this, we introduce the Inceptive Graph Neural Network (IGNN) built on three simple yet effective design principles, which alleviate the dilemma by enabling distinct hop-wise generalization alongside improved overall generalization with adaptive smoothness. Benchmarking against 25 baselines demonstrates IGNN's superiority and reveals notable universality in certain homophilic GNN variants. Our code is available at \href{https://anonymous.4open.science/r/IGNN}{https://anonymous.4open.science/r/IGNN}.



Paperid:5348
Authors:Bicheng Ying, Zhe Li, Haibo Yang
Abstract:
This work tackles the fundamental challenges in Federated Learning (FL) posed by arbitrary client participation and data heterogeneity, prevalent characteristics in practical FL settings. It is well-established that popular FedAvg-style algorithms struggle with exact convergence and can suffer from slow convergence rates since a decaying learning rate is required to mitigate these scenarios. To address these issues, we introduce the concept of stochastic matrix and the corresponding time-varying graphs as a novel modeling tool to accurately capture the dynamics of arbitrary client participation and the local update procedure. Leveraging this approach, we offer a fresh perspective on designing FL algorithms, provide a rigorous quantitative analysis of the limitations inherent in the FedAvg algorithm, and present FOCUS, Federated Optimization with Exact Convergence via Push-pull Strategy, a provably convergent algorithm designed to effectively overcome the previously mentioned two challenges. More specifically, we provide a rigorous proof demonstrating that FOCUS achieves exact convergence with a linear rate regardless of the arbitrary client participation, establishing it as the first work to demonstrate this significant result.



Paperid:5349
Authors:Haoyu Li, Xiangru Zhong, Bin Hu, Huan Zhang
Abstract:
Abstract:Learning-based neural network (NN) control policies have shown impressive empirical performance. However, obtaining stability guarantee and region of attraction estimation of these learned neural controllers is challenging due to the lack of stable and scalable training and verification algorithms. Although previous works in this area have achieved great success, much conservatism remains in their framework. In this work, we propose a novel two-stage training framework to jointly synthesize the controller and Lyapunov function for continuous-time systems. By leveraging a Zubov‑inspired region of attraction characterization to directly estimate stability boundaries, we propose a novel training data sampling strategy and a domain updating mechanism that significantly reduces the conservatism in training. Moreover, unlike existing work on continuous-time systems that relies on an SMT solver to formally verify the Lyapunov condition, we extend state-of-the-art neural network verifier $\alpha,\!\beta$-CROWN with the capability of performing automatic bound propagation through Jacobian over dynamical systems and a novel verification scheme that avoids expensive bisection. To demonstrate the effectiveness of our approach, we conduct numerical experiments by synthesizing and verifying controllers on several challenging nonlinear systems across multiple dimensions. We show that our training can yield a region of attraction with volume $5 - 1.5\cdot 10^{5}$ times larger compared to the baseline, and our verification on continuous systems can be up to $40-10000$ times faster compared to the traditional SMT solver dReal.



Paperid:5350
Authors:Rahul Mahesh, Jun-Jee Chao, Volkan Isler
Abstract:
Abstract:The ability to capture rich representations of combinatorial structures has enabled the application of machine learning to tasks such as analysis and generation of floorplans, terrains, images, and animations. Recent work has primarily focused on understanding structures with well-defined features, neighborhoods, or underlying distance metrics, while those lacking such characteristics remain largely unstudied. Examples of these combinatorial structures can be found in polygons, where a small change in the vertex locations causes a significant rearrangement of the combinatorial structure, expressed as a visibility or triangulation graphs. Current representation learning approaches fail to capture structures without well-defined features and distance metrics.In this paper, we study the open problem of Visibility Reconstruction: Given a visibility graph $G$, construct a polygon $P$ whose visibility graph is $G$. We introduce $\textbf{VisDiff}$, a novel diffusion-based approach to generate polygon $P$ from the input visibility graph $G$. The main novelty of our approach is that, rather than generating the polygon's vertex set directly, we first estimate the signed distance function (SDF) associated with the polygon. The SDF is then used to extract the vertex location representing the final polygon. We show that going through the SDF allows $\textbf{VisDiff}$ to learn the visibility relationship much more effectively than generating vertex locations directly. In order to train $\textbf{VisDiff}$, we create a carefully curated dataset. We use this dataset to benchmark our method and achieve 26\% improvement in F1-Score over standard methods as well as state of the art approaches. We also provide preliminary results on the harder visibility graph recognition problem in which the input $G$ is not guaranteed to be a visibility graph. To demonstrate the applicability of VisDiff beyond visibility graphs, we extend it to the related combinatorial structure of triangulation graph. Lastly, leveraging these capabilties, we show that VisDiff can perform high-diversity sampling over the space of all polygons. In particular, we highlight its ability to perform both polygon-to-polygon interpolation and graph-to-graph interpolation, enabling diverse sampling across the polygon space.



Paperid:5351
Authors:Chen YAN, Weina Wang, Lei Ying
Abstract:
Abstract:We study the finite-horizon Restless Multi-Armed Bandit (RMAB) problem with $N$ homogeneous arms. Prior work has shown that when an RMAB satisfies a non-degeneracy condition, Linear-Programming-based (LP-based) policies derived from the fluid approximation, which captures the mean dynamics of the system, achieve an exponentially small optimality gap. However, it is common for RMABs to be degenerate, in which case LP-based policies can result in a $\Theta(1/\sqrt{N})$ optimality gap per arm. In this paper, we propose a novel Stochastic-Programming-based (SP-based) policy that, under a uniqueness assumption, achieves an $\tilde{\mathcal{O}}(1/N)$ optimality gap for degenerate RMABs. Our approach is based on the construction of a Gaussian stochastic system that captures not only the mean but also the variance of the RMAB dynamics, resulting in a more accurate approximation than the fluid approximation. We then solve a stochastic program for this system to obtain our policy. This is the first result to establish an $\tilde{\mathcal{O}}(1/N)$ optimality gap for degenerate RMABs.



Paperid:5352
Authors:Yue Xu, Chengyan Fu, Li Xiong, Sibei Yang, Wenjie Wang
Abstract:
Abstract:Pre-training large language models (LLMs) on vast text corpora enhances natural language processing capabilities but risks encoding social biases, particularly gender bias. While parameter-modification methods like fine-tuning mitigate bias, they are resource-intensive, unsuitable for closed-source models, and lack adaptability to evolving societal norms. Instruction-based approaches offer flexibility but often compromise general performance on normal tasks. To address these limitations, we propose $\textit{FaIRMaker}$, an automated and model-independent framework that employs an $\textbf{auto-search and refinement}$ paradigm to adaptively generate Fairwords, which act as instructions to reduce gender bias and enhance response quality. $\textit{FaIRMaker}$ enhances the debiasing capacity by enlarging the Fairwords search space while preserving the utility and making it applicable to closed-source models by training a sequence-to-sequence model that adaptively refines Fairwords into effective debiasing instructions when facing gender-related queries and performance-boosting prompts for neutral inputs. Extensive experiments demonstrate that $\textit{FaIRMaker}$ effectively mitigates gender bias while preserving task integrity and ensuring compatibility with both open- and closed-source LLMs.



Paperid:5353
Authors:Dechen Zhang, Zhenmei Shi, Zhang, Yingyu Liang, Difan Zou
Abstract:
Kernel ridge regression (KRR) is a foundational tool in machine learning, with recent work emphasizing its connections to neural networks. However, existing theory primarily addresses the i.i.d. setting, while real-world data often exhibits structured dependencies - particularly in applications like denoising score learning where multiple noisy observations derive from shared underlying signals. We present the first systematic study of KRR generalization for non-i.i.d. data with signal-noise causal structure, where observations represent different noisy views of common signals. Under standard spectral decay assumptions, we develop a novel blockwise decomposition method that enables precise concentration analysis for dependent data. Using this technique, we derive excess risk bounds for KRR that explicitly depend on: (1) the kernel spectrum, (2) causal structure parameters, and (3) sampling mechanisms (including relative sample sizes for signals and noises). We further apply our results to denoising score learning, establishing generalization guarantees and providing principled guidance for sampling noisy data points. This work advances KRR theory while providing practical tools for analyzing dependent data in modern machine learning applications.



Paperid:5354
Authors:Chong Wu, Jiawang Cao, Renjie Xu, Zhuoheng Ran, Maolin Che, Wenbo Zhu, Hong Yan
Abstract:
This paper proposes the Dual-Stage Sparse Attention (DuSA) mechanism for attention acceleration of vision transformers. In the first stage, DuSA performs intrablock sparse attention to aggregate local inductive biases. In the second stage, DuSA performs interblock sparse attention to obtain long-range dependencies. Both stages have low computational complexity and can be further accelerated by memory acceleration attention mechanisms directly, which makes DuSA faster than some extremely fast attention mechanisms. The dual-stage sparse attention design provides a lower error in approximating vanilla scaled-dot product attention than the basic single-stage sparse attention mechanisms and further advances the basic sparse attention mechanisms to match or even outperform vanilla scaled-dot product attention. Even in some plug and play situations, DuSA can still maintain low performance loss. DuSA can be used in both training and inference acceleration. DuSA achieves leading performance in different benchmarks: long range arena, image classification, semantic segmentation, object detection, and text to video generation, and accelerates models of different sizes.



Paperid:5355
Authors:Stephen Zhang, Michael Stumpf
Abstract:
Abstract:We consider the Schrödinger bridge problem which, given ensemble measurements of the initial and final configurations of a stochastic dynamical system and some prior knowledge on the dynamics, aims to reconstruct the ``most likely'' evolution of the system compatible with the data. Most existing literature assume Brownian reference dynamics and are implicitly limited to potential-driven dynamics. We depart from this regime and consider reference processes described by a multivariate Ornstein-Uhlenbeck process with generic drift matrix $\mathbf{A} \in \mathbb{R}^{d \times d}$. When $\mathbf{A}$ is asymmetric, this corresponds to a non-equilibrium system with non-conservative forces at play: this is important for applications to biological systems, which are naturally exist out-of-equilibrium. In the case of Gaussian marginals, we derive explicit expressions that characterise the solution of both the static and dynamic Schrödinger bridge. For general marginals, we propose mvOU-OTFM, a simulation-free algorithm based on flow and score matching for learning the Schrödinger bridge. In application to a range of problems based on synthetic and real single cell data, we demonstrate that mvOU-OTFM achieves higher accuracy compared to competing methods, whilst being significantly faster to train.



Paperid:5356
Authors:Zhibin Gu, weili wang
Abstract:
Deep multi-view clustering (DMVC) has emerged as a promising paradigm for integrating information from multiple views by leveraging the representation power of deep neural networks. However, most existing DMVC methods primarily focus on modeling pairwise relationships between samples, while neglecting higher-order structural dependencies among multiple samples, which may hinder further improvements in clustering performance. To address this limitation, we propose a hypergraph neural network (HGNN)-driven multi-view clustering framework, termed Hypergraph-enhanced cOntrastive learning with hyPEr-Laplacian regulaRization (HOPER), a novel model that jointly captures high-order correlations and preserves local manifold structures across views. Specifically, we first construct view-specific hypergraph structures and employ the HGNN to learn node representations, thereby capturing high-order relationships among samples. Furthermore, we design a hypergraph-driven dual contrastive learning mechanism that integrates inter-view contrastive learning with intra-hyperedge contrastive learning, promoting cross-view consistency while maintaining discriminability within hyperedges. Finally, a hyper-Laplacian manifold regularization is introduced to preserve the local geometric structure within each view, thereby enhancing the structural fidelity and discriminative power of the learned representations. Extensive experiments on diverse datasets demonstrate the effectiveness of our approach.



Paperid:5357
Authors:Donghyun Son, Euntae Choi, Sungjoo Yoo
Abstract:
Abstract:Large Language Model (LLM) inference is typically memory-intensive, especially when processing large batch sizes and long sequences, due to the large size of key-value (KV) cache.Vector Quantization (VQ) is recently adopted to alleviate this issue, but we find that the existing approach is susceptible to distribution shift due to its reliance on calibration datasets.To address this limitation, we introduce $\textbf{NSNQuant}$, a calibration-free Vector Quantization (VQ) technique designed for low-bit compression of the KV cache. By applying a three-step transformation—$\textbf{1)}$ a token-wise normalization ($\textbf{N}$ormalize), $\textbf{2)}$ a channel-wise centering ($\textbf{S}$hift), and $\textbf{3)}$ a second token-wise normalization ($\textbf{N}$ormalize)—with Hadamard transform, NSNQuant effectively aligns the token distribution with the standard normal distribution. This alignment enables robust, calibration-free vector quantization using a single reusable codebook.Extensive experiments show that NSNQuant consistently outperforms prior methods in both 1-bit and 2-bit settings, offering strong generalization and up to 3$\times$ throughput gains over full-precision baselines.



Paperid:5358
Authors:Jiaye Fu, Qiankun Gao, Chengxiang Wen, Yanmin Wu, Siwei Ma, Jiaqi Zhang, Jian Zhang
Abstract:
Online free-viewpoint video (FVV) reconstruction is challenged by slow per-frame optimization, inconsistent motion estimation, and unsustainable storage demands. To address these challenges, we propose the Reconfigurable Continuum Gaussian Stream, dubbed ReCon-GS, a novel storage-aware framework that enables high-fidelity online dynamic scene reconstruction and real-time rendering. Specifically, we dynamically allocate multi-level Anchor Gaussians in a density-adaptive fashion to capture inter-frame geometric deformations, thereby decomposing scene motion into compact coarse-to-fine representations. Then, we design a dynamic hierarchy reconfiguration strategy that preserves localized motion expressiveness through on-demand anchor re-hierarchization, while ensuring temporal consistency through intra-hierarchical deformation inheritance that confines transformation priors to their respective hierarchy levels. Furthermore, we introduce a storage-aware optimization mechanism that flexibly adjusts the density of Anchor Gaussians at different hierarchy levels, enabling a controllable trade-off between reconstruction fidelity and memory usage. Extensive experiments on three widely used datasets demonstrate that, compared to state‐of‐the‐art methods, ReCon-GS improves training efficiency by approximately 15% and achieves superior FVV synthesis quality with enhanced robustness and stability. Moreover, at equivalent rendering quality, ReCon-GS slashes memory requirements by over 50% compared to leading state‑of‑the‑art methods.



Paperid:5359
Authors:粤鹏 郑, Fu Luo, Zhenkun Wang, Yaoxin Wu, Yu Zhou
Abstract:
Multi-Task Learning (MTL) in Neural Combinatorial Optimization (NCO) is a promising approach to train a unified model capable of solving multiple Vehicle Routing Problem (VRP) variants. However, existing Reinforcement Learning (RL)-based multi-task methods can only train light decoder models on small-scale problems, exhibiting limited generalization ability when solving large-scale problems. To overcome this limitation, this work introduces a novel multi-task learning method driven by knowledge distillation (MTL-KD), which enables the efficient training of heavy decoder models with strong generalization ability. The proposed MTL-KD method transfers policy knowledge from multiple distinct RL-based single-task models to a single heavy decoder model, facilitating label-free training and effectively improving the model's generalization ability across diverse tasks. In addition, we introduce a flexible inference strategy termed Random Reordering Re-Construction (R3C), which is specifically adapted for diverse VRP tasks and further boosts the performance of the multi-task model. Experimental results on 6 seen and 10 unseen VRP variants with up to 1000 nodes indicate that our proposed method consistently achieves superior performance on both uniform and real-world benchmarks, demonstrating robust generalization abilities



Paperid:5360
Authors:Toshinori Kitamura, Arnob Ghosh, Tadashi Kozuno, Wataru Kumagai, Kazumi Kasaura, Kenta Hoshino, Yohei Hosoe, Yutaka Matsuo
Abstract:
Abstract:We study the reinforcement learning (RL) problem in a constrained Markov decision process (CMDP), where an agent explores the environment to maximize the expected cumulative reward while satisfying a single constraint on the expected total utility value in every episode. While this problem is well understood in the tabular setting, theoretical results for function approximation remain scarce. This paper closes the gap by proposing an RL algorithm for linear CMDPs that achieves $\widetilde{\mathcal{O}}(\sqrt{K})$ regret with an {\em episode-wise} zero-violation guarantee. Furthermore, our method is computationally efficient, scaling polynomially with problem-dependent parameters while remaining independent of the state space size. Our results significantly improve upon recent linear CMDP algorithms, which either violate the constraint or incur exponential computational costs.



Paperid:5361
Authors:Bogdan Kulynych, Juan Gomez, Jamie Hayes, Borja Balle, Flavio Calmon, Georgios Kaissis, Jean Raisaro
Abstract:
Abstract:Differentially private (DP) mechanisms are difficult to interpret and calibrate because existing methods for mapping standard privacy parameters to concrete privacy risks—re-identification, attribute inference, and data reconstruction—are both overly pessimistic and inconsistent. In this work, we use the hypothesis-testing interpretation of DP ($f$-DP), and determine that bounds on attack success can take the same unified form across re-identification, attribute inference, and data reconstruction risks. Our unified bounds are (1) consistent across a multitude of attack settings, and (2) tunable, enabling practitioners to evaluate risk with respect to arbitrary (including worst-case) levels of baseline risk. Empirically, our results are tighter than prior methods using $\varepsilon$-DP, Renyi DP, and concentrated DP—as a result, calibrating noise using our bounds can reduce the required noise by 20\% at the same risk level, which yields, e.g., up to $18$pp accuracy increase in a text classification task. Overall, this unifying perspective provides a principled framework for interpreting and calibrating the degree of protection in DP against specific levels of re-identification, attribute inference, or data reconstruction risk.



Paperid:5362
Authors:Xueyang Zhou, Guiyao Tie, Guowen Zhang, Hecheng Wang, Pan Zhou, Lichao Sun
Abstract:
Vision-Language-Action (VLA) models have advanced robotic control by enabling end-to-end decision-making directly from multimodal inputs. However, their tightly coupled architectures expose novel security vulnerabilities. Unlike traditional adversarial perturbations, backdoor attacks represent a stealthier, persistent, and practically significant threat—particularly under the emerging Training-as-a-Service paradigm—but remain largely unexplored in the context of VLA models. To address this gap, we proposeBadVLA, a backdoor attack method based on Objective-Decoupled Optimization, which for the first time exposes the backdoor vulnerabilities of VLA models. Specifically, it consists of a two-stage process: (1) explicit feature-space separation to isolate trigger representations from benign inputs, and (2) conditional control deviations that activate only in the presence of the trigger, while preserving clean-task performance. Empirical results on multiple VLA benchmarks demonstrate that BadVLA consistently achieves near-100\% attack success rates with minimal impact on clean task accuracy. Further analyses confirm its robustness against common input perturbations, task transfers, and model fine-tuning, underscoring critical security vulnerabilities in current VLA deployments. Our work offers the first systematic investigation of backdoor vulnerabilities in VLA models, highlighting an urgent need for secure and trustworthy embodied model design practices.



Paperid:5363
Authors:Sepehr Mousavi, Shizheng Wen, Levi Lingsch, Maximilian Herde, Bogdan Raonic, Siddhartha Mishra
Abstract:
Learning the solution operators of PDEs on arbitrary domains is challenging due to the diversity of possible domain shapes, in addition to the often intricate underlying physics. We propose an end-to-end graph neural network (GNN) based neural operator to learn PDE solution operators from data on point clouds in arbitrary domains. Our multi-scale model maps data between input/output point clouds by passing it through a downsampled regional mesh. The approach includes novel elements aimed at ensuring spatio-temporal resolution invariance. Our model, termed RIGNO, is tested on a challenging suite of benchmarks composed of various time-dependent and steady PDEs defined on a diverse set of domains. We demonstrate that RIGNO is significantly more accurate than neural operator baselines and robustly generalizes to unseen resolutions both in space and in time.



Paperid:5364
Authors:Zheng Li, Feng Xie, Xichen Guo, Yan Zeng, Hao Zhang, Zhi Geng
Abstract:
Estimating causal effects from nonexperimental data is a fundamental problem in many fields of science. A key component of this task is selecting an appropriate set of covariates for confounding adjustment to avoid bias. Most existing methods for covariate selection often assume the absence of latent variables and rely on learning the global causal structure among variables. However, identifying the global structure can be unnecessary and inefficient, especially when our primary interest lies in estimating the effect of a treatment variable on an outcome variable. To address this limitation, we propose a novel local learning approach for covariate selection in nonparametric causal effect estimation, which accounts for the presence of latent variables. Our approach leverages testable independence and dependence relationships among observed variables to identify a valid adjustment set for a target causal relationship, ensuring both soundness and completeness under standard assumptions. We validate the effectiveness of our algorithm through extensive experiments on both synthetic and real-world data.



Paperid:5365
Authors:Yuan Wu, Zhiqiang Yan, Yigong Zhang, Xiang Li, Jian Yang
Abstract:
Occupancy prediction aims to estimate the 3D spatial distribution of occupied regions along with their corresponding semantic labels. Existing vision-based methods perform well on daytime benchmarks but struggle in nighttime scenarios due to limited visibility and challenging lighting conditions. To address these challenges, we propose LIAR, a novel framework that learns illumination-affined representations. LIAR first introduces Selective Low-light Image Enhancement (SLLIE), which leverages the illumination priors from daytime scenes to adaptively determine whether a nighttime image is genuinely dark or sufficiently well-lit, enabling more targeted global enhancement. Building on the illumination maps generated by SLLIE, LIAR further incorporates two illumination-aware components: 2D Illumination-guided Sampling (2D-IGS) and 3D Illumination-driven Projection (3D-IDP), to respectively tackle local underexposure and overexposure. Specifically, 2D-IGS modulates feature sampling positions according to illumination maps, assigning larger offsets to darker regions and smaller ones to brighter regions, thereby alleviating feature degradation in underexposed areas. Subsequently, 3D-IDP enhances semantic understanding in overexposed regions by constructing illumination intensity fields and supplying refined residual queries to the BEV context refinement process. Extensive experiments on both real and synthetic datasets demonstrate the superior performance of LIAR under challenging nighttime scenarios. The source code and pretrained models are available.



Paperid:5366
Authors:Kumar Kshitij Patel, Ali Zindari, Sebastian Stich, Lingxiao Wang
Abstract:
Local SGD, or Federated Averaging, is one of the most widely used algorithms for distributed optimization. Although it often outperforms alternatives such as mini-batch SGD, existing theory has not fully explained this advantage under realistic data heterogeneity. Recent work has suggested that a second-order heterogeneity condition may suffice to justify the empirical gains of local SGD and to characterize its communication efficiency. We confirm this conjecture by establishing new upper and lower bounds on the convergence of local SGD. These bounds demonstrate how low second-order heterogeneity, combined with third-order smoothness, enables local SGD to interpolate between heterogeneous and homogeneous regimes while maintaining communication efficiency. Our main technical contribution is a refined analysis of the consensus error, a central quantity in such results. We validate our theory with experiments on a distributed linear regression task.



Paperid:5367
Authors:Feng Lu, Tong Jin, Canming Ye, Xiangyuan Lan, Yunpeng Liu, Chun Yuan
Abstract:
Visual place recognition (VPR) is typically regarded as a specific image retrieval task, whose core lies in representing images as global descriptors. Over the past decade, dominant VPR methods (e.g., NetVLAD) have followed a paradigm that first extracts the patch features/tokens of the input image using a backbone, and then aggregates these patch features into a global descriptor via an aggregator. This backbone-plus-aggregator paradigm has achieved overwhelming dominance in the CNN era and remains extensively used in transformer-based models. In this paper, however, we argue that a dedicated aggregator is not necessary in the transformer era, that is, we can obtain robust global descriptors only with the backbone. Specifically, we introduce some learnable aggregation tokens, which are prepended to the patch tokens before a particular transformer block. All these tokens will be jointly processed and interact globally via the intrinsic self-attention mechanism, implicitly aggregating useful information within the patch tokens to the aggregation tokens. Finally, we only take these aggregation tokens from the last output tokens and concatenate them as the global representation. Although implicit aggregation can provide robust global descriptors in an extremely simple manner, where and how to insert additional tokens, as well as the initialization of tokens, remains an open issue worthy of further exploration. To this end, we also propose the optimal token insertion strategy and token initialization method derived from empirical studies. Experimental results show that our method outperforms state-of-the-art methods on several VPR datasets with higher efficiency and ranks 1st on the MSLS challenge leaderboard. The code will be publicly available.



Paperid:5368
Authors:Zixuan Liu, Siavash H. Khajavi, Guangkai Jiang
Abstract:
Abstract:Recent advances in multi-modal models have demonstrated strong performance in tasks such as image generation and reasoning. However, applying these models to the fire domain remains challenging due to the lack of publicly available datasets with high-quality fire domain annotations. To address this gap, we introduce $\textbf{DetectiumFire}$, a large-scale, multi-modal dataset comprising of 21.7k high-resolution fire-related images and 2.5k real-world fire-related videos covering a wide range of fire types, environments, and risk levels. Each sample is annotated with both traditional computer vision labels (e.g., bounding boxes) and detailed textual prompts describing the scene, enabling applications such as synthetic data generation and fire risk reasoning. DetectiumFire offers clear advantages over existing benchmarks in scale, diversity, and data quality, significantly reducing redundancy and enhancing coverage of real-world scenarios. We validate the utility of DetectiumFire across multiple tasks, including object detection, diffusion-based image generation, and vision-language reasoning. Our results highlight the potential of this dataset to advance fire-related research and support the development of intelligent safety systems. We release DetectiumFire to promote broader exploration of fire understanding in the AI community.



Paperid:5369
Authors:Jinguo Luo, Weihong Ren, Quanlong Zheng, Yanhao Zhang, Zhenlong Yuan, Zhiyong Wang, Haonan Lu, Honghai LIU
Abstract:
Recently, Large Foundation Models (LFMs), e.g., CLIP and GPT, have significantly advanced the Human-Object Interaction (HOI) detection, due to their superior generalization and transferability. Prior HOI detectors typically employ single- or multi-modal prompts to generate discriminative representations for HOIs from pretrained LFMs. However, such prompt-based approaches focus on transferring HOI-specific knowledge, but unexplore the potential reasoning capabilities of LFMs, which can provide informative context for ambiguous and open-world interaction recognition. In this paper, we propose InstructHOI, a novel method that leverages context-aware instructions to guide multi-modal reasoning for HOI detection. Specifically, to bridge knowledge gap and enhance reasoning abilities, we first perform HOI-domain fine-tuning on a pretrained multi-modal LFM, using a generated dataset with 140K interaction-reasoning image-text pairs. Then, we develop a Context-aware Instruction Generator (CIG) to guide interaction reasoning. Unlike traditional language-only instructions, CIG first mines visual interactive context at the human-object level, which is then fused with linguistic instructions, forming multi-modal reasoning guidance. Furthermore, an Interest Token Selector (ITS) is adopted to adaptively filter image tokens based on context-aware instructions, thereby aligning reasoning process with interaction regions. Extensive experiments on two public benchmarks demonstrate that our proposed method outperforms the state-of-the-art ones, under both supervised and zero-shot settings.



Paperid:5370
Authors:Geng Chen, Pengfei Ren, Xufeng Jian, Haifeng Sun, Menghao Zhang, Qi Qi, Zirui Zhuang, Jing Wang, Jianxin Liao, Jingyu Wang
Abstract:
Multi-view 3D human pose estimation (HPE) leverages complementary information across views to improve accuracy and robustness. Traditional methods rely on camera calibration to establish geometric correspondences, which is sensitive to calibration accuracy and lacks flexibility in dynamic settings. Calibration-free approaches address these limitations by learning adaptive view interactions, typically leveraging expressive and flexible continuous representations. However, as the multiview interaction relationship is learned entirely from data without constraint, they are vulnerable to noisy input, which can propagate, amplify and accumulate errors across all views, severely corrupting the final estimated pose.To mitigate this, we propose a novel framework that integrates a noise-resilient discrete prior into the continuous representation-based model. Specifically, we introduce the \textit{UniCodebook}, a unified, compact, robust, and discrete representation complementary to continuous features, allowing the model to benefit from robustness to noise while preserving regression capability.Furthermore, we further propose an attribute-preserving and complementarity-enhancing Discrete-Continuous Spatial Attention (DCSA) mechanism to facilitate interaction between discrete priors and continuous pose features.Extensive experiments on three representative datasets demonstrate that our approach outperforms both calibration-required and calibration-free methods, achieving state-of-the-art performance.



Paperid:5371
Authors:Xixian Yong, Xiao Zhou, Yingying Zhang, Jinlin Li, Yefeng Zheng, Xian Wu
Abstract:
The recent rise of Large Reasoning Models (LRMs) has significantly improved multi-step reasoning performance, but often at the cost of generating excessively long reasoning chains. This paper revisits the efficiency of such reasoning processes through an information-theoretic lens, revealing a fundamental trade-off between reasoning length and semantic efficiency. We propose two metrics—InfoBias and InfoGain—to quantify divergence from ideal reasoning paths and stepwise information contribution, respectively. Empirical analyses show that longer reasoning chains tend to exhibit higher information bias and diminishing information gain, especially for incorrect answers. Motivated by these findings, we introduce an entropy-based Adaptive Think strategy that dynamically halts reasoning once confidence is sufficiently high, improving efficiency while maintaining competitive accuracy. Compared to the Vanilla Think approach (default mode), our strategy yields a 1.10% improvement in average accuracy and a 50.80% reduction in token usage on QwQ-32B across six benchmark tasks spanning diverse reasoning types and difficulty levels, demonstrating superior efficiency and reasoning performance. These results underscore the promise of entropy-based methods for enhancing both accuracy and cost-effiiciency in large language model deployment.



Paperid:5372
Authors:Ethan Hsu, Tony Cao, Lesia Semenova, Chudi Zhong
Abstract:
The Rashomon Effect in machine learning refers to the existence of multiple models that achieve approximately the same predictive performance as the optimal model on a given dataset. Researchers have leveraged this effect for model selection based on desirable properties such as fairness or sparsity. However, a systematic analysis of different properties across models in the Rashomon set (set of near-optimal models) has not been conducted. In this work, we bridge this gap by introducing a framework that evaluates and compares the performance of models from the Rashomon set against models explicitly optimized for a trustworthiness criterion, such as robustness, privacy, or fairness. We focus on decision trees, which are inherently interpretable models. Our results demonstrate that near-optimal decision trees from the Rashomon set can perform comparably to trees explicitly optimized for trustworthiness. This suggests that leveraging the Rashomon set for model selection can be an effective strategy for building reliable machine learning systems. Our framework can be a valuable tool for both benchmarking Rashomon sets of decision trees and studying the trustworthiness properties of interpretable models.



Paperid:5373
Authors:Yiwei Yang, Chung Peng Lee, Shangbin Feng, Dora Zhao, Bingbing Wen, Anthony Liu, Yulia Tsvetkov, Bill Howe
Abstract:
Finetuning can cause spurious correlations to arise between non-essential features and the target labels, but benchmarks to study these effects involve contrived settings and narrow tasks. In contrast, we consider spurious correlations in multi-modal Large Vision Language Models (LVLMs) pretrained on extensive and diverse datasets without explicit task supervision. We develop a benchmark by sourcing GPT-4o errors on real-world visual-question-answering (VQA) benchmarks, then curating a subset through LVLM-human annotation and synthetic counterfactual evaluation to identify errors caused by spurious correlations. This process yields SpuriVerse, a novel benchmark comprised of 124 distinct types of spurious correlations extracted from real-world datasets, each containing 1 realistic and 10 synthetic VQA samples for a total of 1364 multiple choice questions. We evaluate 15 open and closed-source LVLMs on SpuriVerse, finding that even state-of-the-art closed-source models struggle significantly, achieving at best only 37.1\% accuracy. Fine-tuning on synthetic examples that emphasize the spurious correlation improves performance to 78.40\%, suggesting that training on diverse spurious patterns generalizes to unseen situations: models appear to learn to avoid "shortcuts" and attend to the overall image context.



Paperid:5374
Authors:Chaoyue Liu, Han Bi, Like Hui, Xiao Liu
Abstract:
Non-linear activation functions are widely recognized for enhancing the expressivity of neural networks, which is the primary reason for their widespread implementation. In this work, we reveal a novel and intriguing property of non-linear activations. By comparing enabling and disabling the non-linear activations in the neural network, we demonstrate their specific effects on wide neural networks: (a)better feature separation, i.e., a larger angle separation for similar data in the feature space of model gradient, and (b)better NTK conditioning, i.e., a smaller condition number of neural tangent kernel (NTK). Furthermore, we show that the network depth (i.e., with more non-linear activation operations) further magnifies these effects; in addition, in the infinite-width-then-depth limit, all data are equally separated with a fixed angle in the model gradient feature space, regardless of how similar they are originally in the input space. Note that, without the non-linear activation, i.e., in a linear neural network, the data separation remains the same as for the original inputs and NTK condition number is equivalent to the Gram matrix, regardless of the network depth. Due to the close connection between NTK condition number and convergence theories, our results imply that non-linear activation helps to improve the worst-case convergence rates of gradient based methods.



Paperid:5375
Authors:Jiaheng Zhou, Wei Fang, Luyuan Xie, Yanfeng Zhou, Lianyan Xu, Minfeng Xu, Ge Yang, Yuxing Tang
Abstract:
Breast cancer remains a leading cause of death among women, with early detection significantly improving prognosis. Non-contrast computed tomography (NCCT) scans of the chest, routinely acquired for thoracic assessments, often capture the breast region incidentally, presenting an underexplored opportunity for opportunistic breast lesion detection without additional imaging cost or radiation. However, the subtle appearance of lesions in NCCT and the difficulty of jointly modeling lesion detection and malignancy classification pose unique challenges.In this work, we propose Dual-Res Tandem Mamba-3D (DRT-M3D), a novel multitask framework for opportunistic breast cancer analysis on NCCT scans. DRT-M3D introduces a dual-resolution architecture, which captures fine-grained spatial details for segmentation-based lesion detection and global contextual features for breast-level cancer classification. It further incorporates a tandem input mechanism that models bilateral breast regions jointly through Mamba-3D blocks, enabling cross-breast feature interaction by leveraging subtle asymmetries between the two sides. Our approach achieves state-of-the-art performance in both tasks across multi-institutional NCCT datasets spanning four medical centers. Extensive experiments and ablation studies validate the effectiveness of each key component. Code will be released upon publication.



Paperid:5376
Authors:Xueyi Zhang, Ruicong Wang, Jialu Sun, Siqi Cai, Haizhou Li
Abstract:
Speech is a fundamental form of human communication, and speech perception constitutes the initial stage of language comprehension. Although brain-to-speech interface technologies have made significant progress in recent years, most existing studies focus on neural decoding during speech production. Such approaches heavily rely on articulatory motor regions, rendering them unsuitable for individuals with speech motor impairments, such as those with aphasia or locked-in syndrome. To address this limitation, we construct and release NeuroListen, the first publicly available stereo-electroencephalography (sEEG) dataset specifically designed for auditory reconstruction. It contains over 10 hours of neural–speech paired recordings from 5 clinical participants, covering a wide range of semantic categories. Building on this dataset, we propose HyperSpeech, a multi-band neural decoding framework that employs dynamic spatio-temporal hypergraph neural networks to capture high-order dependencies across frequency, spatial, and temporal dimensions. Experimental results demonstrate that HyperSpeech significantly outperforms existing methods across multiple objective speech quality metrics, and achieves superior performance in human subjective evaluations, validating its effectiveness and advancement. This study provides a dedicated dataset and modeling framework for auditory speech decoding, offering foundations for neural language processing and assistive communication systems.



Paperid:5377
Authors:Xiao Cui, Yulei Qin, Wengang Zhou, Hongsheng Li, Houqiang Li
Abstract:
Dataset distillation seeks to synthesize a compact distilled dataset, enabling models trained on it to achieve performance comparable to models trained on the full dataset. Recent methods for large-scale datasets focus on matching global distributional statistics (e.g., mean and variance), but overlook critical instance-level characteristics and intraclass variations, leading to suboptimal generalization. We address this limitation by reformulating dataset distillation as an Optimal Transport (OT) distance minimization problem, enabling fine-grained alignment at both global and instance levels throughout the pipeline. OT offers a geometrically faithful framework for distribution matching. It effectively preserves local modes, intra-class patterns, and fine-grained variations that characterize the geometry of complex, high-dimensional distributions. Our method comprises three components tailored for preserving distributional geometry: (1) OT-guided diffusion sampling, which aligns latent distributions of real and distilled images; (2) label-image-aligned soft relabeling, which adapts label distributions based on the complexity of distilled image distributions; and (3) OT-based logit matching, which aligns the output of student models with soft-label distributions. Extensive experiments across diverse architectures and large-scale datasets demonstrate that our method consistently outperforms state-of-the-art approaches in an efficient manner, achieving at least 4\% accuracy improvement under IPC=10 settings for each architecture on ImageNet-1K. Code is available at https://anonymous.4open.science/r/DGA.



Paperid:5378
Authors:Jikai Wang, Qifan Zhang, Yu-Wei Chao, Bowen Wen, Xiaohu Guo, Yu Xiang
Abstract:
We introduce a data capture system and a new dataset, HO-Cap, for 3D reconstruction and pose tracking of hands and objects in videos. The system leverages multiple RGB-D cameras and a HoloLens headset for data collection, avoiding the use of expensive 3D scanners or motion capture systems. We propose a semi-automatic method for annotating the shape and pose of hands and objects in the collected videos, significantly reducing the annotation time and cost compared to manual labeling. With this system, we captured a video dataset of humans performing various single- and dual-hand manipulation tasks, including simple pick-and-place actions, handovers between hands, and using objects according to their affordance. This dataset can serve as human demonstrations for research in embodied AI and robot manipulation. Our capture setup and annotation framework will be made available to the community for reconstructing 3D shapes of objects and human hands, as well as tracking their poses in videos.



Paperid:5379
Authors:ShuHang Xun, Sicheng Tao, Jungang Li, Yibo Shi, Zhixin Lin, Zhanhui Zhu, Yibo Yan, Hanqian Li, LingHao Zhang, Shikang Wang, Yixin Liu, Hanbo Zhang, Ying Ma, Xuming Hu
Abstract:
Multimodal Large Language Models (MLLMs) increasingly excel at perception,understanding, and reasoning. However, current benchmarks inadequately evaluate their ability to perform these tasks continuously in dynamic, real-world environments. To bridge this gap, we introduce RT V-Bench, a fine-grained benchmark for MLLM real-time video analysis. RTV-Bench includes three key principles: (1) Multi-Timestamp Question Answering (MTQA), where answers evolve with scene changes; (2) Hierarchical Question Structure, combining basic and advanced queries; and (3) Multi-dimensional Evaluation, assessing the ability of continuous perception, understanding, and reasoning. RTV-Bench contains 552 diverse videos (167.2 hours) and 4,631 high-quality QA pairs. We evaluated leading MLLMs, including proprietary (GPT-4o, Gemini 2.0), open-source offline (Qwen2.5-VL, VideoLLaMA3), and open-source real-time (VITA-1.5, InternLM-XComposer2.5-OmniLive) models. Experiment results show open-source real-time models largely outperform offline ones but still trail top proprietary models. Our analysis also reveals that larger model size or higher frame sampling rates do not significantly boost RTV-Bench performance, sometimes causing slight decreases.This underscores the need for better model architectures optimized for video stream processing and long sequences to advance real-time video analysis with MLLMs.



Paperid:5380
Authors:Anthony Fuller, Yousef Yassin, Junfeng Wen, Tarek Ibrahim, Daniel Kyrollos, James Green, Evan Shelhamer
Abstract:
Vision transformers are ever larger, more accurate, and more expensive to compute.At high resolution, the expense is even more extreme as the number of tokens grows quadratically in the image size. We turn to adaptive computation to cope with this cost by learning to predict where to compute.Our LookWhere method divides the computation between a low-resolution selector and a high-resolution extractor without ever processing the full high-resolution input.We jointly pretrain the selector and extractor without task supervision by distillation from a self-supervised teacher, in effect learning where and what to compute at the same time.Unlike prior token reduction methods, which pay to save by pruning already-computed tokens, and prior token selection methods, which require complex and expensive per-task optimization, LookWhere economically and accurately selects and extracts transferrable representations of images.We show that LookWhere excels at sparse recognition on high-resolution inputs (Traffic Signs), maintaining accuracy while reducing FLOPs by 17x and time by 4x, and standard recognition tasks that are global (ImageNet classification) and local (ADE20K segmentation), improving accuracy while reducing time by 1.36x.



Paperid:5381
Authors:Ziyi Yang, Hanyuan Xie, Yinjun Jia, Xiangzhe Kong, Jiqing Zheng, Ziting Zhang, Yang Liu, Lei Liu, Yanyan Lan
Abstract:
Cyclic peptides exhibit better binding affinity and proteolytic stability compared to their linear counterparts. However, the development of cyclic peptide design models is hindered by the scarcity of data. To address this, we introduce CPSea (Cyclic Peptide Sea), a dataset of 2.64 million cyclic peptide-receptor complexes, curated through systematic mining of the AlphaFold Database (AFDB). Our pipeline extracts compact domains from AFDB, identifies cyclization sites using the β-carbon (Cβ) atom distance thresholds, and applies multi-stage filtering to ensure structural validity and binding compatibility. Compared with existing data of cyclic peptides, CPSea shows similar property distributions in terms of structure validity and wet-lab compatibility. To our knowledge, CPSea is the largest cyclic peptide-receptor dataset to date, enabling end-to-end model training for the first time. Our method is scalable, as cyclic peptides with varied cyclization linkages can be obtained by adjusting (Cβ) distance criteria, making it possible to develop specialized models for different cyclization types.



Paperid:5382
Authors:James Oldfield, Shawn Im, Sharon Li, Mihalis Nicolaou, Ioannis Patras, Grigorios Chrysos
Abstract:
Multilayer perceptrons (MLPs) are an integral part of large language models, yet their dense representations render them difficult to understand, edit, and steer. Recent methods learn interpretable approximations via neuron-level sparsity, yet fail to faithfully reconstruct the original mapping--significantly increasing model's next-token cross-entropy loss. In this paper, we advocate for moving to layer-level sparsity to overcome the accuracy trade-off in sparse layer approximation. Under this paradigm, we introduce Mixture of Decoders (MxDs). MxDs generalize MLPs and Gated Linear Units, expanding pre-trained dense layers into tens of thousands of specialized sublayers. Through a flexible form of tensor factorization, each sparsely activating MxD sublayer implements a linear transformation with full-rank weights--preserving the original decoders' expressive capacity even under heavy sparsity. Experimentally, we show that MxDs significantly outperform state-of-the-art methods (e.g., Transcoders) on the sparsity-accuracy frontier in language models with up to 3B parameters. Further evaluations on sparse probing and feature steering demonstrate that MxDs learn similarly specialized features of natural language--opening up a promising new avenue for designing interpretable yet faithful decompositions. Our code is included at: https://anonymous.4open.science/r/MxD-anon-code-6F9D.



Paperid:5383
Authors:Sean McGrath, Debarghya Mukherjee, Rajarshi Mukherjee, Zixiao Wang
Abstract:
Abstract:In this paper, we explore the asymptotically optimal tuning parameter choice in ridge regression for estimating nuisance functions of a statistical functional that has recently gained prominence in conditional independence testing and causal inference. Given a sample of size $n$, we study estimators of the Expected Conditional Covariance (ECC) between variables $Y$ and $A$ given a high-dimensional covariate $X \in \mathbb{R}^p$. Under linear regression models for $Y$ and $A$ on $X$ and the proportional asymptotic regime $p/n \to c \in (0, \infty)$, we evaluate three existing ECC estimators and two sample splitting strategies for estimating the required nuisance functions. Since no consistent estimator of the nuisance functions exists in the proportional asymptotic regime without imposing further structure on the problem, we first derive debiased versions of the ECC estimators that utilize the ridge regression nuisance function estimators. We show that our bias correction strategy yields $\sqrt{n}$-consistent estimators of the ECC across different sample splitting strategies and estimator choices. We then derive the asymptotic variances of these debiased estimators to illustrate the nuanced interplay between the sample splitting strategy, estimator choice, and tuning parameters of the nuisance function estimators for optimally estimating the ECC. Our analysis reveals that prediction-optimal tuning parameters (i.e., those that optimally estimate the nuisance functions) may not lead to the lowest asymptotic variance of the ECC estimator -- thereby demonstrating the need to be careful in selecting tuning parameters based on the final goal of inference. Finally, we verify our theoretical results through extensive numerical experiments.



Paperid:5384
Authors:Yuchen Xia, Yunjian Xu
Abstract:
We introduce MOSDT, the first algorithm designed for multi-agent offline safe reinforcement learning (MOSRL), alongside MOSDB, the first dataset and benchmark for this domain. Different from most existing knowledge distillation-based multi-agent RL methods, we propose policy self-distillation (PSD) with a new global information reconstruction scheme by fusing the observation features of all agents, streamlining training and improving parameter efficiency. We adopt full parameter sharing across agents, significantly slashing parameter count and boosting returns up to 38.4-fold by stabilizing training. We propose a new plug-and-play cost binary embedding (CBE) module, which encodes cumulative costs as safety binary signals and embeds the signals into return features for efficient information aggregation. On the strong MOSDB benchmark, MOSDT achieves state-of-the-art (SOTA) returns in 14 out of 18 tasks (across all base environments including MuJoCo, Safety Gym, and Isaac Gym) while ensuring complete safety, with only 65% of the execution parameter count of a SOTA single-agent offline safe RL method CDT. Code, dataset, and results are available at this anonymous website: https://github.com/MOSDT/MOSDT.git



Paperid:5385
Authors:Xiang Li, Chendi Wang, Buxin Su, Qi Long, Weijie Su
Abstract:
Abstract:Differentially private (DP) decentralized Federated Learning (FL) allows local users to collaborate without sharing their data with a central server. However, accurately quantifying the privacy budget of private FL algorithms is challenging due to the co-existence of complex algorithmic components such as decentralized communication and local updates. This paper addresses privacy accounting for two decentralized FL algorithms within the $f$-differential privacy ($f$-DP) framework. We develop two new $f$-DP–based accounting methods tailored to decentralized settings: Pairwise Network $f$-DP (PN-$f$-DP), which quantifies privacy leakage between user pairs under random-walk communication, and Secret-based $f$-Local DP (Sec-$f$-LDP), which supports structured noise injection via shared secrets. By combining tools from $f$-DP theory and Markov chain concentration, our accounting framework captures privacy amplification arising from sparse communication, local iterations, and correlated noise. Experiments on synthetic and real datasets demonstrate that our methods yield consistently tighter $(\epsilon, \delta)$ bounds and improved utility compared to Rényi DP–based approaches, illustrating the benefits of $f$-DP in decentralized privacy accounting.



Paperid:5386
Authors:Yongxin He, Shan Zhang, Yixuan Cao, Lei Ma, Ping Luo
Abstract:
Detecting AI-involved text is essential for combating misinformation, plagiarism, and academic misconduct. However, AI text generation includes diverse collaborative processes (AI-written text edited by humans, human-written text edited by AI, and AI-generated text refined by other AI), where various or even new LLMs could be involved. Texts generated through these varied processes exhibit complex characteristics, presenting significant challenges for detection. Current methods model these processes rather crudely, primarily employing binary classification (purely human vs. AI-involved) or multi-classification (treating human-AI collaboration as a new class).We observe that representations of texts generated through different processes exhibit inherent clustering relationships.So, we propose DETree, a novel approach that models the relationships among different processes as a Hierarchical Affinity Tree structure, and introduces a specialized loss function that aligns representations to this tree. To facilitate this learning, we developed RealBench, a comprehensive benchmark dataset that automatically incorporates a wide spectrum of hybrid texts produced through various human-AI collaboration processes.Our method improves performance in hybrid text detection tasks and significantly enhances robustness and generalization in out-of-distribution scenarios, particularly in few-shot learning conditions, further demonstrating the promise of training-based approaches in OOD settings. We will publicly release all code, model checkpoints, and the constructed dataset.



Paperid:5387
Authors:Weixiang Zhao, Jiahe Guo, Yang Deng, Tongtong Wu, Wenxuan Zhang, Yulin Hu, Xingyu Sui, Yanyan Zhao, Wanxiang Che, Bing Qin, Tat-Seng Chua, Ting Liu
Abstract:
Multilingual reasoning remains a significant challenge for large language models (LLMs), with performance disproportionately favoring high-resource languages. Drawing inspiration from cognitive neuroscience, which suggests that human reasoning functions largely independently of language processing, we hypothesize that LLMs similarly encode reasoning and language as separable components that can be disentangled to enhance multilingual reasoning. To evaluate this, we perform a causal intervention by ablating language-specific representations at inference time. Experiments on 10 open-source LLMs spanning 11 typologically diverse languages show that this language-specific ablation consistently boosts multilingual reasoning performance. Layer-wise analyses further confirm that language and reasoning representations can be effectively decoupled throughout the model, yielding improved multilingual reasoning capabilities, while preserving top-layer language features remains essential for maintaining linguistic fidelity. Compared to post-training such as supervised fine-tuning or reinforcement learning, our training-free ablation achieves comparable or superior results with minimal computational overhead. These findings shed light on the internal mechanisms underlying multilingual reasoning in LLMs and suggest a lightweight and interpretable strategy for improving cross-lingual generalization.



Paperid:5388
Authors:Xinhao Luo, Zihan Liu, Yangjie Zhou, Shihan Fang, Ziyu Huang, Yu Feng, Chen Zhang, Shixuan Sun, Zhenzhe Zheng, Jingwen Leng, Minyi Guo
Abstract:
Abstract:Large language model (LLM) decoding suffers from high latency due to fragmented execution across operators and heavy reliance on off-chip memory for data exchange and reduction. This execution model limits opportunities for fusion and incurs significant memory traffic and kernel launch overhead.While modern architectures such as NVIDIA Hopper provide distributed shared memory and low-latency intra-cluster interconnects, they expose only low-level data movement instructions, lacking structured abstractions for collective on-chip communication.To bridge this software-hardware gap, we introduce two cluster-level communication primitives, ClusterReduce and ClusterGather, which abstract common communication patterns and enable structured, high-speed data exchange and reduction between thread blocks within a cluster, allowing intermediate results to be on-chip without involving off-chip memory.Building on these abstractions, we design ClusterFusion, an execution framework that schedules communication and computation jointly to expand operator fusion scope by composing decoding stages such as QKV Projection, Attention, and Output Projection into a single fused kernels.Evaluations on H100 GPUs show that ClusterFusion outperforms state-of-the-art inference frameworks by $1.61\times$ on average in end-to-end latency across different models and configurations.



Paperid:5389
Authors:Mizuki Niihori, Shuichi Nishino, Teruyuki Katsuoka, Tomohiro Shiraishi, Kouichi Taji, Ichiro Takeuchi
Abstract:
Abstract:In real-world applications, anomaly detection (AD) often operates without access to anomalous data, necessitating semi-supervised methods that rely solely on normal data. Among these methods, deep $k$-nearest neighbor (deep $k$NN) AD stands out for its interpretability and flexibility, leveraging distance-based scoring in deep latent spaces. Despite its strong performance, deep $k$NN lacks a mechanism to quantify uncertainty—an essential feature for critical applications such as industrial inspection. To address this limitation, we propose a statistical framework that quantifies the significance of detected anomalies in the form of $p$-values, thereby enabling control over false positive rates. A central challenge lies in managing selection bias, which we tackle using Selective Inference—a principled method for conducting inference conditioned on data-driven selections. We evaluate our method on diverse datasets and demonstrate that it provides reliable AD well-suited for industrial use cases.



Paperid:5390
Authors:HantingYan Yan, Pan Mu, SHIQI ZHANG, Yuchao Zhu, jinglin zhang, Cong Bai
Abstract:
Tropical Cyclone (TC) estimation aims to accurately estimate various TC attributes in real time. However, distribution shifts arising from the complex and dynamic nature of TC environmental fields, such as varying geographical conditions and seasonal changes, present significant challenges to reliable estimation. Most existing methods rely on multi-modal fusion for feature extraction but overlook the intrinsic distribution of feature representations, leading to poor generalization under out-of-distribution (OOD) scenarios. To address this, we propose an effective Identity Distribution-Oriented Physical Invariant Learning framework (IDOL), which imposes identity-oriented constraints to regulate the feature space under the guidance of prior physical knowledge, thereby dealing distribution variability with physical invariance. Specifically, the proposed IDOL employs the wind field model and dark correlation knowledge of TC to model task-shared and task-specific identity tokens. These tokens capture task dependencies and intrinsic physical invariances of TC, enabling robust estimation of TC wind speed, pressure, inner-core, and outer-core size under distribution shifts. Extensive experiments conducted on multiple datasets and tasks demonstrate the outperformance of the proposed IDOL, verifying that imposing identity-oriented constraints based on prior physical knowledge can effectively mitigates diverse distribution shifts in TC estimation.



Paperid:5391
Authors:Taehoon Yoon, Yunhong Min, Kyeongmin Yeo, Minhyuk Sung
Abstract:
Abstract:We introduce $\Psi$-Sampler, an SMC-based framework incorporating pCNL-based initial particle sampling for effective inference-time reward alignment with a score-based model. Inference-time reward alignment with score-based generative models has recently gained significant traction, following a broader paradigm shift from pre-training to post-training optimization. At the core of this trend is the application of Sequential Monte Carlo (SMC) to the denoising process. However, existing methods typically initialize particles from the Gaussian prior, which inadequately captures reward-relevant regions and results in reduced sampling efficiency. We demonstrate that initializing from the reward-aware posterior significantly improves alignment performance. To enable posterior sampling in high-dimensional latent spaces, we introduce the preconditioned Crank–Nicolson Langevin (pCNL) algorithm, which combines dimension-robust proposals with gradient-informed dynamics. This approach enables efficient and scalable posterior sampling and consistently improves performance across various reward alignment tasks, including layout-to-image generation, quantity-aware generation, and aesthetic-preference generation, as demonstrated in our experiments.



Paperid:5392
Authors:Yue Guan, Changming Yu, Shihan Fang, Weiming Hu, Zaifeng Pan, Zheng Wang, Zihan Liu, Yangjie Zhou, Yufei Ding, Minyi Guo, Jingwen Leng
Abstract:
Abstract:Speculative decoding improves LLM inference by generating and verifying multiple tokens in parallel, but existing systems suffer from suboptimal performance due to a mismatch between dynamic speculation and static runtime assumptions. We present Yggdrasil, a co-designed system that enables latency-optimal speculative decoding through context-aware tree drafting and compiler-friendly execution. Yggdrasil introduces an equal-growth tree structure for static graph compatibility, a latency-aware optimization objective for draft selection, and stage-based scheduling to reduce overhead. Yggdrasil supports unmodified LLMs and achieves up to $3.98\times$ speedup over state-of-the-art baselines across multiple hardware setups.



Paperid:5393
Authors:Ziwei Shi, Xiaoran Zhang, Wenjing Xu, Yan Xia, Yu Zang, Siqi Shen, Cheng Wang
Abstract:
Abstract:We tackle the challenge of LiDAR-based place recognition, which traditionally depends on costly and time-consuming prior 3D maps. To overcome this, we first construct XA-L\&RSI dataset, which encompasses approximately $110,000$ remote sensing submaps and $13,000$ LiDAR point cloud submaps captured in urban scenes, and propose a novel method, L2RSI, for cross-view LiDAR place recognition using high-resolution Remote Sensing Imagery. This approach enables large-scale localization capabilities at a reduced cost by leveraging readily available overhead images as map proxies. L2RSI addresses the dual challenges of cross-view and cross-modal place recognition by learning feature alignment between point cloud submaps and remote sensing submaps in the semantic domain. Additionally, we introduce a novel probability propagation method based on particle estimation to refine position predictions, effectively leveraging temporal and spatial information. This approach enables large-scale retrieval and cross-scene generalization without fine-tuning. Extensive experiments on XA-L\&RSI demonstrate that, within a $100km^2$ retrieval range, L2RSI accurately localizes $83.27\%$ of point cloud submaps within a $30m$ radius for top-$1$ retrieved location.



Paperid:5394
Authors:Mingyuan Li, Tong Jia, Hao Wang, Bowen Ma, Luhui, Shiyi Guo, Da Cai, Dongyue Chen
Abstract:
Prohibited item detection based on X-ray images is one of the most effective security inspection methods. However, the foreground-background feature coupling caused by the overlapping phenomenon specific to X-ray images makes general detectors designed for natural images perform poorly. To address this issue, we propose a Category Semantic Prior Contrastive Learning (CSPCL) mechanism, which aligns the class prototypes perceived by the classifier with the content queries to correct and supplement the missing semantic information responsible for classification, thereby enhancing the model sensitivity to foreground features. To achieve this alignment, we design a specific contrastive loss, CSP loss, which includes Intra-Class Truncated Attraction (ITA) loss and Inter-Class Adaptive Repulsion (IAR) loss, and outperforms classic N-pair loss and InfoNCE loss. Specifically, ITA loss leverages class prototypes to attract intra-class category-specific content queries while preserving necessary distinctiveness. IAR loss utilizes class prototypes to adaptively repel inter-class category-specific content queries based on the similarity between class prototypes, helping disentangle features of similar categories. CSPCL is general and can be easily integrated into Deformable DETR-based models. Extensive experiments on the PIXray, OPIXray, PIDray, and CLCXray datasets demonstrate that CSPCL significantly enhances the performance of various state-of-the-art models without increasing complexity. The core code is in the supplementary materials, and it will be open sourced once the paper is accepted.



Paperid:5395
Authors:Sajad Khodadadian, Martin Zubeldia
Abstract:
Abstract:Polyak–Ruppert averaging is a widely used technique to achieve the optimal asymptotic variance of stochastic approximation (SA) algorithms, yet its high-probability performance guarantees remain underexplored in general settings. In this paper, we present a general framework for establishing non-asymptotic concentration bounds for the error of averaged SA iterates. Our approach assumes access to individual concentration bounds for the unaveraged iterates and yields a sharp bound on the averaged iterates. We also construct an example, showing the tightness of our result up to constant multiplicative factors. As direct applications, we derive tight concentration bounds for contractive SA algorithms and for algorithms such as temporal difference learning and $Q$-learning with averaging, obtaining new bounds in settings where traditional analysis is challenging.



Paperid:5396
Authors:Tae-Young Lee, Juwon Seo, Jong Hwan Ko, Gyeong-Moon Park
Abstract:
Abstract:Recent advances in diffusion models have enabled high-quality synthesis of specific subjects, such as identities or objects. This capability, while unlocking new possibilities in content creation, also introduces significant privacy risks, as personalization techniques can be misused by malicious users to generate unauthorized images. Although several studies have attempted to counter this by generating adversarially perturbed samples designed to disrupt personalization, they rely on unrealistic assumptions and become ineffective in the presence of even a few clean images or under simple image transformations. To address these challenges, we shift the protection target from the images to the diffusion model itself to hinder the personalization of specific subjects, through our novel framework called $\textbf{A}$nti-$\textbf{P}$ersonalized $\textbf{D}$iffusion $\textbf{M}$odels ($\textbf{APDM}$). We first provide a theoretical analysis demonstrating that a naive approach of existing loss functions to diffusion models is inherently incapable of ensuring convergence for robust anti-personalization. Motivated by this finding, we introduce Direct Protective Optimization (DPO), a novel loss function that effectively disrupts subject personalization in the target model without compromising generative quality. Moreover, we propose a new dual-path optimization strategy, coined Learning to Protect (L2P). By alternating between personalization and protection paths, L2P simulates future personalization trajectories and adaptively reinforces protection at each step.Experimental results demonstrate that our framework outperforms existing methods, achieving state-of-the-art performance in preventing unauthorized personalization.



Paperid:5397
Authors:Kai Wang, Siqiang Luo, Caihua Shan, Yifei Shen
Abstract:
Inspired by the success of large language models, there is a trend toward developing graph foundation models to conduct diverse downstream tasks in various domains. However, current models often require extra fine-tuning to apply their learned structural and semantic representations to new graphs, which limits their versatility. Recent breakthroughs in zero-shot inductive reasoning on knowledge graphs (KGs), offer us a new perspective on extending KG reasoning to general graph applications. In this paper, we introduce SCR, a unified graph reasoning framework designed to train on knowledge graphs and effectively generalize across a wide range of graph tasks and domains. We begin by designing the task-specific KG structures to establish a unified topology for different task formats. Then we propose semantic-conditioned message passing, a novel mechanism addressing the inherent semantic isolation in traditional KG reasoning, by jointly modeling structural and semantic invariance patterns in graph representations. To demonstrate the effectiveness, we evaluate the inductive reasoning capability of SCR using 38 diverse graph datasets, covering node-level, link-level, and graph-level tasks across multiple domains. Our results show substantial performance gains over existing foundation models and supervised baselines, highlighting the efficacy and adaptability of our approach.



Paperid:5398
Authors:Li-Jun Zhao, Zhen-Duo Chen, Yongxin Wang, Xin Luo, Xin-Shun Xu
Abstract:
Recently proposed Fine-Grained Few-Shot Class-Incremental Learning (FG-FSCIL) offers a practical and efficient solution for enabling models to incrementally learn new fine-grained categories under limited data conditions. However, existing methods still settle for the fine-grained feature extraction capabilities learned from the base classes. Unlike conventional datasets, fine-grained categories exhibit subtle inter-class variations, naturally fostering latent synergy among sub-categories. Meanwhile, the incremental learning framework offers an opportunity to progressively strengthen this synergy by incorporating new sub-category data over time. Motivated by this, we theoretically formulate the FSCIL problem and derive a generalization error bound within a shared fine-grained meta-category environment. Guided by our theoretical insights, we design a novel Meta-Environment Learner (MEL) for FG-FSCIL, which evolves fine-grained feature extraction to enhance meta-environment understanding and simultaneously regularizes hypothesis space complexity. Extensive experiments demonstrate that our method consistently and significantly outperforms existing approaches.



Paperid:5399
Authors:Wei Li, Renshan Zhang, Rui Shao, Jie He, Liqiang Nie
Abstract:
Abstract:Recent Vision-Language-Action (VLA) models built on pre-trained Vision-Language Models (VLMs) require extensive post-training, resulting in high computational overhead that limits scalability and deployment. Existing sparsification strategies—such as Mixture-of-Depths, layer skipping, and early exit—fall short by neglecting the semantic coupling across vision-language-action modalities, and focusing narrowly on intra-LLM computation while overlooking end-to-end coherence from perception to control. To address these challenges, we propose **CogVLA**, a Cognition-Aligned Vision-Language-Action framework that leverages instruction-driven routing and sparsification to improve both efficiency and performance. CogVLA draws inspiration from human multimodal coordination and introduces a 3-stage progressive architecture. 1) **Encoder-FiLM based Aggregation Routing (EFA-Routing)** injects instruction information into the vision encoder to selectively aggregate and compress dual-stream visual tokens, forming a instruction-aware latent representation. 2) Building upon this compact visual encoding, **LLM-FiLM based Pruning Routing (LFP-Routing)** introduces action intent into the language model by pruning instruction-irrelevant visually grounded tokens, thereby achieving token-level sparsity. 3) To ensure that compressed perception inputs can still support accurate and coherent action generation, we introduce **V‑L‑A Coupled Attention (CAtten)**, which combines causal vision-language attention with bidirectional action parallel decoding.Extensive experiments on the LIBERO benchmark and real-world robotic tasks demonstrate that CogVLA achieves state-of-the-art performance with success rates of 97.4\% and 70.0\%, respectively, while reducing training costs by 2.5$\times$ and decreasing inference latency by 2.8$\times$ compared to OpenVLA.



Paperid:5400
Authors:Zhenhao Zhong, Zhibin Gu, Pengpeng Yang, Yaqian zhou, Ruiqiang Guo
Abstract:
Abstract:Tensorized Incomplete Multi-View Clustering (TIMVC) algorithms have attracted growing attention for their ability to capture high-order correlations across multiple views. However, most existing TIMVC methods rely on simplistic noise assumptions using specific norms (e.g., $\ell_1$ or $\ell_{2,1}$), which fail to reflect the complex noise patterns encountered in real-world scenarios. Moreover, they primarily focus on modeling the global Euclidean structure of the tensor representation, while overlooking the preservation of local manifold structures. To address these limitations, we propose a novel approach, GaUssian regressIon-driven TIMVC with dual mAnifold Regularization (GUITAR). Specifically, we employ a Gaussian regression model to characterize complex noise distributions in a more realistic and flexible manner. Meanwhile, a dual manifold regularization is introduced in tensor representation learning, simultaneously modeling manifold information at both the view-specific and cross-view consensus levels, thereby promoting intra-view and inter-view consistency in the tensor representation. Furthermore, to better capture the intrinsic low-rank structure, we propose the high-preservation $\ell_{\delta}$-norm tensor rank constraint, which applies differentiated penalties to the singular values, thereby enhancing the robustness of the tensor representation. In addition, an efficient optimization algorithm is developed to solve the resulting non-convex problem with provable convergence. Extensive experiments on six datasets demonstrate that our method outperforms SOTA approaches.



Paperid:5401
Authors:Rui Xu, Dakuan Lu, Zicheng Zhao, Xiaoyu Tan, Xintao Wang, Siyu Yuan, Jiangjie Chen, yinghui xu
Abstract:
Spatial reasoning is a key capability in the field of artificial intelligence, especially crucial in areas such as robotics, computer vision, and natural language understanding. However, evaluating the ability of multimodal large language models (MLLMs) in complex spatial reasoning still faces challenges, particularly in scenarios requiring multi-step reasoning and precise mathematical constraints. This paper introduces ORIGAMISPACE, a new dataset and benchmark designed to evaluate the multi-step spatial reasoning ability and the capacity to handle mathematical constraints of MLLMs through origami tasks. The dataset contains 350 data instances, each comprises a strictly formatted crease pattern (CP diagram), the Compiled Flat Pattern, the complete Folding Process, and the final Folded Shape Image. We propose four evaluation tasks: Pattern Prediction, Multi-step Spatial Reasoning, Spatial Relationship Prediction, and End-to-End CP Code Generation. For the CP code generation task, we design an interactive environment and explore the possibility of using reinforcement learning methods to train MLLMs. Through experiments on existing MLLMs, we initially reveal the strengths and weaknesses of these models in handling complex spatial reasoning tasks.



Paperid:5402
Authors:Bohan Wang, Mingze Zhou, Zhongqi Yue, wang lin, Kaihang Pan, Liyu Jia, Wentao Hu, Wei Zhao, Hanwang Zhang
Abstract:
Reinforcement learning (RL) has become an indispensable post-training step for unlocking the full potential of Large Language Models (LLMs). Its core motivation is to incentivize the model’s inference trajectory via a reward model, effectively balancing the exploration–exploitation trade-off in scenarios where collecting exhaustive input–output ground-truth pairs is infeasible. This motivation naturally extends to visual generation, where perfect alignment between an image and a textual prompt is inherently ambiguous and often unattainable. However, existing visual generative models are not yet ready for RL due to the following two fundamental drawbacks that undermine the foundations of RL: 1) For diffusion-based models, the actual generation trajectories of sampled images cannot be reliably rewarded, as diffusion inversion is notoriously difficult. 2) For autoregressive (AR) models, we show that the widely used spatial visual tokens do not satisfy the Bellman equation and thus violate the policy improvement theorem of RL. To this end, we propose to use Selftok (Self-consistency Tokenizer), which represents each image as a sequential 1D stream of discrete, autoregressive tokens. Together with language, we train a pure AR vision-language model (VLM) for visual generation. Impressively, without using any text-image training pairs, a simple policy gradient algorithm applied to Selftok tokens significantly boosts visual generation performance, surpassing existing models by a large margin. Implementation details are provided in the Appendix.



Paperid:5403
Authors:Sanjoy Chowdhury, Mohamed Elmoghany, Yohan Abeysinghe, Junjie Fei, Sayan Nag, Salman Khan, Mohamed Elhoseiny, Dinesh Manocha
Abstract:
Large multimodal models (LMMs) have shown remarkable progress in audio-visual understanding, yet they struggle with real-world scenarios that requirecomplex reasoning across extensive video collections. Existing benchmarks forvideo question answering remain limited in scope, typically involving one clipper query, which falls short of representing the challenges of large-scale, audio-visual retrieval and reasoning encountered in practical applications. To bridgethis gap, we introduce a novel task named AV-HaystacksQA, where the goalis to identify salient segments across different videos in response to a query andlink them together to generate the most informative answer. To this end, wepresent AVHaystacks, an audio-visual benchmark comprising 3100 annotated QApairs designed to assess the capabilities of LMMs in multi-video retrieval andtemporal grounding task. Additionally, we propose a model-agnostic, multi-agentframework MAGNET to address this challenge, achieving up to 89% and 65%relative improvements over baseline methods on BLEU@4 and GPT evaluationscores in QA task on our proposed AVHaystacks. To enable robust evaluation ofmulti-video retrieval and temporal grounding for optimal response generation, weintroduce two new metrics, STEM, which captures alignment errors between aground truth and a predicted step sequence and MTGS, to facilitate balanced andinterpretable evaluation of segment-level grounding performance. Our code anddataset will be released publicly



Paperid:5404
Authors:Zhengqiang ZHANG, Rongyuan Wu, Lingchen Sun, Lei Zhang
Abstract:
Effective and efficient tokenization plays an important role in image representation and generation. Conventional methods, constrained by uniform 2D/1D grid tokenization, are inflexible to represent regions with varying shapes and textures and at different locations, limiting their efficacy of feature representation. In this work, we proposeGPSToken, a novelGaussianParameterizedSpatially-adaptiveTokenization framework, to achieve non-uniform image tokenization by leveraging parametric 2D Gaussians to dynamically model the shape, position, and textures of different image regions. We first employ an entropy-driven algorithm to partition the image into texture-homogeneous regions of variable sizes. Then, we parameterize each region as a 2D Gaussian (mean for position, covariance for shape) coupled with texture features. A specialized transformer is trained to optimize the Gaussian parameters, enabling continuous adaptation of position/shape and content-aware feature extraction. During decoding, Gaussian parameterized tokens are reconstructed into 2D feature maps through a differentiable splatting-based renderer, bridging our adaptive tokenization with standard decoders for end-to-end training. GPSToken disentangles spatial layout (Gaussian parameters) from texture features to enable efficient two-stage generation: structural layout synthesis using lightweight networks, followed by structure-conditioned texture generation. Experiments demonstrate the state-of-the-art performance of GPSToken, which achieves rFID and FID scores of 0.65 and 1.64 on image reconstruction and generation tasks using 128 tokens, respectively. Codes and models will be released.



Paperid:5405
Authors:Yuena Lin, Hao Wei, Hai-Chun Cai, Bohang Sun, Tao Yang, Zhen Yang, Gengyu Lyu
Abstract:
Graph contrastive learning (GCL) aims to learn self-supervised representations by distinguishing positive and negative sample pairs generated from multiple augmented graph views. Despite showing promising performance, GCL still suffers from two critical biases: (1)Similarity estimation biasarises when feature elements that support positive pair alignment are suppressed by conflicting components within the representation, causing truly positive pairs to appear less similar. (2)Semantic shift biasoccurs when random augmentations alter the underlying semantics of samples, leading to incorrect positive or negative assignments and injecting noise into training. To address these issues, we propose CaliGCL, a GCL model for calibrating the biases by integrating an exponential partitioned similarity measure and a semantics-consistency discriminator. The exponential partitioned similarity computes the similarities among fine-grained partitions obtained through splitting representation vectors and uses exponential scaling to emphasize aligned (positive) partitions while reducing the influence of misaligned (negative) ones. The discriminator dynamically identifies whether augmented sample pairs maintain semantic consistency, enabling correction of misleading contrastive supervision signals. These components jointly reduce biases in similarity estimation and sample pairing, guiding the encoder to learn more robust and semantically meaningful representations. Extensive experiments on multiple benchmarks show that CaliGCL effectively mitigates both types of biases and achieves state-of-the-art performance.



Paperid:5406
Authors:Yuhao Zhou, Yiheng Wang, Xuming He, Ruoyao Xiao, Zhiwei Li, Qiantai Feng, Zijie Guo, Yuejin Yang, Hao Wu, Wenxuan Huang, Jiaqi Wei, Dan Si, YAO XIUQI, Jia Bu, Haiwen Huang, Tianfan Fu, SHIXIANG TANG, Ben Fei, Dongzhan Zhou, Fenghua Ling, Yan Lu, Siqi Sun, Chenhui Li, Guanjie Zheng, Jiancheng Lv, zhangwenlong, LEI BAI
Abstract:
Scientific discoveries increasingly rely on complex multimodal reasoning based on information-intensive scientific data and domain-specific expertise. Empowered by expert-level scientific benchmarks, scientific Multimodal Large Language Models (MLLMs) hold the potential to significantly enhance this discovery process in realistic workflows. However, current scientific benchmarks mostly focus on evaluating the knowledge understanding capabilities of MLLMs, leading to an inadequate assessment of their perception and reasoning abilities. To address this gap, we present the Scientists’ First Exam (SFE) benchmark, designed to evaluate the scientific cognitive capacities of MLLMs through three interconnected levels:scientific signal perception,scientific attribute understanding,scientific comparative reasoning. Specifically, SFE comprises 830 expert-verified VQA pairs across three question types, spanning 66 multimodal tasks across five high-value disciplines. Extensive experiments reveal that currentstate-of-the-artGPT-o3 and InternVL-3 achieve only 34.08% and 26.52% on SFE, highlighting significant room for MLLMs to improve in scientific realms. We hope the insights obtained in SFE will facilitate further developments in AI-enhanced scientific discoveries.



Paperid:5407
Authors:Danial Samadi Vahdati, Tai Nguyen, Ekta Prashnani, Koki Nagano, david luebke, Orazio Gallo, Matthew Stamm
Abstract:
AI-based talking-head videoconferencing systems reduce bandwidth by transmitting a latent representation of a speaker’s pose and expression, which is used to synthesize frames on the receiver's end. However, these systems are vulnerable to “puppeteering” attacks, where an adversary controls the identity of another person in real-time. Traditional deepfake detectors fail here, as all video content is synthetic. We propose a novel biometric defense that detects identity leakage in the transmitted latent representation. Our metric-learning approach disentangles identity cues from pose and expression, enabling detection of unauthorized swaps. Experiments across multiple talking-head models show that our method consistently outperforms prior defenses, operates in real time on consumer GPUs, and generalizes well to out-of-distribution data. By targeting the latent features shared during normal operation, our method offers a practical and robust safeguard against puppeteering.



Paperid:5408
Authors:Yaxin Luo, Zhaoyi Li, Jiacheng Liu, Jiacheng Cui, Xiaohan Zhao, Zhiqiang Shen
Abstract:
CAPTCHAs have been a critical bottleneck for deploying web agents in real-world applications, often blocking them from completing end-to-end automation tasks. While modern multimodal LLM agents have demonstrated impressive performance in static perception tasks, their ability to handle interactive, multi-step reasoning challenges like CAPTCHAs is largely untested. To address this gap, we introduceOpen CaptchaWorld, the first web-based benchmark and platform specifically designed to evaluate the visual reasoning and interaction capabilities of MLLM-powered agents through diverse and dynamic CAPTCHA puzzles. Our benchmark spans 20 modern CAPTCHA types, totaling 225 CAPTCHAs, annotated with a new metric we propose: CAPTCHA Reasoning Depth, which quantifies the number of cognitive and motor steps required to solve each puzzle. Experimental results show that humans consistently achieve near-perfect scores, state-of-the-art MLLM agents struggle significantly, with success rates at most40.0\%by Browser-Use Openai-o3, far below human-level performance,93.3\%. This highlights Open CaptchaWorld as a vital benchmark for diagnosing the limits of current multimodal agents and guiding the development of more robust multimodal reasoning systems.



Paperid:5409
Authors:Xianda Guo, Ruijun Zhang, Yiqun Duan, Yuhang He, Dujun Nie, Wenke Huang, Chenming Zhang, Shuai Liu, Hao Zhao, Long Chen
Abstract:
Accurate spatial reasoning in outdoor environments—covering geometry, object pose, and inter-object relationships—is fundamental to downstream tasks such as mapping, motion forecasting, and high-level planning in autonomous driving. We introduce SURDS, a large-scale benchmark designed to systematically evaluate the spatial reasoning capabilities of vision language models (VLMs). Built on the nuScenes dataset, SURDS comprises 41,080 vision–question–answer training instances and 9,250 evaluation samples, spanning six spatial categories: orientation, depth estimation, pixel-level localization, pairwise distance, lateral ordering, and front–behind relations. We benchmark leading general-purpose VLMs, including GPT, Gemini, and Qwen, revealing persistent limitations in fine-grained spatial understanding. To address these deficiencies, we go beyond static evaluation and explore whether alignment techniques can improve spatial reasoning performance. Specifically, we propose a reinforcement learning–based alignment scheme leveraging spatially grounded reward signals—capturing both perception-level accuracy (location) and reasoning consistency (logic). We further incorporate final-answer correctness and output-format rewards to guide fine-grained policy adaptation. Our GRPO-aligned variant achieves overall score of 40.80 in SURDS benchmark. Notably, it outperforms proprietary systems such as GPT-4o (13.30) and Gemini-2.0-flash (35.71). To our best knowledge, this is the first study to demonstrate that reinforcement learning–based alignment can significantly and consistently enhance the spatial reasoning capabilities of VLMs in real-world driving contexts. We release the SURDS benchmark, evaluation toolkit, and GRPO alignment code through: https://github.com/XiandaGuo/Drive-MLLM.



Paperid:5410
Authors:Yassine El Ouahidi, Jonathan Lys, Philipp Thölke, Nicolas Farrugia, Bastien Pasdeloup, Vincent Gripon, Karim Jerbi, Giulia Lioi
Abstract:
Foundation models have transformed AI by reducing reliance on task-specific data through large-scale pretraining. While successful in language and vision, their adoption in EEG has lagged due to the heterogeneity of public datasets, which are collected under varying protocols, devices, and electrode configurations. Existing EEG foundation models struggle to generalize across these variations, often restricting pretraining to a single setup, resulting in suboptimal performance, in particular under linear probing.We present REVE (Representation for EEG with Versatile Embeddings), a pretrained model explicitly designed to generalize across diverse EEG signals. REVE introduces a novel 4D positional encoding scheme that enables it to process signals of arbitrary length and electrode arrangement. Using a masked autoencoding objective, we pretrain REVE on over 60,000 hours of EEG data from 92 datasets spanning 25,000 subjects, representing the largest EEG pretraining effort to date.REVE achieves state-of-the-art results on 10 downstream EEG tasks, including motor imagery classification, seizure detection, sleep staging, cognitive load estimation, and emotion recognition. With little to no fine-tuning, it demonstrates strong generalization, and nuanced spatio-temporal modeling. We release code, pretrained weights, and tutorials to support standardized EEG research and accelerate progress in clinical neuroscience.



Paperid:5411
Authors:Pukang Ye, Luo Junwei, Saipan Zhou, Jiachen Shen, Shangmin Dou, Hanzhe Yao, Xiaolei Dong, Yunbo Yang, Zhenfu Cao
Abstract:
Abstract:Data duplication within large-scale corpora often impedes large language models' (LLMs) performance and privacy. In privacy-concerned federated learning scenarios, conventional deduplication methods typically rely on trusted third parties to perform uniform deletion, risking loss of informative samples while introducing privacy vulnerabilities. To address these gaps, we propose Federated ReWeighting (FedRW), the first privacy-preserving framework, to the best of our knowledge, that performs soft deduplication via sample reweighting instead of deletion in federated LLM training, without assuming a trusted third party. At its core, FedRW proposes a secure, frequency-aware reweighting protocol through secure multi-party computation, coupled with a parallel orchestration strategy to ensure efficiency and scalability. During training, FedRW utilizes an adaptive reweighting mechanism with global sample frequencies to adjust individual loss contributions, effectively improving generalization and robustness. Empirical results demonstrate that FedRW outperforms the state-of-the-art method by achieving up to $28.78\times$ speedup in preprocessing and approximately $11.42$\% improvement in perplexity, while offering enhanced security guarantees. FedRW thus establishes a new paradigm for managing duplication in federated LLM training.



Paperid:5412
Authors:Tianyi Chen, Pengxiao Lin, Zhiwei Wang, Zhi-Qin Xu
Abstract:
State Space Models (SSMs) have emerged as promising alternatives to attention mechanisms, with the Mamba architecture demonstrating impressive performance and linear complexity for processing long sequences. However, the fundamental differences between Mamba and Transformer architectures remain incompletely understood. In this work, we use carefully designed synthetic tasks to reveal Mamba's inherent limitations. Through experiments, we identify that Mamba's nonlinear convolution introduces an asymmetry bias that significantly impairs its ability to recognize symmetrical patterns and relationships. Using composite function and inverse sequence matching tasks, we demonstrate that Mamba strongly favors compositional solutions over symmetrical ones and struggles with tasks requiring the matching of reversed sequences. We show these limitations stem not from the SSM module itself but from the nonlinear convolution preceding it, which fuses token information asymmetrically. These insights provide a new understanding of Mamba's constraints and suggest concrete architectural improvements for future sequence models.



Paperid:5413
Authors:Feng Jiang, Mangal Prakash, Hehuan Ma, Jianyuan Deng, Yuzhi Guo, Maolaaisha Aminanmu, Tommaso Mansi, Rui Liao, Junzhou Huang
Abstract:
Molecular property prediction aims to learn representations that map chemical structures to functional properties. While multimodal learning has emerged as a powerful paradigm to learn molecular representations, prior works have largely overlooked textual and taxonomic information of molecules for representation learning. We introduce TRIDENT, a novel framework that integrates molecular SMILES, textual descriptions, and taxonomic functional annotations to learn rich molecular representations. To achieve this, we curate a comprehensive dataset of molecule-text pairs with structured, multi-level functional annotations. Instead of relying on conventional contrastive loss, TRIDENT employs a volume-based alignment objective to jointly align tri-modal features at the global level, enabling soft, geometry-aware alignment across modalities. Additionally, TRIDENT introduces a novel local alignment objective that captures detailed relationships between molecular substructures and their corresponding sub-textual descriptions. A momentum-based mechanism dynamically balances global and local alignment, enabling the model to learn both broad functional semantics and fine-grained structure-function mappings. TRIDENT achieves state-of-the-art performance on 11 downstream tasks, demonstrating the value of combining SMILES, textual, and taxonomic functional annotations for molecular property prediction.



Paperid:5414
Authors:Haidong Kang, Lihong Lin, Hanling Wang
Abstract:
Abstract:The success of computer vision tasks is mainly attributed to the architectural design of neural networks. This highlights the need to automatically design high-performance architectures via Neural Architecture Search (NAS). To accelerate the search process, training-free NAS is proposed, which aims to search high-performance architectures at initialization via zero-cost proxies (ZCPs). However, existing zero-cost proxies heavily rely on manual design, which is often labor-intensive and requires extensive expert knowledge. In addition, these crafted proxies often suffer from poor correlation with final model performance and high computational complexity, severely limiting NAS efficiency in real-world applications. To address those issues, this paper proposes a novel Large Language Models (LLMs)-driven $\underline{A}$utomatic $\underline{P}$roxy $\underline{D}$iscovery ($\textbf{APD}$) framework, which revolutionizes the design paradigm of ZCPs by leveraging LLMs to automatically discover optimal ZCPs for Training-Free NAS. Moreover, we utilize actor-critic based reinforcement learning to optimize prompts, enabling to generate better ZCPs in the next generation. We conduct extensive experiments on mainstream NAS benchmarks, demonstrating APD excels in both performance and efficiency. Besides, we firmly believe that our APD will dramatically benefit the deep learning community through providing novel paradigm of design algorithms via LLMs.



Paperid:5415
Authors:Qiwen Gu, Xufei Wang, Junqiao Zhao, Siyue Tao, Tiantian Feng, Ziqiao Wang, Guang Chen
Abstract:
Visual Place Recognition (VPR) enables robust localization through image retrieval based on learned descriptors. However, drastic appearance variations of images at the same place caused by viewpoint changes can lead to inconsistent supervision signals, thereby degrading descriptor learning. Existing methods either rely on manually defined cropping rules or labeled data for view differentiation, but they suffer from two major limitations: (1) reliance on labels or handcrafted rules restricts generalization capability; (2) even within the same view direction, occlusions can introduce feature ambiguity. To address these issues, we propose MutualVPR, a mutual learning framework that integrates unsupervised view self-classification and descriptor learning. We first group images by geographic coordinates, then iteratively refine the clusters using K-means to dynamically assign place categories without manual labeling. Specifically, we adopt a DINOv2-based encoder to initialize the clustering. During training, the encoder and clustering co-evolve, progressively separating drastic appearance variations of the same place and enabling consistent supervision. Furthermore, we find that capturing fine-grained image differences at a place enhances robustness. Experiments demonstrate that MutualVPR achieves state-of-the-art (SOTA) performance across multiple datasets, validating the effectiveness of our framework in improving view direction generalization, occlusion robustness.



Paperid:5416
Authors:ROBERT MC CARTHY, Joey SKAF, Luis Ibanez-Lissen, Vasil Georgiev, Connor Watts, Hannes Whittingham, Lorena Gonzalez-Manzano, Cameron Tice, Edward Young, Puria Radmard, David Lindner
Abstract:
Chain-of-thought (CoT) reasoning not only enhances large language model performance but also provides critical insights into decision-making processes, marking it as a useful tool for monitoring model intent and planning. By proactively preventing models from acting on CoT indicating misaligned or harmful intent, CoT monitoring can be used to reduce risks associated with deploying models. However, developers may be incentivized to train away the appearance of harmful intent from CoT traces, by either customer preferences or regulatory requirements. Recent works have shown that banning mention of a specific example of reward hacking, which may be done either to make CoT presentable to users or as a naive attempt to prevent the behavior, causes obfuscation of the undesired reasoning traces but the persistence of the undesired behavior. Such obfuscation threatens the reliability of CoT monitoring. Here, we provide complementary results. First, we show that penalizing the use of specific strings within load-bearing reasoning traces causes models to substitute alternative strings. Crucially, this does not alter the underlying method by which the model performs the task, demonstrating that the model can learn to steganographically encode its reasoning. We further demonstrate that models can generalize an encoding scheme. When the penalized strings belong to an overarching class, the model learns not only to substitute strings seen in training, but also develops a general encoding scheme for all members of the class which it can apply to held-out testing strings.



Paperid:5417
Authors:Tung Nguyen, Tuan Pham, Troy Arcomano, Rao Kotamarthi, Ian Foster, Sandeep Madireddy, Aditya Grover
Abstract:
Abstract:Accurate weather prediction on the subseasonal-to-seasonal (S2S) scale is critical for anticipating and mitigating the impacts of climate change. However, existing data-driven methods struggle beyond the medium-range timescale due to error accumulation in their autoregressive approach. In this work, we propose SeasonCast, a scalable and skillful probabilistic model for S2S prediction. SeasonCast consists of two components, a VAE model that encodes raw weather data into a continuous, lower-dimensional latent space, and a diffusion-based transformer model that generates a sequence of future latent tokens given the initial conditioning tokens. During training, we mask random future tokens and train the transformer to estimate their distribution given conditioning and visible tokens using a per-token diffusion head. During inference, the transformer generates the full sequence of future tokens by iteratively unmasking random subsets of tokens. This joint sampling across space and time mitigates compounding errors from autoregressive approaches. The low-dimensional latent space enables modeling long sequences of future latent states, allowing the transformer to learn weather dynamics beyond initial conditions. SeasonCast performs competitively with leading probabilistic methods at the medium-range timescale while being 10$\times$ to 20$\times$ faster, and achieves state-of-the-art performance at the subseasonal-to-seasonal scale across accuracy, physics-based, and probabilistic metrics.



Paperid:5418
Authors:Xingliang Wang, Zemin Liu, Junxiao Han, Shuiguang Deng
Abstract:
Graph Foundation Models (GFMs) have demonstrated remarkable potential across graph learning tasks but face significant challenges in knowledge updating and reasoning faithfulness. To address these issues, we introduce the Retrieval-Augmented Generation (RAG) paradigm for GFMs, which leverages graph knowledge retrieval. We propose RAG4GFM, an end-to-end framework that seamlessly integrates multi-level graph indexing, task-aware retrieval, and graph fusion enhancement. RAG4GFM implements a hierarchical graph indexing architecture, enabling multi-granular graph indexing while achieving efficient logarithmic-time retrieval. Furthermore, RAG4GFM implements adaptive retrieval strategies for node, edge, and graph-level tasks, ensuring effective knowledge integration through graph fusion enhancement module. Extensive experiments conducted across diverse GFM applications demonstrate that RAG4GFM significantly enhances both the efficiency of knowledge updating and reasoning faithfulness. Our work establishes a systematic foundation for integrating RAG mechanisms into GFMs, thereby advancing the development of scalable and trustworthy GFM systems.



Paperid:5419
Authors:Wenhao Ding, Yiying Sheng, Choon Hwai Yap, Hwa Leo, Simão de Castro
Abstract:
Hemodynamics has a substantial influence on normal cardiovascular growth and disease formation, but requires time-consuming simulations to obtain. Deep Learning algorithms to rapidly predict hemodynamics parameters can be very useful, but their development is hindered by the lack of large dataset on anatomic geometries and associated fluid dynamics. This paper presents a new large-scale dataset of intracranial aneurysm (IA) geometries and hemodynamics to support the development of neural operators to solve geometry-dependent flow governing partial differential equations. The dataset includes 14,000 steady-flow cases and 200 pulsatile-flow cases simulated with computational fluid dynamics. All cases are computed using a laminar flow setup with more than 3 million cells. Boundary conditions are defined as a parabolic velocity profile with a realistic waveform over time at the inlet, and geometry-dependent mass flow split ratios at the two downstream outlets. The geometries are generated by a deep generative model trained on a cohort of 109 real IAs located at the middle cerebral artery bifurcation, capturing a wide range of geometric variations in both aneurysm sacs and parent vessels. Simulation results shows substantial influence of geometry on fluid forces and flow patterns. In addition to surface mesh files, the dataset provides volume data of velocity, pressure, and wall shear stresses (WSS). For transient cases, spatial and temporal gradients of velocity and pressure are also included. The dataset is tested with PointNet and graph U-Nets for WSS prediction, which showed relative L2 loss of 4.67\% for normalized WSS pattern.



Paperid:5420
Authors:Haichao Zhang, Yun Fu
Abstract:
Token-based video representation has emerged as a promising approach for enabling large language models (LLMs) to interpret video content. However, existing token reduction techniques, such as pruning and merging, often disrupt essential positional embeddings and rely on continuous visual tokens sampled from nearby pixels with similar spatial–temporal locations. By removing only a small fraction of tokens, these methods still produce relatively lengthy continuous sequences, which falls short of the extreme compression required to balance computational efficiency and token count in video LLMs.In this paper, we introduce the novel task ofExtreme Short Token Reduction, which aims to represent entire videos using a minimal set of discrete tokens. We proposeVQToken, a neural discrete token representation framework that(i) applies adaptive vector quantization to continuous ViT embeddings to learn a compact codebook and (ii) preserves spatial–temporal positions via a token hash function by assigning each grid-level token to its nearest codebook entry.On the Extreme Short Token Reduction task, our VQToken compresses sequences to just0.07\%of their original length while incurring only a0.66\%drop in accuracy on NextQA-MC benchmark. It also achieves comparable performance on ActNet-QA, Long Video Bench, and VideoMME. We further introduce theToken Information Density(TokDense) metric and formalize fixed-length and adaptive-length subtasks, achieving state-of-the-art results in both settings. Our approach dramatically lowers theoretical complexity, increases information density, way fewer tokens counts, and enables efficient video large language models in resource-constrained environments.



Paperid:5421
Authors:Ge Wu, Shen Zhang, Ruijing Shi, Shanghua Gao, Zhenyuan Chen, Lei Wang, Zhaowei Chen, Hongcheng Gao, Yao Tang, jian Yang, Ming-Ming Cheng, Xiang Li
Abstract:
Abstract:REPA and its variants effectively mitigate training challenges in diffusion models by incorporating external visual representations from pretrained models, through alignment between the noisy hidden projections of denoising networks and foundational clean image representations. We argue that the external alignment, which is absent during the entire denoising inference process, falls short of fully harnessing the potential of discriminative representations. In this work, we propose a straightforward method called \textit{\textbf{R}epresentation \textbf{E}ntanglement for \textbf{G}eneration} (\textbf{REG}), which entangles low-level image latents with a single high-level class token from pretrained foundation models for denoising. REG acquires the capability to produce coherent image-class pairs directly from pure noise, substantially improving both generation quality and training efficiency.This is accomplished with negligible additional inference overhead, requiring only one single additional token for denoising (<0.5\% increase in FLOPs and latency).The inference process concurrently reconstructs both image latents and their corresponding global semantics, where the acquired semantic knowledge actively guides and enhances the image generation process.On ImageNet 256$\times$256, SiT-XL/2 + REG demonstrates remarkable convergence acceleration, achieving $\textbf{63}\times$ and $\textbf{23}\times$ faster training than SiT-XL/2 and SiT-XL/2 + REPA, respectively. More impressively, SiT-L/2 + REG trained for merely 400K iterations outperforms SiT-XL/2 + REPA trained for 4M iterations ($\textbf{10}\times$ longer). Code is available at: \url{https://anonymous.4open.science/r/REG-6C5B}.



Paperid:5422
Authors:Zhao Song, Jianfei Xue, Lichen Zhang
Abstract:
Abstract:In this paper, we study differentially private mechanisms for functions whose outputs lie in a Euclidean Jordan algebra. Euclidean Jordan algebras capture many important mathematical structures and form the foundation of linear programming, second-order cone programming, and semidefinite programming. Our main contribution is a generic Gaussian mechanism for such functions, with sensitivity measured in $\ell_2$, $\ell_1$, and $\ell_\infty$ norms. Notably, this framework includes the important case where the function outputs are symmetric matrices, and sensitivity is measured in the Frobenius, nuclear, or spectral norm. We further derive private algorithms for solving symmetric cone programs under various settings, using a combination of the multiplicative weights update method and our generic Gaussian mechanism. As an application, we present differentially private algorithms for semidefinite programming, resolving a major open question posed by [Hsu, Roth, Roughgarden, and Ullman, ICALP 2014].



Paperid:5423
Authors:Ya-Wei Eileen Lin, Ronald Coifman, Gal Mishne, Ronen Talmon
Abstract:
High-dimensional data often exhibit hierarchical structures in both modes: samples and features. Yet, most existing approaches for hierarchical representation learning consider only one mode at a time. In this work, we propose an unsupervised method for jointly learning hierarchical representations of samples and features via Tree-Wasserstein Distance (TWD). Our method alternates between the two data modes. It first constructs a tree for one mode, then computes a TWD for the other mode based on that tree, and finally uses the resulting TWD to build the second mode’s tree. By repeatedly alternating through these steps, the method gradually refines both trees and the corresponding TWDs, capturing meaningful hierarchical representations of the data.We provide a theoretical analysis showing that our method converges.We show that our method can be integrated into hyperbolic graph convolutional networks as a pre-processing technique, improving performance in link prediction and node classification tasks. In addition, our method outperforms baselines in sparse approximation and unsupervised Wasserstein distance learning tasks on word-document and single-cell RNA-sequencing datasets.



Paperid:5424
Authors:Lanrui Wang, Mingyu Zheng, Hongyin Tang, Zheng Lin, Yanan Cao, Jingang Wang, Xunliang Cai, Weiping Wang
Abstract:
Processing structured tabular data, particularly large and lengthy tables, constitutes a fundamental yet challenging task for large language models (LLMs). However, existing long-context benchmarks like Needle-in-a-Haystack primarily focus on unstructured text, neglecting the challenge of diverse structured tables. Meanwhile, previous tabular benchmarks mainly consider downstream tasks that require high-level reasoning abilities, and overlook models' underlying fine-grained perception of individual table cells, which is crucial for practical and robust LLM-based table applications. To address this gap, we introduce \textsc{NeedleInATable} (NIAT), a new long-context tabular benchmark that treats each table cell as a ``needle'' and requires models to extract the target cell based on cell locations or lookup questions. Our comprehensive evaluation of various LLMs and multimodal LLMs reveals a substantial performance gap between popular downstream tabular tasks and the simpler NIAT task, suggesting that they may rely on dataset-specific correlations or shortcuts to obtain better benchmark results but lack truly robust long-context understanding towards structured tables. Furthermore, we demonstrate that using synthesized NIAT training data can effectively improve performance on both NIAT task and downstream tabular tasks, which validates the importance of NIAT capability for LLMs' genuine table understanding ability. Our data, code and models will be released to facilitate future research.



Paperid:5425
Authors:Tianfang Zhu, Dongli Hu, Jiandong Zhou, Kai Du, Anan LI
Abstract:
Biological sensorimotor systems process information through spatially organized, functionally specialized modules. A canonical example is the rodent barrel cortex, in which each vibrissa (whisker) projects to a dedicated cortical column, forming a precise somatotopic map. This anatomical organization stands in stark contrast to the architectures of most artificial neural networks, which are typically monolithic or rely on globally routed mixture-of-experts (MoE) mechanisms. In this work, we introduce a brain-inspired modular architecture that treats the barrel cortex as a biologically constrained instantiation of an expert system. Each module (or “expert”) corresponds to a cortical column composed of multiple neuron subtypes spanning vertical cortical layers. Sensory signals are routed exclusively to their corresponding columns, with inter-column communication restricted to local neighbors via a sparse gating mechanism. Despite these anatomical constraints, our model achieves competitive, state-of-the-art performance on challenging 3D tactile object classification benchmarks. Columnar parameter sharing provides inherent regularization, enabling 97\% parameter reduction with improved training stability. Notably, constrained localist routing suppresses inter-module activity correlations, mirroring the barrel cortex's lateral inhibition for sensory differentiation, while suggesting MoE's potential to reduce expert redundancy through collaborative constraints. These results demonstrate how cortical principles of localist-expert routing and topographic organization can be translated into machine learning architectures, providing a step toward next-generation expert systems that bridge neuroscience and artificial intelligence. Code is provided in the supplementary materials and will be made publicly available upon acceptance.



Paperid:5426
Authors:Pengyi Li, Shixiong Kai, Jianye Hao, Ruizhe Zhong, Hongyao Tang, Zhentao Tang, Mingxuan Yuan, Junchi Yan
Abstract:
Floorplanning is the initial step in the physical design process of Electronic Design Automation (EDA), directly influencing subsequent placement, routing, and final power of the chip. However, the solution space in floorplanning is vast, and current algorithms often struggle to explore it sufficiently, making them prone to getting trapped in local optima. To achieve efficient floorplanning, we propose CORE, a general and effective solution optimization framework that synergizes Evolutionary Algorithms (EAs) and Reinforcement Learning (RL) for high-quality layout search and optimization. Specifically, we propose the Clustering-based Diversified Evolutionary Search that directly perturbs layouts and evolves them based on novelty and performance. Additionally, we model the floorplanning problem as a sequential decision problem with B*-Tree representation and employ RL for efficient learning. To efficiently coordinate EAs and RL, we propose the reinforcement-driven mechanism and evolution-guided mechanism. The former accelerates population evolution through RL, while the latter guides RL learning through EAs. The experimental results on the MCNC and GSRC benchmarks demonstrate that CORE outperform other strong baselines in terms of wirelength and area utilization metrics, achieving a 12.9% improvement in wirelength.



Paperid:5427
Authors:Zhongjian Wang, Peng Zhang, Jinwei Qi, wang yuan, Sheng Xu, Bang Zhang
Abstract:
Although significant progress has been made in audio-driven talking head generation, text-driven methods remain underexplored. In this work, we present OmniTalker, a unified framework that jointly generates synchronized talking audio-video content from input text while emulating the target identity's speaking and facial movement styles, including speech characteristics, head motion, and facial dynamics. Our framework adopts a dual-branch diffusion transformer (DiT) architecture, with one branch dedicated to audio generation and the other to video synthesis.At the shallow layers, cross-modal fusion modules are introduced to integrate information between the two modalities. In deeper layers, each modality is processed independently, with the generated audio decoded by a vocoder and the video rendered using a GAN-based high-quality visual renderer. Leveraging DiT’s in-context learning capability through a masked-infilling strategy, our model can simultaneously capture both audio and visual styles without requiring explicit style extraction modules. Thanks to the efficiency of the DiT backbone and the optimized visual renderer, OmniTalker achieves real-time inference at 25 FPS.To the best of our knowledge, OmniTalker is the first one-shot framework capable of jointly modeling speech and facial styles in real time. Extensive experiments demonstrate its superiority over existing methods in terms of generation quality, particularly in preserving style consistency and ensuring precise audio-video synchronization, all while maintaining efficient inference.



Paperid:5428
Authors:Nathaniel S. Keplinger, Baiting Luo, Yunuo Zhang, Kyle H Wray, Aron Laszka, Abhishek Dubey, Ayan Mukhopadhyay
Abstract:
Many real-world applications require decision-making where the environmental dynamics evolve over time. These non-stationary environments pose significant challenges to traditional decision-making models, which typically assume stationary dynamics. Non-stationary Markov decision processes (NS-MDPs) offer a framework to model and solve decision problems under such changing conditions. However, there are no standardized simulation frameworks for NS-MDPs, as opposed to widely popular frameworks for stationary problems. We present NS-Gym, the first simulation toolkit designed explicitly for NS-MDPs, integrated within the popular Gymnasium framework. In NS-Gym, we segregate the evolution of the environmental parameters that characterize non-stationarity from the agent’s decision-making module, allowing for modular and flexible adaptations to dynamic environments. We review prior work in this domain and present a toolkit encapsulating key problem characteristics and types in NS-MDPs. This toolkit is the first effort to develop a set of standardized interfaces and benchmark problems to enable consistent and reproducible evaluation of algorithms under non-stationary conditions. We also benchmark several algorithmic approaches from prior work on NS-MDPs using NS-Gym. We envision that NS-Gym will enable researchers to study decision-making under non-stationarity by providing standardized interfaces, simulation frameworks, and benchmark problems.



Paperid:5429
Authors:Youssef Chaabouni, David Gamarnik
Abstract:
Abstract:We consider the problem of recovering the support of a sparse signal using noisy projections. While extensive work has been done on the dense measurement matrix setting, the sparse setting remains less explored. In this work, we establish sufficient conditions on the sample size for successful sparse recovery using sparse measurement matrices. Bringing together our result with previously known necessary conditions, we discover that, in the regime where $ds/p \rightarrow +\infty$, sparse recovery in the sparse setting exhibits a phase transition at an information-theoretic threshold of $n_{\text{INF}}^{\text{SP}} = \Theta\left(s\log\left(p/s\right)/\log\left(ds/p\right)\right)$, where \(p\) denotes the signal dimension, $s$ the number of non-zero components of the signal, and $d$ the expected number of non-zero components per row of measurement. This expression makes the price of sparsity explicit: restricting each measurement to $d$ non‑zeros inflates the required sample size by a factor of $\log{s}/\log\left(ds/p\right)$, revealing a precise trade‑off between sampling complexity and measurement sparsity. Additionally, we examine the effect of sparsifying an originally dense measurement matrix on sparse signal recovery. We prove in the regime of $s = \alpha p$ and $d = \psi p$ with $\alpha, \psi \in \left(0,1\right)$ and $\psi$ small that a sample of size $n^{\text{Sp-ified}}_{\text{INF}} = \Theta\left(p / \psi^2\right)$ is sufficient for recovery, subject to a certain uniform integrability conjecture, the proof of which is work in progress.



Paperid:5430
Authors:Rongyang Zhang, Yuqing Huang, Chengqiang Lu, Qimeng Wang, Yan Gao, YIWU, Yao Hu, Yin Xu, Wei Wang, Hao Wang, Enhong Chen
Abstract:
In real-world scenarios, providing user queries with visually enhanced responses can considerably benefit understanding and memory, underscoring the great value of interleaved image-text generation. Despite recent progress, like the visual autoregressive model that unifies text and image processing in a single transformer architecture, generating high-quality interleaved content remains challenging. Moreover, evaluations of these interleaved sequences largely remain underexplored, with existing benchmarks often limited by unimodal metrics that inadequately assess the intricacies of combined image-text outputs. To address these issues, we present RAG-IGBench, a thorough benchmark designed specifically to evaluate the task of Interleaved Generation based on Retrieval-Augmented Generation (RAG-IG) in open-domain question answering. RAG-IG integrates multimodal large language models (MLLMs) with retrieval mechanisms, enabling the models to access external image-text information for generating coherent multimodal content. Distinct from previous datasets, RAG-IGBench draws on the latest publicly available content from social platforms and introduces innovative evaluation metrics that measure the quality of text and images, as well as their consistency. Through extensive experiments with state-of-the-art MLLMs (both open-source and proprietary) on RAG-IGBench, we provide an in-depth analysis examining the capabilities and limitations of these models. Additionally, we validate our evaluation metrics by demonstrating their high correlation with human assessments. Models fine-tuned on RAG-IGBench's training set exhibit improved performance across multiple benchmarks, confirming both the quality and practical utility of our dataset. Our benchmark is available at https://github.com/zry13/RAG-IGBench.



Paperid:5431
Authors:Yinghao Zhu, Ziyi He, Haoran Hu, Xiaochen Zheng, Xichen Zhang, Wang, Junyi Gao, Liantao Ma, Lequan Yu
Abstract:
The rapid advancement of Large Language Models (LLMs) has stimulated interest in multi-agent collaboration for addressing complex medical tasks. However, the practical advantages of multi-agent collaboration approaches remain insufficiently understood. Existing evaluations often lack generalizability, failing to cover diverse tasks reflective of real-world clinical practice, and frequently omit rigorous comparisons against both single-LLM-based and established conventional methods. To address this critical gap, we introduce MedAgentBoard, a comprehensive benchmark for the systematic evaluation of multi-agent collaboration, single-LLM, and conventional approaches. MedAgentBoard encompasses four diverse medical task categories: (1) medical (visual) question answering, (2) lay summary generation, (3) structured Electronic Health Record (EHR) predictive modeling, and (4) clinical workflow automation, across text, medical images, and structured EHR data. Our extensive experiments reveal a nuanced landscape: while multi-agent collaboration demonstrates benefits in specific scenarios, such as enhancing task completeness in clinical workflow automation, it does not consistently outperform advanced single LLMs (e.g., in textual medical QA) or, critically, specialized conventional methods that generally maintain better performance in tasks like medical VQA and EHR-based prediction. MedAgentBoard offers a vital resource and actionable insights, emphasizing the necessity of a task-specific, evidence-based approach to selecting and developing AI solutions in medicine. It underscores that the inherent complexity and overhead of multi-agent collaboration must be carefully weighed against tangible performance gains. All code, datasets, detailed prompts, and experimental results are open-sourced atthis link.



Paperid:5432
Authors:Sengim Karayalcin, Marina Krček, Stjepan Picek
Abstract:
Side-channel analysis (SCA) poses a real-world threat by exploiting unintentional physical signals to extract secret information from secure devices. Evaluation labs also use the same techniques to certify device security. In recent years, deep learning has emerged as a prominent method for SCA, achieving state-of-the-art attack performance at the cost of interpretability. Understanding how neural networks extract secrets is crucial for security evaluators aiming to defend against such attacks, as only by understanding the attack can one propose better countermeasures.In this work, we apply mechanistic interpretability to neural networks trained for SCA, revealing how models exploit what leakage in side-channel traces. We focus on sudden jumps in performance to reverse engineer learned representations, ultimately recovering secret masks and moving the evaluation process from black-box to white-box. Our results show that mechanistic interpretability can scale to realistic SCA settings, even when relevant inputs are sparse, model accuracies are low, and side-channel protections prevent standard input interventions



Paperid:5433
Authors:Jiakai Li, Rongzheng Wang, Yizhuo Ma, Shuang Liang, Guangchun Luo, Ke Qin
Abstract:
While large language models (LLMs) show considerable promise across various fields, they have notable limitations in handling multi-document question answering (Multi-doc QA) tasks. The first challenge is long-range dependency modeling, where LLMs struggle to focus on key information in long texts, which weakens important semantic connections. Second, most LLMs suffer from the ''lost-in-the-middle'' issue, where they have difficulty processing and connecting information in the middle of long inputs. Current solutions either truncate global dependencies or demand costly finetuning, ultimately lacking a universal and simple solution for these challenges. To resolve these limitations, we propose Dual-Stage Adaptive Sharpening (\ourmethod) containing two modules. (i) The Contextual Gate Weighting (CGW) module alleviates ''lost-in-the-middle'' by assessing paragraph relevance through layer-wise attention tracking and position-aware weighting. (ii) The Reciprocal Attention Suppression (RAS) module enhances focus on critical paragraphs by suppressing information exchange between key and irrelevant texts, thus mitigating the limitations in long-range dependency modeling. Notably, \ourmethod functions as a plug-and-play solution requiring no architectural modifications or extra training parameters. Extensive experiments on four benchmarks demonstrate DSAS's universality across mainstream LLMs (Llama, Qwen, Mistral, and Deepseek), with an average F1-score improvement of 4.2\% in Multi-doc QA tasks on Llama-3.1-8B-Instruct and Qwen2.5-14B-Instruct. Ablation studies further confirm the essential contributions of both CGW and RAS modules. Our code is available athttps://anonymous.4open.science/r/NIPS_DSAS



Paperid:5434
Authors:Qinghong Ye, Yiqian Chang, Jianing Li, Haoran Xu, Wei Zhang, Xuan Wang, Yonghong Tian, Peixi Peng
Abstract:
Spike camera with high temporal resolution offers a new perspective on high-speed dynamic scene rendering. Most existing rendering methods rely on Neural Radiance Fields (NeRF) or 3D Gaussian Splatting (3DGS) for static scenes using a monocular spike camera. However, these methods struggle with dynamic motion, while a single camera suffers from limited spatial coverage, making it challenging to reconstruct fine details in high-speed scenes. To address these problems, we propose Spike4DGS, the first high-speed dynamic scene rendering framework with 4D Gaussian Splatting using spike camera arrays. Technically, we first build a multi-view spike camera array to validate our solution, then establish both synthetic and real-world multi-view spike-based reconstruction datasets. Then, we design a multi-view spike-based dense initialization module that obtains dense point clouds and camera poses from continuous spike streams. Finally, we propose a spike-pixel synergy constraint supervision to optimize Spike4DGS, incorporating both rendered image quality loss and dynamic spatiotemporal spike loss. The results show that our Spike4DGS outperforms state-of-the-art methods in terms of novel view rendering quality on both synthetic and real-world datasets. We also verify that rendering quality improves with an increasing number of spike cameras. Our demos are available in the supplementary material.



Paperid:5435
Authors:Huaizhi Qu, Inyoung Choi, Zhen Tan, Song Wang, Sukwon Yun, Qi Long, Faizan Siddiqui, Kwonjoon Lee, Tianlong Chen
Abstract:
Abstract:LLM ensembles are widely used for LLM judges. However, how to estimate their accuracy, especially in an efficient way, is unknown. In this paper, we present a principled $\textit{maximum a posteriori}$ (MAP) framework for an economical and precise estimation of the performance of LLM ensemble judgment. We first propose a mixture of Beta-Binomial distributions to model the judgment distribution, revising from the vanilla Binomial distribution. Next, we introduce a conformal prediction-driven approach that enables adaptive stopping during iterative sampling to balance accuracy with efficiency. Furthermore, we design a prior transfer mechanism that utilizes learned distributions on open-source datasets to improve estimation on a target dataset when only scarce annotations are available. Finally, we present $\texttt{BetaConform}$, a framework that integrates our distribution assumption, adaptive stopping, and the prior transfer mechanism to deliver a theoretically guaranteed distribution estimation of LLM ensemble judgment with minimum labeled samples. $\texttt{BetaConform}$ is also validated empirically. For instance, with only $10$ samples from the TruthfulQA dataset, for a Llama ensembled judge, $\texttt{BetaConform}$ gauges its performance with an error margin as small as $3.37\\%$.



Paperid:5436
Authors:Orin Levy, Liad Erez, Alon Peled-Cohen, Yishay Mansour
Abstract:
Abstract:We present regret minimization algorithms for the contextual multi-armed bandit (CMAB) problem over $K$ actions in the presence of delayed feedback, a scenario where loss observations arrive with delays chosen by an adversary. As a preliminary result, assuming direct access to a finite policy class $\Pi$ we establish an optimal expected regret bound of $ O (\sqrt{KT \log |\Pi|} + \sqrt{D \log |\Pi|)} $ where $D$ is the sum of delays. For our main contribution, we study the general function approximation setting over a (possibly infinite) contextual loss function class $ \mathcal{F} $ with access to an online least-square regression oracle $\mathcal{O}$ over $\mathcal{F}$. In this setting, we achieve an expected regret bound of $O(\sqrt{KTR_T(\mathcal{O})} + \sqrt{ d_{\max} D \beta})$ assuming FIFO order, where $d_{\max}$ is the maximal delay, $R_T(\mathcal{O})$ is an upper bound on the oracle's regret and $\beta$ is a stability parameter associated with the oracle. We complement this general result by presenting a novel stability analysis of a Hedge-based version of Vovk's aggregating forecaster as an oracle implementation for least-square regression over a finite function class $\mathcal{F}$ and show that its stability parameter $\beta$ is bounded by $\log |\mathcal F|$, resulting in an expected regret bound of $O(\sqrt{KT \log |\mathcal{F}|} + \sqrt{d_{\max} D \log |\mathcal{F}|})$ which is a $\sqrt{d_{max}}$ factor away from the lower bound of $\Omega(\sqrt{KT \log |\mathcal{F}|} + \sqrt{D \log |\mathcal{F}|})$ that we also present.



Paperid:5437
Authors:Chun Wang, Xiaojun Ye, Xiaoran Pan, Zihao Pan, Haofan Wang, Yiren Song
Abstract:
Recent advances in Visual Language Models (VLMs) have demonstrated exceptional performance in visual reasoning tasks. However, geo-localization presents unique challenges, requiring the extraction of multigranular visual cues from images and their integration with external world knowledge for systematic reasoning. Current approaches to geo-localization tasks often lack robust reasoning mechanisms and explainability, limiting their effectiveness. To address these limitations, we propose the Geo Reason Enhancement (GRE) Suite, a novel framework that augments VLMs with structured reasoning chains for accurate and interpretable location inference. The GRE Suite is systematically developed across three key dimensions: dataset, model, and benchmark. First, we introduce GRE30K, a high-quality geo-localization reasoning dataset designed to facilitate fine-grained visual and contextual analysis. Next, we present the GRE model, which employs a multi-stage reasoning strategy to progressively infer scene attributes, local details, and semantic features, thereby narrowing down potential geographic regions with enhanced precision. Finally, we construct the Geo Reason Evaluation Benchmark (GREval-Bench), a comprehensive evaluation framework that assesses VLMs across diverse urban, natural, and landmark scenes to measure both coarse-grained (e.g., country, continent) and fine-grained (e.g., city, street) localization performance. Experimental results demonstrate that GRE significantly outperforms existing methods across all granularities of geo-localization tasks, underscoring the efficacy of reasoning-augmented VLMs in complex geographic inference. Code and data will be released at https://anonymous.4open.science/r/GRE-74C0.



Paperid:5438
Authors:Wenchang Duan, Yaoliang Yu, Jiwan He, Yi Shi
Abstract:
Recently, deep multi-agent reinforcement learning (MARL) has demonstrated promising performance for solving challenging tasks, such as long-term dependencies and non-Markovian environments. Its success is partly attributed to conditioning policies on large fixed context length. However, such large fixed context lengths may lead to limited exploration efficiency and redundant information. In this paper, we propose a novel MARL framework to obtain adaptive and effective contextual information. Specifically, we design a central agent that dynamically optimizes context length via temporal gradient analysis, enhancing exploration to facilitate convergence to global optima in MARL. Furthermore, to enhance the adaptive optimization capability of the context length, we present an efficient input representation for the central agent, which effectively filters redundant information. By leveraging a Fourier-based low-frequency truncation method, we extract global temporal trends across decentralized agents, providing an effective and efficient representation of the MARL environment. Extensive experiments demonstrate that the proposed method achieves state-of-the-art (SOTA) performance on long-term dependency tasks, including PettingZoo, MiniGrid, and Google Research Football (GRF).



Paperid:5439
Authors:Renzhao Liang, Sizhe Xu, Chenggang Xie, Jingru Chen, Feiyang Ren, Shu Yang, Takahiro Yabe
Abstract:
Time series forecasting plays a pivotal role in critical domains such as energy management and financial markets. Although deep learning-based approaches (e.g., MLP, RNN, Transformer) have achieved remarkable progress, the prevailing "long-sequence information gain hypothesis" exhibits inherent limitations. Through systematic experimentation, this study reveals a counterintuitive phenomenon: appropriately truncating historical data can paradoxically enhance prediction accuracy, indicating that existing models learn substantial redundant features (e.g., noise or irrelevant fluctuations) during training, thereby compromising effective signal extraction. Building upon information bottleneck theory, we propose an innovative solution termed Adaptive Masking Loss with Representation Consistency (AMRC), which features two core components: 1) Dynamic masking loss, which adaptively identified highly discriminative temporal segments to guide gradient descent during model training; 2) Representation consistency constraint, which stabilized the mapping relationships among inputs, labels, and predictions. Experimental results demonstrate that AMRC effectively suppresses redundant feature learning while significantly improving model performance. This work not only challenges conventional assumptions in temporal modeling but also provides novel theoretical insights and methodological breakthroughs for developing efficient and robust forecasting models.



Paperid:5440
Authors:Rudy Morel, Francesco Ramunno, Jeff Shen, Alberto Bietti, Kyunghyun Cho, Miles Cranmer, Siavash Golkar, OLEXANDR GUGNIN, Geraud Krawezik, Tanya Marwah, Michael McCabe, Lucas Meyer, Payel Mukhopadhyay, Ruben Ohana, Liam Parker, Helen Qu, François Rozet, K.D. Leka, Francois Lanusse, David Fouhey, Shirley Ho
Abstract:
Conditional diffusion models provide a natural framework for probabilistic prediction of dynamical systems and have been successfully applied to fluid dynamics and weather prediction. However, in many settings, the available information at a given time represents only a small fraction of what is needed to predict future states, either due to measurement uncertainty or because only a small fraction of the state can be observed. This is true for example in solar physics, where we can observe the Sun’s surface and atmosphere, but its evolution is driven by internal processes for which we lack direct measurements. In this paper, we tackle the probabilistic prediction of stochastic, partially observable dynamical systems, with application to solar dynamics and the evolution of active regions. We show that standard inference schemes, such as autoregressive rollouts, fail to capture long-range dependencies in the data, largely because they do not integrate past information effectively. To overcome this, we propose a multiscale inference scheme for diffusion models, tailored to physical processes. Our method generates trajectories that are temporally fine-grained near the present and coarser as we move farther away, which enables capturing long-range temporal dependencies without increasing computational cost. When this inductive bias is integrated into a diffusion model, we show that our inference scheme significantly reduces the bias of the predicted distributions and improves the rollout stability.



Paperid:5441
Authors:Derong Kong, Zhixiong Yang, Shengxi Li, Shuaifeng Zhi, Li Liu, Zhen Liu, Jingyuan Xia
Abstract:
Low-light image enhancement (LLIE) faces persistent challenges in balancing reconstruction fidelity with cross-scenario generalization. While existing methods predominantly focus on deterministic pixel-level mappings between paired low/normal-light images, they often neglect the continuous physical process of luminance transitions in real-world environments, leading to performance drop when normal-light references are unavailable. Inspired by empirical analysis of natural luminance dynamics revealing power-law distributed intensity transitions, this paper introduces Luminance-Aware Statistical Quantification (LASQ), a novel framework that reformulates LLIE as a statistical sampling process over hierarchical luminance distributions. Our LASQ re-conceptualizes luminance transition as a power-law distribution in intensity coordinate space that can be approximated by stratified power functions, therefore, replacing deterministic mappings with probabilistic sampling over continuous luminance layers. A diffusion forward process is designed to autonomously discover optimal transition paths between luminance layers, achieving unsupervised distribution emulation without normal-light references. In this way, it considerably improves the performance in practical situations, enabling more adaptable and versatile light restoration. This framework is also readily applicable to cases with normal-light references, where it achieves superior performance on domain-specific datasets alongside better generalization-ability across non-reference datasets.



Paperid:5442
Authors:Eun Chang, Zhuangqun Huang, Yiwei Liao, Sagar Bhavsar, Amogh Param, Tammy Stark, Adel Ahmadyan, Xiao Yang, Jiaqi Wang, Ahsan Abdullah, Giang Nguyen, Akil Iyer, David Hall, Elissa Li, Nicolas Scheffer, Ahmed Kirmani, Babak Damavandi, Rakesh Wanga, Anuj Kumar, Rohit Patel, Seungwhan Moon, Xin Luna Dong
Abstract:
We introduce WearVQA, the first benchmark specifically designed to evaluate the visual questionanswering (VQA) capabilities of multi-modal AI assistant on wearable devices like smart glasses. Unlikeprior benchmarks that focus on high-quality, third-person imagery, WearVQA reflects the unique chal-lenges of ego-centric interaction—where visual inputs may be occluded, poorly lit, unzoomed, or blurry,and questions are grounded in realistic wearable use cases. The benchmark comprises 2,500 carefullycurated image-question-answer triplets, spanning 7 diverse image domains including both text-centricand general scenes, 10 cognitive task types ranging from basic recognition to various forms of reasoning,and 6 common wearables-specific image quality issues. All questions are designed to be answerable usingonly the visual input and common senses. WearVQA is paired with a rigorous LLM-as-a-judge evaluationframework with 96% labeling accuracy. Open-source and proprietary multi-modal LLMs achieved a QAaccuracy as low as 24–52% on WearVQA, with substantial drops on lower-quality images and reasoning-heavy tasks. These observations position WearVQA as a comprehensive and challenging benchmark forguiding technicial advancement towards robust, real-world multi-modal wearables AI systems.



Paperid:5443
Authors:Boyu Han, Qianqian Xu, Shilong Bao, Zhiyong Yang, Kangli Zi, Qingming Huang
Abstract:
Abstract:This paper explores a novel lightweight approach LightFair to achieve fair text-to-image diffusion models (T2I DMs) by addressing the adverse effects of the text encoder. Most existing methods either couple different parts of the diffusion model for full-parameter training or rely on auxiliary networks for correction. They incur heavy training or sampling burden and unsatisfactory performance. Since T2I DMs consist of multiple components, with the text encoder being the most fine-tunable and front-end module, this paper focuses on mitigating bias by fine-tuning text embeddings. To validate feasibility, we observe that the text encoder’s neutral embedding output shows substantial skewness across image embeddings of various attributes in the CLIP space. More importantly, the noise prediction network further amplifies this imbalance. To finetune the text embedding, we propose a collaborative distance-constrained debiasing strategy that balances embedding distances to improve fairness without auxiliary references. However, mitigating bias can compromise the original generation quality. To address this, we introduce a two-stage text-guided sampling strategy to limit when the debiased text encoder intervenes. Extensive experiments demonstrate that LightFair is effective and efficient. Notably, on Stable Diffusion v1.5, our method achieves SOTA debiasing at just $1/4$ of the training burden, with virtually no increase in sampling burden.



Paperid:5444
Authors:Qingyue Zhang, Haohao Fu, Guanbo Huang, Yaoyuan Liang, Chang Chu, Tianren Peng, Yanru Wu, Qi Li, Yang Li, Shao-Lun Huang
Abstract:
Multi-source transfer learning provides an effective solution to data scarcity in real-world supervised learning scenarios by leveraging multiple source tasks. In this field, existing works typically use all available samples from sources in training, which constrains their training efficiency and may lead to suboptimal results. To address this, we propose a theoretical framework that answers the question: what is the optimal quantity of source samples needed from each source task to jointly train the target model? Specifically, we introduce a generalization error measure based on K-L divergence, and minimize it based on high-dimensional statistical analysis to determine the optimal transfer quantity for each source task. Additionally, we develop an architecture-agnostic and data-efficient algorithm OTQMS to implement our theoretical results for target model training in multi-source transfer learning. Experimental studies on diverse architectures and two real-world benchmark datasets show that our proposed algorithm significantly outperforms state-of-the-art approaches in both accuracy and data efficiency. The code is available at https://anonymous.4open.science/r/Materials.



Paperid:5445
Authors:Hao Yuan, Wenli Ouyang, Changwen Zhang, Congrui Li, Yong Sun
Abstract:
Abstract:Foundation Models (FMs) have demonstrated remarkable success in fields like computer vision and natural language processing, yet their application to combinatorial optimization remains underexplored. Optimization problems, often modeled as graphs, pose unique challenges due to their diverse structures, varying distributions, and NP-hard complexity. To address these challenges, we propose OPTFM, the first graph foundation model for general combinatorial optimization. OPTFM introduces a scalable multi-view graph transformer with hybrid self-attention and cross-attention to model large-scale heterogeneous graphs in $O(N)$ time complexity while maintaining semantic consistency throughout the attention computation. A Dual-level pre-training framework integrates node-level graph reconstruction and instance-level contrastive learning, enabling robust and adaptable representations at multiple levels. Experimental results across diverse optimization tasks show that models trained on OPTFM embeddings without fine-tuning consistently outperform task-specific approaches, establishing a new benchmark for solving combinatorial optimization problems.



Paperid:5446
Authors:Songhao Wu, Ang Lv, xiao feng, Yufei zhang, Xun Zhang, Guojun Yin, Wei Lin, Rui Yan
Abstract:
The increasing demand for long-context generation has made the KV cache in large language models a bottleneck in memory consumption. Quantizing the cache to lower bit widths is an effective way to reduce memory costs; however, previous methods struggle with key cache quantization due to outliers, resulting in suboptimal performance. We propose a novel quantization approach PolarQuant, which provides a new perspective for key cache quantization and efficiently addresses the outlier dilemma. We observe that the distribution of the key states reveals well-structured patterns under polar transformation. Outliers generally appear in only one of the two dimensions, which are rotated together by a specific angle when rotary position embeddings are applied. When represented as two-dimensional vectors, these dimensions exhibit well-organized patterns, with radii and angles smoothly distributed in polar space. This alleviates the channel-wise outliers, making them well-suited for key cache quantization. PolarQuant divides key vectors into groups of two-dimensional sub-vectors, encoding them as the quantized radius and the polar angle, rather than quantizing original key vectors directly. PolarQuant achieves the superior efficiency in KV cache quantization and accelerates the decoding process by turning the query-key inner product into a table lookup, all while maintaining the downstream performance of full-precision models.



Paperid:5447
Authors:Christian Fruhwirth-Reisinger, Dušan Malić, Wei Lin, David Schinagl, Samuel Schulter, Horst Possegger
Abstract:
We introduce STSBench, a scenario-based framework to benchmark the holistic understanding of vision-language models (VLMs) for autonomous driving. The framework automatically mines pre-defined traffic scenarios from any dataset using ground-truth annotations, provides an intuitive user interface for efficient human verification, and generates multiple-choice questions for model evaluation. Applied to the NuScenes dataset, we present STSnu, the first benchmark that evaluates the spatio-temporal reasoning capabilities of VLMs based on comprehensive 3D perception. Existing benchmarks typically target off-the-shelf or fine-tuned VLMs for images or videos from a single viewpoint and focus on semantic tasks such as object recognition, dense captioning, risk assessment, or scene understanding. In contrast, STSnu evaluates driving expert VLMs for end-to-end driving, operating on videos from multi-view cameras or LiDAR. It specifically assesses their ability to reason about both ego-vehicle actions and complex interactions among traffic participants, a crucial capability for autonomous vehicles. The benchmark features 43 diverse scenarios spanning multiple views and frames, resulting in 971 human-verified multiple-choice questions. A thorough evaluation uncovers critical shortcomings in existing models’ ability to reason about fundamental traffic dynamics in complex environments. These findings highlight the urgent need for architectural advances that explicitly model spatio-temporal reasoning. By addressing a core gap in spatio-temporal evaluation, STSBench enables the development of more robust and explainable VLMs for autonomous driving.



Paperid:5448
Authors:RUIYAO MIAO, Junren Xiao, Shiya Tsang, Hui Xiong, Ying Nian Wu
Abstract:
Existing Bayesian Optimization (BO) methods typically balance exploration and exploitation to optimize costly objective functions. However, these methods often suffer from a significant one-step bias, which may lead to convergence toward local optima and poor performance in complex or high-dimensional tasks. Recently, Black-Box Optimization (BBO) has achieved success across various scientific and engineering domains, particularly when function evaluations are costly and gradients are unavailable. Motivated by this, we propose the Reinforced Energy-Based Model for Bayesian Optimization (REBMBO), which integrates Gaussian Processes (GP) for local guidance with an Energy-Based Model (EBM) to capture global structural information. Notably, we define each Bayesian Optimization iteration as a Markov Decision Process (MDP) and use Proximal Policy Optimization (PPO) for adaptive multi-step lookahead, dynamically adjusting the depth and direction of exploration to effectively overcome the limitations of traditional BO methods. We conduct extensive experiments on synthetic and real-world benchmarks, confirming the superior performance of REBMBO. Additional analyses across various GP configurations further highlight its adaptability and robustness.



Paperid:5449
Authors:Yunfan Li, Eric Liu, Lin Yang
Abstract:
Reinforcement learning (RL) often relies on manually designed reward functions, which are difficult to specify and can lead to issues such as reward hacking and suboptimal behavior. Alternatives like inverse RL and preference-based RL attempt to infer surrogate rewards from demonstrations or preferences but suffer from ambiguity and distribution mismatch. A more direct approach, inspired by imitation learning, avoids reward modeling by leveraging expert demonstrations. However, most existing methods align actions only at individual states, failing to capture the coherence of long-horizon trajectories.In this work, we study the problem of directly aligning policies with expert-labeled trajectories to preserve long-horizon behavior without relying on reward signals. Specifically, we aim to learn a policy that maximizes the probability of generating the expert trajectories. Nevertheless, we prove that, in its general form, this trajectory alignment problem is NP-complete. To address this, we propose Trajectory Graph Learning (TGL), a framework that leverages structural assumptions commonly satisfied in practice—such as bounded realizability of expert trajectories or a tree-structured MDP. These enable a graph-based policy planning algorithm that computes optimal policies in polynomial time under known dynamics. For settings with unknown dynamics, we develop a sample-efficient algorithm based on UCB-style exploration and establish sub-linear regret. Experiments on grid-world tasks demonstrate that TGL substantially outperforms standard imitation learning methods for long-trajectory planning.



Paperid:5450
Authors:Feng Chen, Allan Raventós, Nan Cheng, Surya Ganguli, Shaul Druckmann
Abstract:
Recent progress in large language models (LLMs) highlights the power of scaling test-time compute to achieve strong performance on complex tasks, such as mathematical reasoning and code generation. This raises a critical question: how should model training be modified to optimize performance under a subsequent test-time compute strategy and budget? To explore this, we focus on pass@N, a simple test-time strategy that searches for a correct answer in N independent samples. We show, surprisingly, that training with cross-entropy (CE) can bemisalignedwith pass@N in that pass@N accuracydecreaseswith longer CE training. We explain the origins of this misalignment in terms of model overconfidence induced by CE, and experimentally verify our prediction of overconfidence as an impediment to scaling test-time compute via pass@N. Furthermore we suggest a principled, modified training loss that is better aligned to pass@N by limiting model confidence and rescuing pass@N test performance. Our algorithm demonstrates improved mathematical reasoning on MATH and MiniF2F benchmarks under several scenarios: (1) providing answers to math questions both with and without Chain-of-Thought reasoning traces; and (2) proving theorems by searching over proof trees of varying shapes. Overall our work underscores the importance of co-designing two traditionally separate phases of LLM development: training-time protocols and test-time search and reasoning strategies.



Paperid:5451
Authors:Ruyi An, haicheng huang, Huangjie Zheng, Mingyuan Zhou
Abstract:
Abstract:Generative data augmentation (GDA) leverages generative models to enrich training sets with entirely new samples drawn from the modeled data distribution to achieve performance gains. However, the usage of the mighty contemporary diffusion models in GDA remains impractical: *i)* their thousand-step sampling loop inflates wall-time and energy cost per image augmentation; and *ii)* the divergence between synthetic and real distributions is unknown--classifiers trained on synthetic receive biased gradients.We propose DAR-GDA, a three-stage augmentation pipeline that unites model **D**istillation, **A**dversarial alignment, and importance **R**eweighting that makes diffusion-quality augmentation both fast and optimized for improving downstream learning outcomes. In particular, a teacher diffusion model is compressed into a one-step student via score distillation, slashing the time per-image cost by $>100\times$ while preserving FID.During this distillation (D), the student model additionally undergoes adversarial alignment (A) by receiving direct training signals against real images, supplementing the teacher's guidance to better match the true data distribution.The discriminator from this adversarial process inherently learns to assess the synthetic-to-real data gap. Its calibrated probabilistic outputs are then employed in reweighting (R) by importance weights that quantify the distributional gap and adjust the classification loss when training downstream models; we show that reweighting yields an unbiased stochastic estimator of the real-data risk, fostering training dynamics akin to those of genuine samples.Experiments validate DAR-GDA's synergistic design through progressive accuracy gains with each D-A-R stage. Our approach not only surpasses conventional non-foundation-model GDA baselines but also remarkably matches or exceeds the GDA performance of large, web-pretrained text-to-image models, despite using solely in-domain data.DAR-GDA thus offers diffusion-fidelity GDA samples efficiently, while correcting synthetic-to-real bias to benefit downstream tasks.



Paperid:5452
Authors:Chongkai Gao, Zixuan Liu, Zhenghao Chi, Junshan Huang, Xin Fei, Yiwen Hou, Yuxuan Zhang, Yudi Lin, Zhirui Fang, Lin Shao
Abstract:
Recent studies on Vision-Language-Action (VLA) models have shifted from the end-to-end action-generation paradigm toward a pipeline involving task planning followed by action generation, demonstrating improved performance on various complex, long-horizon manipulation tasks. However, existing approaches vary significantly in terms of network architectures, planning paradigms, representations, and training data sources, making it challenging for researchers to identify the precise sources of performance gains and determine which component is more difficult to learn. To systematically investigate the impacts of different planning paradigms and representations isolating from network architectures and training data, in this paper, we introduce \name, a unified VLA architecture suite capable of various task planning paradigms, and design a comprehensive suite of controlled experiments across diverse object categories (rigid and deformable), visual modalities (2D and 3D), environments (simulation and real-world), and end-effectors (grippers and dexterous hands). Our results demonstrate that: 1) visually grounded planning representations are generally better than language planning representations; 2) the Hierarchical-VLA paradigm generally achieves superior performance than other paradigms, albeit at the cost of slower training and inference speeds.



Paperid:5453
Authors:Jinhee Kim, Jae Jun An, Kang Jeon, Jong Hwan Ko
Abstract:
Multi-bit quantization networks enable flexible deployment of deep neural networks by supporting multiple precision levels within a single model. However, existing approaches suffer from significant training overhead as full-dataset updates are repeated for each supported bit-width, resulting in a cost that scales linearly with the number of precisions. Additionally, extra fine-tuning stages are often required to support additional or intermediate precision options, further compounding the overall training burden. To address this issue, we propose two techniques that greatly reduce the training overhead without compromising model utility: (i) Weight bias correction enables shared batch normalization and eliminates the need for fine-tuning by neutralizing quantization-induced bias across bit-widths and aligning activation distributions; and (ii) Bit-wise coreset sampling strategy allows each child model to train on a compact, informative subset selected via gradient-based importance scores by exploiting the implicit knowledge transfer phenomenon. Experiments on CIFAR-10/100, TinyImageNet, and ImageNet-1K with both ResNet and ViT architectures demonstrate that our method achieves competitive or superior accuracy while reducing training time up to 7.88×.



Paperid:5454
Authors:Niclas Popp, Kevin Alexander Laube, Matthias Hein, Lukas Schott
Abstract:
Large foundation models trained on extensive datasets demonstrate strong zero-shot capabilities in various domains. To replicate their success when data and model size are constrained, knowledge distillation has become an established tool for transferring knowledge from foundation models to small student networks. However, the effectiveness of distillation is critically limited by the available training data. This work addresses the common practical issue of covariate shift in knowledge distillation, where spurious features appear during training but not at test time. We ask the question: when these spurious features are unknown, yet a robust teacher is available, is it possible for a student to also become robust to them? We address this problem by introducing a novel diffusion-based data augmentation strategy that generates images by maximizing the disagreement between the teacher and the student, effectively creating challenging samples that the student struggles with. Experiments demonstrate that our approach significantly improves worst group and mean group accuracy on CelebA and SpuCo Birds as well as the spurious mAUC on spurious ImageNet under covariate shift, outperforming state-of-the-art diffusion-based data augmentation baselines.



Paperid:5455
Authors:Xijun Li, Jiexiang Yang, Jinghao Wang, Bo Peng, Jianguo Yao, Haibing Guan
Abstract:
Abstract:Combinatorial optimization (CO) problems, central to operation research and theoretical computer science, present significant computational challenges due to their $\mathcal{NP}$-hard nature. While large language models (LLMs) have emerged as promising tools for CO—either by directly generating solutions or synthesizing solver-specific codes—existing approaches often $\textit{neglect critical structural priors inherent to CO problems}$, leading to suboptimality and iterative inefficiency. Inspired by human experts’ success in leveraging CO structures for algorithm design, we propose $\texttt{STRCMP}$, a novel structure-aware LLM-based algorithm discovery framework that systematically integrates structure priors to enhance solution quality and solving efficiency. Our framework combines a graph neural network (GNN) for extracting structural embeddings from CO instances with an LLM conditioned on these embeddings to identify high-performed algorithms in the form of solver-specific codes. This composite architecture ensures syntactic correctness, preserves problem topology, and aligns with natural language objectives, while an evolutionary refinement process iteratively optimizes generated algorithm. Extensive evaluations across Mixed Integer Linear Programming and Boolean Satisfiability problems, using nine benchmark datasets, demonstrate that our proposed $\texttt{STRCMP}$ outperforms five strong neural and LLM-based methods by a large margin, in terms of both solution optimality and computational efficiency. The code and learned model will be publicly available upon the acceptance of the paper.



Paperid:5456
Authors:Valia Efthymiou, Chara Podimata, Diptangshu Sen, Juba Ziani
Abstract:
We study strategic classification in binary decision-making settings where agents can modify their features in order to improve their classification outcomes. Importantly, our work considers the causal structure across different features, acknowledging that effort in one feature may affect other features. The main goal of our work is to understandwhen and how much agent effort is invested towards desirable features, and how this is influenced by the deployed classifier, the causal structure of the agent's features, their ability to modify them, and the information available to the agent about the classifier and the feature causal graph. We characterize conditions under which agents with full information about the causal structure and the classifier respond in a way that aligns with the principal's goals of incentivizing effort mostly in desirable features, and identify cases where designing such classifiers (from the principal's side) is still tractable despite general non-convexity. Under incomplete information (about either the causal graph or the principal's classifier), we show that uncertainty leads agents to prioritize features with high expected impact and low variance, which may often be misaligned with the principal's goals. Finally, numerical experiments based on a cardiovascular disease risk study illustrate how to incentivize desirable modifications even under uncertainty.



Paperid:5457
Authors:Yifei Xia, Shuchen Weng, Siqi Yang, Jingqi Liu, Chengxuan Zhu, Minggui Teng, Zijian Jia, Han Jiang, Boxin Shi
Abstract:
Abstract:Panoramic video generation enables immersive 360$^\circ$ content creation, valuable in applications that demand scene-consistent world exploration. However, existing panoramic video generation models struggle to leverage pre-trained generative priors from conventional text-to-video models for high-quality and diverse panoramic videos generation, due to limited dataset scale and the gap in spatial feature representations. In this paper, we introduce PanoWan to effectively lift pre-trained text-to-video models to the panoramic domain, equipped with minimal modules. PanoWan employs latitude-aware sampling to avoid latitudinal distortion, while its rotated semantic denoising and padded pixel-wise decoding ensure seamless transitions at longitude boundaries. To provide sufficient panoramic videos for learning these lifted representations, we contribute PanoVid, a high-quality panoramic video dataset with captions and diverse scenarios. Consequently, PanoWan achieves state-of-the-art performance in panoramic video generation and demonstrates robustness for zero-shot downstream tasks.



Paperid:5458
Authors:Tianxiao He, Malhar Patel, Chenyi Li, Anna Maslarova, Mihály Vöröslakos, Nalini Ramanathan, Wei-Lun Hung, Gyorgy Buzsaki, Erdem Varol
Abstract:
Recent advances in large-scale neural recordings have enabled accurate decoding of behavior and cognitive states, yet decoding anatomical regions remains underexplored, despite being crucial for consistent targeting in multiday recordings and effective deep brain stimulation. Current approaches typically rely on external anatomical information, from atlas-based planning to post hoc histology, which are limited in precision, longitudinal applicability, and real-time feedback. In this work, we develop a self-supervised learning framework, termed \textbf{Lfp2vec}, to infer anatomical regions from the neural signal in vivo. Instead of relying on handcrafted features and extensive preprocessing, we adapt a transformer-based model from the audio domain to embed the Local Field Potential (LFP) signals of individual channels and fine-tune it for downstream decoding of hippocampal sublayers. We demonstrate that our method generalizes across subjects, probe geometries, and lab conditions, achieving strong zero-shot performance and outperforming state-of-the-art self-supervised models on electrophysiology data. The learned embeddings cluster by brain region and generalize to different downstream tasks with minimal fine-tuning. Altogether, our approach enables zero-shot prediction of brain regions in novel subjects, demonstrates that LFP signals encode rich anatomical information, and establishes self-supervised learning on raw LFP as a foundation to learn representations that can be tuned for diverse neural decoding tasks. Code to reproduce our results are found in the anonymous repository at https://anonymous.4open.science/r/lfp2vec-AF87.



Paperid:5459
Authors:Chenhang Cui, Gelei Deng, An Zhang, jingnan zheng, Yicong Li, Lianli Gao, Tianwei Zhang, Tat-Seng Chua
Abstract:
Recent advances in Large Vision-Language Models (LVLMs) have showcased strong reasoning abilities across multiple modalities, achieving significant breakthroughs in various real-world applications.Despite this great success, the safety guardrail of LVLMs may not cover the unforeseen domains introduced by the visual modality. Existing studies primarily focus on eliciting LVLMs to generate harmful responses via carefully crafted image-based jailbreaks designed to bypass alignment defenses.In this study, we reveal that a safe image can be exploited to achieve the same jailbreak consequence when combined with additional safe images and prompts.This stems from two fundamental properties of LVLMs: universal reasoning capabilities and safety snowball effect.Building on these insights, we propose Safety Snowball Agent (SSA), a novel agent-based framework leveraging agents' autonomous and tool-using abilities to jailbreak LVLMs.SSA operates through two principal stages: (1) initial response generation, where tools generate or retrieve jailbreak images based on potential harmful intents, and (2) harmful snowballing, where refined subsequent prompts induce progressively harmful outputs.Our experiments demonstrate that SSA can use nearly any image to induce LVLMs to produce unsafe content, achieving high success jailbreaking rates against the latest LVLMs. Unlike prior works that exploit alignment flaws, SSA leverages the inherent properties of LVLMs, presenting a profound challenge for enforcing safety in generative multimodal systems.



Paperid:5460
Authors:Shuai Yuan, Guowen Xu, Hongwei Li, Xingshuo Han, Wenbo Jiang, Tao Ni, Qingchuan Zhao, Yuguang Fang
Abstract:
Recently, traffic sign recognition (TSR) systems have become a prominent target for physical adversarial attacks. These attacks typically rely on conspicuous stickers and projections, or using invisible light and acoustic signals that can be easily blocked. In this paper, we introduce a novel attack medium, i.e., fluorescent ink, to design a stealthy and effective physical adversarial patch, namely FIPatch, to advance the state-of-the-art. Specifically, we first model the fluorescence effect in the digital domain to identify the optimal attack settings, which guide the real-world fluorescence parameters. By applying a carefully designed fluorescence perturbation to the target sign, the attacker can later trigger a fluorescent effect using invisible ultraviolet light, causing the TSR system to misclassify the sign and potentially leading to traffic accidents. We conducted a comprehensive evaluation to investigate the effectiveness of FIPatch, which shows a success rate of 98.31% in low-light conditions. Furthermore, our attack successfully bypasses five popular defenses and achieves a success rate of 96.72%.



Paperid:5461
Authors:Qingqiu Li, Zihang Cui, Seongsu Bae, Jilan Xu, Runtian Yuan, Yuejie Zhang, Rui Feng, Quanli Shen, Xiaobo Zhang, Shang Gao, Junjun He, Shujun Wang
Abstract:
Chest X-rays (CXRs) are the most frequently performed imaging examinations in clinical settings. Recent advancements in Medical Large Multimodal Models (MLMMs) have enabled automated CXR interpretation, improving diagnostic accuracy and efficiency. However, despite their strong visual understanding, current MLMMs still face two major challenges: (1) Insufficient region-level understanding and interaction, and (2) Limited accuracy and interpretability due to single-step prediction. In this paper, we address these challenges by empowering MLMMs with anatomy-centric reasoning capabilities to enhance their interactivity and explainability. Specifically, we propose an Anatomical Ontology-Guided Reasoning (AOR) framework that accommodates both textual and optional visual prompts, centered on region-level information to enable multimodal multi-step reasoning. We also develop AOR-Instruction, a large instruction dataset for MLMMs training, under the guidance of expert physicians. Our experiments demonstrate AOR's superior performance in both Visual Question Answering (VQA) and report generation tasks. Code and data are available at: https://anonymous.4open.science/r/AOR-48C7/.



Paperid:5462
Authors:Yize Chen, Zhiyuan Yan, Guangliang Cheng, Kangran Zhao, Siwei Lyu, Baoyuan Wu
Abstract:
Abstract:This paper proposes **$\mathcal{X}^2$-DFD**, an **e$\mathcal{X}$plainable** and **e$\mathcal{X}$tendable** framework based on multimodal large-language models (MLLMs) for deepfake detection, consisting of three key stages. The first stage, *Model Feature Assessment*, systematically evaluates the detectability of forgery-related features for the MLLM, generating a prioritized ranking of features based on their intrinsic importance to the model.The second stage, *Explainable Dataset Construction*, consists of two key modules: *Strong Feature Strengthening*, which is designed to enhance the model’s existing detection and explanation capabilities by reinforcing its well-learned features, and *Weak Feature Supplementing*, which addresses gaps by integrating specific feature detectors (e.g., low-level artifact analyzers) to compensate for the MLLM’s limitations.The third stage, Fine-tuning and Inference, involves fine-tuning the MLLM on the constructed dataset and deploying it for final detection and explanation.By integrating these three stages, our approach enhances the MLLM's strengths while supplementing its weaknesses, ultimately improving both the detectability and explainability.Extensive experiments and ablations, followed by a comprehensive human study, validate the improved performance of our approach compared to the original MLLMs.More encouragingly, our framework is designed to be plug-and-play, allowing it to seamlessly integrate with future more advanced MLLMs and specific feature detectors, leading to continual improvement and extension to face the challenges of rapidly evolving deepfakes.



Paperid:5463
Authors:Linlin Fan, Xingyu Liu, Mingliang Zhou, Xuekai Wei, Weizhi Xian, Jielu Yan, Weijia Jia
Abstract:
Effective modelling of uncertain information is crucial for quantifying uncertainty. Dempster–Shafer evidence (DSE) theory is a widely recognized approach for handling uncertain information. However, current methods often neglect the inherent a priori information within data during modelling, and imbalanced data lead to insufficient attention to key information in the model. To address these limitations, this paper presents a fusion framework based on a nonuniform splitting mechanism and Hilbert space mapping. First, the framework uses a nonuniform splitting mechanism to dynamically adjust the weights of data subsets and combines the diffusion factor to effectively incorporate the data a priori information, thereby flexibly addressing uncertainty and conflict. Second, the conflict in the information fusion process is reduced by Hilbert space mapping. Experimental results on multiple tasks show that the proposed method significantly outperforms state-of-the-art methods and effectively improves the performance of classification and low-light image enhancement (LLIE) tasks. The code is available at https://anonymous.4open.science/r/Third-ED16.



Paperid:5464
Authors:Ankan Deria, Adinath Dukre, feilong tang, Sara Atito, Sudipta Roy, Muhammad Awais, Muhammad Haris Khan, Imran Razzak
Abstract:
Abstract:Despite significant advances in inference-time search for vision–language models (VLMs), existing approaches remain both computationally expensive and prone to unpenalized, low-confidence generations which often lead to persistent hallucinations. We introduce Value-guided Inference with Margin-based Reward (ViMaR), a two-stage inference framework that improves both efficiency and output fidelity by combining a temporal-difference value model with a margin-aware reward adjustment. In the first stage, we perform a single pass to identify the highest-value caption among diverse candidates. In the second stage, we selectively refine only those segments that were overlooked or exhibit weak visual grounding, thereby eliminating frequently rewarded evaluations. A calibrated margin-based penalty discourages low-confidence continuations while preserving descriptive richness. Extensive experiments across multiple VLM architectures demonstrate that ViMaR generates captions that are significantly more reliable, factually accurate, detailed, and explanatory, while achieving over 4$\times$ speedup compared to existing value-guided methods. Specifically, we show that ViMaR trained solely on LLaVA Mistral-7B, generalizes effectively to guide decoding in a stronger unseen model. To further validate this, we adapt the ViMaR to steer generation in LLaVA-OneVision-Qwen2-7B, leading to consistent improvements in caption quality and demonstrating robust cross-model guidance. This cross-model generalization highlights ViMaR's flexibility and modularity, positioning it as a scalable and transferable inference-time decoding strategy. Furthermore, when ViMaR-generated captions are used for self-training, the underlying models achieve substantial gains across a broad suite of visual comprehension benchmarks, underscoring the potential of fast, accurate, and self-improving VLM pipelines.



Paperid:5465
Authors:Chuyu Zhou, Tianyu Li, Chenxi Lan, Rongyu Du, Guoguo Xin, Wei Li, Guoqing Wang, Xun Liu, Hangzhou Yang
Abstract:
Physics-informed neural networks (PINN) have achieved notable success in solving partial differential equations (PDE), yet solving the Navier-Stokes equations (NSE) with complex boundary conditions remains a challenging task. In this paper, we introduce a novel Hybrid Boundary PINN (HB-PINN) method that combines a pretrained network for efficient initialization with a boundary-constrained mechanism. The HB-PINN method features a primary network focused on inner domain points and a distance metric network that enhances predictions at the boundaries, ensuring accurate solutions for both boundary and interior regions. Comprehensive experiments have been conducted on the NSE under complex boundary conditions, including the 2D cylinder wake flow and the 2D blocked cavity flow with a segmented inlet. The proposed method achieves state-of-the-art (SOTA) performance on these benchmark scenarios, demonstrating significantly improved accuracy over existing PINN-based approaches.



Paperid:5466
Authors:Yuxing Lu, Gecheng Fu, Wei Wu, Xukai Zhao, Sin Yee Goi, Jinzhuo Wang
Abstract:
Existing medical RAG systems mainly leverage knowledge from medical knowledge bases, neglecting the crucial role of experiential knowledge derived from similar patient cases - a key component of human clinical reasoning. To bridge this gap, we propose DoctorRAG, a RAG framework that emulates doctor-like reasoning by integrating both explicit clinical knowledge and implicit case-based experience. DoctorRAG enhances retrieval precision by first allocating conceptual tags for queries and knowledge sources, together with a hybrid retrieval mechanism from both relevant knowledge and patient. In addition, a Med-TextGrad module using multi-agent textual gradients is integrated to ensure that the final output adheres to the retrieved knowledge and patient query. Comprehensive experiments on multilingual, multitask datasets demonstrate that DoctorRAG significantly outperforms strong baseline RAG models and gains improvements from iterative refinements. Our approach generates more accurate, relevant, and comprehensive responses, taking a step towards more doctor-like medical reasoning systems.



Paperid:5467
Authors:Weikuo Wang, Yue Liao, Huan Luo
Abstract:
Abstract:A major problem of kernel-based methods (e.g., least squares support vector machines, LS-SVMs) for solving linear/nonlinear ordinary differential equations (ODEs) is the prohibitive $O(n^3)$ part of their computational complexity with increasing temporal discretization points $n$. We propose a novel Nyström-accelerated LS-SVMs framework that breaks this bottleneck by reformulating ODEs as primal-space constraints. Specifically, we derive for the first time an explicit Nyström-based mapping and its derivatives from one-dimensional temporal discretization points to a higher $m$-dimensional feature space ($1< m\le n$), enabling the learning process to solve linear/nonlinear equation systems with $m$-dependent complexity. Numerical experiments on fifteen benchmark ODEs demonstrate: 1) 10-3000 times faster computation than classical LS-SVMs and physics-informed neural networks (PINNs), 2) comparable accuracy to LS-SVMs ($<0.13$\% relative MAE, RMSE, and $ \| y-\hat{y} \| _{\infty } $difference) while maximum surpassing PINNs by 72\% in RMSE, and 3) scalability to $n=10^4$ time steps with only $m$=50 features. This work establishes a new paradigm for efficient kernel-based ODEs learning without significantly sacrificing the accuracy of the solution.



Paperid:5468
Authors:Torben Berndt, Benjamin Walker, Tiexin Qin, Jan Stühmer, Andrey Kormilitzin
Abstract:
Dynamic graphs exhibit complex temporal dynamics due to the interplay between evolving node features and changing network structures. Recently, Graph Neural Controlled Differential Equations (Graph Neural CDEs) successfully adapted Neural CDEs from paths on Euclidean domains to paths on graph domains. Building on this foundation, we introduce \textit{Permutation Equivariant Graph Neural CDEs}, which project Graph Neural CDEs onto permutation equivariant function spaces. This significantly reduces the model's parameter count without compromising representational power, resulting in more efficient training and improved generalisation. We empirically demonstrate the advantages of our approach through experiments on simulated dynamical systems and real-world tasks, showing improved performance in both interpolation and extrapolation scenarios.



Paperid:5469
Authors:Grzegorz Gluch, Berkant Turan, Sai Ganesh Nagarajan, Sebastian Pokutta
Abstract:
We formalize and analyze the trade-off between backdoor-based watermarks and adversarial defenses, framing it as an interactive protocol between a verifier and a prover. While previous works have primarily focused on this trade-off, our analysis extends it by identifying transferable attacks as a third, counterintuitive but necessary option. Our main result shows that for all learning tasks, at least one of the three exists: awatermark, anadversarial defense, or atransferable attack. By transferable attack, we refer to an efficient algorithm that generates queries indistinguishable from the data distribution and capable of foolingallefficient defenders. Using cryptographic techniques, specifically fully homomorphic encryption, we construct a transferable attack and prove its necessity in this trade-off. Finally, we show that tasks of bounded VC-dimension allow adversarial defenses against all attackers, while a subclass allows watermarks secure against fast adversaries.



Paperid:5470
Authors:Jiange Wang, Jinxuan Wu, Dongdong Zhang
Abstract:
Effective multiscale medical image segmentation requires simultaneously preserving smooth spatial continuity and accurately delineating high-frequency boundaries, yet pixel-wise decoders often fail to maintain this balance consistently across varying resolutions. We introduce GauSAM, which seamlessly integrates contour‑guided 2D Gaussian probability fields into the Segment Anything Model to address these challenges. In our framework, segmentation masks are parameterized as continuous probability fields of learnable 2D Gaussian primitives, enforcing spatially smooth and structurally consistent. Contourlet transforms extract rich multidirectional frequency information, notably edges and fine textures, which dynamically guide the spatial distribution of Gaussian primitives to substantially improve boundary fidelity in complex structures. The incorporation of these high-frequency contour priors also enriches the expressive capacity of the SAM image encoder. Extensive experiments on diverse 2D medical segmentation tasks confirm that GauSAM consistently delivers robust generalization and state-of-the-art performance with only 1.2M trainable parameters.



Paperid:5471
Authors:Siwei Zhang, Yun Xiong, Yateng Tang, Jiarong Xu, Xi Chen, Zehao Gu, Xuehao Zheng, Jiawei Zhang, Zi'an Jia
Abstract:
Abstract:Temporal graph neural networks (TGNNs) have shown remarkable performance in temporal graph modeling. However, real-world temporal graphs often possess rich textual information, giving rise to temporal text-attributed graphs (TTAGs). Such combination of dynamic text semantics and evolving graph structures introduces heightened complexity. Existing TGNNs embed texts statically and rely heavily on encoding mechanisms that biasedly prioritize structural information, overlooking the temporal evolution of text semantics and the essential interplay between semantics and structures for synergistic reinforcement.To tackle these issues, we present $\textbf{CROSS}$, a flexible framework that seamlessly extends existing TGNNs for TTAG modeling. CROSS is designed by decomposing the TTAG modeling process into two phases: (i) temporal semantics extraction; and (ii) semantic-structural information unification. The key idea is to advance the large language models (LLMs) to $\textit{dynamically}$ extract the temporal semantics in text space and then generate $\textit{cohesive}$ representations unifying both semantics and structures.Specifically, we propose a Temporal Semantics Extractor in the CROSS framework, which empowers LLMs to offer the temporal semantic understanding of node's evolving contexts of textual neighborhoods, facilitating semantic dynamics.Subsequently, we introduce the Semantic-structural Co-encoder, which collaborates with the above Extractor for synthesizing illuminating representations by jointly considering both semantic and structural information while encouraging their mutual reinforcement. Extensive experiments show that CROSS achieves state-of-the-art results on four public datasets and one industrial dataset, with 24.7\% absolute MRR gain on average in temporal link prediction and 3.7\% AUC gain in node classification of industrial application.



Paperid:5472
Authors:Vahan Arsenyan, Elen Vardanyan, Arnak Dalalyan
Abstract:
Generative modeling aims to produce new random examples from an unknown target distribution, given access to a finite collection of examples. Among the leading approaches, denoising diffusion probabilistic models (DDPMs) construct such examples by mapping a Brownian motion via a diffusion process driven by an estimated score function. In this work, we first provide empirical evidence that DDPMs are robust to constant-variance noise in the score evaluations. We then establish finite-sample guarantees in Wasserstein-2 distance that exhibit two key features: (i) they characterize and quantify the robustness of DDPMs to noisy score estimates, and (ii) they achieve faster convergence rates than previously known results. Furthermore, we observe that the obtained rates match those known in the Gaussian case, implying their optimality.



Paperid:5473
Authors:Seungyoo Lee, Giung Nam, Hyungi Lee, Moonseok Choi, Juho Lee
Abstract:
Modern large language models (LLMs) achieve competitive performance across a wide range of natural language processing tasks through zero-shot or few-shot prompting. However, domain-specific tasks often still require fine-tuning, which is frequently hindered by data scarcity, i.e., collecting sufficient domain-specific data remains a practical challenge. A widely adopted solution is to generate synthetic data using LLMs by augmenting a small set of available domain-specific examples. In this work, we first identify fundamental limitations of such approach in terms of both data diversity and quality, particularly when relying on only a handful of domain-specific examples. We then propose our method, PANGEA, which leverages large-scale, publicly available general-purpose data---entirely unrelated to the target domain---to generate more diverse and higher-quality synthetic data. Our extensive experiments on domain-specific benchmarks, including GSM8K, MedQA, and FinQA, as well as a custom domain-specific language task, validate the effectiveness of our approach.



Paperid:5474
Authors:Aadirupa Saha, Robert Schapire
Abstract:
Abstract:We study the contextual dueling bandit problem, where a learner uses contextual information to make two decisions and receives only relative (comparison) feedback. This problem is crucial in reinforcement learning with human feedback (RLHF), widely applied in AI alignment to integrate human preferences into AI models. Unlike prior works, we consider general preference relations and propose the first efficient, near-optimal regret algorithm using an offline regression oracle. Existing approaches rely on online oracles, which are often impractical for complex function classes, leading to poor performance in challenging settings. Our key contribution is analyzing the contextual Best Response regret and developing an $\tilde O(K \sqrt T)$ regret algorithm, explicitly incorporating offline oracle performance. We further extend our results to continuous decision spaces, achieving a regret bound of $\tilde O(\sqrt{dT})$, marking the first such result for contextual dueling bandits with offline oracles in infinite arm spaces. Our work resolves an open problem in efficient contextual dueling bandits, with potential applications in AI alignment and LLM fine-tuning. Theoretical findings are supported by empirical results.



Paperid:5475
Authors:Hao Yin, Guangzong Si, Zilei Wang
Abstract:
Contrastive decoding strategies are widely used to reduce object hallucinations in multimodal large language models (MLLMs). These methods work by constructing contrastive samples to induce hallucinations and then suppressing them in the output distribution. However, this paper demonstrates that such approaches fail to effectively mitigate the hallucination problem. The performance improvements observed on POPE Benchmark are largely driven by two misleading factors: (1) crude, unidirectional adjustments to the model’s output distribution and (2) the adaptive plausibility constraint, which reduces the sampling strategy to greedy search. To further illustrate these issues, we introduce a series of spurious improvement methods and evaluate their performance against contrastive decoding techniques. Experimental results reveal that the observed performance gains in contrastive decoding are entirely unrelated to its intended goal of mitigating hallucinations. Our findings challenge common assumptions about the effectiveness of contrastive decoding strategies and pave the way for developing genuinely effective solutions to hallucinations in MLLMs.



Paperid:5476
Authors:Zhiyuan Ning, Jiawei Shao, Ruge Xu, Xinfei Guo, Jun Zhang, Chi Zhang, Xuelong Li
Abstract:
Abstract:Speculative decoding has become a widely adopted as an effective technique for lossless inference acceleration when deploying large language models (LLMs). While on-the-fly self-speculative methods offer seamless integration and broad utility, they often fall short of the speed gains achieved by methods relying on specialized training. Cascading a hierarchy of draft models promises further acceleration and flexibility, but the high cost of training multiple models has limited its practical application. In this paper, we propose a novel Cascade Adaptive Self-Speculative Decoding (CAS-Spec) algorithm which constructs speculative draft models by leveraging dynamically switchable inference acceleration (DSIA) strategies, including layer sparsity and activation quantization. We further introduce a Dynamic Tree Cascade (DyTC) method that adaptively routes the multi-level draft models and assigns the draft lengths, based on the heuristics of acceptance rates and hardware-aware latency prediction. Our CAS-Spec algorithm achieves state-of-the-art acceleration ($1.6\times$ to $2.1\times$ speedup) compared to existing on-the-fly speculative decoding methods on both edge and server platforms. CAS-Spec can be easily integrated into most existing LLMs and holds promising potential for further acceleration as self-speculative decoding techniques continue to evolve.



Paperid:5477
Authors:Zeyu Zhang, Yiran Wang, Danning Li, Dong Gong, Ian Reid, Richard Hartley
Abstract:
Diffusion models have recently advanced 3D human motion generation by producing smoother and more realistic sequences from natural language. However, existing approaches face two major challenges: high computational cost during training and inference, and limited scalability due to reliance on U-Net inductive bias. To address these challenges, we propose FlashMo, a frequency-aware sparse motion diffusion model that prunes low-frequency tokens to enhance efficiency without custom kernel design. We further introduce MotionSiT, a scalable diffusion transformer based on a joint-temporal factorized interpolant with Lie group geodesics over quaternion manifolds, enabling principled generation of joint rotations. Extensive experiments on the large-scale MotionHub V2 \cite{ling2024motionllama} dataset and standard benchmarks including HumanML3D and KIT-ML demonstrate that our method significantly outperforms previous approaches in motion quality, efficiency, and scalability. Compared to the state-of-the-art 1-step distillation baseline, FlashMo reduces 12.9% inference time and FID by 34.1%.



Paperid:5478
Authors:Hao Cheng, Erjia Xiao, Jing Shao, Yichi Wang, Le Yang, Chao Shen, Philip Torr, Jindong Gu, Renjing Xu
Abstract:
Large Language Models (LLMs) demonstrate impressive zero-shot performance across a wide range of natural language processing tasks. Integrating various modality encoders further expands their capabilities, giving rise to Multimodal Large Language Models (MLLMs) that process not only text but also visual and auditory modality inputs. However, these advanced capabilities may also pose significant security risks, as models can be exploited to generate harmful or inappropriate content through jailbreak attacks. While prior work has extensively explored how manipulating textual or visual modality inputs can circumvent safeguards in LLMs and MLLMs, the vulnerability of audio-specific Jailbreak on Large Audio-Language Models (LALMs) remains largely underexplored. To address this gap, we introduce \textbf{Jailbreak-AudioBench}, which consists of the Toolbox, curated Dataset, and comprehensive Benchmark. The Toolbox supports not only text-to-audio conversion but also a range of audio editing techniques. The curated Dataset provides diverse explicit and implicit jailbreak audio examples in both original and edited forms. Utilizing this dataset, we evaluate multiple state-of-the-art LALMs, establishing the most comprehensive audio jailbreak benchmark to date. Finally, Jailbreak-AudioBench establishes a foundation for advancing future research on LALMs safety alignment by enabling the in-depth exposure of more powerful jailbreak threats, such as query-based audio editing, and by facilitating the development of effective defense mechanisms.



Paperid:5479
Authors:Yichen Yan, Ming Zhong, Qi Zhu, Xiaoling Gu, Jinpeng Chen, Huan Li
Abstract:
Multimodal large language models (MLLMs) rely heavily on instruction tuning to align vision and language capabilities, yet the computational cost of training on large-scale datasets remains a major bottleneck. Existing data selection methods aim to mitigate this by selecting important and diverse subsets, but they often suffer from two critical drawbacks: high computational overhead from processing the entire dataset and suboptimal data selection due to separate treatment of importance and diversity.We introduce CoIDO, a novel dual-objective framework that jointly optimizes data importance and diversity to overcome these challenges. Unlike existing approaches that require costly evaluations across the whole dataset, CoIDO employs a lightweight plug-in scorer. This scorer is trained on just a small random sample of data to learn the distribution of the candidate set, drastically reducing computational demands. By leveraging a homoscedastic uncertainty-based formulation, CoIDO effectively balances importance and diversity during training, enabling the scorer to assign CoIDO scores to all data points. This unified scoring approach allows for direct ranking and selection of the most valuable subsets — completely bypassing the need for specialized algorithms.In our experiments, we trained the CoIDO Scorer using only 20% of randomly sampled data. Once trained, CoIDO was applied to the entire dataset to select a 20% subset for instruction tuning. On the widely-used LLaVA-1.5-7B model across ten downstream tasks, this selected subset achieved an impressive 98.2% of the performance of full-data fine-tuning, on average. Moreover, CoIDO outperforms all competitors in terms of both efficiency (lowest training FLOPs) and aggregated accuracy. Our code is available at: https://anonymous.4open.science/r/CoIDO



Paperid:5480
Authors:Yiran Yang, Rui Chen
Abstract:
Abstract:In this paper, we consider a score-based Integer Programming (IP) approach for solving the Bayesian Network Structure Learning (BNSL) problem. State-of-the-art BNSL IP formulations suffer from the exponentially large number of variables and constraints. A standard approach in IP to address such challenges is to employ row and column generation techniques, which dynamically generate rows and columns, while the complex pricing problem remains a computational bottleneck for BNSL. For the general class of $\ell_0$-penalized likelihood scores, we show how the pricing problem can be reformulated as a difference of submodular optimization problem, and how the Difference of Convex Algorithm (DCA) can be applied as an inexact method to efficiently solve the pricing problems. Empirically, we show that, for continuous Gaussian data, our row and column generation approach yields solutions with higher quality than state-of-the-art score-based approaches, especially when the graph density increases, and achieves comparable performance against benchmark constraint-based and hybrid approaches, even when the graph size increases.



Paperid:5481
Authors:Ahmet Gökmen, Yiğit Ekin, Bahri Batuhan Bilecen, Aysegul Dundar
Abstract:
We propose RoPECraft, a training-free video motion transfer method for diffusion transformers that operates solely by modifying their rotary positional embeddings (RoPE). We first extract dense optical flow from a reference video, and utilize the resulting motion offsets to warp the complex-exponential tensors of RoPE, effectively encoding motion into the generation process. These embeddings are then further optimized during denoising time steps via trajectory alignment between the predicted and target velocities using a flow-matching objective. To keep the output faithful to the text prompt and prevent duplicate generations, we incorporate a regularization term based on the phase components of the reference video’s Fourier transform, projecting the phase angles onto a smooth manifold to suppress high-frequency artifacts. Experiments on benchmarks reveal that RoPECraft outperforms all recently published methods, both qualitatively and quantitatively. Code will be released.



Paperid:5482
Authors:Yujia Chen, Rui Sun, Wangkai Li, Huayu Mai, Si Chen, Zhuoyuan Li, Zhixin Cheng, Tianzhu Zhang
Abstract:
Domain adaptation for LiDAR semantic segmentation remains challenging due to the complex structural properties of point cloud data. While mix-based paradigms have shown promise, they often fail to fully leverage the rich structural priors inherent in 3D LiDAR point clouds. In this paper, we identify three critical yet underexploited structural priors: permutation invariance, local consistency, and geometric consistency. We introduce BeyondMix, a novel framework that harnesses the capabilities of State Space Models (specifically Mamba) to construct and exploit these structural priors while modeling long-range dependencies that transcend the limited receptive fields of conventional voxel-based approaches. By employing space-filling curves to impose sequential ordering on point cloud data and implementing strategic spatial partitioning schemes, BeyondMix effectively captures domain-invariant representations. Extensive experiments on challenging LiDAR semantic segmentation benchmarks demonstrate that our approach consistently outperforms existing state-of-the-art methods, establishing a new paradigm for unsupervised domain adaptation in 3D point cloud understanding.



Paperid:5483
Authors:Hsi-Ling Chen, Chun-Shien Lu, Pau-Choo Chung
Abstract:
The effectiveness of domain translation in addressing image-based problems of Unsupervised Domain Adaptation (UDA) depends on the quality of the translated images and the preservation of crucial discriminative features. However, achieving high-quality and stable translations typically requires paired data, which poses a challenge in scenarios with limited annotations in the target domain. To address this issue, this paper proposes a novel method termed Stain-Guided Cycle Diffusion (SGCD), employing a dual diffusion model with bidirectional generative constraints to synthesize highly realistic data for downstream task fine-tuning. The bidirectional generative constraints ensure that the translated images retain the features critical to the downstream model in properly controlling the generation process. Additionally, a stain-guided consistency loss is introduced to enhance the denoising capability of the dual diffusion model, thereby improving the quality of images translated between different domains using latents from one domain and a diffusion model trained on another. Experiments conducted on four public datasets demonstrate that SGCD can effectively enhance the performance of downstream task models on the target domain.



Paperid:5484
Authors:Zhenjie Yang, Xiaosong Jia, Qifeng Li, Xue Yang, Maoqing Yao, Junchi Yan
Abstract:
Reinforcement Learning (RL) can mitigate the causal confusion and distribution shift inherent to imitation learning (IL). However, applying RL to end-to-end autonomous driving (E2E-AD) remains an open problem for its training difficulty, and IL is still the mainstream paradigm in both academia and industry. Recently Model-based Reinforcement Learning (MBRL) have demonstrated promising results in neural planning; however, these methods typically require privileged information as input rather than raw sensor data. We fill this gap by designing Raw2Drive, a dual-stream MBRL approach. Initially, we efficiently train an auxiliary privileged world model paired with a neural planner that uses privileged information as input. Subsequently, we introduce a raw sensor world model trained via our proposed Guidance Mechanism, which ensures consistency between the raw sensor world model and the privileged world model during rollouts. Finally, the raw sensor world model combines the prior knowledge embedded in the heads of the privileged world model to effectively guide the training of the raw sensor policy. Raw2Drive is so far the only RL based end-to-end method on CARLA Leaderboard 2.0, and Bench2Drive and it achieves state-of-the-art performance.



Paperid:5485
Authors:Yuzhou Cao, Han Bao, Lei Feng, Bo An
Abstract:
Abstract:Surrogate regret bounds, also known as excess risk bounds, bridge the gap between the convergence rates of surrogate and target losses, with linear bounds favorable for their lossless regret transfer. While convex smooth surrogate losses are appealing in particular due to the efficient estimation and optimization, the existence of a trade-off between the smoothness and linear regret bound has been believed in the community. That being said, the better optimization and estimation properties of convex smooth surrogate losses may inevitably deteriorate after undergoing the regret transfer onto a target loss. Strikingly, we overcome this dilemma for arbitrary discrete target losses by constructing a convex smooth surrogate loss, which entails a linear surrogate regret bound composed with a tailored prediction link. The construction is based on Fenchel--Young losses generated by the $\textit{convolutional negentropy}$,which is relevant to the infimal convolution of a generalized negentropy and the target Bayes risk. Consequently, the infimal convolution enables us to derive a smooth loss while maintaining the regret bound linear. We additionally benefit from the infimal convolution to have a consistent estimator of the linear property, which serves as prediction confidence. Our results are overall a novel demonstration of how convex analysis penetrates into optimization and statistical efficiency in risk minimization.



Paperid:5486
Authors:Zhange Zhang, Zhicheng Geng, Yuqing Ma, Tianbo Wang, Kai Lv, Xianglong Liu
Abstract:
Large Language Models (LLMs) have demonstrated broad applications but suffer from issues like hallucinations, erroneous outputs and outdated knowledge. Model editing emerges as an effective solution to refine knowledge in LLMs, yet existing methods typically depend on structured knowledge representations. However, real-world knowledge is primarily embedded within complex, unstructured text. Existing structured knowledge editing approaches face significant challenges when handling the entangled and intricate knowledge present in unstructured text, resulting in issues such as representation ambiguity and editing conflicts. To address these challenges, we propose a Conflict-Aware Knowledge Editing in the Wild (CAKE) framework, the first framework explicitly designed for editing knowledge extracted from wild unstructured text. CAKE comprises two core components: a Semantic-augmented Graph Representation module and a Conflict-aware Knowledge Editing strategy. The Semantic-augmented Graph Representation module enhances knowledge encoding through structural disambiguation, relational enrichment, and semantic diversification. Meanwhile, the Conflict-aware Knowledge Editing strategy utilizes a graph-theoretic coloring algorithm to disentangle conflicted edits by allocating them to orthogonal parameter subspaces, thereby effectively mitigating editing conflicts. Experimental results on the AKEW benchmark demonstrate that CAKE significantly outperforms existing methods, achieving a 16.61\% improvement in accuracy on llama3 editing tasks. Our framework successfully bridges the gap between unstructured textual knowledge and reliable model editing, enabling more robust and scalable updates for practical LLM applications.



Paperid:5487
Authors:Jinwei Hu, Zhenglin Huang, Xiangyu Yin, Wenjie Ruan, Guangliang Cheng, Yi Dong, Xiaowei Huang
Abstract:
Abstract:Large language models have been widely applied, but can inadvertently encode sensitive or harmful information, raising significant safety concerns. Machine unlearning has emerged to alleviate this concern; however, existing training-time unlearning approaches, relying on coarse-grained loss combinations, have limitations in precisely separating knowledge and balancing removal effectiveness with model utility. In contrast, we propose $\textbf{F}$ine-grained $\textbf{A}$ctivation manipu$\textbf{L}$ation by $\textbf{C}$ontrastive $\textbf{O}$rthogonal u$\textbf{N}$alignment (FALCON), a novel representation-guided unlearning approach that leverages information-theoretic guidance for efficient parameter selection, employs contrastive mechanisms to enhance representation separation, and projects conflict gradients onto orthogonal subspaces to resolve conflicts between forgetting and retention objectives. Extensive experiments demonstrate that FALCON achieves superior unlearning effectiveness while maintaining model utility, exhibiting robust resistance against knowledge recovery attempts.



Paperid:5488
Authors:Emmeran Johnson, David Martinez-Rubio, Ciara Pike-Burke, Patrick Rebeschini
Abstract:
Abstract:We study online convex optimization on $\ell_p$-balls in $\mathbb{R}^d$ for $p > 2$. While always sub-linear, the optimal regret exhibits a shift between the high-dimensional setting ($d > T$), when the dimension $d$ is greater than the time horizon $T$ and the low-dimensional setting ($d \leq T$). We show that Follow-the-Regularised-Leader (FTRL) with time-varying regularisation which is adaptive to the dimension regime is anytime optimal for all dimension regimes. Motivated by this, we ask whether it is possible to obtain anytime optimality of FTRL with fixed non-adaptive regularisation. Our main result establishes that for separable regularisers, adaptivity in the regulariser is necessary, and that any fixed regulariser will be sub-optimal in one of the two dimension regimes. Finally, we provide lower bounds which rule out sub-linear regret bounds for the linear bandit problem in sufficiently high-dimension for all $\ell_p$-balls with $p \geq 1$.



Paperid:5489
Authors:Liviu Aolaritei, Oliver Wang, Qianyu Zhu, Michael Jordan, Youssef Marzouk
Abstract:
Conformal prediction provides a powerful framework for constructing prediction intervals with finite-sample guarantees, yet its robustness under distribution shifts remains a significant challenge. This paper addresses this limitation by modeling distribution shifts using Lévy-Prokhorov (LP) ambiguity sets, which capture both local and global perturbations. We provide a self-contained overview of LP ambiguity sets and their connections to popular metrics such as Wasserstein and Total Variation. We show that the link between conformal prediction and LP ambiguity sets is a natural one: by propagating the LP ambiguity set through the scoring function, we reduce complex high-dimensional distribution shifts to manageable one-dimensional distribution shifts, enabling exact quantification of worst-case quantiles and coverage. Building on this analysis, we construct robust conformal prediction intervals that remain valid under distribution shifts, explicitly linking LP parameters to interval width and confidence levels. Experimental results on real-world datasets demonstrate the effectiveness of the proposed approach.



Paperid:5490
Authors:Tong Yang, Bo Dai, Lin Xiao, Yuejie Chi
Abstract:
Online reinforcement learning (RL) with complex function approximations such as transformers and deep neural networks plays a significant role in the modern practice of artificial intelligence. Despite its popularity and importance, balancing the fundamental trade-off between exploration and exploitation remains a long-standing challenge; in particular, we are still in lack of efficient and practical schemes that are backed by theoretical performance guarantees. Motivated by recent developments in exploration via optimistic regularization, this paper provides an interpretation of the principle of optimism through the lens of primal-dual optimization. From this fresh perspective, we set forth a new value-incentivized actor-critic (VAC) method, which optimizes a single easy-to-optimize objective integrating exploration and exploitation --- it promotes state-action and policy estimates that are both consistent with collected data transitions and result in higher value functions. Theoretically, the proposed VAC method has near-optimal regret guarantees under linear Markov decision processes (MDPs) in both finite-horizon and infinite-horizon settings.



Paperid:5491
Authors:Deliang Wei, Peng Chen, Haobo Xu, Jiale Yao, Fang Li, Tieyong Zeng
Abstract:
Plug-and-play (PnP) methods with deep denoisers have shown impressive results in imaging problems. They typically require strong convexity or smoothness of the fidelity term and a (residual) non-expansive denoiser for convergence. These assumptions, however, are violated in Poisson inverse problems, and non-expansiveness can hinder denoising performance. To address these challenges, we propose a cocoercive conservative (CoCo) denoiser, which may be (residual) expansive, leading to improved denoising performance. By leveraging the generalized Helmholtz decomposition, we introduce a novel training strategy that combines Hamiltonian regularization to promote conservativeness and spectral regularization to ensure cocoerciveness. We prove that CoCo denoiser is a proximal operator of a weakly convex function, enabling a restoration model with an implicit weakly convex prior. The global convergence of PnP methods to a stationary point of this restoration model is established. Extensive experimental results demonstrate that our approach outperforms closely related methods in both visual quality and quantitative metrics.



Paperid:5492
Authors:Nanxu Gong, Zijun Li, Sixun Dong, Haoyue Bai, Wangyang Ying, Xinyuan Wang, Yanjie Fu
Abstract:
Feature Transformation (FT) crafts new features from original ones via mathematical operations to enhance dataset expressiveness for downstream models. However, existing FT methods exhibit critical limitations: discrete search struggles with enormous combinatorial spaces, impeding practical use; and continuous search, being highly sensitive to initialization and step sizes, often becomes trapped in local optima, restricting global exploration. To overcome these limitations, DIFFT redefines FT as a reward-guided generative task. It first learns a compact and expressive latent space for feature sets using a Variational Auto-Encoder (VAE). A Latent Diffusion Model (LDM) then navigates this space to generate high-quality feature embeddings, its trajectory guided by a performance evaluator towards task-specific optima. This synthesis of global distribution learning (from LDM) and targeted optimization (reward guidance) produces potent embeddings, which a novel semi-autoregressive decoder efficiently converts into structured, discrete features, preserving intra-feature dependencies while allowing parallel inter-feature generation. Extensive experiments on 14 benchmark datasets show DIFFT consistently outperforms state-of-the-art baselines in predictive accuracy and robustness, with significantly lower training and inference times.



Paperid:5493
Authors:Zhe Zhao, HaiBin Wen, Xianfu Liu, Rui Mao, Pengkun Wang, Liheng Yu, Linjiang Chen, Bo An, Qingfu Zhang, Yang Wang
Abstract:
Abstract:Scientific discovery across diverse fields increasingly grapples with datasets exhibiting pathological long-tailed distributions: a few common phenomena overshadow a multitude of rare yet scientifically critical instances. Unlike standard benchmarks, these scientific datasets often feature extreme imbalance coupled with a modest number of classes and limited overall sample volume, rendering existing long-tailed recognition (LTR) techniques ineffective. Such methods, biased by majority classes or prone to overfitting on scarce tail data, frequently fail to identify the very instances—novel materials, rare disease biomarkers, faint astronomical signals—that drive scientific breakthroughs. This paper introduces a novel, end-to-end framework explicitly designed to address pathological long-tailed recognition in scientific contexts. Our approach synergizes a Balanced Supervised Contrastive Learning (B-SCL) mechanism, which enhances the representation of tail classes by dynamically re-weighting their contributions, with a Smooth Objective Regularization (SOR) strategy that manages the inherent tension between tail-class focus and overall classification performance. We introduce and analyze the real-world ZincFluor chemical dataset ($\mathcal{T}=137.54$) and synthetic benchmarks with controllable extreme imbalances (CIFAR-LT variants). Extensive evaluations demonstrate our method's superior ability to decipher these extremes. Notably, on ZincFluor, our approach achieves a Tail Top-2 accuracy of $66.84\%$, significantly outperforming existing techniques. On CIFAR-10-LT with an imbalance ratio of $1000$ ($\mathcal{T}=100$), our method achieves a tail-class accuracy of $38.99\%$, substantially leading the next best. These results underscore our framework's potential to unlock novel insights from complex, imbalanced scientific datasets, thereby accelerating discovery.



Paperid:5494
Authors:Ameya Daigavane, Bodhi Vani, Darcy Davidson, Saeed Saremi, Joshua Rackers, Joseph Kleinhenz
Abstract:
Conformational ensembles of protein structures are immensely important both to understanding protein function, and for drug discovery in novel modalities such as cryptic pockets. Current techniques for sampling ensembles such as molecular dynamics are computationally inefficient. On the other hand, many recent machine-learning methods do not generalize well outside their training data. We propose JAMUN, that performs MD in a smoothed, noised space of all-atom 3D conformations of molecules by utilizing the framework of walk-jump sampling. JAMUN enables ensemble generation at orders of magnitude faster rates than traditional molecular dynamics or state-of-the-art ML methods. This physical prior enables JAMUN to transfer to systems outside training data. JAMUN is even able to generalize across length scales it was not trained on.



Paperid:5495
Authors:Ziteng Wang, Siqi Yang, Limeng Qiao, Lin Ma
Abstract:
Despite the success of Vision-Language Models (VLMs) like CLIP in aligning vision and language, their proficiency in detailed, fine-grained visual comprehension remains a key challenge. We present CLIP-IN, a novel framework that bolsters CLIP's fine-grained perception through two core innovations. Firstly, we leverage instruction-editing datasets, originally designed for image manipulation, as a unique source of hard negative image-text pairs. Coupled with a symmetric hard negative contrastive loss, this enables the model to effectively distinguish subtle visual-semantic differences. Secondly, CLIP-IN incorporates long descriptive captions, utilizing rotary positional encodings to capture rich semantic context often missed by standard CLIP. Our experiments demonstrate that CLIP-IN achieves substantial gains on the MMVP benchmark and various fine-grained visual recognition tasks, without compromising robust zero-shot performance on broader classification and retrieval tasks. Critically, integrating CLIP-IN's visual representations into Multimodal Large Language Models significantly reduces visual hallucinations and enhances reasoning abilities. This work underscores the considerable potential of synergizing targeted, instruction-based contrastive learning with comprehensive descriptive information to elevate the fine-grained understanding of VLMs.



Paperid:5496
Authors:Lipei Zhang, Rui Sun, Zhongying Deng, Yanqi Cheng, Carola-Bibiane Schönlieb, Angelica Aviles-Rivero
Abstract:
Variations in Magnetic resonance imaging (MRI) scanners and acquisition protocols cause distribution shifts that degrade reconstruction performance on unseen data. Test-time adaptation (TTA) offers a promising solution to address this discrepancies. However, previous single-shot TTA approaches are inefficient due to repeated training and suboptimal distributional models. Self-supervised learning methods may risk over-smoothing in scarce data scenarios. To address these challenges, we propose a novel Dual-Stage Distribution and Slice Adaptation (D2SA) via MRI implicit neural representation (MR-INR) to improve MRI reconstruction performance and efficiency, which features two stages. In the first stage, an MR-INR branch performs patient-wise distribution adaptation by learning shared representations across slices and modelling patient-specific shifts with mean and variance adjustments. In the second stage, single-slice adaptation refines the output from frozen convolutional layers with a learnable anisotropic diffusion module, preventing over-smoothing and reducing computation. Experiments across five MRI distribution shifts demonstrate that our method can integrate well with various self-supervised learning (SSL) framework, improving performance and accelerating convergence under diverse conditions.



Paperid:5497
Authors:Haotian Gao, Zheng Dong, Jiawei Yong, Shintaro Fukushima, Kenjiro Taura, Renhe Jiang
Abstract:
Spatio-temporal forecasting is essential for real-world applications such as traffic management and urban computing. Although recent methods have shown improved accuracy, they often fail to account for dynamic deviations between current inputs and historical patterns. These deviations contain critical signals that can significantly affect model performance. To fill this gap, We propose ST-SSDL, a spatio-temporal forecasting framework that incorporates a Self-Supervised Deviation Learning approach to capture and utilize such deviations. ST-SSDL anchors each input to its historical average and discretizes the latent space using learnable prototypes that represent typical spatio-temporal patterns. Two auxiliary objectives are proposed to refine this structure: a contrastive loss that enhances inter-prototype discriminability and a deviation loss that regularize the distance consistency between input representations and corresponding prototypes to quantify deviation. Optimized jointly with the forecasting objective, these components guide the model to organize its hidden space and improve generalization across diverse input conditions. Experiments on six benchmark datasets show that ST-SSDL consistently outperforms state-of-the-art baselines across multiple metrics. Visualizations further demonstrate its ability to adaptively respond to varying levels of deviation in complex spatio-temporal scenarios. Our code and datasets are available at https://anonymous.4open.science/r/ST-SSDL.



Paperid:5498
Authors:Junlang Huang, Chen Hao, Li Luo, Yong Cai, Lexin Zhang, Tianhao Ma, Yitian Zhang, Zhong Guan
Abstract:
Abstract:Simulation of high-order nonlinear system requires extensive computational resources, especially in modern VLSI backend design where bifurcation-induced instability and chaos-like transient behaviors pose challenges. We present S-Crescendo - a nested transformer weaving framework that synergizes S-domain with neural operators for scalable time-domain prediction in high-order nonlinear networks, alleviating the computational bottlenecks of conventional solvers via Newton-Raphson method. By leveraging the partial-fraction decomposition of an n-th order transfer function into first-order modal terms with repeated poles and residues, our method bypasses the conventional Jacobian matrix-based iterations and efficiently reduces computational complexity from cubic $O(n^3)$ to linear $O(n)$.The proposed architecture seamlessly integrates an S-domain encoder with an attention-based correction operator to simultaneously isolate dominant response and adaptively capture higher-order non-linearities. Validated on order-1 to order-10 networks, our method achieves up to 0.99 test-set \(R^2\) accuracy against HSPICE golden waveforms and accelerates simulation by up to 18\(\times\), providing a scalable, physics-aware framework for high-dimensional nonlinear modeling.



Paperid:5499
Authors:Yan Zhuang, Jiawei Ren, Xiaokang Ye, Jianzhi Shen, Ruixuan Zhang, Tianai Yue, Muhammad Faayez, Xuhong He, Xiyan Zhang, Ziqiao Ma, Lianhui Qin, Zhiting Hu, Tianmin Shu
Abstract:
Recent advances in foundation models have shown promising results in developing generalist robotics that can perform diverse tasks in open-ended scenarios given multimodal inputs. However, current work has been mainly focused on indoor, household scenarios. In this work, we present RoboScape, a simulation platform for embodied AI in large-scale, photorealistic urban environments. Built on Unreal Engine 5, RoboScape procedurally generates unlimited photorealistic urban scenes populated with dynamic elements such as pedestrians and traffic systems, surpassing prior urban simulations in realism, complexity, and scalability. It also supports multi-robot control and communication. With these key features, we build two challenging robot benchmarks: (1) a multimodal instruction-following task, where a robot must follow vision-language navigation instructions to reach a destination in the presence of pedestrians and traffic; and (2) a multi-agent search task, where two robots must communicate to cooperatively locate and meet each other. Unlike existing benchmarks, these two new benchmarks comprehensively evaluate a wide range of critical robot capacities in realistic scenarios, including (1) multimodal instructions grounding, (2) 3D spatial reasoning in large environments, (3) safe, long-range navigation with people and traffic, (4) multi-robot collaboration, and (5) grounded communication. Our experimental results demonstrate that state-of-the-art models, including vision-language models (VLMs), struggle with our tasks, lacking robust perception, reasoning, and planning abilities necessary for urban environments.



Paperid:5500
Authors:Yisen Gao, Xingcheng Fu, Qingyun Sun, Jianxin Li, Xianxian LI
Abstract:
Graph diffusion models have made significant progress in learning structured graph data and have demonstrated strong potential for predictive tasks. Existing approaches typically embed node, edge, and graph-level features into a unified latent space, modeling prediction tasks including classification and regression as a form of conditional generation. However, due to the non-Euclidean nature of graph data, features of different curvatures are entangled in the same latent space without releasing their geometric potential. To address this issue, we aim to construt an ideal Riemannian diffusion model to capture distinct manifold signatures of complex graph data and learn their distribution. This goal faces two challenges: numerical instability caused by exponential mapping during the encoding proces and manifold deviation during diffusion generation. To address these challenges, we proposeGeoMancer: a novel Riemannian graph diffusion framework for both generation and prediction tasks. To mitigate numerical instability, we replace exponential mapping with an isometric-invariant Riemannian gyrokernel approach and decouple multi-level features onto their respective task-specific manifolds to learn optimal representations. To address manifold deviation, we introduce a manifold-constrained diffusion method and a self-guided strategy for unconditional generation, ensuring that the generated data remains aligned with the manifold signature. Extensive experiments validate the effectiveness of our approach, demonstrating superior performance across a variety of tasks.



Paperid:5501
Authors:Ming Hsiu Wu, Ziqian Xie, Shuiwang Ji, Degui Zhi
Abstract:
Advancements in AI for science unlocks capabilities for critical drug discovery tasks such as protein-ligand binding affinity prediction. However, current models overfit to existing oversimplified datasets that does not represent naturally occurring and biologically relevant proteins with modifications. In this work, we curate a complete, modification-aware version of the widely used DAVIS dataset by incorporating 4,032 kinase–ligand pairs involving substitutions, insertions, deletions, and phosphorylation events. This enriched dataset enables benchmarking of predictive models under biologically realistic conditions. Based on this new dataset, we propose three benchmark settings—Augmented Dataset Prediction, Wild-Type to Modification Generalization, and Few-Shot Modification Generalization—designed to assess model robustness in the presence of protein modifications. Through extensive evaluation of both docking-free and docking-based methods, we find that docking-based model generalize better in zero-shot settings. In contrast, docking-free models tend to overfit to wild-type proteins and struggle with unseen modifications but show notable improvement when fine-tuned on a small set of modified examples. We anticipate that the curated dataset and benchmarks offer a valuable foundation for developing models that better generalize to protein modifications, ultimately advancing precision medicine in drug discovery.



Paperid:5502
Authors:Hyunseung Lim, Sooyohn Nam, Sungmin Na, Ji Yong Cho, June Yong Yang, Hyungyu Shin, Yoonjoo Lee, Juho Kim, Moontae Lee, Hwajung Hong
Abstract:
Patent examination remains an ongoing challenge in the NLP literature even after the advent of large language models (LLMs), as it requires an extensive yet nuanced human judgment on whether a submitted claim meets the statutory standards of novelty and non-obviousness against previously granted claims---prior art---in expert domains. Previous NLP studies have approached this challenge as a prediction task (e.g., forecasting grant outcomes) with high-level proxies such as similarity metrics or classifiers trained on historical labels. However, this approach often overlooks the step-by-step evaluations that examiners must make with profound information, including rationales for the decisions provided in office actions documents, which also makes it harder to measure the current state of techniques in patent review processes. To fill this gap, we construct PANORAMA, a dataset of 8,143 U.S. patent examination records that preserves the full decision trails, including original applications, all cited references, Non-Final Rejections, and Notices of Allowance. Also, PANORAMA decomposes the trails into sequential benchmarks that emulate patent professionals' patent review processes and allow researchers to examine large language models' capabilities at each step of them. Our findings indicate that, although LLMs are relatively effective at retrieving relevant prior art and pinpointing the pertinent paragraphs, they struggle to assess the novelty and non-obviousness of patent claims. We discuss these results and argue that advancing NLP, including LLMs, in the patent domain requires a deeper understanding of real-world patent examination. Our dataset is openly available at https://github.com/Hyunseung-Lim/PANORAMA.



Paperid:5503
Authors:Sebastian Kassing, Simon Weissmann, Leif Döring
Abstract:
Abstract:The present article studies the minimization of convex, $L$-smooth functions defined on a separable real Hilbert space. We analyze regularized stochastic gradient descent (reg-SGD), a variant of stochastic gradient descent that uses a Tikhonov regularization with time-dependent, vanishing regularization parameter. We prove strong convergence of reg-SGD to the minimum-norm solution of the original problem without additional boundedness assumptions. Moreover, we quantify the rate of convergence and optimize the interplay between step-sizes and regularization decay. Our analysis reveals how vanishing Tikhonov regularization controls the flow of SGD and yields stable learning dynamics, offering new insights into the design of iterative algorithms for convex problems, including those that arise in ill-posed inverse problems. We validate our theoretical findings through numerical experiments on image reconstruction and ODE-based inverse problems.



Paperid:5504
Authors:Diya Zhang, Mengwei Sun, Xingdan Wang, Cheng Liang, Qiaozhen Meng, Shiqiang Ma, Fei Guo
Abstract:
Accurately modeling the target-drug complex at atom level presents a significant challenge in the computer-aided drug design. Traditional methods that rely solely on rigid transformations often fail to capture the dynamic interactions between targets and drugs, particularly during substantial conformational changes in targets upon ligand binding, which becomes especially critical when learning target-drug interactions in drug design. Accurately modeling these changes is crucial for understanding target-drug interactions and improving drug efficacy. To address these challenges, we introduce DynaPhArM, an SE(3)-Equivariant Transformer model specifically designed to capture dynamic alterations occurring within target-drug interactions. DynaPhArM utilizes the cooperative scalar-vector representation, drug-specific embeddings, and a diffusion process to effectively model the evolving dynamics of interactions between targets and drugs. Furthermore, we integrate physical information and energetic principles that maintain essential geometric constraints, such as bond lengths, bond angles, van der Waals forces (vdW), within a multi-task learning (MTL) framework to enhance accuracy. Experimental results demonstrate that DynaPhArM achieves state-of-the-art performance with an overall root mean square deviation (RMSD) of 2.01 Å and a side-chain RMSD of 0.29 Å while exhibiting higher success rates compared to existing methodologies. Additionally, DynaPhArM shows promise in enhancing drug specificity, thereby simulating how targets adapt to various drugs through precise modeling of atomic-level interactions and conformational flexibility.



Paperid:5505
Authors:Nhan-Phu Chung, Jinhui Han, Bohan Li, Zehao Li
Abstract:
We propose and analyze a new class of unbalanced weak optimal transport (OT) problems with total variation penalties, motivated by spatial resource allocation tasks. Unlike classical OT, our framework accommodates general unbalanced nonnegative measures and incorporates cost objectives that directly capture operational trade-offs between transport cost and supply–demand mismatch. In the general setting, we establish the existence of optimal solutions and a dual formulation. We then focus on the semi-discrete setting, where one measure is discrete and the other is absolutely continuous, a structure relevant to applications such as service area partitioning for facilities like schools or medical stations. Exploiting a tessellation-based structure, we derive the corresponding explicit optimality conditions. We further address a quantization problem that jointly optimizes the locations and weights of discrete support points, applicable to facility location tasks such as the cost-efficient deployment of battery swap stations or e-commerce warehouses, informed by demand-side data. The dual-tessellation structure also yields explicit gradient expressions, enabling efficient numerical optimization in finite dimensions.



Paperid:5506
Authors:Riccardo Alberghi, Elizaveta Demyanenko, Luca Biggio, Luca Saglietti
Abstract:
Recent advances in natural language processing highlight two key factors for improving reasoning in large language models (LLMs): (i) allocating more test-time compute tends to help on harder problems but often introduces redundancy in the reasoning trace, and (ii) compute is most effective when reasoning is systematic and incremental, forming structured chains of thought (CoTs) akin to human problem-solving. To study these factors in isolation, we introduce a controlled setting based on shortest-path tasks in layered graphs. We train decoder-only transformers on question–trace–answer triples using a custom tokenizer, comparing models trained on optimal bottom-up dynamic programming traces with those trained on longer, valid traces involving backtracking. Surprisingly, under the same training-token budget, the latter models generalize better to unseen graphs. This benefit is not due to length alone—injecting arbitrary redundancy into reasoning traces fails to help and can even hurt performance. Instead, we find that generalization correlates with the model's confidence in next-token prediction, suggesting that long, coherent, and locally incremental traces make the training signal easier to optimize.



Paperid:5507
Authors:Haoyu Zhao, Yihan Geng, Shange Tang, Yong Lin, Bohan Lyu, Hongzhou Lin, Chi Jin, Sanjeev Arora
Abstract:
LLM-based formal proof assistants (e.g., in Lean) hold great promise for automating mathematical discovery. But beyond syntactic correctness, do these systems truly understand mathematical structure as humans do? We investigate this question through the lens of mathematical inequalities---a fundamental tool across many domains. While modern provers can solve basic inequalities, we probe their ability to handle \emph{human-intuitive compositionality}. We introduce Ineq-Comp, a benchmark built from elementary inequalities through systematic transformations, including variable duplication, algebraic rewriting, and multi-step composition. Although these problems remain easy for humans, we find that most provers---including Goedel, STP, and Kimina-7B---struggle significantly, with only DeepSeek-Prover-V2 showing relative robustness. Despite being designed to decompose the problems into sub-problems and should easily handle the \emph{human-intuitive compositionality}, DeepSeek-Prover-V2 also suffers a 20\% performance drop on these transformations under pass@32. Strikingly, performance remains poor even when provers are given formal proofs of the constituent parts, revealing a core weakness in compositional reasoning. Our results expose a persistent gap between human mathematical intuition and the generalization abilities of current AI provers---especially in their handling of compositional structure in formal reasoning.



Paperid:5508
Authors:Ethan Trepka, Ruobing Xia, Shude Zhu, Sharif Saleki, Danielle Lopes, Stephen Cital, Konstantin Willeke, Mindy Kim, Tirin Moore
Abstract:
The primate visual system is typically divided into two streams — the ventral stream, responsible for object recognition, and the dorsal stream, responsible for encoding spatial relations and motion. Recent studies have shown that convolutional neural networks (CNNs) pretrained on object recognition tasks are remarkably effective at predicting neuronal responses in the ventral stream, shedding light on the neural mechanisms underlying object recognition. However, similar models of the dorsal stream remain underdeveloped due to the lack of large scale datasets encompassing dorsal stream areas. To address this gap, we present STSBench, a dataset of large-scale, single neuron recordings from over 2,000 neurons in the superior temporal sulcus (STS), a nearly 50-fold increase over existing dorsal stream datasets, collected while Rhesus macaques viewed thousands of unique, natural videos. We show that our dataset can be used for benchmarking encoding models of dorsal stream neuronal responses, reconstructing visual input from neural activity, and comparing the features of the visual world extracted by the dorsal and ventral visual streams.



Paperid:5509
Authors:Qinkai XU, yijin liu, YangChen, Lin Yang, Li Li, Yuxiang Fu
Abstract:
Abstract:Diffusion Transformers (DiTs) have recently demonstrated remarkable performance in visual generation tasks, surpassing traditional U-Net-based diffusion models by significantly improving image and video generation quality and scalability. However, the large model size and iterative denoising process introduce substantial computational and memory overhead, limiting their deployment in real-world applications. Post-training quantization (PTQ) is a promising solution that compresses models and accelerates inference by converting weights and activations to low-bit representations. Despite its potential, PTQ faces significant challenges when applied to DiTs, often resulting in severe degradation of generative quality. To address these issues, we propose VETA-DiT (**V**ariance-**E**qualized and **T**emporal **A**daptation for **Di**ffusion **T**ransformers), a dedicated quantization framework for DiTs. Our method first analyzes the sources of quantization error from the perspective of inter-channel variance and introduces a Karhunen–Loève Transform enhanced alignment to equalize variance across channels, facilitating effective quantization under low bit-widths. Furthermore, to handle the temporal variation of activation distributions inherent in the iterative denoising steps of DiTs, we design an incoherence-aware adaptive method that identifies and properly calibrates timesteps with high quantization difficulty. We validate VETA-DiT on extensive image and video generation tasks, preserving acceptable visual quality under the more aggressive W4A4 configuration. Specifically, VETA-DiT reduces FID by 33.65 on the DiT-XL/2 model and by 45.76 on the PixArt-$\Sigma$ model compared to the baseline under W4A4, demonstrating its strong quantization capability and generative performance. Code is available at the anonymous repository: https://anonymous.4open.science/r/VETA-DiT.



Paperid:5510
Authors:Frank Cole, Yulong Lu, Yuxuan Zhao, Tianhao Zhang
Abstract:
Abstract:This paper investigates approximation-theoretic aspects of the in-context learning capability of the transformers in representing a family of noisy linear dynamical systems. Our first theoretical result establishes an upper bound on the approximation error of multi-layer transformers with respect to an $L^2$-testing loss uniformly defined across tasks. This result demonstrates that transformers with logarithmic depth can achieve error bounds comparable with those of the least-squares estimator. In contrast, our second result establishes a non-diminishing lower bound on the approximation error for a class of single-layer linear transformers, which suggests a depth-separation phenomenon for transformers in the in-context learning of dynamical systems. Moreover, this second result uncovers a critical distinction in the approximation power of single-layer linear transformers when learning from IID versus non-IID data.



Paperid:5511
Authors:Brian Bartoldson, Siddarth Venkatraman, James Diffenderfer, Moksh Jain, Tal Ben-Nun, Seanie Lee, Minsu Kim, Johan Obando Ceron, Yoshua Bengio, Bhavya Kailkhura
Abstract:
Reinforcement learning (RL) is a critical component of large language model (LLM) post-training. However, existing on-policy algorithms used for post-training are not naturally robust to a diversified content of experience replay buffers, which can be populated scalably by distributed off-policy actors to enhance exploration as compute increases. We propose efficiently obtaining this benefit of replay buffers via Trajectory Balance with Asynchrony (TBA), a massively scalable LLM RL system. In contrast to existing approaches, TBA uses a larger fraction of compute on search, constantly generating off-policy data for a central replay buffer. A training node simultaneously samples data from this buffer based on reward or recency to update the policy using Trajectory Balance (TB), a diversity-seeking RL objective introduced for GFlowNets. TBA offers three key advantages: (1) decoupled training and search, speeding up training wall-clock time by 4x or more; (2) improved diversity through large-scale off-policy sampling; and (3) scalable search for sparse reward settings. On mathematical reasoning, preference-tuning, and automated red-teaming (diverse and representative post-training tasks), TBA produces speed and performance improvements over strong baselines.



Paperid:5512
Authors:Yang Li, Aming WU, Zihao Zhang, Yahong Han
Abstract:
In this paper, we focus on Novel Class Discoveryfor Point Cloud Segmentation (3D-NCD), aiming to learn a model that can segment unlabeled(novel) 3D classes using only the supervisionfrom labeled (base) 3D classes. The key of the thistask is to setup the exact correlations between thepoint representations and their base class labels,as well as the representation correlations betweenthe points from base and novel classes. A coarseor statistical correlation learning may lead to theconfusion in novel class inference. lf we imposea casual relationship as a strong correlated constraint upon the learning process, the essentialpoint cloud representations that accurately correspond to the classes should be uncovered. Tothis end, we introduce a structural causal model(SCM) to re-formalize the 3D-NCD problem andpropose a new method, i.e., Joint Learning ofCausal Representation and Reasoning. Specifically, we first analyze hidden confounders in thebase class representations and the causal relationships between the base and novel classes throughSCM. We devise a causal representation prototypethat eliminates confounders to capture the causalrepresentations of base classes. A graph structureis then used to model the causal relationships between the base classes’ casual representation prototypes and the novel class prototypes, enablingcausal reasoning from base to novel classes. Extensive experiments and visualization results on3D and 2D NCD semantic segmentation demonstrate the superiorities of our method.



Paperid:5513
Authors:Hugues Van Assel, Mark Ibrahim, Tommaso Biancalani, Aviv Regev, Randall Balestriero
Abstract:
Reconstruction and joint-embedding have emerged as two leading paradigms in Self‑Supervised Learning (SSL). Reconstruction methods focus on recovering the original sample from a different view in input space. On the other hand, joint-embedding methods align the representations of different views in latent space.Both approaches offer compelling advantages, yet practitioners lack clear guidelines for choosing between them.In this work, we unveil the core mechanisms that distinguish each paradigm. By leveraging closed-form solutions for both approaches, we precisely characterize how the view generation process, e.g. data augmentation, impacts the learned representations.We then demonstrate that, unlike supervised learning, both SSL paradigms require a minimal alignment between augmentations and irrelevant features to achieve asymptotic optimality with increasing sample size. Our findings indicate that in scenarios where these irrelevant features have a large magnitude, joint-embedding methods are preferable because they impose a strictly weaker alignment condition compared to reconstruction-based methods. These results not only clarify the trade-offs between the two paradigms but also substantiate the empirical success of joint-embedding approaches on real-world challenging datasets.



Paperid:5514
Authors:Kanghui Ning, Zijie Pan, Yu Liu, Yushan Jiang, James Zhang, Kashif Rasul, Anderson Schneider, Lintao Ma, Yuriy Nevmyvaka, Dongjin Song
Abstract:
Large Language Models (LLMs) and Foundation Models (FMs) have recently become prevalent for time series forecasting tasks. While fine-tuning LLMs enables domain adaptation, they often struggle to generalize across diverse and unseen datasets. Moreover, existing Time Series Foundation Models (TSFMs) still face challenges in handling non-stationary dynamics and distribution shifts, largely due to the lack of effective mechanisms for adaptation. To this end, we present TS-RAG, a retrieval-augmented generation framework for time series forecasting that enhances the generalization and interpretability of TSFMs. Specifically, TS-RAG leverages pre-trained time series encoders to retrieve semantically relevant segments from a dedicated knowledge base, enriching the contextual representation of the input query. Furthermore, we propose an Adaptive Retrieval Mixer (ARM) module that dynamically fuses the retrieved patterns with the TSFM's internal representation, improving forecasting accuracy without requiring task-specific fine-tuning. Thorough empirical studies on seven public benchmark datasets demonstrate that TS-RAG achieves state-of-the-art zero-shot forecasting performance, outperforming the existing TSFMs by up to 6.84\% across diverse domains while also providing desirable interpretability. To support reproducibility, anonymized code and data are available at: https://anonymous.4open.science/r/TS-RAG-F4DB.



Paperid:5515
Authors:Yan Zhang, Yang Gao, Min Li
Abstract:
Fusing multimodal brain imaging has been a hot topic since different modalities of brain imaging can provide complementary information. However, due to the size of simultaneous recorded fMRI-EEG dataset being limited and the substantial discrepancy between hemodynamic responses of fMRI and neural oscillations of EEG, the joint modeling of fMRI and EEG images is a rarely explored area and has not yielded satisfactory results. Existing studies have also indicated that the relationships between region of interest (ROI) are not one-to-one when synchronizing fMRI and EEG. Current graph-based multimodal modeling methods overlook those information. Based on this, we propose a Neural ODEs-hypergraph based fMRI-EEG modeling framework for asynchronous fMRI-EEG data named FE-NET. To the best of our knowledge, this is the first attempt to jointly model fMRI and EEG as hypergraphs. Extensive experiments have demonstrated that the proposed FE-NET outperforms many state-of-the-art brain imaging modeling methods. Meanwhile, compared to simultaneously recorded fMRI-EEG data, asynchronously acquired fMRI-EEG data is less costly, which demonstrates the practical applicability of our method.



Paperid:5516
Authors:Jiatong Shi, Yifan Cheng, Bo-Hao Su, Hye-jin Shim, Jinchuan Tian, Samuele Cornell, Yiwen Zhao, Siddhant Arora, Shinji Watanabe
Abstract:
Speech signal analysis poses significant challenges, particularly in tasks such as speech quality evaluation and profiling, where the goal is to predict multiple perceptual and objective metrics. For instance, metrics like PESQ (Perceptual Evaluation of Speech Quality), STOI (Short-Time Objective Intelligibility), and MOS (Mean Opinion Score) each capture different aspects of speech quality. However, these metrics often have different scales, assumptions, and dependencies, making joint estimation non-trivial. To address these issues, we introduce ARECHO (Autoregressive Evaluation via Chain-based Hypothesis Optimization), a chain-based, versatile evaluation system for speech assessment grounded in autoregressive dependency modeling. ARECHO is distinguished by three key innovations: (1) a comprehensive speech information tokenization pipeline; (2) a dynamic classifier chain that explicitly captures inter-metric dependencies; and (3) a two-step confidence-oriented decoding algorithm that enhances inference reliability. Experiments demonstrate that ARECHO significantly outperforms the baseline framework across diverse evaluation scenarios, including enhanced speech analysis, speech generation evaluation, and noisy speech evaluation. Furthermore, its dynamic dependency modeling improves interpretability by capturing inter-metric relationships.



Paperid:5517
Authors:YUXIANG WEI, Yanteng Zhang, Xi Xiao, Tianyang Wang, Xiao Wang, Vince Calhoun
Abstract:
Decoding visual experiences from fMRI offers a powerful avenue to understand human perception and develop advanced brain-computer interfaces. However, current progress often prioritizes maximizing reconstruction fidelity while overlooking interpretability, an essential aspect for deriving neuroscientific insight. To address this gap, we propose MoRE-Brain, a neuro-inspired framework designed for high-fidelity, adaptable, and interpretable visual reconstruction. MoRE-Brain uniquely employs a hierarchical Mixture-of-Experts architecture where distinct experts process fMRI signals from functionally related voxel groups, mimicking specialized brain networks. The experts are first trained to encode fMRI into the frozen CLIP space. A finetuned diffusion model then synthesizes images, guided by expert outputs through a novel dual-stage routing mechanism that dynamically weighs expert contributions across the diffusion process. MoRE-Brain offers three main advancements: First, it introduces a novel Mixture-of-Experts architecture grounded in brain network principles for neuro-decoding. Second, it achieves efficient cross-subject generalization by sharing core expert networks while adapting only subject-specific routers. Third, it provides enhanced mechanistic insight, as the explicit routing reveals precisely how different modeled brain regions shape the semantic and spatial attributes of the reconstructed image. Extensive experiments validate MoRE-Brain’s high reconstruction fidelity, with bottleneck analyses further demonstrating its effective utilization of fMRI signals, distinguishing genuine neural decoding from over-reliance on generative priors. Consequently, MoRE-Brain marks a substantial advance towards more generalizable and interpretable fMRI-based visual decoding.



Paperid:5518
Authors:Haipeng Liu, Yang Wang, Meng Wang
Abstract:
Text-guided image inpainting aims at reconstructing the masked regions as per text prompts, where the longstanding challenges lie in the preservation for unmasked regions, while achieving the semantics consistency between unmasked and inpainted masked regions. Previous arts failed to address both of them, always with either of them to be remedied. Such facts, as we observed, stem from the entanglement of the hybrid (e.g., mid-and-low) frequency bands that encode varied image properties, which exhibit different robustness to text prompts during the denoising process. In this paper, we propose a null-text-null frequency-aware diffusion models, dubbed NTN-Diff, for text-guided image inpainting, by decomposing the semantics consistency across masked and unmasked regions into the consistencies as per each frequency band, while preserving the unmasked regions, to circumvent two challenges in a row. Based on the diffusion process, we further divide the denoising process into early (high-level noise) and late (low-level noise) stages, where the mid-and-low frequency bands are disentangled during the denoising process. As observed that, the stable mid-frequency band is progressively denoised to be semantically aligned during text-guided denoising process, which, meanwhile, serves as the guidance to the null-text denoising process to denoise low-frequency band for the masked regions, followed by a subsequent text-guided denoising process at late stage, to achieve the semantics consistency for mid-and-low frequency bands across masked and unmasked regions, while preserve the unmasked regions. Extensive experiments validate the superiority of NTN-Diff over the state-of-the-art diffusion models to text-guided diffusion models. Our code can be accessed from Appendix.



Paperid:5519
Authors:Ao Wang, Hui Chen, Jianchao Tan, Kefeng Zhang, Xunliang Cai, Zijia Lin, Jungong Han, guiguang ding
Abstract:
Recently, large vision-language models (LVLMs) have rapidly gained popularity for their strong generation and reasoning capabilities given diverse multimodal inputs. However, these models incur significant computational and memory overhead during inference, which greatly hinders the efficient deployment in practical scenarios. The extensive key-value (KV) cache, necessitated by the lengthy input and output sequences, notably contributes to the high inference cost. Based on this, recent works have investigated ways to reduce the KV cache size for higher efficiency. Although effective, they generally overlook the distinct importance distributions of KV vectors across layers and maintain the same cache size for each layer during the next token prediction. This results in the significant contextual information loss for certain layers, leading to notable performance decline. To address this, we present PrefixKV. It reframes the challenge of determining KV cache sizes for all layers into the task of searching for the optimal global prefix configuration. With an adaptive layer-wise KV retention recipe based on binary search, the maximum contextual information can thus be preserved in each layer, facilitating the generation. Extensive experiments demonstrate that our method achieves the state-of-the-art performance compared with others. It exhibits superior inference efficiency and generation quality trade-offs, showing promising potential for practical applications. Code will be publicly available.



Paperid:5520
Authors:Xiuning Zhang, Xincheng Lu, Nihong Chen, Yuanyuan Mi
Abstract:
Serial dependence reflects how recent sensory history shapes current perception, producing two opposing biases: repulsion, where perception is repelled from recent stimuli, and attraction, where perception is drawn toward them. Repulsion typically occurs at the sensory perception stage, while attraction arises at the post-perception stage. To uncover the neural basis of these effects, we developed a two-layer continuous attractor neural network model incorporating synaptic short-term plasticity (STP). The lower layer, dominated by synaptic depression, models sensory processing and drives repulsion due to sustained neurotransmitter depletion. The higher layer, dominated by synaptic facilitation, models post-perception processing and drives attraction by sustained high neurotransmitter release probability. Our model successfully explains the serial dependence phenomena observed in the visual orientation judgment experiments, highlighting STP as the critical mechanism, with its time constants defining the temporal windows of repulsion and attraction. Furthermore, the model provides a neural foundation for the Bayesian interpretation of serial dependence. This study advances our understanding of how the neural system leverages STP to balance sensitivity in sensory perception with stability in post-perceptual cognition.



Paperid:5521
Authors:Songchen Fu, Siang Chen, Shaojing Zhao, Letian Bai, Hong Liang, Ta Li, Yonghong Yan
Abstract:
In real-world multi-agent systems (MASs), observation delays are ubiquitous, preventing agents from making decisions based on the environment's true state. An individual agent's local observation often consists of multiple components from other agents or dynamic entities in the environment. These discrete observation components with varying delay characteristics pose significant challenges for multi-agent reinforcement learning (MARL). In this paper, we first formulate the decentralized stochastic individual delay partially observable Markov decision process (DSID-POMDP) by extending the standard Dec-POMDP. We then propose the Rainbow Delay Compensation (RDC), a MARL training framework for addressing stochastic individual delays, along with recommended implementations for its constituent modules. We implement the DSID-POMDP's observation generation pattern using standard MARL benchmarks, including MPE and SMAC. Experiments demonstrate that baseline MARL methods suffer severe performance degradation under fixed and unfixed delays. The RDC-enhanced approach mitigates this issue, remarkably achieving ideal delay-free performance in certain delay scenarios while maintaining generalizability. Our work provides a novel perspective on multi-agent delayed observation problems and offers an effective solution framework. The source code is available at https://anonymous.4open.science/r/RDC-pymarl-4512/.



Paperid:5522
Authors:Yijin Ni, Xiaoming Huo
Abstract:
Abstract:This paper introduces a novel kernel-based formulation of the Equalized Odds (EO) criterion, denoted as $\operatorname{EO}_k$, for fair representation learning (FRL) in supervised settings.The central goal of FRL is to mitigate discrimination regarding a sensitive attribute $S$ while preserving prediction accuracy for the target variable $Y$. Our proposed criterion enables a rigorous and interpretable quantification of three core fairness objectives: independence ($\widehat{Y} \perp S$), separation—also known as equalized odds ($\widehat{Y} \perp S \mid Y$), and calibration ($Y \perp S \mid \widehat{Y}$). Under both unbiased ($Y \perp S$) and biased ($Y \not \perp S$) conditions, we show that $\operatorname{EO}_k$ satisfies both independence and separation in the former, and uniquely preserves predictive accuracy while lower bounding independence and calibration in the latter, thereby offering a unified analytical characterization of the tradeoffs among these fairness criteria.We further define the empirical counterpart, $\widehat{\operatorname{EO}}_k$, a kernel-based statistic that can be computed in quadratic time, with linear-time approximations also available.A concentration inequality for $\widehat{\operatorname{EO}}_k$ is derived, providing performance guarantees and error bounds, which serve as practical certificates of fairness compliance. While our focus is on theoretical development, the results lay essential groundwork for principled and provably fair algorithmic design in future empirical studies.



Paperid:5523
Authors:Tingting Dan, Jiaqi Ding, Guorong Wu
Abstract:
State‑space models (SSMs) have become a cornerstone for unraveling brain dynamics, revealing how latent neural states evolve over time and give rise to observed signals. By combining deep learning’s flexibility with SSMs’ principled dynamical structure, recent studies have achieved powerful fits to functional neuroimaging data. However, most approaches still view the brain as a set of loosely connected regions or impose oversimplified network priors—falling short of a truly holistic, self‐organized dynamical system perspective. Brain functional connectivity (FC) at each time point naturally forms a symmetric positive–definite (SPD) matrix, which lives on a curved Riemannian manifold rather than in Euclidean space. Capturing the trajectories of these SPD matrices is key to understanding how coordinated networks support cognition and behavior. To this end, we introduceGeoDynamics, a geometric state‑space neural network that tracks latent brain‐state trajectories directly on the high‑dimensional SPD manifold.GeoDynamicsembeds each connectivity matrix into a manifold‑aware recurrent framework, learning smooth, geometry‑respecting transitions that reveal task‐driven state changes and early markers of Alzheimer’s, Parkinson’s, and autism. Beyond neuroscience, we validateGeoDynamicson human action recognition benchmarks (UTKinect, Florence, HDM05), demonstrating its scalability and robustness in modeling complex spatiotemporal dynamics across diverse domains.



Paperid:5524
Authors:Qianyue Hao, Yiwen Song, Qingmin Liao, Jian Yuan, Yong Li
Abstract:
Abstract:Policy exploration is critical in reinforcement learning (RL), where existing approaches include $\epsilon$-greedy, Gaussian process, etc.However, these approaches utilize preset stochastic processes and are indiscriminately applied in all kinds of RL tasks without considering task-specific features that influence policy exploration. Moreover, during RL training, the evolution of such stochastic processes is rigid, which typically only incorporates a decay in the variance, failing to adjust flexibly according to the agent's real-time learning status.Inspired by the analyzing and reasoning capability of large language models (LLMs), we design **LLM-Explorer** to adaptively generate task-specific exploration strategies with LLMs, enhancing the policy exploration in RL. In our design, we sample the learning trajectory of the agent during the RL training in a given task and prompt the LLM to analyze the agent's current policy learning status and then generate a probability distribution for future policy exploration. Updating the probability distribution periodically, we derive a stochastic process specialized for the particular task and dynamically adjusted to adapt to the learning process. Our design is a plug-in module compatible with various widely applied RL algorithms, including the DQN series, DDPG, TD3, and any possible variants developed based on them. Through extensive experiments on the Atari and MuJoCo benchmarks, we demonstrate LLM-Explorer's capability to enhance RL policy exploration, achieving an average performance improvement up to 37.27\%. Our code is open-source at https://anonymous.4open.science/r/LLM-Explorer-19BE for reproducibility.



Paperid:5525
Authors:Isha Puri, Shivchander Sudalairaj, Guangxuan Xu, Abhishek Bhandwaldar, Kai Xu, Akash Srivastava
Abstract:
Large language models (LLMs) have achieved significant performance gains via scaling up model sizes and/or data. However, recent evidence suggests diminishing returns from such approaches, motivating a pivot to scaling test-time compute.Existing deterministic inference-time scaling methods, usually with reward models, cast the task as a search problem, but suffer from a key limitation: early pruning. Due to inherently imperfect reward models, promising trajectories may be discarded prematurely, leading to suboptimal performance. We propose a novel inference-time scaling approach by adapting particle-based Monte Carlo methods. Our method maintains a diverse set of candidates and robustly balances exploration and exploitation. Our empirical evaluation demonstrates that our particle filtering methods have a 4--16x better scaling rate over deterministic search counterparts on both various challenging mathematical and more general reasoning tasks. Using our approach, we show that Qwen2.5-Math-1.5B-Instruct surpasses GPT-4o accuracy in only 4 rollouts, while Qwen2.5-Math-7B-Instruct scales to o1 level accuracy in only 32 rollouts.Our work not only presents an effective method to inference-time scaling, but also connects rich literature in probabilistic inference with inference-time scaling of LLMs to develop more robust algorithms in future work.



Paperid:5526
Authors:Po-Chen Kuo, Han Hou, Will Dabney, Edgar Walker
Abstract:
Learning a compact representation of history is critical for planning and generalization in partially observable environments. While meta-reinforcement learning (RL) agents can attain near Bayes-optimal policies, they often fail to learn the compact Bayes-optimal belief states. This representational inefficiency potentially limits the agent's adaptability and generalization capacity. Inspired by predictive coding in neuroscience---which suggests the brain predicts sensory inputs as a neural implementation of Bayesian inference---and by predictive objectives in deep RL, we investigate whether integrating self-supervised predictive coding modules into meta-RL can facilitate the learning of Bayes-optimal representations. Through state machine simulation, we show that meta-RL with predictive modules consistently produces more interpretable representations that better approximate Bayes-optimal belief states compared to conventional meta-RL across a wide variety of tasks, even when both achieve optimal policies. Notably, in challenging tasks requiring active information seeking, only meta-RL with predictive modules successfully learns optimal representations and policies, whereas conventional meta-RL struggles with inadequate representation learning. Our results strongly suggest the role of predictive learning as a guiding principle for effective representation learning for both biological and artificial agents navigating partial observability.



Paperid:5527
Authors:Hongling Zheng, Li Shen, Yong Luo, Deheng Ye, Shuhan Xu, Bo Du, Jialie Shen, Dacheng Tao
Abstract:
The Conditional Sequence Modeling (CSM) paradigm, benefiting from the transformer's powerful distribution modeling capabilities, has demonstrated considerable promise in Reinforcement Learning (RL) tasks. However, much of the work has focused on applying CSM to single online or offline settings, with the general architecture rarely explored. Additionally, existing methods primarily focus on deterministic trajectory modeling, overlooking the randomness of state transitions and the diversity of future trajectory distributions. Fortunately, value-based methods offer a viable solution for CSM, further bridging the potential gap between offline and online RL. In this paper, we propose Value-Guided Decision Transformer (VDT), which leverages value functions to perform advantage-weighting and behavior regularization on the Decision Transformer (DT), guiding the policy toward upper-bound optimal decisions during the offline training phase. In the online tuning phase, VDT further integrates value-based policy improvement with behavior cloning under the CSM architecture through limited interaction and data collection, achieving performance improvement within minimal timesteps. The predictive capability of value functions for future returns is also incorporated into the sampling process. Our method achieves competitive performance on various standard RL benchmarks, providing a feasible solution for developing CSM architectures in general scenarios. Code is available at here.



Paperid:5528
Authors:Alvaro Correia, Christos Louizos
Abstract:
Conformal prediction is a distribution-free uncertainty quantification method that has gained popularity in the machine learning community due to its finite-sample guarantees and ease of use. Its most common variant, dubbed split conformal prediction, is also computationally efficient as it boils down to collecting statistics of the model predictions on some calibration data not yet seen by the model. Nonetheless, these guarantees only hold if the calibration and test data are exchangeable, a condition that is difficult to verify and often violated in practice due to so-called distribution shifts. The literature is rife with methods to mitigate the loss in coverage in this non-exchangeable setting, but these methods require some prior information on the type of distribution shift to be expected at test time. In this work, we study this problem via a new perspective, through the lens of optimal transport, and show that it is possible to estimate the loss in coverage and mitigate it in case of distribution shift.



Paperid:5529
Authors:Kecheng Chen, Pingping Zhang, Hui Liu, Jie Liu, Yibing Liu, Jiaxin Huang, Shiqi Wang, Hong Yan, Haoliang Li
Abstract:
We have recently witnessed that ''Intelligence" and `''Compression" are the two sides of the same coin, where the language large model (LLM) with unprecedented intelligence is a general-purpose lossless compressor for various data modalities. This attribute is particularly appealing to the lossless image compression community, given the increasing need to compress high-resolution images in the current streaming media era. Consequently, a spontaneous envision emerges: Can the compression performance of the LLM elevate lossless image compression to new heights? However, our findings indicate that the naive application of LLM-based lossless image compressors suffers from a considerable performance gap compared with existing state-of-the-art (SOTA) codecs on common benchmark datasets. In light of this, we are dedicated to fulfilling the unprecedented intelligence (compression) capacity of the LLM for lossless image compression tasks, thereby bridging the gap between theoretical and practical compression performance. Specifically, we propose P -LLM, a next-pixel prediction-based LLM, which integrates various elaborated insights and methodologies, \textit{e.g.,} pixel-level priors, the in-context ability of LLM, and a pixel-level semantic preservation strategy, to enhance the understanding capacity of pixel sequences for better next-pixel predictions. Extensive experiments on benchmark datasets demonstrate that P-LLM can beat SOTA classical and learned codecs.



Paperid:5530
Authors:Connor Dunlop, Matthew Zheng, Kavana Venkatesh, Pinar Yanardag Delul
Abstract:
Text-to-image (T2I) diffusion models have made remarkable strides in generating and editing high-fidelity images from text. Yet, these models remain fundamentally generic, failing to adapt to the nuanced aesthetic preferences of individual users. In this work, we present the first framework for personalized image editing in diffusion models, introducing Collaborative Direct Preference Optimization (C-DPO) - a novel method that aligns image edits with user-specific preferences while leveraging collaborative signals from like-minded individuals. Our approach encodes each user as a node in a dynamic preference graph and learns embeddings via a lightweight graph neural network, enabling information sharing across users with overlapping visual tastes. We enhance a diffusion model's editing capabilities by integrating these personalized embeddings into a novel DPO objective, which jointly optimizes for individual alignment and neighborhood coherence. To support this setting, we curate the first personalized image editing benchmark, a large-scale synthetic dataset comprising 3,000 richly structured user profiles with diverse editing histories and aesthetic inclinations. Comprehensive experiments, including user studies and quantitative benchmarks, demonstrate that our method consistently outperforms baselines in generating edits that are not only visually appealing and semantically precise, but also aligned with user identity.



Paperid:5531
Authors:Alistair Shilton, Sunil Gupta, Santu Rana, Svetha Venkatesh
Abstract:
This paper explores the use of Hermite transform based reproducing kernel Banach space methods to construct exact or un-approximated models of feedforward neural networks of arbitrary width, depth and topology, including residual networks, transformers and other attention-based networks, assuming only a feedforward topology, finite energy activations and finite (spectral-) norm weights and biases. Using this model, two straightforward but surprisingly tight bounds on Rademacher complexity are derived, precisely (1) a general bound that is width-independent and scales exponentially with depth; and (2) a width- and depth-independent bound for networks (including ReLU, ResNet and transformer networks) with appropriately constrained (below threshold) weights and biases.



Paperid:5532
Authors:Ziyu Jia, Tingyu Du, Zhengyu Tian, Hongkai Li, Yong Zhang, Chenyu Liu
Abstract:
Emotion recognition based on physiological signals is of considerable significance in fields including psychological health and human-computer interaction, particularly in light of the substantial advances in multimodal emotion recognition techniques. However, two key challenges remain unresolved: 1) how to effectively model the intra-modal long-range dependencies and inter-modal correlations in multimodal physiological emotion signals and 2) how to address the performance limitations resulting from missing multimodal data. In this paper, we propose a multimodal bidirectional Mamba (BiMamba) network with test-time adaptation (TTA) for emotion recognition named BiM-TTA. Specifically, BiM-TTA consists of a multimodal BiMamba network and a multimodal TTA. The former includes intra-modal and inter-modal BiMamba modules, which model long-range dependencies along the time dimension and capture cross-modal correlations along the channel dimension, respectively. The latter (TTA) effectively mitigates the negative impact of the distribution shifts amplified by missing multimodal data through two-level entropy-based sample filtering and mutual information sharing across modalities. Experiments on two multimodal emotion datasets demonstrate that BiM-TTA achieves state-of-the-art performance.



Paperid:5533
Authors:Mingyang Liu, Xinyang Chen, Xiucheng Li, Weili Guan, Liqiang Nie
Abstract:
Unsupervised domain adaptation has emerged as a pivotal paradigm for mitigating distribution shifts in time series analysis. The fundamental challenge in time series domain adaptation arises from the entanglement of domain shifts and intricate temporal patterns. Crucially, the latent continuous-time dynamics, which are often inaccessible due to sensor constraints, are only partially observable through discrete time series from an explicit sensor-limited single view. This partial observability hinders the modeling of intricate temporal patterns, impeding domain invariant representation learning. To mitigate the limitation, we proposeEDEN(multipleExplicitDomainEnhanced adaptationNetwork), expanding the raw dataset to multi-scale explicit domains, multi-subspace explicit domains and multi-segment explicit domains. EDEN enhances domain adaptation with three coordinated modules tailored to integrate multiple explicit domains: (1) Multi-Scale Curriculum Adaptation implements progressive domain alignment from coarse-scale to fine-scale. (2) Quality-Aware Feature Fusion evaluates feature quality in multi-subspace explicit domains and adaptively integrates temporal-frequency features. (3) Temporal Coherence Learning enforces segment-level consistency with multi-segment explicit domains. The representation enriched by multiple explicit domains bridges the gap between partially observed discrete samples and the underlying implicit temporal dynamics, enabling more accurate approximation of implicit temporal patterns for effective cross-domain adaptation. Our comprehensive evaluation across 6 time series benchmarks demonstrates EDEN's consistent superiority, achieving average accuracy improvements of 4.8% over state-of-the-art methods in cross-domain scenarios. Code is available at the anonymous link:https://anonymous.4open.science/r/2025NeurIPS-EDEN.



Paperid:5534
Authors:Pengqi Li, Lizhong Ding, Jiarun Fu, Chunhui Zhang, Guoren Wang, Ye Yuan
Abstract:
Abstract:Kolmogorov-Arnold Networks (KANs) have demonstrated remarkable expressive capacity and predictive power in symbolic learning. However, existing generalization errors of KANs primarily focus on approximation errors while neglecting estimation errors, leading to a suboptimal bias-variance trade-off and poor generalization performance. Meanwhile, the unclear generalization mechanism hinders the design of more effective KANs variants. As the authors of KANs highlighted, they ``would like to explore ways to restrict KANs' hypothesis space so that they can achieve good performance''. To address these challenges, we explore the generalization mechanism of KANs and design more effective KANs with lower model complexity and better generalization. We define \textit{Lipschitz complexity} as the first structural measure for deep functions represented by KANs and derive novel generalization bounds based on \textit{Lipschitz complexity}, establishing a theoretical foundation for understanding their generalization behavior. To reduce \textit{Lipschitz complexity} and boost the generalization mechanism of KANs, we propose Lipschitz-Enhanced KANs ($\textbf{LipKANs}$) by integrating the Lip layer and pioneering the $L_{1.5}$-regularized loss, contributing to tighter generalization bounds. Empirical experiments validate that the proposed LipKANs enhance the generalization mechanism of KANs when modeling complex distributions. We hope our theoretical bounds and LipKANs lay a foundation for the future development of KANs.



Paperid:5535
Authors:Abhishek Sinha, Rahul Vaze
Abstract:
Abstract:We revisit the Online Convex Optimization problem with adversarial constraints (COCO) where, at the beginning of each round, a learner selects an action from a convex decision set. Thereafter, an adversary reveals a convex cost function and a convex constraint function for that round. The goal of the learner is to select a sequence of actions to minimize both regret and the cumulative constraint violation (CCV) over $T$ rounds. The best-known policy for this problem achieves $O(\sqrt{T})$ regret and $\tilde{O}(\sqrt{T})$ CCV. In this paper, we improve upon this result by achieving a significantly smaller CCV by trading it off with regret. Specifically, for any bounded convex cost and constraint functions, we propose an online policy that achieves $\tilde{O}(\sqrt{dT}+ T^\beta)$ regret and $\tilde{O}(dT^{1-\beta})$ CCV, where $d$ is the dimension of the decision set and $\beta \in [0,1]$ is a tunable parameter. We achieve this result by first considering a special case, called the $\texttt{Constrained Expert}$ problem, where the decision set is a probability simplex and the cost and constraint functions are linear. Leveraging a new adaptive small-loss regret bound, we propose a computationally efficient policy for the $\texttt{Constrained Expert}$ problem, that attains $O(\sqrt{T\ln N}+T^{\beta})$ regret and $\tilde{O}(T^{1-\beta} \ln N)$ CCV, where $N$ is the number of experts. The original problem is then reduced to the $\texttt{Constrained Expert}$ problem via a covering argument. Finally, with an additional $M$-smoothness assumption, we propose a computationally efficient gradient-based policy attaining $O(\sqrt{MT}+T^{\beta})$ regret and $\tilde{O}(MT^{1-\beta})$ CCV.



Paperid:5536
Authors:Shogo Iwazaki
Abstract:
We study the noise-free Gaussian Process (GP) bandit problem, in which a learner seeks to minimize regret through noise-free observations of a black-box objective function that lies in a known reproducing kernel Hilbert space (RKHS).The Gaussian Process Upper Confidence Bound (GP-UCB) algorithm is a well-known approach for GP bandits, where query points are adaptively selected based on the GP-based upper confidence bound score.While several existing works have reported the practical success of GP-UCB, its theoretical performance remains suboptimal. However, GP-UCB often empirically outperforms other nearly-optimal noise-free algorithms that use non-adaptive sampling schemes.This paper resolves the gap between theoretical and empirical performance by establishing a nearly-optimal regret upper bound for noise-free GP-UCB. Specifically, our analysis provides the first constant cumulative regret bounds in the noise-free setting for both the squared exponential kernel and the Matérn kernel with some degree of smoothness.



Paperid:5537
Authors:Leqi Zheng, Chaokun Wang, Zixin Song, Cheng Wu, Shannan Yan, Jiajun Zhang, Ziyang Liu
Abstract:
Traditional recommender systems have relied heavily on positive feedback for learning user preferences, while the abundance of negative feedback in real-world scenarios remains underutilized. To address this limitation, recent years have witnessed increasing attention on leveraging negative feedback in recommender systems to enhance recommendation performance. However, existing methods face three major challenges: limited model compatibility, ineffective information exchange, and computational inefficiency. To overcome these challenges, we propose a model-agnostic Signed Dual-Channel Graph Contrastive Learning (SDCGCL) framework that can be seamlessly integrated with existing graph contrastive learning methods. The framework features three key components: (1) a Dual-Channel Graph Embedding that separately processes positive and negative graphs, (2) a Cross-Channel Distribution Calibration mechanism to maintain structural consistency, and (3) an Adaptive Prediction Strategy that effectively combines signals from both channels. Building upon this framework, we further propose a Dual-channel Feedback Fusion (DualFuse) model and develop a two-stage optimization strategy to ensure efficient training. Extensive experiments on four public datasets demonstrate that our approach consistently outperforms state-of-the-art baselines by substantial margins while exhibiting minimal computational complexity. Our source code and data are released at \url{https://anonymous.4open.science/r/SDCGCL-E56E/}.



Paperid:5538
Authors:Noah Krever, Jakub Cerny, Moise Blanchard, Christian Kroer
Abstract:
Game-theoretic algorithms are commonly benchmarked on recreational games, classical constructs from economic theory such as congestion and dispersion games, or entirely random game instances. While the past two decades have seen the rise of security games -- grounded in real-world scenarios like patrolling and infrastructure protection -- their practical evaluation has been hindered by limited access to the datasets used to generate them. In particular, although the structural components of these games (e.g., patrol paths derived from maps) can be replicated, the critical data defining target values -- central to utility modeling -- remain inaccessible. In this paper, we introduce a flexible framework that leverages open-access datasets to generate realistic matrix and security game instances. These include animal movement data for modeling anti-poaching scenarios and demographic and infrastructure data for infrastructure protection. Our framework allows users to customize utility functions and game parameters, while also offering a suite of preconfigured instances. We provide theoretical results highlighting the degeneracy and limitations of benchmarking on random games, and empirically compare our generated games against random baselines across a variety of standard algorithms for computing Nash and Stackelberg equilibria, including linear programming, incremental strategy generation, and self-play with no-regret learners.



Paperid:5539
Authors:Boyang Peng, Sanqing Qu, Tianpei Zou, Fan Lu, Ya Wu, Kai Chen, Siheng Chen, Yong Wu, Guang Chen
Abstract:
In real-world environments, a well-designed model must be capable of handling dynamically evolving distributions, where both in-distribution (ID) and out-of-distribution (OOD) samples appear unpredictably and individually, making real-time adaptation particularly challenging. While open-set test-time adaptation has demonstrated effectiveness in adjusting to distribution shifts, existing methods often rely on batch processing and struggle to manage single-sample data stream in open-set environments. To address this limitation, we propose Open-IRT, a novel open-set Intermediate-Representation-based Test-time adaptation framework tailored for single-image test-time adaptation with vision-language models. Open-IRT comprises two key modules designed for dynamic, single-sample adaptation in open-set scenarios. The first is Polarity-aware Prompt-based OOD Filter module, which fully constructs the ID-OOD distribution, considering both the absolute semantic alignment and relative semantic polarity. The second module, Intermediate Domain-based Test-time Adaptation module, constructs an intermediate domain and indirectly decomposes the ID-OOD distributional discrepancy to refine the separation boundary during the test-time. Extensive experiments on a range of domain adaptation benchmarks demonstrate the superiority of Open-IRT. Compared to previous state-of-the-art methods, it achieves significant improvements on representative benchmarks, such as CIFAR-100C and SVHN — with gains of +8.45\% in accuracy, -10.80\% in FPR95, and +11.04\% in AUROC.



Paperid:5540
Authors:Sihan Wang, Wenjie Du, Qing Zhu, Yang Wang
Abstract:
Artificial intelligence (AI) has played a transformative role in chemical research, greatly facilitating the prediction of small molecule properties, simulation of catalytic processes, and material design. These advances are driven by increases in computing power, open source machine learning frameworks, and extensive chemical datasets. However, a persistent challenge is the limited amount of high-quality real-world data, while models calculated based on large amounts of theoretical data are often costly and difficult to deploy, which hinders the applicability of AI models in real-world scenarios. In this study, we enhance the prediction of solute-solvent properties by proposing a novel sample selection method: the iterative core subset extraction (CSIE) framework. CSIE iteratively updates the core sample subset based on information gain to remove redundant features in theoretical data and optimize the performance of the model on real chemical datasets. Furthermore, we introduce an asymmetric molecular interaction graph neural network (AMGNN) that combines positional information and bidirectional edge connections to simulate real-world chemical reaction scenarios to better capture solute-solvent interactions. Experimental results show that our method can accurately extract the core subset and improve the prediction accuracy.



Paperid:5541
Authors:Haizhou Du, Yiran Xiang, Yiwen Cai, Xiufeng Liu, Zonghan Wu, Huan Huo, Guodong Long
Abstract:
Heterogeneous Federated Learning (HtFL) enables collaborative learning across clients with diverse model architectures and non-IID data distributions, which are prevalent in real-world edge computing applications. Existing HtFL approaches typically employ proxy datasets to facilitate knowledge sharing or implement coarse-grained model-level knowledge transfer. However, such approaches not only elevate risks of user privacy leakage but also lead to the loss of fine-grained model-specific knowledge, ultimately creating barriers to effective knowledge sharing. To address these challenges, we propose FedFree, a novel data-free and model-free HtFL framework featuring two key innovations. First, FedFree introduces a reverse layer-wise knowledge transfer mechanism that aggregates heterogeneous client models into a global model solely using Gaussian-based pseudo data, eliminating reliance on proxy datasets. Second, it leverages Knowledge Gain Entropy (KGE) to guide targeted layer-wise knowledge alignment, ensuring that each client receives the most relevant global updates tailored to its specific architecture. We provide rigorous theoretical convergence guarantees for FedFree and conduct extensive experiments on CIFAR-10 and CIFAR-100. Results demonstrate that FedFree achieves substantial performance gains, with relative accuracy improving up to 46.3% over state-of-the-art baselines. The framework consistently excels under highly heterogeneous model/data distributions and in large scale settings.



Paperid:5542
Authors:Kaichen Xu, Yihang Du, Mianpeng Liu, Zimu Yu, Xiaobo Sun
Abstract:
Positional encoding is essential for supplementing transformer with positional information of tokens. Existing positional encoding methods demand predefined token/feature order, rendering them unsuitable for real-world data with non-sequential yet causally-related features. To address this limitation, we propose CAPE, a novel method that identifies underlying causal structure over non-sequential features as a weighted directed acyclic graph (DAG) using generalized structural equation modeling. The DAG is then embedded in hyperbolic space where its geometric structure is well-preserved using a hyperboloid model-based approach that effectively captures two important causal graph properties (causal strength & causal specificity). This step yields causality-aware positional encodings for the features, which are converted into their rotary form for integrating with transformer's self-attention mechanism. Theoretical analysis reveals that CAPE-generated rotary positional encodings possess three valuable properties for enhanced self-attention, including causal distance-induced attenuation, causal generality-induced attenuation, and robustness to positional disturbances. We evaluate CAPE over both synthetic and real-word datasets, empirically demonstrating its theoretical properties and effectiveness in enhancing transformer for data with non-sequential features.



Paperid:5543
Authors:Xuan Chen, Shiwei Feng, Zikang Xiong, Shengwei An, Yunshu Mao, Lu Yan, Guanhong Tao, Wenbo Guo, Xiangyu Zhang
Abstract:
Assessing the safety of autonomous driving (AD) systems against security threats, particularly backdoor attacks, is a stepping stone for real-world deployment. However, existing works mainly focus on pixel-level triggers which are impractical to deploy in the real world. We address this gap by introducing a novel backdoor attack against the end-to-end AD systems that leverage one or more other vehicles' trajectories as triggers. To generate precise trigger trajectories, we first use temporal logic (TL) specifications to define the behaviors of attacker vehicles. Configurable behavior models are then used to generate these trajectories, which are quantitatively evaluated and iteratively refined based on the TL specifications. We further develop a negative training strategy by incorporating patch trajectories that are similar to triggers but are designated not to activate the backdoor. It enhances the stealthiness of the attack and refines the system’s responses to trigger scenarios. Through extensive experiments on 5 offline reinforcement learning (RL) driving agents with 6 trigger patterns and target actions combinations, we demonstrate the flexibility and effectiveness of our proposed attack, showing the under-exploration of existing end-to-end AD systems' vulnerabilities to such trajectory-based backdoor attacks.Videos of our attack are available at: https://sites.google.com/view/tlbackdoor/home.



Paperid:5544
Authors:Hidde Fokkema, Tim van Erven, Sara Magliacane
Abstract:
Machine learning is a vital part of many real-world systems, but several concernsremain about the lack of interpretability, explainability and robustness of black-boxAI systems. Concept Bottleneck Models (CBM) address some of these challengesby learning interpretable concepts from high-dimensional data, e.g. images, whichare used to predict labels. An important issue in CBMs are spurious correlationbetween concepts, which effectively lead to learning “wrong” concepts. Currentmitigating strategies have strong assumptions, e.g., they assume that the conceptsare statistically independent of each other, or require substantial interaction interms of both interventions and labels provided by annotators. In this paper, wedescribe a framework that provides theoretical guarantees on the correctness ofthe learned concepts and on the number of required labels, without requiring anyinterventions. Our framework leverages causal representation learning (CRL)methods to learn latent causal variables from high-dimensional observations ina unsupervised way, and then learns to align these variables with interpretableconcepts with few concept labels. We propose a linear and a non-parametricestimator for this mapping, providing a finite-sample high probability result in thelinear case and an asymptotic consistency result for the non-parametric estimator.We evaluate our framework in synthetic and image benchmarks, showing that thelearned concepts have less impurities and are often more accurate than other CBMs,even in settings with strong correlations between concepts.



Paperid:5545
Authors:Feng He, Guodong Tan, Qiankun Li, Quan Wen, Jun Yu
Abstract:
Light field microscopy (LFM) has become an emerging tool in neuroscience for large-scale neural imaging in vivo, with XLFM (eXtended Light Field Microscopy) notable for its single-exposure volumetric imaging, broad field of view, and high temporal resolution. However, learning-based 3D reconstruction in XLFM remains underdeveloped due to two core challenges: the absence of standardized datasets and the lack of methods that can efficiently model its angular–spatial structure while remaining physically grounded. We address these challenges by introducing three key contributions. First, we construct the XLFM-Zebrafish benchmark, a large-scale dataset and evaluation suite for XLFM reconstruction. Second, we propose Masked View Modeling for Light Fields (MVM-LF), a self-supervised task that learns angular priors by predicting occluded views, improving data efficiency. Third, we formulate the Optical Rendering Consistency Loss (ORC Loss), a differentiable rendering constraint that enforces alignment between predicted volumes and their PSF-based forward projections. On the XLFM-Zebrafish benchmark, our method improves PSNR by 7.7\% over state-of-the-art baselines. Code, dataset, and evaluation protocol are publicly available at: xxx.



Paperid:5546
Authors:Mucong Ding, Bang An, Tahseen Rabbani, Chenghao Deng, Anirudh Satheesh, Souradip Chakraborty, Mehrdad Saberi, Yuxin Wen, Kyle Sang, Aakriti Agrawal, Xuandong Zhao, Mo Zhou, Mary-Anne Hartley, Lei Li, Yu-Xiang Wang, Vishal Patel, Soheil Feizi, Tom Goldstein, Furong Huang
Abstract:
Abstract:AI-generated images have become pervasive, raising critical concerns around content authenticity, intellectual property, and the spread of misinformation. Invisible watermarks offer a promising solution for identifying AI-generated images, preserving content provenance without degrading visual quality. However, their real-world robustness remains uncertain due to the lack of standardized evaluation protocols and large-scale stress testing. To bridge this gap, we organized “Erasing the Invisible,” a NeurIPS 2024 competition and newly established benchmark designed to systematically stress testing the resilience of watermarking techniques. The competition introduced two attack tracks—Black-box and Beige-box—that simulate practical scenarios with varying levels of attacker knowledge on watermarks, providing a comprehensive assessment of watermark robustness. The competition attracted significant global participation, with 2,722 submissions from 298 teams. Through a rigorous evaluation pipeline featuring real-time feedback and human-verified final rankings, participants developed and demonstrated new attack strategies that revealed critical vulnerabilities in state-of-the-art watermarking methods. On average, the top-5 teams in both tracks could remove watermarks from $\geq$ 89% of the images while preserving high visual quality, setting strong baselines for future research on watermark attacks and defenses. To support continued progress in this field, we summarize the insights and lessons learned from this competition in this paper, and release the benchmark dataset, evaluation toolkit, and competition results. “Erasing the Invisible” establishes a valuable open resource for advancing more robust watermarking techniques and strengthening content provenance in the era of generative AI.



Paperid:5547
Authors:CHUANG MA, Tomoyuki Obuchi, Toshiyuki Tanaka
Abstract:
A phenomenon known as ``Neural Collapse (NC)'' in deep classification tasks, in which the penultimate-layer features and the final classifiers exhibit an extremely simple geometric structure, has recently attracted considerable attention, with the expectation that it can deepen our understanding of how deep neural networks behave. The Unconstrained Feature Model (UFM) has been proposed to explain NC theoretically, and there emerges a growing body of work that extends NC to tasks other than classification and leverages it for practical applications. In this study, we investigate whether a similar phenomenon arises in deep Ordinal Regression (OR) tasks, via combining the cumulative link model for OR and UFM. We show that a phenomenon we call Ordinal Neural Collapse (ONC) indeed emerges and is characterized by the following three properties: (ONC1) all optimal features in the same class collapse to their within-class mean when regularization is applied; (ONC2) these class means align with the classifier, meaning that they collapse onto a one-dimensional subspace; (ONC3) the optimal latent variables (corresponding to logits or preactivations in classification tasks) are aligned according to the class order, and in particular, in the zero-regularization limit, a highly local and simple geometric relationship emerges between the latent variables and the threshold values. We prove these properties analytically within the UFM framework with fixed threshold values and corroborate them empirically across a variety of datasets. We also discuss how these insights can be leveraged in OR, highlighting the use of fixed thresholds.



Paperid:5548
Authors:Wenjie Du, Xuqiang Li, Jinke Feng, Shuai Zhang, Wen Zhang, Yang Wang
Abstract:
Drug recommendation systems aim to identify optimal drug combinations for patient care, balancing therapeutic efficacy and safety. Advances in large-scale longitudinal EHRs have enabled learning-based approaches that leverage patient histories such as diagnoses, procedures, and previously prescribed drugs, to model complex patient-drug relationships. Yet, many existing solutions overlook standard clinical practices that favor certain drugs for specific conditions and fail to fully integrate the influence of molecular substructures on drug efficacy and safety. In response, we propose \textbf{SubRec}, a unified framework that integrates representation learning across both patient and drug spaces. Specifically, SubRec introduces a conditional information bottleneck to extract core drug substructures most relevant to patient conditions, thereby enhancing interpretability and clinical alignment. Meanwhile, an adaptive vector quantization mechanism is designed to generate patient–drug interaction patterns into a condition-aware codebook which reuses clinically meaningful patterns, reduces training overhead, and provides a controllable latent space for recommendation. Crucially, the synergy between condition-specific substructure learning and discrete patient prototypes allows SubRec to make accurate and personalized drug recommendations. Experimental results on the real-world MIMIC III and IV demonstrate our model's advantages. The source code is available at \href{https://anonymous.4open.science/r/DrugRecommendation-5173}{https://anonymous.4open.science/}.



Paperid:5549
Authors:Zechen Wu, Amy Greenwald, Ronald Parr
Abstract:
In off-policy policy evaluation (OPE) tasks within reinforcement learning, Temporal Difference Learning(TD) and Fitted Q-Iteration (FQI) have traditionally been viewed as differing in the number of updates toward the target value function: TD makes one update, FQI makes an infinite number, and Partial Fitted Q-Iteration (PFQI) performs a finite number. We show that this view is not accurate, and provide a new mathematical perspective under linear value function approximation that unifies these methods as a single iterative method solving same linear system, but using different matrix splitting schemes and preconditioners. We show that increasing the number of updates under the same target value function, i.e., the target network technique, is a transition from using a constant preconditioner to using a data-feature adaptive preconditioner. This elucidates, for the first time, why TD convergence does not necessarily imply FQI convergence, and establishes tight convergence connections among TD, PFQI, and FQI. Our framework enables sharper theoretical results than previous work and characterization of the convergence conditions for each algorithm, without relying on assumptions about the features (e.g., linear independence). We also provide an encoder-decoder perspective to better understand TD’s convergence conditions, and prove, for the first time, that when a large learning rate doesn’t work, trying a smaller one may help(for batch TD). Our framework also leads to the discovery of new crucial conditions on features for convergence, and shows how common assumptions about features influence convergence, e.g., the assumption of linearly independent features can be dropped without compromising the convergence guarantees of stochastic TD in the on-policy setting. This paper is also the first to introduce matrix splitting into the convergence analysis of these algorithms.



Paperid:5550
Authors:Elfarouk Harb, Yousef Yassin, Chandra Chekuri
Abstract:
Abstract:We consider the following question: given a submodular or supermodular set function $f:2^V \to \mathbb{R}$, how should one minimize or maximize its average value $f(S)/|S|$ over non-empty subsets $S\subseteq V$? This problem generalizes several well-known objectives including Densest Subgraph (DSG), Densest Supermodular Set (DSS), and Submodular Function Minimization (SFM). Motivated by recent applications [39, 31], we formalize two new broad problems: the Unrestricted Sparsest Submodular Set (USSS) and Unrestricted Densest Supermodular Set (UDSS) which allow negative and non-monotone functions.Using classical results we observe that DSS, SFM, USSS, UDSS, and MNP are all equivalent under strongly polynomial-time reductions. This equivalence enables algorithmic cross-over: methods designed for one problem can be repurposed to solve others efficiently. In particular we use the perspective of the minimum norm point in the base polyhedron of a sub/supermodular function which, via Fujishige's results, yields the dense decomposition as a byproduct. Via this perspective we show that a recent converging heuristic for DSS, \textsc{SuperGreedy++} [15, 29], and Wolfe’s minimum norm point algorithm are both universal solvers for all of these problems. On the theoretical front, we explain the observation made in recent work [39, 31] that \textsc{SuperGreedy++} appears to work well even in settings beyond DSS. Surprisingly, we also show that this simple algorithm can be used for Submodular Function Minimization, including for example that it can act as an effective minimum $st$ cut algorithm. On the empirical front, we explore the utility of several different algorithms including Fujishige-Wolfe min-norm point algorithm for recent problems. We conduct over 400 experiments across seven problem types and large-scale synthetic and real-world datasets (up to $\approx 100$ million edges). Our results reveal that methods historically considered inefficient, such as convex-programming methods, flow-based solvers, and Fujishige-Wolfe’s algorithm, outperform state-of-the-art task-specific baselines by orders of magnitude on concrete problems like HNSN [39]. These findings challenge prevailing assumptions and demonstrate that with the right framing, general optimization algorithms can be both scalable and state-of-the-art for supermodular and submodular ratio problems.



Paperid:5551
Authors:Xiaotang Gai, Jiaxiang Liu, Yichen Li, Zijie Meng, Jian Wu, Zuozhu Liu
Abstract:
Medical Visual Question Answering (Med-VQA) holds significant potential for clinical decision support, yet existing efforts primarily focus on 2D imaging with limited task diversity. This paper presents 3D-RAD, a large-scale dataset designed to advance 3D Med-VQA using radiology CT scans. The 3D-RAD dataset encompasses six diverse VQA tasks: anomaly detection, image observation, medical computation, existence detection, static temporal diagnosis, and longitudinal temporal diagnosis. It supports both open- and closed-ended questions while introducing complex reasoning challenges, including computational tasks and multi-stage temporal analysis, to enable comprehensive benchmarking. Extensive evaluations demonstrate that existing vision-language models (VLMs), especially medical VLMs exhibit limited generalization, particularly in multi-temporal tasks, underscoring the challenges of real-world 3D diagnostic reasoning. To drive future advancements, we release a high-quality training set 3D-RAD-T of 136,195 expert-aligned samples, showing that fine-tuning on this dataset could significantly enhance model performance. Our dataset and code are publicly available at https://github.com/Tang-xiaoxiao/M3D-RAD, aiming to catalyze multimodal medical AI research and establish a robust foundation for 3D medical visual understanding.



Paperid:5552
Authors:Hossein Rajoli Nowdeh, Jie Ji, Xiaolong Ma, Fatemeh Afghah
Abstract:
In multimodal learning, dominant modalities often overshadow others, limiting generalization. We propose Modality-Aware Sharpness-Aware Minimization (M-SAM), a model-agnostic framework that applies to many modalities and supports early and late fusion scenarios. In every iteration, M-SAM in three steps optimizes learning. \textbf{First, it identifies the dominant modality} based on modalities' contribution in the accuracy using Shapley. \textbf{Second, it decomposes the loss landscape}, or in another language, it modulates the loss to prioritize the robustness of the model in favor of the dominant modality, and \textbf{third, M-SAM updates the weights} by backpropagation of modulated gradients. This ensures robust learning for the dominant modality while enhancing contributions from others, allowing the model to explore and exploit complementary features that strengthen overall performance. Extensive experiments on four diverse datasets show that M-SAM outperforms the latest state-of-the-art optimization and gradient manipulation methods and significantly balances and improves multimodal learning. The code will be released.



Paperid:5553
Authors:Yongrui Chen, Yi Huang, Yunchang Liu, Shenyu Zhang, Junhao He, Tongtong Wu, Guilin Qi, Tianxing Wu
Abstract:
Continual Structured Knowledge Reasoning (CSKR) focuses on training models to handle sequential tasks, where each task involves translating natural language questions into structured queries grounded in structured knowledge. Existing general continual learning approaches face significant challenges when applied to this task, including poor generalization to heterogeneous structured knowledge and inefficient reasoning due to parameter growth as tasks increase. To address these limitations, we propose a novel CSKR framework, \textsc{K-DeCore}, which operates with a fixed number of tunable parameters. Unlike prior methods, \textsc{K-DeCore} introduces a knowledge decoupling mechanism that disentangles the reasoning process into task-specific and task-agnostic stages, effectively bridging the gaps across diverse tasks. Building on this foundation, \textsc{K-DeCore} integrates a dual-perspective memory consolidation mechanism for distinct stages and introduces a structure-guided pseudo-data synthesis strategy to further enhance the model's generalization capabilities.Extensive experiments on four benchmark datasets demonstrate the superiority of \textsc{K-DeCore} over existing continual learning methods across multiple metrics, leveraging various backbone large language models.



Paperid:5554
Authors:Yunuo Zhang, Baiting Luo, Ayan Mukhopadhyay, Gabor Karsai, Abhishek Dubey
Abstract:
In Partially Observable Markov Decision Processes (POMDPs), maintaining and updating belief distributions over possible underlying states provides a principled way to summarize action-observation history for effective decision-making under uncertainty. As environments grow more realistic, belief distributions develop complexity that standard mathematical models cannot accurately capture, creating a fundamental challenge in maintaining representational accuracy. Despite advances in deep learning and probabilistic modeling, existing POMDP belief approximation methods fail to accurately represent complex uncertainty structures such as high-dimensional, multi-modal belief distributions, resulting in estimation errors that lead to suboptimal agent behaviors. To address this challenge, we present ESCORT (Efficient Stein-variational and sliced Consistency-Optimized Representation for Temporal beliefs), a particle-based framework for capturing complex, multi-modal distributions in high-dimensional belief spaces. ESCORT extends SVGD with two key innovations: correlation-aware projections that model dependencies between state dimensions, and temporal consistency constraints that stabilize updates while preserving correlation structures. This approach retains SVGD's attractive-repulsive particle dynamics while enabling accurate modeling of intricate correlation patterns. Unlike particle filters prone to degeneracy or parametric methods with fixed representational capacity, ESCORT dynamically adapts to belief landscape complexity without resampling or restrictive distributional assumptions. We demonstrate ESCORT's effectiveness through extensive evaluations on both POMDP domains and synthetic multi-modal distributions of varying dimensionality, where it consistently outperforms state-of-the-art methods in terms of belief approximation accuracy and downstream decision quality.



Paperid:5555
Authors:Weiyu Guo, Ziyang Chen, Shaoguang WANG, Jianxiang He, Yijie Xu, Jinhui Ye, Ying Sun, Hui Xiong
Abstract:
Understanding long video content is a complex endeavor that often relies on densely sampled frame captions or end-to-end feature selectors, yet these techniques commonly overlook the logical relationships between textual queries and visual elements. In practice, computational constraints necessitate coarse frame subsampling, a challenge analogous to “finding a needle in a haystack.” To address this issue, we introduce a semantics-driven search framework that reformulates keyframe selection under the paradigm of Visual Semantic-Logical Search (VSLS). Specifically, we systematically define four fundamental logical dependencies: 1) spatial co-occurrence, 2) temporal proximity, 3) attribute dependency, and 4) causal order. These relations dynamically update frame sampling distributions through an iterative refinement process, enabling context-aware identification of semantically critical frames tailored to specific query requirements. Our method establishes new state-of-the-art performance on the manually annotated benchmark in keyframe selection metrics. Furthermore, when applied to downstream video question-answering tasks, the proposed approach demonstrates the best performance gains over existing methods on LongVideoBench and Video-MME, validating its effectiveness in bridging the logical gap between textual queries and visual-temporal reasoning. The code will be publicly available.



Paperid:5556
Authors:Chanakya Ekbote, Ashok Vardhan Makkuva, Marco Bondaschi, Nived Rajaraman, Michael Gastpar, Jason Lee, Paul Pu Liang
Abstract:
Abstract:In-context learning (ICL) is a hallmark capability of transformers, through which trained models learn to adapt to new tasks by leveraging information from the input context. Prior work has shown that ICL emerges in transformers due to the presence of special circuits called induction heads. Given the equivalence between induction heads and conditional $k$-grams, a recent line of work modeling sequential inputs as Markov processes has revealed the fundamental impact of model depth on its ICL capabilities: while a two-layer transformer can efficiently represent a conditional $1$-gram model, its single-layer counterpart cannot solve the task unless it is exponentially large. However, for higher order Markov sources, the best known constructions require at least three layers (each with a single attention head) - leaving open the question: *can a two-layer single-head transformer represent any $k^{\text{th}}$-order Markov process?* In this paper, we precisely address this and theoretically show that a two-layer transformer with one head per layer can indeed represent any conditional $k$-gram. Thus, our result provides the tightest known characterization of the interplay between transformer depth and Markov order for ICL. Building on this, we further analyze the learning dynamics of our two-layer construction, focusing on a simplified variant for first-order Markov chains, illustrating how effective in-context representations emerge during training. Together, these results deepen our current understanding of transformer-based ICL and illustrate how even shallow architectures can surprisingly exhibit strong ICL capabilities on structured sequence modeling tasks.



Paperid:5557
Authors:Jigang Fan, Quanlin Wu, Shengjie Luo, Liwei Wang
Abstract:
The detection of ligand binding sites for proteins is a fundamental step in Structure-Based Drug Design. Despite notable advances in recent years, existing methods, datasets, and evaluation metrics are confronted with several key challenges: (1) current datasets and methods are centered on individual protein–ligand complexes and neglect that diverse binding sites may exist across multiple complexes of the same protein, introducing significant statistical bias; (2) ligand binding site detection is typically modeled as a discontinuous workflow, employing binary segmentation and subsequent clustering algorithms; (3) traditional evaluation metrics do not adequately reflect the actual performance of different binding site prediction methods. To address these issues, we first introduce UniSite-DS, the first UniProt (Unique Protein)-centric ligand binding site dataset, which contains 4.81 times more multi-site data and 2.08 times more overall data compared to the previously most widely used datasets. We then propose UniSite, the first end-to-end ligand binding site detection framework supervised by set prediction loss with bijective matching. In addition, we introduce Average Precision based on Intersection over Union (IoU) as a more accurate evaluation metric for ligand binding site prediction. Extensive experiments on UniSite-DS and several representative benchmark datasets demonstrate that IoU-based Average Precision provides a more accurate reflection of prediction quality, and that UniSite outperforms current state-of-the-art methods in ligand binding site detection. The dataset and codes will be released.



Paperid:5558
Authors:Zinan Lin, Enshu Liu, Xuefei Ning, Junyi Zhu, Wenyu Wang, Sergey Yekhanin
Abstract:
Generative modeling, representation learning, and classification are three core problems in machine learning (ML), yet their state-of-the-art (SoTA) solutions remain largely disjoint. In this paper, we ask: Can a unified principle address all three? Such unification could simplify ML pipelines and foster greater synergy across tasks. We introduce Latent Zoning Network (LZN) as a step toward this goal. At its core, LZN creates a shared Gaussian latent space that encodes information across all tasks. Each data type (e.g., images, text, labels) is equipped with an encoder that maps samples to disjoint latent zones, and a decoder that maps latents back to data. ML tasks are expressed as compositions of these encoders and decoders: for example, label-conditional image generation uses a label encoder and image decoder; image embedding uses an image encoder; classification uses an image encoder and label decoder. We demonstrate the promise of LZN in three increasingly complex scenarios: (1) LZN can enhance existing models (image generation): When combined with the SoTA Rectified Flow model, LZN improves FID on CIFAR10 from 2.76 to 2.59—without modifying the training objective. (2) LZN can solve tasks independently (representation learning): LZN can implement unsupervised representation learning without auxiliary loss functions, outperforming the seminal MoCo method by 5.4% on downstream linear classification on ImageNet. (3) LZN can solve multiple tasks simultaneously (joint generation and classification): With image and label encoders/decoders, LZN performs both tasks jointly by design, improving FID and achieving SoTA classification accuracy on CIFAR10. Code and models will be released.



Paperid:5559
Authors:jun ling, Yao Qi, Tao Huang, Shibo Zhou, Yanqin Huang, Yang Jiang, Ziqi Song, Ying Zhou, Yang Yang, Hengtao Shen, Peng Wang
Abstract:
In this work, we address the task of table image to LaTeX code generation, with the goal of automating the reconstruction of high-quality, publication-ready tables from visual inputs. A central challenge of this task lies in accurately handling complex tables—those with large sizes, deeply nested structures, and semantically rich or irregular cell content—where existing methods often fail. We begin with a comprehensive analysis, identifying key challenges and highlighting the limitations of current evaluation protocols. To overcome these issues, we propose a reinforced multimodal large language model (MLLM) framework, where a pre-trained MLLM is fine-tuned on a large-scale table-to-LaTeX dataset. To further improve generation quality, we introduce a dual-reward reinforcement learning strategy based on Group Relative Policy Optimization (GRPO). Unlike standard approaches that optimize purely over text outputs, our method incorporates both a structure-level reward on LaTeX code and a visual fidelity reward computed from rendered outputs, enabling direct optimization of the visual output quality.We adopt a hybrid evaluation protocol combining TEDS-Structure and CW-SSIM, and show that our method achieves state-of-the-art performance, particularly on structurally complex tables, demonstrating the effectiveness and robustness of our approach.



Paperid:5560
Authors:Fuyang Liu, Jiaqi Xu, Xiaowei Hu
Abstract:
Adverse weather significantly degrades real-world visual perception, posing a challenge to the robustness of vision systems. Existing methods, often trained on synthetic data with static network parameters, struggle to generalize to complex real-world degradations. To address this, we develop a dual-level reinforcement learning framework that leverages high-quality synthetic data as a cold start and adaptively refines the models using real-world data through reinforcement learning. First, we construct HFLS-Weather, a physics-driven, high-fidelity synthetic dataset that simulates diverse weather phenomena. Second, we formulate a dual-level reinforcement learning architecture: at the local level, weather-specific restoration models are refined using the proposed perturbation-driven image quality optimization, enabling reward-based learning without paired supervision for image restoration; at the global level, a meta-controller dynamically coordinates model selection and execution order based on scene degradation. Our framework continuously adapts to real-world conditions and achieves state-of-the-art performance across a wide range of real-world adverse weather scenarios.



Paperid:5561
Authors:Chengdong Dong, B. V. K. Vijaya Kumar, Zhenyu Zhou, Ajay Kumar
Abstract:
The remarkable progress in neural-network-driven visual data generation, especially with neural rendering techniques like Neural Radiance Fields and 3D Gaussian splatting, offers a powerful alternative to GANs and diffusion models. These methods can generate high-fidelity images and lifelike avatars, highlighting the need for robust detection methods. However, the lack of any large dataset containing images from neural rendering methods becomes a bottleneck for the detection of such sophisticated fake images. To address this limitation, we introduce NeuroRenderedFake, a comprehensive benchmark for evaluating emerging fake image detection methods. Our key contributions are threefold: (1) A large-scale dataset of fake images synthesized using state-of-the-art neural rendering techniques, significantly expanding the scope of fake image detection beyond generative models; (2) A cross-domain evaluation protocol designed to assess the domain gap and common artifacts between generative and neural rendering-based fake images; and (3) An in-depth spectral energy analysis that reveals how frequency domain characteristics influence the performance of fake image detectors. We train representative detectors, based on spatial, spectral, and multimodal architectures, on fake images generated by both generative and neural rendering models. We evaluate these detectors on 15 groups of fake images synthesized by cutting-edge neural rendering models, generative models, and combined methods that can exhibit artifacts from both domains. Additionally, we provide insightful findings through detailed experiments on degraded fake image detection and the impact of spectral features, aiming to advance research in this critical area.



Paperid:5562
Authors:Shizheng Wen, Arsh Kumbhat, Levi Lingsch, Sepehr Mousavi, Yizhou Zhao, Praveen Chandrashekar, Siddhartha Mishra
Abstract:
The very challenging task of learning solution operators of PDEs on arbitrary domains accurately and efficiently is of vital importance to engineering and industrial simulations. Despite the existence of many operator learning algorithms to approximate such PDEs, we find that accurate models are not necessarily computationally efficient and vice versa. We address this issue by proposing geometry aware operator transformer (GAOT) for learning PDEs on arbitrary domains. GAOT combines novel multi-scale attentional graph neural operator encoders and decoders, together with geometry embeddings and (vision) transformer processors to accurately map information about the domain and the inputs into a robust approximation of the PDE solution. Multiple innovations in the implementation of GAOT also ensure computational efficiency and scalability. We demonstrate this significant gain in both accuracy and efficiency of GAOT over several baselines on a large number of learning tasks from a diverse set of PDEs, including achieving state of the art performance on a large scale three-dimensional industrial CFD dataset.



Paperid:5563
Authors:Yi Xiao, Qiangqiang Yuan, Kui Jiang, Wenke Huang, Qiang Zhang, Tingting Zheng, Chia-Wen Lin, Liangpei Zhang
Abstract:
Spiking neural networks (SNNs) are emerging as a promising alternative to traditional artificial neural networks (ANNs), offering biological plausibility and energy efficiency. Despite these merits, SNNs are frequently hampered by limited capacity and insufficient representation power, yet remain underexplored in remote sensing image (RSI) super-resolution (SR) tasks. In this paper, we first observe that spiking signals exhibit drastic intensity variations across diverse textures, highlighting an active learning state of the neurons. This observation motivates us to apply SNNs for efficient SR of RSIs. Inspired by the success of attention mechanisms in representing salient information, we devise the spiking attention block (SAB), a concise yet effective component that optimizes membrane potentials through inferred attention weights, which, in turn, regulates spiking activity for superior feature representation. Our key contributions include: 1) we bridge the independent modulation between temporal and channel dimensions, facilitating joint feature correlation learning, and 2) we access the global self-similar patterns in large-scale remote sensing imagery to infer spatial attention weights, incorporating effective priors for realistic and faithful reconstruction. Building upon SAB, we proposed SpikeSR, which achieves state-of-the-art performance across various remote sensing benchmarks such as AID, DOTA, and DIOR, while maintaining high computational efficiency. The code of SpikeSR will be available upon paper acceptance.



Paperid:5564
Authors:Yihang Zhou, Chen Wei, Minghao Sun, Jin Song, Yang Li, Lin Wang, Yang Zhang
Abstract:
Understanding the conformational landscape of proteins is essential for elucidatingprotein function and facilitating drug design. However, existing protein confor-mation benchmarks fail to capture the full energy landscape, limiting their abilityto evaluate the diversity and physical plausibility of AI-generated structures. Weintroduce ProteinConformers, a large-scale benchmark dataset comprising over381,000 physically realistic conformations for 87 CASP targets. These were de-rived from more than 40,000 structural decoys via extensive all-atom moleculardynamics simulations totaling over 6 million CPU hours. Using this dataset,we propose novel metrics to evaluate conformational diversity and plausibility,and systematically benchmark six protein conformation generative models. Ourresults highlight that leveraging large-scale protein sequence data can enhance amodel’s ability to explore conformational space, potentially reducing reliance onMD-derived data. Additionally, we find that PDB and MD datasets influence modelperformance differently, current models perform well on inter-atomic distance pre-diction but struggle with inter-residue orientation generation. Overall, our dataset,evaluation metrics, and benchmarking results provide the first comprehensive foun-dation for assessing generative models in protein conformational modeling. Datasetand instructions are available at https://huggingface.co/datasets/Jim990908/ProteinConformers/tree/main.



Paperid:5565
Authors:Rohit Kanchi, Benjamin Melanson, Nithin Somasekharan, Shaowu Pan, Sicheng He
Abstract:
Abstract:We present UniFoil, the largest publicly available universal airfoil database based on Reynolds-Averaged Navier–Stokes (RANS) simulations. It contains over 500,000 samples spanning a wide range of Reynolds and Mach numbers, capturing both transitional and fully turbulent flows across incompressible to compressible regimes. UniFoil is designed to support machine learning research in fluid dynamics, particularly for modeling complex aerodynamic phenomena.Most existing datasets are limited to incompressible, fully turbulent flows with smooth field characteristics, thus overlooking the critical physics of laminar–turbulent transition and shock-wave interactions—features that exhibit strong nonlinearity and sharp gradients. UniFoil fills this gap by offering a broad spectrum of realistic flow conditions.In the database, turbulent simulations utilize the Spalart–Allmaras (SA) model, while transitional flows are modeled using an $e^N$-based transition prediction method coupled with the SA model. The database includes a comprehensive geometry set comprising over 4,800 natural laminar flow (NLF) airfoils and 30,000 fully turbulent (FT) airfoils, effectively covering the diversity of airfoil designs relevant to aerospace, wind energy, and marine applications.This database is also highly valuable for scientific machine learning (SciML), enabling the development of data-driven models that more accurately capture the transport processes associated with laminar–turbulent transition. UniFoil is freely available under a permissive CC-BY-SA license.



Paperid:5566
Authors:Andy Bonnetto, Haozhe Qi, Franklin Leong, Matea Tashkovska, Mahdi Rad, Solaiman Shokur, Friedhelm C. Hummel, Silvestro Micera, Marc Pollefeys, Alexander Mathis
Abstract:
Understanding behavior requires datasets that capture humans while carrying out complex tasks. The kitchen is an excellent environment for assessing human motor and cognitive function, as many complex actions are naturally exhibited in kitchens from chopping to cleaning. Here, we introduce the X-Smart-Kitchen-30 dataset (Note: Name "X" anonymized), collected in a noninvasive motion capture platform inside a kitchen environment. Nine static RGB-D cameras, inertial measurement units (IMUs) and one head-mounted HoloLens~2 headset were used to capture 3D hand, body, and eye movements. The X-Smart-Kitchen-30 dataset is a multi-view action dataset with synchronized exocentric, egocentric, depth, IMUs, eye gaze, body and hand kinematics spanning 29.7 hours of 16 subjects cooking four different recipes. Action sequences were densely annotated with 33.78 action segments per minute. Leveraging this multi-modal dataset, we propose four benchmarks to advance behavior understanding and modeling through 1) a vision-language benchmark, 2) a semantic text-to-motion generation benchmark, 3) a multi-modal action recognition benchmark, 4) a pose-based action segmentation benchmark. We expect the X-Smart-Kitchen-30 dataset to pave the way for better methods as well as insights to understand the nature of ethologically-valid human behavior.



Paperid:5567
Authors:Benjamin Ramtoula, Pierre-Yves Lajoie, Paul Newman, Daniele De Martini
Abstract:
Foundation models (FMs) trained with different objectives and data learn diverse representations, making some more effective than others for specific downstream tasks. Existing adaptation strategies – such as parameter-efficient fine-tuning – focus on individual models and do not exploit the complementary strengths across models. Probing methods offer a promising alternative by extracting information from frozen models, but current techniques do not scale well with large feature sets and often rely on dataset-specific hyperparameter tuning. We propose Combined backBones (ComBo), a simple and scalable probing-based adapter that effectively integrates features from multiple models and layers. ComBo compresses activations from layers of one or more FMs into compact token-wise representations and processes them with a lightweight transformer for task-specific prediction. Crucially, ComBo does not require dataset-specific tuning or backpropagation through the backbone models. On the 19 tasks of the VTAB-1k benchmark, ComBo outperforms previous probing methods, matches or surpasses more expensive alternatives, such as distillation-based model merging, and enables efficient probing of tuned models. Our results demonstrate that ComBo offers a practical and general-purpose framework for combining diverse representations from multiple FMs.



Paperid:5568
Authors:Xueqing Deng, Linjie Yang, Qihang Yu, Ali Athar, Chenglin Yang, Xiaojie Jin, Xiaohui Shen, Liang-Chieh Chen
Abstract:
This paper introduces the COCONut-PanCap dataset, created to enhance panoptic segmentation and grounded image captioning. Building upon the COCO dataset with advanced COCONut panoptic masks, this dataset aims to overcome limitations in existing image-text datasets that often lack detailed, scene-comprehensive descriptions. The COCONut-PanCap dataset incorporates fine-grained, region-level captions grounded in panoptic segmentation masks, ensuring consistency and improving the detail of generated captions.Through human-edited, densely annotated descriptions, COCONut-PanCap supports improved training of vision-language models (VLMs) for image understanding and generative models for text-to-image tasks.Experimental results demonstrate that COCONut-PanCap significantly boosts performance across understanding and generation tasks, offering complementary benefits to large-scale datasets. This dataset sets a new benchmark for evaluating models on joint panoptic segmentation and grounded captioning tasks, addressing the need for high-quality, detailed image-text annotations in multi-modal learning.



Paperid:5569
Authors:Chen Qian, Dongrui Liu, Haochen Wen, Zhen Bai, Yong Liu, Jing Shao
Abstract:
Large reasoning models (LRMs) have demonstrated impressive capabilities in complex problem-solving, yet their internal reasoning mechanisms remain poorly understood.In this paper, we investigate the reasoning trajectories of LRMs from an information-theoretic perspective. By tracking how mutual information (MI) between intermediate representations and the correct answer evolves during LRM reasoning, we observe an interesting MI peaks phenomenon: the MI at specific generative steps exhibits a sudden and significant increase during LRM's reasoning process. We theoretically analyze such phenomenon and show that as MI increases, the probability of model's prediction error decreases.Furthermore, these MI peaks often correspond to tokens expressing reflection or transition, such as "Hmm", "Wait" and "Therefore," which we term as the thinking tokens.We then demonstrate that these thinking tokens are crucial for LRM's reasoning performance, while other tokens has minimal impacts.Building on these analyses, we propose two simple yet effective methods to improve LRM's reasoning performance, by delicately leveraging these thinking tokens.Overall, our work provides novel insights into the reasoning mechanisms of LRMs and offers practical ways to improve their reasoning capabilities.



Paperid:5570
Authors:Xingxing Yang, Jie Chen, Zaifeng Yang
Abstract:
Spectral reconstruction from RGB images often suffers from a metameric dilemma, where distinct spectral distributions map to nearly identical RGB values, making them indistinguishable to current models and leading to unreliable reconstructions.In this paper, we present Diff-Spectra that integrates supervised physics-aware spectral estimation and unsupervised high-fidelity spectral regularization for HSI reconstruction.We first introduce an Adaptive illumiChroma Decoupling (AICD) module to decouple illumination and chrominance information, which learns intrinsic and distinctive feature distributions, thereby mitigating the metameric issue.Then, we incorporate the AICD into a learnable spectral response function (SRF) guided hyperspectral initial estimation mechanism to mimic the physical image formation and thus inject physics-aware reasoning into neural networks, turning an ill-posed problem into a constrained, interpretable task. We also introduce a metameric spectra augmentation method to synthesize comprehensive hyperspectral data to pre-train a Spectral Diffusion Module (SDM), which internalizes the statistical properties of real-world HSI data, enforcing unsupervised high-fidelity regularization on the spectral transitions via inner-loop optimization during inference.Extensive experimental evaluations demonstrate that our Diff-Spectra achieves SOTA performance on both Spectral reconstruction and downstream HSI classification. The code and dataset will be released.



Paperid:5571
Authors:Dail Kim, Joon-Hyuk Chang
Abstract:
Target sound extraction aims to isolate target sound sources from an input mixture using a target clue to identify the sounds of interest. To address the challenge posed by the wide variety of sounds, recent work has introduced privileged knowledge distillation (PKD), which utilizes privileged information (PI) about the target sound, available only during training. While PKD has shown promise, existing approaches often suffer from overfitting of the teacher model for the overly rich PI and ineffective knowledge transfer to the student model. In this paper, we propose Disentangled Codec Knowledge Distillation (DCKD) to mitigate these issues by regulating the amount and the flow of target sound information within the teacher model. We begin by extracting a compressed representation of the target sound using a neural audio codec to regulate the amount of PI. Disentangled representation learning is then applied to remove class information and extract fine-grained temporal information as PI. Subsequently, an n-hot vector as the target clue and the class-independent PI are used to condition the early and later layers of the teacher model, respectively, forming a regulated coarse-to-fine target information flow. The resulting representation is transferred to the student model through feature-level knowledge distillation. Experimental results show that DCKD consistently improves existing methods across model architectures under the multi-target selection condition.



Paperid:5572
Authors:Hongyuan Tao, Ying Zhang, Zhenhao Tang, Hongen Peng, Xukun Zhu, Bingchang Liu, Yingguang Yang, Ziyin Zhang, Zhaogui Xu, Haipeng Zhang, Linchao Zhu, Rui Wang, Hang Yu, Jianguo Li, Peng Di
Abstract:
Recent advances in Large Language Models (LLMs) have shown promise in function-level code generation, yet repository-level software engineering tasks remain challenging. Current solutions predominantly rely on proprietary LLM agents, which introduce unpredictability and limit accessibility, raising concerns about data privacy and model customization. This paper investigates whether open-source LLMs can effectively address repository-level tasks without requiring agent-based approaches. We demonstrate this is possible by enabling LLMs to comprehend functions and files within codebases through their semantic information and structural dependencies. To this end, we introduce Code Graph Models (CGMs), which integrate repository code graph structures into the LLM's attention mechanism and map node attributes to the LLM's input space using a specialized adapter. When combined with an agentless graph RAG framework, our approach achieves a 43.00% resolution rate on the SWE-bench Lite benchmark using the open-source Qwen2.5-72B model. This performance ranks first among open weight models, second among methods with open-source systems, and eighth overall, surpassing the previous best open-source model-based method by 12.33%.



Paperid:5573
Authors:Kyra Kadhim, Jonas Beck, Ziwei Huang, Jakob H Macke, Fred Rieke, Thomas Euler, Michael Deistler, Philipp Berens
Abstract:
Visual processing starts in the outer retina where photoreceptors transform light into electrochemical signals. These signals are modulated by inhibition from horizontal cells and sent to the inner retina via excitatory bipolar cells. The outer retina is thought to play an important role in contrast invariant coding of visual information, but how the different cell types implement this computation together remains incompletely understood. To understand the role of each cell type, we developed a fully-differentiable biophysical model of a circular patch of mouse outer retina in JAX. The model includes 200 cone photoreceptors with a realistic phototransduction cascade and ribbon synapses as well as horizontal and bipolar cells, all with cell-type specific ion channels. Going beyond decades of work constraining biophysical models of neurons only by experimental data, we used a dual approach, constraining parts of the model with available measurements and others by a visual task: (1) We fit the parameters of the cone models to whole cell patch-clamp measurements of photocurrents and two-photon glutamate imaging measurements of synaptic release. (2) We then trained the spatiotemporal outer retina model with photoreceptors and the other cell types to perform a visual classification task with varying contrast and luminance levels. We found that our outer retina model could learn to solve the classification task despite contrast and luminance variance in the stimuli. Testing different cell type compositions and connectivity patterns, we found that inhibitory feedback through horizontal cells did not further improve task performance beyond that of excitatory photoreceptors and bipolar cells. Finally, we found that our model generalized better to out of distribution contrast levels than a linear classifier. Our work shows how the nonlinearities found in the outer retina can accomplish contrast invariant classification and teases apart the contributions of different cell types.



Paperid:5574
Authors:Dongguen Kim, Young-Geun Choi, Minwoo Chae
Abstract:
Abstract:Dynamic pricing algorithms typically assume continuous price variables, which may not reflect real-world scenarios where prices are often discrete. This paper demonstrates that leveraging discrete price information within a semi-parametric model can substantially improve performance, depending on the size of the support set of the price variable relative to the time horizon. Specifically, we propose a novel semi-parametric contextual dynamic pricing algorithm, namely BayesCoxCP, based on a Bayesian approach to the Cox proportional hazards model. Our theoretical analysis establishes high-probability regret bounds that adapt to the sparsity level $\gamma$, proving that our algorithm achieves a regret upper bound of $\widetilde{O}(T^{(1+\gamma)/2}+\sqrt{dT})$ for $\gamma < 1/3$ and $\widetilde{O}(T^{2/3}+\sqrt{dT})$ for $\gamma \geq 1/3$, where $\gamma$ represents the sparsity of the price grid relative to the time horizon $T$. Through numerical experiments, we demonstrate that our proposed algorithm significantly outperforms an existing method, particularly in scenarios with sparse discrete price points.



Paperid:5575
Authors:Yisong Fu, Fei Wang, Zezhi Shao, Boyu Diao, Lin Wu, Zhulin An, Chengqing Yu, Yujie Li, Yongjun Xu
Abstract:
Transformers have gained attention in atmospheric time series forecasting (ATSF) for their ability to capture global spatial-temporal correlations. However, their complex architectures lead to excessive parameter counts and extended training times, limiting their scalability to large-scale forecasting. In this paper, we revisit ATSF from a theoretical perspective of atmospheric dynamics and uncover a key insight: spatial-temporal position embedding (STPE) can inherently model spatial-temporal correlations even without attention mechanisms. Its effectiveness arises from integrating geographical coordinates and temporal features, which are intrinsically linked to atmospheric dynamics. Based on this, we proposeSTELLA, aSpatial-Temporal knowledgeEmbeddedLightweight modeLfor ASTF, utilizing only STPE and an MLP architecture in place of Transformer layers. With 10k parameters and one hour of training, STELLA achieves superior performance on five datasets compared to other advanced methods. The paper emphasizes the effectiveness of spatial-temporal knowledge integration over complex architectures, providing novel insights for ATSF.



Paperid:5576
Authors:Ling Zhan, Junjie Huang, Xiaoyao Yu, Wenyu Chen, Tao Jia
Abstract:
Functional brain network (FBN) modeling often relies on local pairwise interactions, whose limitation in capturing high-order dependencies is theoretically analysed in this paper. Meanwhile, the computational burden and heuristic nature of current hypergraph modeling approaches hinder end-to-end learning of FBN structures directly from data distributions. To address this, we propose to extract high-order FBN structures under global constraints, and implement this as a Global Constraints oriented Multi-resolution (GCM) FBN structure learning framework. It incorporates 3 types of global constraint (expected edge numbers, data source, and data labels) to enable learning FBN structures for 4 distinct levels (sample/subject/group/project) of modeling resolution. Experimental results demonstrate that GCM achieves up to a 30.6% improvement in relative accuracy and a 96.3% reduction in computational time across 5 datasets and 2 task settings, compared to 7 baselines and 8 state-of-the-art methods. Extensive experiments validate the contributions of individual components and highlight the interpretability of GCM. This work offers a novel perspective on FBN structure learning and provides a foundation for interdisciplinary applications in cognitive neuroscience.



Paperid:5577
Authors:Nicolas Castanet, Sylvain Lamprier, Olivier Sigaud
Abstract:
Abstract:Goal-Conditioned Reinforcement Learning (GCRL) enables agents to autonomously acquire diverse behaviors, but faces major challenges in visual environments due to high-dimensional, semantically sparse observations. In the online setting, where agents learn representations while exploring, the latent space evolves with the agent's policy, to capture newly discovered areas of the environment. However, without incentivization to maximize state coverage in the representation, classical approaches based on auto-encoders may converge to latent spaces that over-represent a restricted set of states frequently visited by the agent. This is exacerbated in an intrinsic motivation setting, where the agent uses the distribution encoded in the latent space to sample the goals it learns to master.To address this issue, we propose to progressively enforce distributional shifts towards a uniform distribution over the full state space, to ensure a full coverage of skills that can be learned in the environment.We introduce DRAG (Distributionally Robust Auto-Encoding for GCRL), a method that combines the $\beta$-VAE framework with Distributionally Robust Optimization (DRO). DRAG leverage an adversarial neural weighter of training states of the VAE, to account for the mismatch between the current data distribution and unseen parts of the environment. This allows the agent to construct semantically meaningful latent spaces beyond its immediate experience. Our approach improves state space coverage and downstream control performance on hard exploration environments such as mazes and robotic control involving walls to bypass, without relying on pre-training nor prior environment knowledge.



Paperid:5578
Authors:Kijung Yoon
Abstract:
Inferring synaptic connectivity from neural population activity is a fundamental challenge in computational neuroscience, complicated by partial observability and mismatches between inference models and true circuit dynamics. In this study, we propose a graph-based neural inference model that simultaneously predicts neural activity and infers latent connectivity by modeling neurons as interacting nodes in a graph. The architecture features two distinct modules: one for learning structural connectivity and another for predicting future spiking activity via a graph neural network (GNN). Our model accommodates unobserved neurons through auxiliary nodes, allowing for inference in partially observed circuits. We evaluate this approach using synthetic data from ring attractor networks and real spike recordings from head direction cells in mice. Across a wide range of conditions, including varying recurrent connectivity, external inputs, and incomplete observations, our model consistently outperforms standard baselines, resolving spurious correlations more effectively and recovering accurate weight profiles. When applied to real data, the inferred connectivity aligns with theoretical predictions of continuous attractor models. These results highlight the potential of GNN-based models to infer latent neural circuitry through self-supervised structure learning, while leveraging the spike prediction task to flexibly link connectivity and dynamics across both simulated and biological neural systems.



Paperid:5579
Authors:Yaxin Hou, Bo Han, Yuheng Jia, Hui LIU, Junhui Hou
Abstract:
Current long-tailed semi-supervised learning methods assume that labeled data exhibit a long-tailed distribution, and unlabeled data adhere to a typical predefined distribution (i.e., long-tailed, uniform, or inverse long-tailed). However, the distribution of the unlabeled data is generally unknown and may follow an arbitrary distribution. To tackle this challenge, we propose a Controllable Pseudo-label Generation (CPG) framework, expanding the labeled dataset with the progressively identified reliable pseudo-labels from the unlabeled dataset and training the model on the updated labeled dataset with a known distribution, making it unaffected by the unlabeled data distribution. Specifically, CPG operates through a controllable self-reinforcing optimization cycle: (i) At each training step, our dynamic controllable filtering mechanism selectively incorporates reliable pseudo-labels from the unlabeled dataset into the labeled dataset, ensuring that the updated labeled dataset follows a known distribution; (ii) We then construct a Bayes-optimal classifier using logit adjustment based on the updated labeled data distribution; (iii) This improved classifier subsequently helps identify more reliable pseudo-labels in the next training step. We further theoretically prove that this optimization cycle can significantly reduce the generalization error under some conditions. Additionally, we propose a class-aware adaptive augmentation module to further improve the representation of minority classes, and an auxiliary branch to maximize data utilization by leveraging all labeled and unlabeled samples. Comprehensive evaluations on various commonly used benchmark datasets show that CPG achieves consistent improvements, surpassing state-of-the-art methods by up to16.29\%in accuracy.Code is available in the supplementary material.



Paperid:5580
Authors:Vardhan Dongre, Chi Gui, Shubham Garg, Hooshang Nayyeri, Gokhan Tur, Dilek Tur, Vikram Adve
Abstract:
We introduce MIRAGE, a new benchmark for multimodal expert-level reasoning and decision-making in consultative interaction settings. Designed for the domain of agriculture, MIRAGE captures the full complexity of expert consultations by combining natural user queries, expert-authored responses, and image-based context, offering a high-fidelity benchmark for evaluating models on grounded reasoning, clarification strategies, and long-form generation in a real-world, knowledge-intensive domain. Grounded in over 35,000 real user-expert interactions, and curated through a carefully designed multi-step pipeline, MIRAGE spans diverse crop health, pest diagnosis, and crop management scenarios. The benchmark includes more than 7,000 unique biological entities, covering plant species, pests, and diseases, making it one of the most taxonomically diverse benchmarks available for vision-language models in real-world expert-guided domains. Unlike existing benchmarks that rely on well-specified user inputs, MIRAGE features underspecified, context-rich scenarios, requiring models to infer latent knowledge gaps and either proactively guide the interaction or respond. Our benchmark comprises two core components. The Single-turn Challenge to reason over a single user turn and image set, identify relevant entities, infer causal explanations, and generate actionable recommendations; and a Multi-Turn challenge for dialogue state tracking, goal-driven generation, and expert-level conversational decision-making. We evaluate more than 20 closed and open-source frontier vision-language models (VLMs), using three reasoning language models as evaluators, highlighting the significant challenges posed by MIRAGE in both single-turn and multi-turn interaction settings. Even the advanced GPT4.1 and GPT4o models achieve 44.6% and 40.9% accuracy, respectively, indicating significant room for improvement.



Paperid:5581
Authors:Tianchen Zhao, Ke Hong, Xinhao Yang, Xuefeng Xiao, Huixia Li, Feng Ling, Ruiqi Xie, SiQi Chen, Hongyu Zhu, Zhang Yichong, Yu Wang
Abstract:
In visual generation, the quadratic complexity of attention mechanisms results in high memory and computational costs, especially for longer token sequences required in high-resolution image or multi-frame video generation. To address this, prior research has explored techniques such as sparsification and quantization.However, these techniques face significant challenges under low density and reduced bitwidths. Through systematic analysis, we identify that the core difficulty stems from the dispersed and irregular characteristics of visual attention patterns. Therefore, instead of introducing specialized sparsification and quantization design to accommodate such patterns, we propose an alternative strategy: "reorganizing" the attention pattern to alleviate the challenges.Inspired by the local aggregatin nature of visual feature extraction, we design a novelPattern-Aware tokenReOrdering (PARO) technique, which unifies the diverse attention patterns into a hardware-friendly block-wise pattern. This unification substantially simplifies and enhances both sparsification and quantization.We evaluate the performance-efficiency trade-offs of various design choices and finalize a methodology tailored for the unified pattern.Our approach,PAROAttention, achieves video and image generation with lossless metrics, and nearly identical results from full-precision (FP) baselines, while operating at notably lower density (20%-30%) and bitwidth (INT8/INT4), achieving a1.9 - 2.7xend-to-end latency speedup.



Paperid:5582
Authors:Hao Sun, Huailiang Peng, Qiong Dai, Xu Bai, Yanan Cao
Abstract:
Activation steering is an efficient technique for aligning the behavior of large language models (LLMs) by injecting steering vectors directly into a model’s residual stream during inference.A pivotal challenge in this approach lies in choosing the right layers to intervene, as inappropriate selection can undermine behavioral alignment and even impair the model’s language fluency and other core capabilities.While single-layer steering allows straightforward evaluation on held-out data to identify the “best” layer, it offers only limited alignment improvements.Multi-layer steering promises stronger control but faces a combinatorial explosion of possible layer subsets, making exhaustive search impractical.To address these challenges, we propose LayerNavigator, which provides a principled and promising layer selection strategy.The core innovation of LayerNavigator lies in its novel, quantifiable criterion that evaluates each layer's steerability by jointly considering two key aspects: discriminability and consistency.By reusing the activations computed during steering vector generation, LayerNavigator requires no extra data and adds negligible overhead.Comprehensive experiments show that LayerNavigator achieves not only superior alignment but also greater scalability and interpretability compared to existing strategies.Our code is available at https://anonymous.4open.science/r/LayerNavigator



Paperid:5583
Authors:Cynthia Dwork, Lunjia Hu, Han Shao
Abstract:
We study a fundamental question of domain generalization: given a family of domains (i.e., data distributions), how many randomly sampled domains do we need to collect data from in order to learn a model that performs reasonably well on every seen and unseen domain in the family? We model this problem in the PAC framework and introduce a new combinatorial measure, which we call the domain shattering dimension. We show that this dimension characterizes the domain sample complexity. Furthermore, we establish a tight quantitative relationship between the domain shattering dimension and the classic VC dimension, demonstrating that every hypothesis class that is learnable in the standard PAC setting is also learnable in our setting.



Paperid:5584
Authors:Dmitry Lvov, Ilya Pershin
Abstract:
Predicting human eye movements during goal-directed visual search is critical for enhancing interactive AI systems. In medical imaging, such prediction can support radiologists in interpreting complex data, such as chest X-rays. Existing methods often rely on generic vision–language models and saliency-based features, which limits their ability to capture clinical semantics and integrate domain knowledge effectively. We present \textbf{LogitGaze-Med}, a state-of-the-art multimodal transformer framework that unifies (1) domain-specific visual encoders (e.g., CheXNet), (2) textual embeddings of diagnostic labels, and (3) semantic priors extracted via the logit-lens from an instruction-tuned medical vision–language model (LLaVA-Med). By directly predicting continuous fixation coordinates and dwell durations, our model generates clinically meaningful scanpaths. Experiments on the GazeSearch dataset and synthetic scanpaths generated from MIMIC-CXR demonstrate that LogitGaze-Med improves scanpath similarity metrics by 20–30\% over competitive baselines and yields over 5\% gains in downstream pathology classification when incorporating predicted fixations as additional training data.



Paperid:5585
Authors:Qing Zhao, Weijian Deng, Pengxu Wei, ZiYi Dong, hannan lu, Xiangyang Ji, Liang Lin
Abstract:
To improve detection robustness in adverse conditions (e.g., haze and low light), image restoration is commonly applied as a pre-processing step to enhance image quality for the detector. However, the functional mismatch between restoration and detection networks can introduce instability and hinder effective integration---an issue that remains underexplored. We revisit this limitation through the lens of Lipschitz continuity, analyzing the functional differences between restoration and detection networks in both the input space and the parameter space. Our analysis shows that restoration networks perform smooth, continuous transformations, while object detectors operate with discontinuous decision boundaries, making them highly sensitive to minor perturbations. This mismatch introduces instability in traditional cascade frameworks, where even imperceptible noise from restoration is amplified during detection, disrupting gradient flow and hindering optimization. To address this, we propose Lipschitz-regularized object detection (LROD), a simple yet effective framework that integrates image restoration directly into the detector’s feature learning, harmonizing the Lipschitz continuity of both tasks during training. We implement this framework as Lipschitz-regularized YOLO (LR-YOLO), extending seamlessly to existing YOLO detectors. Extensive experiments on haze and low-light benchmarks demonstrate that LR-YOLO consistently improves detection stability, optimization smoothness, and overall accuracy.



Paperid:5586
Authors:Linda Zeng, Rithwik Gupta, Divij Motwani, Diji Yang, Yi Zhang
Abstract:
Retrieval-augmented generation (RAG) has shown impressive capabilities in mitigating hallucinations in large language models (LLMs). However, LLMs struggle to maintain consistent reasoning when exposed to misleading or conflicting evidence, especially in real-world domains such as politics, where information is polarized or selectively framed. Nevertheless, most RAG benchmarks assume a clean retrieval setting, where models succeed by accurately retrieving and generating answers from gold-standard documents. This assumption fails to align with real-world conditions, leading to an overestimation of RAG system performance. To bridge this gap, we introduce \textsc{RAGuard}, a fact-checking dataset designed to evaluate the robustness of RAG systems against misleading retrievals. Unlike prior benchmarks that rely on synthetic noise, our dataset constructs its retrieval corpus from Reddit discussions, capturing naturally occurring misinformation. It categorizes retrieved evidence into three types: \textit{supporting}, \textit{misleading}, and \textit{irrelevant}, providing a realistic and challenging testbed for assessing how well RAG systems navigate different types of evidence. Our experiments reveal that, when exposed to misleading retrievals, all tested LLM-powered RAG systems perform worse than their zero-shot baselines (i.e., no retrieval at all), while human annotators consistently perform better, highlighting LLMs' susceptibility to noisy environments. To our knowledge, \textsc{RAGuard} is the first benchmark to systematically assess the robustness of the RAG against misleading evidence. We expect this benchmark to drive future research toward improving RAG systems beyond idealized datasets, making them more reliable for real-world applications.



Paperid:5587
Authors:Zhenbang Zhang, Jingtong Feng, Hongjia Li, Haythem El-Messiry, Zhiqiang Xu, Renmin Han
Abstract:
Series Section Electron Microscopy (ssEM) has emerged as a pivotal technology for deciphering nanoscale biological architectures. Three-dimensional (3D) registration is a critical step in ssEM, tasked with rectifying axial misalignments and nonlinear distortions introduced during serial sectioning. The core scientific challenge lies in achieving distortion mitigation without erasing the natural morphological deformations of biological tissues, thereby enabling faithful reconstruction of 3D ultrastructural organization. In this study, we present a paradigm-shifting optimization framework that rethinks 3D registration through the lens of manifold trajectory optimization. We propose the first continuous trajectory dynamics formulation for 3D registration and introduce a novel optimization strategy. Specifically, we introduce a dual optimization objective that inherently balances global trajectory smoothness with local structural preservation, while developing a solver that combines Gauss-Seidel iteration with Neural ODEs to systematically integrate biophysical priors with data-driven deformation compensation. A key strength of our method is its fully unsupervised training, which avoids reliance on ground truth and suits ssEM scenarios where annotations are difficult to obtain. Extensive experiments on multiple datasets spanning diverse tissue types demonstrate our method's superior performance in structural restoration accuracy and cross-tissue robustness.



Paperid:5588
Authors:Lianghong Chen, Dongkyu Kim, Mike Domaratzki, Pingzhao Hu
Abstract:
Designing de novo 3D molecules with desirable properties remains a fundamental challenge in drug discovery and molecular engineering. While diffusion models have demonstrated remarkable capabilities in generating high-quality 3D molecular structures, they often struggle to effectively control complex multi-objective constraints critical for real-world applications. In this study, we propose an uncertainty-aware Reinforcement Learning (RL) framework to guide the optimization of 3D molecular diffusion models toward multiple property objectives while enhancing the overall quality of the generated molecules. Our method leverages surrogate models with predictive uncertainty estimation to dynamically shape reward functions, facilitating balance across multiple optimization objectives. We comprehensively evaluate our framework across three benchmark datasets and multiple diffusion model architectures, consistently outperforming baselines for molecular quality and property optimization. Additionally, Molecular Dynamics (MD) simulations and ADMET profiling of top generated candidates indicate promising drug-like behavior and binding stability, comparable to known Epidermal Growth Factor Receptor (EGFR) inhibitors. Our results demonstrate the strong potential of RL-guided generative diffusion models for advancing automated molecular design.



Paperid:5589
Authors:Jiangtong Li, Dongyi Liu, Kun Zhu, Dawei Cheng, changjun jiang
Abstract:
Graph Neural Networks (GNNs) have achieved significant success in various real-world applications, including social networks, finance systems, and traffic management. Recent researches highlight their vulnerability to backdoor attacks in node classification, where GNNs trained on a poisoned graph misclassify a test node only when specific triggers are attached. These studies typically focus on single attack categories and use adaptive trigger generators to create node-specific triggers. However, adaptive trigger generators typically have a simple structure, limited parameters, and lack category-aware graph knowledge, which makes them struggle to handle backdoor attacks across multiple categories as the number of target categories increases. We address this gap by proposing a novel approach for Effective and Unnoticeable Multi-Category (EUMC) graph backdoor attacks, leveraging subgraph from the attacked graph as category-aware triggers to precisely control the target category. To ensure the effectiveness of our method, we construct a Multi-Category Subgraph Triggers Pool (MC-STP) using the subgraphs of the attacked graph as triggers. We then exploit the attachment probability shifts of each subgraph trigger as category-aware priors for target category determination. Moreover, we develop a ``select then attach'' strategy that connects suitable category-aware trigger to attacked nodes for unnoticeability. Extensive experiments across different real-world datasets confirm the efficacy of our method in conducting multi-category graph backdoor attacks on various GNN models and defense strategies.



Paperid:5590
Authors:Pengbo Wang, Chaozhuo Li, Chenxu Wang, Liwen Zheng, Litian Zhang, Xi Zhang
Abstract:
LLMs have demonstrated unprecedented capabilities in natural language processing, yet their practical deployment remains hindered by persistent factuality and faithfulness hallucinations. While existing methods address these hallucination types independently, they inadvertently induce performance trade-offs, as interventions targeting one type often exacerbate the other. Through empirical and theoretical analysis of activation space dynamics in LLMs, we reveal that these hallucination categories share overlapping subspaces within neural representations, presenting an opportunity for concurrent mitigation. To harness this insight, we propose SPACE, a unified framework that jointly enhances factuality and faithfulness by editing shared activation subspaces. SPACE establishes a geometric foundation for shared subspace existence through dual-task feature modeling, then identifies and edits these subspaces via a hybrid probe strategy combining spectral clustering and attention head saliency scoring. Experimental results across multiple benchmark datasets demonstrate the superiority of our approach.



Paperid:5591
Authors:Zhaoliang Wan, Zetong Bi, Zida Zhou, Hao Ren, Yiming Zeng, Yihan Li, Lu Qi, Xu Yang, Ming-Hsuan Yang, Hui Cheng
Abstract:
This paper addresses the scarcity of low-cost but high-dexterity platforms for collecting real-world multi-fingered robot manipulation data towards generalist robot autonomy. To achieve it, we propose the RAPID Hand, a co-optimized hardware and software platform where the compact 20-DoF hand, robust whole-hand perception, and high-DoF teleoperation interface are jointly designed.Specifically, we introduce RAPID Hand, a robust, affordable, and perception-integrated 20-DoF robotic hand co-designed to meet both data and algorithmic requirements for training general-purpose manipulation policies. Specifically, RAPID Hand adopts a compact and practical hand ontology and a hardware-level perception framework that stably integrates wrist-mounted vision, fingertip tactile sensing, and proprioception with sub-7 ms latency and spatial alignment. Collecting high-quality demonstrations on high-DoF hands is challenging, as existing teleoperation methods struggle with precision and stability on complex multi-fingered systems. Collecting high-quality demonstrations on high-DoF hands is challenging. Existing teleoperation methods, primarily designed for under-actuated or low-DoF hands, struggle with precision and stability when applied to complex multi-fingered systems. We address this by co-optimizing hand design, perception integration, and teleoperation interface through a universal actuation scheme, custom perception electronics, and two retargeting constraints. We evaluate the platform’s hardware, perception, and teleoperation interface. Training a diffusion policy on collected data shows superior performance over prior works, validating the system’s capability for reliable, high-quality data collection. The platform is constructed from low-cost and off-the-shelf components and will be made public to ensure reproducibility and ease of adoption.



Paperid:5592
Authors:Sohini Saha, Mezisashe Ojuba, Leslie Collins, Boyla Mainsah
Abstract:
Mixture-of-Experts (MoE) models have become popular in machine learning, boosting performance by partitioning tasks across multiple experts. However, the need for several experts often results in high computational costs, limiting their application on resource-constrained devices with stringent real-time requirements, such as cochlear implants (CIs). We introduce the Omni-Expert (OE) – a simple and efficient solution that leverages feature transformations to achieve the 'divide-and-conquer' functionality of a full MoE ensemble in a single expert model. We demonstrate the effectiveness of the OE using phoneme-specific time-frequency masking for speech dereverberation in a CI. Empirical results show that the OE delivers statistically significant improvements in objective intelligibility measures of CI vocoded speech at different levels of reverberation across various speech datasets at a much reduced computational cost relative to a counterpart MoE.



Paperid:5593
Authors:Fayi Le, Wenwu He, Chentao Cao, Dong Liang, Zhuo-Xu Cui
Abstract:
Pretrained vision-language models (VLMs), such as CLIP, have shown promising zero-shot out-of-distribution (OOD) detection capabilities by leveraging semantic similarities between input images and textual labels. However, most existing approaches focus solely on expanding the label space in the text domain, ignoring complementary visual cues that can further enhance discriminative power. In this paper, we introduce DualCnst, a novel framework that integrates text-image dual consistency for improved zero-shot OOD detection. Specifically, we generate synthetic images from both ID and mined OOD textual labels using a text-to-image generative model, and jointly evaluate each test image based on (i) its semantic similarity to class labels and (ii) its visual similarity to the synthesized images. The resulting unified score function effectively combines multimodal information without requiring access to in-distribution images or additional training.We further provide theoretical analysis showing that incorporating multimodal negative labels reduces score variance and improves OOD separability. Extensive experiments across diverse OOD benchmarks demonstrate that DualCnst achieves state-of-the-art performance while remaining scalable, data-agnostic, and fully compatible with prior text-only VLM-based methods.



Paperid:5594
Authors:Changhan Liu, xunzhi xiang, Zixuan Duan, Wenbin Li, Yang Gao, Qi Fan
Abstract:
Cross-Domain Few-Shot Object Detection (CD-FSOD) seeks to generalize to unseen domains using only a limited number of annotated samples from the target domain. This requires models to possess both strong generalization and regression capabilities. However, existing well-trained detection models and vision foundation models (VFMs) are typically limited, excelling in either localization or generalization, but rarely both, making them unsuitable for the CD-FSOD task. In this paper, we propose a novel CD-FSOD paradigm that retains the structure and localization capabilities of well-trained object detection models while achieving the transfer of generalization abilities from VFMs with minimal additional parameters. Specifically, our proposed Mixture-of-Experts (MoE) based method dynamically routes and modulates VFM features as representation experts to facilitate traditional object detection models for CD-FSOD. Furthermore, our approach eliminates the traditional two-stage paradigm of base training followed by novel fine-tuning. Instead, it extends well-trained object detection models to detect novel classes in unseen domains without requiring re-training on the base data. Experimental results on multiple cross-domain datasets validate the effectiveness of our method.



Paperid:5595
Authors:Martin Marek, Sanae Lotfi, Aditya Somasundaram, Andrew Wilson, Micah Goldblum
Abstract:
Conventional wisdom dictates that small batch sizes make language model pretraining unstable, motivating gradient accumulation which trades off the number of optimizer steps for a proportional increase in batch size. While it is common to decrease learning rate for smaller batch sizes, other hyperparameters are often held fixed. In this work, we revisit small batch sizes, all the way down to batch size one, with appropriate hyperparameters. We find that small batch sizes (1) train stably, (2) are consistently more robust to hyperparameter choices, (3) achieve equal or better per-FLOP performance as larger batch sizes, and (4) notably enable stable language model training with vanilla SGD, even without momentum, despite storing no optimizer state. Building on these results, we provide practical recommendations for selecting a batch size and setting optimizer hyperparameters. We further recommend against gradient accumulation in settings where throughput is not bound by inter-node bandwidth.



Paperid:5596
Authors:Finn Schmidt, Polina Turishcheva, Suhas Shrinivasan, Fabian Sinz
Abstract:
The neural activity in the visual processing is influenced by both external stimuli and internal brain states. Ideally, a neural predictive model should account for both of them.Currently, there are no dynamic encoding models that explicitly model a latent state and the entire neuronal response distribution.We address this gap by proposing a probabilistic model that predicts the joint distribution of the neuronal responses from video stimuli and stimulus-independent latent factors. After training and testing our model on mouse V1 neuronal responses, we find that it outperforms video-only models in terms of log-likelihood and achieves improvements in likelihood and correlation when conditioned on responses from other neurons. Furthermore, we find that the learned latent factors strongly correlate with mouse behavior and that they exhibit patterns related to the neurons’ position on the visual cortex, although the model was trained without behavior and cortical coordinates. Our findings demonstrate that unsupervised learning of latent factors from population responses can reveal biologically meaningful structure that bridges sensory processing and behavior, without requiring explicit behavioral annotations during training.The code is attached to the submission.



Paperid:5597
Authors:Yuxiang Li, Yang Zhang, Li, Mengxuan Chen, Meng Jin, Fang Wang, Haohuan Fu, Juepeng Zheng
Abstract:
Abstract:Modern precipitation forecasting systems—including reanalysis datasets, numerical models, and AI-based approaches—typically produce coarse-resolution gridded outputs that often exhibit substantial spatial biases relative to station-level observations, especially in complex terrains or under extreme conditions. These biases stem from two core challenges: (i) $\textbf{station-level heterogeneity}$, with site-specific temporal and spatial dynamics; and (ii) $\textbf{oversmoothing}$, which blurs fine-scale variability in graph-based models. To address these issues, we propose $\textbf{DiffLiG}$ ($\underline{Diff}$usion-enhanced $\underline{Li}$quid $\underline{G}$raph with Attention Propagation), a graph neural network designed for precise spatial correction from gridded forecasts to station observations. DiffLiG integrates a GeoLiquidNet that adapts temporal encoding via site-aware OU dynamics, a graph neural network with a dynamic edge modulator that learns spatially adaptive connectivity, and a Probabilistic Diffusion Selector that generates and refines ensemble forecasts to mitigate oversmoothing. Experiments across multiple datasets show that DiffLiG consistently outperforms other methods, delivering more accurate and robust corrections across diverse geographic and climatic settings. Moreover, it achieves notable gains on other key meteorological variables, underscoring its generalizability and practical utility.



Paperid:5598
Authors:Suhas BN, Andrew Sherrill, Rosa I. Arriaga, Christopher Wiese, Saeed Abdullah
Abstract:
The advancement of AI systems for mental health support is hindered by limited access to therapeutic conversation data, particularly for trauma treatment. We present Thousand Voices of Trauma, a synthetic benchmark dataset of 3,000 therapy conversations based on Prolonged Exposure therapy protocols for Post-traumatic Stress Disorder (PTSD). The dataset comprises 500 unique cases, each explored through six conversational perspectives that mirror the progression of therapy from initial anxiety to peak distress to emotional processing. We incorporated diverse demographic profiles (ages 18-80, M=49.3, 49.4\% male, 44.4\% female, 6.2\% non-binary), 20 trauma types, and 10 trauma-related behaviors using deterministic and probabilistic generation methods. Analysis reveals realistic distributions of trauma types (witnessing violence 10.6\%, bullying 10.2\%) and symptoms (nightmares 23.4\%, substance abuse 20.8\%). Clinical experts validated the dataset's therapeutic fidelity, highlighting its emotional depth while suggesting refinements for greater authenticity. We also developed an emotional trajectory benchmark with standardized metrics for evaluating model responses. This privacy-preserving dataset addresses critical gaps in trauma-focused mental health data, offering a valuable resource for advancing both patient-facing applications and clinician training tools.



Paperid:5599
Authors:Tom Burgert, Oliver Stoll, Paolo Rota, Begum Demir
Abstract:
The hypothesis that Convolutional Neural Networks (CNNs) are inherently texture-biased has shaped much of the discourse on feature use in deep learning. We revisit this hypothesis by examining limitations in the cue-conflict experiment by Geirhos et al. To address these limitations, we propose a domain-agnostic framework that quantifies feature reliance through systematic suppression of shape, texture, and color cues, avoiding the confounds of forced-choice conflicts. By evaluating humans and neural networks under controlled suppression conditions, we find that CNNs are not inherently texture-biased but predominantly rely on local shape features. Nonetheless, this reliance can be substantially mitigated through modern training strategies or architectures (ConvNeXt, ViTs). We further extend the analysis across computer vision, medical imaging, and remote sensing, revealing that reliance patterns differ systematically: computer vision models prioritize shape, medical imaging models emphasize color, and remote sensing models exhibit a stronger reliance towards texture.



Paperid:5600
Authors:Riley Simmons-Edler, Ryan Badman, Felix Berg, Raymond Chua, John Vastola, Joshua Lunger, William Qian, Kanaka Rajan
Abstract:
Understanding the behavior of deep reinforcement learning (DRL) agents—particularly as task and agent sophistication increase—requires more than simple comparison of reward curves, yet standard methods for behavioral analysis remain underdeveloped in deep RL.We apply tools from neuroscience and ethology to study DRL agents in a novel, complex, partially observable environment, ForageWorld, designed to capture key aspects of real-world animal foraging—including sparse, depleting resource patches, predator threats, and spatially extended arenas.We use this environment as a platform for applying joint behavioral and neural analysis to agents, revealing detailed, quantitatively grounded insights into agent strategies, memory, and planning.Contrary to common assumptions, we find that model-free RNN-based DRL agents can exhibit structured, planning-like behavior purely through emergent dynamics—without requiring explicit memory modules or world models.Our results show that studying DRL agents like animals—analyzing them with neuroethology-inspired tools that reveal structure in both behavior and neural dynamics—uncovers rich structure in their learning dynamics that would otherwise remain invisible.We distill these tools into a general analysis framework that links core behavioral and representational features to diagnostic methods, which can be reused for a wide range of tasks.As agents grow more complex and autonomous, bridging neuroscience, cognitive science, and AI will be essential—not just for understanding their behavior, but for ensuring safe alignment and maximizing desirable behaviors that are hard to measure via reward.We show how this can be done by drawing on lessons from how biological intelligence is studied.



Paperid:5601
Authors:Mengxuan Chen, Li, Zou Ziheng, Fang Wang, Jinxiao Zhang, Runmin Dong, Juepeng Zheng, Haohuan Fu
Abstract:
Seasonal-scale climate prediction is of great importance for agricultural planning, disaster prevention, and decision-making. In particular, extreme precipitation and droughts during the summer monsoon season in East Asia pose significant risks, highlighting the urgent need for reliable forecasts 1-6 months in advance. However, compared to weather forecasting and S2S prediction, seasonal prediction has received limited attention in the AI community. Current approaches primarily rely on ensemble forecasts from numerical prediction systems, which suffer from systematic biases and considerable uncertainty. Motivated by these challenges, we propose SeasonBench-EA, a seasonal prediction benchmark for the East Asia region, featuring multi-resolution and multi-source data with both regional and global coverage. The dataset includes ERA5 reanalysis and ensemble forecasts from several major meteorological agencies. In addition to key atmospheric variables, it also incorporates boundary layer variables, such as ocean, sea ice, soil and solar radiation, which are closely related to seasonal-scale atmospheric processes. The data are provided at hourly, daily and monthly temporal resolutions. Two core tasks are defined and evaluated: 1) machine learning-based prediction using ERA5 data, and 2) post-processing and correction of ensemble forecasts. The benchmark also includes a variety of deterministic and probabilistic evaluation metrics, as well as a hindcast assessment of precipitation prediction during the East Asian summer monsoon, aligned with operational evaluation settings. Overall, SeasonBench-EA aims to provide a systematic data and evaluation framework for seasonal prediction, promoting the development and application of AI methods in this domain.



Paperid:5602
Authors:jingxiao zhang, Shifei Ding, Jian Zhang, Lili Guo, Xuan Li
Abstract:
Graph Neural Networks (GNNs) are powerful models for node classification, but their performance is heavily reliant on manually labeled data, which is often costly and results in insufficient labeling. Recent studies have shown that message-passing neural networks struggle to propagate information in low-degree nodes, negatively affecting overall performance. To address the information bias caused by degree imbalance, we propose a Learnable Enhancement and Label Selection Dynamic Graph Convolutional Network (L2DGCN). L2DGCN consists of a teacher model and a student model. The teacher model employs an improved label propagation mechanism that enables remote label information dissemination among all nodes. The student model introduces a dynamically learnable graph enhancement strategy, perturbing edges to facilitate information exchange among low-degree nodes. This approach maintains the global graph structure while learning graph representations. Additionally, we have designed a label selector to mitigate the impact of unreliable pseudo-labels on model learning. To validate the effectiveness of our proposed model with limited labeled data, we conducted comprehensive evaluations of semi-supervised node classification across various scenarios with a limited number of annotated nodes. Experimental results demonstrate that our data enhancement model significantly contributes to node classification tasks under sparse labeling conditions.



Paperid:5603
Authors:Adrien Vacher, Omar Chehab, Anna Korba
Abstract:
Sampling from multi-modal distributions is challenging, even in low dimensions. We provide the first sampling algorithm for a broad class of distributions --- including all Gaussian mixtures --- with a query complexity that is polynomial in the parameters governing multi-modality, assuming fixed dimension. Our sampling algorithm simulates a time-reversed diffusion process, using a specific Monte Carlo estimator of the intermediate score functions. Unlike previous works, it avoids metastability, requires no prior knowledge of the mode locations, and does not rely on restrictive smoothness assumptions that exclude general Gaussian mixtures. We illustrate this result on a low-dimensional but challenging multi-modal sampling task, where our algorithm significantly outperforms existing approaches.



Paperid:5604
Authors:Haoming Ye, Yunxiao Xiao, Cewu Lu, Panpan Cai
Abstract:
Robotic task planning in real-world environments requires reasoning over implicit constraints from language and vision. While LLMs and VLMs offer strong priors, they struggle with long-horizon structure and symbolic grounding. Existing meth-ods that combine LLMs with symbolic planning often rely on handcrafted or narrow domains, limiting generalization. We propose UniDomain, a framework that pre-trains a PDDL domain from robot manipulation demonstrations and applies it for online robotic task planning. It extracts atomic domains from 12,393 manipulation videos to form a unified domain with 3137 operators, 2875 predicates, and 16481 causal edges. Given a target class of tasks, it retrieves relevant atomics from the unified domain and systematically fuses them into high-quality meta-domains for zero-shot planning. Experiments on diverse real-world tasks show that UniDomain solves complex, unseen tasks in a zero-shot manner, achieving up to 58% higher task success and 160% improvement in plan optimality over state-of-the-art LLM and LLM-PDDL baselines.



Paperid:5605
Authors:Yufan Li, Pragya Sur
Abstract:
Abstract:We study the fundamental problem of calibrating a linear binary classifier of the form \(\sigma(\hat{w}^\top x)\), where the feature vector \(x\) is Gaussian, \(\sigma\) is a link function, and \(\hat{w}\) is an estimator of the true linear weight $w^\star$. By interpolating with a noninformative \emph{chance classifier}, we construct a well-calibrated predictor whose interpolation weight depends on the angle \(\angle(\hat{w}, w_\star)\) between the estimator \(\hat{w}\) and the true linear weight \(w_\star\). We establish that this angular calibration approach is provably well-calibrated in a high-dimensional regime where the number of samples and features both diverge, at a comparable rate. The angle \(\angle(\hat{w}, w_\star)\) can be consistently estimated. Furthermore, the resulting predictor is uniquely \emph{Bregman-optimal}, minimizing the Bregman divergence to the true label distribution within a suitable class of calibrated predictors.Our work is the first to provide a calibration strategy that satisfies both calibration and optimality properties provably in high dimensions. Additionally, we identify conditions under which a classical Platt-scaling predictor converges to our Bregman-optimal calibrated solution. Thus, Platt-scaling also inherits these desirable properties provably in high dimensions.



Paperid:5606
Authors:Jianze Li, Jiezhang Cao, Zichen Zou, Xiongfei Su, Xin Yuan, Yulun Zhang, Yong Guo, Xiaokang Yang
Abstract:
Abstract:Diffusion models have demonstrated excellent performance for real-world image super-resolution (Real-ISR), albeit at high computational costs. Most existing methods are trying to derive one-step diffusion models from multi-step counterparts through knowledge distillation (KD) or variational score distillation (VSD). However, these methods are limited by the capabilities of the teacher model, especially if the teacher model itself is not sufficiently strong. To tackle these issues, we propose a new One-Step \textbf{D}iffusion model with a larger-scale \textbf{D}iffusion \textbf{D}iscriminator for SR, called D$^3$SR. Our discriminator is able to distill noisy features from any time step of diffusion models in the latent space. In this way, our diffusion discriminator breaks through the potential limitations imposed by the presence of a teacher model. Additionally, we improve the perceptual loss with edge-aware DISTS (EA-DISTS) to enhance the model's ability to generate fine details. Our experiments demonstrate that, compared with previous diffusion-based methods requiring dozens or even hundreds of steps, our D$^3$SR attains comparable or even superior results in both quantitative metrics and qualitative evaluations. Moreover, compared with other methods, D$^3$SR achieves at least $3\times$ faster inference speed and reduces parameters by at least 30\%.



Paperid:5607
Authors:Junyi Wang, Yuze Wang, Wantong Duan, Meng Wang, Yue Qi
Abstract:
Visual localization is a critical component across various domains. The recent emergence of novel scene representations, such as 3D Gaussian Splatting (3D GS), introduces new opportunities for advancing localization pipelines. In this paper, we propose a novel 3D GS-based framework for RGB-based, scene-independent camera relocalization, with three main contributions. First, we design a two-stage pipeline with fully exploiting 3D GS. The pipeline consists of an initial stage, which utilizes 2D-3D correspondences between image pixels and 3D Gaussians, followed by pose refinement using the rendered image by 3D GS. Second, we introduce a 3D GS-based Relocalization Network, termed GS-RelocNet, to establish correspondences for initial camera pose estimation. Additionally, we present a refinement network that further optimizes the camera pose. Third, we propose a unidirectional 2D-3D feature fusion module and a bidirectional image feature fusion module, integrated into GS-RelocNet and the refinement network, respectively, to enhance feature sharing across the two stages. Experimental results on public 7 Scenes and Cambridge Landmarks demonstrate state-of-the-art performance. Furthermore, the beneficial effects of the two feature fusion modules and pose refinement are also highlighted. In summary, we believe that the proposed framework can be a novel universal localization pipeline for further research.



Paperid:5608
Authors:Yiming Wang, Qun Li, Dongxia Chang, Jie Wen, Hua Dai, Fu Xiao, Yao Zhao
Abstract:
Abstract:Multi-view clustering aims to enhance clustering performance by leveraging information from diverse sources. However, existing methods typically assume that instances are present in most or all views, which is impractical in real-world scenarios. This paper focuses on the understudied problem of fully incomplete multi-view clustering (FIMC), where each instance contains only a single view or cross-view correspondences are entirely unknown. To address this problem, we propose a Contrastive Prototype Matching Network (CPMN), which pioneers an explicit prototype matching strategy to effectively align cross-view instances through category-level correspondence. CPMN first employs a correspondence-free graph contrastive learning approach, leveraging mutual $k$-nearest neighbors (MNN) to uncover structural correlations and establish initial view-specific prototypes from entirely unpaired views. Building on the prototypes, we introduce a cross-view prototype graph matching stage to resolve category misalignment and forge a unified clustering structure. Finally, guided by this alignment, we devise a prototype-aware contrastive learning mechanism to promote semantic consistency, replacing the reliance on the initial MNN-based structural similarity. Extensive experiments on benchmark datasets demonstrate that our method significantly outperforms various baselines and ablation variants, validating its effectiveness.



Paperid:5609
Authors:Qinglun Li, Yingqi Liu, Miao Zhang, Xiaochun Cao, Quanjun Yin, Li Shen
Abstract:
Abstract:Decentralized training removes the centralized server, making it a communication-efficient approach that can significantly improve training efficiency, but it often suffers from degraded performance compared to centralized training.Multi-Gossip Steps (MGS) serve as a simple yet effective bridge between decentralized and centralized training, significantly reducing experiment performance gaps. However, the theoretical reasons for its effectiveness and whether this gap can be fully eliminated by MGS remain open questions.In this paper, we derive upper bounds on the generalization error and excess error of MGS using stability analysis, systematically answering these two key questions.1). Optimization Error Reduction: MGS reduces the optimization error bound at an exponential rate, thereby exponentially tightening the generalization error bound and enabling convergence to better solutions.2). Gap to Centralization: Even as MGS approaches infinity, a non-negligible gap in generalization error remains compared to centralized mini-batch SGD ($\mathcal{O}(T^{\frac{c\beta}{c\beta +1}}/{n m})$ in centralized and $\mathcal{O}(T^{\frac{2c\beta}{2c\beta +2}}/{n m^{\frac{1}{2c\beta +2}}})$ in decentralized).Furthermore, we provide the first unified analysis of how factors like learning rate, data heterogeneity, node count, per-node sample size, and communication topology impact the generalization of MGS under non-convex settings without the bounded gradients assumption, filling a critical theoretical gap in decentralized training. Finally, promising experiments on CIFAR datasets support our theoretical findings.



Paperid:5610
Authors:Honglin Li, Zhongyi Shui, Yunlong Zhang, Chenglu Zhu, Lin Yang
Abstract:
Abstract:Pathology whole slide image (WSI) analysis is vital for disease diagnosis and understanding. While foundation models (FMs) have driven recent advances, their scalability in pathology remains a key challenge. In particular, vision-language (VL) pathology FMs align visual features with language annotation for downstream tasks, but they rely heavily on large-scale image-text paired data, which is scarce thus limiting generalization. On the other hand, vision-only pathology FMs can leverage abundant unlabeled data via self-supervised learning (SSL). However, current approaches often use the [CLS] token from tile-level ViTs as slide-level input for efficiency (a tile with 224×224 pixels composed of 196 patches with 16×16 pixels). This SSL pretrained [CLS] token lacks alignment with downstream objectives, limiting effectiveness. We find that spatial patch tokens retain a wealth of informative features beneficial for downstream tasks, but utilizing all of them incurs up to 200× higher computation and storage costs compared [CLS] token only (e.g., 196 tokens per ViT$_{224}$). This highlights a fundamental trade-off between efficiency and representational richness to build scalable pathology FMs. To address this, we propose a feature distillation framework via vector-quantization (VQ) that compresses patch tokens into discrete indices and reconstructs them via a decoder, achieving 64× compression (1024 → 16 dimensions) while preserving fidelity. We further introduce a multi-scale VQ (MSVQ) strategy, enhancing both reconstruction and providing SSL supervision for slide-level pretraining. Built upon MSVQ features and supervision signals, we design a progressive convolutional module and a slide-level SSL objective to learn spatially rich representations for downstream WSI tasks. Extensive experiments across multiple datasets demonstrate that our approach achieves state-of-the-art performance, offering a scalable and effective solution for high-performing pathology FMs in WSI analysis.



Paperid:5611
Authors:Mingzhi Zhu, Ding Shang, Sai Qian Zhang
Abstract:
Abstract:Photorealistic Codec Avatars (PCA), which enable high-fidelity human face rendering, are increasingly adopted in AR/VR applications to support immersive communication and interaction via deep learning-based generative models. However, these models impose significant computational demands, making real-time inference challenging on resource-constrained AR/VR devices such as head-mounted displays (HMDs), where latency and power efficiency are critical. To address this challenge, we propose an efficient post-training quantization (PTQ) method tailored for Codec Avatar models, enabling low-precision execution without compromising output quality. In addition, we design a custom hardware accelerator that can be integrated into the system-on-chip (SoC) of AR/VR devices to further enhance processing efficiency. Building on these components, we introduce~\textit{ESCA}, a full-stack optimization framework that accelerates PCA inference on edge AR/VR platforms. Experimental results demonstrate that ESCA boosts FovVideoVDP quality scores by up to $+0.39$ over the best 4-bit baseline, delivers up to $3.36\times$ latency reduction, and sustains a rendering rate of 100 frames per second in end-to-end tests, satisfying real-time VR requirements. These results demonstrate the feasibility of deploying high-fidelity codec avatars on resource-constrained devices, opening the door to more immersive and portable VR experiences.



Paperid:5612
Authors:xichen xu, Yanshu Wang, Jinbao Wang, XiaoNing Lei, Guoyang Xie, GUANNAN JIANG, Zhichao Lu
Abstract:
Industrial anomaly segmentation relies heavily on pixel-level annotations, yet real-world anomalies are often scarce, diverse, and costly to label. Segmentation-oriented industrial anomaly synthesis (SIAS) has emerged as a promising alternative; however, existing methods struggle to balance sampling efficiency and generation quality. Moreover, most approaches treat all spatial regions uniformly, overlooking the distinct statistical differences between anomaly and background areas. This uniform treatment hinders the synthesis of controllable, structure-specific anomalies tailored for segmentation tasks. In this paper, we propose FAST, a foreground-aware diffusion framework featuring two novel modules: the Anomaly-Informed Accelerated Sampling (AIAS) and the Foreground-Aware Reconstruction Module (FARM). AIAS is a training-free sampling algorithm specifically designed for segmentation-oriented industrial anomaly synthesis, which accelerates the reverse process through coarse-to-fine aggregation and enables the synthesis of state-of-the-art segmentation-oriented anomalies in as few as 10 steps. Meanwhile, FARM adaptively adjusts the anomaly-aware noise within the masked foreground regions at each sampling step, preserving localized anomaly signals throughout the denoising trajectory. Extensive experiments on multiple industrial benchmarks demonstrate that FAST consistently outperforms existing anomaly synthesis methods in downstream segmentation tasks. We release the code in: https://anonymous.4open.science/r/NeurIPS-938.



Paperid:5613
Authors:Xiaolong Xu, Xudong Zhao, Haolong Xiang, Xuyun Zhang, Wei Shen, Hongsheng Hu, Lianyong Qi
Abstract:
The long-tail problem in sequential recommender systems stems from imbalanced interaction data, resulting in suboptimal model performance for tail users and items. Recent studies have leveraged head data to enhance tail data for diminish the impact of the long-tail problem. However, these methods often adopt ad-hoc strategies to distinguish between head and tail data, which fails to capture the underlying distributional characteristics and structural properties of each category. Moreover, due to a substantial representational gap exists between head and tail data, head-to-tail enhancement strategies are susceptible to negative transfer, often leading to a decline in overall model performance. To address these issues, we propose a hierarchical partitioning and stepwise enhancement framework, called HPSERec, for long-tailed sequential recommendation. HPSERec partitions the item set into subsets based on a data imbalance metric, assigning an expert network to each subset to capture user-specific local features. Subsequently, we apply knowledge distillation to progressively improve long-tail interest representation, followed by a Sinkhorn optimal transport-based feedback module, which aligns user representations across expert levels through a globally optimal and softly matched mapping. Extensive experiments on three real-world datasets demonstrate that HPSERec consistently outperforms all baseline methods. The implementation code is available at https://anonymous.4open.science/r/HPSERec-2404.



Paperid:5614
Authors:Xiaolei Wang, Tianhong Dai, Huihui Bai, Yao Zhao, Jimin XIAO
Abstract:
Due to the difficulty in collecting all unexpected abnormal patterns, One-Class Classification (OCC) has become the most popular approach to anomaly detection (AD). Reconstruction-based AD method relies on the discrepancy between inputs and reconstructed results to identify unobserved anomalies. However, recent methods trained only on normal samples may generalize to certain abnormal inputs, leading to well-reconstructed anomalies and degraded performance. To address this, we constrain reconstructions to remain on the normal manifold using a novel AD framework based on contraction mapping. This mapping guarantees that any input converges to a fixed point through iterations of this mapping. Based on this property, training the contraction mapping using only normal data ensures that its fixed point lies within the normal manifold. As a result, abnormal inputs are iteratively transformed toward the normal manifold, increasing the reconstruction error. In addition, the inherent invertibility of contraction mapping enables flow-based density estimation, where a prior distribution learned from the previous reconstruction is used to estimate the input likelihood for anomaly detection, further improving the performance. Using both mechanisms, we propose a bidirectional structure with forward reconstruction and backward density estimation. Extensive experiments on tabular data, natural image, and industrial image data demonstrate the effectiveness of our method. The code will be made publicly available.



Paperid:5615
Authors:Rahul Vaze, Abhishek Sinha
Abstract:
Abstract:The constrained version of the standard online convex optimization (OCO) framework, called COCO is considered, where on every round, a convex cost function and a convex constraint function are revealed to the learner after it chooses the action for that round.The objective is to simultaneously minimize the static regret and cumulative constraint violation (CCV). An algorithm is proposed that guarantees a static regret of $O(\sqrt{T})$ and a CCV of $\min\{{\cal V}, O(\sqrt{T}\log T) \}$, where ${\cal V}$ depends on the distance between the consecutively revealed constraint sets, the shape of constraint sets, dimension of action space and the diameter of the action space. When constraint sets have additional structure, ${\cal V}=O(1)$. Compared to the state of the art results, static regret of $O(\sqrt{T})$ and CCV of $O(\sqrt{T}\log T)$, that were universal, the new result on CCV is instance dependent, which is derived by exploiting the geometric properties of the constraint sets.



Paperid:5616
Authors:Andreas Kontogiannis, Vasilis Pollatos, Gabriele Farina, Panayotis Mertikopoulos, Ioannis Panageas
Abstract:
We examine the problem of efficiently learning coarse correlated equilibria (CCE) in polyhedral games, that is, normal-form games with an exponentially large number of actions per player and an underlying combinatorial structure. Prominent examples of such games are the classical Colonel Blotto and congestion games. To achieve computational efficiency, the learning algorithms must exhibit regret and per iteration complexity that scale polylogarithmically in the size of the players' action sets. This challenge has recently been addressed in the full-information setting, primarily through the use of kernelization. However, in the case of the realistic, but mathematically challenging, partial-information setting, existing approaches result in suboptimal and impractical runtime complexity to learn CCE. We tackle this limitation by building a framework based on the kernelization paradigm.We apply this framework to prominent examples of polyhedral games---namely the Colonel Blotto, graphic matroid and network congestion games --- and provide computationally efficient payoff-based learning algorithms, which significantly improve upon prior works in terms of the runtime for learning CCE in these settings.



Paperid:5617
Authors:Shiyuan Zuo, Xingrun Yan, Rongfei Fan, Li Shen, Puning Zhao, Jie Xu, Han Hu
Abstract:
Abstract:Federated Learning (FL) enables multiple clients to collaboratively train models without sharing raw data, but is vulnerable to Byzantine attacks and data heterogeneity, which can severely degrade performance. Existing Byzantine-robust approaches tackle data heterogeneity, but incur high computational overhead during gradient aggregation, thereby slowing down the training process. To address this issue, we propose a simple yet effective Federated Normalized Gradients Algorithm (Fed-NGA), which performs aggregation by merely computing the weighted mean of the normalized gradients from each client. This approach yields a favorable time complexity of $\mathcal{O}(pM)$, where $p$ is the model dimension and $M$ is the number of clients. We rigorously prove that Fed-NGA is robust to both Byzantine faults and data heterogeneity. For non-convex loss functions, Fed-NGA achieves convergence to a neighborhood of stationary points under general assumptions, and further attains zero optimality gap under some mild conditions, which is an outcome rarely achieved in existing literature. In both cases, the convergence rate is $\mathcal{O}(1/T^{\frac{1}{2} - \delta})$, where $T$ denotes the number of iterations and $\delta \in (0, 1/2)$. Experimental results on benchmark datasets confirm the superior time efficiency and convergence performance of Fed-NGA over existing methods.



Paperid:5618
Authors:Ziliang Chen, Yongsen Zheng, Zhao-Rong Lai, Zhanfu Yang, Cuixi Li, Yang Liu, Liang Lin
Abstract:
Instruction-Tuning (IT) was recently found the impressive data efficiency in post-training large language models (LLMs). While the pursuit of efficiency predominantly focuses on sequence-level curation, often overlooking the nuanced impact of critical tokens and the inherent risks of token noise and biases. Drawing inspiration from bi-level coreset selection, our work provides the principled view of the motivation behind selecting instructions' responses. It leads to our approach Quadratic Coreset Selection (QCS) that reconciles sequence-level and token-level influence contributions, deriving more expressive LLMs with established theoretical result. Despite the principled QCS framework challenged by prohibitive computational costs from inverting LLM-scale Hessian matrices, we overcome this barrier by proposing a novel QCS probabilistic solver, which relaxes the original formulation through re-parameterized densities. This innovative solver is efficiently learned using hierarchical policy gradients without requiring back-propagation, achieving provable convergence and certified asymptotic equivalence to the original objective. Our experiments demonstrate QCS's superior sequence-level data efficiency and reveal how strategically leveraging token-level influence elevates the performance ceiling of data-efficient IT. Furthermore, QCS's adaptability is showcased through its successful application in challenging targeted IT and continual IT scenarios, underscoring its potential for a wide array of versatile post-training applications.



Paperid:5619
Authors:Chen Chen, Majid Abdolshah, Violetta Shevchenko, Hongdong Li, Chang Xu, Pulak Purkait
Abstract:
Existing Stable-Diffusion-based super-resolution approaches often exhibit semantic ambiguities due to inaccuracies and incompleteness in their text prompts, coupled with the inherent tendency for cross-attention to divert towards irrelevant pixels. To address these, we propose a novel, plug-and-play spatially re-focused super-resolution (SRSR) framework, which refines text conditioning at inference time by applying visually-grounded segmentation masks to guide cross-attention and selectively bypassing text influences on ungrounded pixels to prevent hallucinations. Extensive experiments on both synthetic and real-world datasets demonstrate that SRSR consistently outperforms seven state-of-the-art baselines in standard fidelity metrics (PSNR and SSIM) across all datasets, and in perceptual quality measures (LPIPS and DISTS) on two real-world benchmarks, underscoring its effectiveness in achieving both high semantic fidelity and perceptual quality in super-resolution.



Paperid:5620
Authors:Seonghwan Park, Jueun Mun, Donghyun Oh, Namhoon Lee
Abstract:
Concept bottleneck models (CBMs) ensure interpretability by decomposing predictions into human interpretable concepts.Yet the annotations used for training CBMs that enable this transparency are often noisy, and the impact of such corruption is not well understood.In this study, we present the first systematic study of noise in CBMs and show that even moderate corruption simultaneously impairs prediction performance, interpretability, and the intervention effectiveness.Our analysis identifies a susceptible subset of concepts whose accuracy declines far more than the average gap between noisy and clean supervision and whose corruption accounts for most performance loss.To mitigate this vulnerability we propose a two-stage framework.During training, sharpness-aware minimization stabilizes the learning of noise-sensitive concepts.During inference, where clean labels are unavailable, we rank concepts by predictive entropy and correct only the most uncertain ones, using uncertainty as a proxy for susceptibility.Theoretical analysis and extensive ablations elucidate why sharpness-aware training confers robustness and why uncertainty reliably identifies susceptible concepts, providing a principled basis that preserves both interpretability and resilience in the presence of noise.



Paperid:5621
Authors:Weilun Feng, Haotong Qin, Chuanguang Yang, Xiangqi Li, Han Yang, Yuqi Li, Zhulin An, Libo Huang, Michele Magno, Yongjun Xu
Abstract:
Abstract:Diffusion transformers have emerged as the mainstream paradigm for video generation models. However, the use of up to billions of parameters incurs significant computational costs. Quantization offers a promising solution by reducing memory usage and accelerating inference. Nonetheless, we observe that the joint modeling of spatial and temporal information in video diffusion models (V-DMs) leads to extremely long token sequences, which introduces high calibration variance and learning challenges. To address these issues, we propose **$\text{S}^2$Q-VDiT**, a post-training quantization framework for V-DMs that leverages **S**alient data and **S**parse token distillation. During the calibration phase, we identify that quantization performance is highly sensitive to the choice of calibration data. To mitigate this, we introduce *Hessian-aware Salient Data Selection*, which constructs high-quality calibration datasets by considering both diffusion and quantization characteristics unique to V-DMs. To tackle the learning challenges, we further analyze the sparse attention patterns inherent in V-DMs. Based on this observation, we propose *Attention-guided Sparse Token Distillation*, which exploits token-wise attention distributions to emphasize tokens that are more influential to the model's output. Under W4A6 quantization, $\text{S}^2$Q-VDiT achieves lossless performance while delivering $3.9\times$ model compression and $1.3\times$ inference acceleration.



Paperid:5622
Authors:Geonho Hwang
Abstract:
Abstract:In this paper, we investigate the universal approximation property of deep, narrow multilayer perceptrons (MLPs) for $C^1$ functions under the Sobolev norm, specifically the $W^{1, \infty}$ norm. Although the optimal width of deep, narrow MLPs for approximating continuous functions has been extensively studied, significantly less is known about the corresponding optimal width for $C^1$ functions. We demonstrate that \textit{the optimal width} can be determined in a wide range of cases within the $C^1$ setting. Our approach consists of two main steps. First, leveraging control theory, we show that any diffeomorphism can be approximated by deep, narrow MLPs. Second, using the Borsuk-Ulam theorem and various results from differential geometry, we prove that the optimal width for approximating arbitrary $C^1$ functions via diffeomorphisms is $\min(d_x + d_y, \max(2d_x + 1, d_y))$ in infinitely many cases. Our results apply to a broad class of activation functions.



Paperid:5623
Authors:Mahmoud Selim, Sriharsha Bhat, Karl H. Johansson
Abstract:
Modeling and forecasting nonlinear dynamics under distribution shifts is essential for robust decision-making in real-world systems. In this work, we proposeMetaKoopman, a Bayesian meta-learning framework for modeling nonlinear dynamics through linear latent representations. MetaKoopman learns a Matrix Normal-Inverse Wishart (MNIW) prior over the Koopman operator, enabling closed-form Bayesian updates conditioned on recent trajectory segments. Moreover, it provides a closed-form posterior predictive distribution over future state trajectories, capturing both epistemic and aleatoric uncertainty in the learned dynamics. We evaluate MetaKoopman on a full-scale autonomous truck and trailer system across a wide range of adverse winter scenarios—including snow, ice, and mixed-friction conditions—as well as in simulated control tasks with diverse distribution shifts. MetaKoopman consistently outperforms prior approaches in multi-step prediction accuracy, uncertainty calibration and robustness to distributional shifts. Field experiments further demonstrate its effectiveness in dynamically feasible motion planning, particularly during evasive maneuvers and operation at the limits of traction.



Paperid:5624
Authors:Ling Fu, Zhebin Kuang, Jiajun Song, Mingxin Huang, Biao Yang, Yuzhe Li, Linghao Zhu, Qidi Luo, Xinyu Wang, Hao Lu, Zhang Li, Guozhi Tang, Bin Shan, Chunhui Lin, Qi Liu, Binghong Wu, Hao Feng, Hao Liu, Can Huang, Jingqun Tang, Wei Chen, Lianwen Jin, Yuliang Liu, Xiang Bai
Abstract:
Abstract:Scoring the Optical Character Recognition (OCR) capabilities of Large Multimodal Models (LMMs) has witnessed growing interest. Existing benchmarks have highlighted the impressive performance of LMMs in text recognition; however, their abilities in certain challenging tasks, such as text localization, handwritten content extraction, and logical reasoning, remain underexplored. To bridge this gap, we introduce OCRBench v2, a large-scale bilingual text-centric benchmark with currently the most comprehensive set of tasks ($4\times$ more tasks than the previous multi-scene benchmark OCRBench), the widest coverage of scenarios ($31$ diverse scenarios), and thorough evaluation metrics, with $10,000$ human-verified question-answering pairs and a high proportion of difficult samples. Moreover, we construct a private test set with $1,500$ manually annotated images. The consistent evaluation trends observed across both public and private test sets validate the \datasetname's reliability. After carefully benchmarking state-of-the-art LMMs, we find that most LMMs score below $50$ ($100$ in total) and suffer from five-type limitations, including less frequently encountered text recognition, fine-grained perception, layout perception, complex element parsing, and logical reasoning. The benchmark and evaluation scripts are available at https://anonymous.4open.science/r/qytest-5FC4.



Paperid:5625
Authors:jusheng zhang, Kaitong Cai, Yijia Fan, Ningyuan Liu, Keze Wang
Abstract:
Abstract:We propose a novel collaborative multi-agent optimization framework for adaptive training in multi-label image classification, fundamentally advancing beyond static decision rules and isolated automation. Our method deploys a set of distributed, task-specific agents, each responsible for dynamically orchestrating critical training components—including data augmentation, optimization methods, learning rate schedules, and loss functions—according to evolving visual-semantic relationships and training states. Each agent employs an advanced non-stationary multi-armed bandit algorithm, integrating both $\epsilon$-greedy and upper confidence bound strategies, to judiciously balance exploration with exploitation throughout the training lifecycle. A hierarchical composite reward mechanism synergizes overall classification accuracy, rare class recognition, and training stability, fostering both independent optimization and implicit collaborative behavior among agents. The framework further leverages refined techniques such as dual-rate exponential moving average smoothing and structured mixed-precision training to enhance robustness and computational efficiency. Extensive experiments across benchmarks including Pascal VOC, COCO, Yeast, and Mediamill demonstrate that our approach achieves superior mean average precision and rare-class F1 scores compared to state-of-the-art methods, while also exhibiting rapid convergence and remarkable cross-domain generalization. Our results indicate that collaborative multi-agent adaptive optimization offers a scalable and principled solution for self-optimizing deep learning in complex multi-label scenarios.



Paperid:5626
Authors:Yiqi Tian, Pengfei Jin, Mingze Yuan, Na Li, Bo Zeng, Quanzheng Li
Abstract:
Diffusion models have achieved state-of-the-art performance in generative modeling, yet their sampling procedures remain vulnerable to hallucinations—often stemming from inaccuracies in score approximation. In this work, we reinterpret diffusion sampling through the lens of optimization and introduce RODS (Robust Optimization–inspired Diffusion Sampler), a novel method that detects and corrects high-risk sampling steps using geometric cues from the loss landscape. RODS enforces smoother sampling trajectories and \textit{adaptively} adjusts perturbations, reducing hallucinations without retraining and at minimal additional inference cost. Experiments on AFHQv2, FFHQ, and 11k-hands demonstrate that RODS improves both sampling fidelity and robustness, detecting over 70\% of hallucinated samples and correcting more than 25\%, all while avoiding the introduction of new artifacts.



Paperid:5627
Authors:Matthias Jammot, Björn Braun, Paul Streli, Rafael Wampfler, Christian Holz
Abstract:
Understanding affect is central to anticipating human behavior, yet current egocentric vision benchmarks largely ignore the person’s emotional states that shape their decisions and actions. Existing tasks in egocentric perception focus on physical activities, hand-object interactions, and attention modeling—assuming neutral affect and uniform personality. This limits the ability of vision systems to capture key internal drivers of behavior. In this paper, we present egoEMOTION, the first dataset that couples egocentric visual and physiological signals with dense self-reports of emotion and personality across controlled and real-world scenarios. Our dataset includes over 50 hours of recordings from 43 participants, captured using Meta’s Project Aria glasses. Each session provides synchronized eye-tracking video, head-mounted photoplethysmography, inertial motion data, and physiological baselines for reference. Participants completed emotion-elicitation tasks and naturalistic activities while self-reporting their affective state using the Circumplex Model and Mikels’ Wheel as well as their personality via the Big Five model. We define three benchmark tasks: (1) continuous affect classification (valence, arousal, dominance); (2) discrete emotion classification; and (3) trait-level personality inference. We show that a classical learning-based method, as a simple baseline in real-world affect prediction, produces better estimates from signals captured on egocentric vision systems than processing physiological signals. Our dataset establishes emotion and personality as core dimensions in egocentric perception and opens new directions in affect-driven modeling of behavior, intent, and interaction.



Paperid:5628
Authors:Wenju Sun, Qingyong Li, Wen Wang, Yang Liu, Yangliao Geng, Boyang Li
Abstract:
Multi-task model merging aims to consolidate knowledge from multiple fine-tuned task-specific experts into a unified model while minimizing performance degradation. Existing methods primarily approach this by minimizing differences between task-specific experts and the unified model, either from a parameter-level or a task-loss perspective. However, parameter-level methods exhibit a significant performance gap compared to the upper bound, while task-loss approaches entail costly secondary training procedures. In contrast, we observe that performance degradation closely correlates with feature drift, i.e., differences in feature representations of the same sample caused by model merging. Motivated by this observation, we propose Layer-wise Optimal Task Vector Merging (LOT Merging), a technique that explicitly minimizes feature drift between task-specific experts and the unified model in a layer-by-layer manner. LOT Merging can be formulated as a convex quadratic optimization problem, enabling us to analytically derive closed-form solutions for the parameters of linear and normalization layers. Consequently, LOT Merging achieves efficient model consolidation through basic matrix operations. Extensive experiments across vision and vision-language benchmarks demonstrate that LOT Merging significantly outperforms baseline methods, achieving improvements of up to 4.4% (ViT-B/32) over state-of-the-art approaches.



Paperid:5629
Authors:Haoyuan Liang, Shilei Cao, Li, Zhiyu Ye, Haohuan Fu, Juepeng Zheng
Abstract:
Federated learning (FL) has recently emerged as the primary approach to overcoming data silos, enabling collaborative model training without sharing sensitive or proprietary data.Parallel federated learning (PFL) aggregates models trained independently on each client’s local data, which can lead to suboptimal convergence due to limited data exposure.In contrast, Sequential Federated Learning (SFL) allows models to traverse client datasets sequentially, enhancing data utilization. However, SFL effectiveness is limited in real-world non-IID scenarios characterized by category shift (inconsistent class distributions) and domain shift (distribution discrepancies).These shifts cause two critical issues: update order sensitivity, where model performance varies significantly with the sequence of client updates, and catastrophic forgetting, where the model forgets previously learned features when trained on new client data.We propose SPFL, a novel updating method that can be integrated into existing FL methods, integrating sequential updates with parallel aggregation to enhance data utilization and ease update order sensitivity. At the same time, we give the convergence analysis of SPFL under strong convex, general convex, and non-convex conditions, proving that this update scheme is significantly better than PFL and SFL.Additionally, we introduce the Global-Local Alignment Module to mitigate catastrophic forgetting by aligning the predictions of the global model with those of the local and previous models during training. Our extensive experiments demonstrate that integrating SPFL into existing PFL methods significantly improves performance under category and domain shifts.



Paperid:5630
Authors:Tianyuan Jia, Ziyu Li, Qing Li, Xiuxing Li, Xiang Li, Chen Wei, Li Yao, Xia Wu
Abstract:
Motion planning in high-dimensional continuous spaces remains challenging due to complex environments and computational constraints. Although learning-based planners, especially graph neural network (GNN)-based, have significantly improved planning performance, they still struggle with inaccurate graph construction and limited structural reasoning, constraining search efficiency and path quality. The human brain exhibits efficient planning through a two-stage Perception-Decision model. First, egocentric spatial representations from visual and proprioceptive input are constructed, and then semantic–episodic synergy is leveraged to support decision-making in uncertainty scenarios. Inspired by this process, we propose NeuroMP, a brain-inspired planning framework that learns to plan like the human brain. NeuroMP integrates a Perceptive Segment Selector inspired by visuospatial perception to construct safer graphs, and a Global Alignment Heuristic guide search in weakly connected graphs by modeling semantic-episodic synergistic decision-making. Experimental results demonstrate that NeuroMP significantly outperforms existing planning methods in efficiency and quality while maintaining a high success rate.



Paperid:5631
Authors:Ibragim Badertdinov, Alexander Golubev, Maksim Nekrashevich, Anton Shevtsov, Simon Karasik, Andrei Andriushchenko, Maria Trofimova, Daria Litvintseva, Boris Yangel
Abstract:
LLM-based agents have shown promising capabilities in a growing range of software engineering (SWE) tasks. However, advancing this field faces two critical challenges. First, high-quality training data is scarce, especially data that reflects real-world SWE scenarios, where agents must interact with development environments, execute code and adapt behavior based on the outcomes of their actions. Existing datasets are either limited to one-shot code generation or comprise small, manually curated collections of interactive tasks, lacking both scale and diversity. Second, the lack of fresh interactive SWE tasks affects evaluation of rapidly improving models, as static benchmarks quickly become outdated due to contamination issues. To address these limitations, we introduce a novel, automated, and scalable pipeline to continuously extract real-world interactive SWE tasks from diverse GitHub repositories. Using this pipeline, we construct SWE-rebench, a public dataset comprising over 21,000 interactive Python-based SWE tasks, suitable for reinforcement learning of SWE agents at scale. Additionally, we use continuous supply of fresh tasks collected using SWE-rebench methodology to build a contamination-free benchmark for agentic software engineering. We compare results of various LLMs on this benchmark to results on SWE-bench Verified and show that performance of some language models might be inflated due to contamination issues.



Paperid:5632
Authors:Zilin Shen, Xinyu Luo, Imtiaz Karim, Elisa Bertino
Abstract:
Abstract:Accurately extracting protocol-state machines (PSMs) from the long, densely written Request-for-Comments (RFC) standards that govern Internet‐scale communication remains a bottleneck for automated security analysis and protocol testing. In this paper, we introduce RFC2PSM, the first large-scale dataset that pairs 1,580 pages of cleaned RFC text with 108 manually validated states and 297 transitions covering 14 widely deployed protocols spanning the data-link, transport, session, and application layers. Built on this corpus, we propose PsmBench, a benchmark that (i) feeds chunked RFC to an LLM, (ii) prompts the model to emit a machine-readable PSM, and (iii) scores the output with structure-aware, semantic fuzzy-matching metrics that reward partially correct graphs.A comprehensive baseline study of nine state-of-the-art open and commercial LLMs reveals a persistent state–transition gap: models identify many individual states (up to $0.82$ F1) but struggle to assemble coherent transition graphs ($\leq 0.38$ F1), highlighting challenges in long-context reasoning, alias resolution, and action/event disambiguation. We release the dataset, evaluation code, and all model outputs as open-sourced, providing a fully reproducible starting point for future work on reasoning over technical prose and generating executable graph structures. RFC2PSM and PsmBench aim to catalyze cross-disciplinary progress toward LLMs that can interpret and verify the protocols that keep the Internet safe.



Paperid:5633
Authors:Chenghao Liu, Enming Liang, Minghua Chen
Abstract:
Abstract:Projection-free first-order methods offer compelling advantages for constrained optimization. These methods excel on limited convex sets, demonstrating scalability, fast convergence, and iteration-wise feasibility without projection costs. However, their extension to general convex sets faces significant challenges. For example, Frank-Wolfe methods require linear optimization oracles at each iteration, while penalty-based approaches often exhibit poor solution feasibility. This raises a fundamental question: can projection-free methods be extended to optimization over a general convex set without expensive oracles?We address this challenge by introducing a novel projection-free method: **Hom-PGD**. It leverages a homeomorphism between the constraint set and a unit ball to transform the original optimization problem into a ball-constrained formulation. This transformation enables efficient projection-free first-order optimization in the transformed space while maintaining the original problem's structure. We establish theoretical guarantees that Hom-PGD achieves *optimal* convergence rates matching gradient descent with constant step-size: $\mathcal{O}(\log 1/\epsilon)$ for strongly convex objectives,$\mathcal{O}(1/\epsilon)$ for convex objectives, and $\mathcal{O}(1/\epsilon^2)$ for non-convex objectives, with $\mathcal{O}(n^2)$ per-iteration complexity. The framework also extends to certain non-convex sets, broadening its applicability to challenging optimization problems. Extensive numerical experiments on both convex and non-convex problems show that Hom-PGD achieves a similar convergence rate but with significantly lower per-iteration time compared with existing projection-free algorithms.This advancement significantly expands the scope of projection-free methods, making them viable for a broader range of constrained optimization applications.



Paperid:5634
Authors:Yiyang Cai, Zhengkai Jiang, Yulong Liu, Chunyang Jiang, Wei Xue, Yike Guo, Wenhan Luo
Abstract:
Facial personalization faces challenges to maintain identity fidelity without disrupting the foundation model's prompt consistency. The mainstream personalization models employ identity embedding to integrate identity information within the attention mechanisms. However, our preliminary findings reveal that identity embeddings compromise the effectiveness of other tokens in the prompt, thereby limiting high prompt consistency and attribute-level controllability. Moreover, by deactivating identity embedding, personalization models still demonstrate the underlying foundation models' ability to control facial attributes precisely. It suggests that such foundation models' knowledge can be leveraged to cure the ill-aligned prompt consistency of personalization models. Building upon these insights, we propose FreeCure, a framework that improves the prompt consistency of personalization models with their latent foundation models' knowledge. First, by setting a dual inference paradigm with/without identity embedding, we identify attributes (e.g., hair, accessories, etc.) for enhancements. Second, we introduce a novel foundation-aware self-attention module, coupled with an inversion-based process to bring well-aligned attribute information to the personalization process. Our approach is training-free, and can effectively enhance a wide array of facial attributes; and it can be seamlessly integrated into existing popular personalization models based on both Stable Diffusion and FLUX. FreeCure has consistently demonstrated significant improvements in prompt consistency across these facial personalization models while maintaining the integrity of their original identity fidelity. Project page: https://freecure.github.io/.



Paperid:5635
Authors:Tangwen Qian, Junhe Li, Yile Chen, Gao Cong, Zezhi Shao, Jun Zhang, Tao Sun, Fei Wang, Yongjun Xu
Abstract:
Abstract:Spatio-temporal trajectory representation learning plays a crucial role in various urban applications such as transportation systems, urban planning, and environmental monitoring. Existing methods can be divided into single-view and multi-view approaches, with the latter offering richer representations by integrating multiple sources of spatio-temporal data. However, these methods often struggle to generalize across diverse urban scenes due to multi-city structural heterogeneity, which arises from the disparities in road networks, grid layouts, and traffic regulations across cities, and the amplified seesaw phenomenon, where optimizing for one city, view, or task can degrade performance in others. These challenges hinder the deployment of trajectory learning models across multiple cities, limiting their real-world applicability. In this work, we propose SMARTraj$^2$, a novel stable multi-city adaptive method for multi-view spatio-temporal trajectory representation learning. Specifically, we introduce a feature disentanglement module to separate domain-invariant and domain-specific features, and a personalized gating mechanism to dynamically stabilize the contributions of different views and tasks. Our approach achieves superior generalization across heterogeneous urban scenes while maintaining robust performance across multiple downstream tasks. Extensive experiments on benchmark datasets demonstrate the effectiveness of SMARTraj$^2$ in enhancing cross-city generalization and outperforming state-of-the-art methods. See our project website at \url{https://anonymous.4open.science/r/SMARTraj-515}.



Paperid:5636
Authors:Junyoung Park, Dalton Jones, Matthew Morse, Raghavv Goel, Mingu Lee, Christopher Lott
Abstract:
We demonstrate that geometrically distinctive keys during LLM inference tend to have high attention scores. Based on the phenomenon we propose KeyDiff, a training-free KV cache eviction method based solely on key similarity. Unlike other KV cache eviction methods, KeyDiff can process arbitrarily long prompts within strict resource constraints and efficiently generate responses.We provide a theoretical basis for KeyDiff by relating key diversity with attention scores. These results imply KeyDiff can efficiently identify the most important tokens to retain. Notably KeyDiff does not rely on attention scores, allowing the use of optimized attention mechanisms like FlashAttention. Under a strict memory allowance, we demonstrate the effectiveness of KeyDiff for the Llama and Qwen model families by observing a performance gap of less than 0.04\% with 8K cache budget (~23\% KV cache reduction) from the non-evicting baseline on LongBench for Llama 3.1-8B and Llama 3.2-3B. We also observe near baseline performance for Deepseek-R1-Distill-Llama-8B on the Math500 reasoning benchmark and decrease end-to-end inference latency by up to 30\% compared to the other token-eviction methods.



Paperid:5637
Authors:Elena Celledoni, Brynjulf Owren, Chong Shen, Baige Xu, Takaharu Yaguchi
Abstract:
Physical phenomena in the real world are often described by energy-based modeling theories, such as Hamiltonian mechanics or the Landau theory. It is known that physical phenomena based on these theories have an energy conservation law or a dissipation law. Therefore, in the simulations of such physical phenomena, numerical methods that preserve the energy-conservation or dissipation laws are desirable. However, because various energy-behavior-preserving numerical methods have been proposed, it is difficult to discover the best one. In this study, we propose a method for learning highly accurate energy-behavior-preserving integrators from data. Numerical results show that our approach certainly learns energy-behavior-preserving numerical methods that are more accurate than existing numerical methods for various differential equations, including chaotic Hamiltonian systems, dissipative systems, and a nonlinear partial differential equation. We also provide universal approximation theorems for the proposed approach.



Paperid:5638
Authors:Ki-Joong Kwon, Junho So, Sang-Hoon Lee
Abstract:
While substantial advances have been achieved in TTS for languages such as English and Mandarin, Korean remains comparatively underrepresented due to the lack of rigorous preprocessing methods, systematically constructed datasets, a shortage of standardized Korean TTS benchmarks, and explicitly optimized models for Korean. To address these limitations, we propose a Korean-tailored data-refinement and coreset selection pipeline. It refines speech data and performs textual normalization especially for numerals and English terms, followed by a novel coreset selection strategy that leverages Jamo-based linguistic and phonological features unique to Korean. As a result, we release CoreaSpeech, an efficient and robust Korean speech corpus comprising 700 hours across 21,449 speakers. This refined core subset, evenly balanced across utterances ranging from 0 to 30 seconds, is derived from 2,058 hours of widely used Korean datasets. Building on this, we conducted extensive experiments via cross-lingual fine-tuning with our CoreaSpeech dataset. Furthermore, we introduce a new universal Korean TTS benchmark dataset including clean, noisy, and numeric subsets. Additionally, we demonstrate that our Korean-specific text normalization serves as a plug-and-play module, reliably improving performance regardless of the underlying TTS architecture. We publicly release our dataset, pipeline code, and evaluation benchmarks to support reproducible research and further advances in Korean and multilingual speech synthesis.



Paperid:5639
Authors:Yi Huang, Zhan Qu, Lihui Jiang, Hongbo Zhang, Bingbing Liu
Abstract:
End-to-end autonomous driving systems, predominantly trained through imitation learning, have demonstrated considerable effectiveness in leveraging large-scale expert driving data. Despite their success in open-loop evaluations, these systems often exhibit significant performance degradation in closed-loop scenarios due to causal confusion. This confusion is fundamentally exacerbated by the overreliance of the imitation learning paradigm on expert trajectories, which often contain unattributable noise and interfere with the modeling of causal relationships between environmental contexts and appropriate driving actions.To address this fundamental limitation, we propose Perception-Guided Self-Supervision (PGS)—a simple yet effective training paradigm that leverages perception outputs as the primary supervisory signals, explicitly modeling causal relationships in decision-making. The proposed framework aligns both the inputs and outputs of the decision-making module with perception results—such as lane centerlines and the predicted motions of surrounding agents—by introducing positive and negative self-supervision for the ego trajectory. This alignment is specifically designed to mitigate causal confusion arising from the inherent noise in expert trajectories.Equipped with perception-driven supervision, our method—built on a standard end-to-end architecture—achieves a Driving Score of 77.63 and a mean success rate of 48.18% on the challenging closed-loop Bench2Drive benchmark, significantly outperforming existing state-of-the-art methods, including those employing more complex network architectures and inference pipelines. These results underscore the effectiveness and robustness of the proposed PGS framework, and point to a promising direction for addressing causal confusion and enhancing real-world generalization in autonomous driving.



Paperid:5640
Authors:Zeyue Zhang, Heng Ping, Xiongye Xiao, Peiyu Zhang, Nikos Kanakaris, Xiaoling LU, Paul Bogdan
Abstract:
With AI advancement and increasing circuit complexity, efficient chip design through Electronic Design Automation (EDA) is critical. Fast and accurate congestion prediction in chip layout and routing can significantly enhance automated design performance. Existing congestion modeling methods are limited by(i)ineffective processing and fusion of multi-view circuit data information, and(ii)insufficient reliability and interpretability in the prediction process. To address these challenges, We proposeMulti-viewInterpretableHypergraph forChip (MIHC), a trustworthy 'multi-view hypergraph neural network'-based framework that(i)processes both graph and image information in unified hypergraph representations, capturing topological and geometric circuit data, and(ii)implements a novel subgraph Information Bottleneck mechanism identifying critical congestion-correlated regions to guide predictions. This represents the first attempt to incorporate such interpretability into congestion prediction through informative graph reasoning. Experiments show our model reduces NMAE by 16.67% and 8.57% in cell-based and grid-based predictions on ISPD2015, and 5.26% and 2.44% on CircuitNet-N28, respectively, compared to state-of-the-art methods.



Paperid:5641
Authors:Shanawaj Madarkar, Mahajabeen Madarkar, Madhumitha Venkatesh, TELI PRAKASH, Konda Reddy Mopuri, Vinaykumar MV, Kota Sathwika, Adarsh Kasturi, Gandla Raj, Padharthi Supranitha, Harsh Udai
Abstract:
Artificial intelligence is poised to augment dermatological care by enabling scalable image-based diagnostics. Yet, the development of robust and equitable models remains hindered by datasets that fail to capture the clinical and demographic complexity of real-world practice. This complexity stems from region-specific disease distributions, wide variation in skin tones, and the underrepresentation of outpatient scenarios from non-Western populations. We introduce DermaCon-IN, a prospectively curated dermatology dataset comprising over 5,450 clinical images from approximately 3,000 patients across outpatient clinics in South India. Each image is annotated by board-certified dermatologists with over 240 distinct diagnoses, structured under a hierarchical, etiology-based taxonomy adapted from Rook’s classification. The dataset captures a wide spectrum of dermatologic conditions and tonal variation commonly seen in Indian outpatient care. We benchmark a range of architectures—including convolutional models (ResNet, DenseNet, EfficientNet), transformer-based models (ViT, MaxViT, Swin), and Concept Bottleneck Models to establish baseline performance and explore how anatomical and concept-level cues may be integrated. These results are intended to guide future efforts toward interpretable and clinically realistic models. DermaCon-IN provides a scalable and representative foundation for advancing dermatology AI in real-world settings.



Paperid:5642
Authors:Yuwu Lu, Chunzhi Liu
Abstract:
Domain Adaptation of Black-box Predictors (DABP) transfers knowledge from a labeled source domain to an unlabeled target domain, without requiring access to either source data or source model. Common practices of DABP leverage reliable samples to suppress negative information about unreliable samples. However, there are still some problems: i) Excessive attention to reliable sample aggregation leads to premature overfitting; ii) Valuable information in unreliable samples is often overlooked. To address them, we propose a novel spatial learning approach, called Controlled Visual Hallucination via Thalamus-driven Decoupling Network (CVH-TDN). Specifically, CVH-TDN is the first work that introduces the thalamus-driven decoupling network in the visual task, relying on its connection with hallucination to control the direction of sample generation in feature space. CVH-TDN is composed of Hallucination Generation (HG), Hallucination Alignment (HA), and Hallucination Calibration (HC), aiming to explore the spatial relationship information between samples and hallucinations. Extensive experiments confirm that CVH-TDN achieves SOTA performance on four standard benchmarks.



Paperid:5643
Authors:Pucheng Dang, Di Huang, Dong Li, Kang Chen, Yuanbo Wen, Qi Guo, Xing Hu
Abstract:
Abstract:Out-of-tree kernel patches are essential for adapting the Linux kernel to new hardware or enabling specific functionalities. Maintaining and updating these patches across different kernel versions demands significant effort from experienced engineers. Large language models (LLMs) have shown remarkable progress across various domains, suggesting their potential for automating out-of-tree kernel patch migration. However, our findings reveal that LLMs, while promising, struggle with incomplete code context understanding and inaccurate migration point identification. In this work, we propose MigGPT, a framework that employs a novel code fingerprint structure to retain code snippet information and incorporates three meticulously designed modules to improve the migration accuracy and efficiency of out-of-tree kernel patches. Furthermore, we establish a robust benchmark using real-world out-of-tree kernel patch projects to evaluate LLM capabilities. Evaluations show that MigGPT significantly outperforms the direct application of vanilla LLMs, achieving an average completion rate of 72.59\% ($\uparrow 50.74\%$) for migration tasks.



Paperid:5644
Authors:Xuan Liu, Siru Ouyang, Xianrui Zhong, Jiawei Han, Huimin Zhao
Abstract:
Large language models (LLMs) have gained significant attention in chemistry. However, most existing datasets center on molecular-level property prediction and overlook the role of fine-grained functional group (FG) information. Incorporating FG-level data can provide valuable prior knowledge that links molecular structures with textual descriptions, which can be used to build more interpretable, structure-aware LLMs for reasoning on molecule-related tasks. Moreover, LLMs can learn from such fine-grained information to uncover hidden relationships between specific functional groups and molecular properties, thereby advancing molecular design and drug discovery. Here, we introduce FGBench, a dataset comprising 625K molecular property reasoning problems with functional group information. Functional groups are precisely annotated and localized within the molecule, which ensures the dataset's interoperability thereby facilitating further multimodal applications. FGBench includes both regression and classification tasks on 245 different functional groups across three categories for molecular property reasoning: (1) single functional group impacts, (2) multiple functional group interactions, and (3) direct molecular comparisons. In the benchmark of state-of-the-art LLMs on 7K curated data, the results indicate that current LLMs struggle with FG-level property reasoning, highlighting the need to enhance reasoning capabilities in LLMs for chemistry tasks. We anticipate that the methodology employed in FGBench to construct datasets with functional group-level information will serve as a foundational framework for generating new question–answer pairs, enabling LLMs to better understand fine-grained molecular structure–property relationships. The dataset and evaluation code are available at this \href{https://github.com/xuanliugit/FGBench}{link}.



Paperid:5645
Authors:Chenrui Wang, Junyi Shu, Billy Chiu, YU LI, Saleh Alharbi, Min Zhang, Jing Li
Abstract:
The rapid development of LLMs has raised concerns about their potential misuse, leading to various watermarking schemes that typically offer high detectability. However, existing watermarking techniques often face trade-off between watermark detectability and generated text quality. In this paper, we introduce Learning to Watermark (LTW), a novel selective watermarking framework that leverages multi-objective optimization to effectively balance these competing goals. LTW features a lightweight network that adaptively decides when to apply the watermark by analyzing sentence embeddings, token entropy, and current watermarking ratio. Training of the network involves two specifically constructed loss functions that guide the model toward Pareto-optimal solutions, thereby harmonizing watermark detectability and text quality.By integrating LTW with two baseline watermarking methods, our experimental evaluations demonstrate that LTW significantly enhances text quality without compromising detectability. Our selective watermarking approach offers a new perspective for designing watermarks for LLMs and a way to preserve high text quality for watermarks.



Paperid:5646
Authors:Kai Hu, Zhang Yu, Yuan Zhang, Zhineng Chen, Xieping Gao
Abstract:
Non-exemplar class incremental learning (NECIL) aims to continuously assimilate new knowledge while retaining previously acquired knowledge in scenarios where prior examples are unavailable. A prevalent strategy within NECIL mitigates knowledge forgetting by freezing the feature extractor after training on the initial task. However, this freezing mechanism does not provide explicit training to differentiate between new and old classes, resulting in overlapping feature representations. To address this challenge, we propose aConfusion-driven seLf-supervised prOgressiVely weightedEnsemble leaRning (CLOVER) framework for NECIL. Firstly, we introduce a confusion-driven self-supervised learning approach that enhances representation extraction by guiding the model to distinguish between highly confusable classes, thereby reducing class representation overlap. Secondly, we develop a progressively weighted ensemble learning method that gradually adjusts weights to integrate diverse knowledge more effectively, further minimizing representation overlap. Finally, extensive experiments demonstrate that our proposed method achieves state-of-the-art results on the CIFAR100, TinyImageNet, and ImageNet-Subset NECIL benchmarks.



Paperid:5647
Authors:Shaopeng Fu, Liang Ding, Jingfeng ZHANG, Di Wang
Abstract:
Abstract:Jailbreak attacks against large language models (LLMs) aim to induce harmful behaviors in LLMs through carefully crafted adversarial prompts.To mitigate attacks, one way is to perform adversarial training (AT)-based alignment, i.e., training LLMs on some of the most adversarial prompts to help them learn how to behave safely under attacks.During AT, the length of adversarial prompts plays a critical role in the robustness of aligned LLMs.While long-length adversarial prompts during AT might lead to strong LLM robustness, their synthesis however is very resource-consuming, which may limit the application of LLM AT.This paper focuses on adversarial suffix jailbreak attacks and unveils that to defend against a jailbreak attack with an adversarial suffix of length $\Theta(M)$, it is enough to align LLMs on prompts with adversarial suffixes of length $\Theta(\sqrt{M})$.Theoretically, we analyze the adversarial in-context learning of linear transformers on linear regression tasks and prove a robust generalization bound for trained transformers.The bound depends on the term $\Theta(\sqrt{M_{\text{test}}}/M_{\text{train}})$, where $M_{\text{train}}$ and $M_{\text{test}}$ are the numbers of adversarially perturbed in-context samples during training and testing.Empirically, we conduct AT on popular open-source LLMs and evaluate their robustness against jailbreak attacks of different adversarial suffix lengths.Results confirm a positive correlation between the attack success rate and the ratio of the square root of the adversarial suffix length during jailbreaking to the length during AT.Our findings show that it is practical to defend against "long-length" jailbreak attacks via efficient "short-length" AT.



Paperid:5648
Authors:shurong yang, Dong Wei, Yihuang Hu, Qiong Peng, Hong Liu, Yawen Huang, Xian Wu, Yefeng Zheng, Liansheng Wang
Abstract:
Diffusion-based virtual staining methods of histopathology images have demonstrated outstanding potential for stain normalization and cross-dye staining (e.g., hematoxylin-eosin to immunohistochemistry). However, achieving pathology-correct cross-dye virtual staining with versatile tone controls poses significant challenges due to the difficulty of decoupling the given pathology and tone conditions. This issue would cause non-pathologic regions to be mistakenly stained like pathologic ones, and vice versa, which we term “pathology leakage.” To address this issue, we propose diffusion virtual staining Transformer (D-VST), a new framework with versatile tone control for cross-dye virtual staining. Specifically, we introduce a pathology encoder in conjunction with a tone encoder, combined with a two-stage curriculum learning scheme that decouples pathology and tone conditions, to enable tone control while eliminating pathology leakage. Further, to extend our method for billion-pixel whole slide image (WSI) staining, we introduce a novel frequency-aware adaptive patch sampling strategy for high-quality yet efficient inference of ultra-high resolution images in a zero-shot manner. Integrating these two innovative components facilitates a pathology-correct, tone-controllable, cross-dye WSI virtual staining process. Extensive experiments on three virtual staining tasks that involve translating between four different dyes demonstrate the superiority of our approach in generating high-quality and pathologically accurate images compared to existing methods based on generative adversarial networks and diffusion models. Our code and trained models will be released.



Paperid:5649
Authors:JaeYun Lee, Jae Hyeon Park, Gyoomin Lee, Bogyeong Kim, Min Hee Cha, Hyeok Nam, Joo Jeon, Hyunse Lee, Sung In Cho
Abstract:
Source-free domain adaptation (SFDA) aims to adapt a pre-trained source model to an unlabeled target domain without requiring labeled source data. In a self-supervised manner, relying on pseudo labels on target domain samples facilitates the domain adaptation performance providing strong supervision.However, the critical problem of this approach is the inherent instability of the pre-trained source model in the target domain, leading to unreliable pseudo labels for the target domain data. To tackle this, we propose a novel Dual-facet pseudo labeling strategy that jointly leverages a task-specific perspective and a domain-invariant perspective, assigning pseudo labels only to target samples with agreement between a target model and CLIP. To further enhance representation learning without introducing noisy supervision, we apply consistency training to uncertain samples. Additionally, we introduce a Tsallis mutual information (TMI)-based vision optimization strategy guided by an Uncertainty-based adaptation index (UAI), which dynamically modulates entropy sensitivity based on the model’s adaptation uncertainty. The UAI-based training paradigm enables stable and adaptive domain alignment by effectively balancing exploration and exploitation processes during the optimization process. Our proposed method achieves state-of-the-art performance on domain adaptation benchmark datasets, improving adaptation accuracy by 1.6\% on Office-Home, 1.4\% on VisDA-C, and 2.9\% on DomainNet-126, demonstrating its effectiveness in SFDA.



Paperid:5650
Authors:Rishal Aggarwal, Jacky Chen, Nicholas Boffi, David Koes
Abstract:
Efficient sampling from the Boltzmann distribution defined by an energy function is a key challenge in modeling physical systems such as molecules. Boltzmann Generators tackle this by leveraging Continuous Normalizing Flows that transform a simple prior into a distribution that can be reweighted to match the Boltzmann distribution using sample likelihoods. However, obtaining likelihoods requires computing costly Jacobians during integration, making it impractical for large molecular systems. To overcome this, we propose learning the likelihood of the generated distribution via an energy-based model trained with noise contrastive estimation and score matching. By using stochastic interpolants to anneal between the prior and generated distributions, we combine both the objective functions to efficiently learn the density function. On the alanine dipeptide system, we demonstrate that our method yields free energy profiles and energy distributions comparable to those obtained with exact likelihoods. Additionally, we show that free energy differences between metastable states can be estimated accurately with orders-of-magnitude speedup.



Paperid:5651
Authors:Dapeng Zhang, Fei Shen, Rui Zhao, Yinda Chen, Peng Zhi, Chenyang Li, Rui Zhou, Qingguo Zhou
Abstract:
Autonomous driving represents a prominent application of artificial intelligence. Recent approaches have shifted from focusing solely on common scenarios to addressing complex, long-tail situations such as subtle human behaviors, traffic accidents, and non-compliant driving patterns. Given the demonstrated capabilities of large language models (LLMs) in understanding visual and natural language inputs and following instructions, recent methods have integrated LLMs into autonomous driving systems to enhance reasoning, interpretability, and performance across diverse scenarios. However, existing methods typically rely either on real-world data, which is suitable for industrial deployment, or on simulation data tailored to rare or hard case scenarios. Few approaches effectively integrate the complementary advantages of both data sources. To address this limitation, we propose a novel VLM-guided, end-to-end adversarial transfer framework for autonomous driving that transfers long-tail handling capabilities from simulation to real-world deployment. The framework comprises a teacher VLM model, a student VLM model, and a discriminator. Both the teacher and student VLM models utilize a shared base architecture, termed the Chain-of-Causality Visual–Language Model (CoC VLM), which integrates temporal information via an end-to-end text adapter. This architecture supports chain-of-thought reasoning to infer complex driving logic. The teacher and student VLM models are pre-trained separately on simulated and real-world datasets. The discriminator is trained adversarially to facilitate the transfer of long-tail handling capabilities from simulated to real-world environments by the student VLM model, using a novel backpropagation strategy. Experimental results show that our method effectively bridges the gap between simulation and real-world autonomous driving, indicating a promising direction for future research.



Paperid:5652
Authors:Zhiqiang Kou, Yucheng Xie, Hailin Wang, Junyang Chen, Jingq Wang, Ming-Kun Xie, Shuo Chen, Yuheng Jia, Tongliang Liu, Xin Geng
Abstract:
Pseudo label based semi-supervised learning (SSL) for single-label and multi-label classification tasks has been extensively studied; however, semi-supervised label distribution learning (SSLDL) remains a largely unexplored area. Existing SSL methods fail in SSLDL because the pseudo-labels they generate only ensure overall similarity to the ground truth but do not preserve the ranking relationships between true labels, as they rely solely on KL divergence as the loss function during training. These skewed pseudo-labels lead the model to learn incorrect semantic relationships, resulting in reduced performance accuracy. To address these issues, we propose a novel SSLDL method called \textit{RankMatch}. \textit{RankMatch} fully considers the ranking relationships between different labels during the training phase with labeled data to generate higher-quality pseudo-labels. Furthermore, our key observation is that a flexible utilization of pseudo-labels can enhance SSLDL performance. Specifically, focusing solely on the ranking relationships between labels while disregarding their margins helps prevent model overfitting. Theoretically, we prove that incorporating ranking correlations enhances SSLDL performance and establish generalization error bounds for \textit{RankMatch}. Finally, extensive real-world experiments validate its effectiveness.



Paperid:5653
Authors:Yidi Liu, Xueyang Fu, Jie Huang, Jie Xiao, Dong Li, zhangwenlong, LEI BAI, Zheng-Jun Zha
Abstract:
Abstract:Ultra-High Definition (UHD) image restoration struggles to balance computational efficiency and detail retention.While Variational Autoencoders (VAEs) offer improved efficiency by operating in the latent space, with the Gaussian variational constraint, this compression preserves semantics but sacrifices critical high-frequency attributes specific to degradation and thus compromises reconstruction fidelity. % This compromises reconstruction fidelity, even when global semantics are preserved.Consequently, a VAE redesign is imperative to foster a robust semantic representation conducive to generalization and perceptual quality, while simultaneously enabling effective high-frequency information processing crucial for reconstruction fidelity.To address this, we propose \textit{Latent Harmony}, a two-stage framework that reinvigorates VAEs for UHD restoration by concurrently regularizing the latent space and enforcing high-frequency-aware reconstruction constraints. Specifically, Stage One introduces the LH-VAE, which fortifies its latent representation through visual semantic constraints and progressive degradation perturbation for enhanced semantics robustness; meanwhile, it incorporates latent equivariance to bolster its high-frequency reconstruction capabilities. Then, Stage Two facilitates joint training of this refined VAE with a dedicated restoration model.This stage integrates High-Frequency Low-Rank Adaptation (HF-LoRA), featuring two distinct modules: an encoder LoRA, guided by a fidelity-oriented high-frequency alignment loss, tailored for the precise extraction of authentic details from degradation-sensitive high-frequency components; and a decoder LoRA, driven by a perception-oriented loss, designed to synthesize perceptually superior textures. These LoRA modules are meticulously trained via alternating optimization with selective gradient propagation to preserve the integrity of the pre-trained latent structure. This methodology culminates in a flexible fidelity-perception trade-off at inference, managed by an adjustable parameter $\alpha$.Extensive experiments demonstrate that \textit{Latent Harmony} effectively balances perceptual and reconstructive objectives with efficiency, achieving superior restoration performance across diverse UHD and standard-resolution scenarios.



Paperid:5654
Authors:Pengyi Li, Jianye Hao, Yifu Yuan, Hongyao Tang, Zibin Dong, YAN ZHENG
Abstract:
Many real-world control problems require continual policy adjustments to balance multiple objectives, which requires the acquisition of high-quality policies to cover diverse preferences. Multi-Objective Reinforcement Learning (MORL) provides a general framework to solve such problems. However, current MORL methods suffer from high sample complexity, primarily due to the neglect of efficient knowledge sharing and conflicts in optimization with different preferences. To this end, this paper introduces a novel framework, Conflict Objective Regularization in Latent Space (COLA). To enable efficient knowledge sharing, COLA establishes a shared latent representation space for common knowledge, which can avoid redundant learning under different preferences. Besides, COLA introduces a regularization term for the value function to mitigate the negative effects of conflicting preferences on the value function approximation, thereby improving the accuracy of value estimation. Experimental results across various multi-objective continuous control tasks demonstrate the significant superiority of COLA over the state-of-the-art MORL baselines.



Paperid:5655
Authors:Yingtao Zhang, Diego Cerretti, Jialin Zhao, Wenjing Wu, Ziheng Liao, Umberto Michieli, Carlo Cannistraci
Abstract:
Abstract:This study aims to enlarge our current knowledge on the application of brain-inspired network science principles for training artificial neural networks (ANNs) with sparse connectivity. Dynamic sparse training (DST) emulates the synaptic turnover of real brain networks, reducing the computational demands of training and inference in ANNs. However, existing DST methods face difficulties in maintaining peak performance at high connectivity sparsity levels. The Cannistraci-Hebb training (CHT) is a brain-inspired method that is used in DST for growing synaptic connectivity in sparse neural networks. CHT leverages a gradient-free, topology-driven link regrowth mechanism, which has been shown to achieve ultra-sparse (1\% connectivity or lower) advantage across various tasks compared to fully connected networks. Yet, CHT suffers two main drawbacks: (i) its time complexity is $\mathcal{O}(N\cdot d^3)$- N node network size, d node degree - hence it can be efficiently applied only to ultra-sparse networks. (ii) it rigidly selects top link prediction scores, which is inappropriate for the early training epochs, when the network topology presents many unreliable connections. Here, we design the first brain-inspired network model - termed bipartite receptive field (BRF) - to initialize the connectivity of sparse artificial neural networks. Then, we propose a matrix multiplication GPU-friendly approximation of the CH link predictor, which reduces the computational complexity to $\mathcal{O}(N^3)$, enabling a fast implementation of link prediction in large-scale models. Moreover, we introduce the Cannistraci-Hebb training soft rule (CHTs), which adopts a flexible strategy for sampling connections in both link removal and regrowth, balancing the exploration and exploitation of network topology. To further enhance performance, we integrate CHTs with a brain-inspired network topology initialization method known as the bipartite receptive field (BRF). Additionally, we propose a sigmoid-based gradual density decay strategy, leading to an advanced framework referred to as CHTss. Empirical results show that BRF offers performance advantages over previous network science models. Using 1\% of connections, CHTs outperforms fully connected networks in MLP architectures on visual classification tasks, compressing some networks to less than 30\% of the nodes. Using 5\% of the connections, CHTss outperforms fully connected networks in two Transformer-based machine translation tasks. Finally, using 30\% of the connections, CHT and CHTss achieve superior performance compared to other dynamic sparse training methods in language modeling across different sparsity levels, and it surpasses the fully connected counterpart in zero-shot evaluations.



Paperid:5656
Authors:Faeze Moradi Kalarde, Ali Bereyhi, Ben Liang, Min Dong
Abstract:
Differential privacy (DP) limits the impact of individual training data samples by clipping the sample gradients. Conventional clipping methods often assign unequal weights to the gradients, which prevents excessive magnitude attenuation after clipping. This however leads to a notable direction mismatch between the true batch gradient and the aggregation of the clipped gradients, causing training steps to move towards less desirable areas of the loss landscape. This work proposes a novel clipping method, termed adaptive sigmoid (AdaSig), which uses a sigmoid function with an adjustable saturation slope to clip the sample gradients. The slope is learned from data during the training phase to account for variations in sample gradient statistics over iterations. This enables the algorithm to control the direction mismatch while avoiding excessive reduction in the magnitude of the aggregation of the clipped gradients. Despite AdaSig's adaptive nature, our convergence analysis shows that DP-SGD with AdaSig clipping retains the best-known convergence rate in the non-convex loss setting. Evaluating AdaSig on sentence and image classification tasks across eight datasets shows that it consistently improves learning performance compared with established clipping methods.



Paperid:5657
Authors:Fu Luo, Yaoxin Wu, Zhi Zheng, Zhenkun Wang
Abstract:
Recent neural combinatorial optimization (NCO) methods have shown promising problem-solving ability without requiring domain-specific expertise. Most existing NCO methods use training and testing data with a fixed constraint value and lack research on the effect of constraint tightness on the performance of NCO methods. This paper takes the capacity-constrained vehicle routing problem (CVRP) as an example to empirically analyze the NCO performance under different tightness degrees of the capacity constraint. Our analysis reveals that existing NCO methods overfit the capacity constraint, and they can only perform satisfactorily on a small range of the constraint values but poorly on other values. To tackle this drawback of existing NCO methods, we develop an efficient training scheme that explicitly considers varying degrees of constraint tightness and proposes a multi-expert module to learn a generally adaptable solving strategy. Experimental results show that the proposed method can effectively overcome the overfitting issue, demonstrating superior performances on the CVRP and CVRP with time windows (CVRPTW) with various constraint tightness degrees.



Paperid:5658
Authors:Sophia Simeng Han, Stephen Xia, Grant Zhang, Howard Dai, Chen Liu, Lichang Chen, Hoang H Nguyen, Hongyuan Mei, Jiayuan Mao, R. Thomas McCoy
Abstract:
Accuracy remains a standard metric for evaluating AI systems, but it offers limited insight into how models arrive at their solutions. In this work, we introduce a benchmark based on brainteasers written in long narrative form to probe more deeply into the types of reasoning strategies that models use. Brainteasers are well-suited for this goal because they can be solved with multiple approaches, such as a few-step solution that uses a creative insight or a longer solution that uses more brute force. We investigate large language models (LLMs) across multiple layers of reasoning, focusing not only on correctness but also on the quality and creativity of their solutions. We investigate many aspects of the reasoning process: (1) semantic parsing of the brainteasers into precise mathematical competition style formats; (2) self-correcting solutions based on gold solutions; (3) producing step-by-step sketches of solutions; and (4) making use of hints.We find that LLMs are in many cases able to find creative, insightful solutions to brainteasers, suggesting that they capture some of the capacities needed to solve novel problems in creative ways. Nonetheless, there also remain situations where they rely on brute force despite the availability of more efficient, creative solutions, highlighting a potential direction for improvement in the reasoning abilities of LLMs.



Paperid:5659
Authors:boheng liu, Ziyu Li, Chenghua Duan, YuTian Liu, Zhuo Wang, Xiuxing Li, Qing Li, Xia Wu
Abstract:
Open-world reinforcement learning challenges agents to develop intelligent behavior in vast exploration spaces. Recent approaches like LS-Imagine have advanced the field by extending imagination horizons through jumpy state transitions, yet remain limited by fixed exploration mechanisms and static jump thresholds that cannot adapt across changing task phases, resulting in inefficient exploration and lower completion rates. Humans demonstrate remarkable capabilities in open-world decision-making through a chain-like process of task decomposition, selective memory utilization, and adaptive uncertainty regulation. Inspired by human decision-making processes, we present Cognitive Predictive Processing (CPP), a novel framework that integrates three neurologically-inspired systems: a phase-adaptive cognitive controller that dynamically decomposes tasks into exploration, approach, and completion phases with adaptive parameters; a dual-memory integration system implementing dual-modal memory that balances immediate context with selective long-term storage; and an uncertainty-modulated prediction regulator that continuously updates environmental predictions to modulate exploration behavior. Comprehensive experiments in MineDojo demonstrate that these human-like decision-making strategies enhance performance over recent techniques, with success rates improving by an average of 4.6\% across resource collection tasks while reducing task completion steps by an average of 7.1\%. Our approach bridges cognitive neuroscience and reinforcement learning, excelling in complex scenarios that require sustained exploration and strategic adaptation while demonstrating how neural-inspired models can solve key challenges in open-world AI systems. Our main code has been anonymously uploaded to \url{https://anonymous.4open.science/r/CPP} without any author information.



Paperid:5660
Authors:Vivek Myers, Bill Zheng, Anca Dragan, Kuan Fang, Sergey Levine
Abstract:
Effective task representations should facilitate compositionality, such that after learning a variety of basic tasks, an agent can perform compound tasks consisting of multiple steps simply by composing the representations of the constituent steps together. While this is conceptually simple and appealing, it is not clear how to automatically learn representations that enable this sort of compositionality. We show that learning to associate the representations of current and future states with a temporal alignment loss can improve compositional generalization, even in the absence of any explicit subtask planning or reinforcement learning. We evaluate our approach across diverse robotic manipulation tasks as well as in simulation, showing substantial improvements for tasks specified with either language or goal images.



Paperid:5661
Authors:Yingqing Guo, Yukang Yang, Hui Yuan, Mengdi Wang
Abstract:
Training-free guidance enables controlled generation in diffusion and flow models, but most methods rely on gradients and assume differentiable objectives. This work focuses on training-free guidance addressing challenges from non-differentiable objectives and discrete data distributions. We propose TreeG: Tree Search-Based Path Steering Guidance, applicable to both continuous and discrete settings in diffusion and flow models. TreeG offers a unified framework for training-free guidance by proposing, evaluating, and selecting candidates at each step, enhanced with tree search over active paths and parallel exploration. We comprehensively investigate the design space of TreeG over the candidate proposal module and the evaluation function, instantiating TreeG into three novel algorithms. Our experiments show that TreeG consistently outperforms top guidance baselines in symbolic music generation, small molecule design, and enhancer DNA design with improvements of 29.01%, 16.6%, and 18.43%. Additionally, we identify an inference-time scaling law showing TreeG's scalability in inference-time computation.



Paperid:5662
Authors:Reece Keller, Alyn Kirsch, Felix Pei, Xaq Pitkow, Leo Kozachkov, Aran Nayebi
Abstract:
Autonomy is a hallmark of animal intelligence, enabling adaptive and intelligent behavior in complex environments without relying on external reward or task structure.Existing reinforcement learning approaches to exploration in sparse reward and reward-free environments, including class of methods known asintrinsic motivation, exhibit inconsistent exploration patterns and thus fail to produce robust autonomous behaviors observed in animals.Moreover, systems neuroscience has largely overlooked the neural basis of autonomy, focusing instead on experimental paradigms where animals are motivated by external reward rather than engaging in unconstrained, naturalistic and task-independent behavior. To bridge these gaps, we introduce a novel model-based intrinsic drive explicitly designed to capture robust autonomous exploration observed in animals. Our method (3M-Progress) motivates naturalistic behavior by tracking divergence between the agent's current world model and an ethological prior. We demonstrate that artificial embodied agents trained with 3M-Progress capture the explainable variance in behavioral patterns and whole-brain neural-glial dynamics recorded from autonomously-behaving larval zebrafish, introducing the first goal-driven model of neural-glial computation.Our findings establish a computational framework connecting model-based intrinsic motivation to naturalistic behavior, providing a foundation for building artificial agents with animal-like autonomy.



Paperid:5663
Authors:Hanyin Wang, Zhenbang Wu, Gururaj Kolar, Hariprasad Korsapati, Brian Bartlett, Bryan Hull, Jimeng Sun
Abstract:
Diagnosis-Related Group (DRG) codes are essential for hospital reimbursement and operations but require labor-intensive assignment. Large Language Models (LLMs) struggle with DRG coding due to the out-of-distribution (OOD) nature of the task: pretraining corpora rarely contain private clinical or billing data. We introduce DRG-Sapphire, which uses large-scale reinforcement learning (RL) for automated DRG coding from clinical notes. Built on Qwen2.5-7B and trained with Group Relative Policy Optimization (GRPO) using rule-based rewards, DRG-Sapphire introduces a series of RL enhancements to address domain-specific challenges not seen in previous mathematical tasks. Our model achieves state-of-the-art accuracy on the MIMIC-IV benchmark and generates physician-validated reasoning for DRG assignments, significantly enhancing explainability. Our study further sheds light on broader challenges of applying RL to knowledge-intensive, OOD tasks. We observe that RL performance scales approximately linearly with the logarithm of the number of supervised fine-tuning (SFT) examples, suggesting that RL effectiveness is fundamentally constrained by the domain knowledge encoded in the base model. For OOD tasks like DRG coding, strong RL performance requires sufficient knowledge infusion prior to RL. Consequently, scaling SFT may be more effective and computationally efficient than scaling RL alone for such tasks.



Paperid:5664
Authors:Runze Xia, Richard Jiang
Abstract:
Recent advances in Vision Transformers (ViTs) have shown remarkable performance across vision tasks, yet their deep, uniform layer structure introduces significant computational overhead. In this work, we explore the emergent dynamics of ViT layers through the lens of energy-based memory systems, drawing a connection between self-attention and modern Hopfield networks. We introduce a novel metric—Layer Instability Index (LII)—derived from the operational softmax mode and its variability, to quantify the metastability of each Transformer layer over time. Our analysis reveals that certain layers exhibit consistent convergence to attractor-like states, suggesting functional specialisation and early stabilisation. Leveraging this insight, we propose an adaptive training framework that dynamically freezes or skips stable layers based on their energy landscape behavior. Our method reduces training costs while maintaining or improving accuracy. Extensive experiments on ViT-S/B/L on CUB-200-2011, CIFAR-10/100, Food-101, Stanford Dogs, and Beans demonstrate the generality and efficiency of our approach. This work provides new theoretical and practical perspectives for energy-aware optimisation of deep Transformer models.



Paperid:5665
Authors:Yu Li, Jin Jiang, Jianhua Zhu, Shuai Peng, Baole, Yuxuan Zhou, Liangcai Gao
Abstract:
Handwritten Mathematical Expression Recognition (HMER) remains a persistent challenge in Optical Character Recognition(OCR) due to the inherent freedom of symbol layout and variability in handwriting styles. Prior methods have faced performance bottlenecks, proposing isolated architectural modifications that are difficult to integrate coherently into a unified framework. Meanwhile, recent advances in pretrained vision-language models (VLMs) have demonstrated strong cross-task generalization, offering a promising foundation for developing unified solutions. In this paper, we introduce Uni-MuMER, which fully fine-tunes the Qwen2.5-VL-3B model for the HMER task without modifying its architecture, effectively injecting domain-specific knowledge into a generalist framework. Our method integrates three data-driven tasks: Tree-Aware Chain-of-Thought (Tree-CoT) for structured spatial reasoning, Error-Driven Learning (EDL) for reducing confusion among visually similar characters, and Symbol Counting (SC) for improving recognition consistency in long expressions. Experiments on the CROHME and HME100K datasets show that Uni-MuMER achieves new state-of-the-art performance, surpassing the best lightweight specialized model SSAN by 16.31\% and the top-performing VLM Gemini2.5-flash by 24.42\% in the zero-shot setting.



Paperid:5666
Authors:Dongsheng Yuan, Xie Zhang, Weiying Hou, Sheng Lyu, Luca Jiang-Tao Yu, Yuemin Yu, Chengxiao Li, Chenshu Wu
Abstract:
We introduce OctoNet, a large-scale, multi-modal, multi-view human activity dataset designed to advance human activity understanding and multi-modal learning. OctoNet comprises 12 heterogeneous modalities (including RGB, depth, thermal cameras, infrared arrays, audio, millimeter-wave radar, Wi-Fi, IMU, and more) recorded from 41 participants under multi-view sensor setups, yielding over 67.72M synchronized frames. The data encompass 62 daily activities spanning structured routines, freestyle behaviors, human-environment interaction, healthcare tasks, etc. Critically, all modalities are annotated by high-fidelity 3D pose labels captured via a professional motion-capture system, allowing precise alignment and rich supervision across sensors and views. OctoNet is one of the most comprehensive datasets of its kind, enabling a wide range of learning tasks such as human activity recognition, 3D pose estimation, multi-modal fusion, cross-modal supervision, and sensor foundation models. Extensive experiments have been conducted to demonstrate the sensing capacity using various baselines. OctoNet offers a unique and unified testbed for developing and benchmarking generalizable, robust models for human-centric perceptual AI.



Paperid:5667
Authors:Yixiao Chen, Shikun Sun, Jianshu Li, Ruoyu Li, Zhe Li, Junliang Xing
Abstract:
Adversarial attacks are widely used to evaluate model robustness, and in black-box scenarios, the transferability of these attacks becomes crucial. Existing generator-based attacks have excellent generalization and transferability due to their instance-agnostic nature. However, when training generators for multi-target tasks, the success rate of transfer attacks is relatively low due to the limitations of the model's capacity. To address these challenges, we propose a novel Dual-Flow framework for multi-target instance-agnostic adversarial attacks, utilizing Cascading Distribution Shift Training to develop an adversarial velocity function. Extensive experiments demonstrate that Dual-Flow significantly improves transferability over previous multi-target generative attacks. For example, it increases the success rate from Inception-v3 to ResNet-152 by 34.58%. Furthermore, our attack method shows substantially stronger robustness against defense mechanisms, such as adversarially trained models.



Paperid:5668
Authors:Zhihao Wang, Yiqun Xie, Cooper Li, Ruichen Wang, Lei Ma, George Hurtt, Xiaowei Jia, Gengchen Mai, Zhili Li
Abstract:
Monitoring individual tree mortality at scale has been found crucial for understanding forest loss, ecosystem resilience, carbon fluxes, and climate-induced impacts. However, the fine-granularity monitoring faces major challenges on both the data and methodology side because: (1) finding isolated individual-level tree deaths require high-resolution remote sensing images with broad coverage, and (2) compared to regular geo-objects (e.g., buildings), dead trees often exhibit weaker contrast and high variability across tree types, landscapes and ecosystems. Existing datasets on tree mortality primarily rely on moderate-resolution satellite imagery (e.g., 30m resolution), which aims to detect large-patch wipe-outs but is unable to recognize individual-level tree mortality events. Several efforts explored alternatives via very-high-resolution drone imagery. However, drone images are highly expensive and can only be collected at local scales, which are not suitable for national-scale applications and beyond. To bridge the gaps, we introduce TreeFinder, the first high-resolution remote sensing benchmark dataset designed for individual-level tree mortality mapping across the Contiguous United States (CONUS). Specifically, the dataset uses NAIP imagery at 0.6m-resolution that provides wall-to-wall coverage of the entire CONUS. TreeFinder contains images with pixel-level labels generated via extensive manual annotation that covers forested areas in 48 different states with over 23,000 hectares. All annotations are rigorously validated using multi-temporal NAIP images and auxiliary vegetation indices from remote sensing imagery. Moreover, TreeFinder includes multiple evaluation scenarios to test the models' ability in generalizing across different geographic regions, climate zones, and forests with different plant function types. Finally, we develop benchmarks using a suite of semantic segmentation models, including both convolutional architectures and more recent foundation models based on vision transformers for general and remote sensing images.



Paperid:5669
Authors:Rishabh Mondal, Jeet Parab, Heer Kubadia, Shataxi Dubey, Shardul Junagade, Zeel Bharatkumar Patel, Nipun Batra
Abstract:
Air pollution was responsible for 2.6 million deaths across South Asia in 2021 alone, with brick manufacturing contributing significantly to this burden. In particular, the Indo-Gangetic Plain; a densely populated and highly polluted region spanning northern India, Pakistan, Bangladesh, and parts of Afghanistan sees brick kilns contributing 8--14\% of ambient air pollution. Traditional monitoring approaches, such as field surveys and manual annotation using tools like Google Earth Pro, are time and labor-intensive. Prior ML-based efforts for automated detection have relied on costly high-resolution commercial imagery and non-public datasets, limiting reproducibility and scalability. In this work, we introduce SentinelKilnDB, a publicly available, hand-validated benchmark of 62,671 brick kilns spanning three kiln types Fixed Chimney Bull’s Trench Kiln (FCBK), Circular FCBK (CFCBK), and Zigzag kilns—annotated with oriented bounding boxes (OBBs) across 2.8 million square km. using free and globally accessible Sentinel-2 imagery. We benchmark state-of-the-art oriented object detection models and evaluate generalization across in-region, out-of-region and super-resolution settings. SentinelKilnDB enables rigorous evaluation of geospatial generalization and robustness for low-resolution object detection, and provides a new testbed for ML models addressing real-world environmental and remote sensing challenges at a continental scale.



Paperid:5670
Authors:Sicong Leng, Yun Xing, Zesen Cheng, Yang Zhou, Hang Zhang, Xin Li, Deli Zhao, Shijian Lu, Chunyan Miao, Lidong Bing
Abstract:
Recent advancements in large multimodal models (LMMs) have significantly enhanced performance across diverse tasks, with ongoing efforts to further integrate additional modalities such as video and audio. However, most existing LMMs remain vulnerable to hallucinations, the discrepancy between the factual multimodal input and the generated textual output, which has limited their applicability in various real-world scenarios. This paper presents the first systematic investigation of hallucinations in LMMs involving the three most common modalities: language, visual, and audio. Our study reveals two key contributors to hallucinations: overreliance on unimodal priors and spurious inter-modality correlations. To address these challenges, we introduce the benchmark The Curse of Multi-Modalities (CMM), which comprehensively evaluates hallucinations in LMMs, providing a detailed analysis of their underlying issues. Our findings highlight key vulnerabilities, including imbalances in modality integration and biases from training data, underscoring the need for balanced cross-modal learning and enhanced hallucination mitigation strategies. Based on our observations and findings, we suggest potential research directions that could enhance the reliability of LMMs.



Paperid:5671
Authors:Minje Kim, Tae-Kyun Kim
Abstract:
Reconstructing detailed hand avatars plays a crucial role in various applications. While prior works have focused on capturing high-fidelity hand geometry, they heavily rely on high-resolution multi-view image inputs and struggle to generalize on low-resolution images. Multi-view image super-resolution methods have been proposed to enforce 3D view consistency. These methods, however, are limited to static objects/scenes with fixed resolutions and are not applicable to articulated deformable hands. In this paper, we propose SRHand (Super-Resolution Hand), the method for reconstructing detailed 3D geometry as well as textured images of hands from low-resolution images. SRHand leverages the advantages of implicit image representation with explicit hand meshes. Specifically, we introduce a geometric-aware implicit image function (GIIF) that learns detailed hand prior by upsampling the coarse input images. By jointly optimizing the implicit image function and explicit 3D hand shapes, our method preserves multi-view and pose consistency among upsampled hand images, and achieves fine-detailed 3D reconstruction (wrinkles, nails). In experiments using the InterHand2.6M and Goliath datasets, our method significantly outperforms state-of-the-art image upsampling methods adapted to hand datasets, and 3D hand reconstruction methods, quantitatively and qualitatively. The code will be publicly available.



Paperid:5672
Authors:Xiatoian Liu, Ali Pesaranghader, Jaehong Kim, Tanmana Sadhu, Hyejeong Jeon, Scott Sanner
Abstract:
The ability to actively acquire information is essential for open-world planning under partial observability and incomplete knowledge. Existing embodied AI systems typically rely on passive strategies that exhaustively collect object and relational information. However, such passive knowledge acquisition becomes impractical in visually complex domains. For instance, a typical household may contain hundreds of uniquely configured objects with unique configurations. Therefore, open-world agents must be able to actively identify which objects are relevant to the task at hand. In this work, we introduce ActiveVOI, a novel zero-shot framework for open-world embodied planning that emphasizes object-centric active knowledge acquisition. ActiveVOI leverages Lifted Regression to generate compact subgoal descriptions that identify task-relevant objects. It also provides a principled approach to quantify the utility of sensing objects using the theory of Value of Information (VOI), guided by commonsense knowledge from large language and vision-language models (LLMs/VLMs). ActiveVOI is evaluated on the visual ALFWorld benchmark, showing substantial improvements over existing LLM- and VLM-based planning methods, and notably even outperforming VLMs that are fine-tuned on ALFWorld data. This work establishes a principled foundation for building embodied agents that actively and efficiently acquire knowledge to plan in open-world environments.



Paperid:5673
Authors:Sima Noorani, Shayan Kiyani, George J. Pappas, Hamed Hassani
Abstract:
Uncertainty quantification (UQ) is essential for safe deployment of generative AI models such as large language models (LLMs), especially in high-stakes applications. Conformal prediction (CP) offers a principled UQ framework, but classical methods focus on regression and classification, relying on geometric distances or softmax scores—tools that presuppose structured outputs. We depart from this paradigm by studying CP in a query-only setting, where prediction sets must be constructed solely from finite queries to a black-box generative model, introducing a new trade-off between coverage, test-time query budget, and informativeness. We introduceConformal Prediction with Query Oracle(CPQ), a framework characterizing the optimal interplay between these objectives. Our finite-sample algorithm is grounded in two principles: the first characterizes the optimal query policy, and the second the optimal mapping from queried samples to prediction sets, remarkably connecting both to the classicalmissing mass problemin statistics. Fine-grained experiments on three real-world open-ended tasks and two LLMs, show CPQ's applicability toany black-box LLMand highlight: (1) individual contribution of each principle to CPQ’s performance, and (2) CPQ's ability to yield significantly more informative prediction sets than existing conformal methods for language UQ.



Paperid:5674
Authors:Xiang Fang, Wanlong Fang, Changshuo Wang
Abstract:
Vision-Language Navigation in Continuous Environments (VLN-CE) poses a formidable challenge for autonomous agents, requiring seamless integration of natural language instructions and visual observations to navigate complex 3D indoor spaces. Existing approaches often falter in long-horizon tasks due to limited scene understanding, inefficient planning, and lack of robust decision-making frameworks. We introduce the \textbf{Hierarchical Semantic-Augmented Navigation (HSAN)} framework, a groundbreaking approach that redefines VLN-CE through three synergistic innovations. First, HSAN constructs a dynamic hierarchical semantic scene graph, leveraging vision-language models to capture multi-level environmental representations—from objects to regions to zones—enabling nuanced spatial reasoning. Second, it employs an optimal transport-based topological planner, grounded in Kantorovich's duality, to select long-term goals by balancing semantic relevance and spatial accessibility with theoretical guarantees of optimality. Third, a graph-aware reinforcement learning policy ensures precise low-level control, navigating subgoals while robustly avoiding obstacles. By integrating spectral graph theory, optimal transport, and advanced multi-modal learning, HSAN addresses the shortcomings of static maps and heuristic planners prevalent in prior work. Extensive experiments on multiple challenging VLN-CE datasets demonstrate that HSAN achieves state-of-the-art performance, with significant improvements in navigation success and generalization to unseen environments.



Paperid:5675
Authors:Parshin Shojaee, Iman Mirzadeh, Keivan Alizadeh vahid, Maxwell Horton, Samy Bengio, Mehrdad Farajtabar
Abstract:
Recent generations of frontier language models have introduced Large Reasoning Models (LRMs) that generate detailed thinking processes before providing answers. While these models demonstrate improved performance on reasoning benchmarks, their fundamental capabilities, scaling properties, and limitations remain insufficiently understood. Current evaluations primarily focus on established mathematical and coding benchmarks, emphasizing final answer accuracy. However, this evaluation paradigm often suffers from data contamination and does not provide insights into the reasoning traces' structure and quality. In this work, we systematically investigate these gaps with the help of controllable puzzle environments that allow precise manipulation of compositional complexity while maintaining consistent logical structures. This setup enables the analysis of not only final answers but also the internal reasoning traces, offering insights into how LRMs think. Through extensive experimentation across diverse puzzles, we show that frontier LRMs face a complete accuracy collapse beyond certain complexities. Moreover, they exhibit a counterintuitive scaling limit: their reasoning effort increases with problem complexity up to a point, then declines despite having an adequate token budget. By comparing LRMs with their standard LLM counterparts under equivalent inference compute, we identify three performance regimes: (1) low-complexity tasks where standard models surprisingly outperform LRMs, (2) medium-complexity tasks where additional thinking in LRMs demonstrates advantage, and (3) high-complexity tasks where both models experience complete collapse. We also investigate the reasoning traces in more depth, studying the patterns of explored solutions and analyzing the models' computational behavior, shedding light on their strengths, limitations, and ultimately raising crucial questions about their true reasoning capabilities.



Paperid:5676
Authors:Daking Rai, Samuel Miller, Kevin Moran, Ziyu Yao
Abstract:
Abstract:Despite remarkable advances in coding capabilities, language models (LMs) still struggle with simple syntactic tasks such as generating balanced parentheses. In this study, we investigate the underlying mechanisms behind the persistence of these errors across LMs of varying sizes (124M–7B) to both understand and mitigate the errors. Our study reveals that LMs rely on a number of components (attention heads and FF neurons) that independently make their own predictions. While some components reliably promote correct answers across a generalized range of inputs (i.e., implementing "sound mechanisms''), others are less reliable and introduce noise by promoting incorrect tokens (i.e., implementing "faulty mechanisms''). Errors occur when the faulty mechanisms overshadow the sound ones and dominantly affect the predictions. Motivated by this insight, we introduce RASteer, a steering method to systematically identify and increase the contribution of reliable components for improving model performance. RASteer substantially improves performance on balanced parentheses tasks, boosting accuracy of some models from $0$\% to around $100$\% without impairing the models' general coding ability. We further demonstrate its broader applicability in arithmetic reasoning tasks, achieving performance gains of up to around $20$\%.



Paperid:5677
Authors:Ruiqi Wang, Dezhong Zhao, Ziqin Yuan, Tianyu Shao, Guohua Chen, Dominic Kao, Sungeun Hong, Byung-Cheol Min
Abstract:
Preference-based reinforcement learning (PbRL) has emerged as a promising paradigm for teaching robots complex behaviors without reward engineering. However, its effectiveness is often limited by two critical challenges: the reliance on extensive human input and the inherent difficulties in resolving query ambiguity and credit assignment during reward learning. In this paper, we introduce PRIMT, a PbRL framework designed to overcome these challenges by leveraging foundation models (FMs) for multimodal synthetic feedback and trajectory synthesis. Unlike prior approaches that rely on single-modality FM evaluations, PRIMT employs a hierarchical neuro-symbolic fusion strategy, integrating the complementary strengths of vision-language models (VLMs) and large language models (LLMs) in evaluating robot behaviors for more reliable and comprehensive feedback. PRIMT also incorporates foresight trajectory generation to warm-start the trajectory buffer with bootstrapped samples, reducing early-stage query ambiguity, and hindsight trajectory augmentation for counterfactual reasoning with a causal auxiliary loss to improve credit assignment. We evaluate PRIMT on 2 locomotion and 6 manipulation tasks on various benchmarks, demonstrating superior performance over FM-based and scripted baselines. Website at https://sites.google.com/view/PRIMT.



Paperid:5678
Authors:Shubo Lin, Yutong Kou, Zirui Wu, Shaoru Wang, Bing Li, Weiming Hu, Jin Gao
Abstract:
Abstract:While existing query-based 3D end-to-end visual trackers integrate detection and tracking via the $\textit{tracking-by-attention}$ paradigm, these two chicken-and-egg tasks encounter optimization difficulties when sharing the same parameters. Our findings reveal that these difficulties arise due to two inherent constraints on the self-attention mechanism, i.e., over-deduplication for object queries and self-centric attention for track queries. In contrast, removing self-attention mechanism not only minimally impacts regression predictions of the tracker, but also tends to generate more latent candidate boxes. Based on these analyses, we present SynCL, a novel plug-and-play synergistic training strategy designed to co-facilitate multi-task learning for detection and tracking. Specifically, we propose a Task-specific Hybrid Matching module for a weight-shared cross-attention-based decoder that matches the targets of track queries with multiple object queries to exploit promising candidates overlooked by the self-attention mechanism. To flexibly select optimal candidates for the one-to-many matching, we also design a Dynamic Query Filtering module controlled by model training status. Moreover, we introduce Instance-aware Contrastive Learning to break through the barrier of self-centric attention for track queries, effectively bridging the gap between detection and tracking. Without additional inference costs, SynCL consistently delivers improvements in various benchmarks and achieves state-of-the-art performance with $58.9\%$ AMOTA on the nuScenes dataset. Code and raw results will be publicly available.



Paperid:5679
Authors:Hongkang Zhang, Shao-Lun Huang, Ercan KURUOGLU, Yanlong Wang
Abstract:
Significant progress has been achieved using Vision Transformers (ViTs) in computer vision. However, challenges persist in modeling multi-scale spatial relationships, hindering effective integration of fine-grained local details and long-range global dependencies. To address this limitation, a Multi-Kernel Correlation-Attention Vision Transformer (MK-CAViT) grounded in the Hirschfeld-Gebelein-Rényi (HGR) theory was proposed, introducing three key innovations. A parallel multi-kernel architecture was utilized to extract multi-scale features through small, medium, and large kernels, overcoming the single-scale constraints of conventional ViTs. The cross-scale interactions were enhanced via an HGR-correlation attention mechanism, which modeled nonlinear dependencies and applied adaptive scaling to weigh connections and refine contextual reasoning. Additionally, a stable multi-scale fusion strategy was adopted, integrating dynamic normalization and staged learning to mitigate gradient variance, progressively fusing local and global contexts, and improving training stability. The experimental results on ImageNet, COCO, and ADE20K validated the superiority of MK-CAViT in classification, detection, and segmentation, surpassing state-of-the-art baselines in capturing complex spatial relationships while maintaining efficiency. These contributions can establish a robust framework for visual representation learning and address the longstanding limitations of ViTs.



Paperid:5680
Authors:Yangtao Zhou, Hua Chu, chen, Ziwen Wang, Jiacheng Liu, Jianan Li, Yueying Feng, Xiangming Li, Zihan Han, Qingshan Li
Abstract:
Adaptive learning, which requires the in-depth understanding of students' learning processes and rational planning of learning resources, plays a crucial role in intelligent education. However, how to effectively model these two processes and seamlessly integrate them poses significant implementation challenges for adaptive learning. As core learning resources, exercises have the potential to diagnose students' knowledge states during the learning processes and provide personalized learning recommendations to strengthen students' knowledge, thereby serving as a bridge to boost student-oriented adaptive learning. Therefore, we introduce a novel task called Knowledge-aware Exercise Generative Recommendation (KEGR). It aims to dynamically infer students' knowledge states from their past exercise responses and customizably generate new exercises. To achieve KEGR, we propose an adaptive multi-agent cooperation framework, called ExeGen, inspired by the excellent reasoning and generative capabilities of LLM-based AI agents. Specifically, ExeGen coordinates four specialized agents for supervision, knowledge state perception, exercise generation, and quality refinement through an adaptive loop workflow pipeline. More importantly, we devise two enhancement mechanisms in ExeGen: 1) A human-simulated knowledge perception mechanism mimics students' cognitive processes and generates interpretable knowledge state descriptions via demonstration-based In-Context Learning (ICL). In this mechanism, a dual-matching strategy is further designed to retrieve highly relevant demonstrations for reliable ICL reasoning. 2) An exercise generation-adversarial mechanism collaboratively refines exercise generation leveraging a group of quality evaluation expert agents via iterative adversarial feedback. Finally, a comprehensive evaluation protocol is carefully designed to assess ExeGen. Extensive experiments on real-world educational datasets and a practical deployment in college education demonstrate the effectiveness and superiority of ExeGen. The code is available at https://anonymous.4open.science/r/exeGen-6F48.



Paperid:5681
Authors:Jiawei Sun, Shuai Zhang, Hongkang Li, Meng Wang
Abstract:
Contrastive learning is a powerful framework for learning discriminative representations from image-text pairs. Despite its success, its theoretical foundations, especially when the image-text pair exhibits misalignment, remain underexplored. This paper provides the first theoretical analysis of contrastive learning under data misalignment, proving how the ground-truth modality-paired features are amplified while spurious features are suppressed through the training dynamics analysis. Specifically, we study two nonlinear encoders trained jointly with a contrastive loss and demonstrate that noisy (or misaligned) data pairs result in mixed representations and degrade the model's generalization ability. In contrast, recaptioning and filtering improve the data alignment, which in turn purifies the features learned by neurons and subsequently enhances generalization. Our analysis identifies feature purity as a key factor in the success of contrastive learning and offers insights into how data quality and training procedures impact representation learning and downstream generalization. Theoretical insights are supported by experiments on standard benchmarks.



Paperid:5682
Authors:Qiwei Wang, Shaoxun Wu, Yujiao Shi
Abstract:
This paper addresses the problem of weakly supervised cross-view localization, where the goal is to estimate the pose of a ground camera relative to a satellite image with noisy ground truth annotations. A common approach to bridge the cross-view domain gap for pose estimation is Bird’s-Eye View (BEV) synthesis. However, existing methods struggle with height ambiguity due to the lack of depth information in ground images and satellite height maps. Previous solutions either assume a flat ground plane or rely on complex models, such as cross-view transformers.We propose BevSplat, a novel method that resolves height ambiguity by using feature-based Gaussian primitives. Each pixel in the ground image is represented by a 3D Gaussian with semantic and spatial features, which are synthesized into a BEV feature map for relative pose estimation.We validate our method on the widely used KITTI and VIGOR datasets, which include both pinhole and panoramic query images. Experimental results show that BevSplat significantly improves localization accuracy over prior approaches.



Paperid:5683
Authors:Maonan Wang, Yirong Chen, Aoyu Pang, Yuxin Cai, Chung Shue Chen, Yuheng Kan, Man On Pun
Abstract:
Traffic signal control (TSC) is a core challenge in urban mobility, where real-time decisions must balance efficiency and safety. Existing methods—ranging from rule-based heuristics to reinforcement learning (RL)—often struggle to generalize to complex, dynamic, and safety-critical scenarios. We introduce \textbf{VLMLight}, a novel TSC framework that integrates vision-language meta-control with dual-branch reasoning. At the core of VLMLight is the first image-based traffic simulator that enables multi-view visual perception at intersections, allowing policies to reason over rich cues such as vehicle type, motion, and spatial density. A large language model (LLM) serves as a safety-prioritized meta-controller, selecting between a fast RL policy for routine traffic and a structured reasoning branch for critical cases. In the latter, multiple LLM agents collaborate to assess traffic phases, prioritize emergency vehicles, and verify rule compliance. Experiments show that VLMLight reduces waiting times for emergency vehicles by up to 65% over RL-only systems, while preserving real-time performance in standard conditions with less than 1% degradation. VLMLight offers a scalable, interpretable, and safety-aware solution for next-generation traffic signal control.



Paperid:5684
Authors:Jiacai Liu, Chaojie Wang, Chris Liu, Liang Zeng, Rui Yan, Yiwen Sun, Yang Liu
Abstract:
The role of reinforcement learning (RL) in enhancing the reasoning of large language models (LLMs) is becoming increasingly significant. Despite the success of RL in many scenarios, there are still many challenges in improving the reasoning of LLMs. One key challenge is the sparse reward, which introduces more training variance in policy optimization and makes it difficult to obtain a good estimation for value function in Actor-Critic (AC) methods. To address these issues, we introduce Direct Advantage-Based Policy Optimization (DAPO), a novel step-level offline RL algorithm with theoretical guarantees for enhancing the reasoning abilities of LLMs. Unlike response-level methods (such as DPO and GRPO) that the update directions of all reasoning steps are governed by the outcome reward uniformly, DAPO employs a critic function to provide step-level dense signals for policy optimization. Additionally, the actor and critic in DAPO are trained independently, ensuring that critic is a good estimation of true state value function and avoiding the co-training instability observed in standard AC methods. We train DAPO on mathematical and code problems and then evaluate its performance on multiple benchmarks. Our results show that DAPO can effectively enhance the mathematical and code capabilities on both SFT models and RL models, demonstrating the effectiveness of DAPO.



Paperid:5685
Authors:Tianle Zhang, Wanlong Fang, Jonathan Woo, Paridhi Latawa, Deepak Subramanian, Alvin Chan
Abstract:
The remarkable performance of Large Language Models (LLMs) can be enhanced with test-time computation, which relies on external tools and even other deep learning models.However, existing approaches for integrating non-text modality representations into LLMs typically require additional costly supervised training, restricting on-the-fly adaptation to new domains and modalities. In this work, we explore the feasibility of integrating representations from non-text foundational models (FMs) into text-based LLMs in a training-free manner. We propose In-Context Representation Learning (ICRL) as a proof-of-concept to allow LLMs to adaptively utilize non-text modality representations with few-shot learning. Unlike traditional in-context learning, which incorporates text-label pairs, ICRL replaces text inputs with FM representations, enabling the LLM to perform multi-modal inference without fine-tuning. We evaluate ICRL on a suite of tasks in the molecular domain, investigating three core research questions: (i) how to map FM representations into LLMs in a training-free manner, (ii) what factors influence ICRL performance, and (iii) what mechanisms underlie the effectiveness of ICRL. To the best of our knowledge, ICRL is the first training-free framework for integrating non-text modality representations into text-based LLMs, presenting a promising direction for adaptable, multi-modal generalization.



Paperid:5686
Authors:Jinluan Yang, Dingnan Jin, Anke Tang, Li Shen, Didi Zhu, Zhengyu Chen, Ziyu Zhao, Daixin Wang, Qing Cui, Zhiqiang Zhang, Jun Zhou, Fei Wu, Kun Kuang
Abstract:
Achieving balanced alignment of large language models (LLMs) in terms of Helpfulness, Honesty, and Harmlessness (3H optimization) constitutes a cornerstone of responsible AI. Existing methods like data mixture strategies face limitations, including heavy reliance on expert knowledge and conflicting optimization signals. While model merging offers parameter-level conflict-resolution strategies through integrating specialized models' parameters, its potential for 3H optimization remains underexplored. This paper systematically compares the effectiveness of model merging and data mixture methods in constructing 3H-aligned LLMs for the first time, revealing previously overlooked collaborative and conflict relationships among the 3H dimensions and discussing the advantages and drawbacks of data mixture (\textit{data-level}) and model merging (\textit{parameter-level}) methods in mitigating the conflict for balanced 3H optimization. Specially, we propose a novel \textbf{R}eweighting \textbf{E}nhanced task \textbf{S}ingular \textbf{M}erging method, \textbf{RESM}, through outlier weighting and sparsity-aware rank selection strategies to address the challenges of preference noise accumulation and layer sparsity adaptation inherent in 3H-aligned LLM merging. Extensive evaluations can verify the effectiveness and robustness of RESM compared to previous data mixture (2\%-5\% gain) and model merging (1\%-3\% gain) methods in achieving balanced LLM alignment.



Paperid:5687
Authors:Ismael Villanueva Miranda, Zifan Gu, Donghan Yang, Kuroush Nezafati, Jingwei Huang, Peifeng Ruan, Xiaowei Zhan, Guanghua Xiao, Yang Xie
Abstract:
Large Language Models (LLMs) offer substantial promise for clinical natural language processing (NLP); however, a lack of standardized benchmarking methodologies limits their objective evaluation and practical translation. To address this gap, we introduce ClinBench, an open-source, multi-model, multi-domain benchmarking framework. ClinBench is designed for the rigorous evaluation of LLMs on important structured information extraction tasks (e.g., tumor staging, histologic diagnoses, atrial fibrillation, and social determinants of health) from unstructured clinical notes. The framework standardizes the evaluation pipeline by: (i) operating on consistently structured input datasets; (ii) employing dynamic, YAML-based prompting for uniform task definition; and (iii) enforcing output validation via JSON schemas, supporting robust comparison across diverse LLM architectures. We demonstrate ClinBench through a large-scale study of 11 prominent LLMs (e.g., GPT-4o series, LLaMA3 variants, Mixtral) across three clinical domains using configurations of public datasets (TCGA for lung cancer, MIMIC-IV-ECG for atrial fibrillation, and MIMIC notes for SDOH). Our results reveal significant performance-efficiency trade-offs. For example, when averaged across the four benchmarked clinical extraction tasks, GPT-3.5-turbo achieved a mean F1 score of 0.83 with a mean runtime of 16.8 minutes. In comparison, LLaMA3.1-70b obtained a similar mean F1 of 0.82 but required a substantially longer mean runtime of 42.7 minutes. GPT-4o-mini also presented a favorable balance with a mean F1 of 0.81 and a mean runtime of 13.4 minutes. ClinBench provides a unified, extensible framework and empirical insights for reproducible, fair LLM benchmarking in clinical NLP. By enabling transparent and standardized evaluation, this work advances data-centric AI research, informs model selection based on performance, cost, and clinical priorities, and supports the effective integration of LLMs into healthcare. The framework and evaluation code are publicly available at https://github.com/ismaelvillanuevamiranda/ClinBench/.



Paperid:5688
Authors:Raphael Boige, Amine Boumaza, Bruno Scherrer
Abstract:
Deterministic game-solving algorithms are conventionally analyzed in the light of their average-case complexity against a distribution of random game-trees, where leaf values are independently sampled from a fixed distribution. This simplified model enables uncluttered mathematical analysis, revealing two key properties: root value distributions asymptotically collapse to a single fixed value for finite-valued trees, and all reasonable algorithms achieve global optimality. However, these findings are artifacts of the model’s design—its long criticized independence assumption strips games of structural complexity, producing trivial instances where no algorithm faces meaningful challenges. To address this limitation, we introduce a new probabilistic model that incrementally constructs game-trees using a fixed level-wise conditional distribution. By enforcing ancestor dependency, a critical structural feature of real-world games, our framework generates problems with adjustable difficulty while retaining some form of analytical tractability. For several algorithms, including AlphaBeta and Scout, we derive recursive formulas characterizing their average-case complexities under this model. These allow us to rigorously compare algorithms on deep game-trees, where Monte-Carlo simulations are no longer feasible. While asymptotically, all algorithms seem to converge to identical branching factor (a result analogous to those of independence-based models), deep finite trees reveal stark differences: AlphaBeta incurs a significantly larger constant multiplicative factor compared to algorithms like Scout, leading to a substantial practical slowdown. Our framework sheds new light on classical game-solving algorithms, offering rigorous evidence and analytical tools to advance the understanding of these methods under a more realistic, challenging, and yet tractable model.



Paperid:5689
Authors:Enzhi Zhang, Peng Chen, Rui Zhong, Du Wu, Jun Igarashi, Isaac Lyngaas, Xiao Wang, Masaharu Munetomo, Mohamed Wahib
Abstract:
Abstract:This paper presents a novel approach to addressing the long-sequence problem in high-resolution medical images for Vision Transformers (ViTs). Using smaller patches as tokens can enhance ViT performance, but quadratically increases computation and memory requirements. Therefore, the common practice for applying ViTs to high-resolution images is either to: (a) employ complex sub-quadratic attention schemes or (b) use large to medium-sized patches and rely on additional mechanisms within the model to capture the spatial hierarchy of details. We propose Symmetrical Hierarchical Forest (SHF), a lightweight approach that adaptively patches the input image to increase token information density and encode hierarchical spatial structures into the input embedding. We then apply a reverse depatching scheme to the output embeddings of the transformer encoder, eliminating the need for convolution-based decoders. Unlike previous methods that modify attention mechanisms \wahib{or use a complex hierarchy of interacting models}, SHF can be retrofitted to any ViT model to allow it to learn the hierarchical structure of details in high-resolution images without requiring architectural changes. Experimental results demonstrate significant gains in computational efficiency and performance: on the PAIP WSI dataset, we achieved a 3$\sim$32$\times$ speedup or a 2.95\% to 7.03\% increase in accuracy (measured by Dice score) at a $64K^2$ resolution with the same computational budget, compared to state-of-the-art production models. On the 3D medical datasets BTCV and KiTS, training was 6$\times$ faster, with accuracy gains of 6.93\% and 5.9\%, respectively, compared to models without SHF.



Paperid:5690
Authors:Leslie Hwang
Abstract:
Physics-informed neural networks (PINNs) have emerged as a promising framework to develop complex scientific surrogate models, yet their scalability and accuracy often degrade in non-canonical geometries, such as non-rectangular domains or three-dimensional (3D) domains with high aspect ratios. These limitations hinder the broader adoption of vanilla PINNs in real-world, practical systems. In this work, we introduce a multi-domain PINN (MDPINN) framework designed to address the scalability and generalization challenges inherent in 3D non-rectangular domains governed by nonlinear fluid dynamics. The target domain consists of intersecting 3D fluid channels with a high aspect ratio, inducing complex flow features such as deflections, mixing, and recirculations. Our approach is grounded in two key innovations: 1) domain decomposition, which partitions the channel volumes into multiple cubic-like subdomains, each modeled by an individual PINN, 2) enforcement of global dynamics (MDPINN-GD), which ensures that the total mass flow rate entering the domain equals that exiting. These innovations reduce the complexity of the problem imposed on individual PINNs and guide effective network optimization toward physically consistent solutions throughout the domain. We demonstrate that our method achieves: 1) 74.8\% accuracy improvement over a single-network PINN, and 2) 52.9\% accuracy improvement over MDPINN that do not enforce global mass conservation. Furthermore, the MDPINN-GD framework exhibits accurate prediction even in highly complex regions-such as the channel intersecting zone and the outlet zone characterized by intense flow mixing and large velocity gradients-achieving maximum normalized mean absolute errors below 14.9\% for velocity predictions compared to simulation results. This work establishes a path towards scalable, physically grounded surrogate modeling approach that is extensible to multiphysics and high-dimensional scientific problems.



Paperid:5691
Authors:Shinsaku Sakaue, Taira Tsuchiya, Han Bao, Taihei Oki
Abstract:
Abstract:In online inverse linear optimization, a learner observes time-varying sets of feasible actions and an agent's optimal actions, selected by solving linear optimization over the feasible actions. The learner sequentially makes predictions of the agent's true linear objective function, and their quality is measured by the *regret*, the cumulative gap between optimal objective values and those achieved by following the learner's predictions. A seminal work by Bärmann et al. (2017) obtained a regret bound of $O(\sqrt{T})$, where $T$ is the time horizon. Subsequently, the regret bound has been improved to $O(n^4 \ln T)$ by Besbes et al. (2021, 2025) and to $O(n \ln T)$ by Gollapudi et al. (2021), where $n$ is the dimension of the ambient space of objective vectors. However, these logarithmic-regret methods are highly inefficient when $T$ is large, as they need to maintain regions specified by $O(T)$ constraints, which represent possible locations of the true objective vector. In this paper, we present the first logarithmic-regret method whose per-round complexity is independent of $T$; indeed, it achieves the best-known bound of $O(n \ln T)$. Our method is strikingly simple: it applies the online Newton step (ONS) to appropriate exp-concave loss functions. Moreover, for the case where the agent's actions are possibly suboptimal, we establish a regret bound of $O(n\ln T + \sqrt{\Delta_T n\ln T})$, where $\Delta_T$ is the cumulative suboptimality of the agent's actions. This bound is achieved by using MetaGrad, which runs ONS with $\Theta(\ln T)$ different learning rates in parallel. We also present a lower bound of $\Omega(n)$, showing that the $O(n\ln T)$ bound is tight up to an $O(\ln T)$ factor.



Paperid:5692
Authors:Ke Li, Yan Ding, Zhiqin Zhu, Shenhai Zheng
Abstract:
Heterogeneous federated learning enables collaborative training across clients under dual heterogeneity of models and data, posing challenges for effective knowledge transfer. Federated mutual learning employs proxy models to bridge cross-model knowledge exchange; however, existing methods remain limited to direct alignment between the outputs of private and proxy models, ignoring the deep discrepancies in representation and decision spaces between them. Such cognitive biases cause knowledge to be transferred only at shallow levels and trigger performance bottlenecks. To address this, this paper proposes FedKWAZ to identify and exploit Knowledge Weak-Aware Zones (KWAZ)—spatial zones of deep knowledge misalignment between private and proxy models, further refined into Semantic Weak-Aware Zones and Decision Weak-Aware Zones, which characterize cognitive misalignments in representation and decision spaces as focal targets for enhanced bidirectional distillation. FedKWAZ designs a Hierarchical Adaptive Patch Mixing (HAPM) mechanism to generate multiple mixed samples and employs a Knowledge Discrepancy Perceptron (KDP) to select the samples exhibiting the largest representation and decision discrepancies, thereby mining critical KWAZ. These modules are integrated into a two-stage mutual learning framework, achieving global class-level representation-decision consistency alignment and local KWAZ-guided refinement, structurally bridging cognitive biases across heterogeneous mutual learning models. Experimental results on multiple datasets and model configurations demonstrate the superior performance of FedKWAZ.



Paperid:5693
Authors:Yan Shu, Hangui Lin, Yexin Liu, Yan Zhang, Gangyan Zeng, Yan Li, Yu Zhou, Ser Nam Lim, Harry Yang, Nicu Sebe
Abstract:
Large Multimodal Models (LMMs) have achieved impressive progress in visual perception and reasoning. However, when confronted with visually ambiguous or non-semantic scene text, they often struggle to accurately spot and understand the content, frequently generating semantically plausible yet visually incorrect answers, which we refer to as semantic hallucination.In this work, we investigate the underlying causes of semantic hallucination and identify a key finding: Transformer layers in LLM with stronger attention focus on scene text regions are less prone to producing semantic hallucinations. Thus, we propose a training-free semantic hallucination mitigation framework comprising two key components: (1) ZoomText, a coarse-to-fine strategy that identifies potential text regions without external detectors; and (2) Grounded Layer Correction, which adaptively leverages the internal representations from layers less prone to hallucination to guide decoding, correcting hallucinated outputs for non-semantic samples while preserving the semantics of meaningful ones. To enable rigorous evaluation, we introduce TextHalu-Bench, a benchmark of over 1,730 samples spanning both semantic and non-semantic cases, with manually curated question–answer pairs designed to probe model hallucinations.Extensive experiments demonstrate that our method not only effectively mitigates semantic hallucination but also achieves strong performance on public benchmarks for scene text spotting and understanding.



Paperid:5694
Authors:Changyeol Lee, Yongho Shin, Hyung-Chan An
Abstract:
Clustering is a fundamental task in both machine learning and data mining. Among various methods, edge-colored clustering (ECC) has emerged as a useful approach for handling categorical data. Given a hypergraph with (hyper)edges labeled by colors, ECC aims to assign vertex colors to minimize the number of edges where the vertex color differs from the edge's color. However, traditional ECC has inherent limitations, as it enforces a nonoverlapping and exhaustive clustering. To tackle these limitations, three versions of ECC have been studied: Local ECC and Global ECC, which allow overlapping clusters, and Robust ECC, which accounts for vertex outliers. For these problems, both linear programming (LP) rounding algorithms and greedy combinatorial algorithms have been proposed. While these LP-rounding algorithms provide high-quality solutions, they demand substantial computation time; the greedy algorithms, on the other hand, run very fast but often compromise solution quality. In this paper, we present an algorithmic framework that combines the strengths of LP with the computational efficiency of combinatorial algorithms. Both experimental and theoretical analyses show that our algorithms efficiently produce high-quality solutions for all three problems: Local, Global, and Robust ECC. We complement our algorithmic contributions with complexity-theoretic inapproximability results and integrality gap bounds, which suggest that significant theoretical improvements are unlikely. Our results also answer two open questions previously raised in the literature.



Paperid:5695
Authors:Chensheng Dai, Shengjun Zhang, Min Chen, Yueqi Duan
Abstract:
3D Gaussian Splatting (3DGS) has demonstrated impressive performance in 3D scene reconstruction. Beyond novel view synthesis, it shows great potential for multi-view surface reconstruction. Existing methods employ optimization-based reconstruction pipelines that achieve precise and complete surface extractions. However, these approaches typically require dense input views and high time consumption for per-scene optimization. To address these limitations, we propose SurfaceSplat, a feed-forward framework that generates efficient and generalizable pixel-aligned Gaussian surfel representations from sparse-view images. We observe that conventional feed-forward structures struggle to recover accurate geometric attributes of Gaussian surfels because the spatial frequency of pixel-aligned primitives exceeds Nyquist sampling rates. Therefore, we propose a cross-view feature aggregation module based on the Nyquist sampling theorem. Specifically, we first adapt the geometric forms of Gaussian surfels with spatial sampling rate-guided low-pass filters. We then project the filtered surfels across all input views to obtain cross-view feature correlations. By processing these correlations through a specially designed feature fusion network, we can finally regress Gaussian surfels with precise geometry. Extensive experiments on DTU reconstruction benchmarks demonstrate that our model achieves comparable results with state-of-the-art methods, and predict Gaussian surfels within 1 second, offering a 100× speedup without costly per-scene training.



Paperid:5696
Authors:Yixuan Yang, Zhen Luo, Tongsheng Ding, Junru Lu, Mingqi Gao, Jinyu Yang, Victor Sanchez, Feng Zheng
Abstract:
Automatic indoor layout generation has attracted increasing attention due to its potential in interior design, virtual environment construction, and embodied AI. Existing methods fall into two categories: prompt-driven approaches that leverage proprietary LLM services (e.g., GPT APIs), and learning-based methods trained on layout data upon diffusion-based models. Prompt-driven methods often suffer from spatial inconsistency and high computational costs, while learning-based methods are typically constrained by coarse relational graphs and limited datasets, restricting their generalization to diverse room categories. In this paper, we revisit LLM-based indoor layout generation and present 3D-SynthPlace, a large-scale dataset that combines synthetic layouts generated via a `GPT synthesize, Human inspect' pipeline, upgraded from the 3D-Front dataset. 3D-SynthPlace contains nearly 17,000 scenes, covering four common room types—bedroom, living room, kitchen, and bathroom—enriched with diverse objects and high-level spatial annotations. We further introduce OptiScene, a strong open-source LLM optimized for indoor layout generation, fine-tuned based on our 3D-SynthPlace dataset through our two-stage training. For the warum-up stage I, we adopt supervised fine-tuning (SFT), which is taught to first generate high-level spatial descriptions then conditionally predict concrete object placements. For the reinforcing stage II, to better align the generated layouts with human design preferences, we apply multi-turn direct preference optimization (DPO), which significantly improving layout quality and generation success rates. Extensive experiments demonstrate that OptiScene outperforms traditional prompt-driven and learning-based baselines. Moreover, OptiScene shows promising potential in interactive tasks such as scene editing and robot navigation, highlighting its applicability beyond static layout generation.



Paperid:5697
Authors:Jiao Tang, WEI SHAO, Daoqiang Zhang
Abstract:
The whole-slide pathology images (WSIs) are widely recognized as the golden standard for cancer survival analysis. However, due to the high-resolution of WSIs, the existing studies require dividing WSIs into patches and identify key components before building the survival prediction system, which is time-consuming and cannot reflect the overall spatial organization of WSIs. Inspired by the fact that the spatial interactions among different tumor microenvironment (TME) components in WSIs are associated with the cancer prognosis, some studies attempt to capture the complex interactions among different TME components to improve survival predictions. However, they require extra efforts for building the TME segmentation model, which involves substantial annotation workloads on different TME components and is independent to the construction of the survival prediction model. To address the above issues, we propose ZTSurv, a novel end-to-end cancer survival analysis framework via efficient zero-shot TME segmentation on low-resolution WSIs. Specifically, by leveraging tumor infiltrating lymphocyte (TIL) maps on the 50x down-sampled WSIs, ZTSurv enables zero-shot segmentation on other two important TME components (i.e., tumor and stroma) that can reduce the annotation efforts from the pathologists. Then, based on the visual and semantic information extracted from different TME components, we construct a heterogeneous graph to capture their spatial intersections for clinical outcome prediction. We validate ZTSurv across four cancer cohorts derived from The Cancer Genome Atlas (TCGA), and the experimental results indicate that our method can not only achieve superior prediction results but also significantly reduce the computational costs in comparison with the state-of-the-art methods.



Paperid:5698
Authors:Zicheng Hu, Cheng Chen
Abstract:
Abstract:We investigate various stochastic bandit problems in the presence of adversarial corruption. A seminal work in this area is the BARBAR algorithm, which is both robust and efficient. However, it suffers from a regret of $O(KC)$, which does not match the lower bound $\Omega(C)$. In this paper, we first improve the BARBAR algorithm by proposing a novel framework called BARBAT, which eliminates the factor of $K$ to achieve an optimal regret bound up to a logarithmic factor. We also extend BARBAT to various settings, including multi-agent bandits, graph bandits, combinatorial semi-bandits and batched bandits. Compared to the Follow-The-Regularized-Leader (FTRL) framework, our methods offer the advantages of being parallelizable (making it suitable for multi-agent bandits and batched bandits) and having lower computational costs (especially in semi-bandits). Numerical experiments verifies the efficiency of proposed methods.



Paperid:5699
Authors:Weixing Wang, Zifeng Ding, Jindong Gu, RUI CAO, Christoph Meinel, Gerard de Melo, Haojin Yang
Abstract:
Large Vision-Language Models (LVLMs) with discrete image tokenizers unify multimodal representations by encoding visual inputs into a finite set of tokens. Despite their effectiveness, we find that these models still hallucinate non-existent objects. We hypothesize that one reason is due to visual priors induced during training: when certain image tokens frequently co-occur in the same spatial regions and represent shared objects, they become strongly associated with the verbalizations of those objects. As a result, the model may hallucinate by evoking visually absent tokens that often co-occur with present ones. To test this assumption, we construct a co-occurrence graph of image tokens using a segmentation dataset and employ a Graph Neural Network (GNN) with contrastive learning followed by a clustering method to group tokens that frequently co-occur in similar visual contexts. We find that hallucinations predominantly correspond to clusters whose tokens dominate the input, and more specifically, that the visually absent tokens in those clusters show much higher correlation with hallucinated objects compared to tokens present in the image. Based on this observation, we propose a hallucination mitigation method that suppresses the influence of visually absent tokens by modifying latent image embeddings during generation. Experiments show our method reduces hallucinations while preserving expressivity.



Paperid:5700
Authors:Guoji Fu, Wee Sun Lee
Abstract:
Abstract:This paper studies the approximation and generalization abilities of score-based neural network generative models (SGMs) in estimating an unknown distribution $P_0$ from $n$ i.i.d. observations in $d$ dimensions. Assuming merely that $P_0$ is $\alpha$-sub-Gaussian, we prove that for any time step $t \in [t_0, n^{\mathcal{O}(1)}]$, where $t_0 \geq \mathcal{O}(\alpha^2n^{-2/d}\log n)$, there exists a deep ReLU neural network with width $\leq \mathcal{O}(\log^3n)$ and depth $\leq \mathcal{O}(n^{3/d}\log_2n)$ that can approximate the scores with $\tilde{\mathcal{O}}(n^{-1})$ mean square error and achieve a nearly optimal rate of $\tilde{\mathcal{O}}(n^{-1}t_0^{-d/2})$ for score estimation, as measured by the score matching loss. Our framework is universal and can be used to establish convergence rates for SGMs under milder assumptions than previous work. For example, assuming further that the target density function $p_0$ lies in Sobolev or Besov classes, with an appropriately early stopping strategy, we demonstrate that neural network-based SGMs can attain nearly minimax convergence rates up to logarithmic factors. Our analysis removes several crucial assumptions, such as Lipschitz continuity of the score function or a strictly positive lower bound on the target density.



Paperid:5701
Authors:Matthias Kümmerer, Federico D'Agostino, Lisa Schwetlick, Matthias Bethge
Abstract:
Understanding how humans move their eyes to gather visual information is a central question in neuroscience, cognitive science, and vision research. While recent deep learning (DL) models achieve state-of-the-art performance in predicting human scanpaths, their underlying decision processes remain opaque. In contrast, cognitively inspired mechanistic models aim to explain scanpath behavior through interpretable cognitive mechanisms but lag far behind in predictive accuracy. In this work, we bridge this gap by using a high-performing deep model—DeepGaze III—to discover and test mechanisms that improve a leading mechanistic model, SceneWalk. By identifying individual fixations where DeepGaze III succeeds and SceneWalk fails, we isolate behaviorally meaningful discrepancies and use them to motivate targeted extensions of the mechanistic framework. These include time-dependent temperature scaling, saccadic momentum and an adaptive cardinal attention bias: Simple, interpretable additions that substantially boost predictive performance. With these extensions, SceneWalk’s explained variance doubles from 35% to 70%, setting a new state of the art in mechanistic scanpath prediction. Our findings show how performance-optimized neural networks can serve as tools for cognitive model discovery, offering a new path toward interpretable and high-performing models of visual behavior.



Paperid:5702
Authors:Wenbin An, Jiahao Nie, Feng Tian, Haonan Lin, mingxiang cai, Yaqiang Wu, QianYing Wang, Xiaoqin Zhang, Shijian Lu
Abstract:
Despite their recent progress, Multimodal Large Language Models (MLLMs) often struggle in knowledge-intensive tasks due to the limited and outdated parametric knowledge they acquired during training. Multimodal Retrieval Augmented Generation addresses this issue by retrieving contextual knowledge from external databases, thereby enhancing LVLMs with expanded knowledge sources. However, existing LVLMs often fail to fully leverage the retrieved contextual knowledge. We examine representative LVLMs and identify two major causes, namely, attention bias toward different tokens and knowledge conflicts between parametric and contextual knowledge. To this end, we design Adaptive Logits Fusion and Attention Reallocation (ALFAR), a training-free and plug-and-play approach that improves LVLM responses by maximizing the utility of the retrieved knowledge. Specifically, ALFAR tackles the challenges from two perspectives. First, it alleviates attention bias by adaptively shifting attention from visual tokens to relevant context tokens according to query-context relevance. Second, it decouples and weights parametric and contextual knowledge at output logits, mitigating conflicts between the two types of knowledge. As a plug-and-play method, ALFAR achieves superior performance across diverse datasets without requiring additional training or external tools. Extensive experiments over multiple LVLMs and benchmarks show that ALFAR consistently outperforms the state-of-the-art by large margins. Our code and data will be made publicly available.



Paperid:5703
Authors:Fujun Han, Peng Ye
Abstract:
Vision-based intrusion detection has multiple applications in practical scenarios, e.g., autonomous driving, intelligent monitoring, and security. Previous works mainly focus on improving the intrusion detection performance, without a comprehensive and in-depth understanding of the intrusion scene. To fill this gap, we explore a novel task called Multimodal Large Language Models based Intrusion Scene Understanding (MLLM-ISU) and report a comprehensive benchmark for the task. Specifically, we first design an effective and automatic visual question-answer generation strategy, constructing a new MLLM-ISU dataset, with 3000 VQA evaluation Pairs, 8925 training Pairs, and six relevant subtasks. Then, we perform a comprehensive assessment on various state-of-the-art proprietary and open-source MLLMs, e.g., DeepSeek-VL2, GPT-4o, Qwen2.5-VL, etc, and find that current MLLMs have weak abilities for this task. Further, in order to improve the intrusion understanding capabilities of current MLLMs, we propose a Post-Training Framework with three sequential training stages, i.e., Intrusion-aware Visual Instruction Pre-training, Intrusion Chain of Thought tuning, and Intrusion-centric VQA tuning, and sufficient experiments and comparisons are conducted to verify the effectiveness of the proposed three-stages training framework. All datasets, codes, and baselines will be publicly available.



Paperid:5704
Authors:Alistair Benford, Per Kristian Lehre
Abstract:
Abstract:Competitive coevolutionary algorithms (CoEAs) have a natural application to problems that are adversarial or feature strategic interaction. However, there is currently limited theoretical insight into how to avoid pathological behaviour associated to CoEAs. In this paper we use impartial combinatorial games as a challenging domain for CoEAs and provide corresponding runtime analysis. By analysing how individuals capitalise on the mistakes of their opponents, we prove that the Univariate Marginal Distribution Algorithm finds (with high probability) an optimal strategy for a game called Reciprocal LeadingOnes within $O(n^2\log^3{n})$ game evaluations, a significant improvement over the best known bound of $O(n^5\log^2{n})$. Critical to the analysis is the introduction of a novel stabilising operator, the impact of which we study both theoretically and empirically.



Paperid:5705
Authors:Jiashuo Sun, Xianrui Zhong, Sizhe Zhou, Jiawei Han
Abstract:
Retrieval-augmented generation (RAG) systems combine large language models (LLMs) with external knowledge retrieval, making them highly effective for knowledge-intensive tasks. A crucial but often under-explored component of these systems is the reranker, which refines retrieved documents to enhance generation quality and explainability. The challenge of selecting the optimal number of documents (k) remains unsolved: too few may omit critical information, while too many introduce noise and inefficiencies. Although recent studies have explored LLM-based rerankers, they primarily leverage internal model knowledge and overlook the rich supervisory signals that LLMs can provide, such as using response quality as feedback for optimizing reranking decisions. In this paper, we propose DynamicRAG, a novel RAG framework where the reranker dynamically adjusts both the order and number of retrieved documents based on the query. We model the reranker as an agent optimized through reinforcement learning (RL), using rewards derived from LLM output quality. Across seven knowledge-intensive datasets, DynamicRAG demonstrates superior performance, achieving state-of-the-art results.



Paperid:5706
Authors:Yi Liu, Yang Liu, Leqian Zheng, Jue Hong, Junjie Shi, Qingyou Yang, Ye Wu, Cong Wang
Abstract:
Abstract:With the rapid advancement of the digital economy, data collaboration between organizations has become a well-established business model, driving the growth of various industries. However, privacy concerns make direct data sharing impractical. To address this, Two-Party Split Learning (a.k.a. Vertical Federated Learning (VFL)) has emerged as a promising solution for secure collaborative learning. Despite its advantages, this architecture still suffers from low computational resource utilization and training efficiency. Specifically, its synchronous dependency design increases training latency, while resource and data heterogeneity among participants further hinder efficient computation. To overcome these challenges, we propose \texttt{PubSub-VFL}, a novel VFL paradigm with a Publisher/Subscriber architecture optimized for two-party collaborative learning with high computational efficiency. \texttt{PubSub-VFL} leverages the decoupling capabilities of the Pub/Sub architecture and the data parallelism of the parameter server architecture to design a hierarchical asynchronous mechanism, reducing training latency and improving system efficiency. Additionally, to mitigate the training imbalance caused by resource and data heterogeneity, we formalize an optimization problem based on participants’ system profiles, enabling the selection of optimal hyperparameters while preserving privacy. We conduct a theoretical analysis to demonstrate that \texttt{PubSub-VFL} achieves stable convergence and is compatible with security protocols such as differential privacy. Extensive case studies on five benchmark datasets further validate its effectiveness, showing that \texttt{PubSub-VFL} compared to state-of-the-art baselines not only accelerates training by $2 \sim 7\times$ without compromising accuracy but also achieves computational resource utilization by up to 91.07\%.



Paperid:5707
Authors:Ke Sun, Yingnan Zhao, Enze Shi, Yafei Wang, Xiaodong Yan, Bei Jiang, Linglong Kong
Abstract:
The remarkable empirical performance of distributional reinforcement learning~(RL) has garnered increasing attention to understanding its theoretical advantages over classical RL. By decomposing the categorical distributional loss commonly employed in distributional RL, we find that the potential superiority of distributional RL can be attributed to a derived distribution-matching entropy regularization. This less-studied entropy regularization aims to capture additional knowledge of return distribution beyond only its expectation, contributing to an augmented reward signal in policy optimization. In contrast to the vanilla entropy regularization in MaxEnt RL, which explicitly encourages exploration by promoting diverse actions, the novel entropy regularization derived from categorical distributional loss implicitly updates policies to align the learned policy with (estimated) environmental uncertainty. Finally, extensive experiments verify the significance of this uncertainty-aware regularization from distributional RL on the empirical benefits over classical RL. Our study offers an innovative exploration perspective to explain the intrinsic benefits of distributional learning in RL.



Paperid:5708
Authors:Nayel Bettache
Abstract:
Abstract:This paper studies a bilinear matrix-valued regression model where both predictors and responses are matrix-valued. For each observation \( t = 1, \ldots, T \), the response \( Y_t \in \mathbb{R}^{n \times p} \) and predictor \( X_t \in \mathbb{R}^{m \times q} \) satisfy the model $Y_t = A^* X_t B^* + E_t,$with \( A^* \in \mathbb{R}_+^{n \times m} \) (row-wise \(\ell_1\)-normalized), \( B^* \in \mathbb{R}^{q \times p} \), and \( E_t \) independent Gaussian noise matrices. The goal is to estimate \( A^* \) and \( B^* \) from the observed pairs \( (X_t, Y_t) \).We propose explicit, optimization-free estimators and establish non-asymptotic error bounds, including sparse settings. Simulations confirm the theoretical rates and demonstrate strong finite-sample performance. We further illustrate the practical utility of our method through an image denoising application on real data.



Paperid:5709
Authors:Di Wu, Liu Liu, Zhou Linli, Anran Huang, Liangtu Song, Qiaojun Yu, Qi Wu, Cewu Lu
Abstract:
Articulated objects, as prevalent entities in human life, their 3D representations play crucial roles across various applications. However, achieving both high-fidelity textured surface reconstruction and dynamic generation for articulated objects remains challenging for existing methods. In this paper, we present REArtGS, a novel framework that introduces additional geometric and motion constraints to 3D Gaussian primitives, enabling realistic surface reconstruction and generation for articulated objects. Specifically, given multi-view RGB images of arbitrary two states of articulated objects, we first introduce an unbiased Signed Distance Field (SDF) guidance to regularize Gaussian opacity fields, enhancing geometry constraints and improving surface reconstruction quality. Then we establish deformable fields for 3D Gaussians constrained by the kinematic structures of articulated objects, achieving unsupervised generation of surface meshes in unseen states. Extensive experiments on both synthetic and real datasets demonstrate our approach achieves high-quality textured surface reconstruction for given states, and enables high-fidelity surface generation for unseen states. Codes can be found in the supplementary materials and will be made publicly available.



Paperid:5710
Authors:Bo Peng, Jie Lu, Guangquan Zhang, Zhen Fang
Abstract:
Out-of-distribution (OOD) detection, recognized for its ability to identify samples of unknown classes, provides solid advantages in ensuring the reliability of machine learning models. Among existing OOD detection methods, pre-trained vision-language models have emerged as powerful post-hoc OOD detectors by leveraging textual and visual information. Despite the empirical success, there still remains a lack of research on a formal understanding of their effectiveness. This paper bridges the gap by theoretically demonstrating that existing CLIP-based post-hoc methods effectively perform a stochastic estimation of the point-wise mutual information (PMI) between the input image and each in-distribution label. This estimation is then utilized to construct energy functions for modeling in-distribution distributions.Different from prior methods that inherently consider PMI estimation as a whole task, we, motivated by the divide-and-conquer philosophy, decompose PMI estimation into multiple easier sub-tasks by applying the chain rule of PMI, which not only reduces the estimation complexity but also provably increases the estimation upper bound to reduce the underestimation bias. Extensive evaluations across mainstream benchmarks empirically manifest that our method establishes a new state-of-the-art in a variety of OOD detection setups.



Paperid:5711
Authors:Pei-Shuo Wang, Jian-Jia Chen, Chun-Che Yang, Chi-Chih Chang, Ning-Chi Huang, Mohamed Abdelfattah, Kai-Chiang Wu
Abstract:
Abstract:The immense model sizes of large language models (LLMs) challenge deployment on memory-limited consumer GPUs. Model compression and parameter offloading are common strategies to address the memory limitation. While compression may degrade quality, offloading maintains quality but suffers from slow inference. Speculative decoding presents a promising avenue to accelerate parameter offloading, utilizing a fast draft model to propose multiple draft tokens, which are then verified by the target LLM in parallel with a single forward pass. This reduces the time-consuming data transfers in forward passes involving offloaded weight transfers. Existing methods often rely on pretrained weights of the same family, but actually require additional training to align with custom-trained models in reality. Some other approaches involving draft model training often yield only modest speedups. This limitation arises from insufficient alignment with the target model, which prevents achieving higher token acceptance lengths. To address these challenges and realize greater speedups, we propose SubSpec, a lossless and training-free approach to accelerate parameter offloading. SubSpec constructs a highly aligned draft model by generating low-bit quantized substitute layers from offloaded target LLM portions. Additionally, our method shares remaining GPU-resident layers and the KV-Cache. These sharing strategies further reduce memory overhead and enhance alignment. SubSpec achieves a high average acceptance length, delivering 9.1$\times$ speedup for Qwen2.5 7B on MT-Bench (8GB VRAM limit) and an average of 12.5$\times$ speedup for Qwen2.5 32B on popular generation benchmarks (24GB VRAM limit). The code will be made publicly available.



Paperid:5712
Authors:Nibir Mandal, Oishee Hoque, Mandy Wilson, Samarth Swarup, Sayjro Nouwakpo, Abhijin Adiga, Madhav Marathe
Abstract:
The lack of fine-grained, large-scale datasets on water availability presents a critical barrier to applying machine learning (ML) for agricultural water management. Since there are multiple natural and anthropogenic factors that influence water availability, incorporating diverse multimodal features can significantly improve modeling performance. However, integrating such heterogeneous data is challenging due to spatial misalignments, inconsistent formats, semantic label ambiguities, and class imbalances. To address these challenges, we introduce IRRISIGHT, a large-scale, multimodal dataset spanning 20 U.S. states. It consists of 1.4 million pixel-aligned 224×224 patches that fuse satellite imagery with rich environmental attributes. We develop a robust geospatial fusion pipeline that aligns raster, vector, and point-based data on a unified 10m grid, and employ domain-informed structured prompts to convert tabular attributes into natural language. With irrigation type classification as a representative problem, the dataset is AI-ready, offering a spatially disjoint train/test split and extensive benchmarking with both vision and vision–language models. Our results demonstrate that multimodal representations substantially improve model performance, establishing a foundation for future research on water availability.https://github.com/Nibir088/IRRISIGHThttps://huggingface.co/datasets/OBH30/IRRISIGHT



Paperid:5713
Authors:Tal Barami, Nimrod Berman, Ilan Naiman, Amos Hason, Rotem Ezra, Omri Azencot
Abstract:
Learning disentangled representations in sequential data is a key goal in deep learning, with broad applications in vision, audio, and time series. While real-world data involves multiple interacting semantic factors over time, prior work has mostly focused on simpler two-factor (static-dynamic) settings, primarily due to simpler data collection requirements—thus overlooking the inherently multi-factor nature of real-world data. We introduce the first standardized benchmark for evaluating multi-factor sequential disentanglement across six diverse datasets spanning video, audio, and time series. Our benchmark includes modular tools for dataset integration, model development, and evaluation metrics tailored to multi-factor analysis. We additionally propose a post-hoc Latent Exploration Stage to automatically align latent dimensions with semantic factors, and introduce a Koopman-inspired model that achieves state-of-the-art results. Moreover, we show that Vision-Language Models can automate dataset annotation and serve as zero-shot disentanglement evaluators—removing the need for manual labels and human intervention. Together, these contributions provide a robust and scalable foundation for advancing multi-factor sequential disentanglement.



Paperid:5714
Authors:Sunny Duan, Ling Dong, Ila Fiete
Abstract:
How precisely does circuit wiring specify function? This fundamental question is particularly relevant for modern neuroscience, as large-scale electron microscopy now enables the reconstruction of neural circuits at single-synapse resolution across many organisms. To interpret circuit function from such datasets, we must understand the extent to which [measured] structure constrains dynamics. We investigate this question in the drosophila head direction (HD) circuit, which maintains an internal heading estimate through attractor dynamics that integrate self-motion velocity cues. This circuit serves as a sensitive assay for functional specification: continuous attractor networks are theoretically known to require finely tuned wiring, whereas connectomes reveal that biological wiring can be variable and omit key cellular parameters such as synaptic gains, neuronal thresholds, and time constants. We introduce a method that combines self-supervised and unsupervised learning objectives to estimate unknown parameters at the level of cell types, rather than individual neurons and synapses. Given the raw connectivity matrix, our approach recovers a network that robustly exhibits continuous attractor dynamics and accurately integrates a range of velocity inputs, despite minimal parameter tuning on a connectome which notably departs from the symmetric regularity of an idealized ring attractor. We characterize how deviations from the original connectome shape the space of viable solutions. We also perform in-silico ablation experiments to probe the distinct functional roles of specific cell types in the circuit, demonstrating how connectome-derived structure, when augmented with minimal, biologically grounded tuning, can replicate known physiology and elucidate circuit function.



Paperid:5715
Authors:Boyuan Cao, Jiaxin Ye, Yujie Wei, Hongming Shan
Abstract:
While latent diffusion models (LDMs), such as Stable Diffusion, are designed for high-resolution image generation, they often struggle with significant structural distortions when generating images at resolutions higher than their training one.Instead of relying on extensive retraining, a more resource-efficient approach is to reprogram the pretrained model for high-resolution (HR) image generation; however, existing methods often result in poor image quality and long inference time.We introduce RepLDM, a novel reprogramming framework for pretrained LDMs that enables high-quality, high-efficiency, high-resolution image generation; see Fig. 1. RepLDM consists of two stages: (i) an attention guidance stage, which generates a latent representation of a higher-quality training-resolution image using a novel parameter-free self-attention mechanism to enhance the structural consistency; and (ii) a progressive upsampling stage, which progressively performs upsampling in pixel space to mitigate the severe artifacts caused by latent space upsampling. The effective initialization from the first stage allows for denoising at higher resolutions with significantly fewer steps, improving the efficiency.Extensive experimental results demonstrate that RepLDM significantly outperforms state-of-the-art methods in both quality and efficiency for HR image generation, underscoring its advantages for real-world applications.



Paperid:5716
Authors:Santiago Galella, Maren Wehrheim, Matthias Kaschube
Abstract:
Biological and artificial vision systems both rely on hierarchical architectures, yet it remains unclear how their representational geometry evolves across processing stages, and what functional consequences may arise from potential differences. In this work, we systematically quantify and compare the linear and non-linear (intrinsic) dimensionality of human brain activity (fMRI) and artificial neural networks (ANNs) during natural image viewing. In the human ventral visual stream, both dimensionality measures increase super-linearly along the visual hierarchy, supporting the emergence of semantic and abstract representations. For linear dimensionality, ANNs show a similar increase, but only for pooled features, reinforcing the importance of appropriate feature readouts in brain–model comparisons. In contrast, nonlinear dimensionality shows a collapse in the later layers of ANNs, pointing at a mismatch in representational geometry between the human and artificial visual systems. Moreover, we show that this mismatch may have important functional consequences: high-dimensional brain representations support flexible generalization to abstract, behaviorally rated features, while ANNs lose access to such information in deeper layers despite achieving high categorization accuracy. We show that in the brain, dimensionality increases along the visual hierarchy and supports the encoding of abstract features. In contrast, ANNs exhibit late-stage compression, reflected by the collapse of nonlinear dimensionality, which appears to limit their capacity for abstraction. Our findings suggest that representational geometry provides a functional signature of the ability of abstraction, and propose dimensionality alignment as a benchmark for building more flexible and biologically grounded vision models.



Paperid:5717
Authors:Antonio Guillen-Perez, Avisek Naug, Vineet Gundecha, Sahand Ghorbanpour, Ricardo Luna Gutierrez, Ashwin Ramesh Babu, Munther Salim, Shubhanker Banerjee, Eoin Essink, Damien Fay, Soumyendu Sarkar
Abstract:
The escalating energy demands and carbon footprint of large-scale AI necessitate intelligent workload management across globally distributed data centers. However, developing and evaluating sustainable scheduling strategies is hindered by the lack of benchmarks that realistically capture the complex interplay of time-varying environmental factors (grid carbon intensity, electricity price, weather variations), detailed data center physics (including CPUs, GPUs, memory, and crucial HVAC energy consumption), and geo-distributed network dynamics (costs and latency). To address this critical gap, we introduce SustainCluster: an open-source, high-fidelity simulation benchmark environment designed for research in sustainable, geo-temporal task scheduling. SustainCluster integrates curated, real-world datasets covering AI workload traces, grid carbon intensity, electricity price, weather data for over 20 global regions, cloud transmission costs, and empirical network delay parameters with physics-informed models of data center operations. It presents a challenging scheduling problem where a top-level coordinating agent must reassign or defer dynamically arriving tasks with resource and SLA requirements across a configurable cluster of data centers to optimize multiple objectives. The environment also supports additional components like heat recovery. SustainCluster features a modular reward system enabling explicit study of trade-offs such as carbon emissions, energy cost, SLA, and water usage, and provides a Gymnasium API with baseline controllers such as Reinforcement Learning and rule-based examples, facilitating reproducible machine learning research and fair comparison of diverse scheduling algorithms. By offering a realistic, configurable, and accessible testbed, SustainCluster aims to accelerate the development and validation of next-generation sustainable computing solutions for distributed data centers.



Paperid:5718
Authors:Yuanyao Chen, Rongsheng Chen, Fu Luo, Zhenkun Wang
Abstract:
Neural Combinatorial Optimization (NCO) has emerged as a promising learning-based paradigm for addressing Vehicle Routing Problems (VRPs) by minimizing the need for extensive manual engineering. While existing NCO methods, trained on small-scale instances (e.g., 100 nodes), have demonstrated considerable success on problems of similar scale, their performance significantly degrades when applied to large-scale scenarios. This degradation arises from the distributional shift between training and testing data, rendering policies learned on small instances ineffective for larger problems. To overcome this limitation, we introduce a novel learning framework driven by Large Language Models (LLMs). This framework learns a projection between the training and testing distributions, which is then deployed to enhance the scalability of the NCO model. Notably, unlike prevailing techniques that necessitate joint training with the neural network, our approach operates exclusively during the inference phase, obviating the need for model retraining. Extensive experiments demonstrate that our method enables a backbone model (trained on 100-node instances) to achieve superior performance on large-scale Traveling Salesman Problem (TSP) and Capacitated Vehicle Routing Problem (CVRP) of up to 100K nodes from diverse distributions.



Paperid:5719
Authors:Siyuan Wu, Yongzhe Jia, Haolong Xiang, Xiaolong Xu, Xuyun Zhang, Lianyong Qi, Wanchun Dou
Abstract:
Federated learning (FL) is a distributed machine learning paradigm that enables multiple clients to collaboratively train a shared model without exposing their raw data. However, existing FL research has primarily focused on optimizing learning performance based on the assumption of uniform client participation, with few studies delving into performance fairness under inconsistent client participation, particularly in model-heterogeneous FL environments. In view of this challenge, we propose PHP-FL, a novel model-heterogeneous FL method that explicitly addresses scenarios with varying client participation probabilities to enhance both model accuracy and performance fairness. Specifically, we introduce a Dual-End Aligned ensemble Learning (DEAL) module, where small auxiliary models on clients are used for dual-end knowledge alignment and local ensemble learning, effectively tackling model heterogeneity without a public dataset. Furthermore, to mitigate update conflicts caused by inconsistent participation probabilities, we propose an Importance-driven Selective Parameter Update (ISPU) module, which accurately updates critical local parameters based on training progress. Finally, we implement PHP-FL on a lightweight FL platform with heterogeneous clients across three different client participation patterns. Extensive experiments under heterogeneous settings and diverse client participation patterns demonstrate that PHP-FL achieves state-of-the-art performance in both accuracy and fairness. Our code is available at: https://anonymous.4open.science/r/PHP-FL-8E3D.



Paperid:5720
Authors:Shixiang Liu
Abstract:
The sparse additive model (SpAM) offers a trade-off between interpretability and flexibility, and hence is a powerful model for high-dimensional research.This paper focuses on the variable selection, i.e., the univariate function selection problem in SpAM.We establish the minimax separation rates from both the perspectives of sparse multiple testing (FDR + FNR control) and support recovery (wrong recovery probability control).We further study how adaptation to unknown smoothness affects the minimax separation rate, and propose an adaptive selection procedure.Finally, we discuss the difference between estimation and selection in SpAM: Procedures achieving minimax optimal estimation may fail to achieve minimax optimal univariate function selection.



Paperid:5721
Authors:Yunhai Hu, Tianhua Xia, Zining Liu, Rahul Raman, Xingyu Liu, BO BAO, Eric Sather, Vithursan Thangarasa, Sai Qian Zhang
Abstract:
Speculative decoding (SD) has emerged as a powerful method for accelerating autoregressive generation in large language models (LLMs), yet its integration into vision-language models (VLMs) remains underexplored. We introduce DREAM, a novel speculative decoding framework tailored for VLMs that combines three key innovations: (1) a cross-attention-based mechanism to inject intermediate features from the target model into the draft model for improved alignment, (2) adaptive intermediate feature selection based on attention entropy to guide efficient draft model training, and (3) visual token compression to reduce draft model latency. DREAM enables efficient, accurate, and parallel multimodal decoding with significant throughput improvement. Experiments across a diverse set of recent popular VLMs, including LLaVA, Pixtral, SmolVLM and Gemma3, demonstrate up to 3.6x speedup over conventional decoding and significantly outperform prior SD baselines in both inference throughput and speculative draft acceptance length across a broad range of multimodal benchmarks.



Paperid:5722
Authors:Yifeng Yu, Lu Yu
Abstract:
Abstract:Score-based diffusion models have emerged as powerful tools in generative modeling, yet their theoretical foundations remain underexplored. In this work, we focus on the Wasserstein convergence analysis of score-based diffusion models. Specifically, we investigate the impact of various discretization schemes, including Euler discretization, exponential integrators, and midpoint randomization methods. Our analysis provides the first quantitative comparison of these discrete approximations, emphasizing their influence on convergence behavior. Furthermore, we explore scenarios where Hessian information is available and propose an accelerated sampler based on the local linearization method. We establish the first Wasserstein convergence analysis for such a Hessian-based method, showing that it achieves an improved convergence rate of order $\widetilde{\mathcal{O}}\left(\frac{1}{\varepsilon}\right)$, which significantly outperforms the standard rate $\widetilde{\mathcal{O}}\left(\frac{1}{\varepsilon^2}\right)$ of vanilla diffusion models. Numerical experiments on synthetic data and the MNIST dataset validate our theoretical insights.



Paperid:5723
Authors:Seungyeon Choi, Hwanhee Kim, Chihyun Park, Dahyeon Lee, Seungyong Lee, Yoonju Kim, Hyoungjoon Park, Sein Kwon, Youngwan Jo, Sanghyun Park
Abstract:
Abstract:Recent advances in Structure-based Drug Design (SBDD) have leveraged generative models for 3D molecular generation, predominantly evaluating model performance by binding affinity to target proteins. However, practical drug discovery necessitates high binding affinity along with synthetic feasibility and selectivity, critical properties that were largely neglected in previous evaluations. To address this gap, we identify fundamental limitations of conventional diffusion-based generative models in effectively guiding molecule generation toward these diverse pharmacological properties. We propose $\texttt{CByG}$, a novel framework extending Bayesian Flow Network into a gradient-based conditional generative model that robustly integrates property-specific guidance. Additionally, we introduce a comprehensive evaluation scheme incorporating practical benchmarks for binding affinity, synthetic feasibility, and selectivity, overcoming the limitations of conventional evaluation methods. Extensive experiments demonstrate that our proposed $\texttt{CByG}$, framework significantly outperforms baseline models across multiple essential evaluation criteria, highlighting its effectiveness and practicality for real-world drug discovery applications.



Paperid:5724
Authors:Ruskin Raj Manku, Yuzhi Tang, Xingjian Shi, Mu Li, Alexander Smola
Abstract:
Abstract:Text-to-Speech (TTS) benchmarks often fail to capture how well models handle nuanced and semantically complex text. Building on $\textit{EmergentTTS}$, we introduce $\textit{EmergentTTS-Eval}$, a comprehensive benchmark covering six challenging TTS scenarios: emotions, paralinguistics, foreign words, syntactic complexity, complex pronunciation (e.g. URLs, formulas), and questions. Crucially, our framework automates both test-generation and evaluation, making the benchmark easily extensible. Starting from a small set of human-written seed prompts, we iteratively extend them using LLMs to target specific structural, phonetic and prosodic challenges, resulting in 1,645 diverse test samples. Moreover, we employ a model-as-a-judge approach, using a Large Audio Language Model (LALM) to assess the speech across multiple dimensions such as expressed emotion, prosodic, intonational, and pronunciation accuracy. We evaluate state-of-the-art open-source and proprietary TTS systems, such as 11Labs, Deepgram, and OpenAI's 4o-mini-TTS, on EmergentTTS-Eval, demonstrating its ability to reveal fine-grained performance differences. Results show that the model-as-a-judge approach offers robust TTS assessment and a high correlation with human preferences. We open-source the code and the benchmark.



Paperid:5725
Authors:Zhiyuan Zhou, Yueming Yin, Yiming Yang, Yuguang Mu, Hoi-Yeung Li, Adams Wai Kin Kong
Abstract:
Strong ligands can prevent endogenous substrates from binding to their target pockets, inhibiting or activating target proteins, making them promising drug candidates. The ligand's binding affinity determines the competitive binding outcome; stronger binding results in more effective drugs. Some ligands bind strongly to specific target pockets; however, occupancy and the corresponding target regions vary based on sizes and interactions. This variability may result in unoccupied pocket space, allowing endogenous substrates to compete with the ligand, which diminishes the binding interaction and reduces the ligand’s effectiveness as a drug candidate. Existing AI drug discovery methods highly concentrate on local information in the ligand or protein-ligand complex but neglect the unoccupied space, the spatial emptiness. To enhance bioactivity prediction, in this paper, we propose Geometric Representation of Spatial Emptiness in Protein-ligand Complex (GeoREC), which measures the empty space around each atom in the protein-ligand complex and Union of Multiple Pockets (Union-Pockets), which is composed of multiple pockets of the protein. They together provide a consistent and global view of the binding sites to drug discovery network. Furthermore, we note that the mean-square error (MSE) loss commonly employed in existing methods is not an ideal loss function for drug discovery because it considers each prediction independently. To alleviate this problem, we exploit a pairwise loss function. Extensive experiments on multiple datasets with diverse bioactivity types demonstrate that integrating the proposed representations and the loss function into state-of-the-art models leads to significant performance gains, highlighting their effectiveness.



Paperid:5726
Authors:Ke Ji, Jiahao Xu, Tian Liang, Qiuzhi Liu, Zhiwei He, Xiaoyuan Liu, Xingyu Chen, Junying Chen, Benyou Wang, Zhaopeng Tu, Haitao Mi, Dong Yu
Abstract:
Improving the reasoning capabilities of large language models (LLMs) typically requires supervised fine-tuning with labeled data or computationally expensive sampling. We introduce Unsupervised Prefix Fine-Tuning (UPFT), which leverages the observation of Prefix Self-Consistency -- the shared initial reasoning steps across diverse solution trajectories -- to enhance LLM reasoning efficiency. By training exclusively on the initial prefix substrings (as few as 8 tokens), UPFT removes the need for labeled data or exhaustive sampling. Experiments on reasoning benchmarks show that UPFT matches the performance of supervised methods such as Rejection Sampling Fine-Tuning, while reducing training time by 75\% and sampling cost by 99\%. Further analysis reveals that errors tend to appear in later stages of the reasoning process and that prefix-based training preserves the model’s structural knowledge. This work demonstrates how minimal unsupervised fine-tuning can unlock substantial reasoning gains in LLMs, offering a scalable and resource-efficient alternative to conventional approaches.



Paperid:5727
Authors:Zewei Zhou, Tianhui Cai, Seth Zhao, Yun Zhang, Zhiyu Huang, Bolei Zhou, Jiaqi Ma
Abstract:
Recent advancements in Vision-Language-Action (VLA) models have shown promise for end-to-end autonomous driving by leveraging world knowledge and reasoning capabilities. However, current VLA models often struggle with physically infeasible action outputs, complex model structures, and unnecessarily long reasoning. In this paper, we propose AutoVLA, a novel VLA framework that unifies reasoning and action generation within a single autoregressive generation model. AutoVLA performs semantic reasoning and trajectory planning directly from raw visual inputs and language instructions. We tokenize continuous trajectories into discrete, feasible actions, enabling direct integration into the language model. For training, we employ supervised fine-tuning to equip the model with dual thinking modes: fast thinking (trajectory-only) and slow thinking (enhanced with chain-of-thought reasoning). To further enhance planning performance and efficiency, we introduce a reinforcement fine-tuning method based on Group Relative Policy Optimization (GRPO), reducing unnecessary reasoning in straightforward scenarios. Extensive experiments across real-world and simulated datasets and benchmarks, including nuPlan, nuScenes, Waymo, and CARLA, demonstrate the competitive performance of AutoVLA in both open-loop and closed-loop settings. Qualitative results further showcase the adaptive reasoning and accurate planning capabilities of AutoVLA in diverse scenarios. We will release the code, model weights, and datasets to facilitate future research in the field.



Paperid:5728
Authors:Yutong Wang, Haiyu Wang, Sai Qian Zhang
Abstract:
Abstract:Vision-Language Models (VLMs) are integral to tasks such as image captioning and visual question answering, but their high computational cost, driven by large memory footprints and processing time, limits their scalability and real-time applicability. In this work, we propose leveraging Singular-Value Decomposition (SVD) over the joint query (Q), key (K), and value (V) weight matrices to reduce KV cache size and computational overhead. We in addition introduce a efficient rank allocation strategy that dynamically adjusts the SVD rank based on its impact on VLM accuracy, achieving a significant reduction in both memory usage and computational cost. Finally, we extend this approach by applying quantization to both VLM weights and activations, resulting in a highly efficient VLM. Our method outperforms previous approaches that rely solely on quantization or SVD by achieving more than $10\%$ accuracy improvement while consuming less hardware cost, making it better for real-time deployment on resource-constrained devices.



Paperid:5729
Authors:Jiahao Yu, Haozhuang Liu, Yeqiu Yang, Lu Chen, Jian Wu, Yuning Jiang, Bo Zheng
Abstract:
Regression models are crucial in recommender systems. However, retransformation bias problem has been conspicuously neglected within the community. While many works in other fields have devised effective bias correction methods, all of them are post-hoc cures externally to the model, facing practical challenges when applied to real-world recommender systems. Hence, we propose a preemptive paradigm to eradicate the bias intrinsically from the models via minor model refinement. Specifically, a novel TranSUN method is proposed with a joint bias learning manner to offer theoretically guaranteed unbiasedness under empirical superior convergence. It is further generalized into a novel generic regression model family, termed Generalized TranSUN (GTS), which not only offers more theoretical insights but also serves as a generic framework for flexibly developing various bias-free models. Comprehensive experimental results demonstrate the superiority of our methods across data from various domains, which have been successfully deployed in two core recommendation scenarios on an e-commerce platform to serve the major online traffic. Codes will be released after this paper is accepted.



Paperid:5730
Authors:Lorenzo Baldassari, Josselin Garnier, Knut Solna, Maarten V. de Hoop
Abstract:
Designing algorithms for solving high-dimensional Bayesian inverse problems directly in infinite‑dimensional function spaces – where such problems are naturally formulated – is crucial to ensure stability and convergence as the discretization of the underlying problem is refined. In this paper, we contribute to this line of work by analyzing a widely used sampler for linear inverse problems: Langevin dynamics driven by score‑based generative models (SGMs) acting as priors, formulated directly in function space. Building on the theoretical framework for SGMs in Hilbert spaces, we give a rigorous definition of this sampler in the infinite-dimensional setting and derive, for the first time, error estimates that explicitly depend on the approximation error of the score. As a consequence, we obtain sufficient conditions for global convergence in Kullback–Leibler divergence on the underlying function space. Preventing numerical instabilities requires preconditioning of the Langevin algorithm and we prove the existence and form of an optimal preconditioner. The preconditioner depends on both the score error and the forward operator and guarantees a uniform convergence rate across all posterior modes. Our analysis applies to both Gaussian and a general class of non‑Gaussian priors. Finally, we present examples that illustrate and validate our theoretical findings.



Paperid:5731
Authors:Jianqi Feng, Wei Zhao, Zhenke Wu, Chengchun Shi, Xiaodong Yan
Abstract:
Constructing multistage optimal decisions for alternating recurrent event data is critically important in medical and healthcare research. Current reinforcement learning (RL) methods have only been applied to time-to-event data, and maximize the expectation. However, recurrent event data exhibit a distinct structure and emphasize the probability of event occurrences. In this paper, we incorporate recurrent event data and for the first time propose a RL objective focused on maximizing probabilities. To apply recurrent event data within the RL framework, we formulate a Decision Process optimization framework. During the optimization, we address the challenge of heterogeneous stage counts across individuals by reformulating an auxiliary problem. The proposed optimal policy can be efficiently implemented using Bellman optimality operators. Additionally, we establish the equivalence properties of the optimal policy under the new objective and the unbiasedness of the estimated Q-function. Experiments show that proposed method can converge faster and reduce the variance, and achieve a larger probability compared with the traditional objective.



Paperid:5732
Authors:Batuhan Altundas, Shengkang Chen, Shivika Singh, Shivangi Deo, Minwoo Cho, Matthew Gombolay
Abstract:
Abstract:Coordinating large teams of heterogeneous mobile agents to perform complex tasks efficiently has scalability bottlenecks in feasible and optimal task scheduling, with critical applications in logistics, manufacturing, and disaster response. Existing task allocation and scheduling methods, including heuristics and optimization-based solvers, often fail to scale and overlook inter-task dependencies and agent heterogeneity. We propose a novel Simultaneous Decision-Making model for Heterogeneous Multi-Agent Task Allocation and Scheduling (HM-MATAS), built on a Residual Heterogeneous Graph Transformer with edge and node-level attention. Our model encodes agent capabilities, travel times, and temporospatial constraints into a rich graph representation and is trainable via reinforcement learning. Trained on small-scale problems (10 agents, 20 tasks), our model generalizes effectively to significantly larger scenarios (up to 40 agents and 200 tasks), enabling fast, one-shot task assignment and scheduling. Our simultaneous model outperforms classical heuristics by assigning 47.10\% more feasible tasks given temporal constraints in 3.83\% of the time, metaheuristics by 68.60\% in 0.01\% of the time and exact solver by 101.48\% in 0.03\% of the time, while achieving $20\times$-to-$250\times$ speedup from prior graph-based methods.



Paperid:5733
Authors:Jie Zhao, Kang Cheong
Abstract:
While evolutionary algorithms (EAs) have proven effective in exploring the vast solution spaces characteristic of such problems, traditional encoding schemes, such as binary or numerical representations, often fail to capture the intricate structural properties of networks. Through employing image-based encodings to preserve topological context, this study utilizes multimodal large language models (MLLMs) as evolutionary operators to facilitate structure-aware optimization over graph data. To address the visual clutter inherent in large-scale network visualizations, we leverage graph sparsification techniques to simplify representations while maintaining essential structural features. To further enhance robustness and mitigate biases introduced by any single sparsification view, we propose a cooperative evolutionary framework that operates across multiple sparsified variants from diverse structural perspectives to achieve knowledge transfer among various domains. Additionally, acknowledging the sensitivity of MLLMs to network layout, we introduce an ensemble strategy that aggregates outputs from various layout configurations through consensus voting. Finally, experiments on real-world networks, illustrated through the classical influence maximization task, demonstrate that our approach substantially improves both the quality and resilience of solutions in MLLM-driven evolutionary optimization.



Paperid:5734
Authors:Joel Dapello, Marcel Nassar, Ridvan Eksi, Ban Wang, Jules Gagnon-Marchand, Kenneth Gao, akram Baharlouei, Kyra Thrush, Nina Riehs, Amy Peterson, Aniket Tolpadi, Abhejit Rajagopal, Henry Miller, Ashley Conard, David Alvarez-Melis, Rory Stark, Simone Bianco, Morgan Levine, Ava Amini, Alex X Lu, Nicolo Fusi, Ravi Pandya, Valentina Pedoia, Hana El-Samad
Abstract:
Understanding cellular responses to chemical interventions is critical to the discovery of effective therapeutics. Because individual biological techniques often measure only one axis of cellular response at a time, high-quality multimodal datasets are needed to unlock a holistic understanding of how cells respond to treatments and to advance computational methods that integrate modalities. However, many techniques destroy cells and thus preclude paired measurements, and attempts to match disparate unimodal datasets are often confounded by data being generated in incompatible experimental settings. Here we introduce scGeneScope, a multimodal single‑cell RNA sequencing (scRNA-seq) and Cell Painting microscopy image dataset conditionally paired by chemical treatment, designed to facilitate the development and benchmarking of unimodal, multimodal, and multiple profile machine learning methods for cellular profiling. 28 chemicals, each acting on distinct biological pathways or mechanisms of action (MoAs), were applied to U2-OS cells in two experimental data generation rounds, creating paired sets of replicates that were then profiled independently by scRNA‑seq or Cell Painting. Using scGeneScope, we derive a replicate- and experiment-split treatment identification benchmark simulating MoA discovery under realistic laboratory variability conditions and evaluate unimodal, multimodal, and multiprofile models ranging in complexity from linear approaches to recent foundation models. Multiprofile integration improved performance in both the unimodal and multimodal settings, with gains more consistent in the former. Evaluation of unimodal models for MoA identification demonstrated that recent scRNA-seq foundation models deployed zero-shot were consistently outperformed by classic fit-to-data methods, underscoring the need for careful, realistic benchmarking in machine learning for biology. We release the scGeneScope dataset and benchmarking code to support further research.



Paperid:5735
Authors:SHULI ZHANG, Hao Zhou, Jiaqi Zheng, Guibin Jiang, Cheng Bing, Wei Lin, Guihai Chen
Abstract:
Online Internet platforms require sophisticated marketing strategies to optimize user retention and platform revenue — a classical resource allocation problem. Traditional solutions adopt a two-stage pipeline: machine learning (ML) for predicting individual treatment effects to marketing actions, followed by operations research (OR) optimization for decision-making. This paradigm presents two fundamental technical challenges. First, the prediction-decision misalignment: Conventional ML methods focus solely on prediction accuracy without considering downstream optimization objectives, leading to improved predictive metrics that fail to translate to better decisions. Second, the bias-variance dilemma: Observational data suffers from multiple biases (e.g., selection bias, position bias), while experimental data (e.g., randomized controlled trials), though unbiased, is typically scarce and costly --- resulting in high-variance estimates. We proposeBi-levelDecision-FocusedCausalLearning (Bi-DFCL) that systematically addresses these challenges. First, we develop an unbiased estimator of OR decision quality using experimental data, which guides ML model training through surrogate loss functions that bridge discrete optimization gradients. Second, we establish a bi-level optimization framework that jointly leverages observational and experimental data, solved via implicit differentiation. This novel formulation enables our unbiased OR estimator to correct learning directions from biased observational data, achieving optimal bias-variance tradeoff. Extensive evaluations on public benchmarks, industrial marketing datasets, and large-scale online A/B tests conducted on one of the world's largest online food delivery platforms demonstrate the effectiveness of Bi-DFCL, showing statistically significant improvements over state-of-the-art baselines. Our code is now available at:https://anonymous.4open.science/r/Bi-DFCL.



Paperid:5736
Authors:Xiaorui Wu, Xiaofeng Mao, Xin Zhang, Fei Li, Donghong Ji, Chong Teng, Yuxiang Peng, Li Zheng, Zhuang Li
Abstract:
Large language models (LLMs) frequently refuse to respond to pseudo-malicious instructions: semantically harmless input queries triggering unnecessary LLM refusals due to conservative safety alignment, significantly impairing user experience. Collecting such instructions is crucial for evaluating and mitigating over-refusals, but existing instruction curation methods, like manual creation or instruction rewriting, either lack scalability or fail to produce sufficiently diverse and effective refusal-inducing prompts. To address these limitations, we introduce EVOREFUSE, a prompt optimization approach that generates diverse pseudo-malicious instructions consistently eliciting confident refusals across LLMs. EVOREFUSE employs an evolutionary algorithm exploring the instruction space in more diverse directions than existing methods via mutation strategies and recombination, and iteratively evolves seed instructions to maximize evidence lower bound on LLM refusal probability. Using EVOREFUSE, we create two novel datasets: EVOREFUSE-TEST, a benchmark of 582 pseudo-malicious instructions that outperforms the next-best benchmark with 140.41% higher average refusal triggering rate across 9 LLMs, 34.86% greater lexical diversity, and 40.03% improved LLM response confidence scores; and EVOREFUSE-ALIGN, which provides 3,000 pseudo-malicious instructions with responses for supervised and preference-based alignment training. LLAMA3.1-8B-INSTRUCT supervisedly fine-tuned on EVOREFUSE-ALIGN achieves up to 14.31% fewer over-refusals than models trained on the second-best alignment dataset, without compromising safety. Our analysis with EVOREFUSE-TEST reveals models trigger over-refusals by overly focusing on sensitive keywords while ignoring broader context.



Paperid:5737
Authors:Matthew Bull, Po-Chen Kuo, Andrew Smith, Michael Buice
Abstract:
The modern connectivist framing of neural computation emphasizes the primacy of synaptic communication at the risk of neglecting the influence of the surrounding neuromodulatory environment --- a neuron's 'biophysical context.' Decades of experimental work has established two views of neuromodulatory (NMs) influence: 1) NMs significantly alter circuit dynamics and 2) NMs gate synaptic plasticity, acting as a 'third factor' in learning. Here, we unify these perspectives, proposing that neuromodulation via volume transmission implements a powerful computational principle: context factorization. We derive an endogenously neuromodulated Recurrent Neural Network (e-nmRNN) from a rate reduction of NM release, showing how NM concentrations dynamically factorize network connectivity. This framework reveals how multiplicative NM gating distinctly influences dynamical regimes compared to additive input. Crucially, this context factorization enables targeted online credit assignment: learning rules derived for the e-nmRNN are naturally gated by NM concentrations, localizing updates to relevant contexts. We demonstrate that e-nmRNN dynamics can learn to approximate gradient descent, facilitating rapid in-context adaptation akin to meta-learning. Empirically, e-nmRNNs achieve strong compositional generalization in sequence-to-sequence tasks, outperforming baselines and exhibiting greater hyperparameter robustness. Furthermore, when trained on complex multitasking benchmarks, e-nmRNNs develop emergent properties mirroring biological observations, including modularity, cell-type specialization based on NM release, and distinct neuromodulatory timescales encoding task context. The model's interpretability allows us to reverse engineer these emergent structures. Notably, in reinforcement learning tasks, the e-nmRNN learns to encode context and signals like Reward Prediction Error (RPE) within its neuromodulator dynamics, demonstrating a mechanism for RPE-gated online credit assignment essential for learning how to learn. By bridging biophysical mechanisms with computational principles and empirical validation, our work presents e-nmRNNs as a performant, interpretable model for understanding the computational role of neuromodulation in flexible and compositional learning.



Paperid:5738
Authors:Mouïn Ben Ammar, David Brellmann, Arturo Mendoza, Antoine Manzanera, Gianni Franchi
Abstract:
Out-of-distribution (OOD) detectionis essential for ensuring the reliability and safety of machine learning systems. In recent years, it has received increasing attention, particularly through post-hoc detection and training-based methods. In this paper, we focus onpost-hoc OOD detection, which enables identifying OOD samples without altering the model's training procedure or objective. Our primary goal is to investigate the relationship betweenmodel capacityand its OOD detection performance. Specifically, we aim to answer the following question:Does the Double Descent phenomenon manifest in post-hoc OOD detection?This question is crucial, as it can reveal whether overparameterization, which is already known to benefit generalization, can also enhance OOD detection.Despite the growing interest in these topics by the classic supervised machine learning community, this intersection remains unexplored for OOD detection.We empirically demonstrate that the Double Descent effect does indeed appear in post-hoc OOD detection. Furthermore, we provide theoretical insights to explain why this phenomenon emerges in such setting. Finally, we show that the overparameterized regime does not yield superior results consistently, and we propose a method to identify the optimal regime for OOD detection based on our observations.



Paperid:5739
Authors:Tianxiang Zhao, Youqing Wang, Jinlu Wang, Jiapu Wang, Mingliang Cui, Junbin Gao, Jipeng Guo
Abstract:
Due to its powerful capability of self-supervised representation learning and clustering, contrastive attributed graph clustering (CAGC) has achieved great success, which mainly depends on effective data augmentation and contrastive objective setting. However, most CAGC methods utilize edges as auxiliary information to obtain node-level embedding representation and only focus on node-level embedding augmentation. This approach overlooks edge-level embedding augmentation and the interactions between node-level and edge-level embedding augmentations across various granularity. Moreover, they often treat all contrastive sample pairs equally, neglecting the significant differences between hard and easy positive-negative sample pairs, which ultimately limits their discriminative capability. To tackle these issues, a novel robust attributed graph clustering (RAGC), incorporating hybrid-collaborative augmentation (HCA) and contrastive sample adaptive-differential awareness (CSADA), is proposed. First, node-level and edge-level embedding representations and augmentations are simultaneously executed to establish a more comprehensive similarity measurement criterion for subsequent contrastive learning. In turn, the discriminative similarity further consciously guides edge augmentation. Second, by leveraging pseudo-label information with high confidence, a CSADA strategy is elaborately designed, which adaptively identifies all contrastive sample pairs and differentially treats them by an innovative weight modulation function. The HCA and CSADA modules mutually reinforce each other in a beneficent cycle, thereby enhancing discriminability in representation learning. Comprehensive graph clustering evaluations over six benchmark datasets demonstrate the effectiveness of the proposed RAGC against several state-of-the-art CAGC methods.



Paperid:5740
Authors:Xiaoyang Zhang, He Fang, Yang Deng, Dan Wang
Abstract:
Abstract:The rapid growth of large AI models has raised significant environmental concerns due to their substantial carbon footprint. Existing carbon accounting methods for AI models are fundamentally deterministic and fail to account for inherent uncertainties in embodied and operational carbon emissions. Our work aims to investigate the effect of these uncertainties on embodied and operational carbon footprint estimates for large AI models. We propose a Probabilistic Carbon Accounting Model (PCAM), which quantifies uncertainties across three dimensions in the carbon accounting for AI models: geotemporal manufacturing variability, dynamic manufacturing evolution, and dynamic operating context. We develop parameter models to quantify key components (processors, memory, storage) in the carbon footprint of AI models. To characterize the distribution of the parameters, we develop a carbon dataset by aggregating related data from various sources. Then, we implement a simple yet effective dual-stage distribution modeling to generate the distribution of the parameters from the collected dataset through Kernel Density Estimate (KDE). We compare the performance of PCAM with LLMCarbon, the state-of-the-art carbon accounting method for large AI models. PCAM achieves $\leq7.44$% error compared to LLMCarbon’s $\leq108.51$%.



Paperid:5741
Authors:Yunlong Tang, Pinxin Liu, Mingqian Feng, Zhangyun Tan, Rui Mao, Chao Huang, Jing Bi, Yunzhong Xiao, Susan Liang, Hang Hua, Ali Vosoughi, Luchuan Song, Zeliang Zhang, Chenliang Xu
Abstract:
Understanding perspective is fundamental to human visual perception, yet the extent to which multimodal large language models (MLLMs) internalize perspective geometry remains unclear. We introduce MMPerspective, the first benchmark specifically designed to systematically evaluate MLLMs' understanding of perspective through 10 carefully crafted tasks across three complementary dimensions: Perspective Perception, Reasoning, and Robustness. Our benchmark comprises 2,711 real-world and synthetic image instances with 5,083 question-answer pairs that probe key capabilities, such as vanishing point perception and counting, perspective type reasoning, line relationship understanding in 3D space, invariance to perspective-preserving transformations, etc. Through a comprehensive evaluation of 43 state-of-the-art MLLMs, we uncover significant limitations: while models demonstrate competence on surface-level perceptual tasks, they struggle with compositional reasoning and maintaining spatial consistency under perturbations. Our analysis further reveals intriguing patterns between model architecture, scale, and perspective capabilities, highlighting both robustness bottlenecks and the benefits of chain-of-thought prompting. MMPerspective establishes a valuable testbed for diagnosing and advancing spatial understanding in vision-language systems.



Paperid:5742
Authors:Weiming Liu, Xinting Liao, Jun Dan, Fan Wang, Hua Yu, Junhao Dong, Shunjie Dong, Lianyong Qi, Yew Soon Ong
Abstract:
Semi-Unbalanced Optimal Transport (SemiUOT) shows great promise in matching two probability measures by relaxing one of the marginal constraints. Previous solvers often incorporate an entropy regularization term, which can result in inaccurate matching solutions. To address this issue, we focus on determining the marginal probability distribution of SemiUOT with KL divergence using the proposed Equivalent Transformation Mechanism (ETM) approach. Furthermore, we extend the ETM-based method into exploiting the marginal probability distribution of Unbalanced Optimal Transport (UOT) with KL divergence for validating its generalization. Once the marginal probabilities of UOT/SemiUOT are determined, they can be transformed into a classical Optimal Transport (OT) problem. Moreover, we propose a KKT-Multiplier regularization term combined with Multiplier Regularized Optimal Transport (MROT) to achieve more accurate matching results. We conduct several numerical experiments to demonstrate the effectiveness of our proposed methods in addressing UOT/SemiUOT problems.



Paperid:5743
Authors:Dengfeng Xue, Wenjuan Li, Yifan Lu, chunfeng yuan, Yufan Liu, Wei Liu, Man Yao, Li Yang, Guoqi Li, Bing Li, Stephen Maybank, Weiming Hu, Zhetao Li
Abstract:
Brain-inspired spiking neural networks (SNNs) provide energy-efficient computation through event-driven processing. However, the shared weights across multiple timesteps lead to serious temporal feature redundancy, limiting both efficiency and performance. This issue is further aggravated when processing static images due to the duplicated input. To mitigate this problem, we propose a parameter-free and plug-and-play module named Mutual Information-based Temporal Redundancy Quantification and Reduction (MI-TRQR), constructing energy-efficient SNNs. Specifically, Mutual Information (MI) is properly introduced to quantify redundancy between discrete spike features at different timesteps on two spatial scales: pixel (local) and the entire spatial features (global). Based on the multi-scale redundancy quantification, we apply a probabilistic masking strategy to remove redundant spikes. The final representation is subsequently recalibrated to account for the spike removal. Extensive experimental results demonstrate that our MI-TRQR achieves sparser spiking firing, higher energy efficiency, and better performance concurrently with different SNN architectures in tasks of neuromorphic data classification, static data classification, and time-series forecasting. Notably, MI-TRQR increases accuracy by \textbf{1.7\%} on CIFAR10-DVS with 4 timesteps while reducing energy cost by \textbf{37.5\%}.



Paperid:5744
Authors:Karthikeyan Chandra Sekaran, Markus Geisler, Dominik Rößle, Adithya Mohan, Daniel Cremers, Wolfgang Utschick, Michael Botsch, Werner Huber, Torsten Schön
Abstract:
Recent cooperative perception datasets have played a crucial role in advancing smart mobility applications by enabling information exchange between intelligent agents, helping to overcome challenges such as occlusions and improving overall scene understanding. While some existing real-world datasets incorporate both vehicle-to-vehicle and vehicle-to-infrastructure interactions, they are typically limited to a single intersection or a single vehicle. A comprehensive perception dataset featuring multiple connected vehicles and infrastructure sensors across several intersections remains unavailable, limiting the benchmarking of algorithms in diverse traffic environments. Consequently, overfitting can occur, and models may demonstrate misleadingly high performance due to similar intersection layouts and traffic participant behavior. To address this gap, we introduce UrbanIng-V2X, the first large-scale, multi-modal dataset supporting cooperative perception involving vehicles and infrastructure sensors deployed across three urban intersections in Ingolstadt, Germany. UrbanIng-V2X consists of 34 temporally aligned and spatially calibrated sensor sequences, each lasting 20 seconds. All sequences contain recordings from one of three intersections, involving two vehicles and up to three infrastructure-mounted sensor poles operating in coordinated scenarios. In total, UrbanIng-V2X provides data from 12 vehicle-mounted RGB cameras, 2 vehicle LiDARs, 17 infrastructure thermal cameras, and 12 infrastructure LiDARs. All sequences are annotated at a frequency of 10 Hz with 3D bounding boxes spanning 13 object classes, resulting in approximately 712k annotated instances across the dataset. We provide comprehensive evaluations using state-of-the-art cooperative perception methods and publicly release the codebase, dataset, HD map, and a digital twin of the complete data collection environment.



Paperid:5745
Authors:Kunyu Peng, Junchao Huang, Xiangsheng Huang, Di Wen, Junwei Zheng, Yufan Chen, Kailun Yang, Jiamin Wu, Chongqing Hao, Rainer Stiefelhagen
Abstract:
Action segmentation is a core challenge in high-level video understanding, aiming to partition untrimmed videos into segments and assign each a label from a predefined action set. Existing methods primarily address single-person activities with fixed action sequences, overlooking multi-person scenarios. In this work, we pioneer textual reference-guided human action segmentation in multi-person settings, where a textual description specifies the target person for segmentation. We introduce the first dataset for Referring Human Action Segmentation, i.e., RHAS133, built from 133 movies and annotated with 137 fine-grained actions with 33h video data, together with textual descriptions for this new task. Benchmarking existing action recognition methods on RHAS133 using VLM-based feature extractors reveals limited performance and poor aggregation of visual cues for the target person. To address this, we propose aholistic-partial aware Fourier-conditioned diffusion framework, i.e., HopaDIFF, leveraging a novel cross-input gate attentional xLSTM to enhance holistic-partial long-range reasoning and a novel Fourier condition to introduce more fine-grained control to improve the action segmentation generation. HopaDIFF achieves state-of-the-art results on RHAS133 in diverse evaluation settings. The dataset and code are available in the supplementary.



Paperid:5746
Authors:Sina Malakouti, Adriana Kovashka
Abstract:
Text-to-image (T2I) diffusion models exhibit impressive photorealistic image generation capabilities, yet they struggle in compositional image generation. In this work, we introduce RoleBench, a benchmark focused on evaluating compositional generalization in action-based relations (e.g., mouse chasing cat). We show that state-of-the-art T2I models and compositional approaches consistently default to frequent reversed relations (i.e., cat chasing mouse), a phenomenon we call Role-Collapse. Related works attribute this to the model's architectural limitation or being underrepresented in the data. Our key insight reveals that while models fail on rare compositions when their inversions are common, they can successfully generate similar intermediate compositions (e.g., ``mouse chasing boy"), suggesting that this limitation is due to the presence of frequent counterparts rather than the absence of rare compositions. Motivated by this, we hypothesize that directional decomposition can gradually mitigate role collapse. We test this via ReBind, a lightweight framework that teaches role bindings using carefully selected active/passive intermediaries. Experiments suggest that intermediate compositions through intermediate fine-tuning can significantly mitigate role bias, with humans preferring more than 78% compared to state-of-the-art methods. Our findings highlight the role of distributional asymmetries in compositional failures and offer a simple, effective path to improving generalization.



Paperid:5747
Authors:Yifan Sun, Jingyan Shen, Yibin Wang, Tianyu Chen, Zhendong Wang, Mingyuan Zhou, Huan Zhang
Abstract:
Abstract:Reinforcement learning (RL) has become an effective approach for fine-tuning large language models (LLMs), particularly to enhance their reasoning capabilities. However, RL fine-tuning remains highly resource-intensive, and existing work has largely overlooked the problem of data efficiency. In this paper, we propose two techniques to improve data efficiency in LLM RL fine-tuning: \textit{difficulty-targeted online data selection} and \textit{rollout replay}. We introduce the notion of adaptive difficulty to guide online data selection, prioritizing questions of moderate difficulty that are more likely to yield informative learning signals. To estimate adaptive difficulty efficiently, we develop an attention-based framework that requires rollouts for \textit{only} a small reference set of questions. The adaptive difficulty of the remaining questions is then estimated based on their similarity to this set. To further reduce rollout cost, we introduce a rollout replay mechanism that reuses recent rollouts, lowering per-step computation while maintaining stable updates. Extensive experiments across $6$ LLM-dataset combinations show that our method reduces RL fine-tuning time by $25\\%$ to $65\\%$ to reach the same level of performance as the original GRPO algorithm.



Paperid:5748
Authors:Youmin Ko, Sungjong Seo, Hyunjoon Kim
Abstract:
Since large language models (LLMs) have a tendency to generate factually inaccurate output, retrieval-augmented generation (RAG) has gained significant attention as a key means to mitigate this downside of harnessing only LLMs. However, existing RAG methods for simple and multi-hop question answering (QA) are still prone to incorrect retrievals and hallucinations. To address these limitations, we propose CoopRAG, a novel RAG framework for the question answering task in which a retriever and an LLM work cooperatively with each other by exchanging informative knowledge, and the earlier and later layers of the retriever model work cooperatively with each other to accurately rank the retrieved documents relevant to a given query. In this framework, we (i) unroll a question into sub-questions and a reasoning chain in which uncertain positions are masked, (ii) retrieve the documents relevant to the question augmented with the sub-questions and the reasoning chain, (iii) rerank the documents by contrasting layers of the retriever, and (iv) reconstruct the reasoning chain by filling the masked positions via the LLM. Our experiments demonstrate that CoopRAG consistently outperforms state-of-the-art QA methods on three multi-hop QA datasets as well as a simple QA dataset in terms of both the retrieval and QA performances. Our code is available.



Paperid:5749
Authors:Chenyu Zhu, Yefeng Liu, Hao Zhang, Aowen Wang, Yangxue, Guanhua Chen, Longyue Wang, Weihua Luo, Kaifu Zhang
Abstract:
Despite their outstanding performance in numerous applications, large language models (LLMs) remain prone to hallucinations, generating content inconsistent with their pretraining corpora. Currently, almost all contrastive decoding approaches alleviate hallucinations by introducing a model susceptible to hallucinations and appropriately widening the contrastive logits gap between hallucinatory tokens and target tokens. However, although existing contrastive decoding methods mitigate hallucinations, they lack enough confidence in the factual accuracy of the generated content. In this work, we propose Multi-Model Contrastive Decoding (MCD), which integrates a pretrained language model with an evil model and a truthful model for contrastive decoding. Intuitively, a token is assigned a high probability only when deemed potentially hallucinatory by the evil model while being considered factual by the truthful model. This decoding strategy significantly enhances the model’s confidence in its generated responses and reduces potential hallucinations. Furthermore, we introduce a dynamic hallucination detection mechanism that facilitates token-by-token identification of hallucinations during generation and a tree-based revision mechanism to diminish hallucinations further. Extensive experimental evaluations demonstrate that our MCD strategy effectively reduces hallucinations in LLMs and outperforms state-of-the-art methods across various benchmarks.



Paperid:5750
Authors:Shihao Yang, Feng Liu
Abstract:
In the past decades, brain imaging research underwent a shift from mapping tasked evoked brain regions of activations towards identifying and characterizing the dynamic brain networks of multiple coordinating brain regions. Electrophysiological signals are the direct manifestation of brain activities, thus, characterizing the whole brain electrophysiological networks (WBEN) can serve as a fundamental tool for neuroscience studies and clinical applications. In this work, we introduce the first framework for the integration of scalp EEG and intracranial EEG (iEEG) for the WBEN estimation with a principled estimation framework based on state-space models, where an Expectation-Maximization (EM) algorithm is designed to infer the state variables and brain connectivity simultaneously. We validated the proposed method on synthetic data, and the results revealed improved performance compared to traditional two-step methods using scalp EEG only, demonstrating the importance of including iEEG signal for WBEN estimation. For real data with simultaneous EEG and iEEG, we applied the developed framework to understand the information flows of the encoding and maintenance phases during the working memory task. The information flows between the subcortical and cortical regions are delineated, highlighting more significant information flows from cortical to subcortical regions than maintenance phases. The results are consistent with previous research findings but with the view of the whole brain scope, which underscores the unique utility of the proposed framework.



Paperid:5751
Authors:Hanyu Zhu, Lance Fiondella, Jiawei Yuan, Kai Zeng, Long Jiao
Abstract:
Retrieval-Augmented Generation (RAG) empowers Large Language Models (LLMs) to dynamically integrate external knowledge during inference, improving their factual accuracy and adaptability. However, adversaries can inject poisoned external knowledge to override the model’s internal memory. While existing attacks iteratively manipulate retrieval content or prompt structure of RAG, they largely ignore the model’s internal representation dynamics and neuron-level sensitivities. The underlying mechanism of RAG poisoning hasn’t been fully studied and the knowledge conflict effect with strong parametric knowledge in RAG is not considered. In this work, we propose NeuroGenPoisoning, a novel attack framework that generates adversarial external knowledge in RAG guided by LLM internal neuron attribution and genetic optimization. Our method first identifies a set of \textbf{Poison-Responsive Neurons} whose activation strongly correlates with contextual poisoning knowledge. We then employ a genetic algorithm to evolve adversarial passages that maximally activate these neurons. Crucially, our framework enables massive-scale generation of effective poisoned RAG knowledge by identifying/reusing promising but initially unsuccessful external knowledge variants via observed attribution signals. At the same time, poison-responsive neurons guided poisoning can effectively resolves knowledge conflict. Experimental results across models and datasets demonstrate consistently achieving high Population Overwrite Success Rate (POSR) of over 90\% while preserving fluency. Empirical evidence shows that our method effectively resolves knowledge conflict.



Paperid:5752
Authors:Junchao Gong, Jingyi Xu, Ben Fei, Fenghua Ling, zhangwenlong, Kun Chen, Wanghan Xu, Weidong Yang, Xiaokang Yang, LEI BAI
Abstract:
Weather prediction is a critical task for human society, where impressive progress has been made by training artificial intelligence weather prediction (AIWP) methods with reanalysis data. However, reliance on reanalysis data limits the AIWPs with shortcomings, including data assimilation biases and temporal discrepancies. To liberate AIWPs from the reanalysis data, observation forecasting emerges as a transformative paradigm for weather prediction. One of the key challenges in observation forecasting is learning spatiotemporal dynamics across disparate measurement systems with irregular high-resolution observation data, which constrains the design and prediction of AIWPs. To this end, we propose our DAWP as an innovative framework to enable AIWPs to operate in a complete observation space by initialization with an artificial intelligence data assimilation (AIDA) module. Specifically, our AIDA module applies a mask multi-modality autoencoder (MMAE) for assimilating irregular satellite observation tokens encoded by mask ViT-VAEs. For AIWP, we introduce a spatiotemporal decoupling transformer with cross-regional boundary conditioning (CBC), learning the dynamics in observation space, to enable sub-image-based global observation forecasting. Comprehensive experiments demonstrate that AIDA initialization significantly improves the roll-out and efficiency of AIWP. Additionally, we show that DAWP holds promising potential to be applied in global precipitation forecasting.



Paperid:5753
Authors:Chau Pham, Juan C. Caicedo, Bryan Plummer
Abstract:
Prior work using Masked Autoencoders (MAEs) typically relies on random patch masking based on the assumption that images have significant redundancies across different channels, allowing for the reconstruction of masked content using cross-channel correlations. However, this assumption does not hold in Multi-Channel Imaging (MCI), where channels may provide complementary information with minimal feature overlap. Thus, these MAEs primarily learn local structures within individual channels from patch reconstruction, failing to fully leverage cross-channel interactions and limiting their MCI effectiveness. In this paper, we present ChA-MAEViT, an MAE-based method that enhances feature learning across MCI channels via four key strategies: (1) dynamic channel-patch masking, which compels the model to reconstruct missing channels in addition to masked patches, thereby enhancing cross-channel dependencies and improving robustness to varying channel configurations; (2) memory tokens, which serve as long-term memory aids to promote information sharing across channels, addressing the challenges of reconstructing structurally diverse channels; (3) hybrid token fusion module, which merges fine-grained patch tokens with a global class token to capture richer representations; and (4) Channel-Aware Decoder, a lightweight decoder utilizes channel tokens to effectively reconstruct image patches. Experiments on satellite and microscopy datasets, CHAMMI, JUMP-CP, and So2Sat, show that ChA-MAEViT significantly outperforms state-of-the-art MCI-ViTs by 3.0-21.5%, highlighting the importance of cross-channel interactions in MCI.



Paperid:5754
Authors:H M Sabbir Ahmad, Ehsan Sabouni, Alexander Wasilkoff, Param Budhraja, Zijian Guo, Songyuan Zhang, Chuchu Fan, Christos G Cassandras, Wenchao Li
Abstract:
Abstract:We address the problem of safe policy learning in multi-agent safety-critical autonomous systems.In such systems, it is necessary for each agent to meet the safety requirements at all times while also cooperating with other agents to accomplish the task. Toward this end, we propose a safe Hierarchical Multi-Agent Reinforcement Learning (HMARL) approach based on Control Barrier Functions (CBFs). Our proposed hierarchical approach decomposes the overall reinforcement learning problem into two levels –- learning joint cooperative behavior at the higher level and learning safe individual behavior at the lower or agentlevel conditioned on the high-level policy. Specifically, we propose a skill-based HMARL-CBF algorithm in which the higher-level problem involves learning a joint policy over the skills for all the agents and the lower-level problem involveslearning policies to execute the skills safely with CBFs. We validate our approach on challenging environment scenarios whereby a large number of agents have to safely navigate through conflicting road networks. Compared with existing state-of-the-art methods, our approach significantly improves the safety achieving near perfect (within $5\%$) success/safety rate while also improving performance across all the environments.



Paperid:5755
Authors:Martin Pelikan, Shams Azam, Vitaly Feldman, Jan Silovsky, Kunal Talwar, Christopher Brinton, Tatiana Likhomanenko
Abstract:
Abstract:While federated learning (FL) and differential privacy (DP) have been extensively studied, their application to automatic speech recognition (ASR) remains largely unexplored due to the challenges in training large transformer models. Specifically, large models further exacerbate issues in FL as they are particularly susceptible to gradient heterogeneity across layers, unlike the relatively uniform gradient behavior observed in shallow models. As a result, prior works struggle to converge with standard optimization techniques, even in the absence of DP mechanisms. To the best of our knowledge, no existing work establishes a competitive, practical recipe for FL with DP in the context of ASR. To address this gap, we establish **the first benchmark for FL with DP** in end-to-end ASR. Our approach centers on per-layer clipping and layer-wise gradient normalization: theoretical analysis reveals that these techniques together mitigate clipping bias and gradient heterogeneity across layers in deeper models. Consistent with these theoretical insights, our empirical results show that FL with DP is viable under strong privacy guarantees, provided a population of at least several million users. Specifically, we achieve user-level ($7.2$, $10^{-9}$)-**DP** (resp. ($4.5$, $10^{-9}$)-**DP**) with only a 1.3\% (resp. 4.6\%) absolute drop in word error rate when extrapolating to high (resp. low) population scales for **FL with DP in ASR**.Although our experiments focus on ASR, the underlying principles we uncover — particularly those concerning gradient heterogeneity and layer-wise gradient normalization — offer broader guidance for designing scalable, privacy-preserving FL algorithms for large models across domains.



Paperid:5756
Authors:Shuaiqi Wang, Vikas Raunak, Arturs Backurs, Victor Reis, Pei Zhou, Sihao Chen, Longqi Yang, Zinan Lin, Sergey Yekhanin, Giulia Fanti
Abstract:
Differentially private (DP) synthetic data generation (SDG) is a promising technique for making use of private datasets that otherwise cannot be exposed for model training or other analytics. While much research literature has focused on generating private unstructured text and image data, in enterprise settings, structured data (e.g., tabular) is more common, often including natural language fields or components. Such datasets have structural properties and correlations that are difficult to evaluate using only existing synthetic data evaluation techniques (e.g., FID). In this work, we propose Struct-Bench, a framework and benchmark for evaluating private synthetic datasets based on structured datasets that contain natural language data. The Struct-Bench framework requires users to provide a representation of their dataset structure as a Context-Free Grammar (CFG). Our benchmark comprises 5 real-world and 2 synthetically generated datasets, each annotated with Context-Free Grammars (CFGs). We show that these datasets demonstrably present a great challenge even for state-of-the-art DP synthetic data generation methods. Struct-Bench also includes reference implementations of different metrics and a leaderboard, thereby providing researchers a standardized evaluation platform to benchmark and investigate privacy-preserving synthetic data generation methods. Further, we also present a case study showing how to use Struct-Bench to improve the synthetic data quality of Private Evolution on structured data. The benchmark and the leaderboard have been publicly made available at https://struct-bench.github.io.



Paperid:5757
Authors:Hao Fang, Changle Zhou, Jiawei Kong, Kuofeng Gao, Bin Chen, Tao Liang, Guojun Ma, Shu-Tao Xia
Abstract:
Large Vision-Language Models (LVLMs) are susceptible to hallucinations, where generated responses seem semantically plausible yet exhibit little or no relevance to the input image. Previous studies reveal that this issue primarily stems from LVLMs' over-reliance on language priors while disregarding the visual information during decoding. To alleviate this issue, we introduce a novel Conditional Pointwise Mutual Information (C-PMI) calibrated decoding strategy, which adaptively strengthens the mutual dependency between generated texts and input images to mitigate hallucinations. Unlike existing methods solely focusing on text token sampling, we propose to jointly model the contributions of visual and textual tokens to C-PMI, formulating hallucination mitigation as a bi-level optimization problem aimed at maximizing mutual information. To solve it, we design a token purification mechanism that dynamically regulates the decoding process by sampling text tokens remaining maximally relevant to the given image, while simultaneously refining image tokens most pertinent to the generated response. Extensive experiments across various benchmarks reveal that the proposed method significantly reduces hallucinations in LVLMs while preserving decoding efficiency.



Paperid:5758
Authors:Yide Qiu, Tong Zhang, Shaoxiang Ling, Xing Cai, Ziqi Gu, Zhen Cui
Abstract:
Irregular data in the real world are usually organized as heterogeneous graphs consisting of multiple types of nodes and edges. However, current heterogeneous graph research confronts three fundamental challenges: i) Benchmark Deficiency, ii) Semantic Disalignment, and iii) Propagation Degradation. In this paper, we construct a large-scale, universal, and joint multi-domain heterogeneous graph dataset named UniHG to facilitate heterogeneous graph representation learning and cross-domain knowledge mining. Overall, UniHG contains 77.31 million nodes and 564 million directed edges with thousands of labels and attributes, which is currently the largest universal heterogeneous graph dataset available to the best of our knowledge. To perform effective learning and provide comprehensively benchmarks on UniHG , two key measures are taken, including i) the semantic alignment strategy for multi-attribute entities, which projects the feature description of multi-attribute nodes and edges into a common embedding space to facilitate information aggregation; ii) proposing the novel Heterogeneous Graph Decoupling (HGD) framework with a specifically designed Anisotropy Feature Propagation (AFP) module for learning effective multi-hop anisotropic propagation kernels. These two strategies enable efficient information propagation among a tremendous number of multi-attribute entities and meanwhile mine multi-attribute association adaptively through the multi-hop aggregation in large-scale heterogeneous graphs. Comprehensive benchmark results demonstrate that our model significantly outperforms existing methods with an accuracy improvement of 28.93\%. And the UniHG can facilitate downstream tasks, achieving an NDCG@20 improvement rate of 11.48\% and 11.71\%. The UniHG dataset and benchmark codes have been released at https://anonymous.4open.science/r/UniHG-AA78.



Paperid:5759
Authors:Jiaming Li, Zhijia Liang, Weikai Chen, Lin Ma, Guanbin Li
Abstract:
Fine-grained open-vocabulary object detection (FG-OVD) aims to detect novel object categories described by attribute-rich texts. While existing open-vocabulary detectors show promise at the base-category level, they underperform in fine-grained settings due to the semantic entanglement of subjects and attributes in pretrained vision-language model (VLM) embeddings -- leading to over-representation of attributes, mislocalization, and semantic drift in embedding space. We propose GUIDED, a decomposition framework specifically designed to address the semantic entanglement between subjects and attributes in fine-grained prompts. By separating object localization and fine-grained recognition into distinct pathways, GUIDED aligns each subtask with the module best suited for its respective roles. Specifically, given a fine-grained class name, we first use a language model to extract a coarse-grained subject and its descriptive attributes. Then the detector is guided solely by the subject embedding, ensuring stable localization unaffected by irrelevant or overrepresented attributes. To selectively retain helpful attributes, we introduce an attribute embedding fusion module that incorporates attribute information into detection queries in an attention-based manner. This mitigates over-representation while preserving discriminative power.Finally, a region-level attribute discrimination module compares each detected region against full fine-grained class names using a refined vision-language model with a projection head for improved alignment.Extensive experiments on FG-OVD and 3F-OVD benchmarks show that GUIDED achieves new state-of-the-art results, demonstrating the benefits of disentangled modeling and modular optimization.



Paperid:5760
Authors:Yuhan Huang, Keren Gao, Dongping Yang, Sen Song, Guozhang Chen
Abstract:
The brain's diverse spatiotemporal activity patterns are fundamental to cognition and consciousness, yet how these macroscopic dynamics emerge from microscopic neural circuitry remains a critical challenge. We address this by developing a spatially extended neural network model integrated with a spectral theory of its connectivity matrix. Our theory quantitatively demonstrates how local structural parameters, such as E/I neuron projection ranges, connection strengths and density determine distinct features of the eigenvalue spectrum, specifically outlier eigenvalues and a bulk disk. These spectral signatures, in turn, precisely predict the network's emergent global dynamical regime, encompassing asynchronous states, synchronous states, oscillations, localized activity bumps, traveling waves, and chaos. Motivated by observations of shifting cortical dynamics in mice across arousal states, our framework not only explains this repertoire of behaviors but also offers a principled approach to inferring underlying effective connectivity changes from macroscopic brain activity. By mechanistically linking neural structure to dynamics, this work provides a powerful tool for understanding brain function and paves they way for identifying potential biomarkers for neurological disorders.



Paperid:5761
Authors:Maximilian Forstenhäusler, Daniel Külzer, Christos Anagnostopoulos, Shameem Puthiya Parambath, Natascha Weber
Abstract:
Understanding user intent is essential for situational and context-aware decision-making. Motivated by a real-world scenario, this work addresses intent predictions of smart device users in the vicinity of vehicles by modeling sequential spatiotemporal data. However, in real-world scenarios, environmental factors and sensor limitations can result in non-stationary and irregularly sampled data, posing significant challenges. To address these issues, we propose STaRFormer, a Transformer-based approach that can serve as a universal framework for sequential modeling. STaRFormer utilizes a new dynamic attention-based regional masking scheme combined with a novel semi-supervised contrastive learning paradigm to enhance task-specific latent representations. Comprehensive experiments on 19 datasets varying in types (including non-stationary and irregularly sampled), domains, sequence lengths, training samples, and applications demonstrate the efficacy of STaRFormer. We achieve notable improvements over state-of-the-art approaches. Code and data are available via https://anonymous.4open.science/r/STaRFormer-78D8/.



Paperid:5762
Authors:Tianwei Wang, Wei Pang, Xinhui Ma
Abstract:
Abstract:Motivated by the geometric advantages of quaternions in representing rotations and postures, we propose a quaternion-valued supervised learning Hopfield-structured neural network (QSHNN) with a fully connected structure inspired by the classic Hopfield neural network (HNN). Starting from a continuous-time dynamical model of HNNs, we extend the formulation to the quaternionic domain and establish the existence and uniqueness of fixed points with asymptotic stability. For the learning rules, we introduce a periodic projection strategy that modifies standard gradient descent by periodically projecting each $4\times 4$ block of the weight matrix onto the closest quaternionic structure in the least-squares sense. This approach preserves both convergence and quaternionic consistency throughout training. Benefiting from this rigorous mathematical foundation, the experimental model implementation achieves high accuracy, fast convergence, and strong reliability across randomly generated target sets. Moreover, the evolution trajectories of the QSHNN exhibit well-bounded curvature, i.e., sufficient smoothness, which is crucial for applications such as control systems or path planning modules in robotic arms, where joint postures are parameterized by quaternion neurons. Beyond these application scenarios, the proposed model offers a practical implementation framework and a general mathematical methodology for designing neural networks under hypercomplex or non-commutative algebraic structures.



Paperid:5763
Authors:Haoran Sun, Yankai Jiang, Wenjie Lou, Yujie Zhang, Wenjie Li, Lilong Wang, Mianxin Liu, Lei Liu, Xiaosong Wang
Abstract:
Multimodal large language models (MLLMs) have begun to demonstrate robust reasoning capabilities on general tasks, yet their application in the medical domain remains in its early stages. Constructing chain-of-thought (CoT) training data is essential for bolstering the reasoning abilities of medical MLLMs. However, existing approaches exhibit a deficiency in offering a comprehensive framework for searching and evaluating effective reasoning paths towards critical diagnosis. To address this challenge, we propose Mentor-Intern Collaborative Search (MICS), a novel reasoning-path searching scheme to generate rigorous and effective medical CoT data. MICS first leverages mentor models to initialize the reasoning, one step at a time, then prompts each intern model to continue the thinking along those initiated paths, and finally selects the optimal reasoning path according to the overall reasoning performance of multiple intern models. The reasoning performance is determined by an MICS-Score, which assesses the quality of generated reasoning paths. Eventually, we construct MMRP, a multi-task medical reasoning dataset with ranked difficulty, and Chiron-o1, a new medical MLLM devised via a curriculum learning strategy, with robust visual question-answering and generalizable reasoning capabilities. Extensive experiments demonstrate that Chiron-o1, trained on our CoT dataset constructed using MICS, achieves state-of-the-art performance across a list of medical visual question answering and reasoning benchmarks. Our model and code will be publicly available.



Paperid:5764
Authors:Mengru Wang, Xingyu Chen, Yue Wang, Zhiwei He, Jiahao Xu, Tian Liang, Qiuzhi Liu, Yunzhi Yao, Wenxuan Wang, Ruotian Ma, Haitao Mi, Ningyu Zhang, Zhaopeng Tu, Xiaolong Li, Dong Yu
Abstract:
Mixture-of-Experts (MoE) architectures within Large Reasoning Models (LRMs) have achieved impressive reasoning capabilities by selectively activating experts to facilitate structured cognitive processes. Despite notable advances, existing reasoning models often suffer from cognitive inefficiencies like overthinking and underthinking. To address these limitations, we introduce a novel inference-time steering methodology called Reinforcing Cognitive Experts (RICE), designed to improve reasoning depth and efficiency without additional training or complex heuristics. Leveraging normalized Pointwise Mutual Information (nPMI), we systematically identify specialized experts, termed cognitive experts that orchestrate meta-level reasoning operations characterized by tokens like ``''. Empirical evaluations with leading MoE-based LRMs (DeepSeek-R1 and Qwen3-235B) on rigorous quantitative and scientific reasoning benchmarks (AIME and GPQA Diamond) demonstrate noticeable and consistent improvements in reasoning accuracy, cognitive efficiency, and cross-domain generalization. Crucially, our lightweight approach substantially outperforms prevalent reasoning-steering techniques, such as prompt design and decoding constraints, while preserving the model's general instruction-following skills. These results highlight reinforcing cognitive experts as a promising, practical, and interpretable direction to enhance cognitive efficiency within advanced reasoning models.



Paperid:5765
Authors:Luca Ghafourpour, Valentin Duruisseaux, Bahareh Tolooshams, Philip Wong, Costas Anastassiou, Animashree Anandkumar
Abstract:
Abstract:Characterizing the diverse computational properties of human neurons via multimodal electrophysiological, transcriptomic, and morphological data provides the foundation for constructing and validating bio-realistic neuron models that can advance our understanding of fundamental mechanisms underlying brain function. However, current modeling approaches remain constrained by the limited availability and intrinsic variability of experimental neuronal data. To capture variability, ensembles of deterministic models are often used, but are difficult to scale as model generation requires repeating computationally expensive optimization for each neuron. While deep learning is becoming increasingly relevant in this space, it fails to capture the full biophysical complexity of neurons, their nonlinear voltage dynamics, and variability. To address these shortcomings, we introduce $\texttt{NOBLE}$, a neural operator framework that learns a mapping from a continuous frequency-modulated embedding of interpretable neuron features to the somatic voltage response induced by current injection. Trained on data generated from biophysically realistic neuron models, $\texttt{NOBLE}$ predicts distributions of neural dynamics accounting for the intrinsic experimental variability. Unlike conventional bio-realistic neuron models, interpolating within the embedding space offers models whose dynamics are consistent with experimentally observed responses. $\texttt{NOBLE}$ enables the efficient generation of synthetic neurons that closely resemble experimental data and exhibit trial-to-trial variability, offering a $4200$× speedup over the numerical solver. $\texttt{NOBLE}$ is the first scaled-up deep learning framework that validates its generalization with real experimental data. To this end, $\texttt{NOBLE}$ captures fundamental neural properties in a unique and emergent manner that opens the door to a better understanding of cellular composition and computations, neuromorphic architectures, large-scale brain circuits, and general neuroAI applications.



Paperid:5766
Authors:Zijian Guo, İlker Işık, H M Sabbir Ahmad, Wenchao Li
Abstract:
Abstract:Generalizing to complex and temporally extended task objectives and safety constraints remains a critical challenge in reinforcement learning (RL). Linear temporal logic (LTL) offers a unified formalism to specify such requirements, yet existing methods are limited in their abilities to handle nested long-horizon tasks and safety constraints, and cannot identify situations when a subgoal is not satisfiable and an alternative should be sought. In this paper, we introduce GenZ-LTL, a method that enables zero-shot generalization to arbitrary LTL specifications. GenZ-LTL leverages the structure of Büchi automata to decompose an LTL task specification into sequences of reach-avoid subgoals. Contrary to the current state-of-the-art method that conditions on subgoal sequences, we show that it is more effective to achieve zero-shot generalization by solving these reach-avoid problems $\textit{one subgoal at a time}$ through proper safe RL formulations. In addition, we introduce a novel subgoal-induced observation reduction technique that can mitigate the exponential complexity of subgoal-state combinations under realistic assumptions. Empirical results show that GenZ-LTL substantially outperforms existing methods in zero-shot generalization to unseen LTL specifications.



Paperid:5767
Authors:Yuxuan Zhou, Xien Liu, Xiao Zhang, Chen Ning, Shijin Wang, Guoping Hu, Ji Wu
Abstract:
Large Language Models (LLMs) must possess sufficient and comprehensive medical knowledge to effectively address real-world clinical problems. Though domain-specific fine-tuning can effectively inject medical knowledge into LLMs, it often leads to catastrophic forgetting of previously acquired knowledge and instruction following abilities. In this paper, we investigate catastrophic forgetting in medical knowledge injection and uncover a notable proximity-dependent forgetting nature: knowledge that is closely related to the injected knowledge is more likely to be forgotten, whereas distant knowledge is less affected. Moreover, we observe that existing mitigation techniques fail to effectively address this proximity-dependent forgetting. Inspired by this finding as well as human learning process, we propose a novel \textbf{Inter}nal K\textbf{n}owledge \textbf{A}ugmentation \textbf{L}earning (\textbf{InternAL}) approach that mitigates catastrophic forgetting by exploiting LLM's internal knowledge. Specifically, we first extract internal knowledge from the target LLM that is directly related to the injected knowledge, using questions generated from the injected knowledge. Then we augment the original knowledge injection dataset with the extracted internal knowledge, and train the LLM on this augmented dataset to enhance the model's ability to retain prior knowledge that is closely related to the injected knowledge. Experimental results on several representative LLMs (Llama, Qwen) show that our method significantly alleviates the forgetting of prior knowledge, especially the knowledge that is closely related to the injected knowledge. The datasets and codes will be released for future research.



Paperid:5768
Authors:Yuan Chiang, Tobias Kreiman, Christine Zhang, Matthew Kuner, Elizabeth Weaver, Ishan Amin, Hyunsoo Park, Yunsung Lim, Jihan Kim, Daryl Chrzan, Aron Walsh, Samuel Blau, Aditi Krishnapriyan, Mark Asta
Abstract:
Machine learning interatomic potentials (MLIPs) have revolutionized molecular and materials modeling, but existing benchmarks suffer from data leakage, limited transferability, and an over-reliance on error-based metrics tied to specific density functional theory (DFT) references. We introduce MLIP Arena, a benchmark platform that evaluates force field performance based on physics awareness, chemical reactivity, stability under extreme conditions, and predictive capabilities for thermodynamic properties and physical phenomena. By moving beyond static DFT references and revealing the important failure modes of current foundation MLIPs in real-world settings, MLIP Arena provides a reproducible framework to guide the next-generation MLIP development toward improved predictive accuracy and runtime efficiency while maintaining physical consistency. The Python package and online leaderboard are available at https://github.com/atomind-ai/mlip-arena.



Paperid:5769
Authors:Maria-Florina Balcan, Anh Nguyen, Dravyansh Sharma
Abstract:
Modern machine learning algorithms, especially deep learning-based techniques, typically involve careful hyperparameter tuning to achieve the best performance. Despite the surge of intense interest in practical techniques like Bayesian optimization and random search-based approaches to automating this laborious and compute-intensive task, the fundamental learning-theoretic complexity of tuning hyperparameters for deep neural networks is poorly understood. Inspired by this glaring gap, we initiate the formal study of hyperparameter tuning complexity in deep learning through a recently introduced data-driven setting. We assume that we have a series of learning tasks, and we have to tune hyperparameters to do well on average over the distribution of tasks. A major difficulty is that the utility function as a function of the hyperparameter is very volatile, and furthermore, it is given implicitly by an optimization problem over the model parameters. To tackle this challenge, we introduce a new technique to characterize the discontinuities and oscillations of the utility function on any fixed problem instance as we vary the hyperparameter; our analysis relies on subtle concepts, including tools from algebraic geometry, differential geometry, and constrained optimization. We use this to show that the learning-theoretic complexity of the corresponding family of utility functions is bounded. We instantiate our results and provide sample complexity bounds for concrete applications—tuning a hyperparameter that interpolates neural activation functions and setting the kernel parameter in graph neural networks.



Paperid:5770
Authors:Tien Nguyen, Dac Nguyen, Duc Nguyen The Minh, Trung Thanh Nguyen, Truong Thao Nguyen, Hieu Pham, Johan Barthelemy, Tran Minh Quan, Quoc Viet Hung Nguyen, Thanh Tam Nguyen, Mai Son, Chau Anh, Thanh Nguyen, Phi Le Nguyen
Abstract:
Vision-Language Foundation Models (VLMs), trained on large-scale multimodal datasets, have driven significant advances in Artificial Intelligence by enabling rich cross-modal reasoning. Despite their success in general domains, applying these models to medical imaging remains challenging due to the limited availability of diverse imaging modalities and multilingual clinical data. Most existing medical VLMs are trained on a subset of imaging modalities and focus primarily on high-resource languages, thus limiting their generalizability and clinical utility. To address these limitations, we introduce a novel Vietnamese-language multimodal medical dataset comprising 1,567,062 paired CT-PET images and corresponding full-length clinical reports. This dataset is designed to fill two pressing gaps in medical AI development: (1) the lack of PET/CT imaging data in existing VLMs training corpora, which hinders the development of models capable of handling functional imaging tasks; and (2) the underrepresentation of low-resource languages, particularly the Vietnamese language, in medical vision-language research. To the best of our knowledge, this is the first dataset to provide comprehensive PET/CT-report pairs in Vietnamese. We further introduce a training framework to enhance VLMs' learning, including data augmentation and expert-validated test sets. We conduct comprehensive experiments benchmarking state-of-the-art VLMs on downstream tasks, including medical report generation and visual question answering. The experimental results show that incorporating our dataset significantly improves the performance of existing VLMs. However, despite these advancements, the models still underperform on clinically critical criteria, particularly the diagnosis of lung cancer, indicating substantial room for future improvement. We believe this dataset and benchmark will serve as a pivotal step in advancing the development of more robust VLMs for medical imaging, particularly in low-resource languages, and improving their clinical relevance in Vietnamese healthcare.



Paperid:5771
Authors:Yijun Liang, Ming Li, Chenrui Fan, Ziyue Li, Dang Nguyen, Kwesi Cobbina, Shweta Bhardwaj, Jiuhai Chen, Fuxiao Liu, Tianyi Zhou
Abstract:
Color plays an important role in human perception and usually provides critical clues in visual reasoning. However, it is unclear whether and how vision-language models (VLMs) can perceive, understand, and leverage color as humans.This paper introduces ColorBench, an innovative benchmark meticulously crafted to assess the capabilities of VLMs in color understanding, including color perception, reasoning, and robustness. By curating a suite of diverse test scenarios, with grounding in real applications, ColorBench evaluates how these models perceive colors, infer meanings from color-based cues, and maintain consistent performance under varying color transformations. Through an extensive evaluation of 32 VLMs with varying language models and vision encoders, our paper reveals some undiscovered findings: (i) The scaling law (larger models are better) still holds on ColorBench, while the language model plays a more important role than the vision encoder. (ii) However, the performance gaps across models are relatively small, indicating that color understanding has been largely neglected by existing VLMs. (iii) CoT reasoning improves color understanding accuracies and robustness, though they are vision-centric tasks. (iv) Color clues are indeed leveraged by VLMs on ColorBench but they can also mislead models in some tasks.These findings highlight the critical limitations of current VLMs and underscore the need to enhance color comprehension. Our ColorBench can serve as a foundational tool for advancing the study of human-level color understanding of multimodal AI.



Paperid:5772
Authors:Jonggwon Park, Soobum Kim, Byungmu Yoon, Kyoyun Choi
Abstract:
Abstract:Recent advancements in multi-modal models have significantly improved vision-language (VL) alignment in radiology. However, existing approaches struggle to effectively utilize complex radiology reports for learning and offer limited interpretability through attention probability visualizations. To address these challenges, we introduce $\textbf{RadZero}$, a novel framework for VL alignment in radiology with zero-shot multi-task capability. A key component of our approach is $\textbf{VL-CABS}$ ($\textbf{V}$ision-$\textbf{L}$anguage $\textbf{C}$ross-$\textbf{A}$ttention $\textbf{B}$ased on $\textbf{S}$imilarity),which aligns text embeddings with local image features for interpretable, fine-grained VL reasoning.RadZero leverages large language models to extract concise semantic sentences from radiology reports and employs multi-positive contrastive training to effectively capture relationships between images and multiple relevant textual descriptions.It uses a pre-trained vision encoder with additional trainable Transformer layers, allowing efficient high-resolution image processing. By computing similarity between text embeddings and local image patch features, VL-CABS enables zero-shot inference with similarity probability for classification, and pixel-level VL similarity maps for grounding and segmentation. Experimental results on public chest radiograph benchmarks show that RadZero outperforms state-of-the-art methods in zero-shot classification, grounding, and segmentation. Furthermore, VL similarity map analysis highlights the potential of VL-CABS for improving explainability in VL alignment. Additionally, qualitative evaluation demonstrates RadZero’s capability for open-vocabulary semantic segmentation, further validating its effectiveness in medical imaging.



Paperid:5773
Authors:Chaitanya Kapoor, Sudhanshu Srivastava, Meenakshi Khosla
Abstract:
Understanding convergent learning — the degree to which independently trained neural systems — whether multiple artificial networks or brains and models — arrive at similar internal representations — is crucial for both neuroscience and AI. Yet, the literature remains narrow in scope—typically examining just a handful of models with one dataset, relying on one alignment metric, and evaluating networks at a single post-training checkpoint. We present a large-scale audit of convergent learning, spanning dozens of vision models and thousands of layer-pair comparisons, to close these long-standing gaps. First, we pit three alignment families against one another---linear regression (affine-invariant), orthogonal Procrustes (rotation-/reflection-invariant), and permutation/soft-matching (unit-order-invariant). We find that orthogonal transformations align representations nearly as effectively as more flexible linear ones, and although permutation scores are lower, they significantly exceed chance, indicating a privileged representational basis. Tracking convergence throughout training further shows that nearly all eventual alignment crystallizes within the first epoch---well before accuracy plateaus---indicating it is largely driven by shared input statistics and architectural biases, not by the final task solution. Finally, when models are challenged with a battery of out-of-distribution images, early layers remain tightly aligned, whereas deeper layers diverge in proportion to the distribution shift. These findings fill critical gaps in our understanding of representational convergence, with implications for neuroscience and AI.



Paperid:5774
Authors:Zhenyu Tao, Wei Xu, Xiaohu You
Abstract:
The bisimulation metric (BSM) is a powerful tool for computing state similarities within a Markov decision process (MDP), revealing that states closer in BSM have more similar optimal value functions. While BSM has been successfully utilized in reinforcement learning (RL) for tasks like state representation learning and policy exploration, its application to multiple-MDP scenarios, such as policy transfer, remains challenging. Prior work has attempted to generalize BSM to pairs of MDPs, but a lack of rigorous analysis of its mathematical properties has limited further theoretical progress. In this work, we formally establish a generalized bisimulation metric (GBSM) between pairs of MDPs, which is rigorously proven with the three fundamental properties: GBSM symmetry, inter-MDP triangle inequality, and the distance bound on identical states. Leveraging these properties, we theoretically analyse policy transfer, state aggregation, and sampling-based estimation in MDPs, obtaining explicit bounds that are strictly tighter than those derived from the standard BSM. Additionally, GBSM provides a closed-form sample complexity for estimation, improving upon existing asymptotic results based on BSM. Numerical results validate our theoretical findings and demonstrate the effectiveness of GBSM in multi-MDP scenarios.



Paperid:5775
Authors:Lin Zhu, Yifeng Yang, Xinbing Wang, Qinying Gu, Nanyang Ye
Abstract:
Abstract:Recent approaches for vision-language models (VLMs) have shown remarkable success in achieving fast downstream adaptation. When applied to real-world downstream tasks, VLMs inevitably encounter both the in-distribution (ID) data and out-of-distribution (OOD) data. The OOD datasets often include both covariate shifts (e.g., known classes with changes in image styles) and semantic shifts (e.g., test-time unseen classes). This highlights the importance of improving VLMs' generalization ability to covariate-shifted OOD data, while effectively detecting open-set semantic-shifted OOD classes. In this paper, inspired by the substantial energy change observed in closed-set data when re-aligning vision-language modalities—specifically by directly reducing the maximum cosine similarity to a low value—we introduce a novel OOD score, named $\Delta\mathrm{Energy}$. $\Delta\mathrm{Energy}$ significantly outperforms the vanilla energy-based OOD score and provides a more reliable approach for OOD detection. Furthermore, $\Delta\mathrm{Energy}$ can simultaneously improve OOD generalization under covariate shifts, which is achieved by lower-bound maximization for $\Delta\mathrm{Energy}$ (termed EBM). EBM is theoretically proven to not only enhance OOD detection but also yields a domain-consistent Hessian, which serves as a strong indicator for OOD generalization. Based on this finding, we developed a unified fine-tuning framework that allows for improving VLMs' robustness in both OOD generalization and OOD detection. Extensive experiments on challenging OOD detection and generalization benchmarks demonstrate the superiority of our method, outperforming recent approaches by 10\%–25\% in AUROC.



Paperid:5776
Authors:Runyu Lu, Peng Zhang, Ruochuan Shi, Yuanheng Zhu, Dongbin Zhao, Yang Liu, Dong Wang, Cesare Alippi
Abstract:
Equilibrium learning in adversarial games is an important topic widely examined in the fields of game theory and reinforcement learning (RL). Pursuit-evasion game (PEG), as an important class of real-world games from the fields of robotics and security, requires exponential time to be accurately solved. When the underlying graph structure varies, even the state-of-the-art RL methods require recomputation or at least fine-tuning, which can be time-consuming and impair real-time applicability. This paper proposes an Equilibrium Policy Generalization (EPG) framework to effectively learn a generalized policy with robust cross-graph zero-shot performance. In the context of PEGs, our framework is generally applicable to both pursuer and evader sides in both no-exit and multi-exit scenarios. These two generalizability properties, to our knowledge, are the first to appear in this domain. The core idea of the EPG framework is to train an RL policy across different graph structures against the equilibrium policy for each single graph. To construct an equilibrium oracle for single-graph policies, we present a dynamic programming (DP) algorithm that provably generates pure-strategy Nash equilibrium with near-optimal time complexity. To guarantee scalability with respect to pursuer number, we further extend DP and RL by designing a grouping mechanism and a sequence model for joint policy decomposition, respectively. Experimental results show that, using equilibrium guidance and a distance feature proposed for cross-graph PEG training, the EPG framework guarantees desirable zero-shot performance in various unseen real-world graphs. Besides, when trained under an equilibrium heuristic proposed for the graphs with exits, our generalized pursuer policy can even match the performance of the fine-tuned policies from the state-of-the-art PEG methods.



Paperid:5777
Authors:Zhaoyi Li, Xiaohan Zhao, Dong-Dong Wu, Jiacheng Cui, Zhiqiang Shen
Abstract:
Despite promising performance on open-source large vision-language models (LVLMs), transfer-based targeted attacks often fail against black-box commercial closed-source LVLMs. Analyzing failed adversarial perturbations reveals that the learned perturbations typically originate from a uniform distribution and lack clear semantic details, resulting in unintended responses. This critical absence of semantic information leads commercial LVLMs to either ignore the perturbation entirely or misinterpret its embedded semantics, thereby causing the attack to fail. To overcome these issues, we propose to refine semantic clarity by encoding explicit semantic details within local regions, thus ensuring interoperability and capturing finer-grained features, and by concentrating modifications on semantically rich areas rather than applying them uniformly. To achieve this, we proposea simple yet highly effective baseline: at each optimization step, the adversarial image is cropped randomly by a controlled aspect ratio and scale, resized, and then aligned with the target image in the embedding space. While the na\"ive source-target matching method has been utilized before in the literature, we are the first to provide a tight analysis, which establishes a close connection between perturbation optimization and semantics. Experimental results confirm our hypothesis. Our adversarial examples crafted with local-aggregated perturbations focused on crucial regions exhibit surprisingly good transferability to commercial LVLMs, including GPT-4.5, GPT-4o, Gemini-2.0-flash, Claude-3.5/3.7-sonnet, and even reasoning models like o1, Claude-3.7-thinking and Gemini-2.0-flash-thinking. Our approach achieves success rates exceeding 90\% on GPT-4.5, 4o, and o1, significantly outperforming all prior state-of-the-art attack methods. Our code and optimized adversarial examples are available in supplementary materials.



Paperid:5778
Authors:Qi Xin, Robert E Kass
Abstract:
Advances in large-scale recording technologies now enable simultaneous measurements from multiple brain areas, offering new opportunities to study signal transmission across interacting components of neural circuits. However, neural responses exhibit substantial trial-to-trial variability, often driven by unobserved factors such as subtle changes in animal behavior or internal states. To prevent evolving background dynamics from contaminating identification of functional coupling, we developed a hybrid neural spike train model, GLM-Transformer, that incorporates flexible, deep latent variable models into a point process generalized linear model (GLM) having an interpretable component for cross-population interactions. A Transformer-based variational autoencoder captures nonstationary individual-neuron dynamics that vary across trials, while standard nonparametric regression GLM coupling terms provide estimates of directed interactions between neural populations. We incorporate a low-rank structure on population-to-population coupling effects to improve scalability. Across synthetic datasets and mechanistic simulations, GLM-Transformer recovers known coupling structure and remains robust to shared background fluctuations. When applied to the Allen Institute Visual Coding dataset, it identifies feedforward pathways consistent with established visual hierarchies. This work offers a step toward improved identification of neural population interactions, and contributes to ongoing efforts aimed at achieving interpretable results while harvesting the benefits of deep learning.



Paperid:5779
Authors:Qiuchen Wang, Ruixue Ding, Yu Zeng, Zehui Chen, Lin Chen, Shihang Wang, Pengjun Xie, Fei Huang, Feng Zhao
Abstract:
Effectively retrieving, reasoning and understanding visually rich information remains a challenge for traditional Retrieval-Augmented Generation (RAG) methods. On the one hand, traditional text-based methods cannot handle visual-related information. On the other hand, current vision-based RAG approaches are often limited by fixed pipelines and frequently struggle to reason effectively due to the insufficient activation of the fundamental capabilities of models. As reinforcement learning (RL) has been proven to be beneficial for model reasoning, we introduce \textbf{VRAG-RL}, a novel RL framework tailored for complex reasoning across visually rich information. With this framework, VLMs interact with search engines, autonomously sampling single-turn or multi-turn reasoning trajectories with the help of visual perception tokens and undergoing continual optimization based on these samples. Our approach highlights key limitations of RL in RAG domains: (i) Prior Multi-modal RAG approaches tend to merely incorporate images into the context, leading to insufficient reasoning token allocation and neglecting visual-specific perception; and (ii) When models interact with search engines, their queries often fail to retrieve relevant information due to the inability to articulate requirements, thereby leading to suboptimal performance. To address these challenges, we define an action space tailored for visually rich inputs, with actions including cropping and scaling, allowing the model to gather information from a coarse-to-fine perspective. Furthermore, to bridge the gap between users' original inquiries and the retriever, we employ a simple yet effective reward that integrates query rewriting and retrieval performance with a model-based reward. Our VRAG-RL optimizes VLMs for RAG tasks using specially designed RL strategies, aligning the model with real-world applications. Extensive experiments on diverse and challenging benchmarks show that our VRAG-RL outperforms existing methods by 20\% (Qwen2.5-VL-7B) and 30\% (Qwen2.5-VL-3B), demonstrating the effectiveness of our approach.



Paperid:5780
Authors:Yu Wang, Yang Li, Junchi Yan, Yi Chang
Abstract:
Abstract:Diffusion models have recently emerged as powerful neural solvers for combinatorial optimization (CO). However, existing approaches fail to reveal how variables are progressively determined during inference, making the final solution opaque until the last step. To address this limitation, we propose a structured denoising diffusion model, StruDiCO, which incrementally constructs solutions through step-wise variable selection. This is achieved via a variable-absorption noising model, wherein the forward process simulates gradual variable deactivation, converging to an empty solution, while the reverse process incrementally selects variables to reconstruct the final solution. This design induces structural continuity across intermediate states, enabling interpretable and trajectory-consistent partial solutions throughout inference. To further improve the reliability of reverse inference, we introduce a constrained consistency sampling strategy, which suppresses low-confidence variable selection at each step to stabilize the reverse process. Leveraging the structure-preserving reverse process, we further propose a lightweight, gradient-free, objective-aware refinement framework, which iteratively reconstructs lower-cost feasible solutions by applying structure-aware perturbations to the current solution, performing reverse inference through the constraint consistency model, and decoding with an objective-guided scoring scheme. Extensive experiments on two canonical CO tasks, the Traveling Salesman Problem (TSP) and Maximal Independent Set (MIS), show that StruDiCO outperforms state-of-the-art diffusion-based solvers, achieving up to $3.5\times$ faster inference, 70\% lower GPU memory usage, and significantly improved solution quality, with up to 37.7\% drop reduction on TSP and an average 38.1\% improvement on MIS.



Paperid:5781
Authors:Run Luo, Ting-En Lin, Haonan Zhang, Yuchuan Wu, Xiong Liu, Min Yang, Yongbin Li, Longze Chen, Jiaming Li, Lei Zhang, Xiaobo Xia, Hamid Alinejad-Rokny, Fei Huang
Abstract:
Abstract:Recent advancements in omnimodal learning have significantly improved understanding and generation across images, text, and speech, yet these developments remain predominantly confined to proprietary models. The lack of high-quality omnimodal datasets and the challenges of real-time emotional speech synthesis have notably hindered progress in open-source research. To address these limitations, we introduce OpenOmni, a two-stage training framework that integrates omnimodal alignment and speech generation to develop a state-of-the-art omnimodal large language model. In the alignment phase, a pretrained speech model undergoes further training on image-text tasks, enabling (near) zero-shot generalization from vision to speech, outperforming models trained on tri-modal datasets. In the speech generation phase, a lightweight decoder is trained on speech tasks with direct preference optimization, which enables real-time emotional speech synthesis with high fidelity. Extensive experiments demonstrate that OpenOmni surpasses state-of-the-art models across omnimodal, vision-language, and speech-language benchmarks. It achieves a 4-point absolute improvement on OmniBench over the leading open-source model VITA, despite using 5$\times$ fewer training examples and a smaller model size (7B vs. 7$\times$8B). Besides, OpenOmni achieves real-time speech generation with less than 1 second latency at non-autoregressive mode, reducing inference time by 5$\times$ compared to autoregressive methods, and improves emotion classification accuracy by 7.7\%. The codebase is available at https://anonymous.4open.science/r/OpenOmni-1544.



Paperid:5782
Authors:Zhixiao Wu, Yao Lu, Jie Wen, Hao Sun, Qi Zhou, Guangming Lu
Abstract:
Poison-only Clean-label Backdoor Attacks (PCBAs) aim to covertly inject attacker-desired behavior into DNNs by merely poisoning the dataset without changing the labels. To effectively implant a backdoor, multiple triggers are proposed for various attack requirements of Attack Success Rate (ASR) and stealthiness. Additionally, sample selection enhances clean-label backdoor attacks' ASR by meticulously selecting "hard'' samples instead of random samples to poison. Current methods, however, 1) usually handle the sample selection and triggers in isolation, leading to severely limited improvements on both ASR and stealthiness. Consequently, attacks exhibit unsatisfactory performance on evaluation metrics when converted to PCBAs via a mere stacking of methods. Therefore, we seek to explore the bi-directional collaborative relations between the sample selection and triggers to address the above dilemma. 2) Since the strong specificity within triggers, the simple combination of sample selection and triggers fails to substantially enhance both evaluation metrics, with generalization preserved among various attacks. Therefore, we seek to propose a set of components to significantly improve both stealthiness and ASR based on the commonalities of attacks. Specifically, Component A ascertains two critical selection factors, and then makes them an appropriate combination based on the trigger scale to select more reasonable "hard'' samples for improving ASR. Component B is proposed to select samples with similarities to relevant trigger implanted samples to promote stealthiness. Component C reassigns trigger poisoning intensity on RGB colors through distinct sensitivity of the human visual system to RGB for higher ASR, with stealthiness ensured by sample selection including Component B. Furthermore, all components can be strategically integrated into diverse PCBAs, enabling tailored solutions that balance ASR and stealthiness enhancement for specific attack requirements. Extensive experiments demonstrate the superiority of our components in stealthiness, ASR, and generalization. Our code will be released as soon as possible.



Paperid:5783
Authors:YUXUAN SUN, Yixuan Si, Chenglu Zhu, Kai Zhang, Zhongyi Shui, Bowen Ding, Tao Lin, Lin Yang
Abstract:
Recent advances in computational pathology have led to the emergence of numerous foundation models. However, these approaches fail to replicate the diagnostic process of pathologists, as they either simply rely on general-purpose encoders with multi-instance learning for classification or directly apply multimodal models to generate reports from images. A significant limitation is their inability to emulate the diagnostic logic employed by pathologists, who systematically examine slides at low magnification for overview before progressively zooming in on suspicious regions to formulate comprehensive diagnoses. To address this gap, we introduce CPathAgent, an innovative agent-based model that mimics pathologists' reasoning processes by autonomously executing zoom-in/out and navigation operations across pathology images based on observed visual features. To achieve this, we develop a multi-stage training strategy unifying patch-level, region-level, and whole-slide capabilities within a single model, which is essential for mimicking pathologists, who require understanding and reasoning capabilities across all three scales. This approach generates substantially more detailed and interpretable diagnostic reports compared to existing methods, particularly for huge region understanding. Additionally, we construct an expert-validated PathMMU-HR², the first benchmark for huge region analysis, a critical intermediate scale between patches and whole slides, as diagnosticians typically examine several key regions rather than entire slides at once. Extensive experiments demonstrate that CPathAgent consistently outperforms existing approaches across three scales of benchmarks, validating the effectiveness of our agent-based diagnostic approach and highlighting a promising direction for the future development of computational pathology.



Paperid:5784
Authors:Muqi Han, Xiaobin Li, Kai Wu, Xiaoyu Zhang, Handing Wang
Abstract:
Zero‑shot optimization, which solves new black‑box problems without task‑specific tuning, is critical for efficient decision making across science and engineering. Pre‑trained optimization models (POMs) address this need by leveraging large‑scale learned priors to optimize unseen objective functions out of the box, reducing reliance on expert knowledge. However, existing POMs suffer from three key weaknesses: (1) limited population modeling that hinders convergence and generalization, (2) static information exchange that restricts solution diversity, and (3) noisy gradient estimates that destabilize training. To overcome these challenges, we introduce Efficient Pretrained Optimization Models, featuring: (1) LMM/LCM Tokenizer, which embeds landscape cues into population tokens for richer modeling and seamless feature fusion; (2) Efficient LMM, a deformable attention module enabling adaptive, diversity‑promoting interactions; (3) Differentiable LCM, which yields precise gradient approximations for stable, reliable training. Our model sets a new benchmark in zero‑shot black‑box optimization by delivering faster convergence, higher diversity, and robust performance.



Paperid:5785
Authors:Jingmin An, Yilong Song, Ruolin Yang, Nai Ding, Lingxi Lu, Yuxuan Wang, Wei Wang, Chu Zhuang, Qian Wang, Fang Fang
Abstract:
Large Language Models (LLMs) demonstrate human-level or even superior language abilities, effectively modeling syntactic structures, yet the specific computational modules responsible remain unclear. A key question is whether LLM behavioral capabilities stem from mechanisms akin to those in the human brain. To address these questions, we introduce the Hierarchical Frequency Tagging Probe (HFTP), a tool that utilizes frequency-domain analysis to identify neuron-wise components of LLMs (e.g., individual Multilayer Perceptron (MLP) neurons) and cortical regions (via intracranial recordings) encoding syntactic structures. Our results show that models such as GPT-2, Gemma, Gemma 2, Llama 2, Llama 3.1, and GLM-4 process syntax in analogous layers, while the human brain relies on distinct cortical regions for different syntactic levels. Representational similarity analysis reveals a stronger alignment between LLM representations and the left hemisphere of the brain (dominant in language processing). Notably, upgraded models exhibit divergent trends: Gemma 2 shows greater brain similarity than Gemma, while Llama 3.1 shows less alignment with the brain compared to Llama 2. These findings offer new insights into the interpretability of LLM behavioral improvements, raising questions about whether these advancements are driven by human-like or non-human-like mechanisms, and establish HFTP as a valuable tool bridging computational linguistics and cognitive neuroscience.



Paperid:5786
Authors:Hisham Temmar, Yixuan Wang, Nina Gill, Nicholas Mellon, Chang Liu, Luis Cubillos, Rio Parsons, Joseph Costello, Matteo Ceradini, Madison Kelberman, Matthew Mender, Aren Hite, Dylan Wallace, Samuel Nason-Tomaszewski, Parag Patil, Matt Willsey, Anne Draelos, Cynthia Chestek
Abstract:
Intracortical brain-machine interfaces (iBMIs) have enabled movement and speech in people living with paralysis by using neural data to decode behaviors in real-time. However, intracortical neural recordings exhibit significant instabilities over time, which poses problems for several domains related to iBMIs, neuroscience, and machine learning. For iBMIs, neural instabilities require frequent decoder recalibration to maintain high performance, a critical bottleneck for real-world translation. Several approaches have been developed to address this issue, and the field has recognized the need for standardized datasets on which to compare them, but no standard dataset exists for testing over year-long timescales. In neuroscience, a growing body of research attempts to elucidate the latent computations performed by populations of neurons. Nonstationarity in neural recordings imposes significant challenges to the design of these studies, so a dataset containing recordings over large time spans would improve methods to account for instabilities. In machine learning, continuous domain adaptation of temporal data is an area of active research, and a dataset containing shift distributions on long time scales would be beneficial to researchers. To address these gaps, we present the LINK Dataset (Long-term Intracortical Neural activity and Kinematics), which contains intracortical spiking activity and kinematic data from 303 sessions of a non-human primate performing a dexterous, 2 degree-of-freedom finger movement task, spanning 1,242 days. We also present longitudinal analyses of the dataset’s neural spiking activity and its relationship to kinematics, as well as overall decoding performance using linear and neural network models. The LINK dataset (https://dandiarchive.org/dandiset/001201) and code (https://github.com/chesteklab/LINK_dataset) are freely available for the public.



Paperid:5787
Authors:Xin Lu, Yanyan Zhao, Si Wei, Shijin Wang, Bing Qin, Ting Liu
Abstract:
Pre-trained language models represented by the Transformer have been proven to possess strong base capabilities, and the representative self-attention mechanism in the Transformer has become a classic in sequence modeling architectures. Different from the work of proposing sequence modeling architecture to improve the efficiency of attention mechanism, this work focuses on the impact of sequence modeling architectures on base capabilities. Specifically, our concern is: How exactly do sequence modeling architectures affect the base capabilities of pre-trained language models? In this work, we first point out that the mixed domain pre-training setting commonly adopted in existing architecture design works fails to adequately reveal the differences in base capabilities among various architectures. To address this, we propose a limited domain pre-training setting with out-of-distribution testing, which successfully uncovers significant differences in base capabilities among architectures at an early stage. Next, we analyze the base capabilities of stateful sequence modeling architectures, and find that they exhibit significant degradation in base capabilities compared to the Transformer. Then, through a series of architecture component analysis, we summarize a key architecture design principle: A sequence modeling architecture need possess full-sequence arbitrary selection capability to avoid degradation in base capabilities. Finally, we empirically validate this principle using an extremely simple Top-1 element selection architecture and further generalize it to a more practical Top-1 chunk selection architecture. Experimental results demonstrate our proposed sequence modeling architecture design principle and suggest that our work can serve as a valuable reference for future architecture improvements and novel designs.